pmu for smartpmu for smart grid

8/12/2019 PMU for SmartPMU for Smart Grid

1/47


2/47

KTH Royal Institute of TechnologyStockholm, Sweden

E-mail: [email protected]

Web: http://www.vanfretti.com

International Workshop:

The use of Synchrophasors in Power Systems

D 12

th

2012

Ri d J i B il

Developing Real-Time PMU Applications for

Smart Transmission Grids

Scientific Advisor

Smart Operation R&D Program

R&D Department

Statnett SF, Oslo Norway

Dr.-Ing. Luigi VanfrettiDocent and Assistant Professor

In a Real-Time Hardware-in-the-Loop Laboratory
mailto:[email protected]://www.vanfretti.com/http://www.vanfretti.com/mailto:[email protected]


3/47

Outline

Motivation:

- The paradigm shiftor whyare Smart Grids needed (also at the Power Transmission Level)

- STRONgrid project

Goals, partners and LV PMU Network

Development of Smart Grid Apps:

- SmarTS Lab: A real-time Hardware-in-the-Loop Laboratory

- Flexible Tools for Detecting and Monitoring Power SystemOscillations

Q&A


4/47

Intermitted generationbrings renewableenergy capacity

New production ofelectricity:

- Wind: variable - uncertain

- Solar: variable - uncertain

Location:

- Most new wind is located

at distant locations (Long-distance transmissionrequired)

- Most new solar sources

sited closer to the

customeron all voltage

levels (becoming bigger)

And more to come!

- In Norwaygreat potentialfor micro hydro plants atremote locations.

Small Hydro

PV

Wind


5/47


6/47

To deal with new generation sources requiresa paradigm shift also at the transmission level

Future generation patternsbringfaster power transferinteractions

It will become necessary toquickly attain a system wide

visibility andawareness

of thesystems condition

Real-time monitoring canprovide real-time visibility ofHV networks across traditionaloperational boundaries

Real-time controlcan helphandling operation understringent conditions


7/47

System dynamics become increasingly important for system operation

Flexibilitybecomes availablewith advancements in

-d d l

Need for Flexibility:

Near-voltage-collapse example

1 2 3 4Voltage

North

Center

South

1: Line tripped2 & 3:OLTC activation

4: Manual load

Shedding

Events


8/47

Smart Grids Also at the Transmission LevelMonitoring, Operation, Control and Protection in Real-Time

A challenge also in

Scandinavia:

Picture of the RegionalControl Center at Alta,Norway.

Operators have the challengeof running a power systemunder severe weatherconditions and guaranteeingsupply to critical heavyindustries


9/47

Smart Grids Also at the Transmission LevelMonitoring, Operation, Control and Protection in Real-Time

A challenge also in

Scandinavia:

Picture of the RegionalControl Center at Alta,Norway.

Operators have the challengeof running a power systemunder severe weatherconditions and guaranteeingsupply to critical heavyindustries


10/47

Develop innovative applications that will enable operation and control of the Nordic

power grid more reliably and with better information about security margins. Develop a research platform comprised by a power systems emulator (software and

hardware labs), PMUs, PDCs and specialized software.

Develop a set of software interfaces allowing PMU- data application development,and implementation

Partners:

Project Goals and Partners

Institute of Physical

Energetics

Industry

Academia


11/47

Build up from initiative by Lund University. PMUs are connected at the LV network

- Inputs: 3 phase voltage

- Outputs: Voltage magnitude, angle, frequencyand df/dt

LV PMU Network

Extension in STRONg2rid

On-line PDC

Planned PMU

On-line PMU

NI ABB SEL

SEL PDCOpen PDC

PMUs

PDCs

PRL


12/47

Build up from initiative by Lund University. PMUs are connected at the LV network

- Inputs: 3 phase voltage

- Outputs: Voltage magnitude, angle, frequencyand df/dt

LV PMU Network

Extension in STRONg2rid

On-line PDC

Planned PMU

On-line PMU

NI ABB SEL

SEL PDCOpen PDC

PMUs

PDCs

PRL

5 On-line PMUs:- KTH (Stockholm), CTH (Gothemburg), LTH

(Lund), LTH (Lule), Tampere

2 On-line PDCs:

- Lund (homebrew PDC)

- KTH (openPDC) 4 New PMUs will be installed on 2012:

- NTNU (Trondheim), DTU (Bingby), TUT(Estonia), IPE (Latvia)


13/47

The incident happened on Nov 6th 2012 at 2pm

All newspapers wrote about it:http://www.dn.se/sthlm/totalstopp-for-sl

- Fault due to bad reconnection during plannedmaintenance work in lsj (Stockholm area). Thistriggered protection on one line

- All railway systems affected (T-bana, Pendeltg andlocal trains)

- Some households (84000) without power during onehour

More stories available under #strmavbrott on

Twitter How we noticed:

- The lights blinked in KTH, some electric protectiontriggered in our Lab

Sample Event: Totalstopp fr SLOutage affecting the whole railway system in Stockholm


14/47

The fault had an impactacross the whole Nordicnetwork.

- Approx. 0.15 Hz deviation!

The grid mode frequencymeasured on the graph isapprox 0,04 Hz

- Frequency/load governingmode of the network

Measurements and preliminary analysis:The Nordic Grid is quite sensitive!

0 50 100 150 200 250

49.85

49.9

49.95

50

50.05

50.1

50.15

50.2

50.25

T i me (sec)

f

(Hz)

LT H

CT H

Tampere

LT U

K T H

The total generated capacity for the Nordic grid is roughly 50 GW.

- The lost load during the fault was 300 MW. This is less than 1% of the generation capacityin the grid.

Such a small fraction of disconnected load should not have such far reaching

impact in the entire grid.

With a larger availability of detailed and time-syncrhonized measurements (fromPMUs) spread over the grid there is a possibility of using them as inputs to controls.


15/47

SmarTS Lab

A real-time hardware-in-the-loop laboratory for WAMPAC /PMU Application Development


16/47

Smart Grid require Smart Operation, Smart Control and Smart Protection:

- The ultimate goal should be to attain an automatic-feedback self-healing control system

MeasureCommunicateAnalyze (System Assessment and reallimits)DeterminePreventive/Corrective ActionsCommunicateControl and protect

To achieve this vision, new applications need to be developed in a controlled environment, allowing

testing and considering the ICT chain

How to develop a controlled environment fordeveloping Smart Transmission Apps?

Sync. Timing

Phasor Measurement Data

Controllable and Protective Device

Measurement and Status Data

Data Alignment

and

Concentration

CommunicationNetworks

Data Storage

and Mining

Monitoring

Transmission System

Control Center

AdvancedDisplays Early Warning

System

Near Real-Time

Security

Assessment Automatic

Determination of

Control Actions

Decision/Control Support System

Smart-Automatic Control Actions

Smarter Operator

Decision Support

Smarter Operator Control

Actions


17/47

The SmarTS Lab ArchitectureOpal-RT

Real-Time Simulator

CommunicationNetwork

(WAN /NetworkEmulator)

StationBus

Low-LevelAnalog I/Os

Physical DevicesLow-Power

Prototypes

Vand IAmplifiers

Digital I/Os

Amplified analog outputs(current/voltage)

SEL PMUs /Relays

ABB Relays

NI- cRIO 9076

SynchrophasorData

WAMS

Applications

PDC(s)

WAPS

Applications

PMUs

External Controllers

Vand IAmplifiers

GOOSE

Sampled Values

GOOSE

Sampled Values

ProtectiveRelays

WACS

Applications

WAMPACApplication

Host Platform

GOOSE

ProcessBus

GOOSE

DigitalI/Os


18/47

SmarTSLa

b

Hard

wareImplem

entation


19/47

SmarTS LabComm. and Synchronization Architecture and Implementation

GPS - Signal

IRIG-B

Process Bus

(IEC 61850-9-2)

Station Bus

(IEC 61850-8-1)

Synchrophasor Bus

(IEEE C37.118)


20/47

MATLAB/SimulinkSimPowerSystems

Model

Model Splitting intoSub-systems forRT-Simulation

Real-Time ModelSimulation

RT-Lab SoftwareInterface Compiles andLoads the Model into

RT-Targets

EthernetPort

OP5949 Active Monitoring Panel

OP 5600 I/O Extension Chasis

eMEGAsim (12 Cores)

64 Analog Out

16 Analog In

Current Inputs Voltage Inputs

OPAL-RT

MATLAB/Simulink

Design models for real-

time simulation

Simulator Analog

and Digital I/Os

1

2

4

5

Ethernet

Switch

The model is compiled and loaded

into the simulator using Opal-RTLab Software

Real-Time Digital simulation is

converted to Analog / Digital Signals

through I/O s

The Analog outputs of the

Simulator are fed into the CT

Inputs of the SEL-487E

RealTime simulations are accessed from

the console generated by OPAL-RT Lab

software

OP 5251 (128

DIgital I/O))

Digital Outputs Digital Inputs

Receiving Data

from SEL 487E

using SEL

AcSELerator

Quickset

The current and voltage from the

analog outputs of the simulator

amplified by using Megger SMRT-1

Amplifier and fed into the CT/VT

inputs of the relay

3

Off-line to RT ModelCode Generation

Model-to-DataWorkflow


21/47

Model-to-Data Proof of Concept Experiment(Hardware-in-Open-Loop)

Real-TimeSimulator

Amplifier

PMU/Relay

Media Converter

SEL Synchrowave PDCHardwired Ethernet

Generate 3-Phase RTSignals (Voltages and

Currents)

Output signals, low

level throughsimulator I/Os

(+/- 16 V., +/- 20 mA)

Amplifying from low-level to standard

rating(1-60 Amp, 60-300 V)

PMU Transmits

ComputedPhasors through

Serial Port

RS232 to TCP/IPover Ethernet

l i h f i li


22/47

Megger SMRT1(Back Panel)

Ethernet Port

Data Stream on

IEEE C37.118

GPS Antenna

InputThree-Phase

Voltage &

Current Signals

Opal-RT OP5600 Computational Target

Analog Outputs from IO to

Megger SMRT1

SEL-421 Relay with PMU functionality

Monitoring OutputMeasurement

(3-Phase signals)

GPS Antenna

Megger SMRT1(Front Panel)

GPS SignalsSpliter


23/47

Modeling for Real-TimeHardware-in-the-Loop

Simulation

Video!
http://www.youtube.com/watch?v=vAywcqYYhLAhttp://www.youtube.com/watch?v=vAywcqYYhLA


24/47

What is observed at the PMU at 50 fpsreporting rate?

Generator mechanicalpower perturbation

Generator mechanical power perturbation

The whole process in real-time:Interaction with the model in real-time (Hardware-in-open-loop)

Synchronous

Generartor

500MVA , 20kV

Step Up Transformer

20kV / 380 kV

Bus 2

Bus 3

Thevenin

Equivalent

Bus 1Transmission Line

L 1-3

Transmission Line

L 1-3 a

Transmission

Line

L 3-5

M

`

Bus 4

Step Down

Transformer

380 kV / 6.3 kV

Bus 5

Induction Motor

4.9 MVA , 6.3 kV

OLTC Controlled Load

Damped oscillations captured by the PMU at 50 fps inthe frequency computed by the PMU

One-Line Diagram


25/47

A LabView PMU Application

Software Development Toolkit (SDK)


26/47

Connection mechanims for the IEEE C37.118 protocoland IEC 61850-9-5

Make PMU-data available in a buffer

Allow data access from the buffer with adjustable

update rate

Allow selection of channels

Allow receiving data on a queue

PMU App. SDKA LabView-Based PMU Application SDK


27/47

Building Smart Appsthe last mile

PMU Data Bus

(IEEE C37.118)

Phasor Data

Concentrator(s)

Real-Time

Simulator

+

Amplifiers

+

PMUs

+

Comm.

Network

000101 Parser

Labview PMU Application

Software Development Toolkit

(SDK)


28/47

SDK Platform

C37.118

PDC

DLL

Live Buffer

Access Buffer

PRL

SnapShooter

Data- Time Stamp- Voltage Phasors- Current Phasors- Frequency

Selector

Custom Application

Remote


29/47

PMU Recorder Light (PRL)

DLL

?

?

?

SnapShooter

PRL


30/47

Prototype Implementation (PMU App. SDK Beta)

Connection with PDC

Configuration

PC Loading MonitorData Channel Selection


31/47

Real-Time Data Access

Straightforward Development

of Monitoring Application


32/47

FAST REAL TIME OSCILLATION DETECTION AND

MONITORING TOOL

32


33/47

BackgroundProblem that we are aiming to address

Awareness of system dynamics is becoming increasingly important for system operationdue to uncertainties of intermittent generation:

This is an example of the dynamic operation challenges that intermittent generationbrings.

NEED for innovation: SCADA is too slow, can never capture the phenomena.

Operators can benefit from new real-time operation tools, that can allow them tomonitor, track and control these oscillations without resorting to curtail power orisolated wind eneration.

In December 2010, OG&E monitoredoscillations on the transmission system innorthwestern Oklahoma.

Occurring during periods of high wind

generation.

5% fluctuation at a frequency of 13-15 Hz.

The oscillations were product of

interactions between controllers in two

different wind farms. Countermeasures:

Switching to electrically isolate the

wind farms.

Curtail the power output!

~2sec

PMU Data


34/47

The PMUs stream data to the PDC. PDC concentrates and sends to Bable Fish, afterhand-shaking (config.), the application runs in real-time.

34

Fast Real-Time Oscillation Detectionand Monitoring Tool

PDC

PRL

PMU PMU

Fast Oscillations

RT Monitoring

Tool

Fast Oscillations

RT Monitoring Tool

Upper-level overview of

the software developed


35/47

Algorithms used: oscillation detectorsand spectral estimators.

Oscillation Detector:- A tool used to detect oscillatory activity.

- Provides fast generalized alarm of oscillation activity in real-time for a given frequency range.

- Once initialized, it requires only the latest data sampleto provide a new indication (recursive filtering )

Spectral Estimator:- A tool used to estimate the spectrum of a signal

- Uses digital signal processing methodsnon-parametric Welchs method (fft averaging), FFT and others

- It requires a larger parcel of data (few minutes, slower update)speed up by recursive data parceling.

Fast RT Oscillation Detection and Monitoring Tool Components:

- Real-time data stream display (measured variable to display to be selected by the user)

Provides real-time monitoring of the selected measurement and its corresponding overall spetrum estimation (range [0 fs/2])

- Oscillation detectors at four different and configurable ranges (low frequency inter-area modes, local modes,and fast modes (up to Nyquist freq.))

- Detection and Alarming:

Energy of the oscillation at the given frequency

Alarm levels at configurable thresholds (green = ok, orange = high activity, red = dangerous activity)

Fast Real-Time Oscillation Detection and Monitoring Tool

Fast Real Time Oscillation Detection


36/47

36

Illustration of where the algorithms are used


PDC

PRL

PMU PMU

Fast Oscillations

RT Monitoring

Tool

Real-Time Data Stream

Overall Spectrum Estimation(shows the full frequency spectrum and activity of

the oscillationsnon-parametric Welch metod)

PMUat KTHs LV.

Network, shows16.7 Hz freq. from

railway supply.

Hand-shaking (config.)with Bable Fish

Oscillation Detector and

Spectrum Estimator 1

(inter-area freqs.)



(local freqs.)

Oscillation Detector andSpectrum Estimator 3

(other freqs.)


(fast freqs.)

l ll


37/47

37

Real-Time Execution Example


PDC

BabelFish

PMU PMU

Fast Oscillations

RT Monitoring

Tool

Real-Time Data Stream

Overall Spectrum Estimation(shows the full frequency spectrum and

activity of the oscillations)

PMUat KTHs LV.

Network, shows16.7 Hz freq. from

railway supply.

Hand-shakingwith Bable Fish



(inter-area freqs.)



(local freqs.)


(other freqs.)


(fast freqs.)

Energy of the Detected Oscillation is Low becausethis is a Harmonic! (no power swings)

PMUat KTHs LV.Network, shows 16.7Hz freq. from railway

supply.

Alarms are not set off (this is correct!), andrailway frequency component is correctlydetected.


38/47

Oscillation Detectors:

- RMS Energy Computation

- Frequency bands:

- Configurable frequency range for detecting oscillations forparticular dynamic behaviour.

- Recommended frequency bands:

FB1: 0.01 Hz to 0.15 Hz, 200 sec. response time. (Slow governor-loaddynamics, i.e. balancing)

FB2: 0.15 Hz to 1.00 Hz, 12 sec. response time. (Low frequency inter-areamodes, i.e. power swings)

FB3: 1.00 Hz to 5.00 Hz, 6 sec. response time. (Low frequency intra-areaand local area modes, i.e. internal plant swings)

FB4: 5 Hz to 15 Hz, 3 sec. response time. (High frequency oscillations, i.e.

wind farm controller interactions)

FB5: 15 Hz to 25 Hz, less than 3 sec. response time. (Sub-synchronousoscillations)

38


InputsSynchrophasor

Signals from PMUs

Pre-Processing(Outlier removal,

interpolation, etc.)

Band Pass

Filter(Selection of

frequency range ofinterest)

Moving Average

Filter(Smooth out fast fluctuations

and highlight dominanttrend)

OutputRMS

Energy

Detrending(Remove quasi-steady

state variations)

Input(PMU SignalV. mag)

OutputRMS Energy

Wind Farm Oscillations

Detected at 13.4 Hz

Oscillations Detector


39/47


40/47

Spectral Estimators:

- Estimation of the spectrum of the signal

- Uses a configurable parametric YW AR method to estimate

the spectrum in each frequency band.

- Frequency bands:- Configurable frequency range for estimating the spectrum of the signal at the

given frequency range.

- Recommended frequency bandssame as with the oscillation detector

Aim: allow correlation between the input signal, its spectrum and

the oscillation frequency that is active (and detected)40


InputsSynchrophasor

Signals from PMUs

Pre-Processing(Outlier removal,

interpolation, etc.)

Band Pass Filter(Selection of frequency

range of interest)

Non-parametric

Spectral Estimation(Determine the spectrum using

YW Method)

OutputPower Spectrum

Density

40

Input(PMU SignalV. mag)

Spectral

Estimator

Wind Farm Oscillations

Detected at 13.4 Hz

Signal

d


41/47

KTH and COPPE/UFRJ have obtained a grant for collaboration from STINT

and CAPES.

We will work in implementing methods for voltage stability monitoring.

KTH will try to aid COPPE in developing their own RT lab.

The openPMU will be provided to COPPE next year for educational andresearch purposes.

We hope this collaboration can continue into the future!

Going Forward:Collaboration with COPPE/UFRJ and KTH


42/47

Thank [email protected]://www.vanfretti.com
mailto:[email protected]://www.vanfretti.com/http://www.vanfretti.com/mailto:[email protected]


43/47

Software and HardwareDemostration


44/47

Software and Hardware Demonstration

Devices:

- PMUs: openPMU, cRIO PMU (National Instruments)

- PDCs: openPDC and SEL PDC

Software:

- PMU Connection Tester- PRLPMU Recorder Light

- Real-time and off-line fast oscillation detection tool

Real-time- Wind farm models with Hardware-in-the-Loop Simulation

- HVDC Models


45/47

Developed by Queens University of Belfast

Replicated at KTH in 2012.

MeasurementPhase

EstimationTelecoms

Voltages

Currents IEEE C37.118.2SynchrophasorTransient Detection

Local Storage


46/47

OpenPMU KTHs Replica


47/47

How welldoes it work?

pmu for smartpmu for smart grid

Documents