Simulations and Studies Report Update David Piper, P.E. Operations Planning & Analysis SCE Grid Control Center May 22, 2019 1 Fault Voltage impact on inverters Many inverters momentarily cease current injection when voltage is outside 0.9-1.1 pu. 3 PH fault can cause widespread voltage depression below 0.9 pu. >9,000 MW of impacted inverter based resources for most impactful faults System Separation Area 2 Center Area 1 Inverter Based Resource Load Gen Load Gen VArea 1 VCenter VArea 2 Loss of inverter based resource can cause systems to separate. Distance relays may isolate the two systems. Dynamic Voltage Stability Excessive power swings can cause voltage collapse Flow on transfer path Voltage along transfer path Dangerous first swing voltage/uncontrolled loss of load (fast voltage collapse followed by overvoltage) Momentary Cessation=Energy Loss Frequency Stability The total energy loss (MW∙seconds) due to Momentary Cessation is proportional to the change in system frequency • Magnitude and duration of gen loss matters! UFLS W ide Area Studies Studies sought to identify potential issues: Instability caused by Momentary Cessation Limited reactive capability during disturbance Lack of frequency response capability or headroom Assumptions included: Stressed path flows into Southern California High renewable output Interconnection-wide summer mid-day (noon) MC assumptions: Low Voltage Threshold: 0.9 pu Recovery Delay: 0.5 s Active Current Recovery Ramp Rate: 1.0 pu/sec I nitial Study Results Two fault scenarios identified potential problems with two contingencies: PV did not recover Voltage along the Pacific AC intertie collapsed within 1.5 seconds Chicken or the Egg Scenario – Does voltage collapse cause inverters to go into MC, or does MC cause voltage collapse? Study Results Further analysis revealed that DC line blocked during the fault due to low voltage and remained blocked for 1 second DC line did not restart due to sustained low voltage associated with the initiation of voltage collapse. This response triggers existing RAS actions to protect the Pacific AC intertie Sequence of Events Fault Momentary Cessation DC line blocks due to fault voltage Fault clears normally PV Sequence of Events PACI starts to increase Voltage along PACI starts to decrease PV Sequence of Events RAS activates Alleviates loading on Pacific AC Intertie Gen PV Sequence of Events Voltage recovers DC Line Restarts PV generation recovers Gen No Voltage Collapse if M C is Elim inated Bus voltage along Pacific AC Intertie, with momentary cessation causing voltage collapse without RAS action simulated Prelim inary Conclusions Unintended interactions between MC and existing RAS Further evidence that MC can have a negative impact on system performance References Reliability Guideline – PPMV Inverter-Based Resources https://www.nerc.com/comm/OC_Reliability_Guidelines_DL/PPMV_for_Inverter-Based_Resources.pdf Reliability Guideline – Inverter Based Resource Performance Guideline https://www.nerc.com/comm/OC_Reliability_Guidelines_DL/Inverter-Based_Resource_Performance_Guideline.pdf Reliability Guideline – Modeling Distributed Energy Resources in Dynamic Load Models https://www.nerc.com/comm/PC_Reliability_Guidelines_DL/Reliability_Guideline_-_Modeling_DER_in_Dynamic_Load_Models_-_FINAL.pdf Reliability Guideline – Integrating Inverter-Based Resources into Low Short Circuit Strength Systems https://www.nerc.com/comm/PC_Reliability_Guidelines_DL/Item_4a._Integrating%20_Inverter-Based_Resources_into_Low_Short_Circuit_Strength_Systems_-_2017-11-08-FINAL.pdf Reliability Guideline – Power Plant Model Verification Using PMUs https://www.nerc.com/comm/PC_Reliability_Guidelines_DL/Reliability%20Guideline%20-%20Power%20Plant%20Model%20Verification%20using%20PMUs%20-%20Resp.pdf BPA Presentation - Power Plant Model Verification Using PMUs https://www.naspi.org/sites/default/files/2016-09/01%20-%20PPMV_NASPI%20_March_2015.pdf NERC Alert Momentary Cessation Assessment for Existing Solar Generation Resources John Schmall ERCOT Transmission Planning ROS Meeting January 10, 2019 NERC Alert • NERC Alert issued in May 2018 to recommend assessment of potential reliability risks due to adverse characteristics of solar PV resources – Assess the use and impact of momentary cessation – Assess model updates to eliminate/reduce momentary cessation • Momentary Cessation – Inverter does not inject current under voltage conditions where the resource is not allowed to trip (e.g. “no trip” zone specified in Nodal Operating Guide, Section 2.9.1) PUBLIC 2 TRE Recommendations • In order to efficiently address NERC Alert (with recommendations addressed to Transmission Planners, Planning Coordinators, Transmission Operators, and Reliability Coordinators): – ERCOT survey solar resources – ERCOT assess models and perform studies – Share results with Transmission Planners and Transmission Operators • Minimum 5-day comment period (November 16-28, 2018) – Address all comments and finalize report by December 7, 2018 PUBLIC 3 NERC Alert Survey Response • ERCOT sent out the survey to 16 solar resources (~1400 MW total) and all 16 resources responded to survey – Three resources (~200 MW total) indicated the use of momentary cessation – One resource does not have plans to eliminate momentary cessation • ERCOT staff reviewed models for all 16 solar resources – Worked with Resource Entities to confirm appropriate model response PUBLIC 4 Model Review • Individual model performance tested – Voltage profiles per ERCOT Nodal Operating Guide, Section 2.9.1 • Model adjustments confirmed with Resource Entities as necessary – Expected response with respect to momentary cessation – Protection relay model settings – Specific model parameter tuning • All reviewed models required some adjustment – Most updates reflect data corrections – Most updates not directly associated with momentary cessation PUBLIC 5 Grid Impact • ERCOT DWG 2020 HWLL case – All sixteen solar generation projects were modeled at full output – Applied a nine cycle 3-phase fault at each POI • Simulation results did not indicate instability, cascading, or uncontrolled separation associated with the use of momentary cessation • Updated models for two solar resources exhibited suspect behavior contributing to over-voltage conditions and subsequent resource tripping (~300 MW) – ERCOT will work further with the Resource Entities to improve this undesirable response PUBLIC 6 Questions? PUBLIC 7 Modeling Improvement Efforts NERC IRPTF Meeting May 22, 2019 Songzhe Zhu California ISO ISO Public NERC Alert II Generator Owner Data Submission Total MW 7198 No MC Existing MC 1233 5965 MC Can Be Eliminated 3250 MC Remaining 2714 • Data submission from 65 solar plants • 13 submission included dynamic models for existing momentary cessation settings – All have deficiencies • No dynamic models for proposed momentary cessation settings ISO Public Page 2 NERC Alert Data Follow-up • The CAISO sent letters to the owners of the solar PV plants that can eliminate or reduce use of momentary cessation. – Requested the GO to provide updated models when the changes are completed. • The CAISO received updated models for 5 plants (1400 MW). – The models are being validated. ISO Public CAISO Modeling Enhancement Efforts • Solar PV plants operational on or before September 1st, 2018 – Model requests were scheduled in phases – The schedule and model request were posted on the CAISO website • Solar PV plants achieving commercial operation after September 1st, 2018 – As-build models are required within 120 days of achieving commercial operation • Periodical updates on the test reports and models ISO Public Page 4 Model Submission Categories • Category 1 – Connecting to BES – Individual resource > 20 MVA or aggregate resource > 75 MVA • Category 2 – Connecting to 60kV and above – Individual resource > 10 MVA or aggregate resource > 20 MVA ISO Public Page 5 Model Submission Category • Category 3 – Connecting to BES or 60kV and above – Individual resource <= 10 MVA and aggregate resource <= 20 MVA • Category 4 – Connecting to below 60kV – Modeled explicitly in the planning base case • Category 5 – Connecting to below 60kV – Aggregated in the planning base case ISO Public Page 6 General Model Requirements for Category 1 through 4 • Steady state electrical characteristics and operating parameters – One-line diagram, power flow model, etc • Dynamic models – Generating units, control devices, protections, etc • Short circuit models – Subtransient reactance, transient reactance, sequence impedance, grounding data, fault currents, etc ISO Public Page 7 Dynamic Model Acceptance Criteria • Initialize without error; • A no-disturbance simulation results in negligible transients; and • A disturbance simulation results in the models exhibiting positive damping and reasonable expected performance. ISO Public Page 8 Specific Model Requirements for Category 1 & 2 • Test reports for volt/var control and active power/frequency control less than 10-year old • Active and reactive capability test report less than 5-year old • EMT model for resources close to a series compensated transmission line • Geo-magnetic disturbance data • Ride-through performance compliant with PRC-024 ISO Public Page 9 Model Submission Schedule for IBRs Due Date Phase 1 Phase 2 Phase 3 Phase 4 May 31, 2019 Oct 30, 2019 Mar 30, 2020 Aug 30, 2020 Type MW Solar PV 2730 Wind 2530 Solar PV 2070 Wind 410 Solar PV 320 Wind 760 Solar PV 1550 Wind 1090 ISO Public Page 10 Issues with TOV & Phase Jump in WTGs In TX close proximity to series compensation Background • Type 4 WTGs are located in TX with close proximity of heavy series compensation • Highlights the issue of WTG tripping due to • High TOV • DC over voltage resulting from phase jump and ROCOF Issue with high TOV It is 690V (~563V phase to ground) system, one can see the instantaneous phase to voltage reached close to 1000V (~1.8pu) at ~5.07s. Voltage wave form is very distorted. This has triggered instantaneous DC voltage protection and tripped WTG ~5.158s Fault is applied at 5s with a clearing time of 100ms One can see even after WTG is tripped the voltage wave form is very distorted and reaching voltage close to 1000V Underlying cause of Trips DC Over voltage Trips Same WTGs but two different converters (manufacturers) Phase Jump (Unbalanced L-L fault) Phase Angle Protection Flag One can see two aspects of phase jump: • Rapid change • Magnitude of change • Trip did not just depend on pre-contingent and post contingent phases but rapidly changing phase during transient at WTG terminal which is very different from POI phase information, something TOs can provide In this simulation the trip signal was disabled to force the simulation to continue. One can see following: • Converter was able to manage the initial phase jump and rapid rate of change • But couldn’t survive next one • Protection flags were raised in repeated manner when the phase kept changing over time even though the magnitude is relatively lower Frequency Estimation (erroneous) Phase Angle Phase Angle Frequency estimation based on phase angle is erroneous Frequency Protection Flag Unbalanced Fault (L-L) No tripping should be allowed Instantaneous on: 1. Frequency estimated by converter 2. Or ROCOF Frequency Protection Flag Balanced Fault (3Ph-G) Talking Points • In WTGs there is no protection on PLL phase jumps or ROCOF • However such phase jumps or ROCOF may very well results in • DC over voltage due to energy imbalance • Sometime AC over voltage • As shown in previous slide, phase jump changes rapidly at WTG terminal which is responsible for the consequential trip • TOs can provide worst case phase jump at POI in post contingent condition, however: • They won’t be able do that at WTG’s terminal • They won’t be able to predict max phase angle variation during transient at WTG terminal Thanks Simulations and Studies Report Update NERC IPRTF Meeting May 22, 2019 Songzhe Zhu California ISO ISO Public Use NERC Alert Data in the Study – Current MC Settings 14,560 MW Solar PV Yes (7,198 MW) Yes (5965 MW) Current MC Setting Use MC currently? Submitted Data? No (1,233 MW) No (7,362 MW) Default MC Setting Dyd Model from WECC Database ISO Public Page 2 Use NERC Alert Data in the Study – Proposed MC Settings ISO Public Result Summary • Instability were found under P1 and P6 contingency conditions previously with default MC settings for all solar PV resources. • The P1 contingency is stable in both current MC setting and proposed MC setting scenarios. • The P6 contingency is unstable in the current MC setting scenario but stable in the proposed MC setting scenario – P6 contingency results in isolating ~1500 MW wind generation from the grid. ISO Public Inverter Control Comparison Degradation of V as Ip takes priority over Iq Slower recovery in P due to inverter controls and recovery of V Post-fault overvoltage. Plant level V control & local coordinated Q/V control, dominated by kvi = 40; iqcmd rises to 1.3 during fault; Ipcmd drops as Iqcmd ramps up to provide reactive support. This is from Kvp parameter, with no Kqv enabled. Hence, slower type response. Q-Priority Control with Imax = 1.3, Kqv disabled P-Priority Control with Imax = 1.3, Kqv disabled ISO Public Page 5 Inverter Control Comparison No post-fault overvoltage. Kqv loop has fast proportional gain response to clamp voltage down by responding quickly with change in Iqcmd. Kqv=2 preventing overshoot since Iqcmd only rises to 0.5 during fault. Ipcmd able to rise, and only slightly clamped since voltage drop is not all that severe. Fast recovery in Pg since Iqcmd able to quickly respond and provide additional voltage support. Post-fault overvoltage. Plant level V control & local coordinated Q/V control, dominated by kvi = 40; iqcmd rises to 1.3 during fault; Ipcmd drops as Iqcmd ramps up to provide reactive support. This is from Kvp parameter, with no Kqv enabled. Hence, slower type response. Q-Priority Control with Imax = 1.3, Kqv disabled Q-Priority Control with Imax = 1.3, Vdip = 0.9, Kqv = 2 ISO Public Page 6 Study Report Updates • The study report is being drafted to address – Modeling of IBR – Reliability assessment of IBR impacts • Reliability concerns • Scenario development • Case studies • Mitigation of reliability risks – Tuning of IBR to support the grid reliability need – Target date of completion: end of July ISO Public Page 7 Unexpected Wind Farm Voltage Ride Through Performance Shun Hsien (Fred) Huang, Jeff Billo ERCOT Transmission Planning NERC IRPTF Meeting May 21-22, 2019 Background • During transmission faults, the neighboring wind generation resources experienced partial wind turbines tripped due to various causes. • The recorded transmission voltage stayed within voltage ride through (VRT) profile required by ERCOT Operating Guide. • ERCOT is working with the affected Resource Entities to identify the root of causes. PUBLIC 2 Events Description Event Fault Total WGR Tripped (MW) Direct Connect WGRs (MW) Partial Tripping in the Neighboring WGRs (MW) Potential Causes of Partial Tripping Wind Type 1 SLG, 138 kV 636 395 MW 241 MW Smart crowbar, high tower vibration 3 2 SLG, 138 kV 382 0 MW 382 MW Smart crowbar, UPS, frequency ride through, unknown 3, 4 3 SLG, 138 kV 800 391 MW 409 MW Smart crowbar, GSU setting, high tower vibration, false alarm due to voltage spike, unknown 3 Voltages stayed within the Voltage Ride Through limits in these events PUBLIC 3 Faulted Line and Tripped Turbines in Event 1 Legend: Partial Trips Disconnected --- Unit Other Wind Units Series compensated 345 kV lines Faulted 138 kV Line PUBLIC 4 Event 1: Voltage Response (345 kV) PUBLIC 5 Event 1: Recorded Frequency and Wind Generation PUBLIC 6 Issues With Smart Crowbar and UPS Failures • Unit is unable to operate outside of the voltage range 0.9 p.u. - 1.1 p.u. • Lack of VAR support to quickly recover from voltage deviations • Smart crowbar/UPS function does not affect normal operations, and therefore is not prioritized as a maintenance issue – Some turbines that tripped offline had known malfunctioning components PUBLIC 7 Corrective Actions Identified • Increasing Crowbar and UPS battery inspection to 12 months – Turbines that had recently gone through maintenance did not fail • Prioritizing Smart/Active Crowbar and UPS battery repairs • A single resource indicated there may be a software issue with the Active Crowbar function – The owner reaching out to the manufacturer PUBLIC 8 Questions? • Contact: – Jeff Billo, Jeff.Billo@ercot.com – Shun Hsien (Fred) Huang, shuang@ercot.com PUBLIC 9