1 presented at international symposium on smart grid and renewable generation impacts on the power...

79
1 Presented at International Symposium on Smart Grid and Renewable Generation Impacts on the Power System Taiwan National University, Institute of Applied Mechanics Taipei by Robert Blohm 卜卜卜 Member North American Electric Reliability Corp. (NERC) Managing Director KEEN Resources Asia Ltd. http://www.blohm.cnc.net Operating and Planning the US & China Grids for Reliability and Economic Efficiency to Enable Optimal Use of Smart Grid

Upload: job-basil-tucker

Post on 26-Dec-2015

213 views

Category:

Documents


0 download

TRANSCRIPT

1

Presented atInternational Symposium on Smart Grid and

Renewable Generation Impacts on the Power SystemTaiwan National University, Institute of Applied Mechanics

Taipeiby

Robert Blohm卜若柏Member

North American Electric Reliability Corp. (NERC)

Managing DirectorKEEN Resources Asia Ltd.

http://www.blohm.cnc.net

November 18, 2010年 11月 28日

Operating and Planning the US & China Grids for Reliability and Economic

Efficiency to Enable Optimal Use of Smart Grid

2

OUTLINE

1. How the electric grid is planned and operated determines which Smart Grid investments will be made. Planning and operating models

2. Planning models.

3. Operating models

4. Challenges in grid configuration and management

5. Challenges for Mainland China

3

1. How the electric grid is planned and operated determines which Smart Grid investments will be made. Planning and operating models

• are based on:– power-flow model. Model of electricity flow on the system

in response to changes in system components.– models to forecast

• economic/scheduled usage and • contingencies/emergencies.

• are combined with 2 economic efficiency principles– choosing the least-cost (or market-chosen) alternative

project– economic transfer from the causer of the cost to the bearer

of the cost

4

2. Planning models. Competitive market time horizon: less than 5 years. Two kinds often confused together.

• Economic. Of scheduled power. – Based on an economic forecasting model

and a targeted, controlled LOLP (loss-of-load probability)

– Criticism: • biased toward excess supply• ignores price-rationing• attempts to target an unstated low price not

clearly related to LOLP.

5

2. Planning models. Competitive market time horizon: less than 5 years. Two kinds often confused together. (cont.d)

• Reliability. Transmission. Of unscheduled power. – Criticism: usually based on robustness to any

contingency (a fault or loss of any element) without regard to probability of occurrence.

• Probability must be based on historical data. The US still doesn’t have a complete database

– NERC has a generation outage database, whose software was translated by China in the 1980s

– DOE has a transmission outage database but no load loss database

• Two events could be more probable than a single one

6

2. Planning models. Competitive market time horizon: less than 5 years. Two kinds often confused together. (cont.d)• Reliability. Transmission. Of unscheduled power.

(cont.d)– Used to determine TRM (transmission reliability margin) to

subtract from TTC (total transmission capacity) to get ATC (available transmission capacity)

– TRM is designed to meet the NERC performance requirement of withstanding

• a single uncontrolled contingency, and • a second controlled contingency to prevent an uncontrolled

contingency

– Reliability is strictly defined as avoidance of uncontrolled outages

7

3. Operating models• Transmission.

– Economic. Rationing of ATC (available transmission capacity) by redispatch when there is congestion. By • out-of-“merit”-order dispatch of generation (SEE NEXT SLIDE), and possibly by • allocation of FTRs (fixed transmission rights) in markets.

– Reliability. Based in the US on the IDC (interchange distribution calculator) which is a DC model of the powerflow identifying source and sink and updated every hour (but updatable every minute).• Invocation of TLR (transmission loading relief) to reroute/redispatch power in the event of a contingency.• (Under development) Allocation of unscheduled congestion by ACE Distribution Factors. Allocation of

allowable ACE by IDC distribution factors determined by Kirchoff’s law.

• Generation.– Economic. By “merit” order of

• least cost (“system lambda”) in non-markets, or• least LMP (“locational marginal price”) in markets

congestion

energy price to generator on cheap sideof constraint

congested price to consumer on expensive sideof constraint

Congestion chargenormally to

transmission owners

Supply curve

Demand curve

Generators on thecheap side of theconstraint colludeto raise prices tocapture congestionrents fromtransmission owner

Only physical/forward transmissionrights can prevent generators fromcapturing congestion rents fromtransmission owners. Cannot beprevented by congestion contractsthat value congested transmission asthe difference in energy pricesacross the congested interface.

quantity

price

9

3. Operating models (cont.d)

• Generation (cont.d)– Frequency control. Dispatch of power generation held as

“reliability reserve”• Definition. Balancing of unscheduled supply (generation) and

demand (load), measured by closeness to scheduled frequency (60 Hz in US, 50 Hz in China).

– Upward deviation• High frequency means insufficient supply• Excessive high frequency means surge in power flow which spreads

throughout the system (SEE NEXT SLIDE) – Downward deviation

• Low frequency means insufficient load• Excessive low frequency means generator vibration and explosion

E a s te rn Inte rco nne c tio n B la cko ut

5 9 .8

5 9 .8 5

5 9 .9

5 9 .9 5

6 0

6 0 .0 5

6 0 .1

6 0 .1 5

6 0 .2

6 0 .2 5

6 0 .315

:05:

42

15:0

6:08

15:0

6:30

15:0

6:54

15:0

7:16

15:0

7:36

15:0

8:00

15:0

8:20

15:0

8:46

15:0

9:06

15:0

9:30

15:0

9:52

15:1

0:12

15:1

0:38

15:1

0:58

15:11

:22

15:11

:42

15:1

2:06

15:1

2:32

15:1

2:53

15:1

3:18

15:1

3:40

15:1

4:03

15:1

4:26

15:1

4:56

15:1

5:16

15:1

5:38

T im e o n A ug us t 1 4 , 2 0 0 3

Hertz

F re q ue ncy

11

3. Operating models (cont.d)

• Generation (cont.d)– Frequency control. Dispatch of power generation held as

“reliability reserve” (cont.d)• Means. Deployment of “reliability reserve” power generation to

manage unscheduled changes.– Instantaneous governor response (within seconds): socialized

responsibility because all operating generators are equipped to “respond” to any frequency change in order to maintain frequency.

• The most expensive reserve• The most difficult to measure and require

– Regulation & AGC (automatic generation control) by more slowly deployable reserve (within ten minutes) that is deployed to free-up the governor response for response to the next frequency change.

Regulation

Operating Reserves

Load Following Following

Energy Market Energy

Capacity

FrequencyResponse

Response Timein order of quickness

\\

Seconds

A Few to Several Minutes

10 to 15 Minutes

30 Minutes

Market Interval – One Hour

Response Not Defined

The value of resources lies not just in the amount of energy but also inhow readily the energy is available to counter sudden frequency error.

Resourcesstacked by value

Adapted from Energy Mark, Inc.

Rapidly Stopping a Frequency Drop

AGC & Regulation Gradually Restore Frequency

Hz

40

6059.925

1

Hz

59.925

60

10

Seconds

Minutes

AGC & Regulation (including by the party responsible for the power loss )(Shared)Rapid Response

All generators rapidly slow down when generation is suddenly lost while inertia, shared governor response, and load-response counter and arrest the slowdown

within seconds.

Some generators subsequently gradually provide additional excess power to replace the original power loss, while all shared governor response is gradually

withdrawn in readiness for the next sudden power loss. Source: “Author’s analysis and Robert W. Cummings “Overview Frequency Response Initiative”, North American Electric Reliability Corp. (Princeton, NJ) 2010. http://www.spp.org/publications/NERC%20Frequency%20Response%20Initiative%20Overview.pdf

3 9 .2

3 9 .4

3 9 .6

3 9 .8

4 0

4 0 .2

4 0 .4

4 0 .6

4 0 .8

8 :2 4 :0 0 8 :3 1 :1 2 8 :3 8 :2 4 8 :4 5 :3 6 8 :5 2 :4 8 9 :0 0 :0 0 9 :0 7 :1 2 9 :1 4 :2 4 9 :2 1 :3 6 9 :2 8 :4 8 9 :3 6 :0 0

5 9 .3

5 9 .4

5 9 .5

5 9 .6

5 9 .7

5 9 .8

5 9 .9

6 0

6 0 .1

6 0 .2

PO PL A R H.A 7 9 0 S PO PL A R HIL L G EN .A V Fre q

Fre q u e n c y

Po p la r Hills MW O u tp u t

P e rv e rse G o v e rn o r R e sp o n se

15

3. Operating models (cont.d)

• Generation (cont.d)– Frequency control. Dispatch of power generation held as

“reliability reserve” (cont.d) • US Measure/standard:

– Instantaneous response: • based on a self-measure of responsiveness • included as an obligation in the non-instantaneous response measure• Result: • --Cheaper, non-instantaneous response is being used to meet the • instantaneous response obligation• --frequency response deterioration in North America (SEE NEXT • SLIDES)• --2003 Northeast blackout because the instantaneous response • obligation was not strictly met

16

3. Operating models (cont.d)

-

500

1,000

1,500

2,000

2,500

3,000

3,500

4,000

1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009

MW

/ 0.

01 H

z

Year

Eastern Interconnection Mean Primary Frequency Response

*

* 1999 Data Interpolated

Source: J. Ingleson & E. Allen, "Tracking the Eastern Interconnection Frequency Governing Characteristic" to be presented at 2010 IEEE PES Source: J. Ingleson & E. Allen, “Tracking the Eastern Interconnection Frequency Governing Characteristic”, IEEE Power & Energy Society, Minneapolis, July 26, 2010.

Deteriorating Eastern Interconnection Frequency Response

WECC Frequency Response

11001150120012501300135014001450150015501600

1998 1999 2000 2001 2002

Year

MW

/0.1

Hz

20Source: Robert W. Cummings “Overview Frequency Response Initiative”, North American Electric Reliability Corp. (Princeton, NJ) 2010, slide 13. Slide 30 of http://ewh.ieee.org/cmte/pes/etcc/B_Cummings_Latest_Developments_on_NERC_Standards.pdf

Deadbanding All the Governors on an Interconnection Fattens the Tails of the Probability Distribution of Frequency Error by at least a 1/2 Standard Deviation Worth of Extra Probability Mass

Flat-top Near Normal Distribution stretched by ½ SD of Near Normal Distribution

Deadbanding of Governors Worsens the Frequency Volatility/Risk on an Interconnection

Asleep at the Switch. Robbing Peter to Pay Paul.

-0.0001

0.0029

0.0059

0.0089

0.0119

0.0149

0.0179

0.0209

0.0239

Co

un

t as

a S

ha

re o

f 1

-0.2

50

-0.2

00

-0.1

50

-0.1

00

-0.0

50

0.00

0

0.05

0

0.10

0

0.15

0

0.20

0

0.25

0

Frequency Error (Hz)

Deadbanding Transfers Central Probability Mass from the Old Distribution to Tail Mass of the New Distribution

Deadbanding Transfers Central Probability Mass from the Old Distribution to Tail Mass of the New Distribution

Within the Deadband, Frequency Error is Distributed Uniformly with No Central Tendency

Source: Author’s analysis and H.F. Illian & S.L. Niemeyer, “Integrating Variable Renewable Energy Resour-ces into the Smart Grid”, Carnegie Mellon University Transmission Conference, Pittsburgh, March 10, 2009. http://www.ece.cmu.edu/~electricityconference/2009/2009%20CMU%20Smart%20Grids%20Conf%20Disk/Presentations/Day%201/Session%201/P11_H%20Illian_Integrating%20Renewable%20Resources.pdf

-0.04000 59.96000 4.69287 8.52357%-0.03900 59.96100 4.49175 8.68258%-0.03800 59.96200 4.29064 8.85650%-0.03700 59.96300 4.08952 9.04752%-0.03600 59.96400 3.88840 9.25830%-0.03500 59.96500 3.68728 9.49208%-0.03400 59.96600 3.48617 9.75283%-0.03300 59.96700 3.28505 10.04551%-0.03200 59.96800 3.08393 10.37636%-0.03100 59.96900 2.88281 10.75338%-0.03000 59.97000 2.68170 11.18694%-0.02900 59.97100 2.48058 11.69081%-0.02800 59.97200 2.27946 12.28359%-0.02700 59.97300 2.07835 12.99110%-0.02600 59.97400 1.87723 13.85020%-0.02500 59.97500 1.67611 14.91548%-0.02400 59.97600 1.47499 16.27125%-0.02300 59.97700 1.27388 18.05512%-0.02200 59.97800 1.07276 20.50786%-0.02100 59.97900 0.87164 24.09245%-0.02000 59.98000 0.67052 29.82737%-0.01900 59.98100 0.46941 40.47654%-0.01800 59.98200 0.26829 67.09147%-0.01700 59.98300 0.06717 100.00000%-0.01600 59.98400 0.00000 100.00000%

Frequency Grid FrequencyDeviation Frequency Response

Hz Hz MW Droop %

-0.04000 59.96000 8.00000 5.00000%-0.03900 59.96100 7.80000 5.00000%-0.03800 59.96200 7.60000 5.00000%-0.03700 59.96300 7.40000 5.00000%-0.03600 59.96400 7.20000 5.00000%-0.03500 59.96500 0.00000 100.00000%-0.03400 59.96600 0.00000 100.00000%-0.03300 59.96700 0.00000 100.00000%-0.03200 59.96800 0.00000 100.00000%-0.03100 59.96900 0.00000 100.00000%-0.03000 59.97000 0.00000 100.00000%-0.02900 59.97100 0.00000 100.00000%-0.02800 59.97200 0.00000 100.00000%-0.02700 59.97300 0.00000 100.00000%-0.02600 59.97400 0.00000 100.00000%-0.02500 59.97500 0.00000 100.00000%-0.02400 59.97600 0.00000 100.00000%-0.02300 59.97700 0.00000 100.00000%-0.02200 59.97800 0.00000 100.00000%-0.02100 59.97900 0.00000 100.00000%-0.02000 59.98000 0.00000 100.00000%-0.01900 59.98100 0.00000 100.00000%-0.01800 59.98200 0.00000 100.00000%-0.01700 59.98300 0.00000 100.00000%-0.01600 59.98400 0.00000 100.00000%

Frequency Grid FrequencyDeviation Frequency Response

Hz Hz MW Droop %

Governor response “Steps” to the 5% droop curve at the dead-band

Governor response is proportional at the dead-

band reaching 5% at 3 Hz deviation

Dead-band

Dead-band

600 MW Steam Turbine 5% Droop Setting

0.01666 Hz Dead-Band 0.036 Hz Dead-Band

Do you call this response “continuous but non-linear change inside the dead band? No. It’s “discontinuous”.

ERCOT

600

MW Response

-1 Hz 1 Hz 2 Hz-2 Hz

-3 Hz 3 Hz0 Hz

MW Response±16 mHz deadband

±36 mHz deadband

Responsiveness=Slope=-20 MW/0.1HzDroop: linear (5%)

Responsiveness=Slope=-20.112 MW/0.1HzDroop: geometrically decreasing (to 5% at ±3Hz)

-600

ERCOT

0

24

3. Operating models (cont.d)• Generation (cont.d)

– Frequency control. Dispatch of power generation held as “reliability reserve” (cont.d)

• US Measure/standard: (cont.d)– Non-instantaneous response: CPS1 Control Performance Standard. (SEE EQUATION & GRAPH)

• Measures: • --a statistical average over time• --a combination of (“covariance” between) system frequency • deviation and individual deviation. • ----When the system deviates very much, the individual member is • allowed to deviate very little• ----When the system deviates very little, the individual member is • allowed to deviate very much• The deviation limits can be statistically determined (SEE GRAPH)• --by the US standard of a once-in-ten-years probability of a major • blackout event. (Economies have experienced recessions at• approximately the same rate in the past century.)• --by a database of scanned frequency and tie-line error

25

year_a_in_utesmin_of_number:600,525

erargl_getting_from_deviations_frequency_eoustantanins_prevent_to_response__________

governor_eoustantanins_provide_to_obligation_shared_the:FB10

FB10TACE

i_Authority_Balancing_of_Error_Control_Area:ACE

other_each_to_derconnecteint_slysynchronou_are_iiI

BB.fleet_generation_s'I_ctionInterconne_of_"bias_"Hz1.0/MW:B

fleet_generation_s'i_Authority_Balancing_of_"bias_"Hz1.0/MW:B

0T.i_Authority_Balancing_of_Error_lineTie_averageutemin1:T

t_utemin_at_deviation_frequency_averageutemin1:F

deviation_frequency_average_on_itlim_etargt:

F600,525

1

B

BF

B10

BF10T

600,525

1

B

BF

B10

FB10T

600,525

1

:I_ctionInterconne_for_dardtanS_ePerformanc_Control

B10

B10F

B10

FB10T

600,525

1

i_AUTHORITY_BALANCING_FOR_MEASURE_EPERFORMANC_CONTROL_s'NERC

ti

tiii

i

IIi

iI

i

Iiii

t

2

yeart

2t

2

I

Iii

t

yeart I

Iiit

Ii

i

Ii Ii

2

I

it

yeart I

tii

2

I

it

yeart I

tii

CPS1

(mHz)+

+

FB i 10

-

MW, or

Control area i's maximum allowed 1-minute average tie-lineerror (plus response obligation) in direction of the frequency error:

:

:

: One-year probability density limit on 1-minuteaverages of frequency error, adjusted for deviation of

the mean from 0

FBi 10

F

F

:

: InstantaneousProbability

"No inadvertent allowedin the direction of

Frequency error when"

F :

On average over the past year:Approximate

0iB :

F

FBT

ii10

1-minute average ofFrequency error - + F

50-50

Annual standarddeviation of F

Year's Mean of F

Control area i's bias

F 22 RMSLimit :

: in same direction as+ F

NERC’s Control Performance Standard

Reducing the Standard Deviation Bandwidth to Reduce the Area/Probability under the Tails of the Distribution

-0.0001

0.0029

0.0059

0.0089

0.0119

0.0149

0.0179

0.0209

0.0239

-0.2

50

-0.2

00

-0.1

50

-0.1

00

-0.0

50

0.0

00

0.0

50

0.1

00

0.1

50

0.2

00

0.2

50

FrequencyError(Hz)

Pro

ba

bili

ty

NormalDistribution4SD

NormalDistribution

28

3. Operating models (cont.d)• Generation (cont.d)

– Frequency control. Dispatch of power generation held as “reliability reserve” (cont.d)

• Deterioration in North American frequency response performance – Specifically

• upward drift in frequency to near the control limits (SEE NEXT SLIDE)• persistent imbalance accumulations by system participants

– Due to no “price” for unscheduled power• no cost or penalty assessment to the causers of imbalance (users of somebody

else’s reliability reserve)• no reward to the remediators of imbalance (holders of reliability reserve).

– Result• Cost is socialized in fixed transmission fees paid by consumers• Imbalance increases because it is free power to the entity who produces it or

consumes it.

10

in The New York Times, August 20, 2003:

Upward slopes pay for their slopes to the Downward slopes and alllines’ slopes add up to zero slope of horizontal line. Frequency

Contribution Component always clears.

The 17 Frequency-Contribution-Com-ponent lines of the 17 colored Control

Areas on an actual Interconnected system

-.03 .04

Measurement of an Interconnection’s FCCs over an 11-Day PeriodCombined scatters of each

colored CA’s 264 pairs(points) of hourly inadvertent

and hourly averagefrequency error

MW of Inadvertent Interchange

Bad quadrant:Inadvertent and frequencyerror in the same direction.

Bad quadrant:Inadvertent and frequencyerror in the same direction.

Good quadrant:Inadvertent and frequencyerror in opposite directions.

Good quadrant:Inadvertent and frequencyerror in opposite directions.

h,orange10

Hz of hF

31

Frequency Response Measure for BA i: MWi/f. MWi is response provided by i.

Decomposition of Near-Normal Distribution of Frequency Error into Normal Distribution of Normal Errors, & Back-to-Back Chi-Square Distribution of Events.

-0.0001

0.0029

0.0059

0.0089

0.0119

0.0149

0.0179

0.0209

0.0239

-0.2

50

-0.2

00

-0.1

50

-0.1

00

-0.0

50

0.0

00

0.0

50

0.1

00

0.1

50

0.2

00

0.2

50

FrequencyError(Hz)

Pro

ba

bili

ty

Back-to-Back Chi-Square Distributions

NormalDistribution

33

Distribution/Decomposition of Frequency Response Performance/Responsibility

34

4. Challenges in grid configuration and management• Radial versus networked. Active power versus reactive power

– A radial grid consists of remote large generation such as coal, hydro, and nuclear– A networked grid is a mix of

• remote generation to take advantage of trade, and• local, typically gas-fired generation to provide

– “peaking” plants to serve infrequent peak load at low capital cost at a high enough electricity price to pay the capital cost (SEE NEXT SLIDE)

– local reliability reserve to deploy to• meet local disturbances/contingencies without • --disturbing the systemwide powerflow by deploying remote reserve • generation, and thereby• --making a local disturbance rapidly spread to collapse a wide area • of the power system and• --adding a second possible contingency (loss of transmission) to a• single contingency (loss of generation accessed by that• transmission) (SEE SLIDE)

35

30 %

Electric systems based on marginal-cost pricing use low-capital-cost gas turbines to most efficiently meet peak load. High natural gas prices just lower the “capacity factor”

which is how long the plant can operate before the coal plant becomes cheaper.

Load

High gas prices just rotate the GTline upward around the pivot pointof intersection with the price axis

Peak, short-duration load servedby gas turbines

G

L

G

L

Sudden local generator loss

G

L

RISO

G

L

RISO

L

RISOG

Build transmission

Congested transmission

ISO deploysresponsive reserve

from big centralsource

Sudden local generator loss

Rlocal

Rlocal

Local BalancingMay Require Less

Transmission

Centralized BalancingMay Require More

Transmission

G

L

Sudden remote generator loss

Rlocal

G

Local BalancingAuthority

deploys localresponsive

reserve

Local responsivereserve still

available to thesystem

LWithout congesting transmission

37

4. Challenges in grid configuration and management (cont.d)• Radial versus networked. Active power versus reactive power (cont.d)

– A networked grid is a mix of (cont.d)• local, typically gas-fired generation to provide (cont.d)

– local reliability reserve to deploy to (cont.d)• provide a source of “reactive” power to prevent reduction of transmission capacity due to the

import of “active” power.• --Reactive power supports voltage and is determined by the phase-• shift between alternating voltage and alternating current. It• ----can travel only a short distance (150 km)• ----affects the electrical capacity of transmission lines, and • ----must be provided locally by capacitors, synchronous condensers, • generation or extra unused transmission capacity.• --Active power is scheduled energy. • ----Too much long-distance power flowing from the Midwest to the • Northeast on a grid network not designed for long-distance • power flow was a factor in the US Northeast Blackout of 2003• ----This fact was expressly omitted in the final DOE blackout report • on the basis that long-term power transactions are economics, • not reliability.

38

4. Challenges in grid configuration and management (cont.d)

• Integrating the natural-gas pipeline network with the electricity grid– A robust open-access natural-gas pipeline grid

• stabilizes an electricity grid by providing the natural gas to fire local gas-fired power plants

• dramatically reduces coal-mining accidents by enabling the sale and distribution of all coal-bed methane gas before coal mines are dug.

– The presence or not of local gas-fired electric power affects the configuration of the electric grid providing remote power

– A gas pipeline grid and the electric transmission grid are complements to each other, not competitors of each other

39

4. Challenges in grid configuration and management (cont.d)• Interconnection or isolation

– US transmission interconnections between integrated utilities, and eventually control areas

• were originally entirely reliability reserve (begun in the late 1920s), then• became used for economic transactions (after WWII). When the concepts of

economics and reliability were separated, an unused portion of capacity was assigned for reliability/contingencies

– Advantages & disadvantages of interconnection • Interconnected power systems

– benefit from the • economics of long-distance power trade, and• the reliability benefit of frequency support among the interconnected regions, but

– suffer from vulnerability that a local disturbance can spread to become a collapse of an entire wide-area system

40

4. Challenges in grid configuration and management (cont.d)

• Interconnection or isolation (cont.d)– Advantages & disadvantages of interconnection (cont.d)

• Non interconnected systems– are completely robust to disturbances from neighbors– do not benefit from gains from trade between systems

• The world financial and economic system is experiencing this– China’s banking system was

• largely immune from the financial crisis because not integrated into the global financial system, but

• did not benefit from the previous advantages of financial integration– China’s economy is

• affected by the economic recession because integrated into the world trade system, and• threatened by protectionism by other countries to substitute their increased national

production/jobs for imports/jobs from China.

• DC ties provide – the economic advantages of interconnection, but– not the reliability advantages or disadvantages of interconnection

41

4. Challenges in grid configuration and management (cont.d)

• Congestion management– Transmission congestion is efficiently managed only on an

economic basis through market-based • long-term transmission contracts/rights and • spot locational marginal pricing of purchased power

– Transmission congestion pricing provides an objective economic basis for eliminating or not the congestion bottleneck by building more transmission or building more generation on the expensive side of the constraint

– For an efficient power grid, price-based congestion management applies to railroad transportation, a key means of transporting coal to power plants.

42

4. Challenges in grid configuration

and management (cont.d)• Centralized versus decentralized control

– Transmission congestion management is best centralized into a single control-room/area because actions have a systemwide effect on powerflow and locational prices

– Frequency control is best decentralized because the control error by a single central control center is greater than the combined errors of multiple control-centers which cancel each other out

• This was evident when the ERCOT (Electric Reliability Council of Texas) centralized frequency control into a single control center and frequency performance deteriorated

• There is an unfortunate trend in the US to centralize frequency control into ever larger ISOs (Independent System Operators) which operate centralized spot markets for pricing congestion. The 2003 Northeast Blackout originated in the hastily-organized, largest and newest of those, the Midwest ISO.

43

4. Challenges in grid configuration and management (cont.d)

• Adequate reserve generation– “Economic” reserve can be handled by a robust market whose

participants use their own market-based planning models – Reliability reserve is driven by a system “requirement”

enforceable by a penalty since reliability is a “public good” like clean air. The requirement can be

• a direct reserve requirement not directly relatable to – operating performance, and – causation of cost

• or a(n operating) performance requirement – whereby the entity decides the level of reserves/risk to bear in order to perform

within the targeted performance requirement, and – can be directly related to cost causation.

44

4. Challenges in grid configuration and management (cont.d) • Market pricing to address the environment

– Market pricing of energy • curtails energy consumption when prices are high because demand is too great• increases energy consumption when prices are low because the economy is

depressed• is called Demand Response, or Demand-Side Management

– Global Warming and renewables have displaced other environmental concerns, such as with hydro and nuclear

– The reliability cost of the frequency instability caused by wind and solar • is not being included in their economic cost• needs to be measured and allocated directly to the specific wind and solar

generators and paid to the providers of frequency support• instantaneous systemwide governor response cannot be self-provided

economically by a wind or solar renewable generator.

45

4. Challenges in grid configuration and management (cont.d)

• Market pricing to address the environment (cont.d)– Global warming and renewables are likely to be

ignored if a sustained global recession takes hold. A likely at least 5 % drop in global GDP for at least a year meets the global warming reduction goal of 1 % GDP reduction for the next 5 years!

46

4. Challenges in grid configuration and management (cont.d)• Smart Grid does not by itself solve the hardest grid configuration &

management challenges. Smart Grid– provides uiseful tools like

• FACTs devices to route power to avoid congestion• Frequency responsive hot-water heaters that nevertheless cannot provide

instantaneous frequency response• Batteries to drain and discharge variation excesses and deficits caused by wind

power. – does help solve the challenge of limited US grid expansion– does not address cost causation and allocation for these devices, nor least

cost.– should not be used to subsidize and socialize the costs of devices and to

override attempts to market price and allocate those costs. For example, • instead of congestion cost being borne by the parties causing it, and• without comparing the cost of the FACTS device with the cost of other remedies, such

as increasing transmission capacity or adding generation.

47

5. Challenges for China

• Greater efficiency is needed in the resources sector of the economy, including electric power (SEE NEXT SLIDE)

• Greater transparency is needed in the models used to plan and operate the grid, to – enable providers to propose hardware and software

solutions and not just respond to requests from central grid management

– enable market participants to better forecast, plan and manage risk

48http://ihome.ust.hk/~socholz/China-productivity-measures-web-22July06.pdf

49

5. Challenges for China (cont.d)• End below-cost and below-market price regulation

– Below world-commodity-market-cost based pricing, such as for electricity, coal and oil-products, prompts excess Chinese consumption that

• pushes world prices higher and only hurts China because China has now become a net coal importer, not just oil importer (SEE NEXT SLIDES)

• creates unnecessary environmental problems.

– Consequently, China imposes administrative demand reduction measures to compensate for the adverse environmental effect of too-low prices, when the simplest solution is

• eliminate the below-market, below-cost pricing• subsidize poor people directly by giving them cash not related to electricity usage.

They will consume less electricity and use the cash for something else.

– Begin demand-side pricing/bidding in the electric power market, now that world energy prices were lower for a while.

50

China's Coal Imports and Exports, 2002-2007

Source: National Bureau of Statisticshttp://www.researchinchina.com/report/UploadFiles_8547/200708/20070802151002331.gif

51Source: International Energy Agency World Energy Outlook 2007

52

According to the Energy Watch analysis, world coal production will peak in around 2025. In that case output would undershoot official forecasts from the International Energy Agency’s World Energy

Outlook (WEO) by a substantial margin. Source: (1) Energy Watch Group, of scientists led by the German renewable energy consultancy Ludwig Bölkow Systemtechnik (LBST) , & (2) Energy Data

Associates, Dorset, UK http://www.energybulletin.net/39236.html

Coal Production

Peak production in 2025

53

5. Challenges for China (cont.d)• Build an interconnected national natural gas pipeline grid, China’s

only missing piece of world-class infrastructure (A good economic stimulus infrastructure project. Planning of the electric transmission system and the gas pipeline system should take each other into account and not fight each other. SEE NEXT SLIDES)– to provide the local power generation needed to stabilize the radial grid from

outage, especially if the grid is elevated to 1000 kV transmission, by bringing• Western gas to the East and• Eastern gasified LNG to inward areas

– to provide an energy delivery system that is immune to winter icing– to provide environmentally friendly gas power– to increase the amount of peaking generation on the system

• to improve the system economics of too much base-load generation (coal, nuclear and hydro) SEE SLIDES

• to reduce the use of peak-load shedding corresponding to too big a share of base-loaded generation in the fleet

54

3000MW

7200MW

9000MW

2000MW

Hydro Power Base

Thermal Base

AC

Regional Grids Interconnection in 2005

DC

2500MW

10000MW

3000MW

1800MW

55

3000MW

7200MW

9000MW

2000MW

Hydro Power Base

Thermal Base

AC

Regional Grids Interconnection in 2010

DC

10000MW

3000MW

1800MW2500MW

Possible International connection

56

3000MW

7200MW

9000MW

2000MW

Hydro Power Base

Thermal Base

AC

Regional Grids Interconnection in 2015-2020

DC

10000MW

3000MW

1800MW2500MW

Possible International connection

57

0 300 公里Km

Existing and planned, North & South China,

but excluding CNOOC’s proposed

Coastal Grid

58SOURCE http://www.iea.org/textbase/work/2005/LNGGasMarkets/session_5/1_Yugao_Xu.pdf

Existing and planned, North & South China,plus CNOOC’s proposed Coastal Grid

59Source: http://www.ieej.or.jp/aperc/final/ne.pdf

60Source: Investment in China’s Demanding and Deregulating Power Market, Capgemini Consulting 2005.

China’s natural-gas fired power generation capacity expected at least to exceed nuclear and new energies, and use 30-40 % of natural gas supply capacity of 100 BCF in 2010 and 200 BCF in 2020

gas-fired

61SOURCE: http://www.ieej.or.jp/aperc/pdf/GRID_COMBINED_DRAFT.pdf

62SOURCE: http://www.ieej.or.jp/aperc/pdf/CHINA_COMBINED_DRAFT.pdf

63SOURCE: http://www.ieej.or.jp/aperc/pdf/CHINA_COMBINED_DRAFT.pdf

64

5. Challenges for China (cont.d)• Build an interconnected national natural gas pipeline grid, China’s

only missing piece of world-class infrastructure (A good economic stimulus infrastructure project. Planning of the electric transmission system and the gas pipeline system should take each other into account and not fight each other.) cont.d– to eliminate coal-mining fatalities by enabling all the coal-bed methane gas

to be removed from from mines for sale and distribution to consumers. (SEE NEXT SLIDE)

– to protect areas like Guizhou from winter blackouts like 2007’s by enabling it to extract and distribute coal-bed-methane gas that

• enables mining abundant local coal to fuel abundant local coal-fired power plants and end remote coal delivery interruptible by winter icing conditions

• fuels sufficient local gas-fired reserve power generation deployed if the transmission system collapses from icing, or during winter peaking or during low rainfall/reservoir periods (SEE SLIDES)

65http://www.worldcoal.org/assets_cm/files/PDF/coalmining.pdf

66http://www.mapsofworld.com/business/industries/coal-energy/china_coal_deposits.jpg

67http://www.american.com/graphics/2007/may-june-2007/coal-in-china/China%20Map.JPG

68

http://www.platts.com/Coal/Resources/News%20Features/ctl/images/chinamap.gif

69

5. Challenges for China (cont.d)• Build an interconnected national natural gas pipeline grid, China’s

only missing piece of world-class infrastructure (A good economic stimulus infrastructure project. Planning of the electric transmission system and the gas pipeline system should take each other into account and not fight each other) cont.d– The current NPC will vote the new energy law that will make the gas pipeline

grid like the electric transmission grid by• opening access to gas pipelines to any producer or consumer under a single regulated

tariff, • by separating the pipeline operation function from the production and sales function

• Develop congestion pricing, including for railways to enable contract and supply certainty during times of congestion, such as coal delivery during winter. The 2007 winter blackout occurred partly because remote coal delivery to coal-fired power plants was interrupted by congestion.

70

Appendix

71

North American Synchro Phasor Initiative

• Funded by DOE & NERC• At first stage: providing better measurement of

– frequency transients– excessive angular separation between PMUs– voltage drop– oscillations: often precursors seen minutes or hours before a major

disturbance– MW flows– size & location of large generation trips– & other signs of grid stress

• by a Phasor Measurement Unit (PMU): high speed– 30 samples per second– versus 1 sample per 4 seconds

72

North American Synchro Phasor Initiative (cont.d)

• In order to – provide

• wide-area monitoring. Now (see slides) achieved in the – Eastern Interconnection by Real Time Dynamic Monitoring System from central data maintained

by Tennessee Valley Authority– Western Interconnection by Wide Area Monitoring System (WAMS)

• forensic analysis of grid disturbances– to trigger

• corrective action ahead of time (see slides), not just• post mortem analysis

• Data – is time-stamped to a common time reference for the entire interconnection, but– Format must be converted to conform across the interconnection

• Problem: PMUs are not “plug N play”– Not all the same output– No cookbooks: much engineering & IT time to set up

73http://www.naspi.org

74

Why Phasors?• Wide-Area and Sub-SCADA Visibility• Time Synchronization Allows us to see Dynamics not

Visible in SCADA Environment• Analysis of Major Events Typically Show Angular or

Dynamic Warning Signs Minutes to Hours Beforehand

8

4000

4200

4400

4600

0 10 20 30 40 50 60 70 80 90

4000

4200

4400

4600

Time in Seconds

Simulated COI Power (initial WSCC base case)

Observed COI Power (Dittmer Control Center)

Real event

Dynamicsimulations

No confidence in dynamic database

August 10, 1996 WSCC Outage

August 14 Angular Separation

-170-160

-150-140-130-120

-110-100

-90

-80-70-60

-50-40-30-20-10

0

15:05:00 15:32:00 15:44:00 15:51:00 16:05:00 16:06:01 16:09:05 16:10:38

Time (EDT)

Rel

ati

ve

Ph

as

e A

ng

le

Cleveland West MI

Normal Angle ~ -25º

Reference: Browns Ferry

-170-160

-150-140-130-120

-110-100

-90

-80-70-60

-50-40-30-20-10

0

15:05:00 15:32:00 15:44:00 15:51:00 16:05:00 16:06:01 16:09:05 16:10:38

Time (EDT)

Rel

ati

ve

Ph

as

e A

ng

le

Cleveland West MI

Normal Angle ~ -25º

Reference: Browns Ferry

Source: www.nerc.com

75

30min plot: 9/18/2007 MRO Event

Source: Virginia Tech FNet Data

76

Fill the Gaps

77

Current DashboardCA Independent System Operator (CAISO)

Real Time Dynamics Monitoring System (RTDMS)

78

Observable Mode Clusters

Spectral Monitoring of Select Signal

Small Signal Monitoring

Damping Ratio (%)

Mod

e Fr

eque

ncy

(Hz)

Poorly Damped ModeAlarm Threshold

<3% damping

79

Download this presentation:

• In English at http://www.blohm.cnc.net/SmartGrid.ppt

• In Chinese at http://www.blohm.cnc.net/SmartGridHanyu.ppt

Note: .ppt file name is case sensitive