WECC JSIS Update Dmitry Kosterev, BPA 1. January 15-17, 2012 Meeting WECC JSIS met on January 15-17, 2012 in Tempe, AZ 2.

• Slide 1

WECC JSIS Update Dmitry Kosterev, BPA 1 Slide 2 January 15-17, 2012 Meeting WECC JSIS met on January 15-17, 2012 in Tempe, AZ 2 Slide 3 4. Event Analysis JSIS is working with MVWG on model validation studies of July 4, 2012 under-frequency event 3 System Frequency CO Power Next step: verify the amount of generation tripped, confirm with SRP and APS Slide 4 5. Frequency Response Analysis JSIS worked with PNNL on improvements to Frequency Response tool JSIS populated Frequency Response Tool with historic events 4 Slide 5 5. Frequency Response Analysis Frequency response baseline (by date) 5 Slide 6 5. Frequency Response Analysis Frequency response baseline (by time of day) 6 Slide 7 6. Disturbance-Based Model Validation Slide 8 Test A Test B 8 Slide 9 Test A Test B Reality 9 Slide 10 Using PMU Data for Model Validation Install PMUs at power plant POIs BPA developed Power Plant Model Validation (PPMV) application using PMU data ~ Record: -POI bus voltage -POI bus frequency -Power plant MWs and MVARs Point of Interconnection VI Substation Power Plant 10 Slide 11 Power Plant Model Validation What a good models looks like: Blue line = actual recording Red line = model Voltage and frequency are inputs Active and reactive power are measures of success 11 Slide 12 Power Plant Model Validation What a bad model looks like: Blue line = actual recording Red line = model Voltage and frequency are inputs Active and reactive power are measures of success 12 Slide 13 Model Validation State PlantActive PowerReactive H1(14) H2(18) H3(16) H4(12) H5(27) T1(1)* T2(2)* T3(3) T4(1) W2 Acceptable Good Needs Major Revisions Needs Review 13 Slide 14 Experience with Disturbance-Based Model Validation Most common model issues: Power System Stabilizer models Turbine control mode of operation / governor models Generator inertia Deficiencies in model structure Other reasons for model mismatch Automatic Generation Controls Clinical experience: Plants with modern digital systems have good models that stay accurate over time Plants with legacy analog controls have most errors and tend to change in time 14 Slide 15 Model Calibration BPA has PMUs at POIs of 14 power plants About 40 monitoring points, 112 generators About 19 GW capacity More PMUs are to be added at wind sites PMU data is used to validate dynamic models of power plants: used for pass / fail checking no model adjustments are made should the model be wrong DOE is funding a research project by Bernie Lesieutre at University of Wisconsin to determine how the PMU data can be used for model calibration 15 Slide 16 Model (in)Validation Blue = actual Red = simulated Simulations done using a model from WECC dynamic data base Slide 17 Model (in)Validation Blue = actual Red = simulated Simulations done using a model from WECC dynamic data base Slide 18 Model Calibration Blue = actual Red = simulated Simulations done using a calibrated model Slide 19 Model Calibration Blue = actual Red = simulated Simulations done using a calibrated model Slide 20 Performance Monitoring Event 1 PSS OFF PSS ON Slide 21 Performance Monitoring Event 3 PSS OFF PSS ON Slide 22 Performance Monitoring Event 4 PSS OFF PSS ON Slide 23 Performance Monitoring Event 7 PSS OFF PSS ON Slide 24 6. Disturbance-Based Model Validation 24 Recommendations: Use disturbance data recordings for power plant model verification Disseminate within WECC applications for model validation and calibration using PMU data Slide 25 7. Oscillation Baselining 25 PDCI probing and Chief Joseph brake tests are performed in 2012 The test will continue in 2013, starting March 20. Three brake tests will be done. Tests are useful for calibrating Mode Meter application We can work with MVWG to use the tests for model validation More baselining is planned as WECC-wide data becomes available PCC Assignment: Develop a paper on modes of power oscillations in Western Interconnection Slide 26 Power System Models 26 Accurate Uncertain Inaccurate Transmission Lines, Transformers Generators (Machine) Excitation Systems Turbine controls Power System Stabilizers Load Models Generator Protection Slide 27 2008 vs. 2012 NS Mode A Starts at midnight Slide 28 2008 vs. 2012 NS Mode B Starts at midnight Slide 29 8. Ringdown Damping and Oscillation Detection 29 WECC JSIS prepared test cases using actual and simulated data: Disturbance ringdown (60 second files) Ambient data containing sustained oscillations, disturbances and bad data Developers and vendors were asked to show case their applications Slide 30 Montana Tech 30 Ringdown Script-driven general Prony tool Easy to repeat results Script-driven batch WECC Prony tool Specifically designed for WECC system modes Automates mode selection Designed to batch analyze MANY cases Writes results to XLS and DOC files Good signal conditioning Linear algorithm, sometimes have mode splititng Ambient Oscillation Detection Got most, not all, oscillations, got tricked by bad data Slide 31 University of Wisconsin 31 Ringdown Uses a nonlinear estimator to fit prony model Provided very good fit with a minimum model size Got good results for challenging cases when conventional linear Prony produced mode splitting. Ambient Oscillation Detection Under analysis Slide 32 Washington State University 32 Event Analysis Engine Three algorithms: Prony, Matrix Pencil and Hankel Total Least Square. Aimed at events resulting in sudden changes in damping Looks at consistency among three engines to alarm Damping Monitor Engine Ambient noise based. Continuous. Frequency Domain Decomposition Provides early warning on poorly damped modes Slide 33 Other demonstrations 33 Other demonstrations included: Alstom Grid Electric Power Group Pacific Northwest National Lab San Diego State University Presentatons and demo movies are available on WECC JSIS website Slide 34 Other Items Control Room applications Phase Angle Alarming Voltage Stability Applications Linear state estimation Oscillation Detection and Low Damping Alarming Operating Procedures Application and Operating procedure Approval process Data Management 34 Slide 35 35 0.25-Hz Mode 0.4-Hz Mode Measured during Aug. 22, 2009 PDCI probe test. Alberta Connected 15. Oscillation Mitigation Procedures 0.4-Hz Mode Slide 36 36 Alberta Disconnected 0.32-Hz Mode Measured during Sep 14, 2005 PDCI probe test.