Voltage Management - Comprehensive Solution Requirements and Test Results

Vahid Madani, Ron Markham, Ronnie Lau, Joe Betro (PG&E)

Damir Novosel, Dino Lelic (QT)

Manu Parashar, Vijay Sukhavasi, Jay Giri (Alstom)

WECC JSIS Tempe, January 2014

Copyright ALSTOM Grid

WARNING: PG&E Proprietary Confidential Information COPYRIGHT NOTICE: This document contains information which may be proprietary and/or

confidential information of Pacific Gas and Electric Company and is intended for use only by authorized persons. Unless specifically authorized to do so, do not remove this document from Pacific Gas and Electric Company premises, make copies of this document, or distribute it to anyone other than persons authorized to use this document.

e-terravisionTM is Trademark of ALSTOM Grid, Quanta Technology (QT) is

trademark of the Quanta Technology. This document contains intellectual property of PG&E along with ALSTOM and QT. All Rights reserved.

Disclaimer

2

Voltage Instability Related Work by Others? A small sample • Sandro Corsi, and Glauco N. Taranto – IEEE

Transaction 2008 • Vijay Vittal, e.t.al - ASU, Sharma Kolluri, et. al –

Entergy –- Decision Tree Assisted Online Security Assessment using PMU measurements; PSERC Project 2008

• Venkataramana Ajjarapu - Iowa State Univ - Computational Techniques for Voltage Stability Assessment and Control – 2007 – Publisher, Springer

• IEEE and CIGRE technical committee on Stability Terms and Definitions

• A lot of work within the WECC Dynamic Modeling and Validation and Technical Study Subcommittee

• K. Vu and D. Novosel, “Voltage Instability Predictor (VIP) - Method and System for Performing Adaptive Control to Improve Voltage Stability in Power Systems,” US Patent, April 2001.

• K. Vu, M. M. Begovic, D. Novosel, and M. M. Saha, “Use of Local Measurements to Estimate Voltage-Stability Margin”, IEEE Trans., Aug. 1999

3

• J. A. Diaz de Leon, C. W. Taylor, “Understanding and Solving Short Term Voltage Stability Problems,” Proc. IEEE PES Summer Meeting 2002, Jul. 2002.

• T. Van Cutsem, C. D. Vournas, “Emergency Voltage Stability Controls: An Overview,” Proc. IEEE PES General Meeting, Tampa, Jun. 2007.

• M. Glavic, D. Novosel, E. Heredia, D. Kosterev, A. Salazar, F. Habibi-Ashrafi, M. Donnelly, “See It Fast to Keep Calm,” IEEE Power and Energy Magazine, July/August 2012.

• M. Glavic, D. Lelic, and D. Novosel, "Real-Time Monitoring of Electric Power System Voltage Stability Margins,” US Patent

• ……. • ……….

• Reactive Reserve Monitoring (RRM) or Alerts o Definition (usually offline) of reactive power zones o Real-time accounting of reactive margin in each of the zones by adding up unused

reactive power capability of generators, and summing up potential output of cap/reactor banks (those that have not been switched on yet)

o Contingency reactive power demand, which is the difference of the Reactive Reserve in the base case and after the contingency

• SIL Monitoring – A single line limit based on the value of a flow for which the reactive power production (due to capacitance of the line) equals the loss of reactive power along the line (does not take into account VAR support from the neighbors).

• Voltage Stability Assessment - SE-based system dynamic model with relevant contingencies and stress patterns to detect how far the system can be stressed o Defines a margin (e.g. corridor MW and MVAR power) as difference of the

monitored quantity at the collapse and the base case

Voltage Stability Management – Model Based

Dynamic voltage & transient stability studies o Use of time domain simulation tools Validate model correctness Include dynamic load and reactive support

device models o Study tools need to support: Contingency

Analysis, EMS/State Estimation packages; Off-line; Real-time

Comprehensive Dynamic Analysis

Worst N - 1 Contingency, Base Loador Interface Flow

Worst N - 1 Contingency, Base Load or

Interface Flow + 5%

Q

N - 0, Base Loador Interface Flow

500

200

V

100

100

Worst N - 1 Contingency, Base Loador Interface Flow +5%,Short-term Load Model

Dynamic

Static

P

Q

V Trajectory (P,Q,V)

Point of voltage Instability, or collapse

An operating point

Active power margin

Reactive power margin

Thevenin voltage could be lower that Vl (see PV for negative tan(phi))

P-V, V-Q, and PVQ Curves - Reactive Power Margin determination

• Reactive Reserve Margin (RRM) and VSA margins should get closer as the system moves toward a collapse

• Real-time tracking of the relative distance from voltage instability boundary o Distance to the PV curve nose o State Estimation based stability boundary

• Provide predictive capability • Important to validate model accuracy

Source: V&R Energy

Voltage Instability Monitoring - VSA

Source: ABB

Source: V&R Energy

Designed for fast Dynamic and Steady-State Phenomena

Accuracy comparable to model-based methods that require accurate models

• Distance of the load's apparent impedance to the Thevenin impedance (VIP, REI, RVII) o Detecting closeness to instability and local reactive power margins – Can be applied for local UVLS o Accuracy improves as closer to instability o Could trigger detailed contingency analysis

• Monitor available reactive power levels (capacitor/reactor reserves, tap-changers)

• Singular Value Decomposition & Sensitivity Analysis

• Predictive capability may be beneficial o Contingency Analysis comparison

Maximum power transfer ⇔ |Zapp | = |ZThev | Point of collapse

Z

Thev Z

app

Thevenin Load

E

Voltage Instability Tool – Measurement Based

x Under-voltage signal

#1

#2

r

Voltage instability region

#1: Inaccurate under-voltage detection

#2: Under-voltage fails to detect

8

Implementation Options – Measurement Based

EMS & Online-DSA Applications

VSAT

CA

RAS

A Hybrid Approach to Voltage Stability Assessment An integrated “MEASUREMENT-BASED” and “MODEL-BASED“ approach

Real-Time Voltage Instability Indicator (RVII)

Real-time Alerts

PMU measurements

Estimate equivalent parameters in real-time

PMUs

Real-Time Voltage Instability Indicator (RVII)

Predict HOW to respond / Advance Arming (accurate model).

Bus Voltage (V) Equiv. Imped. (Zeq) Reactive Margin (Qmargin)

Predict Qmargin changes under “worst case” contingency.

Provide recommendations on corrective actions.

MEAS

UREM

ENT-B

ASED

MO

DEL-B

ASDE

D

Comprehensive Voltage Stability Alarms Linking “Wide-Area” Low Voltage Alarms to Operator Guides in EMS

WAMS indicate the simultaneous occurrence of Low Voltage over a broad region. AND

Operator Guides in EMS

Decision making: Operator Guides (e.g. “Switch On Capacitor Banks”). • “Arm” Special Protection Schemes

Reactive Reserve (MVAR) monitoring for user-defined areas in EMS.

Detect WHEN to Take Action... and HOW to Respond!

Measurement-based voltage instability monitoring applicable to:

1. Bus (measurements at the local station)

- May be applied as UVLS tool - Requires solution to be available as part of a product (e.g: Relay or server).

2. Corridor measurements at both ends

3. Load Center all tie lines)

Reactive Margin Tool Applications

Bus

Corridor

Load Center

• Voltage stability algorithm needs to be validated: o Single line and Double-line outages o Simultaneous line outages – including adjacent systems o Comparison of results against model based solutions,

e.g. reactive reserves, load flow, contingency Analysis

• M-class data is sufficient

• When PMU data from both ends of the line are available, line impedance is calculated and validated by the program

• Q-Margins for pre-identified buses and the corridors desired for comparison

• Number of PMUs and the data needed o More measurements better accuracy, e.g. synchrophasor data from both ends of the line o Determine tuning parameters - Types of system data beyond real-time data, such as

impedances, CT / PT ratio, etc. o Initial voltage computations

Deployment Requirements and Evaluation Factors

New Q-Margin

Loading margin

Validating RVII at Proof-of-Concept (PoC)- Results Increased accuracy with improved observability when tracking Qmargin • Having additional information provides security provision in computing

reactive margin values

CASE 1: PMUs at both sending & receiving ends CASE 2: PMU at sending end only CASE 3: PMU at receiving end only

Consistent results!

Reactive Margin Tool Requirements and Validation – Flags for Improved Accuracy

For accurate calculation of Equivalent Impedance & Reactive Margin important to consider: 1. Observability – Absence of a PMU 2. Multiple iteration computation 3. Flags help

• Identify Unloaded and open ended systems • Incorporate switching or outages, e.g. line or

equipment, bypassing capacitors • Incorporate loss of a PMU data (e.g: momentary bad

network connection)

15

Polarity (power flow directions) help determining particular end of the corridor as SENDING or RECEIVING • Active and reactive powers

(aggregated over the corridor) can flow in different directions

• From PV curves o Thevenin voltage could be

lower that Vl (see PV curves for negative tan(phi))

o Eth is equal to Vl only for no-loading conditions

Reactive Margin Tool Requirements and Validation at PoC - Using Flags to Improve Accuracy (cont.)

16

Using Information in the Corridor – Calculation Steps A solid real-time reactive voltage stability monitoring solution should at minimum compute in three steps: 1. Equivalent at sending end 2. Append equivalent impedance of the corridor 3. Equivalent at receiving end using the values

obtained from sending end and the corridor as initial values (no changes in the algorithm)

4. Use computed values of the equivalent (Thevenin voltage and impedance) and append new equivalent impedance computed by considering additional measurements flows as equivalent load.

5. Ability to discriminate between changes in Thevenin voltage and Impedance in case any important change is happening in the system (outage of a line, generator hitting the limit, switching reactive power sources, etc.),

RVII Engineering User Interface (UI)

Monitored Corridors

(Sending & Recvn’g

ends)

Monitored

Busses

Real Time

PQ Curve and P,Q Operating

Point

Real Time Voltage(kV)

Real Time Equivalent Impedance and Load Impedance(p.u)

Real Time Active Power(MW)

Real Time Reactive Power(MVAR)

Input Frame Rate Display Duration

Monitored Bus

or Corridor Status/Alarms

Historical Data

Settings

Real Time Chart Display

Copyright ALSTOM Grid

User Selectable “Corridors” & Outputs

User Selectable “Buses” & Outputs

RVII Configuration Editor

• Sending Busses • Receiving Busses

Corridor

Bus Branches

Branch Parameters

Real Time Bus and Corridor Status Monitor

Real Time Status/Alarms Status bar

Copyright ALSTOM Grid

Real Time Bus and Corridor Status Monitor

Real Time Status/Alarms Status bar

Copyright ALSTOM Grid

Other Functionalities

Sub Second Zoom

Cursor Point Display

P

Q

V Trajectory (P,Q,V)

Point of voltage Instability, or collapse

An operating point

Active power margin

Reactive power margin

Voltage Stability Assessment in e-terravision (Pacific North West) Voltage Contours, MW Margins, Weak Elements, Remedial Actions

Identify weak elements (i.e. regions most prone

to voltage instability)

MW Transfer Margins

Controls

Remedial Actions

Voltage Management Conclusions • Comprehensive solution for voltage management is to use a combination of selected

methods o Each method offers some benefits, as they reflect a particular aspect of system operation,

various manifestations of instability, measurement configurations, etc.

• Real-time, model-free methods are faster compared to EMS based Systems o Good for trend and status monitoring o Should offer predictive capabilities - e.g.

wind hub prediction for reactive power support needs

• Reactive margin means different thing in different contexts

• Recognizing the Criticality & Importance of Flags

• Contingency Analysis critical to identifying vulnerabilities

• Invaluable - Importance of observability in reactive margin computation

Alarm

Source: BPA

25

1) What is the current voltage management scheme at PG&E, BPA, SCE, NV Energy,

SRP, IP…? 2) What is the current voltage management scheme at the CASIO? 3) Any voltage management systems at the RC level? 5) Is the proposed solution robust enough to implement? 6) Any integration problems as we rely on different software products? 7) How far are we from full deployment? What will it take to get there? 8) What actions and when are anticipated through use of the application.

a) Will the application take actions on its own, and if so, how would the appropriate actions be identified?

b) If actions are expected to be manually initiated, what criteria and limits should be imposed? Based on what the application is reporting?

9) Distinction between pre-contingency analysis and post contingency (or steady state) analysis, Does the application look only at the current state of the system, or does it predict the future state of the system based on contingency analysis? a) What actions are anticipated? Recognizing time frame to act may be different

for contingency analysis results versus current system state results b) Also, the same issue of 'at what limits should actions be taken' applies to both

contingency results and steady state results.

Topics for Further Discussions

Integrating Lessons Learned - FERC Report, Situational Awareness Angular Separation • Ability to determine, in real time, the

standing angles that would result following major transmission line outages

Real-Time External Visibility • Lack of adequate awareness of external

contingencies that could impact one’s system

Real-Time Tools • Phase angle difference between the two

terminals of a line after the line tripped, one should not / cannot commit to restore the line quickly.

• Having, but not using the real-time tools to monitor system conditions

Questions?

Contacts: Vahid Madani, vxm6@pge.com