Real Time Stability Analysis at Peak Reliability. Slaven Kincic, Hongming Zhang JSIS May 2017, SLC

Transcription

1 Real Time Stability Analysis at Peak Reliability Slaven Kincic, Hongming Zhang JSIS May 2017, SLC

2 Overview: Overview of Peak s DSA Application; o Set up o User Cases Transient Stability Criteria; TSAT Results Validation; 2

3 Overview of Peak s DSA Application 14 Western States, 1 Canadian Province, and Northern Baja California, Mexico 170,000 MW peak load 37 Balancing Authorities (BAs) 56 Transmission Operators We do have stability issues; We have full topology model representing Western Interconnection (WSM) What is relation to PMUs? 3 British Columbia Washington Oregon Nevada California Mexico Alberta Montana Idaho Wyoming Utah Colorado Arizona New Mexico

4 Overview of Peak s DSA Application Powerflow Cases EMS PTI 30 RAW File with EMS Label Contingency Data DSA Manager Dynamic Data Criteria Transfer Scenarios Online TSAT Stability Results Archive Files Stability Limit Archive Files Dynamic Model (GEN, PSS, GOV, etc) RAS Model Data and HVDC 4 4 Prod Servers cycle at 15 min for tool test and solution validation run since mid Targets Prod cutoff in late 2017.

5 Scenarios: Base-case scenario o Assessment of current system for N-1 and some credible N-2. Transfers o Source and Sink concept- to determine limit. o Same sources and sinks as in VSA. o Difference is in load model. 5

6 User Cases: 6 Units out of the step due to fault; Frequency response; Damping for N-1, N-2; RAS Interactions; RAS optimization Fast Voltage Collapse; N-1/N-2 that cannot solve in RTCA due to impossibility to accurately model RAS in power flow;

7 Transient Stability Criteria There is no continent wide criteria; Each of the area may have its own transient stability performance criteria 7 TPL-001 old

8 8 Transient Stability Criteria (TPL-001 new) There is no frequency criteria; R1:1.3 Following fault clearing, the voltage shall recover to 80% of the pre-contingency voltage within 20 seconds of the initiating event for all P1 through P7 events, for each applicable BES bus serving load. Note: it addresses only load buses R1:1.6 All oscillations that do not show positive damping within 30-seconds after the start of the studied event shall be deemed unstable.

9 Transient Stability Criteria TPL-001 changed; Anyway TPL-001 does not address how system operates; SOL Methodology Peak hosted multiple meeting with CAISO, BPA, SCE, BC-Hydro, SDGE White Paper Drafted 9

10 Peak new SOL Methodology The BES must remain transiently stable, and must not cascade or experience uncontrolled separation as described in the SOL Methodology. System frequency in the interconnected system as a whole must not trigger UFLS. Any controlled islands formed must remain stable. Real Time We added an option allowing users to specify a threshold for the maximum pre-contingency MW tripped before declaring the contingency unstable. 10

11 Peak new SOL Methodology Transient voltage or frequency dips and settling points shall not violate in magnitude and duration: 11 o Generator ride-through capabilities as specified by PRC-024-2; no BES generating unit shall pull out of synchronism (or trip) in response to transient system performance; UFLS shall not be triggered. o Nuclear plant interface requirements. o Known BES equipment trip or failure levels, e.g. surge arrestors, transformer saturation levels, generator over-excitation.

12 Real-time Monitoring o Unit ride through capability PRC

13 Real-time Monitoring-frequency UFLS settings are based on the WECC Off- Nominal Frequency Load Shedding and Restauration Plan. We want to monitor frequency in real time: 13 o We do not want to see UFLS (2617 LSDT9 and LSDT1); o We set up frequency threshold at 59.6 Hz for 0.1 s at 69 PMU telemetered buses (for alarming only); o We need to account for frequency offset

14 Real-time Monitoring-frequency We monitor AC measure (response at nadir) 14 (source Dmitry)

15 Real-time Monitoring-frequency BA Frequency Response Measure By NERC ERSTF Measure 4, calculate WECC A-B measure, A-C measure, C-B measure 15

16 Real-time Monitoring-other Per the revised Peak Reliability SOL methodology for the Operations Horizon the power, angle, voltage and/or frequency signals need to be evaluated to ensure that damping is 3% or greater for any four cycles after all RAS actions have occurred. If a TOP has additional criteria they desired to be observed, these should be implemented for the respective TOP for the real-time transient stability assessment. 16

17 TSAT Results Validation 17 Use of Distributed Swing Bus

18 Examples: 18 before after Reason: number of wind farms in WNDTGE PWmin negative (-0.4). As a result, the pitch angle limits are violated, and this causes the large rotor angle deviations, as the controllers bring the pitch angles back to within limits. After set at 0.05 pu issue was gone. Problem reported to WECC staff.

19 TSAT Results Validation We want to benchmark WSM against PMU measurements for different level of system stress: o For major transmission paths flow; o Frequency response; o Major bus voltages; RAS Models Validation We only assume the model is accurate but we really do not know until we benchmark it against PMUs 19

20 System Model Validation by TSAT 20 Validated TSAT dynamic model accuracy against recent disturbance events using PMU data (examples): PDCI RAS validated by multiple PDCI events Multiple Colstrip units tripping o Newly added ATR model was validated by the event Multiple power plant units tripping examples: o Jim Bridger, Colstrip trappings o BC Hydro gen drop o PV tripping

21 Event Sequence (Jan. X th 2017) XXX units have been tripped by RAS action in this event TSAT Log: 21

22 22 Event example:

23 Simulation Results Gen Drop and PV Unit #1 Tripping 09/07/ ~100 MW gap between PMU data and TSAT simulated xxx flow at the point of 10s after the event

24 Simulation Results Gen Drop and PV Unit #1 Tripping 09/07/

25 Slaven Kincic,