BAL-001-TRE-01 BAL-001-TRE-01 ERCOT CPS2 R2 Waiver ERCOT CPS2 R2 Waiver Regional Variance Regional Variance March 31, 2009 March 31, 2009 Sydney Niemeyer Sydney Niemeyer
Mar 15, 2016
BAL-001-TRE-01BAL-001-TRE-01ERCOT CPS2 R2 WaiverERCOT CPS2 R2 Waiver
Regional VarianceRegional VarianceMarch 31, 2009March 31, 2009
Sydney NiemeyerSydney Niemeyer
Work ProductWork Product Draft standard consistent with SAR on which it Draft standard consistent with SAR on which it
was based.was based.
include requirements concerning frequency response contained in the ERCOT Protocols, section 5
include Requirements, Measures, and Levels of Non-Compliance sections
determining the minimum frequency response needed for reliability and requiring appropriate generators to have specific governor droop to be a more stringent practice than requirement R2 in BAL-001-0
Impact on Existing Impact on Existing StandardsStandards
Identification of any existing Standard Identification of any existing Standard that will be deleted, in part or whole, or that will be deleted, in part or whole, or otherwise impacted by the otherwise impacted by the implementation of the draft Standardimplementation of the draft Standard Protocol Section 5.9.2, 5.9.2.1, 5.9.3Protocol Section 5.9.2, 5.9.2.1, 5.9.3 Protocol Section 6.8.1.15.3 Uninstructed Protocol Section 6.8.1.15.3 Uninstructed
Charge must account for all frequency Charge must account for all frequency response.response.
Operating Guides Section 2.2.5 (deadband)Operating Guides Section 2.2.5 (deadband) Operating Guides Section 6.2.1.6 (deadband)Operating Guides Section 6.2.1.6 (deadband)
Goals of the VarianceGoals of the Variance Document the perceived reliability Document the perceived reliability
impact should the Standard be impact should the Standard be approved.approved. Improved frequency profile will increase Improved frequency profile will increase
probability that frequency will be near probability that frequency will be near 60.000 Hz when a perturbation occurs.60.000 Hz when a perturbation occurs.
Improved frequency stability even with an Improved frequency stability even with an increase in variable generation.increase in variable generation.
Support for Voltage and Automatic Voltage Support for Voltage and Automatic Voltage Regulators by supplying appropriate Regulators by supplying appropriate energy delivery during perturbations.energy delivery during perturbations.
Goals of the Variance, Goals of the Variance, continuedcontinued
If an electrical island forms within If an electrical island forms within the ERCOT Interconnection, this the ERCOT Interconnection, this variance will provide the best variance will provide the best probability of both islands surviving.probability of both islands surviving. Resources, through correct frequency Resources, through correct frequency
response, will know immediately the response, will know immediately the correct action to take to balance the correct action to take to balance the generation to the demand of each generation to the demand of each island.island.
Comparing September 2008 vs March 2008 profile of frequency in 5 mHz bins
00.010.020.030.040.050.060.070.08
Sep-08 Mar-08
Existing +/-0.036 Hz allowable deadband
ERCOT ACE +/- 237.6 MW
CPS2 10 minute average
ACE Limit = 143.2 MW (0.0217 Hz)
Due to the small amount of frequency response within the deadband, MW needed in Resource movement to return to schedule frequency is much less than ERCOT ACE.
Comparing November 2008 vs March 2008 profile of frequency in 5 mHz bins
00.010.020.030.040.050.060.070.080.09
Nov-08 Mar-08
ERCOT ACE +/- 110 MW
Proposed +/-0.0166 Hz allowable
deadband
CPS2 10 minute average
ACE Limit = 143.2 MW (0.0217 Hz)
Smaller deadband on Resources provides resistance to frequency change sooner. This results in three things 1) Overall frequency response needed from the Resource is less. 2) ERCOT ACE more closely represents the MW needed to return frequency to schedule. 3) Greater frequency stability.
High-Speed Frequency Recorder Data
59.75
59.8
59.85
59.9
59.95
60
60.05
02/09/0910:40:31
02/09/0910:40:39
02/09/0910:40:48
02/09/0910:40:57
02/09/0910:41:05
02/09/0910:41:14
02/09/0910:41:23
02/09/0910:41:31
02/09/0910:41:40
02/09/0910:41:48
02/09/0910:41:57
Date and Time (mm/dd/yy hh:mm:ss)
Freq
uenc
y (H
z)10:41:02.975
10:41:04.829
10:41:08.785
1.854 seconds to reach 59.900 Hz (3594 rpm)
5.810 seconds to reach 59.785 Hz (3587.1 rpm) from pre-event
60.025 Hz (3601.5 rpm)
59.80
59.82
59.84
59.86
59.88
59.90
59.92
59.94
59.96
59.98
60.00
60.02
60.04
60.06
60.08
60.10
13:32 13:37 13:42 13:47 13:52 13:57 14:02 14:07 14:12 14:17 14:22 14:27 14:3228200
28300
28400
28500
28600
28700
28800
28900
29000
29100
29200
29300
29400
29500
29600
29700
HZ GENMISC.ND.ERCOT.ANLG.FLD Total Generation ERCOT 60 HZ Calculated Load
12/28/08
Frequency, Load, & 60 Hz Load
59.80
59.82
59.84
59.86
59.88
59.90
59.92
59.94
59.96
59.98
60.00
60.02
60.04
60.06
60.08
60.10
13:44 13:45 13:46 13:47 13:48 13:49 13:50 13:51 13:52 13:53 13:5428600
28630
28660
28690
28720
28750
28780
28810
28840
28870
28900
28930
28960
28990
29020
29050
HZ GENMISC.ND.ERCOT.ANLG.FLD Total Generation ERCOT 60 HZ Calculated Load
12/28/08
Frequency, Load, & 60 Hz Load
786 MW Unit Tripped. Difference in Load MW = 95. Seconds after the trip, where did the other 691 MW come from?
28910 MW
28815 MW
= 95 MW difference.
59.80
59.82
59.84
59.86
59.88
59.90
59.92
59.94
59.96
59.98
60.00
60.02
60.04
60.06
60.08
60.10
13:44 13:45 13:46 13:47 13:48 13:49 13:50 13:51 13:52 13:53 13:5428600
28630
28660
28690
28720
28750
28780
28810
28840
28870
28900
28930
28960
28990
29020
29050
HZ GENMISC.ND.ERCOT.ANLG.FLD Total Generation ERCOT 60 HZ Calculated Load
12/28/08
Frequency, Load, & 60 Hz Load
If the 691 MW of frequency response from other generators is not sustained what would frequency do?
Load responded 95 MW (decreased) for a 0.13 Hz change in frequency or 73 MW/0.1 Hz.
If 409 MW of the 691 MW of frequency response was withdrawn and other generators did not respond further, frequency would reach 59.300 Hz. (firm load shed)
59.860 – 59.300 = 0.560 Hz. At 73 MW/0.1 Hz load response 73 x 5.6 = 409 MW
ERCOT has LaaR that trip at 59.700 Hz so firm load shed would not occur, but this demonstrates the importance of sustained frequency response from all Resources.
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
13:30 13:35 13:40 13:45 13:50 13:55 14:00 14:05 14:10 14:15 14:20 14:25 14:30
MW
59.76
59.78
59.80
59.82
59.84
59.86
59.88
59.90
59.92
59.94
59.96
59.98
60.00
60.02
60.04
60.06
60.08
60.10
HZ
Net MW Base Schedule HZ
MP ADecember 28, 2008
3.5 MW/0.1 Hz Frequency Response
MP A providing 5% droop with no governor deadband.This provides the Interconnection the greatest stability in frequency control.
Net Generation
105 MW Capacity Resource.
-10
-9
-8
-7
-6
-5
-4
-3
-2
-1
0
1
2
3
4
5
6
7
13:30 13:35 13:40 13:45 13:50 13:55 14:00 14:05 14:10 14:15 14:20 14:25 14:30
MW
59.76
59.78
59.80
59.82
59.84
59.86
59.88
59.90
59.92
59.94
59.96
59.98
60.00
60.02
60.04
60.06
60.08
60.10
HZ
SCE HZ
MP ADecember 28, 2008
3.5 MW/0.1 Hz Bias Setting in SCE Equation
MP A providing 5% droop with no governor deadband and Bias closely matching actual frequency response.
SCE of previous graph.
SCE
-10
-9
-8
-7
-6
-5
-4
-3
-2
-1
0
1
2
3
4
5
6
7
13:30 13:35 13:40 13:45 13:50 13:55 14:00 14:05 14:10 14:15 14:20 14:25 14:30
MW
59.76
59.78
59.80
59.82
59.84
59.86
59.88
59.90
59.92
59.94
59.96
59.98
60.00
60.02
60.04
60.06
60.08
60.10
HZ
SCE HZ
MP ADecember 28, 2008
MW/0.1 Hz Bias Setting in SCE Equation0
MP A providing 5% droop with no governor deadband and Bias NOT matching actual frequency response.
If QSE and the Resource is on AGC, control actions will reverse frequency response
SCE
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
13:30 13:35 13:40 13:45 13:50 13:55 14:00 14:05 14:10 14:15 14:20 14:25 14:30
MW
59.76
59.78
59.80
59.82
59.84
59.86
59.88
59.90
59.92
59.94
59.96
59.98
60.00
60.02
60.04
60.06
60.08
60.10
HZ
Net MW Base Schedule HZ
MP ADecember 28, 2008
0 MW/0.1 Hz Frequency Response
Resource with governor out of service and not providing frequency response.
Does not meet current Protocols, Operating Guides or this Regional Variance.
Net Generation
-10
-9
-8
-7
-6
-5
-4
-3
-2
-1
0
1
2
3
4
5
6
7
13:30 13:35 13:40 13:45 13:50 13:55 14:00 14:05 14:10 14:15 14:20 14:25 14:30
MW
59.76
59.78
59.80
59.82
59.84
59.86
59.88
59.90
59.92
59.94
59.96
59.98
60.00
60.02
60.04
60.06
60.08
60.10
HZ
SCE HZ
MP ADecember 28, 2008
MW/0.1 Hz Bias Setting in SCE Equation0
Resource with governor out of service and bias set to zero, which matches frequency response.
Does not meet current Protocols, Operating Guides or this Regional Variance and does not result in excessive SCE.
SCE
-10
-9
-8
-7
-6
-5
-4
-3
-2
-1
0
1
2
3
4
5
6
7
13:30 13:35 13:40 13:45 13:50 13:55 14:00 14:05 14:10 14:15 14:20 14:25 14:30
MW
59.76
59.78
59.80
59.82
59.84
59.86
59.88
59.90
59.92
59.94
59.96
59.98
60.00
60.02
60.04
60.06
60.08
60.10
HZ
SCE HZ
MP ADecember 28, 2008
MW/0.1 Hz Bias Setting in SCE Equation3.5
Resource with governor out of service and bias set to non zero value, which does not match frequency response.
Does not meet current Protocols, Operating Guides or this Regional Variance and will result in AGC that will “shadow” frequency response.
SCE
GOP GO
50 MW
50 MW
59.9 Hz
53.5 MW
3.5 MW/0.1 Hz
3.5 MW/0.1 Hz
3.5 MW/0.1 Hz
105 MW Capacity, 5% droop, Zero 105 MW Capacity, 5% droop, Zero DeadbandDeadband
Short fall of VarianceShort fall of Variance Does not measure Frequency Does not measure Frequency
Response at Point C (lowest Response at Point C (lowest frequency point during first 5 to 6 frequency point during first 5 to 6 seconds of the perturbation).seconds of the perturbation).
Actual & Natural SimulatedActual & Natural Simulated
In addition, I recently performed some simulations on the Texas Interconnection to investigate the effect of a delay in the delivery of Frequency Response.
First I simulated a disturbance event. The actual event is shown in Red and the simulation is shown in Blue.
Natural & DelayedNatural & Delayed
I then modified the Governor Response to deliver it on a delayed basis. In this case the additional delay was about five seconds.
It can be seen that the delay significantly affected the nature of the response.
Simulations were performed to evaluate how much this kind of delay would affect the interconnection reliability as measured by the size of a disturbance that would cause underfrequency relays to trip at 59.3 Hz.
The limit changed from 3150 MW without a delay to 1930 MW with a delay.The risk changed from one event in 250 years to about 2 events per year.This indicates that we must address the Frequency Response component of balancing. It also indicates that we
may not be able to use controllable loads unless they act as independent agents controlled locally without remote communications. This is necessary to avoid the detrimental effects of unavoidable communications delays.
Underfrequency EventUnderfrequency Event
ReferencesReferences Energy Mark: Link to Howard Energy Mark: Link to Howard
Illian’s complete presentation Illian’s complete presentation “Integrating Renewable Resources” “Integrating Renewable Resources” http://www.ece.cmu.edu/~tanddconfhttp://www.ece.cmu.edu/~tanddconf_2004/2009%20CMU%20Smart_2004/2009%20CMU%20Smart%20Grids%20Conf%20Disk/%20Grids%20Conf%20Disk/Presentations/Day%201/SessionPresentations/Day%201/Session%201/%201/