analyse et simulation de stratégies de juste-à-temps dans ... · depuis plusieurs années, les...
TRANSCRIPT
© Baroma Tidoumba Flora Bamana, 2018
Analyse et simulation de stratégies de juste-à-temps dans le domaine de la construction - Application à un
bâtiment multi-étages en bois
Mémoire
Baroma Tidoumba Flora Bamana
Maîtrise en génie mécanique - avec mémoire
Maître ès sciences (M. Sc.)
Québec, Canada
Analyse et simulation de stratégies de juste-à-temps dans le domaine de la construction – Application à un
bâtiment multi-étages en bois
Mémoire
Baroma Tidoumba Flora Bamana
Sous la direction de :
Nadia Lehoux, directrice de recherche
iii
Résumé
L’industrie de la construction est l’un des secteurs vers lesquels se tournent les pays pour
stimuler la création d’emplois dans les moments les plus difficiles. Important moteur dans
la croissance économique de la province de Québec et du Canada, l’industrie de la
construction recherche constamment des mesures lui permettant d’éliminer toute forme de
gaspillage dans ses activités, dès la conception jusqu’à l’achèvement d’un projet. C’est dans
cette optique que s’intègre cette maîtrise, dont le but est d’étudier la philosophie du juste-à-
temps (JAT) en vue d’application dans le domaine de la construction. Pour ce faire, une
revue systématique de la littérature a été faite afin d’établir un état de l’art sur le JAT et les
modalités de son application dans la construction. Ensuite, la simulation de la construction
d’un bâtiment multi-étages en bois a été réalisée afin de tester les possibilités
d’implantation du JAT pour un projet de construction réelle tout en faisant varier différents
paramètres clés. Des analyses statistiques ont par la suite été effectuées afin de déterminer
l’impact et les interactions de ces paramètres sur la productivité du projet de construction.
Finalement, l’étape de synthèse et de recommandations a permis, grâce à l’analyse des
résultats, de mettre en lumière le meilleur scénario d’implantation du JAT pour le projet à
l’étude. En effet, le meilleur scénario s’est avéré réduire la durée de la construction de
26,09 à 22,31 semaines, en plus d’éliminer les risques de ruptures et d’augmenter le taux
d’utilisation des travailleurs. Somme toute, l’étude a permis d’expliciter les concepts
essentiels à la mise en application du JAT dans la construction tout en démontrant, au
moyen de la simulation, les retombées possibles d’une telle application dans un projet de
construction réelle. En effet, peu d’études discutent du sujet et encore moins démontrent les
bénéfices quantitatifs liés à son implantation. Cette étude en fait l’illustration et par
conséquent contribue respectivement à la science et à l’industrie en rapportant clairement
les avenues d’implantation du JAT dans la construction et en déterminant la mesure dans
laquelle les livraisons JAT, les méthodes Lean, et la préfabrication pourraient être mis en
œuvre en vue d’en retirer des bénéfices sur les chantiers de construction.
iv
Abstract
As a driving force in the economic growth of the province of Quebec and Canada, the
construction industry is constantly looking for measures to eliminate all forms of waste in
its activities, from the design stage to completion of a project. This research aims to study
the philosophy of just-in-time (JIT) for application in the construction field. To do so, a
systematic literature review was conducted to establish a state of the art on JIT and the
modalities of its application in construction. Afterward, a simulation model was developed
to test different possibilities of JIT implementation for a real construction project with
different key parameters. Statistical analyzes were then executed to determine the impact
and interactions of these parameters on productivity in the construction project under
investigation. Finally, the synthesis and recommendations step highlighted, through the
analysis of the results, the best scenario to implement. Indeed, the best scenario allowed to
reduce the construction duration from 26.09 to 22.31 weeks, to eliminate the risk of
shortages and to increase workers utilisation rate. In sum, the study has provided thorough
enlightenments on concepts essential to JIT implementation in construction. Few studies
discussed the subject and even less demonstrated the quantitative benefits of its
implementation. This study therefore contributes to science and the industry by reporting
pathways of JIT implementation in construction while determining the extent to which JIT
deliveries, Lean methods, and prefabrication could be implemented to derive benefits on
construction sites.
v
Table des matières
Résumé .............................................................................................................................................................. 3
Abstract ............................................................................................................................................................... 4
Liste des figures .................................................................................................................................................. 7
Liste des tableaux ............................................................................................................................................... 8
Liste des abréviations ......................................................................................................................................... 9
Remerciements ................................................................................................................................................. 10
Avant-propos .................................................................................................................................................... 11
Introduction générale .......................................................................................................................................... 1
Revue de la littérature ......................................................................................................................................... 3
Le juste à temps ............................................................................................................................................. 3
Le juste à temps dans la construction ............................................................................................................ 6
Le concept de la simulation ............................................................................................................................ 7
Objectifs et méthodologie ................................................................................................................................. 13
Phase 1 : Revue systématique de la littérature ............................................................................................ 14
Phase 2 : Simulation..................................................................................................................................... 15
Phase 3 : Analyses statistiques .................................................................................................................... 16
Phase 4 : Synthèse et recommandations ..................................................................................................... 17
Just in time in construction: description and implementation insights ............................................................... 18
Résumé ........................................................................................................................................................ 19
Abstract ........................................................................................................................................................ 20
Introduction ................................................................................................................................................... 21
JIT in construction: systematic literature review ........................................................................................... 21
Planning the review .................................................................................................................................. 21
Conducting the review .............................................................................................................................. 22
Discussion .................................................................................................................................................... 23
Reporting and dissemination .................................................................................................................... 23
Description of different implementation scenarios of JIT .......................................................................... 25
Conclusion .................................................................................................................................................... 26
References ................................................................................................................................................... 27
Appendix ...................................................................................................................................................... 29
Simulation of a prefabricated wooden building using JIT and Lean strategies to reduce construction duration 32
Résumé ........................................................................................................................................................ 33
Abstract ........................................................................................................................................................ 34
Introduction ................................................................................................................................................... 35
Literature review ........................................................................................................................................... 36
Methodology ................................................................................................................................................. 37
Case study ............................................................................................................................................... 38
Data collection .......................................................................................................................................... 38
Current state simulation model ................................................................................................................ 39
Model verification and validation .............................................................................................................. 42
Results ......................................................................................................................................................... 43
vi
Simulation of various scenarios ............................................................................................................... 43
Results per scenario ................................................................................................................................ 45
Conclusion .................................................................................................................................................... 47
Acknowledgment .......................................................................................................................................... 48
References ................................................................................................................................................... 48
Just-in-time and Lean in Construction: Simulation of a Six-story Building to Improve Productivity ................... 50
Résumé ........................................................................................................................................................ 51
Abstract ........................................................................................................................................................ 52
Introduction ................................................................................................................................................... 53
Literature review ........................................................................................................................................... 54
Research methodology ................................................................................................................................. 57
Pre-analysis ............................................................................................................................................. 58
Data collection .......................................................................................................................................... 59
Current state simulation model ................................................................................................................ 61
Model verification and validation .............................................................................................................. 64
Results ......................................................................................................................................................... 65
Simulation of various scenarios ............................................................................................................... 65
Results per Scenario ................................................................................................................................ 66
Design of Experiments ............................................................................................................................. 71
Best scenario for the case study .............................................................................................................. 74
Implementation framework ....................................................................................................................... 75
Conclusion .................................................................................................................................................... 77
Appendix ...................................................................................................................................................... 79
Notation ........................................................................................................................................................ 82
References ................................................................................................................................................... 82
Conclusion ........................................................................................................................................................ 85
Bibliographie ..................................................................................................................................................... 89
vii
Liste des figures
Figure 1 Méthodologie de recherche. .............................................................................................................. 14
Figure 2 Number of occurrences of JIT elements and KPI mentioned in the SLR. .......................................... 25
Figure 3 Scenarios of implementation of JIT in construction. .......................................................................... 26
Figure 4 Mechanism on run initialized. ............................................................................................................ 40
Figure 5 Creation of order and choice of path.................................................................................................. 41
Figure 6 Decision to replenish. ........................................................................................................................ 42
Figure 7 Main steps of the analysis. ................................................................................................................. 58
Figure 8 Root causes of low productivity. ......................................................................................................... 59
Figure 9 The building in terms of its seven divisions. ........................................................................................ 60
Figure 10 Different elements intervening on the construction site activities. ..................................................... 61
Figure 11 Mechanism on run initialized. ........................................................................................................... 62
Figure 12 Regular replenishment decision process. ......................................................................................... 63
Figure 13 Creation of order and choice of path................................................................................................. 64
Figure 14 Graph of construction duration vs stock for different scenarios with one work shift. ......................... 68
Figure 15 Graph of shortages vs. stock for different scenarios with one work shift. ......................................... 69
Figure 16 Graph of labor utilization vs stock for different scenarios with one work shift. .................................. 69
Figure 17 Pareto diagram of normalized effects on construction duration. ....................................................... 72
Figure 18 Graph of main effects on construction duration. ............................................................................... 73
Figure 19 Graph of interactions of effects for construction duration. ................................................................ 74
Figure 20 Implementation framework. .............................................................................................................. 76
viii
Liste des tableaux
Table 1 Details concerning the variations of each parameter. .......................................................................... 44
Table 2 Description of the different scenarios tested. ....................................................................................... 44
Table 3 Results of different scenarios tested. ................................................................................................... 47
Table 4 Details of the variations of each parameter.......................................................................................... 66
Table 5 Results for the different scenarios for one shift. ................................................................................... 67
Table 6 All the different combinations for the DOE. .......................................................................................... 71
Table 7 Results of the P value analysis for each effect on construction duration, labor utilization, and
shortages. ......................................................................................................................................................... 73
Table 8 Results of the current state vs. the best scenario for this case study. ................................................. 75
Table A 1 Results of the SLR. ........................................................................................................................... 29
Table A 2 Description of the 60 scenarios tested and their results. .................................................................. 79
ix
Liste des abréviations
DOE : Design d’expériences
DOE : Design of experiment
JAT : Juste-à-temps
JIT : Just-in-time
QEC : Quantité économique à commander
RSL : Revue systématique de la littérature
SLR : Systematic literature review
x
Remerciements
Je tiens à remercier tous ceux qui, de près ou de loin, ont contribué à la réalisation de ce
travail.
Je tiens particulièrement à exprimer mes sincères remerciements à ma directrice de maîtrise
Mme Lehoux pour sa disponibilité et l’excellent encadrement qu’elle m’a prodigué pendant
mes deux années au second cycle. Grâce à sa grande clarté d’esprit et toujours dans une
bonne humeur contagieuse, elle a su me guider tout au long de mon projet, me permettant
ainsi de mieux avancer. Elle me laisse comme cadeau son empreinte de qualité qui teintera
à jamais ma vie professionnelle.
Mes remerciements s’étendent également à Mme Cloutier, l’une des co-auteures de mes
trois articles, qui de par son expertise a participé et veillé activement à la réussite de nos
articles scientifiques. J’adresse aussi mes vifs remerciements à M. Blanchet et M. Gagné
dont le travail au CIRCERB a grandement assuré et facilité l’avancement de mon projet par
le biais de mon partenaire industriel Nordic Structures.
Je ne laisserai cette occasion passer sans remercier tous les partenaires industriels du
CIRCERB, sans qui cette étude n’aurait pas vu le jour, spécialement Nordic Structures (par
le biais de M. Faelli) pour leur temps et leurs données lorsque j’en avais besoin durant
l’étude.
xi
Avant-propos
Ce travail intitulé « Analyse et simulation de stratégies de juste-à-temps dans le domaine de
la construction – Application à un bâtiment multi-étages en bois », est réalisé afin d’obtenir
le diplôme de maîtrise en Génie Mécanique (MSc.) de l’Université Laval. Il a été effectué
sous la direction de Mme Nadia Lehoux au sein de la Chaire industrielle de recherche sur la
construction écoresponsable en bois, CIRCERB.
Ce mémoire est rédigé selon le principe d’insertion d’articles avec deux articles de
conférence et un article de journal, tous coécrits avec Mme Nadia Lehoux et Mme Caroline
Cloutier. Pour chacun des trois articles, j’ai œuvré en tant qu’auteure principale
responsable de toutes les recherches, rédactions, travaux et analyses relatifs à l’étude.
Le premier article intitulé « Just in time in Construction: Description and Implementation
Insights » a été soumis le 23 novembre 2016 à la conférence « LC3 : Lean & Computing in
Construction Congress » et présenté à la 25e édition de cette conférence le 10 juillet 2017 à
Crète, Grèce. La version publiée est identique à la version présentée dans ce mémoire.
Le second article intitulé « Simulation of a Prefabricated Wooden Building Using JIT and
Lean Strategies to Reduce Construction Duration » a été soumis le 08 décembre 2017 à la
conférence « International Conference on Information Systems, Logistics and Supply Chain
(ILS) » et sera présenté à la 7e édition de cette conférence le 08 juillet 2018 à Lyon, France.
La version publiée est identique à la version présentée dans ce mémoire.
Le troisième article intitulé « Just-in-time and Lean in Construction: Simulation of a Six-
story Building to Improve Productivity » a été soumis au journal « Construction
Engineering and Management » le 29 mai 2018. La version soumise est identique à la
version présentée dans ce mémoire.
xii
Afin de respecter les politiques de confidentialité du partenaire industriel, Nordic
Structures, les données présentées dans l’étude sont fictives. En effet, les données
quantitatives collectées et les résultats de la simulation ont été modifiés
proportionnellement avec un facteur mathématique. Ce mémoire ne contient donc aucun
détail précis sur les paramètres du chantier de construction ainsi que sur le savoir-faire de la
compagnie.
1
Introduction générale
Depuis plusieurs années, les succès du juste-à-temps (JAT) ont fait de cette philosophie
d’origine manufacturière un sujet de curiosité pour les professionnels et chercheurs des
autres domaines. Dans le domaine de la construction, les entreprises voulant se démarquer
et assurer une pérennité économique de leur organisation, elles se sont alors tournées vers
cette philosophie dans le but d’éliminer les gaspillages et les pertes de temps. Cependant,
les entreprises ne peuvent plus de nos jours aspirer au succès simplement en tenant compte
de l’aspect économique. En effet, l’industrie de la construction étant une industrie clé de
l’économie mondiale, canadienne, et québécoise, elle est également celle qui engendre le
plus haut niveau de pollution (Horvath, 2004). Sur une base par unité, le ciment est le
composant le plus énergivore et le plus polluant du domaine, alors que la construction
représente à elle seule jusqu'à 30 % des émissions mondiales de gaz à effet de serre par an
tout en consommant jusqu'à 40 % de l'énergie mondiale (UNEP, 2018). Compte tenu de la
croissance massive des nouvelles constructions à l’étendue du globe, du fait que deux
millions d’habitants urbains supplémentaires sont attendus d’ici 2030 et de l'inefficacité de
certains bâtiments existants, si aucune action n’est prise, les émissions de gaz à effet de
serre des bâtiments risquent de doubler dans les 20 prochaines années (UNEP, 2018). Les
aspects sociaux et environnementaux se doivent donc d’être considérés dans un souci de
préservation des ressources pour les générations futures et de sain développement des
sociétés dans lesquelles évolueront ces générations. L’usage du bois dans la construction
devient alors une planche de salut gagnant en popularité partout sur le globe.
Le but de la recherche consiste à évaluer comment le JAT, philosophie ayant engendré des
gains dans le domaine manufacturier, peut être appliqué dans la construction et s’avérer une
pratique durable. Pour ce faire, les travaux ont donc voulu expliciter le JAT en construction
tout matériau confondu, évaluer différents scénarios d’implantation du JAT, et ultimement
proposer la meilleure combinaison de paramètres pour maximiser la productivité dans un
projet de construction.
2
La méthodologie utilisée se ramifie en quatre phases, la phase 1, la revue systématique de
la littérature, a permis de comprendre en quoi consiste le JAT dans la construction. La
phase 2, la simulation, a rendu possible la conception d’un modèle de simulation reflétant la
construction réelle d’un bâtiment multi-étages et l’expérimentation de différents scénarios.
La phase 3 a reposé sur la réalisation d’analyses statistiques et sur l’établissement d’un plan
d’expérience, alors que la phase 4 a veillé à faire une synthèse des résultats et à formuler
des recommandations en fonction de la construction à l’étude.
À la lueur des résultats obtenus de la simulation, le meilleur scénario pour la compagnie
partenaire du projet impliquerait d’avoir 3 jours de stock tampon sur le chantier, de garder
le même niveau de préfabrication et de faire du Lean sur le chantier en réduisant de 4 à 8
min le temps de déplacement sur le chantier, afin d’obtenir une durée de la construction
passant de 26,09 à 22,31 semaines, un taux d’utilisation des travailleurs augmentant de
68,9% à 80,9% et l’élimination des risques de ruptures de stock sur le chantier.
Cette étude contribue au milieu académique en explicitant les modalités d'application du
JAT dans l’industrie de la construction. Même si l’étude porte sur un bâtiment en bois
préfabriqué spécifique, les concepts exploités restent assez généraux pour être appliqués à
d'autres projets de construction. Cette étude contribue également à l’industrie en
déterminant la mesure dans laquelle les entreprises pourraient mettre en œuvre les
livraisons JAT, les méthodes Lean dans les activités de chantier et la préfabrication des
matériaux dans un projet de construction, afin d’augmenter la productivité
Le présent mémoire comporte six sections. La première présente une revue littéraire
succincte des grands concepts utilisés dans l’étude, étant donné qu’une revue exhaustive
pour chaque grand thème abordé dans la recherche sera incluse à l’intérieur des trois
articles constituant le cœur du mémoire. La deuxième section explique la méthodologie
utilisée dans la réalisation de l’étude, tandis que les sections trois, quatre et cinq
introduisent les articles rédigés dans le cadre de la maîtrise. La sixième section conclut le
mémoire, suivie de la bibliographie.
3
Revue de la littérature
L’objectif de cette section est d’introduire les concepts généraux utilisés dans cette étude,
grâce à la documentation scientifique disponible sur le JAT, le JAT dans la construction,
ainsi que la simulation. Le premier article portera par la suite principalement sur une revue
systématique du JAT dans le domaine de la construction tandis que les deux autres articles
proposeront une revue sur la construction au Canada et l’intérêt porté au bois dans le
domaine de la construction.
Le juste à temps
Avant de se lancer dans l’utilisation du concept du JAT, il est important d’en comprendre
les origines et de saisir l’essence même de son fonctionnement. Le JAT est une philosophie
développée au Japon au tour des années 50 par Taichi Ohno et ses collègues de chez
Toyota, dont la principale motivation était l’élimination des gaspillages à tous les niveaux
de la chaîne de production (Ghada, 2012). La philosophie est décrite par Sugimori et al.
(1977) comme étant une méthode permettant de réduire considérablement le délai de
production en faisant en sorte que tous les processus produisent les pièces requises au
moment nécessaire, tout en ayant en stock que le minimum pour maintenir le flot du
processus.
Schonberger (1982), cité dans Ebrahimpour and Schonberger (1984), rapporte le concept
comme étant la production et la livraison de produits finis juste à temps pour être vendus,
de sous-assemblages juste à temps pour être assemblés dans les produits finis, d’éléments
fabriqués juste à temps pour être utilisés dans des sous-assemblages et de matériaux achetés
juste à temps pour être transformés en éléments fabriqués. Ebrahimpour and Schonberger
(1984) tirent leurs propres conclusions par la suite et présentent le JAT comme étant la
production d’une unité dans un processus à incorporer juste à temps dans un processus
subséquent.
Une autre vision intéressante est celle de Koskela (2000). Ce dernier souligne que les gains
en productivité, obtenus grâce à la mise en place d’une meilleure politique d’organisation et
4
à l’application de certaines méthodes managériales, connues au départ sous le JAT, ont été
appelées la production Lean depuis le début des années 70. Plusieurs définitions sont ainsi
attribuables au JAT, chacune d’elles restant néanmoins fidèles au but de la philosophie, soit
de produire et de fournir la pièce voulue au moment requis, dans les quantités et la qualité
requises.
Dans la philosophie du JAT, tous les efforts sont concentrés pour offrir le produit au bon
moment, à bas coûts, et à bonne qualité au client. L’harmonisation de ces trois critères,
délais-qualité-prix, présente toujours des difficultés de juste équilibre pour les
organisations. Afin d’équilibrer ces critères sans entraîner des dépenses faramineuses, les
entreprises ont tout intérêt à améliorer leur efficacité opérationnelle.
Suite aux échos des performances de la compagnie Toyota en matière d’efficacité
opérationnelle, plusieurs entreprises se sont inspirées de leurs méthodes. Afin de maximiser
le rapport délais-qualité-prix, les sept catégories de gaspillages identifiées par le système de
production de Toyota et présentées aux paragraphes suivants, deviennent alors très
intéressantes à considérer (Koskela, 2000; Hohmann, 2009).
La surproduction
La surproduction peut être due à la perte d’un lot de produits ou à une fabrication de non
qualité. Indépendamment de son origine, la surproduction entraîne l’utilisation
supplémentaire de matériel, d’énergie et de temps. La surproduction peut être observée
dans le cas où les besoins en matériaux sont mal communiqués à un fournisseur d’éléments
préfabriqués et que ce dernier soit contraint de lancer une nouvelle production pour
éventuellement livrer le bon lot à temps sur le chantier.
L’attente
Les attentes dans un système témoignent souvent d’un manque d’organisation ou de
synchronisation entre les activités. Ne pouvant pas aller plus vite que le goulot dans un
système, les attentes résultent en des processus inefficients et des pertes de temps. Dans
l’exemple mentionné ci-dessus, le lancement en production d’un mauvais lot engendrera un
5
délai de livraison sur le chantier. Les équipes en place n’auront d’autre choix que d’attendre
la livraison des bons matériaux afin qu’elles puissent poursuivre leurs activités.
Le transport
Tout déplacement de matériel, n’apportant aucune valeur ajoutée aux produits, engendre
des pertes d’énergie et des temps d’attente inutiles en aval. En effet, dans le cas où le
mauvais lot est livré sur le chantier avant que les équipes ne s’en aperçoivent, cette
livraison est considérée comme étant inutile puisqu’elle crée des pertes de temps et
d’énergie.
Les processus inappropriés
L’utilisation d’équipements, de méthodes, ou de processus très hautement technologiques et
par conséquent chers sont inutiles si un équipement, une méthode, ou un processus plus
simple permettrait d’effectuer les mêmes activités de fabrication ou d’édification.
Les stocks inutiles
Les stocks inutiles résultent souvent d’une surproduction ou encore d’une mauvaise
prévision de la demande à un moment donné. L’accumulation des stocks engendre
d’énormes coûts d’entreposage et de maintenance. Ainsi, l’espace occupé par des stocks
inutiles sur les chantiers encombrent les espaces déjà restreints et empêchent l’utilisation de
ces espaces pour effectuer d’autres activités à valeur ajoutée, sans mentionner la
détérioration de certains matériaux qui n’est pas à prendre à la légère.
Les mouvements non nécessaires
Les mouvements non nécessaires entraînent des pertes d’énergie et de temps pour les
travailleurs qui doivent se déplacer ou parcourir des distances afin d’effectuer leur travail.
Aussi, les stocks retrouvés sur les chantiers de construction sont généralement composés de
matériaux lourds nécessitant l’usage de grues pour les déplacer. Lorsque ces matériaux sont
déplacés inutilement, des pertes d’énergie et de temps sont observables.
6
Les défectuosités
Les défectuosités entraînent des coûts supplémentaires et des pertes de temps à relancer un
nouveau lot de fabrication ou à effectuer du contrôle qualité sur le lot touché par la
défectuosité.
En récapitulatif à cette introduction sur le JAT, il a été observé que le JAT implique la
fabrication et la livraison du bon produit exactement au moment où le client en a besoin
dans le but de réduire les gaspillages (temps, énergie, et coûts) liés à la logistique, à la
production et au stockage. La section suivante présente le concept général du JAT dans la
construction ainsi que les observations faites sur son existence dans ce secteur d’activité.
Le juste à temps dans la construction
Au cours des dernières années, les manufacturiers ont travaillé pour améliorer la
productivité et la qualité de leurs processus tout en réduisant les délais. Les chercheurs et
les industriels du milieu de la construction ont quant à eux effectué des progrès dans
l’amélioration des pratiques Lean sur le terrain (Zimmer, 2008). Cependant, la majorité de
leurs efforts furent axés sur le chantier en tant que tel, négligeant ainsi la chaîne de valeur
complète des opérations de construction (Zimmer, 2008).
D’après la définition proposée par Cossio and Cossio (2012), le JAT dans le domaine de la
construction consiste à fabriquer de plus petits lots de chaque pièce et à les envoyer au
moment requis afin de respecter les échéanciers et de réduire l’espace d’entreposage sur le
chantier. En effet, Akintoye (1995) cité par Pheng (2001) souligne que le JAT peut être
appliqué dans la logistique de management des sites de construction afin d’augmenter le
niveau de productivité. Une autre définition avancée par Opfer (1998) rapporte le JAT
comme étant une stratégie dans laquelle le rapprochement des matériaux et des composants,
le plus près possible des chantiers, permettrait de les avoir sur les chantiers au moment où
ils sont nécessaires pour l’installation.
Suite aux différents articles lus des auteurs Bjornfot, A. and Sarden, Y. (2006), Salem
(2006), Koskela (2000), Koskela et al. (2004), Bo, J. and Stephen, E. (2008), et Eriksson
7
(2010), la définition suivante peut également être proposée au JAT dans le domaine de la
construction. Le JAT dans le domaine de la construction consiste à s’assurer de la livraison
de produits par les fournisseurs, sur les chantiers de construction, au bon moment, afin que
les temps alloués à chaque tâche soient respectés. Ainsi à postériori, le temps prévu pour la
réalisation du projet équivaudrait au temps véritablement investi pour effectuer la
construction.
Selon Lim et Low (1992) cités par Pheng (2001), avec le système de management JAT en
place, les matériaux peuvent être livrés sur le chantier de construction le jour même de leur
utilisation ou un jour avant. Les changements et imprévus étant inhérents au domaine de la
construction, dans la pratique, n’avoir aucun stock présenterait des risques élevés de
ruptures de matériaux, d’où le concept de JAT modifié introduit par Pheng and Chuan
(2001). Le JAT modifié consiste à garder un certain niveau de stock sur le chantier, mais la
quantité de ce stock devra être définie stratégiquement. Cependant, il a été observé que
lorsque les matériaux sont livrés en JAT, s’ils nécessitent une sorte de conditionnement sur
le chantier avant leur utilisation, ou si la localisation des matériaux sur le chantier engendre
beaucoup de recherche et de déplacements inutiles pour les travailleurs, les livraisons JAT
perdraient alors de leur intérêt. D’où l’importance de la préfabrication des matériaux et du
Lean dans les activités du chantier.
Les projets de construction étant des systèmes complexes, de par la présence de différents
intervenants et les interactions multiples qu’on y retrouve, la simulation se positionne
comme un outil parfait dans l’étude de ce type de systèmes. La section suivante introduit le
concept de la simulation et la structure de la simulation à événements discrets.
Le concept de la simulation
La simulation est définie comme étant le développement d’un modèle mathématique
logique d’un système réel et de manipulations expérimentales du modèle (Biles 1987). Les
termes « modèle » et « système » sont importants dans la compréhension de la simulation,
dans la mesure où le modèle s’avère le moyen choisi pour illustrer les caractéristiques
importantes du système (Biles 1987). Comme expliqué par Biles (1987), un système est une
8
collection d'éléments d'un secteur circonscrit de la réalité faisant l’objet de l’étude. Les
frontières du système sont définies afin d'inclure les éléments qui sont jugés les plus
importants et d'exclure les éléments de moindre importance à l’étude. Le modèle doit
posséder des représentations des entités ou objets du système et refléter les activités dans
lesquelles s’engagent ces entités.
Comme décrit par Banks (1999), la simulation est l'imitation des opérations d’un système
du monde réel au fil du temps. La simulation implique la génération d'une histoire
artificielle du système et l’observation de cette histoire artificielle pour tirer des conclusions
concernant les caractéristiques de fonctionnement du système réel qui est représenté. La
simulation est également définie comme étant une méthodologie de résolution de problème
indispensable pour la résolution de nombreux problèmes du monde réel. Elle est utilisée
pour décrire et analyser le comportement d’un système, poser des questions au conditionnel
du type « si » sur le système réel, et aider à la conception de systèmes réels (Banks, 1999).
La simulation, technique utilisée dans cette étude, s’appuie donc sur une représentation du
fonctionnement d’un système réel, soit la construction d’un bâtiment multi-étages en bois,
afin d’anticiper l’évolution des caractéristiques du système si soumis à différentes
configurations d’implantation de la philosophie du JAT.
Les simulations peuvent être classées en fonction de leur stratégie de mise en œuvre. La
simulation de systèmes continus, la simulation de Monte Carlo, la simulation à événements
discrets, la simulation hybride et la simulation basée sur des agents sont toutes des
stratégies de mise en œuvre particulières (White and Ingalls, 2015). La simulation à
événements discrets est celle utilisée dans cette étude. Dans ce type de simulation, les
variables d'état changent à des points discrets dans le temps où les événements se
produisent. Les événements se produisent en conséquence à des temps d'activité et des
délais. Les entités peuvent être en compétition pour l’obtention des ressources du système,
éventuellement en joignant des files d'attente où elles attendent que la ressource soit
disponible (Banks, 1999). Un modèle de simulation à événements discrets est conduit au fil
du temps par un mécanisme qui déplace vers l’avant le temps simulé. L'état du système est
9
mis à jour à chaque événement avec la saisie et la libération des ressources qui peuvent se
produire à ce même temps.
Comme expliqué par Ingalls (2008) et White and Ingalls (2015), peu importe la complexité
de la simulation, un modèle conçu à l’aide de la simulation à événements discrets comporte
les composants de base présentés ci-dessous.
Entrées, sorties, et état
Les actions de l'environnement sur un système sont appelées les entrées du système. Ces
entrées provoquent des changements dans la condition interne du système, appelée l'état du
système. Les sorties sont ces quantités mesurées, qui peuvent être dérivées de l'état du
système. Ce sont les sorties qui permettent de répondre aux interrogations ayant initié
l'étude de la simulation. En d'autres termes, les entrées provoquent des changements dans
l’état du système et ceux-ci sont reflétés par des changements dans la sortie.
Entités et attributs
Les entrées sont créées par l’arrivée dynamique des entités dans le système. Sans entités,
rien ne se passerait dans le système. Ces entités se déplacent à travers le système et sont les
éléments structurels qui affectent les changements au niveau des variables d’état du
système. Des exemples d’entités dans le système à l’étude sont les réapprovisionnements de
matériaux de l’usine au chantier et les commandes de matériaux de la zone de stock tampon
au chantier. Un réapprovisionnement comme une commande possède des caractéristiques
lui permettant d’obtenir des quantités de différents matériaux en fonction de l’avancement
de la construction. En effet, une condition d'arrêt pour une simulation est le fait qu’il n'y ait
pas d'entités actives dans le système. Dans le système à l’étude, la condition d’arrêt repose
sur le moment où les quantités nécessaires de matériaux pour achever la construction sont
toutes consommées.
Les entités ont des attributs. Ces attributs sont des caractéristiques qui sont uniques à cette
entité. Dans la simulation à l’étude par exemple, chaque réapprovisionnement ou chaque
10
commande possède des attributs uniques tels que leurs heures d’arrivée exactes dans le
système, leurs tailles, et leurs destinations.
Activités et événements
Les activités sont des processus et de la logique dans la simulation. Les événements sont
des conditions qui se produisent à un moment donné et qui provoquent un changement dans
l'état du système. Une entité interagit avec des activités pour créer des événements. Il y a
trois principaux types d'activités dans une simulation: les délais, les files d'attente et la
logique.
Une activité de délai se produit lorsque le déplacement d'une entité est suspendu pendant
une période définie. Dans le système à l’étude, l’on peut noter comme activité de délai le
temps de déchargement des matériaux qui suspend le déplacement de l’entité de
réapprovisionnement à la zone de stock tampon.
Les queues sont des endroits dans le système où les entités attendent pour une période non
déterminée. Les entités peuvent attendre que des ressources soient disponibles ou qu’une
condition spécifique du système se produise. Dans le système à l’étude, un exemple est la
queue dans laquelle les matériaux préparés à la zone de stock tampon attendent jusqu’à leur
utilisation.
Les activités de logique permettent aux entités d’affecter l’état du système grâce à la
manipulation des variables d'état ou de décisions logiques. Dans le système à l’étude, un
exemple serait la décision logique dont le but est de vérifier s’il y a assez de matériaux en
inventaire. Si ce n’est pas le cas, un réapprovisionnement est lancé, affectant ainsi l’état du
système avec une nouvelle entrée de réapprovisionnement.
Ressources
Les ressources représentent n’importe quel élément dans une simulation qui a une
contrainte de capacité. Les ressources sont partagées en temps par les entités et les entités
sont généralement retardées après avoir saisi une ressource. Des exemples communs de
11
ressources incluent les travailleurs, les machines, les nœuds dans un réseau de
communication et les intersections de trafic. La grue, le grutier, le chariot élévateur, le
conducteur du chariot élévateur, et les travailleurs sont des exemples de ressources dans le
modèle à l’étude.
Variables globales
Les variables globales permettent de suivre tout ce qui pourrait être d’intérêt. Les quantités
économiques à commander (QEC) des différents matériaux par division en sont des
exemples. En modifiant ces variables, le simulateur peut analyser l'efficacité de différentes
configurations de QEC.
Générateur de nombres aléatoires
Chaque outil de simulation a un générateur de nombres aléatoires. Le générateur de
nombres aléatoires (techniquement appelé un générateur de nombres pseudo-aléatoires) est
un logiciel de routine qui génère un nombre aléatoire indépendant uniformément réparti
entre 0 et 1. Ce nombre est ensuite utilisé pour échantillonner d'autres distributions
aléatoires. Par exemple, dans le système à l’étude, il a été déterminé que le temps de
déchargement des camions est uniformément distribué entre 10 minutes et 20 minutes.
Ainsi, à chaque fois qu'une entité de réapprovisionnement passe par ce processus, le
générateur de nombres aléatoires génère un nombre entre 0 et 1 et évalue la formule de la
distribution uniforme qui a un minimum de 10 et un maximum de 20 unités de temps. À
titre d'exemple, supposons que le nombre aléatoire généré est 0,8134. Le temps de
déchargement serait alors 10 + (0,8124) * (20-10) = 18,124 unités de temps. Le générateur
de nombres aléatoires est donc utilisé comme une entrée pour déterminer tout ce qui est
valeurs aléatoires dans la simulation.
L’horloge et le calendrier
L'horloge est une variable globale qui porte la valeur de l'heure actuelle dans la simulation.
Le calendrier est une liste d'événements programmés pour se produire dans le futur, c'est-à-
dire, à des heures d'horloge plus tard que l’heure actuelle. Dans chaque simulation, il n'y a
qu'un seul calendrier et il est ordonné en fonction du temps planifié le plus ancien. À
12
n’importe quel temps x dans la simulation, chaque événement qui a déjà été planifié à se
produire dans le futur, est tenu sur le calendrier.
Les collecteurs statistiques
Les collecteurs de statistiques recueillent des statistiques sur la valeur des variables
globales, les attributs des entités, etc. Trois différents types de statistiques peuvent être
collectées, celles utilisées pour compter, comme pour compter le nombre de
réapprovisionnements effectués dans le système, celles utilisées pour calculer une valeur à
travers le temps, comme le taux d’utilisation des travailleurs, et celles de pointage qui
recueillent des valeurs, une observation à la fois, sans égard à la quantité de temps entre les
observations.
La simulation permet de mieux comprendre les interactions entre les variables, de
diagnostiquer les problèmes et d’avoir un meilleur aperçu de l’importance de ces variables;
ceci dans le but d’obtenir une compréhension plus éclairée des effets importants de ces
variables sur la performance globale du système (Banks, 1999). Cependant, en plus du
niveau élevé de compétences et d’expérience requises pour développer un bon modèle de
simulation et interpréter ses résultats (Banks, Carson and Nelson, 1996; Law and Kelton,
1991; Pegden, Shannon and Sadowski, 1995; and Schriber, 1991 cité par Banks, 1999), la
vérification et la validité du modèle demeurent la clé sans quoi une faible confiance sera
accordée aux conclusions issues de la simulation. La vérification du modèle est le processus
permettant de déterminer que le modèle fonctionne comme prévu (Centeno, 1996). La
validation, quant à elle, est un processus permettant de s'assurer qu'un modèle est une
représentation acceptable du système réel (Centeno, 1996).
13
Objectifs et méthodologie
La recherche effectuée dans le cadre de la maitrise avait pour objectif général d’évaluer
comment le JAT, philosophie ayant apporté des retombées positives dans le domaine
manufacturier, pouvait être mise en pratique dans le domaine de la construction et générer
des bénéfices. Les objectifs spécifiques poursuivis étaient :
• D’expliciter le JAT en construction tout matériau confondu;
• D’évaluer différents scénarios d’implantation du JAT;
• De proposer la meilleure combinaison de paramètres pour maximiser la productivité
dans un projet de construction.
Pour ce faire, l’étude a été réalisée en quatre grandes phases présentées ci-bas. Les quatre
phases incluent la revue systématique de la littérature, la simulation, les analyses
statistiques, ainsi que la synthèse et les recommandations.
La phase 1 entièrement présentée dans l’article 1 a consisté en la réalisation d’une revue
systématique de la littérature sur le JAT dans la construction afin de mieux saisir le concept
et de bien cerner comment les organisations peuvent le mettre en pratique. À partir de cette
revue, un cadre préliminaire proposant des pistes de déploiement du JAT a alors pu être
réalisé. La phase 2 discutée en partie dans l’article 2 a reposé sur la conception d’un modèle
de simulation reflétant la construction d’un bâtiment en bois de six étages. À partir de ce
modèle, il devenait alors possible de tester différents scénarios de déploiement de stratégies
de JAT tout en mesurant leur valeur. La phase 3 a fait appel à des analyses statistiques afin
de mieux cerner les stratégies ayant le plus d’effet sur trois indicateurs clés. Une synthèse
des résultats obtenus ainsi que des recommandations pour le futur ont finalement été
formulées, ce qui constitue la phase 4 de la recherche. L’expérimentation, l’analyse des
résultats, la synthèse et les recommandations se retrouvent à l’intérieur de l’article 3 du
présent mémoire. La figure 1 résume chacune des phases exécutées.
14
Figure 1 Méthodologie de recherche.
Phase 1 : Revue systématique de la littérature
À la première phase de l’étude, une recherche systématique sur le JAT a été effectuée dans
la documentation scientifique. Comme mentionné par Tranfield et al. (2003), la revue
systématique de la littérature consiste à « améliorer la qualité du processus de revue en
synthétisant les recherches de manière systématique et reproductible ». Il a tout d’abord été
question de définir les objectifs poursuivis en effectuant la revue systématique. En effet, la
revue systématique de la littérature a été menée afin de mieux comprendre le JAT et
d’évaluer comment le JAT permettait de résoudre les problèmes observés dans la
construction. Ainsi, l’analyse des résultats issus de la littérature a permis de confirmer la
15
présence du JAT en construction, de cibler les facteurs tels que le Lean et la préfabrication
comme étant des moyens de déploiement de la philosophie et les indicateurs clés à utiliser
pour mesurer les bénéfices liés à l’implantation du JAT dans la construction.
Phase 2 : Simulation
En parallèle à la phase 1, la collecte de données et la modélisation du système réel ont
débuté, impliquant des allers-retours entre l’une et l’autre des deux étapes afin de concevoir
un modèle de simulation qui reflète bien la réalité. Le logiciel utilisé fut Simio, un logiciel
3D reposant sur la simulation de flux à événements discrets (Simio, 2018). La collecte de
données a consisté à recueillir auprès du partenaire industriel les plans de la construction,
les horaires des livraisons de matériaux sur le chantier et les données relatives aux quantités
de matériaux, d’équipements et d’équipes de travail utilisés.
À l’étape de modélisation, un modèle conceptuel a tout d’abord été effectué, puis converti
en model digital dans Simio. Le modèle digital initial était constitué de l’usine, de la zone
de stockage et du bâtiment.
En parallèle, la collecte de données a continué, en effectuant des rencontres avec le
partenaire industriel et des visites de chantiers. Le flux de matériaux a ensuite été ajouté,
puis validé en collaboration avec le partenaire industriel. En effet, la conversion du model
conceptuel en model digital en soi est insignifiante à moins que ce dit model digital soit
minutieusement vérifié et validé (Centeno, 1996). La fiabilité du modèle est directement
affectée par la qualité du processus de vérification et de validation.
La vérification implique d'assurer que les codes et logiques du modèle fonctionnent comme
prévu, et la validation cherche à montrer que le comportement du modèle représente celui
du système réel à l’étude (Centeno, 1996). Les validations ont valu à plusieurs reprises des
modifications sur le modèle, la bonification du modèle digital initial avec de nouveaux
intrants. En effet, les bonifications ont impliqué entre autres, l’explosion du bâtiment en
divisions puis en sous-assemblages, l’ajout des équipes de travail, de leur horaire, de leur
pause dîner, et des divers équipements utilisés (la grue, le chariot élévateur, et l’abri des
travailleurs) dans le but d’instaurer un dynamisme plus réel des activités sur le chantier. Les
différents groupes de logique comme la logique de réapprovisionnement de l’usine au
chantier ont ensuite été développés.
16
Les expérimentations des différents scénarios ont dès lors pu être effectuées, permettant
d’obtenir les résultats présentés dans les articles 2 et 3. Les scénarios testés concernaient
toutes les combinaisons possibles de niveau de Lean (0, 50, ou 100%), de niveau de
préfabrication (80 ou 100%), de nombre de jours de stock tampon sur le chantier (1, 2,
2,33, 3, ou 4 jours), et de nombres d’équipes sur le chantier (1 ou 2). Les scénarios testés
dans l’article 2 incluent les scénarios avec une équipe de travail uniquement. Les scénarios
avec deux équipes, soit une équipe le matin et une équipe le soir ont été testés et présentés
dans l’article 3. Grâce à l’étape de validation impliquant diverses consultations avec
l’expert du projet auprès de la compagnie, plusieurs modifications ont été apportées au
modèle dont une qui a permis de mettre à jour le séquençage des activités des travailleurs et
des équipements sur le chantier. Après l’étape de validation revisitée plus d’une fois, les
différents scénarios d’implantation ont été suivis de l’analyse des résultats obtenus et ainsi
que de discussions des conclusions préalables avec le partenaire industriel.
Phase 3 : Analyses statistiques
À la phase 3, un design d’expériences a été fait afin de déterminer l’influence des différents
paramètres sur les indicateurs de performance utilisés dans les analyses, soit la durée de la
construction, le taux d’utilisation des travailleurs et les ruptures de stock. Le design
d’expériences a consisté en l’utilisation d’une approche factorielle sous Minitab, un logiciel
statistique conçu pour analyser des données, déterminer les causes des problèmes observés
dans le système et ultimement améliorer le système. La première analyse statistique
effectuée a permis de déterminer les facteurs ayant un effet significatif sur chacun des
indicateurs de performance, la seconde analyse de déterminer quels facteurs influençaient
négativement, positivement ou de manière neutre les indicateurs et la troisième de
déterminer si l’interaction entre les différents facteurs avait une influence sur chacun des
indicateurs de performance. L’impact de la variation des facteurs définis a ensuite été
mesuré sur les trois indicateurs clés. Grâce au design d’expériences (DOE), il a été possible
de déterminer comment chaque paramètre et leurs interactions influençaient la productivité
du projet de construction, ce qui a permis d’établir des recommandations.
17
Phase 4 : Synthèse et recommandations
À la phase 4, les grandes conclusions issues de l’interprétation des résultats de la simulation
et des analyses statistiques ont été dégagées. Le meilleur scénario d’implantation du JIT a
été proposé au partenaire industriel, accompagné des recommandations telles que la mise
de l’avant du Lean sur le chantier et dans les activités de construction pour maximiser les
gains. Le cadre d’implantation préliminaire proposé suite à la revue littéraire a ensuite été
revisité et amélioré à la lueur des résultats issus de l’étude.
18
Just in time in construction: description and
implementation insights
L’article intitulé « Just in time in construction: description and implementation insights »
est inséré dans cette section du mémoire. Il a été soumis le 23 novembre 2016 à la
conférence « LC3 : Lean & Computing in Construction Congress » et présenté à la 25e
édition de cette conférence le 10 juillet 2017 à Crète, Grèce. La version publiée est
identique à la version présentée dans ce mémoire.
19
Résumé
L'industrie de la construction a depuis longtemps été en proie aux problèmes de
productivité et de gestion des gaspillages sur les chantiers de construction. Cet article étudie
comment la philosophie JAT pourrait aider à pallier à ces problèmes. L'article illustre
quatre scénarios de mise en œuvre du JAT en construction selon le niveau de coordination
requis, la gestion sur le chantier visée et le partage de l'information désiré. La méthodologie
qui consiste en une revue systématique de la littérature sur le JAT dans la construction
confirme la nécessité d'adapter cette philosophie pour un déploiement adéquat dans la
construction. Les conclusions issues de la revue systématique de la littérature confirment
également les liens étroits entre le JAT, le Lean et la préfabrication pour une gestion
efficace d’un projet de construction.
20
Abstract
The construction industry has long been plagued with productivity and waste management
issues on construction sites, unmet deadlines, and client dissatisfaction over the quality of
the construction delivered. Having greatly aided the manufacturing industry, this paper
investigates how the JIT philosophy could help with these difficulties. The paper illustrates
four scenarios of JIT implementation in construction according to the level of coordination
required, on-site management, and information sharing. The methodology which consists of
a systematic literature review on JIT in construction confirms the need to adapt this
philosophy for an adequate deployment in this industry. It also confirms the close ties
between JIT, lean construction, and prefabrication for successful construction project
management.
21
Introduction
Several authors address the fact that productivity in the construction industry has improved
significantly but is still lagging behind other industries (Pheng and Chuan 2001; Asri et al.
2016). A way to address this issue is to embrace the just in time (JIT) philosophy which
consists of providing the right materials, in the right quantities and quality when it is
needed in order to reduce waste and to provide maximum value (Tommelein and Li 1999;
Tommelein and Weissenberger 1999; Vokurka and Davis 1996). More research indicates
that JIT in construction consists of producing smaller batches of each component and
sending them on site at the required installation time in order to reduce waste, diminish on-
site storage space, and meet the deadlines and high standards of the construction industry
(Cossio and Cossio 2012). The goal of the paper is to show how JIT may reduce costs,
waste, and quality problems encountered in a construction project. To achieve this goal,
different scenarios of JIT implementation are proposed determined by the level of
coordination required, construction site management, and information sharing. The paper
contributes to scholars by identifying elements that influence JIT implementation in
construction and to practitioners by presenting requirements for its successful
implementation. The paper contains six parts. An introduction, a systematic literature
review (SLR), a discussion of results, a conclusion, references, and an appendix.
JIT in construction: systematic literature review
In order to show how JIT can help to better deal with issues in construction, a scientific
methodology called systematic literature review (SLR) is applied. The goal of a SLR is to
“improve the quality of the review process by synthesizing research in a systematic and
reproducible manner” (Tranfield et al. 2003). As these authors agreed, conducting a
systematic review passes through three stages: planning, conducting, and reporting.
Planning the review
The first phase in the planning stage is to state the need for the review. This review is
needed in order to present studies in which JIT has been implemented in a construction
project. The studies will be used to equip practitioners with proper strategies of
22
implementation and if not, to highlight gaps in those studies that could be filled in future
studies. The second phase consists of preparing questions that should be answered by the
review. For this research, the following questions are identified:
1) Is there a presence of JIT in construction mentioned in scientific literature?
2) Which shape does JIT take in construction?
3) Which indicators measure performance when JIT is implemented in
construction?
The third phase presents the methodology used while conducting the review. Step one of
phase three consists of an identification of keywords that will be used in the review. In the
second step, a search in all engineering databases with established keyword combinations is
done. The third step consists of an analysis of titles and abstracts of results. In the fourth
step, articles that do not respect logical pre-established criteria are eliminated. The fifth step
refers to a complete analysis and classification of the remaining articles.
Conducting the review
Identification of research
The databases used in the search of scientific articles are Compendex and Inspec from the
research databases Engineering Village and Web of science. Engineering Village has been
selected because it is an excellent choice for rigorous scientific research specific to
engineering. Web of science has also been chosen because it covers a wide variety of
scientific areas. The research coverage period has been left by default, from 1884 to 2016
in Engineering village and from 1900 to 2016 in Web of science in order to cover all
potential articles. The research was done for the last time on September 27, 2016. Key
concepts of the subject matter are “just in time” and “construction”. Synonyms found for
each of them in the synonym database Thesaurus are "JIT" and "Build-to-order" for "just in
time", and "building" and "prefabrication" for "construction" (an asterisk is used after
prefabricat to search all words starting with prefabricat and finishing with all its suffixes).
The research consisted of equations which are combinations of key concepts and their
synonyms while specifying their appearance in the Subject, the Title or the Abstract. The
nine research equations are as follow: “just in time” Near/6 construction, “just in time”
Near/6 building, “just in time” Near/6 prefabricat*, “JIT” Near/6 construction, “JIT” Near/6
23
building, “JIT” Near/6 prefabricat*, “Buid-to-order” Near/6 construction, “Buid-to-order”
Near/6 building, and “Buid-to-order” Near/6 prefabricat*. In all research equations,
"Near/6" is used to specify the presence of at most six words between the two key concepts.
After testing all combinations, the Boolean operator OR was used in order to present results
of all combinations together. A third database from the International Group for Lean
Construction (IGLC) was also used because IGLC has been working on the subject for
several years. IGLC conference papers relating to JIT in construction were searched on
January 16, 2017.
Selection of studies
A total of 164 articles were found with Engineering Village, 227 with Web of science, and
83 with IGLC. In order to refine the results in Web of science, only articles from
categories: operations research management science, industrial engineering, management,
manufacturing engineering, civil engineering, and multidisciplinary engineering are kept
since they relate more to JIT in construction. Of the 146 articles left in Web of science, 164
in Engineering Village, and 83 in IGLC, titles and abstracts are scanned by using five
logical criteria. Articles kept are articles discussing only JIT in building construction,
construction supply chain, tools used to implement JIT in construction, JIT success factors,
and barriers to the presence of JIT in construction. This leaves us with a result of 23 articles
from which critical information is extracted such as performance indicators used to measure
the impact of the implementation of JIT in a construction project and tools used to undergo
the implementation. Table A1 in the appendix summarizes the findings.
Discussion
Reporting and dissemination
Analysis of resulting papers showed that scientific literature covers Lean in construction
(eight papers) and JIT in construction (7) more than JIT implementation in construction (3).
Indeed, few articles discuss an implementation of JIT in construction even though some of
them underlined the impact of prefabrication and Lean techniques used on construction
sites. No paper showed quantitative results on the impact of different levels of
prefabrication as well as buffer stock and construction site organisation on the outcome of a
24
construction. Eighteen out of twenty-three articles discussed JIT in construction, JIT
implementation in construction along with Lean in construction and prefabrication.
Response to research questions
As the first research question concerned the presence of JIT in construction in scientific
literature, we can now confirm that it is the case. The second question dealt with the form
JIT typically takes in construction. The case studies from the SLR mention and describe
how JIT has been applied in several construction projects whether big or small. However,
Tommelein and Weissenberger (1999) and Pheng and Chuan (2001) maintain that in
practice a buffer is necessary and its size should be determined strategically. In
construction, JIT mostly covers the management of deliveries (Tommelein and Li 1999;
Asri et al. 2016) and control of buffer stock levels on construction sites (Pheng 2001;
Pheng and Chuan 2001; Ng et al. 2009; Roos et al. 2010; Amornsawadwatana 2011; Cossio
and Cossio 2012) through a pull system (Akintoye 1995; Tommelein and Li 1999; Low and
Wu 2005; Viana et al. 2015). Furthermore, the implementation of JIT principles in
construction seems to require prefabrication (Cossio and Cossio 2012), Lean techniques, an
integration of the materials procurement time frame with the construction project schedule,
and evaluation of supplier performance to ensure quality of delivered materials while
avoiding rework on site (Opfer 1998). Figure 2a presents the number of times these
elements are mentioned to support an implementation of JIT in construction. Question three
was about the indicators used to measure performance when JIT is implemented in
construction. Figure 2b illustrates the number of times KPIs such as costs are mentioned to
assess the results of an implementation of JIT in construction. The most frequently
mentioned KPIs, in descending order, are costs, productivity, project duration, amount of
buffer stock, quality of construction, and waste quantity. Five out of the six KPIs are
elements that impact only a portion of the building’s life cycle. However, the fifth KPI:
quality of construction, which depends on client satisfaction, should have a greater weight
since it lasts throughout the project's entire life cycle. Moreover, the adoption of JIT in
construction seems to generate qualitative results such as better partnership between
suppliers and contractors and improved system of deliveries (Cossio and Cossio 2012).
25
Figure 2 Number of occurrences of JIT elements and KPI mentioned in the SLR.
Description of different implementation scenarios of JIT
According to the literature, JIT benefits on a construction project’s value chain depend on
the prefabrication plant, the construction site, and the flows between them. As pointed out
by Pheng (2001), one party trying to adopt a JIT philosophy while the other does not, will
realize his efforts are futile and will not achieve the potential benefits. The situation is also
depicted in Viana et al. (2015). Figure 3 illustrates four scenarios of JIT in construction.
Scenario I shows that low information sharing between the plant and the site and low
supply chain coordination are respectively less favourable to (-) prefabrication of
components and (-) JIT deliveries while low construction site management suggests (-) low
presence of Lean principles on the site. In scenarios II and III, one or two elements are less
present, making it difficult to obtain gains on the overall value chain. Scenario IV illustrates
the best scenario for JIT implementation in construction where high information sharing
and high supply chain coordination are respectively more favourable to (+) prefabrication
of components and (+) JIT deliveries while high construction site management suggests (+)
the presence of Lean principles on the site.
26
Figure 3 Scenarios of implementation of JIT in construction.
Pheng (2001) states that prefabrication has come with “the hope of reaping the benefits of
factory-styled operations” since it entails benefits such as improved quality, waste
reduction, and faster erection of buildings. However, even if an efficient plant makes JIT
deliveries, if there are no resources available to unload them on site, the truck and its driver
will remain monopolized. Such a situation illustrates the need for JIT deliveries to ensure
the fast erection resulting from prefabrication but also the need to use Lean in construction
site activities to avoid waste of time (Friblick et al. 2009). For example, Khalfan et al.
(2008) used Kanban, “one of the Lean approaches” in their construction project “to pull
construction materials through their production systems on a just-in-time basis”. Moreover,
Lean tools such as 5S (Deshpande et al. 2012) are used to ensure that the construction site
is well organized to improve performance of workers and optimal movements of materials
on site. It can be concluded that successful prefabrication needs successful implementation
of JIT while successful JIT implementation in construction requires adoption of Lean in
construction site activities. However, the scenarios present some limitations. It is supposed
that the plant is able to provide materials regardless of the demand and that it can
manufacture different levels of prefabricated materials. These scenarios do not apply if the
plant does not meet such expectations.
Conclusion
The research showed through the SLR that JIT exists in construction and helped identify
the JIT elements applied as well as the KPIs used to measure the impact of their
implementation. Moreover, the research proposed four scenarios to illustrate the influence
27
of three JIT elements (prefabrication, JIT deliveries, and presence of Lean principles) on
the successful implementation of JIT in construction. However, the aforementioned
scenarios are purely qualitative. Future work will use activity-based software to simulate
different scenarios presenting various levels of prefabrication, on-site buffer stock, and
Lean on-site activities in order to obtain quantitative KPI measurements.
References
Akintoye, A. (1995). Just-in-time application and implementation for building material
management. Constr. Management and Economics, 13, 105-113.
Amornsawadwatana, S. (2011). Effective design of the construction supply chain: A case of
small buildings in Thailand. In: IEEE Int. Conf. on Industrial Engineering and
Engineering Management. [online] Piscataway: IEEE Computer Society, 206-210.
Available at: http://dx.doi.org/10.1109/IEEM.2011.6117908 [Accessed 02 Sept. 2016].
Anglin, R. & Popescu, C. (1995). Just-in-time: Applicability in construction. In: 39th Ann.
Meeting of the AACE-Int. [online] Morgantown: Amer Assoc Cost Engineers, 48-50.
Asri, M. a. N. M., Nawi, M. N. M. & Nadarajan, S. (2016). Key factors of successful JIT
integration with IBS - An overview. In: Int. Conf. on Applied Science and Technology.
[online] Huntington: AIP - American Institute of Physics. Available at:
http://dx.doi.org/10.1063/1.4960905 [Accessed 01 Sep. 2016].
Christensen, W. (2005). Build-to-order and just-in-time as predictors of applied supply
chain knowledge and market performance. J. of operations management, 23, 470-481.
Cossio, J. G. & Cossio, A. G. (2012). Application of just in time to the fabrication and
installation of prefabricated concrete facades in buildings. In: 20th Conf. of the Int.
Group for Lean Constr. [online] United States: The Int. Group for Lean Constr.
Deshpande, A. S., Filson, L. E., Salem, O. M. & Miller, R. A. (2012). Lean techniques in
the management of the design of an industrial project. J. of Management in
Engineering, 28, 221-223.
Friblick, F., Tommelein, I.D., Mueller, E. & Falk, J.H. (2009). Development of an
integrated facade system to improve the high-rise building process. In: 17th Conf. of the
Int. Group for Lean Constr. [online] United States: The Int. Group for Lean Constr.
Available at: http://iglc.net/Papers/Details/636. [Accessed 16 Jan. 2017].
Horman, M. J. & Thomas, H. R. (2005). Role of inventory buffers in construction labor
performance. J. of Constr. Engineering and Management, 131, 834-843.
Khalfan, M. M. A., Mcdermott, P., Oyegoke, A. S., Dickinson, M. T., Li, X. & Neilson, D.
(2008). Application of kanban in the UK construction industry by public sector clients.
In: 16th Ann. Conf. of the Int. Group for Lean Constr. [online] United States: University
of Salford, 347-358. Available at:
https://iglcstorage.blob.core.windows.net/papers/attachment-28869d08-a5b3-419c-
9112-ce67baca9e25.pdf. [Accessed 02 Sep. 2016].
Low, S. P. & Wu, M. (2005). Just-in-time management in the ready mixed concrete
industries of Chongqing, China and Singapore. Constr. Management and Economics,
23, 815-829.
28
Ng, S. T., Shi, J. & Fang, Y. (2009). Enhancing the logistics of construction materials
through activity-based simulation approach. Engineering, Constr. and Architectural
Management, 16, 224-237.
Opfer, N. D. (1998). Just-in-time construction materials management. In: Proceedings of
the 1998 42nd Ann. Meeting of AACE Int. [online] Morgantown: AACE Inc. Available
at: http://connection.ebscohost.com/c/articles/2449026/just-in-time-construction-
materials-management [Accessed 01 Sep. 2016].
Pheng, L. S. (2001). Just-in-time management in precast concrete construction: A survey of
the readiness of main contractors in Singapore. Integrated manufacturing systems, 12,
416-429.
Pheng, L. S. & Chuan, C. J. (2001). Just-in-time management in precast concrete
construction: A survey of the readiness of main contractors in Singapore. J. of Constr.
Engineering and Management, 127, 494-501.
Roos, A., Woxblom, L. & Mccluskey, D. (2010). The influence of architects and structural
engineers on timber in construction - perceptions and roles. Silva Fennica, 44, 871-884.
Tommelein, I.D. & Li, A. E. Y. (1999). Just-In-Time concrete delivery: Mapping
alternatives for vertical supply chain integration. In: 7th Conf. of the Int. Group for
Lean Constr. [online] United States: The Int. Group for Lean Constr. Available at:
http://iglc.net/Papers/Details/86. [Accessed 16 Jan. 2017].
Tommelein, I.D. & Weissenberger, M. (1999). More Just-In-Time: Location of buffers in
structural steel supply and construction processes. In: 7th Conf. of the Int. Group for
Lean Constr. [online] United States: The Int. Group for Lean Constr. Available at:
http://iglc.net/Papers/Details/87. [Accessed 16 Jan. 2017].
Tranfield, D., Denyer, D. & Smart, P. (2003). Towards a methodology for developing
evidence-informed management knowledge by means of systematic review. British J. of
Management, 14, 207-222.
Trivedi, J. & Kumar, R. (2014). Optimisation of construction resources using lean
construction technique. Int. J. of Engineering Management and Economics, 4, 213-28.
Viana, D.D., Tillmann, P.A., Zach, S., Tommelein, I.D., & Formoso, C.T. (2015). Analysis
of HVAC subcontractor mechanisms for JIT Materials supply to a construction site. In:
23rd Conf. of the Int. Group for Lean Constr. [online] United States: The Int. Group for
Lean Constr. Available at: http://iglc.net/Papers/Details/1171. [Accessed 16 Jan. 2017].
Vokurka, R. J. & Davis, R. A. (1996). Just-in-time: The evolution of a philosophy.
Production and Inventory Management J., 37, 56.
Yong-Woo, K., Azari-N, R., June-Seong, Y. & Jinwoo, B. (2013). Environmental impacts
comparison between on-site vs. prefabricated Just-In-Time (prefab-JIT) rebar supply in
construction projects. J. of Civil Engineering and Management, 19, 647-55.
Zimmer, E., Salem, O., Genaidy, A. & Shell, R. (2008). Case study: Lean supply chain
management in construction projects. In: 16th Ann. Conf. of the Int. Group for Lean Constr.
[online] United States: The Int. Group for Lean Constr., 381-388.
29
Appendix
Table A 1 Results of the SLR.
References Origin Focus / KPI Tools
(Akintoye 1995) UK JIT in construction/ KPI: - Materials management, production
planning
(Anglin and
Popescu 1995)
US JIT in construction
KPI: Quantity of waste
Waste management
(Tommelein
1997)
US Lean in construction (LC)
KPI: Waste quantity, time
saving
Lean techniques
(Opfer 1998) US JIT in construction/ KPI: - Materials standardization, advanced
procurement technology, material
supplier evaluation
(Tommelein and
Li 1999)
US JIT in construction/ KPI: - JIT deliveries
(Tommelein and
Weissenberger
1999)
US JIT in construction/ KPI: - JIT deliveries
(Pheng and Chuan
2001)
SG JIT implementation in
construction, LC, Prefabrication
KPI: Costs
JIT delivery, Lean techniques, on-
site storage, waste management
(Pheng 2001) SG JIT implementation in
construction, LC, Prefabrication
KPI: Costs
JIT delivery, Lean techniques, on-
site storage, waste management
(Low and Wu
2005)
SG JIT implementation in plant,
Prefabrication
KPI: Lead time, machine setup
time, quality
JIT vendor strategy, JIT production
strategy, quality control strategy
30
Table A1 Results of the SLR (Continued).
References Origin Focus / KPI Tools
(Horman and Thomas
2005)
US JIT in construction
KPI: Labour performance,
productivity
Different levels of inventories
(Christensen 2005) US JIT in construction/ KPI: - -
(Hamzeh et al. 2007) US LC/ KPI: Costs, time average
inventory
Different level of inventory
(Khalfan et al. 2008) UK LC/ KPI: Waste quantity, costs,
time saving
E-procurement system
(Zimmer et al. 2008) US LC/Construction supply chain
KPI: Observations
Lean techniques, JIT delivery,
material management, last planner
(Friblick et al. 2009) SE LC
KPI: Costs, time saving, delays
Lean techniques
(Ng et al. 2009) CN JIT implementation in
construction/ KPI: Project
duration, amount of buffer stock
JIT delivery, on site storage
management
(Amornsawadwatana
2011)
TH LC/ KPI: Number of loading
activities
JIT delivery
(Deshpande et al.
2012)
US LC/ KPI: Costs, efficiency,
waste quantity
5S
(Cossio and Cossio
2012)
MX JIT implementation in
construction, Prefabrication
KPI: Project duration, costs,
productivity, quality of
construction
JIT delivery, on site storage
management
(Tillmann et al. 2014) US LC/ KPI: Induction of desired
Lean behaviours
Lean techniques
31
Table A1 Results of the SLR (Continued and end).
References Origin Focus / KPI Tools
(Trivedi and Kumar
2014)
IN LC/ KPI: Quantity of waste,
costs, quantity of inventory
Pull system, continuous process flow,
standardization
(Viana et al. 2015) BR JIT implementation in
construction/ KPI: -
Production planning/JIT delivery
(Asri et al. 2016) MY JIT, LC, Prefabrication/ KPI: - -
32
Simulation of a prefabricated wooden building
using JIT and Lean strategies to reduce
construction duration
L’article intitulé « Simulation of a Prefabricated Wooden Building Using JIT and Lean
Strategies to Reduce Construction Duration » est inséré dans cette section du mémoire. Il a
été soumis le 08 décembre 2017 à la conférence « International Conference on Information
Systems, Logistics and Supply Chain (ILS) » et sera présenté à la 7e édition de cette
conférence le 08 juillet 2018 à Lyon, France. La version publiée est identique à la version
présentée dans ce mémoire.
33
Résumé
Traitant des délais serrés, de la gestion complexe des gaspillages sur les chantiers et des
attentes élevées des clients vis-à-vis de la qualité des projets livrés, l'industrie de la
construction recherche constamment des moyens d'améliorer sa productivité tout en
réduisant son empreinte écologique. Un modèle de simulation a donc été développé pour
examiner l'impact dans la construction de concepts à succès du domaine manufacturier, tels
que les livraisons juste-à-temps, les outils Lean et la préfabrication. Le modèle développé
est basé sur la construction d’un bâtiment en bois de six étages, érigé dans la province de
Québec, au Canada. Les résultats de la simulation montrent les avantages et les
inconvénients quantitatifs liés à la tenue d’un, de deux ou de plusieurs jours de stock
tampon sur le chantier, ainsi que la meilleure combinaison des concepts susmentionnés
pour obtenir une productivité plus élevée.
34
Abstract
Dealing with tight deadlines, complex waste management on construction sites, and clients’
high expectations over the quality of constructions delivered, the construction industry is
constantly searching for better ways to manage construction projects and improve
productivity while leaving less ecological footprint. To help with this goal, a simulation
model was developed to examine the impact of successful concepts from the manufacturing
industry, such as JIT deliveries, Lean tools, and prefabrication in the construction industry.
The model developed was based on a real six-storey wooden building erected in the
Province of Quebec, Canada. The results of the simulation show quantitative advantages
and disadvantages of having one, two, or more days of buffer stock on the construction site,
as well as the best combination of aforementioned concepts to obtain better productivity.
35
Introduction
Discussions have been held on the fact that productivity in the construction industry has
long lagged behind other industries [12], [2]. This situation originates from difficulties for
the construction industry to handle the logistics of material deliveries and on-site operations
management. To tackle this issue, concepts that have brought successes in the
manufacturing industry are explored for the construction sector. Nevertheless, the nature of
the process delivery of a construction project differs from the manufacturing field in terms
of operations and the planning of activities [4]. These differences suggest that productivity
improvement methods profitable for the manufacturing industry might require some
adaptation to be successful in construction. The goal of this research is to examine how
concepts from the manufacturing industry, such as just in time (JIT) deliveries, Lean
methods, and prefabrication can improve productivity in the construction industry. To do
so, a simulation model is developed using the Simio software to determine the impact of
the combination of different levels of i) prefabrication, ii) on-site buffer stock, and iii) Lean
on-site activities. The performance indicators monitored are construction duration,
shortages, and labour utilization. Results from the simulation highlighted advantages as
well as disadvantages of keeping one, two, or more days of buffer stock on the construction
site. Furthermore, by combining different best practices, it became possible to find the best
combination to obtain better productivity and reduce the construction duration. The
systematic literature review from [3] on JIT in construction showed that a few articles
discuss an implementation of JIT in construction or proposed quantitative results
concerning the impact of different levels of prefabrication as well as buffer stock and
construction site organisation on the outcome of a construction project. This paper therefore
contributes to scholars by showing modalities of applying concepts from manufacturing to
the construction industry. Even though they are applied to a specific prefabricated wooden
building, the concepts exploited remain general enough to be applied to other construction
projects. The paper also contributes to practitioners by helping determine the right amount
of buffer stock to keep on site and the extent to which they might implement Lean methods
and prefabrication on a construction site. The paper is structured as follows. Section 2
proposes a literature review and Section 3 summarizes the methodology and the model
36
developed. Section 4 describes the results obtained from the simulation model while
Section 5 concludes the paper.
Literature review
The construction industry in Canada employs more than 1.38 million workers, making it
the fifth-largest employer in the country [15]. In the Province of Quebec, it accounts for
more than 40 billion of investments each year [5], representing 12% of Quebec Gross
Domestic Product [6]. As an important economic player for the Quebec Province and for
the country, significant gains could be expected from improving its productivity. In their
research, Asri et al. [2] mentioned how prefabricated components offer tremendous
advantages in rapid erection of a construction. Nevertheless, these benefits of time saving
seem futile if logistics of material transportation and deliveries are not well managed. Thus,
JIT philosophy adapted for the construction sector becomes an interesting avenue to
consider. It may consist of producing smaller batches of each component and sending them
on site at the required installation time in order to reduce waste, diminish on-site storage
space, and meet the deadlines and high standards of the construction industry [7]. However,
if a truck arrives on the construction site and there are no space or resources available to
unload materials, the truck and its driver will be monopolized. This monopolization is
costly for either the supplier or the main contractor. To avoid such situations which
counteract gains made by prefabricating and JIT deliveries, Lean construction approaches
should also be adopted on the construction site. Despite the worthiness of JIT deliveries,
Pheng and Chuan [13] mentioned that applying perfect JIT deliveries in construction is a bit
idealistic due to the interdependence of stakeholders and the harsh and varied conditions on
construction sites. Therefore, they introduced the concept of a modified JIT in construction
to ensure a minimum level of buffer stock in order to cushion potential delays. They also
surveyed 32 main contractors and concluded that contractors preferred bigger quantities
(seven days) of buffer stock on site so that materials are in their area of control and idleness
on the construction site is avoided. Of the 32 contractors, 19% were in favor of 3-5 days of
buffer stock, 38% thought 1-2 days was acceptable and only 1 contractor applied 0-1 day.
Material inventories on site generate several costs such as storage costs, insurance costs,
costs related to obsolescence and monitoring against theft [1]. Increasing buffer quantities
37
can therefore create time and costs issues while Lean techniques still remain challenging to
implement on a construction site. The case study considered in this research concerns the
construction of a wooden building. Wood has been recognized as a sustainable material
presenting a low ecological footprint due to its high carbon retention [8]. It is therefore
understandable that constructors show a growing interest toward products made from wood.
In fact, according to Gosselin et al. [9], factors that motivate companies to opt for a wooden
construction are, in order of importance: durability, speed of erection, cost reduction,
visibility and lightness of wooden structures. Wood is particularly appropriate for
prefabrication since it is a light and easy-to-use material for which quality is better
controlled in mills. However, wood being sensitive to weather, keeping it in stock on
construction sites can be a source of cost increase in order to ensure its protection.
Methodology
Construction sites being dynamic systems with complex interrelationships and high doses
of uncertainty, simulation is an effective tool in determining the impact of different
parameters on output results. In this research, simulation is therefore used to model the
erection of a real wooden building and test the impact of applying JIT and Lean techniques
on the construction duration, labour utilization, and shortages. The simulation software
used is Simio, a 3D simulation modelling framework based on intelligent objects [11]. The
company partnering with the research team selected the construction project to simulate
based on the size of the project as well as availability and accuracy of data and
documentation. Data, plans, and documentation of the project were given to the research
team by the company. Techniques used to understand the dynamic and processes on the
construction site were visits and interviews with one of the company’s experts. Since on-
site observations and studies are important to building an effective simulation model, the
company allowed visits and measurements on site. The company and its team were
extremely collaborative during the data collection process, observations on the field, and
clarifications on the construction site routine. Many discussions and exchanges of
information were held between the research team and theirs throughout the project. Due to
the company’s confidentiality policies, all quantitative data and results are modified
38
proportionally with a mathematical factor. The real construction project was finished two
months prior to the beginning of the simulation.
Case study
The simulation model is based on the construction of a six-storey commercial timber
complex, constructed by one of the leaders in wood products and construction systems in
Canada. The structure of the building is composed of a post-and-beam system as well as
glulam decking, left visible under the floor system. Cross-laminated timber elements are
used for elevator shafts and stairways. The building has acquired the Leed (Leadership in
Energy and Environmental Design) certificate which is used widely and internationally for
rating building on metrics such as energy savings, carbon dioxide emissions, water use,
indoor environmental quality, use of resources, and attention to their impacts. A total of
156 cross-laminated timber (CLT), 1,836 decking and structure timber and 5,434 fittings
were used for the erection of the wooden structure. The installation of the structure is made
of seven divisions, with a team of six workers and one crane. Each division is composed of
nine sub-assemblies. The construction site is located in a spacious area, unlike the majority
of construction sites, allowing for a flexible storage space on site. Materials are delivered
on the construction site in 54 feet trucks. Deliveries arrive from a manufacturing plant
owned by the project partner. As mentioned by Pheng [12], the construction site then
becomes an assembly station where the factory-fabricated components are brought on site
to be assembled and installed. The manufacturing plant produces standard parts in advance
and then personalizes the required components once the building plans are confirmed. This
second portion of the fabrication process as well as all deliveries are based on the
construction schedule in a pull system.
Data collection
Since the research project started after the six-storey timber building was constructed,
qualitative data were collected through observations on the partner’s other on-going
construction sites. The selection criteria for construction sites observed were their similarity
to the one discussed in this paper in terms of size, cost of project, type of materials used,
and type of location. As for numerical data, such as the delivery schedule of materials, they
were collected by going through the project’s files and information given by the partner.
39
Then, all data were validated by one of the experts of the company. The construction
process starts at the plant. When the plant receives an order from the construction site, it
takes at least 11 hours, at most half a day, and on average 11.5 hours to get a full truck
ready to leave the plant with materials. Since the plant is in a remote area and the
construction site is on the outskirts of the city, it takes approximately 7.5 hours (at least 7
hours and at most 8 hours) to travel from the plant to the site. Once the truck arrives on site,
the unloading of materials takes approximately 7 min per pallet (at least 6 mins and at most
8 mins). Once materials are unloaded, they can be used to erect the building. According to
the nature of these data, triangular distributions are used (min, mode, max). CLT materials
are kept in the container left on site while decking, structure timber, and fittings are stored
on the ground. Depending on the size of materials picked in the buffer stock, the travel time
between the buffer stock and the exact location where materials are used takes between 20
and 40 min, leading to the use of a uniform distribution (min, max). Each distribution is
given its own random stream which is, for a distribution, a specific series of random
observations from that distribution. This ensures that no extra randomness adds additional
variance to the output performance measures. Total quantity of materials used on the
construction site were extracted from the project files and quantities of each material per
assembly deducted along with the reorder point of each division. The average reorder point
corresponds to 2.33 days of buffer as maintained on the actual construction site.
Current state simulation model
In order to focus on the logistics and dynamics of the construction site, it is assumed that
the plant providing the timber materials has a flexible production capacity. Since the
building is located in a wide area and the buffer stock is one of the varying parameters, the
buffer stock area has no capacity limitation. The simulation of the construction project
involves developing two main sub systems in the model: the replenishment sub system and
the execution of activities on the construction site sub system. In the first sub system, the
plant replenishes the buffer stock area, based on the reorder point of the division on which
workers are working on at the construction site. Like in Khalfan et al.’s [10] paper, the
company uses Kanban in the construction project to signal the need for materials delivery
on site. Each delivery from the plant to the construction site is a full load. In the second sub
system, the crane, driven by a worker, goes to the buffer stock area where another worker
40
awaits to put materials on it. Once the crane gets the materials, they are brought to the
location where they are needed on the construction site. The location where materials are
needed depends on the assembly and the division on which workers are working. The first
day of construction starts with materials delivery to ensure that human resources and
equipment such as the crane, which is rented expensively per hour, do not stay idle while
waiting for materials. In the simulation, state variables for each material are assigned the
appropriate initial inventory and received materials are sent to the buffer stock area to be
unloaded. Approximately 30 pallets fit into one truck. Currently, the total time required to
unload a truck is two and a half hours. The simulation thus uses a gradual process where
pallets are unloaded one at a time. When the first quantities are unloaded, the assigned
workers begin installation. The process illustrated in Figure 4 shows the mechanism on run
initialized.
Figure 4 Mechanism on run initialized.
On the simulated construction site, all workers start at 8am. Worker 1 stays at the buffer
stock area and prepares materials for the crane which is driven by the crane operator:
worker 2. Worker 3 is the site foreman. Once the crane has travelled to the desired location
of installation (several paths to choose from), worker 4 gets materials off the crane.
Workers 5 and 6 use aerial lifts to install the structure. Workers 3 and 4 also do other
necessary activities on site like nailing. The simulation then validates if the construction is
finished or if there are more divisions to install. If the quantity of materials required to
41
complete the installation is consumed, the construction is finished so the simulation run
ends. If there are more divisions to install, there is a validation to see if more materials
should be ordered from the plant. The validation considers the current inventory of
materials and those in transit. If it appears under the average reorder point or under the
quantity of a particular material needed to proceed at a desired assembly, a list of required
materials is made so as to order a full truck. If an order has been already made, that order’s
status is in transit. Then, all resources wait for the arrival of materials. Notice that this
waiting time engenders enormous waste of time and therefore, of money for the company.
Figure 5 shows a part of the process that launches the creation of orders on the construction
site. This process also manages the choice for paths the crane must follow to get to the right
location of installation. With nine assemblies per division, 54 different paths are created in
the model to go from the buffer stock area to the concerned area of assembly on site.
Figure 5 Creation of order and choice of path.
Figure 6 shows a part of the process that manages the decision to order a replenishment
from the plant and if so, to put the ordered quantity as being on order and in transit. Once
the truck leaves the plant, the material is considered as being only in transit.
42
Figure 6 Decision to replenish.
Model verification and validation
Model verification is defined as certifying “that the computer program of the computerized
model and its implementation are correct” [14]. Throughout the development process, the
addition of each new function was tested and verified before adding a new one. To verify
and validate the model, a decision-making approach “based on results of the various tests
and evaluations conducted” was used as described by Sargent [14]. Model validation is
described as the “substantiation that a model within its domain of applicability possesses a
satisfactory range of accuracy consistent with the intended application of the model” [14].
Constant involvement from the company’s expert ensured that the simulation model
developed was valid. Validation of the model also involved comparing the simulation
output values with the actual construction. For example, the actual project lasted 26 weeks
and the simulated construction has an average duration of 26.04 weeks after 10 runs, which
is acceptable given the variance of distributions. In fact, statistical tests like the hypothesis
test was done to test the ability of the model in predicting the real system given the same
conditions (input applied, policies, and strategies) but different independent random
observations. The hypothesis test went as follow. Let be the null hypothesis and the
alternative hypothesis. 26 weeks and 26 weeks. If is rejected,
43
then the model should be revised. Otherwise, there is no reason to think the model is not
valid. If then should be rejected and one should conclude that the model is
inadequate to predict the construction duration. The average construction duration for 10
replications is 26.04 weeks with a standard deviation of 0.274 weeks. The significance level
α=0.05 and the number of replications n=10. With the student distribution,
. The statistical value One
can notice that Then cannot be rejected and the model is
accepted as being valid.
Results
Simulation of various scenarios
Some scenarios being tested were proposed by the company’s expert while others were
chosen by the research team and then validated by the company’s expert. Since the goal of
the research was to evaluate how concepts such as JIT deliveries, Lean methods, and
prefabrication can improve productivity, three parameters were used to illustrate and
quantify the impact of these best practices on the construction project studied. The level of
prefabrication of materials is quantified by the time of installation of each sub assembly,
the use of Lean principles by the travel time between the buffer stock and the construction
site, and JIT deliveries by the number of days of buffer stock on the construction site. All
travel times go by a random uniform distribution while all installation times go by a
random triangular distribution. Each scenario represents a combination of aforementioned
parameters. Based on the construction site’s actual parameters that are equivalent to 2.33
days of buffer stock, 50% Lean and 80% prefabrication, it was possible to create scenarios
by making variations on these parameters, as shown in Table 1.
44
Table 1 Details concerning the variations of each parameter.
0% Lean 50% Lean 100% Lean 80% Prefabrication 100% Prefabrication
(20,40) min
travel time
(10,20) min
travel time
(6,12) min
travel time
(82.6,83.8,85) min
installation time
(42.6 ,43.8,45) min
installation time
-Materials stored
where there is
space on site
-No
classification
system to locate
the needed parts
-Workers need
to search for
equipment and
materials.
Workers take a
long time to find
parts
-Area defined
for storage of
materials
-No
classification
system to
locate the
needed parts
-Workers take
a while to find
parts
-Area defined
for storage of
materials,
equipment,
and work
tools
-Classification
system to
locate the
needed parts
-Workers find
parts quickly
-Prefabricated
materials in the plant
so that once on
construction site,
installation is done
directly (actual level
of prefabrication held
by the company on its
projects). Plumbing,
electricity, and outlet
cages cut on
construction site
-Prefabricated
materials in the plant
so that one on
construction site,
installation is done
directly. Plumbing,
electricity, and outlet
cages are already cut
at the plant
Table 2 presents the different scenarios tested. All combination levels of Lean and
prefabrication were tested for a buffer stock of 1, 2, 2.33, 3, and 4 days.
Table 2 Description of the different scenarios tested.
Scenarios JIT: buffer
stock (days)
Lean
level (%)
Prefabrication
level (%)
Lean:
travel time
(min)
Prefabrication:
installation time
(min)
S1 1 0 80 (20,40) (82.6,83.8,85)
S2 1 0 100 (20,40) (42.6 ,43.8,45)
S3 1 50 80 (10,20) (82.6,83.8,85)
S4 1 50 100 (10,20) (42.6 ,43.8,45)
S5 1 100 80 (6,12) (82.6,83.8,85)
S6 1 100 100 (6,12) (42.6 ,43.8,45)
S7 2 0 80 (20,40) (82.6,83.8,85)
S8 2 0 100 (20,40) (42.6 ,43.8,45)
S9 2 50 80 (10,20) (82.6,83.8,85)
S10 2 50 100 (10,20) (42.6 ,43.8,45)
S11 2 100 80 (6,12) (82.6,83.8,85)
S12 2 100 100 (6,12) (42.6 ,43.8,45)
S13 2.33 0 80 (20,40) (82.6,83.8,85)
S14 2.33 0 100 (20,40) (42.6 ,43.8,45)
45
Table 2 Description of the different scenarios tested (continued and end).
Scenarios JIT: buffer
stock (days)
Lean
level (%)
Prefabrication
level (%)
Lean:
travel time
(min)
Prefabrication:
installation time
(min)
S15 2.33 50 80 (10,20) (82.6,83.8,85)
S16 2.33 50 100 (10,20) (42.6 ,43.8,45)
S17 2.33 100 80 (6,12) (82.6,83.8,85)
S18 2.33 100 100 (6,12) (42.6 ,43.8,45)
S19 3 0 80 (20,40) (82.6,83.8,85)
S20 3 0 100 (20,40) (42.6 ,43.8,45)
S21 3 50 80 (10,20) (82.6,83.8,85)
S22 3 50 100 (10,20) (42.6 ,43.8,45)
S23 3 100 80 (6,12) (82.6,83.8,85)
S24 3 100 100 (6,12) (42.6 ,43.8,45)
S25 4 0 80 (20,40) (82.6,83.8,85)
S26 4 0 100 (20,40) (42.6 ,43.8,45)
S27 4 50 80 (10,20) (82.6,83.8,85)
S28 4 50 100 (10,20) (42.6 ,43.8,45)
S29 4 100 80 (6,12) (82.6,83.8,85)
S30 4 100 100 (6,12) (42.6 ,43.8,45)
Results per scenario
Table 3 presents the results for the different scenarios. The current state simulation model
(Scenario 15) reflects the reality of the six-storey timber building’s construction. A buffer
stock of 2.33 days is kept on the construction site throughout the construction, the travel
time between the buffer stock area and the location of the installation is
Random.Uniform(10,20) min, and the installation time Random.Triangular(82.6,83.8,85)
min on average. The simulation run ended at 26.04 weeks similar to the real project. The
average waiting time on site for the arrival of materials is 19.92h, including nights but
excluding week-ends. Twice in the simulation, resources were idle on site while waiting for
materials, leaving workers’ average utilization at 72.6%.
With the same parameters as the current state model, except for the buffer stock reduced to
one day, scenario 3 results in a longer construction time (28.04 weeks). This is due to the
fact that there is shortage of materials 5 times so resources have to wait for materials
delivery. Again, with equal parameters, but two days of buffer stock (S9), the construction
duration (26.1 weeks) stays relatively similar to the current state (26.04 weeks) with two
shortages in both cases. On the other hand, with three days of buffer stock (S21), there is no
46
shortage and the construction duration reduces to 25.72 weeks. It is interesting to notice
that workers’ scheduled utilization goes from 70.80% at 1-day buffer stock to 73.90% at 3-
day buffer stock and stays the same at four days (S27). Indeed, up to three days of buffer
stock, shortages diminish to 0, reducing the idle time of resources on site. At four days,
there is no shortage and the scheduled utilization stays the same because its maximum is
reached. The rest of the idle time happens when the crane does its movement from the
buffer stock area to the installation area. The scheduled utilization is calculated by the ratio
of the busy time on the available total time.
With the same level of prefabrication of materials made in the company’s plant (80%), the
best combination is obtained in scenario S23 where the travel time is reduced to a (6,12)
min distribution with three days of buffer stock. No shortages occur, the deadlines are met,
and the construction duration is reduced to 22.34 weeks while the utilization of workers,
which represents a considerable part of the project costs, is maximized at 82%. By
introducing more Lean methods to help reduce travel time by 4 minutes, one can save up to
23.66 days on the construction duration. One thing to notice, with the same parameters as
S23 but with more prefabrication at the plant (S24), the simulated construction finishes in
14 weeks but the scheduled utilization drops to 69.90% because it takes workers less time
and effort to install the same parts, leaving them idle while waiting for the next materials
delivery. Furthermore, in reality, this scenario requires a lot of time and energy upstream.
In fact, the challenge with having a higher level of prefabrication lies in the extreme
coordination needed in the design phase with all stakeholders: clients, structural and
electrical engineers, main contractors, and architects. Because of the huge energy required
to reach this level of prefabrication, the company prefers to explore other beneficial
scenarios.
47
Table 3 Results of different scenarios tested.
Scenarios
Construct.
duration
(weeks)
Workers'
utilization
(%)
Shortages Scenarios
Construct.
duration
(weeks)
Workers'
utilization
(%)
Shortages
S1 34.64 59.40 3 S16 16.60 59.80 0
S2 28.30 42.40 8 S17 22.58 81.40 1
S3 28.04 70.80 5 S18 14.38 70.70 0
S4 21.74 51.90 13 S19 34.00 60.10 1
S5 25.52 76.70 6 S20 24.38 45.70 0
S6 18.58 60.00 14 S21 25.72 73.90 0
S7 34.30 59.50 2 S22 16.60 59.80 0
S8 25.48 44.50 2 S23 22.34 82.00 0
S9 26.10 72.20 2 S24 14.06 69.90 1
S10 17.50 59.20 1 S25 33.76 60.50 0
S11 22.66 80.30 2 S26 24.38 45.70 0
S12 14.38 70.60 0 S27 25.72 73.90 0
S13 32.76 60.50 0 S28 16.60 59.80 0
S14 25.48 44.50 2 S29 22.34 82.00 0
S15 26.04 72.60 2 S30 14.00 70.60 0
Conclusion
This paper shows, through the use of a 3D simulation modelling framework, how concepts
from the manufacturing industry such as JIT deliveries, Lean methods, and prefabrication
can improve productivity in the construction industry. The simulation reveals that one to
two days of buffer stock with 100% Lean and prefabrication can significantly reduce (by
weeks) the project duration. However, these levels bring out disadvantages such as
shortages and less utilization of on-site workers since they are idle during shortages.
Idleness on site, however small, is something pricey every general contractor wants to
avoid. In the light of the simulation’s results, one can conclude that the reaction time of the
system does not allow these amounts of buffer stock to fully cover variations and hazards
encountered on construction sites. With current buffer stocks (2.33 days), one can obtain a
decent project duration and worker utilization but some risks of shortage depending on the
other parameters. Three days of buffer stock seems to be ideal for storage on site with the
same resources as in the current construction. Indeed, the company could improve its
productivity from 72.6% in the current state to 82% and reduce its construction duration
from 26.04 to 22.34 weeks by making slight changes such as adding 0.66 days of buffer
48
stock and reorganizing its storage activities (4 min less on each travel time). In further
scenarios, the company is interested in validating if doubling the number of cranes or
workers on the previously tested scenarios will increase productivity even more.
Acknowledgment
The authors are grateful to Natural Sciences and Engineering Research Council of Canada
for the financial support through its ICP and CRD programs (IRCPJ 461745-12 and RDCPJ
445200-12) as well as the industrial partners of the NSERC industrial chair on eco-
responsible wood construction (CIRCERB).
References
1. Akintoye, A.: Just-in-time Application and Implementation for Building Material
Management. Construction Management and Economics, 13, 105--113 (1995)
2. Asri, M. a. N. M., Nawi, M. N. M., Nadarajan, S.: Key Factors of Successful JIT
Integration with IBS - An Overview. In: International Conference on Applied Science
and Technology, 020065 (5 pp). ICAST Press, College Park (2016)
3. Bamana, F., Lehoux, N., Cloutier, C.: Just in Time in Construction: Description and
Implementation insights. In: 25th Conference of the International Group for Lean
Construction, IGLC25. IGLC25 Press, Heraklion (2017)
4. Bashir, A. M., Suresh, S., Oloke, D. A., Proverbs, D. G., Gameson, R.: The Application
of Lean Construction Tools in United Kingdom Construction Organisations: Findings
from a Qualitative Inquiry. In: 2013 Architectural Engineering National Conference, pp.
63--72. AEI Press, Reston (2013)
5. Commission de la construction du Québec, http://www.ccq.org/fr-
CA/A_QuiSommesNous
6. Commission de la construction du Québec, http://www.ccq.org/fr-
CA/A_QuiSommesNous/A05_IndustrieConstruction
7. Cossio, J. G., Cossio, A. G.: Application of Just in Time to the Fabrication and
Installation of Prefabricated Concrete Facades in Buildings. In: 20th Conference of the
International Group for Lean Construction, IGLC20. IGLC20 Press, San Diego (2012)
8. Forest Products Association of Canada,
http://www.fpac.ca/publications/Bio%20materials_FR.pdf
9. Gosselin, A., Lehoux, N., Cimon, Y., Blanchet, P.: Main Motivations and Barriers for
Using Wood as a Structural Building Material – A Case Study. In: 11e Congres
International de Genie Industriel – CIGI2015. CIGI2015 Press, Québec (2015)
10. Khalfan, M. M. A., Mcdermott, P., Oyegoke, A. S., Dickinson, M. T., Li, X.,
Neilson, D.: Application of Kanban in the UK Construction Industry by Public Sector
Clients. In: 16th Annual Conference of the International Group for Lean Construction,
IGLC16, pp. 347--358. IGLC16 Press, Salford (2008)
11. Pegden, C. D., Sturrock, D. T.: Introduction to Simio. In: 2011 Winter Simulation
Conference, WSC 2011, pp.29--30. Institute of Electrical and Electronics Engineers Inc.,
Phoenix (2011)
49
12. Pheng, L. S.: Just-in-time Management in Precast Concrete Construction: A Survey
of the Readiness of Main Contractors in Singapore. Integrated Manufacturing Systems,
12, 416--429 (2001)
13. Pheng, L. S., Chuan, C. J.: Just-in-time Management in Precast Concrete
Construction: A Survey of the Readiness of Main Contractors in Singapore. Journal of
Construction Engineering and Management, 127, 494--501(2001)
14. Sargent, R. G.: Verification and Validation of Simulation Models. Journal of
Simulation, 7, 12--24 (2013)
15. Statistique Canada, http://www.statcan.gc.ca/tables-tableaux/sum-
som/l01/cst01/econ40-eng.htm
50
Just-in-time and Lean in Construction: Simulation
of a Six-story Building to Improve Productivity
L’article intitulé « Just-in-time and Lean in Construction: Simulation of a Six-story
Building to Improve Productivity » est inséré dans cette section du mémoire. Il a été soumis
au journal « Construction Engineering and Management » le 29 mai 2018. La version
soumise est identique à la version présentée dans ce mémoire.
51
Résumé
L'industrie de la construction recherche constamment de meilleurs moyens de gérer ses
projets de construction et d'améliorer sa productivité tout en réduisant son empreinte
écologique. Un modèle de simulation développé dans un cadre de modélisation de
simulation 3D est ici proposé afin d’examiner l'impact dans la construction de concepts à
succès de l'industrie manufacturière, tels que les livraisons juste-à-temps, les outils Lean et
la préfabrication. Le modèle développé est basé sur la construction d’un bâtiment en bois
de six étages, érigé dans la province de Québec, au Canada. Grâce au design d’expérience,
les résultats de la simulation ont montré l'impact de chaque concept et de leurs interactions
sur la durée de la construction, les pénuries et le taux d'utilisation de la main-d'œuvre. Les
résultats ont également mis en évidence les avantages et inconvénients quantitatifs de la
mise en œuvre de 0%, 50% ou 100% de méthodes de gestion Lean, du maintien de 1 à 4
jours de stock tampon sur le chantier, de l’utilisation de 80% ou 100% de préfabrication des
matériaux, et de l’utilisation de 1 ou 2 équipes de construction.
52
Abstract
The construction industry is always searching for better ways to manage construction
projects and improve productivity while leaving less ecological footprint. To help with this
goal, a simulation model was developed in a 3D simulation modelling framework to
examine the impact of successful concepts from the manufacturing industry, such as just-
in-time (JIT) deliveries, Lean tools, and prefabrication in construction. The model
developed was based on a real six-story wooden building erected in the Province of
Quebec, Canada. Through a design of experiment, results of the simulation showed the
impact of each concept and their interactions on the construction duration, shortages, and
labor utilization. They also put into light the quantitative advantages and disadvantages of
implementing 0%, 50% or 100% Lean management methods, keeping between 1 and 4
days of buffer stock on the construction site, having 80% or 100% prefabrication of
components, and using 1 or 2 construction teams.
53
Introduction
Several authors address the fact that productivity in the construction industry has improved
significantly but is still lagging behind other industries (Asri et al. 2016; Pheng and Chuan
2001). Regularly, deficient management and organization has been pointed out as the main
reasons for the burden (Koskela 2000). A way to address this issue is to embrace the just-
in-time (JIT) philosophy which consists of providing the right materials, in the right
quantities and quality when it is needed in order to reduce waste and provide maximum
value (Tommelein and Li 1999; Tommelein and Weissenberger 1999; Vokurka and Davis
1996). In Canada, the construction industry employs more than 1.38 million workers,
making it the fifth-largest employer in the country (Statistics Canada 2018). In the Province
of Quebec, it accounts for more than 40 billion dollars of investments each year
(Commission de la construction du Québec 2018a) representing 12% of Quebec’s Gross
Domestic Product (Commission de la construction du Québec 2018b). As an important
worldwide economic player, significant gains could be expected from improving the
productivity of this industry. Indeed, under the widely reported success in the
manufacturing sector, the idea of a centralized and pull production conveyed in the Lean
thinking, as promoted by scientific management, has been broadly adopted as a model in
construction (Koskela 2000). Unlike in previous years (Koskela 2000), more contractors
caress the idea of profitability through an efficient management of site operations, material
management as well as thorough preliminary factory-based activities. Successful projects
reported in subsequent publications (Ballard and Reiser 2004; Ballard 2006; and Ballard
2008 cited by Ballard 2011) have given evidence confirming Lean in construction as
substantially improving performance of construction projects (Ballard 2011). However, the
systematic literature review from Bamana et al. (2017) on JIT in construction showed that
few articles depict the critical elements essential in coping with the performance of
construction projects or elaborate on the measurable results needed to convince and
encourage contractors to adopt the JIT philosophy.
Therefore, the goal of this research is to examine how the JIT philosophy could improve
productivity in the construction industry. More specifically, Lean methods could be used to
manage construction sites more efficiently, JIT deliveries could limit on-site inventory
54
(buffer stock), and prefabrication in a factory environment could reduce on-site
construction time. Hence, to provide measurable results, a simulation model was developed
to determine the impact of combining different levels of i) Lean methods, ii) on-site buffer
stock and iii) prefabrication. The simulation also compared the use of only a day shift with
the use of a day and a night shifts. The performance indicators monitored for each scenario
were construction duration, shortages, and labor utilization. Results from the simulation
highlighted advantages as well as disadvantages of implementing 0%, 50% or 100% Lean
management methods, keeping between 1 and 4 days of buffer stock on the construction
site, having 80% or 100% prefabrication of components, and using 1 or 2 construction
teams. Furthermore, by combining different levels of best practices, it became possible to
find the best combination that maximized performance (i.e. higher productivity but lower
shortages and construction duration). This paper therefore contributes to scholars by
showing modalities of application of concepts from manufacturing to the construction
industry. Even though they are applied to a specific prefabricated wooden building, the
concepts exploited remain general enough to be applied to other construction projects. The
paper also contributes to practitioners by helping determine the right amount of buffer stock
to keep on site and the extent to which they might implement Lean methods and
prefabrication on a construction site to reduce the construction duration.
The paper is structured as follows. The next section proposes a literature review followed
by a section on the methodology and the model developed. Then the results obtained from
the simulation model are described and concluding remarks are presented.
Literature review
Since its early promotion in the manufacturing industry, JIT has been described as a
method, a concept, a goal, a belief, a philosophy, a system, an approach, a process, a
strategy, a program, and a state of mind (Vokurka and Davis 1996). As mentioned by these
authors, for the last 120 years, the production system has evolved from a mass production
objective toward single personalized production and thus a vision of continuous
improvement and elimination of waste, which has gradually been accepted as a vision of
managerial practices. Therefore, JIT is no longer simply defined as a specific management
technique but a management philosophy. Mortimer states that JIT philosophy aims to make
55
businesses lean, more simple and effective to operate and with a higher degree of
integration (as cited by Pheng and Hui 2010).
Bamana et al. (2017) studied how JIT could help address issues in construction through a
systematic literature review (SLR). This SLR confirmed the need to adapt JIT philosophy
for an adequate deployment in this industry. It also confirmed the close ties between JIT,
Lean construction, and prefabrication for successful construction project management. In
their research, Asri et al. (2016) mentioned how industrialized building systems which use
prefabricated components offer tremendous advantages in rapid erection of a construction.
Moreover, Pheng (2001, 417) stated that prefabrication has come with “the hope of reaping
the benefits of factory-styled operations” since it entails benefits such as improved quality,
waste reduction, and faster erection of buildings. Nevertheless, benefits of time saving from
rapid erection seem futile if logistics of material transportation and deliveries are not well
managed. Thus, JIT philosophy adapted for the construction sector becomes an interesting
avenue to consider. It may consist of producing smaller batches of each component and
sending them on site at the required installation time in order to reduce waste, diminish on-
site storage space, and meet the deadlines and high standards of the construction industry
(Cossio and Cossio 2012). However, if a truck arrives on the construction site and there are
no space or resources available to unload materials, the truck and its driver will be
monopolized. This monopolization is costly for either the supplier or the main contractor.
To avoid such situations which counteract gains of time made by prefabricating and JIT
deliveries, Lean construction tools and thinking might also be adopted on construction site
activities (Friblick et al. 2009). Lean tools such as 5S (Deshpande et al. 2012) are used to
ensure that the construction site is well organized to improve performance of workers and
optimal movements of materials on site.
The main element in JIT is stockless production which means reducing inventories both in
raw materials and work-in-progress products (Pheng and Hui 2010). Tommelein and
Weissenberger (1999) and Pheng and Chuan (2001) maintained that in practice a buffer is
necessary and its size should be determined strategically. Moreover, Pheng and Chuan
(2001) mentioned that applying perfect JIT deliveries in construction is a bit idealistic due
to the interdependence of stakeholders and the harsh and varied conditions on construction
sites. Therefore, they introduced the concept of a modified JIT in construction to ensure a
56
minimum level of buffer stock in order to cushion potential delays. They also surveyed 32
main contractors and concluded that contractors preferred bigger quantities (seven days) of
buffer stock on site so that materials are in their area of control and idleness on the
construction site is avoided. Of the 32 contractors, 19% were in favor of 3-5 days of buffer
stock and 38% agreed to 2 days or less. Material inventories on site generate several costs
such as storage costs, insurance costs, costs related to obsolescence, and monitoring against
theft (Akintoye 1995). Increasing buffer quantities can therefore create time and cost issues
while Lean techniques still remain challenging to implement on a construction site. Another
aspect to consider in order to get benefits from JIT is to pay particular attention to the
implication of all stakeholders of the supply chain. In general, each of the stakeholders
work without taking into consideration the effects of their actions downstream of the
process (Zimmer et al. 2008). Activities of each stakeholder gain to be synchronized
otherwise misunderstandings and bad synchronizations are an expensive price to pay in
construction.
A construction site being different from the manufacturing environment in terms of the
duration, the approach, the frequency of occurrence of unforeseen complications, the
dependency between stakeholders, and the nature of the tasks being executed especially on
the construction site, researchers have developed a set of Lean tools (Bashir et al. 2013)
suitable for the construction environment. The tools comprise the root-causes diagrams, the
5 whys, the 5S, the last planner, standardization of teams, Kaizen, Kanban cards, increased
visualization, poka-yokes, visual management with the Pareto chart, and daily huddle
meetings and employee empowerment (Salem et al. 2006; Bashir et al. 2013). When
applied, their benefits include a reduction and even elimination of all types of waste and an
increase in the output quality of the construction.
The case study considered in this research concerns the construction of a medium-sized
wooden building. Wood has been recognized as a sustainable material presenting a low
ecological footprint due to its high carbon retention (Forest Products Association of Canada
2018). It is therefore understandable that constructors show a growing interest toward
products made from wood. In fact, according to Gosselin et al. (2015), factors that motivate
companies to opt for a wooden construction are, in order of importance: durability, speed of
erection, cost reduction, visibility and lightness of wooden structures. Wood is particularly
57
appropriate for prefabrication since it is a light and easy-to-use material for which quality is
better controlled in mills. However, wood being sensitive to weather, keeping it in stock on
construction sites can be a source of cost increase to ensure its protection.
The literature helped to identify JIT elements applicable to construction as well as the key
performance indicators used to measure the impact of their implementation. However, a
need to broaden the scientific knowledge on the implementation of these concepts in
construction was observed in order to quantify the influence of each of these concepts,
separately or combined, on the productivity of a real construction site. Therefore, the
following section presents the methodology used to investigate the implementation of JIT
for the construction of a six-story wooden building.
Research methodology
Construction sites being dynamic systems with complex interrelationships and high doses
of uncertainty, simulation is an effective tool in determining the impact of different
parameters on output results. As defined by Banks (1999), simulation involves the
generation of an artificial history of the system, and the observation of that artificial history
to draw inferences concerning the operating characteristics of the real system that is
represented. In this research, simulation was therefore used to measure the effect of using
prefabricated components and applying JIT and Lean techniques on the construction
duration, labor utilization, and shortages of a six-story commercial timber complex. The
simulation software used to model the construction of the building was Simio, a 3D
simulation modelling framework based on intelligent objects (Pegden and Sturrock 2011).
The construction project simulated was constructed by one of the leaders in sustainable
wood solutions and construction systems in Canada.
To fully understand the activities on the construction site and create an effective simulation
model, visits and measurements on site as well as interviews with the company’s experts
were conducted. The company shared the data and documentation related to the
construction and was very collaborative with the research team. Due to the company’s
confidentiality policies, all quantitative data and results presented in the article were
modified proportionally with a mathematical factor. The real construction project was
finished two months prior to the beginning of the simulation.
58
A pre-analysis was first conducted to identify the productivity issues in this construction
project. This helped validate the choice of parameters to focus on for the simulation and the
selection of scenarios to investigate. The next step involved data collection so as to develop
a valid and reliable simulation model. The actual way that the construction project was
carried out was therefore reproduced with the simulation model (current state) to ensure the
accuracy of the model through a verification and validation exercise. Then, the selected
scenarios were simulated to test the impact of JIT deliveries, Lean methods, and
prefabrication on construction productivity. A design of experiment (DOE) approach was
used to understand the effect of applying these concepts separately or simultaneously on
construction productivity. Finally, an implementation framework was developed to guide
JIT implementation in construction. Figure 7 presents the main steps of the analysis.
Figure 7 Main steps of the analysis.
Pre-analysis
A pre-analysis with an Ishikawa diagram put into perspective the root causes of low
productivity in the construction project under investigation, allowing to identify the
parameters of focus in order to improve productivity. Parameters such as weather, depend
59
more on hazards, others such as the location of the construction site, depend primarily on
the client’s will, and others like material delivery, depend more on the logistics behind the
construction project. Indeed, based on the pre-analysis and the information available from
the case study, it was possible to target parameters and combinations of scenarios to be
tested with the simulation. Figure 8 summarizes all these parameters of influence on
productivity.
Figure 8 Root causes of low productivity.
Data collection
To develop a model that would reflect the real construction, data collection was thoroughly
undertaken by using documentation as well as on site observations. A total of 156 cross-
laminated timber (CLT), 1,836 decking and structure timber, and 5,434 fittings were used
for the erection of the wooden structure. The installation of the structure was made of seven
divisions, with a team of six workers and one crane. Each division is composed of nine sub-
assemblies. Figure 9 illustrates the building in terms of its seven divisions.
60
Figure 9 The building in terms of its seven divisions.
The construction site was located in a spacious area, allowing for a flexible storage space
on site. Materials were delivered on the construction site in 54 feet trucks. Deliveries
arrived from a manufacturing plant owned by the project partner. As mentioned by Pheng
(2001), the construction site then becomes an assembly station where the factory-fabricated
components are brought on site to be assembled and installed. The manufacturing plant
produced standard parts in advance and then personalized the required components once the
building plans were confirmed. This second portion of the fabrication process as well as all
deliveries were based on the construction schedule in a pull system. Since the research
project started after the six-story timber building was constructed, qualitative data were
collected through observations on the industrial partner’s other on-going construction sites.
The selection criteria for construction sites observed were their similarity to the one
discussed in this paper in terms of size, cost of project, type of materials used, type of
location as well as accuracy of data and documentation. As for numerical data, such as the
delivery schedule of materials, they were collected by going through the project’s files and
information given by the industrial partner. Then, all data were validated by the experts of
the company.
The construction process starts at the plant. When the plant received an order from the
construction site, it took on average 11.5 hours (+/- 0.5 hrs) to get a full truck ready to leave
61
the plant with materials. It took approximately 7.5 hours (+/- 0.5 hrs) to travel from the
plant to the construction site. Whenever the truck arrived on site, the unloading of materials
took about 7 minutes per pallet (+/- 1 min). Once materials were unloaded, they could be
used to erect the building. CLT materials were kept in the container left on site while
decking, structure timber, and fittings were stored on the ground. Depending on the size of
materials, the travel time between the buffer stock area and the exact location where
materials were consumed, took between 20 and 40 minutes. The average reorder point
corresponded to 2.33 days of buffer as maintained on the actual construction site.
Current state simulation model
In order to focus on the logistics and dynamics of the construction site, it was assumed that
the plant providing the timber materials had a flexible production capacity. Since the
construction site was located in a wide area and that buffer stock was one of the varying
parameters, the buffer stock area in the simulation had no capacity limitation. Figure 10
shows the different elements intervening on the construction site activities.
Figure 10 Different elements intervening on the construction site activities.
62
At the beginning of the simulation, the on run initialized process is created to have
materials on site when the construction starts on the first day. The process illustrated in
Figure 11 shows the mechanism on run initialized.
Figure 11 Mechanism on run initialized.
The simulation of the construction project involves developing two main sub systems in the
model: the replenishment sub system and the execution of activities on the construction site
sub system. Each sub system requires different processes to be functional. In the first sub
system, the plant replenishes the buffer stock area, based on the reorder point of the
division on which workers are working on at the construction site. The first day of
construction starts with materials delivery to ensure that human resources and equipment
such as the crane, which is rented expensively per hour, do not stay idle while waiting for
materials. In the simulation, state variables for each material are assigned the appropriate
initial inventory and received materials are sent to the buffer stock area to be unloaded. As
in Khalfan et al. (2008), the company uses Kanban in the construction project to signal the
need for materials delivery on site. Inventory replenishment is triggered by either being
below the average reorder point (regular replenishment) or below the quantity needed to
proceed construction at a desired sub-assembly (irregular replenishment). A list of required
materials is made so as to order a full truck from the plant. Approximately 30 pallets fit into
one truck. Currently, the total time required to unload a truck is three and a half hours. The
simulation thus uses a gradual process where pallets are unloaded one at a time. Figure 12
shows a part of the process that manages regular replenishment decisions.
63
Figure 12 Regular replenishment decision process.
In the second sub section for the execution of activities, all workers start at 8am, as on the
real construction site: the crane operator, the skytracker, the site foreman, the worker who
awaits to put materials on the crane, and the other two workers who use aerial lifts to install
the structure. The exact location where materials are needed depends on the sub-assembly
and the division on which workers are working. Hence, this process also manages the
choice for paths that the skytracker must follow to get to the right location of installation.
With nine sub-assemblies per division, 54 different paths are created in the model to go
from the buffer stock area to the concerned area of assembly on site. Depending on the size
of materials, the travel time between the buffer stock area and the exact location where
materials are used, takes between 20 and 40 minutes, leading to the use of a uniform
distribution (min, max). Each distribution is given its own random stream which is, for a
distribution, a specific series of random observations from that distribution. This ensures
that no extra randomness adds additional variance to the output performance measures.
Figure 13 shows a part of the process that launches the creation of orders on the
construction site.
64
Figure 13 Creation of order and choice of path.
Model verification and validation
Model verification is defined as certifying “that the computer program of the computerized
model and its implementation are correct” (Sargent 2013, 14). Throughout the development
process, the addition of each new function was tested and verified before adding a new one.
To verify and validate the model, a decision-making approach “based on results of the
various tests and evaluations conducted” was used as described by Sargent (2013,12).
Sargent (2013,12) also defined model validation as the “substantiation that a model within
its domain of applicability possesses a satisfactory range of accuracy consistent with the
intended application of the model.” Constant involvement from the company’s expert
ensured that the simulation model developed was valid. Validation of the model also
involved comparing the simulation output values with the actual construction. For example,
the actual project lasted 26 weeks and the simulated construction had an average duration
of 25.92 weeks after 10 runs, which was acceptable given the variance of distributions. In
fact, statistical tests like the hypothesis test were done to test the ability of the model in
predicting the real system given the same conditions (input applied, policies, and strategies)
but different independent random observations. The hypothesis test went as follows. Let
be the null hypothesis and the alternative hypothesis. 26 weeks and
26 weeks. If is rejected, then the model should be revised. Otherwise, there
is no reason to think that the model is not valid. If then should be rejected and
65
one should conclude that the model is inadequate to predict the construction duration. The
average construction duration for 10 replications was 25.92 weeks with a standard
deviation of 0.30 weeks. The significance level α=0.05 and the number of replications
n=10. With the student distribution, . The statistical value
. One can notice that
Hence could not be rejected and the model was accepted as being valid. With 10
replications, one could confirm with 95% confidence that the real value of construction
duration was between 25.70 and 26.13 weeks while with 20 replications, one could confirm
with 95% confidence that the real value of construction duration was between 25.78 and
26.06 weeks.
Results
Simulation of various scenarios
Since the goal of the research was to evaluate how concepts such as JIT deliveries, Lean
methods, and prefabrication could improve productivity, three parameters were used to
illustrate and quantify the impact of these best practices on the construction project studied.
The level of prefabrication for materials was quantified at either 80% or 100%, based on
the time of installation of each sub-assembly. The use of Lean methods was estimated at
0%, 50%, or 100% based on the travel time needed between the buffer stock area and the
assembly site. The JIT deliveries were quantified by the number of days of buffer stock
kept on the construction site. The values evaluated were 1, 2, 2.33, 3, or 4 days of buffer
stock. All travel times were determined by a random uniform distribution while all
installation times by a random triangular distribution. Each scenario represented a different
combination of aforementioned parameters for a total of 30 scenarios for all possible
combinations with only 1 work shift and 30 other scenarios for these same combinations
with 2 work shifts. Table A2, presented in the appendix, summarizes all the different
scenarios tested for one and two work shifts along with their parameters, for a total of 60
scenarios. Based on the construction site’s actual parameters that were equivalent to 2.33
days of buffer stock, 50% Lean and 80% prefabrication, and represented by scenario 0, it
66
was possible to create scenarios by making variations on these parameters, as shown in
Table 4.
Table 4 Details of the variations of each parameter.
0% Lean 50% Lean 100% Lean 80%
Prefabrication
100%
Prefabrication
(10,20) min (10,20) min (6,12) min (82.6,83.8,85)
min
(42.6 ,43.8,45)
min - Materials stored
where there is space
on site;
- No classification
system to locate the
needed parts;
- Workers need to
search for equipment
and materials;
- Workers take a long
time to find parts.
- Area defined
for storage of
materials;
- No
classification
system to locate
the needed
parts;
- Workers take a
while to find
parts.
- Area defined
for storage of
materials,
equipment, and
work tools;
- Classification
system to locate
the needed
parts;
- Workers find
parts quickly.
-Materials
prefabricated at
the plant so that
on construction
site, installation is
done directly.
Plumbing,
electricity, and
outlet cages cut
on construction
site.
- Materials
prefabricated at
the plant so that
on construction
site, installation
is done directly.
Plumbing,
electricity, and
outlet cages are
already cut at
the plant.
Results per Scenario
Table 5 presents the results of the different scenarios tested in the simulation software for
one work shift. As seen in the table, the current state simulation model (S0) reflects the
reality of the six-story timber building’s construction. A buffer stock of 2.33 days is kept on
the construction site throughout the construction, the travel time between the buffer stock
area and the location of the installation is Random.Uniform(10,20) min, and the installation
time Random.Triangular(82.6,83.8,85) min on average. The simulation run ended at 26.09
weeks similar to the real project. The average waiting time on site for the arrival of
materials was 19.92h, including nights but excluding weekends. Twice in the simulation,
resources were idle on site while waiting for materials, leaving average labor utilization at
68.9%. Table A2 in the appendix presents the defining characteristics for all 60 scenarios as
well as the experimental results for each one.
67
Table 5 Results for the different scenarios for one shift.
Scenarios Key performance indicators
Construct. duration (weeks) Labor utilization (%) Shortages
S0 26.09 68.9% 2
S1a 33.73 54.9% 3
S2a 27.74 36.3% 11
S3a 28.33 69.3% 5
S4a 21.45 51.5% 10
S5a 25.76 76.0% 6
S6a 19.51 57.0% 14
S7a 33.72 54.2% 2
S8a 26.36 35.8% 2
S9a 26.09 68.9% 2
S10a 17.51 55.0% 0
S11a 22.67 78.5% 2
S12a 14.39 67.7% 0
S13a 33.72 54.2% 0
S14a 26.08 36.3% 0
S15a 17.47 54.9% 0
S16a 22.29 80.9% 0
S17a 14.39 67.8% 0
S18a 32.35 55.4% 0
S19a 26.08 36.3% 0
S20a 24.59 72.7% 0
S21a 17.49 54.8% 0
S22a 22.31 80.9% 0
S23a 14.10 67.4% 0
S24a 32.06 56.1% 0
S25a 26.08 36.3% 0
S26a 24.59 72.7% 0
S27a 17.47 99.8% 0
S28a 22.31 80.5% 0
S29a 14.04 67.6% 0
68
To better analyze and interpret the results, graphs aggregating all scenarios with one work
shift were designed in terms of each performance indicator vs the number of days of buffer
stock (Figures 14, 15, and 16). Each line in the graph represents constant Lean and
prefabrication levels. In Figure 14, the duration of the construction project can then be
given as a function of the number of days of buffer stock. For example, Figure 14 shows
that the lowest duration was 14.04 weeks obtained using 4 days of buffer stock with 100%
Lean and prefabrication levels. Figure 14 also shows that construction duration for 1 day of
buffer stock with 50% Lean and 80% prefabrication (scenario S3a) had a construction time
of 28.33 weeks. These are the same Lean and prefabrication levels as the current state
model but with a buffer stock reduced to one day which resulted in a longer construction
time. This is due to the fact that there was shortage of materials 5 times during the
simulation, so resources had to wait for delivery of materials. Again, with the same Lean
and prefabrication levels, but two days of buffer stock (S9a), the construction duration
(26.09 weeks) was similar to the current state (26.09 weeks) with two shortages in both
cases.
Figure 14 Graph of construction duration vs stock for different scenarios with one work shift.
As for Figure 15 below, each line still represents a constant level of Lean and
prefabrication, however the graph focuses on shortages as a function of the number of days
of buffer stock. Figure 15 clearly illustrates that with 3 and 4 days of buffer stock, all
combinations of Lean and prefabrication levels result in zero shortages.
69
Figure 15 Graph of shortages vs. stock for different scenarios with one work shift.
Figure 16 below uses the same approach for the third indicator, labor utilization. It is
interesting to notice that with current state, labor utilization varied from 69.3% at 1-day
buffer stock to 72.7% at 3- and 4-day buffer stock. Indeed, up to three days of buffer stock,
shortages diminished to 0, reducing the idle time of resources on site. At four days, there
was still no shortage and the scheduled utilization stayed the same because its maximum
has been reached. The rest of the idle time occurred when the skytracker travelled from the
buffer stock area to the installation area. The scheduled utilization was calculated using the
ratio of occupied time on the total available time.
Figure 16 Graph of labor utilization vs stock for different scenarios with one work shift.
With the same level of prefabrication of materials observed in the current state (80%), the
best combination is obtained in scenario S22a with 3 days of buffer stock and 100% Lean,
70
where the travel time is reduced to a (6,12) minute uniform distribution. No shortages
occurred, the deadlines were met, and the construction duration was reduced to 22.3 weeks
while labor utilization, which represents a considerable part of the project costs, was
maximized at 80.9%. By introducing more Lean methods to help reduce travel time by 4 to
8 minutes, 26.46 days could be saved on the construction duration. Furthermore, by
increasing prefabrication to 100% with the same levels of Lean and buffer stock, the
simulated construction finished in 14.1 weeks but the scheduled utilization dropped to
67.4% because it took workers less time and effort to install the same parts, leaving them
idle while waiting for the next materials delivery. However, in reality, 100% prefabrication
requires a lot of time and energy upstream. In fact, the challenge with having a higher level
of prefabrication lies in the extreme coordination needed in the design phase with all
stakeholders: clients, structural and electrical engineers, main contractors, and architects.
Because of the huge energy required to reach this level of prefabrication, the company
preferred to explore other beneficial scenarios.
The same thirty scenarios were also tested with 2 work shifts per day as presented in table
A1 in the appendix (scenarios S1b to S30b). The second team continued the installation on
a night shift when the morning team finished at 5pm. In these scenarios, the construction
duration was drastically reduced while shortages tended to increase. An illustration of these
findings was reflected in scenario S17b, with parameters at 100% Lean, 80%
prefabrication, and 2.33 days of buffer stock. In this scenario, the construction finished in
15.5 weeks, 6.8 weeks earlier than with only one work shift. However, shortages were
higher, 7 instead of 0, and tended to reduce when stock increased. The labor utilization also
decreased with 80.9% with one morning shift compared to 70.5% with one morning shift
and one night shift. This drop in labor utilization for scenarios with two working teams
makes these scenarios less interesting for implementation.
As for the utilization rate of the crane, it was between 95.2% and 99.8% in all scenarios.
The crane was the bottleneck in the system. Adding a second team on the night shift
allowed labor utilization rate to go up but then dropped drastically when the crane reached
its maximum utilization rate. Therefore, doubling the number of cranes was proposed,
however the company decided against this option as the pace would be too fast for the plant
and the different stakeholders.
71
Design of Experiments
In the quest for defining the influence of manufacturing concepts such as JIT, Lean and
prefabrication on construction sites, design of experiments (DOE) was perfectly suitable to
systematically conduct the investigation. More specifically, DOE helped determine exactly
how each of these concepts or their combinations influenced productivity, which was
quantified in the simulation model as construction duration, labor utilization, and shortages.
A factorial approach to DOE was undertaken in Minitab, a powerful statistical program
used to analyze the data, uncover flaws in processes, and improve them. With the factorial
approach, multiple variables were tested simultaneously and common cause variation was
used to determine which factors were important through replication of trials. As the testing
parameters (low and high) for the JIT factor are 1 and 4 days; Lean factor 0 and 100%; and
prefabrication factor 80 and 100%, the number of levels is 2 and the number of factors 3. A
full factorial which involves all (8) possible combinations: 3 factors, each at 2 levels for a
total of 8 combinations. Results of an additional replication were used in the experiment to
measure variation among runs performed under the same experimental conditions and
analyze the mean and the standard deviation of responses. Table 6 shows all the different
combinations for the DOE.
Table 6 All the different combinations for the DOE.
Replications Scenario JIT Lean
(%)
Prefabrication
(%)
Duration
(weeks)
Labor
use (%)
Shortages
1 1 1 0 80 33.73 54.90 3
1 2 1 0 100 27.74 36.30 11
1 3 1 100 80 22.29 80.90 6
1 4 1 100 100 14.39 70.01 13
1 5 4 0 80 32.06 56.10 0
1 6 4 0 100 26.08 36.31 0
1 7 4 100 80 22.29 80.90 0
1 8 4 100 100 14.04 67.60 0
2 1 1 0 80 33.72 54.20 2
2 2 1 0 100 23.72 40.38 2
2 3 1 100 80 25.76 76.00 6
2 4 1 100 100 19.51 57.00 14
2 5 4 0 80 33.72 54.20 0
2 6 4 0 100 26.08 36.30 0
2 7 4 100 80 22.30 80.91 0
2 8 4 100 100 14.39 67.70 0
72
Identifying Significant Effects and their Impacts on Each Indicator
With the Pareto diagram and the normal plot, the magnitude and importance of each factor
were determined. This statistic helped determine the most significant parameter, between
Lean, prefabrication, or JIT, on productivity. The Pareto chart displays the absolute value
of the effects and draws a reference line for the effect. On the Pareto diagram shown in
Figure 17, any effect that extends past the reference line is potentially significant. One can
notice on the Pareto diagram that Lean (factor B) had the most significant effect on the
construction duration and that prefabrication (factor C) was second. This statistic revealed
Lean to be the vital base of improvement of productivity in construction.
Figure 17 Pareto diagram of normalized effects on construction duration.
Another way to determine a significant effect is through the P value of the effect. When the
P value of an effect is less than the significance level alpha = 0.05, the effect is significant.
Table 7 presents the results of the P value analysis for each effect on construction duration,
labor utilization, and shortages. Results show that Lean and prefabrication had the most
significant effect on project duration and labor utilization, whereas JIT had the most
significant effect on shortages, followed by prefabrication and the combination
JIT*prefabrication.
73
Table 7 Results of the P value analysis for each effect on construction duration, labor utilization, and shortages.
Responses/Factors Construction duration Labor utilization Shortages
P value: JIT 0.228 0.500 0.00
P value: Lean 0.00 0.00 0.05
P value: Prefabrication 0.00 0.00 0.036
P value: JIT*Lean 0.322 0.304 0.050
P value: JIT*Prefabrication 0.962 0.900 0.036
P value: Lean*Prefabrication 0.928 0.376 0.464
P value: JIT*Lean*Prefabrication 0.579 0.571 0.464
Most Important Effects
The graph in figure 18 illustrates how each factor affects the construction duration
(positively, negatively, or neutrally). As shown, Lean made the construction duration
decrease the most, followed closely by prefabrication.
Figure 18 Graph of main effects on construction duration.
This same statistical test for labor utilization rate and shortages showed that Lean made the
labor utilization rate increase while prefabrication made it decrease, and that more stock
tended to reduce the number of shortages.
Significant Interactions between Factors
Looking at interactions between factors helped in understanding if implementing two or
three changes at the same time such as JIT and Lean or JIT, Lean, and prefabrication had an
impact on the output response: here the construction duration. As seen in Figure 19, the JIT
74
and Lean combination appeared to have an influence, even though small, in reducing the
construction duration since the slope had changed. Because the slopes for JIT and
prefabrication as well as Lean and prefabrication remained the same, their combination did
not seem to help in reducing construction duration.
Figure 19 Graph of interactions of effects for construction duration.
As for labor utilization rate, the JIT and Lean combination also have a positive impact. The
analysis for shortages shows that the combination of JIT*Lean, JIT*prefabrication, as well
as Lean*prefabrication all have a positive influence on shortages.
Best scenario for the case study
Based on the objective of the company in this case study which was to reduce costs
associated with the installation and its realities with prefabrication, the best scenario
appeared to be S22a. Indeed, this scenario combining one work shift, 80% prefabrication,
100% Lean and 3 days of buffer stock presented the best compromise for the company as it
helped in highly reducing the construction duration, increasing labor utilization and
eliminating the risks of shortages, hence reducing costs related to installation. Table 8
shows the results of the current state and the best scenario for this case study.
75
Table 8 Results of the current state vs. the best scenario for this case study.
Scenario Current state
One work team, 80% prefabrication,
50% Lean, 2.33 days of buffer stock
Best scenario
One work team, 80%
prefabrication, 100% Lean, 3
days of buffer stock
Construction duration (weeks) 26.09 weeks 22.31 weeks
Labor utilization rate (%) 68.9% 80.9%
Shortages 2 0
Therefore, it was recommended that the company increase its buffer stock on the
construction site to three days, keep its same level of prefabrication at 80%, while
maximizing the Lean on site by reducing the travel time from the buffer stock area to the
construction site by 4 to 8 min. Focusing on applying Lean on the construction site revealed
itself to be extremely beneficial in terms of considerable time and resource savings in the
installation process, the most value adding activity for the client. As the buffer stock is
strategically managed, meeting deadlines and maintaining good labor utilization reside in
integrating Lean in construction site operations while using prefabricated materials at the
80% level which maintains the best time/benefits ratio.
Implementation framework
Some particularities in the construction sector make the application of JIT more challenging
and requires some adaptations before its implementation. In fact, the objective of JIT is that
all activities performed on a product must create value for the client while construction
projects create value more in the long term. In the manufacturing sector, a standardization
of production activities is adopted to reduce differentiation activities while in the
construction industry, each project is unique, making it difficult to standardize (Behera et
al. 2015). Furthermore, in the manufacturing industry, production does not depend on the
place of consumption of the product. The fabrication, the assembly, and the packing of the
product can be done in different countries, and each of these steps adds value to the
product. In the construction industry, only a small part of the project is done off-site, and
this, because of the constraints present on the construction site. Both off-site and on-site
76
activities are important, however on-site installation and erection activities add the greatest
value to the product (Salem et al. 2006).
Due to the specific characteristics of the construction industry, the benefits of JIT for
construction projects depend on the overall value chain, which includes the plant supplying
materials, the construction site, and the physical as well as virtual flows between them. As
pointed out by Pheng (2001) and depicted in Viana et al. (2015), one party trying to adopt
the JIT philosophy while the other does not, will realize its efforts are futile and will not
achieve the potential benefits. Based on the findings in the literature on the JIT elements
required for its successful implementation in construction and the validation on its
application modalities through the results obtained from the simulation, an implementation
framework with four scenarios of JIT in construction may be proposed as illustrated in
Figure 20.
Figure 20 Implementation framework.
Scenario I shows that low information sharing, low coordination between the plant and the
site and poor management of materials are respectively less favorable to (-) prefabrication
of components and (-) JIT deliveries. Low construction site management and poor use of
Lean tools suggest (-) low presence of Lean principles on the site. In scenarios II and III,
still one or two elements are less present, making it difficult to obtain gains on the overall
77
value chain. Scenario IV illustrates the best scenario for JIT implementation in construction
where information sharing, supply chain coordination and management of materials are
respectively more favorable to (+) prefabrication of components and (+) JIT deliveries
while high construction site management suggests (+) the presence of Lean principles on
the site.
As observed, to implement a successful JIT in construction, the construction site should be
well organized for reception of the materials ordered and, once on site, these materials
should be easily accessible for usage. Therefore, at the start of the project, it is important to
establish good communication between the stakeholders. Moreover, determining the right
amount of materials to order for the construction site and the ordering frequency are highly
significant for JIT deliveries. Once on the construction site, to avoid waiting for availability
of equipment and/or wasting time by searching for the right materials, Lean construction
tools such as 5S should be used to define storage location for parts and a system to localize
them. The expected outcomes for the best scenario are a reduction of costs, waste of time,
and waste of materials, as well as an increase of labor utilization, and ultimately, client
satisfaction.
Conclusion
This paper shows, through the use of a 3D simulation modelling framework, how concepts
from the manufacturing industry such as JIT deliveries, Lean methods, and prefabrication
can improve productivity in the construction industry. The simulation revealed that one to
two days of buffer stock with 100% Lean and prefabrication could significantly reduce (by
weeks) the project duration. However, these levels brought out disadvantages such as
shortages and less utilization of on-site workers since they were idle during shortages.
Idleness on site, however small, is something costly every general contractor wants to
avoid. Considering the simulation’s results, one could conclude that the reaction time of the
system did not allow these amounts of buffer stock to fully cover variations and hazards
encountered on construction sites.
Results of the DOE confirmed that prefabrication had a significant impact on construction
duration. The company may therefore evaluate the possibility to increase prefabrication in
order to reduce the duration of the construction projects. Applying Lean also contributed to
78
increase productivity while offering a good compromise between the construction duration
and shortages. The increase of buffer stock tended to reduce the number of shortages down
to 0 with 3 days of buffer. More than that did not lead to significant benefits on
construction duration or labor utilization while increasing costs related to storage of buffer
stock on the construction site. Moreover, in the best scenario (3 days buffer, 100% Lean
and 80% prefabrication), adding a night shift helped reduce the construction duration by 5
weeks (19%). However, the risks of shortages were higher and the costs related to adding
one more team, or renting the crane for extra hours and for having more stock on the
construction site, should be taken into consideration.
With current buffer stocks (2.33 days), a decent project duration and labor utilization were
obtained but it could involve some risks of shortage depending on the other parameters.
Three days of buffer stock seemed to be ideal for storage on site with the same resources as
in the current construction. Indeed, the company could improve its productivity from 68.9%
in the current state to 80.9% and reduce its construction duration from 26.09 to 22.31 weeks
by making slight changes such as adding 0.66 days of buffer stock and reorganizing its
storage activities on the construction site (4 to 8 minutes less on each travel time).
The results obtained in this study give scientific numerical ground to confirm the benefits
of applying JIT in construction projects with consideration of specific elements ensuring a
prosperous adoption. Indeed, the quantitative results of the study allow to convince even
more contractors on the profitability of JIT in construction. This paper contributes to
science by illustrating through the literature, factors that influence productivity in
construction and through a simulation model, the quantitative results of the impact of these
factors: prefabrication, Lean, and JIT on the outcome performance of a construction. This is
quantified in the simulation model as construction duration, labor utilization, and shortages.
The paper also contributes in identifying through a DOE, if these factors, whether used
separately or together, influence productivity negatively, neutrally, or positively.
This case study remains limited to the construction of a medium-sized commercial wooden
building. The conclusions may vary in a different frame size building made in different
materials. In future works, it would be interesting to evaluate and simulate different
construction sites in terms of size and complexity to further the findings on JIT in
construction.
79
Appendix
Table A 2 Description of the 60 scenarios tested and their results.
The scenarios presented in this table are all possible combinations of the number of work shifts, JIT, Lean, and prefabrication levels
for a total of 60 scenarios. The results obtained by experimenting these scenarios are also indicated (project duration, labor utilization
and shortages). A drop in labor utilization for scenarios with two work shifts (S1b to S30b), make these scenarios less interesting for
implementation.
Scenarios Nb of
teams
JIT: buffer
stock (days)
Lean
level (%)
Prefabrication
level (%)
Lean: travel
time (min)
Prefabrication:
installation time
(min)
Construct.
duration (weeks)
Labor utilization
(%) Shortages
S0 1 2.33 50 80 (10,20) (82.6,83.8,85) 26.09 68.9% 2
S1a 1 1 0 80 (20,40) (82.6,83.8,85) 33.73 54.9% 3
S2a 1 1 0 100 (20,40) (42.6 ,43.8,45) 27.74 36.3% 11
S3a 1 1 50 80 (10,20) (82.6,83.8,85) 28.33 69.3% 5
S4a 1 1 50 100 (10,20) (42.6 ,43.8,45) 21.45 51.5% 10
S5a 1 1 100 80 (6,12) (82.6,83.8,85) 25.76 76.0% 6
S6a 1 1 100 100 (6,12) (42.6 ,43.8,45) 19.51 57.0% 14
S7a 1 2 0 80 (20,40) (82.6,83.8,85) 33.72 54.2% 2
S8a 1 2 0 100 (20,40) (42.6 ,43.8,45) 26.36 35.8% 2
S9a 1 2 50 80 (10,20) (82.6,83.8,85) 26.09 68.9% 2
S10a 1 2 50 100 (10,20) (42.6 ,43.8,45) 17.51 55.0% 0
S11a 1 2 100 80 (6,12) (82.6,83.8,85) 22.67 78.5% 2
S12a 1 2 100 100 (6,12) (42.6 ,43.8,45) 14.39 67.7% 0
S13a 1 2.33 0 80 (20,40) (82.6,83.8,85) 33.72 54.2% 0
S14a 1 2.33 0 100 (20,40) (42.6 ,43.8,45) 26.08 36.3% 0
80
Table A2. Description of the 60 scenarios tested and their results (continued).
Scenarios Nb of
teams
JIT: buffer
stock (days)
Lean
level (%)
Prefabrication
level (%)
Lean: travel
time (min)
Prefabrication:
installation time
(min)
Construct.
duration (weeks)
Labor utilization
(%) Shortages
S15a 1 2.33 50 100 (10,20) (42.6 ,43.8,45) 17.47 54.9% 0
S16a 1 2.33 100 80 (6,12) (82.6,83.8,85) 22.29 80.9% 0
S17a 1 2.33 100 100 (6,12) (42.6 ,43.8,45) 14.39 67.8% 0
S18a 1 3 0 80 (20,40) (82.6,83.8,85) 32.35 55.4% 0
S19a 1 3 0 100 (20,40) (42.6 ,43.8,45) 26.08 36.3% 0
S20a 1 3 50 80 (10,20) (82.6,83.8,85) 24.59 72.7% 0
S21a 1 3 50 100 (10,20) (42.6 ,43.8,45) 17.49 54.8% 0
S22a 1 3 100 80 (6,12) (82.6,83.8,85) 22.31 80.9% 0
S23a 1 3 100 100 (6,12) (42.6 ,43.8,45) 14.10 67.4% 0
S24a 1 4 0 80 (20,40) (82.6,83.8,85) 32.06 56.1% 0
S25a 1 4 0 100 (20,40) (42.6 ,43.8,45) 26.08 36.3% 0
S26a 1 4 50 80 (10,20) (82.6,83.8,85) 24.59 72.7% 0
S27a 1 4 50 100 (10,20) (42.6 ,43.8,45) 17.47 99.8% 0
S28a 1 4 100 80 (6,12) (82.6,83.8,85) 22.31 80.5% 0
S29a 1 4 100 100 (6,12) (42.6 ,43.8,45) 14.04 67.6% 0
S1b 2 1 0 80 (20,40) (82.6,83.8,85) 24.02 44.6% 17
S2b 2 1 0 100 (20,40) (42.6 ,43.8,45) 21.62 30.8% 11
S3b 2 1 50 80 (10,20) (82.6,83.8,85) 20.45 54.1% 17
S4b 2 1 50 100 (10,20) (42.6 ,43.8,45) 16.00 45.5% 12
S5b 2 1 100 80 (6,12) (82.6,83.8,85) 20.04 61.2% 16
S6b 2 1 100 100 (6,12) (42.6 ,43.8,45) 15.56 46.0% 17
81
Table A2. Description of the 60 scenarios tested and their results (continued and end).
Scenarios Nb of
teams
JIT: buffer
stock (days)
Lean
level (%)
Prefabrication
level (%)
Lean: travel
time (min)
Prefabrication:
installation time
(min)
Construct.
duration (weeks)
Labor utilization
(%) Shortages
S7b 2 2 0 80 (20,40) (82.6,83.8,85) 17.91 55.8% 1
S8b 2 2 0 100 (20,40) (42.6 ,43.8,45) 18.10 33.0% 9
S9b 2 2 50 80 (10,20) (82.6,83.8,85) 17.60 65.6% 5
S10b 2 2 50 100 (10,20) (42.6 ,43.8,45) 14.47 46.4% 11
S11b 2 2 100 80 (6,12) (82.6,83.8,85) 16.33 73.3% 6
S12b 2 2 100 100 (6,12) (42.6 ,43.8,45) 14.42 51.2% 12
S13b 2 2.33 0 80 (20,40) (82.6,83.8,85) 17.60 53.2% 2
S14b 2 2.33 0 100 (20,40) (42.6 ,43.8,45) 14.72 36.4% 6
S15b 2 2.33 50 80 (10,20) (82.6,83.8,85) 15.90 69.5% 1
S16b 2 2.33 50 100 (10,20) (42.6 ,43.8,45) 15.61 43.0% 10
S17b 2 2.33 100 80 (6,12) (82.6,83.8,85) 15.51 70.5% 7
S18b 2 2.33 100 100 (6,12) (42.6 ,43.8,45) 14.42 54.3% 10
S19b 2 3 0 80 (20,40) (82.6,83.8,85) 16.08 57.5% 1
S20b 2 3 0 100 (20,40) (42.6 ,43.8,45) 12.15 35.8% 1
S21b 2 3 50 80 (10,20) (82.6,83.8,85) 12.76 72.9% 0
S22b 2 3 50 100 (10,20) (42.6 ,43.8,45) 13.60 50.5% 4
S23b 2 3 100 80 (6,12) (82.6,83.8,85) 12.39 75.0% 2
S24b 2 3 100 100 (6,12) (42.6 ,43.8,45) 12.75 58.0% 8
S25b 2 4 0 80 (20,40) (82.6,83.8,85) 16.19 56.8% 0
S26b 2 4 0 100 (20,40) (42.6 ,43.8,45) 13.54 36.8% 0
S27b 2 4 50 80 (10,20) (82.6,83.8,85) 13.47 72.8% 0
S28b 2 4 50 100 (10,20) (42.6 ,43.8,45) 10.74 53.8% 1
S29b 2 4 100 80 (6,12) (82.6,83.8,85) 11.80 81.2% 0
S30b 2 4 100 100 (6,12) (42.6 ,43.8,45) 11.57 97.6% 2
82
Acknowledgments
The authors are grateful to the Natural Sciences and Engineering Research Council of
Canada for its financial support through the ICP and CRD programs (IRCPJ 461745-12
and RDCPJ 445200-12) as well as the industrial partners of the NSERC industrial chair
on eco-responsible wood construction (CIRCERB).
Notation
The following symbols are used in this paper:
= null hypothesis;
= alternative hypothesis;
S= standard deviation;
= student statistical value;
α= significance level;
= nu zero.
References
Akintoye, A., (1995). “Just-in-time application and implementation for building material
management.” Construction Management and Economics, 13, 105-113.
Asri, M. a. N. M., Nawi, M. N. M. and Nadarajan, S., (2016). “Key factors of successful
JIT integration with IBS - An overview.” Proc., International Conference on Applied
Science and Technology 2016, AIP Publishing, College Park, 020065 (5 pp.).
Bamana, F., Lehoux, N. and Cloutier, C., (2017). “Just in time in construction:
description and implementation insights.” Proc., 25th Conference of the International
Group for Lean Construction, IGLC 2017, The International Group for Lean
Construction.
Banks, J., (1999). “Introduction to simulation.” 1999 Winter Simulation Conference,
WSC 1999, Institute of Electrical and Electronics Engineers Inc, Piscataway, 7-13.
Bashir, A. M., Suresh, S., Oloke, D. A., Proverbs, D. G., and Gameson, R., (2013). “The
Application of Lean Construction Tools in United Kingdom Construction
Organisations: Findings from a Qualitative Inquiry.” Proc., 2013 Architectural
Engineering National Conference, AEI Press, Reston, 63-72.
Behera, P., Mohanty, R. P. and Prakash, A., (2015). “Understanding Construction Supply
Chain Management.” Production Planning & Control - The Management of
Operations, 26, 1332-50.
Commission de la construction du Québec, (2018a). “Qui sommes-
nous? ”〈http://www.ccq.org/fr-CA/A_QuiSommesNous〉 (Oct. 01, 2017).
83
Commission de la construction du Québec, (2018b). “L’industrie de la construction.”
〈http://www.ccq.org/frCA/A_QuiSommesNous/A05_IndustrieConstruction〉 (Oct.
01, 2017).
Cossio, J. G. and Cossio, A. G., (2012). “Application of just in time to the fabrication and
installation of prefabricated concrete facades in buildings.” Proc., 20th Conference of
the International Group for Lean Construction, IGLC 2012, The International Group
for Lean Construction.
Deshpande, A. S., A. S., Filson, L. E., Salem, O. M. & Miller, R. A., (2012). “Lean
techniques in the management of the design of an industrial project.” Journal of
Management in Engineering, 28, 221-223.
Forest Products Association of Canada, (2018). “Les biomatériaux en Amérique du Nord
: synthèse. ” 〈http://www.fpac.ca/publications/Bio%20materials_FR.pdf〉 (June 01,
2016).
Friblick, F., Tommelein, I.D., Mueller, E. & Falk, J.H., (2009). “Development of an
integrated facade system to improve the high-rise building process.” Proc., 17th
Conference of the International Group for Lean Construction, The International
Group for Lean Construction, 359-370.
Gosselin, A., Lehoux, N., Cimon, Y., Blanchet, P., (2015). “Main Motivations and
Barriers for Using Wood as a Structural Building Material – A Case Study.” Proc.,
11e Congres International de Génie Industriel – CIGI2015, BioResources, Québec,
546-570.
Khalfan, M. M. A., Mcdermott, P., Oyegoke, A. S., Dickinson, M. T., Li, X., and
Neilson, D., (2008). “Application of kanban in the UK construction industry by public
sector clients.” Proc., 16th Annual Conference of the International Group for Lean
Construction, IGLC16, The International Group for Lean Construction, 347-358.
Pegden, C. D. and Sturrock, D. T., (2011). “Introduction to Simio”. 2011 Winter
Simulation Conference, WSC 2011, Institute of Electrical and Electronics Engineers
Inc., Piscataway, 29-38.
Pheng, L. S., (2001). “Just-in-time management in precast concrete construction: A
survey of the readiness of main contractors in Singapore.” Integrated manufacturing
systems, 12, 416-429.
Pheng, L. S. and Chuan, C. J., (2001). “Just-in-time management in precast concrete
construction: A survey of the readiness of main contractors in Singapore.” Journal of
Construction Engineering and Management, 127, 494-501.
Pheng, L. S. and Hui, M. S., (1999). “The application of JIT philosophy to construction: a
case study in site layout.” Construction Management and Economics, 17:5, 657-668.
Salem, O., Solomon, J., Genaidy, A. and Minkarah, I., (2006). “Lean construction: from
theory to implementation.” Journal of Management in Engineering, 22, 168-75.
Sargent, R. G., 2013. Verification and validation of simulation models. Journal of
Simulation, 7, 12-24.
Statistics Canada, (2018). “Employment by industry.” 〈http://www.statcan.gc.ca/tables-
tableaux/sum-som/l01/cst01/econ40-eng.htm〉 (May 22, 2018).
Tommelein, I.D. and Li, A. E. Y., (1999). “Just-In-Time concrete delivery: Mapping
alternatives for vertical supply chain integration.” Proc., 7th Conference of the
International Group for Lean Construction, The International Group for Lean
Construction, 97-108.
84
Tommelein, I.D. and Weissenberger, M., (1999). “More Just-In-Time: Location of
buffers in structural steel supply and construction processes.” Proc., 7th Conference of
the International Group for Lean Construction, The International Group for Lean
Construction, 109-120.
Tranfield, D., Denyer, D. and Smart, P., (2003). “Towards a methodology for developing
evidence-informed management knowledge by means of systematic review.” British
Journal of Management, 14, 207-222.
Viana, D.D., Tillmann, P.A., Zach, S., Tommelein, I.D., and Formoso, C.T., (2015).
“Analysis of HVAC subcontractor mechanisms for JIT Materials supply to a
construction site.” Proc., 23rd Conference of the International Group for Lean
Construction, IGLC23, The International Group for Lean Construction, 246-255.
Vokurka, R. J. and Davis, R. A., (1996). “Just-in-time: The evolution of a philosophy.”
Production and Inventory Management Journal, 37, 56.
Womack, J. P. and Jones, D. T., (1996). “Beyond Toyota: How to root out waste and
pursue perfection.” Harvard Business Review, 74, 140-158.
Zimmer, E., Salem, O., Genaidy, A. and Shell, R., (2008). “Case study: Lean supply
chain management in construction projects.” Proc., 16th Annual Conference of the
International Group for Lean Construction, IGLC16, The International Group for
Lean Construction, 381-388.
85
Conclusion
Depuis son apparition, la philosophie du JAT a outillé les compagnies manufacturières
d’approches leur permettant d’obtenir d’énormes gains en productivité dans leurs
opérations. Le domaine de la construction étant continuellement à l’affût de nouvelles
manières de s’améliorer, cette étude avait pour but d’étudier comment les principes du
JAT pouvaient être mis en pratique dans le domaine de la construction afin d’apporter des
retombées similaires. Les principaux objectifs poursuivis étaient donc d’expliciter la
philosophie du JAT en construction tout matériau confondu, d’évaluer différents
scénarios d’implantation du JAT, et ultimement de proposer la meilleure combinaison de
paramètres pour maximiser la productivité dans un projet de construction.
Afin d’atteindre le but visé, une revue systématique de la littérature a été réalisée sur le
JAT dans la construction. La revue a permis de bien cerner comment les entreprises
peuvent adopter cette philosophie tout en conduisant à l’élaboration d’un cadre
d’implantation préliminaire du JAT pour le secteur de la construction. Par la suite, un
modèle de simulation de la construction d’un bâtiment en bois de six étages a été conçu,
permettant d’expérimenter différents scénarios de déploiement de stratégies de JAT tout
en mesurant leurs impacts sur les trois indicateurs de performance suivants : la durée de
la construction, le taux d’utilisation des travailleurs et les ruptures de stock. Ensuite, des
analyses statistiques ont été effectuées afin d’établir les éléments ayant le plus d’effets
sur ces trois indicateurs de performance. Finalement, il a été possible d’agrémenter le
cadre d’implantation grâce aux conclusions tirées, de formuler une synthèse des résultats
obtenus ainsi que de proposer des recommandations pour le futur.
L’analyse des résultats a montré qu’avec 1 jour de stock tampon, 50% Lean et 80%
préfabrication, la durée de la construction est de 28,33 semaines. Ce scénario présente les
mêmes niveaux de Lean et de préfabrication que le modèle de base, mais un stock
tampon réduit à un jour, ce qui a entraîné un temps de construction plus long puisqu’il y a
eu manque de matériaux 5 fois durant la simulation. Toujours avec les mêmes niveaux de
Lean et de préfabrication, mais 2 jours de stock tampon, la durée de construction (26,09
semaines) est similaire au modèle de base avec 2 pénuries dans les deux cas. Grâce aux
86
résultats, on observe que toutes les combinaisons de niveaux de Lean et de préfabrication
pour 3 et 4 jours de stock tampon n’entraînent aucune pénurie.
Il est intéressant de noter qu'avec le modèle de base, le taux d’utilisation de la main-
d’œuvre varie de 69,3% pour 1 jour de stock tampon à 72,7% pour 3 et 4 jours de stock
tampon. En effet, jusqu’à 3 jours de stock tampon, les pénuries tombent à 0, ce qui réduit
le temps d’inactivité des ressources sur le chantier. En introduisant davantage de
méthodes Lean pour réduire les temps de déplacement de 4 à 8 minutes, une réduction de
26,46 jours pourrait être obtenue sur la durée de la construction. De plus, en augmentant
le niveau de préfabrication à 100% avec les mêmes niveaux de Lean et de stock tampon
que le modèle de base, la construction simulée se termine en 14,1 semaines, mais le taux
d’utilisation des travailleurs tombe à 67,4% puisque les travailleurs mettent moins de
temps à installer les mêmes pièces, les laissant libres en attendant la prochaine livraison
de matériaux. Cependant, le niveau de préfabrication de 100% nécessite en réalité
beaucoup de temps et d'énergie en amont. En fait, le niveau élevé de préfabrication
repose sur l'extrême coordination nécessaire à la phase de conception avec toutes les
parties prenantes: clients, ingénieurs en structure et ingénieurs électriques, principaux
contracteurs et architectes. En raison de l'énergie significative requise pour atteindre ce
niveau de préfabrication, l'entreprise a préféré explorer d'autres scénarios bénéfiques.
Les mêmes 30 scénarios ont également été testés avec 2 quarts de travail par jour. La
deuxième équipe poursuit l'installation après l'équipe de jour. Dans ces scénarios, la
durée de la construction est considérablement réduite, tandis que les pénuries ont
tendance à augmenter. Une illustration de ces résultats se reflète dans le scénario avec
100% Lean, 80% de préfabrication et 2,33 jours de stock tampon. Dans ce scénario, la
construction se termine en 15,5 semaines, 6,8 semaines plus tôt qu'avec un seul quart de
travail. Cependant, les pénuries sont plus importantes, 7 au lieu de 0, et tendent à
diminuer lorsque le stock augmente. Le taux d’utilisation de la main-d’œuvre diminue
également avec 80,9% pour une équipe de jour contre 70,5% pour 2 équipes. Cette baisse
du taux d'utilisation de la main-d'œuvre pour les scénarios comportant deux équipes rend
ces scénarios moins intéressants dans leur mise en œuvre. Dans tous les scénarios, le taux
d'utilisation de la grue se situe entre 95,2% et 99,8%. La grue représente le goulot
87
d'étranglement du système. L’ajout d’une deuxième équipe de travail a permis
d’augmenter le taux d’utilisation de la main-d’œuvre qui a ensuite considérablement
baissé lorsque la grue a atteint son taux d’utilisation maximal. Par conséquent, les
scénarios illustrant l’ajout d’une grue supplémentaire ont été proposés. La compagnie a
préféré explorer d’autres alternatives tout en gardant l’option de la grue supplémentaire
pour un moyen terme car actuellement le rythme serait trop rapide pour l’usine qui
fournit plusieurs chantiers en même temps et pour les différentes parties prenantes.
Suite aux investigations et analyses effectuées, les résultats présentés dans ce mémoire
démontrent que l’application du JAT dans le projet de construction à l’étude a permis de
réduire la durée de la construction d’environ 14,49% (de 26,09 à 22,31 semaines),
d’augmenter le taux d’utilisation des travailleurs de 17,42% (de 68,9% à 80,9%) et
d’éliminer les risques de ruptures de stock sur le chantier. En effet, grâce au modèle de
simulation, il a été possible de déterminer ces meilleures performances qui sont issues du
même niveau de préfabrication qu’effectué par la compagnie, d’une légère augmentation
du stock tampon de 2,33 jours à 3 jours et de l’adoption des méthodes Lean sur le
chantier. Ces conclusions ont été reçues avec beaucoup d’entrain par le partenaire
industriel qui, dès lors, s’intéresse à l’opérationnalisation du Lean sur ses chantiers. Avec
les résultats du DOE, il a été possible de confirmer que la préfabrication a un impact
significatif sur la durée de construction. Il a également été possible de confirmer que le
déploiement du Lean dans les activités de chantier augmente le taux d’utilisation des
travailleurs tout en offrant un bon compromis entre la durée de construction et les
pénuries. Les résultats obtenus sont d’autant plus pertinents pour l’industrie de la
construction en bois où la rapidité d’érection et la réduction des coûts demeurent des
éléments clés motivant les compagnies à opter pour des constructions en bois.
La simulation s’est révélée très efficace dans cette étude de par les variétés des scénarios
qu’elle a rendus possible d’expérimenter. En effet, pour observer les effets des différentes
combinaisons des paramètres d’intérêts dans l’étude, il s’agissait d’apporter des
modifications au modèle de simulation de départ. Il était alors possible d’obtenir des
résultats rapidement, directement à l’ordinateur, sans nécessiter des investissements
88
faramineux dans la réalité. De toute évidence, le modèle de simulation et par conséquent
les résultats issus de cette étude n’auraient vu le jour sans l’apport de données obtenues
auprès du partenaire Nordic Structures dont la collaboration fut réellement présente tout
au long de l’étude. Les résultats obtenus des 60 scénarios testés pourraient donc faire
l’objet de changements de pratiques au sein de la compagnie dans le futur (niveau de
stock sur le chantier, organisation Lean, et niveau de préfabrication). Cependant, si la
compagnie désire tester de nouveaux paramètres à partir du modèle de simulation, cela
impliquera pour la personne en charge une connaissance de l’utilisation du logiciel Simio
afin d’apporter correctement des modifications au modèle.
La présente étude a permis de tirer des conclusions intéressantes quant à l’application du
JAT dans la construction par le biais de la simulation d’un projet de construction réelle.
Cependant, l’étude s’est limitée à la construction d'un bâtiment commercial en bois de
taille moyenne. Les conclusions tirées pourraient varier d’un type de projets de
construction à un autre. Dans le futur, il serait intéressant de simuler différents chantiers
de construction en termes de taille et de complexité pour élargir la compréhension des
différentes parties prenantes sur l’application du JAT dans la construction. Aussi, en se
basant sur la préanalyse effectuée dans le troisième article de cette étude, l’ajout de
paramètres supplémentaires dans le modèle de simulation pourrait également être
exploité à l’avenir. Par exemple, l’ajout en paramètres de l’arrivée des camions de
livraisons permettrait de tenir compte de l’incertitude créée par les embouteillages dans
lesquelles se retrouvent fréquemment les camionneurs. Aussi, en tenant compte de
l’ancienneté aux postes des équipes de travail, les temps d’installation varieraient en
fonction de l’expérience des travailleurs. Par ailleurs, il y aurait un réel potentiel à
examiner dans quelle mesure les intrants du modèle de simulation comme celui
développé dans cette étude pourrait être obtenus et introduits directement à travers le
BIM (Building Information Modeling). En effet, il deviendrait possible d’intégrer en
temps réel des intrants variables comme les embouteillages et les déviations de trajet de
livraisons dues aux travaux routiers, aux accidents, ou aux intempéries. Ainsi, un modèle
beaucoup plus dynamique permettrait de mieux saisir l’influence de la philosophie dans
la construction.
89
Bibliographie
Banks, J., 1999. Introduction to simulation. 1999 Winter Simulation Conference, WSC
1999, 5-8 December 1999 Phoenix. Piscataway: Institute of Electrical and Electronics
Engineers Inc., 7-13.
Biles, W. E. Introduction to simulation. 1987 Winter Simulation Conference, WSC 1987,
14-16 December 1987 San Diego. Piscataway: Institute of Electrical and Electronics
Engineers Inc., 7-15.
Bjornfot, A. and Sarden, Y., 2006. Prefabrication: A lean strategy for value generation in
construction. In: 14th Annual Conference of the International Group for Lean
Construction, IGLC 2014, 25-27 July 2006 Santiago. The International Group for Lean
Construction, 265-277.
Centeno, M. A., 1996. An introduction to simulation modeling. 1996 Winter Simulation
Conference, WSC 1996, 8-11 December 1996 San Diego. Piscataway: Institute of
Electrical and Electronics Engineers Inc., 15-22.
Cossio, J. G. and Cossio, A. G., 2012. Application of just in time to the fabrication and
installation of prefabricated concrete facades in buildings. In: 20th Conference of the
International Group for Lean Construction, IGLC 2012, 18-20 July 2012 San Diego. The
International Group for Lean Construction.
Ebrahimpour, M. and Schonberger, R. J.,1984. The Japanese just-in-time/total quality
control production system: potential for developing countries. International Journal of
Production Research, 22, 421-30.
Eriksson, P. E., 2010. Improving construction supply chain collaboration and
performance: a lean construction pilot project. Supply Chain Management, 15, 394-403.
Ghada, Z., 2012. Partie théorique concernant : Généralités sur le JAT et le Kanban. [En
ligne]
Disponible sur : https://fr.scribd.com/doc/88416317/Histoire-du-juste-a-temps
[Accès le 24/05/2015]
Gunasekaran, A., Goyal, S. K., Martikainen, T. and Yli-Olli, P., 1993. Determining
economic inventory polices in a multi-stage just-in-time production system. International
Journal of Production Economics, 30-31, 531-542.
Hohmann, C., 2009. Les 5S et les sept gaspillages selon TOYOTA. [En ligne]
Disponible sur : http://chohmann.free.fr/5S/gaspillages.htm [Accès le 12/10/2015]
90
Horvath, A., 2004. Construction materials and the environment. Annual Review of
Environment and Resources, 29, 181-204.
Ingalls, R. G., 2008. Introduction to simulation. 2008 Winter Simulation Conference,
WSC 2008, 7-10 December 2008 Miami. Piscataway: Institute of Electrical and
Electronics Engineers Inc., 17-26.
Jørgensen, B. and Emmitt, S., 2008. Lost in transition: the transfer of lean manufacturing
to construction. Engineering, Construction and Architectural Management, 15, 383-398.
Koskela, L., 2000. An exploration towards a production theory and its application to
construction. Technical Research Center of Finland, Espoo 15, 02150.
Koskela, L., Howell, G., Ballard, G., and Tommelein, I., 2004. The foundations of lean
construction. [En ligne]
Disponible sur:
http://www.researchgate.net/publication/28578914_The_foundations_of_lean_constructi
on [Accès le 21/10/2015]
Naso, D., Surico, M., Turchiano, B. and Kaymak, U., 2004. Just-in-time production and
delivery in supply chains: a hybrid evolutionary approach. International Conference on
Systems, Man and Cybernetics, 10-13 October 2004 The Hague. Piscataway: Institute of
Electrical and Electronics Engineers Inc., 1932-7.
Opfer, N. D., 1998. Just-in-time construction materials management. In: Proceedings of
the 1998 42nd Annual Meeting of AACE International, 28June-1 July 1998 Cincinnati.
Morgantown: AACE International Transactions, PC.05.
Pegden, C. D. and Sturrock, D. T., 2011. Introduction to Simio. 2011 Winter Simulation
Conference, WSC 2011, 11-14 December 2011 Phoenix. Piscataway: Institute of
Electrical and Electronics Engineers Inc., 29-38.
Pheng, L. S., 2001. Just-in-time management in precast concrete construction: A survey
of the readiness of main contractors in Singapore. Integrated manufacturing systems, 12,
416-429.
Pheng, L. S. and Chuan, C. J., 2001. Just-in-time management in precast concrete
construction: A survey of the readiness of main contractors in Singapore. Journal of
Construction Engineering and Management, 127, 494-501.
Pheng, L. S. and Hui, M. S., 1999. The application of JIT philosophy to construction: a
case study in site layout. Construction Management and Economics, 17:5, 657-668.
Salem, O., Solomon, J., Genaidy, A. and Minkarah, I., 2006. Lean construction: from
theory to implementation. Journal of Management in Engineering, 22, 168-75.
91
Sanchez, P. J., 2006. As simple as possible, but no simpler: a gentle introduction to
simulation modeling. 2006 Winter Simulation Conference, WSC 2006, 3-6 December
2006 Monterey. Piscataway: Institute of Electrical and Electronics Engineers Inc., 2-10.
Sargent, R. G., 2013. Verification and validation of simulation models. Journal of
Simulation, 7, 12-24.
Schmidt, J. W., 1984. Introduction to simulation. 1984 Winter Simulation Conference,
WSC 1984, 28-30 November 1984 Dallas. Piscataway: Institute of Electrical and
Electronics Engineers Inc., 65-73.
Simio, 2018. About Simio. [En ligne]
Disponible sur: https://www.simio.com/about-simio/
[Accès le 05/04/2018]
Sugimori, Y., Kusunoki, K., Cho, F. and Uchikawa, S., 1977. Toyota production system
and Kanban system. Materialization of just-in-time and respect-for-human system.
International Journal of Production Research, 15, 553-564.
Tranfield, D., Denyer, D. and Smart, P., 2003. Towards a methodology for developing
evidence-informed management knowledge by means of systematic review. British
Journal of Management, 14, 207-222.
UNEP-United Nations Environment Program, 2018. Sustainable buildings and
construction programme. [En ligne]
Disponible sur:
http://www.oneplanetnetwork.org/sites/default/files/brochure_10yfp_sbc_prog_final.pdf
[Accès le 01/04/2018]
Vokurka, R. J. and Davis, R. A., 1996. Just-in-time: The evolution of a philosophy.
Production and Inventory Management Journal, 37, 56.
White, K. P. and Ingalls, R. G., 2015. Introduction to simulation. 2015 Winter Simulation
Conference, WSC 2015, 6-9 December 2015 Huntington Beach. Piscataway: Institute of
Electrical and Electronics Engineers Inc., 1741-55.