Adamantios Marinakis, Scientist, 12 th IEEE SB Power Engineering Symposium, Leuven, Enhancing Power System Operation with WAMS

Transcription

1 Adamantios Marinakis, Scientist, 12 th IEEE SB Power Engineering Symposium, Leuven, Enhancing Power System Operation with WAMS

2 Presentation Outline 1. Introduction to WAMS 2. Present WAMS applications: offered by ABB PSGuard 3. Future WAMS applications: data-based root-cause analysis and decision support Slide 2

3 Introduction to WAMS

4 Motivation From traditional to future grids Traditional grids: Centralized power generation One-directional power flows Generation follows load Operation based on historical experience Future grids: More distributed power generation More intermittent renewable power generation Some consumers become prosumers More multi-directional power flows Load adapted to production Operation based more on real-time data Need for WAMS Slide 4

5 Wide Area Monitoring Systems Basic Idea PMU I 3 U 1 Phasor Measurement Unit PMU Group Slide 5 PMU PMU PMU PMU PMU PMU Slide 5

6 Wide Area Monitoring Systems Phasor measurements High accuracy provides the basis for an accurate monitoring of power networks RES670 V.2.0 Timestamp accuracy: 1 microsecond Absolute angle accuracy error: < 0.1 degree CT/VT: % Slide 6

7 Wide Area Monitoring Systems Positioning in system operation SCADA / EMS Coordination Level Locally Network Component Protection Direct local actions by on-line status confirmation Dynamic WAMS 1 PSGuard Reaction Time IEC Coordinated measures based on dynamic view for monitoring, protection and control of power systems Monitoring at SCADA / EMS cycle rates actions initiated by long term phenomena Static Slide 7

8 Wide Area Monitoring Systems Typical architecture PMUs at substations Phasor Data Concentrators (PDC) PSGuard at the control center Data from other control centers Interface with SCADA/EMS Slide 8

9 WAMS applications

10 WAMS applications Overview Advanced visualization of raw measurements Voltage and phase angle profiles Real-time power swing display Phasor-assisted state estimation Monitoring and prediction of transmission capacity Line thermal monitoring Voltage stability monitoring Power oscillation & damping monitoring Coordination of actions in emergency situations Emergency FACTS/HVDC setpoint rescheduling Wide-area control for damping oscillations Slide 10

11 WAMS applications Visualization of raw PMU measurements VIDEO Slide 11

12 WAMS applications Event driven data archiving Wide area disturbance recorder based on logical trigger conditions Central triggering by observing network-wide data Configurable archiving length and resolution Archives are provided in CSV file format Continuous archiving provides daily archiving for long-term data storage Slide 12

13 WAMS applications Line thermal monitoring (LTM) Measurement of Current and Voltage Phasors Estimation of line resistance Determination of conductor temperature Real-time display of average temperature of conductor Patented method Slide 13

14 WAMS applications Line thermal monitoring (LTM) Slide 14

15 WAMS applications LTM: pilot installation in the Alpes Line Mettlen-Lavorgo First line to trip at the Italian blackout in 2003 Line length ~120 km Considerable differentiation of altitude over line length Slide 15

16 WAMS applications LTM: pilot installation in the Alpes Three line temperature monitoring technologies: 1) surface acoustic wave, 2) tension sensors, 3) ABB PMU-based Slide 16

17 WAMS applications Voltage stability monitoring (VSM) Assessment of distance to Point of Maximum Load ability (in MWs) Identify network equivalent Stay on top section of PV Curve! Trigger emergency actions when Power Margin too small Patented Method Slide 17

18 WAMS applications Voltage stability monitoring (VSM) Identification of three areas: generation area transmission corridor load area Strategic placement of PMUs Summation of the currents in each cut gives the two currents, i 1 and i 2 The voltages v 1 and v 2 can be computed as follows = + Slide 18

19 WAMS applications Power oscillation monitoring (POM) Damping Mode Frequency Real-time detection of power swings Algorithm is fed with selected voltage and current phasors Detection of various swing (power oscillation) modes Quickly identifies amplitude and frequency of oscillations In service since 2005 Field experience in Switzerland, Croatia, Mexico, Thailand, Finland, Norway, Austria Bus Voltage Mode Amplitude Slide 19

20 WAMS applications Power damping monitoring Ambient vs. Transient Oscillations POM detects transient oscillations PDM determines modes based on ambient variations Frequency (Hz) PDM capabilities: Use of multiple input signals Mode shape determination G1 1 G2 2 G3 3 G4 4 G5 5 G6 6 G7 7 G8 8 G9 9 G10 10 G11 11 G time/sample interval Accurate determination of damping level Simultaneous detection of multiple modes Possibility to incorporate probing signals G13 G14 G15 G16 T01 T16 T61 T62 T74 T76 T77 T86 normalised trend MW ambient ambient transient Time (sec) Slide 20

21 WAMS applications POM & PDM: actual case PSGuard at Swissgrid Major oscillation modes identified by PDM North-south mode East-west mode Slide 21

22 WAMS applications POM & PDM: actual case From ENTSO-E report: Sat. 19/02/2011 ~8:00: inter-area oscillations within the Continental Europe power system. Highest impact observed in middle-south with amplitudes of +/- 100 mhz in S. Italy, power oscillation on several northsouth corridor lines of up to +/- 150 MW & voltage oscillation on the 400 kv system of +/- 5 kv. Duration was ~15 minutes. The oscillations started and finished without direct correlation to known disturbances or forced outages. Frequency PDM output POM warning Damping (%) POM alarm at 8:02 am POM alarm condition cleared PDM was correctly identifying the presence of the poorly damped mode Slide 22

23 WAMS applications First closed-loop wide-area control in Europe Wide area power oscillation damping control (WA-POD) Chose feedback signals from any PMU equipped substation in Nordel Coordinated POD action from several actuators SVC, FACTS, generators, HVDC Prototype WACS implemented and tested Integration of PMUs with FACTS control system Wide area power oscillation damper with local signal based POD as backup First successful pilots carried out in 2011 by Statnett Slide 23

24 Data-based root-cause analysis and decision support

25 Mining WAMS data Motivation Stability indices: LTM (line thermal monitoring) VSM (voltage stability monitoring) POM (power oscillation monitoring) PDM (power damping monitoring) What we have: An operator knows in real-time the stability indices in its system The operator knows the system s security status What is additionally needed: Given a candidate operating point predict its expected security status Given an observed insecure operating point determine the reason ; modify the operating point to make it secure Slide 25

26 Mining WAMS data Outline of proposed approach WAMS PDC Stability indices alignment/ cleaning, etc. SCADA/ EMS WAMS Archive SCADA/WAMS Archive SCADA Archive Data mining / Machine Learning 1) Associate security with system state 3) Correct operating point Regression / Classification Model 2) Predict stability indices Candidate operating point Slide 26

27 Mining WAMS data Illustrative example 13k samples, produced by simulations Slide 27

28 Mining WAMS data Illustrative example Typical action taken by system operators: Reduce the intertie flow if damping ratio is below a selected value. Slide 28

29 Mining WAMS data Illustrative example: Results Feature selection objective: Find a subset of features that correlate well with the target output but have little inter-correlation. Features that are most relevant to the inter-area oscillation damping as identified by various algorithms Intertie flow PSS 63 PSS 51 Gen 63 Slide 29

30 Mining WAMS data Illustrative example: Results Model training objective: Build a model that is able to predict whether a candidate operating point is expected to correspond to a poorly damped oscillation mode. Input features Accuracy (in %) Intertie flow Intertie flow & PSS status ALL available features Slide 30

31 Closure

32 Closure Why do we need WAMS? Better observability Real-time monitoring of power systems using synchrophasors, better use of existing equipment Enhanced operation Know the limits Supervision, alarms, events Detection of incipient abnormal system conditions, reduced risk of instability, increased transmission capacity Investment protection and step-wise improvement Easy access to alarm, event list and disturbance information, integration into SCADA / EMS Real-time system monitoring Improved system planning Compare to off-line Optimization through history data analysis, data storage for enhanced planning, post-mortem analysis Slide 32

33