controller and power hardware-in-loop methods for accelerating

Controller and Power Hardware-In-Loop Methods for Accelerating Renewable

Energy IntegrationMichael „Mischa“ Steurer, F. Bogdan, W. Ren, M. Sloderbeck, and S. Woodruff

Center for Advanced Power Systems, Florida State University, Tallahassee, FL

• FSU/CAPS System Simulation and Power Testing Competence

• Controller Hardware-In-Loop (CHIL) and Power Hardware-In-Loop (PHIL) Simulations

• Examples of CHIL and PHIL Experiments

06/26/2007 Innovative Simulation and Prototyping Labs for Renewable Energy Integration Panel @ IEEE PES GM Tampa 2

FSU - Center for Advanced Power Systems

• Established at Florida State University in 2000 under a grant from the US Office of Naval Research (ONR)

• Focus on research and education related to application of new technologies to electric power systems

• Member of ONR Electric Ship R&D ConsortiumFSU, MSU, USC, UT-Austin, MIT, Purdue, Naval Academy/Navy Post Graduate School

Staffing:

– 36 scientific, engineering and supporting staff, including

– 9 faculty (9 mo teach and 12 mo none teach)

– 5 post docs

– 4 visiting scientists

– >20 graduate students from FAMU-FSU CoE– 34,000 square feet offices and laboratories– Facility based on grants from the US Office of Naval Research (ONR) and the US Department of

Energy (DOE): $20 million capital investment

– Annual Operating budget: $4 million

– Preparing for the the ability to perform classified research

Grant No. N0014-02-1-0623

U.S. Department of EnergyGrant No DE FG02 05CH11292


CAPS Large-Scale High-Fidelity TransientPower System Simulation

• Largest real-time digital simulator (RTDS) installation in any university, worldwide• Systems studies sized up to 250 three-phase buses at 50/2μs time steps• High-speed analog I/O to enable realistic control and power HIL experiments• Additional off-line simulation tools, i.e.: EMTDC, Matlab, PSS/E• Established expertise in understanding the details of novel and legacy power

system apparatus and their interaction with the system• Knowledge in system simulation methods, analysis, and interpretation of results


CAPS High Power Laboratories

• Power apparatus and systems laboratory up to 5 MW @ 12.47 & 4.16 kV• Integrated with the RTDS for unique power and control HIL experiments• Established expertise in advanced testing of apparatus under simulated system

conditions


Controller Hardware in Loop (CHIL) and Power hardware in loop (PHIL)

• Controller HIL Simulation– Controller under test– Low level transmitting signals (+/-15V, mA)– A/D and D/A converters are adequate for

the interface

SimulatorD/A A/D

A/D D/AController under Test

SimulatorD/A A/D

A/D D/APower Device

under Test

Power Interface

• Power HIL Simulation– Power device (load, sink) under test– High level transmitting signals (kV, kA, MW)– Power amplifiers required for interface


Why CHIL Simulations?

• Mitigate risks for all stakeholders• Will allow optimal tuning of control

parameters for fastest possible commissioning

• Will potentially reveal– Hidden issues in control algorithm and real-

time controller run-time environment– Unpredicted interactions between and

surrounding utility system (i.e. protection equipment, capacitor-bank switching, etc.)

– Unpredicted interactions between the equipment and other controlled devices (i.e. near by FACTS)

• The key for success of HIL testing is a system model that contains sufficient detail such as– Realistic model capturing system behavior– Sufficient detail of surrounding control and

protection systems– Sensor characteristics (e.g. saturation)– …

RT Simulator Hardware

NCSU Controller Hardware

GTDI

NCSU Controller Terminal

DDAC

GTAO

ODAC

GTAI

SIF


RT-HIL Simulation Based Optimization with RTDS Small Δt-Loop Capability

G

M

Motor Drive1

45MVA

36MW

G

45MVA

Port Motor

M

Motor Drive2

36MW

Starboard Motor

Gen1

Gen2

Small Δt (1.5 μs)

Large Δt (100 μs)

RTDS internal controller

Dual Δt Interface

ABB Power Electronics Controller

(AC 800PEC)

D/A

FPGA Digital InputFiring

pulses

Simulation responses

RTDS Simulation Environment

Matlab / Simulink / Real Time Workshop

RSCAD / Runtime (Optimization script)

Runtime Monitoring and Control

Parameter tuning

HIL Simulation


Why PHIL Simulations?

• Will allow highly dynamic testing of R&D prototype power apparatus– Complex operating scenarios (i.e.

faults, pulse loads, etc.) become possible in a controlled lab environment

• Will potentially reveal– Thermal management performance– Power control performance under

realistic system conditions– Problems with any details not

considered in software model of the hardware under test (HUT)

• The key for realistic PHIL testing of electrical equipment is the unique 4.16kV/6.25MVA variable AC/DC bus facility– Complete commissioning expected

3Q of 2007

5 MW HTS Propulsion Motor at CAPS during electromechanical PHIL testing

in 2004


2.5 MW

5 MW RTDS-PHIL Facility at CAPS

=

~

~

=

M2

=

~

~

=

M1

2.5MW

Dynos: 5 MW @ 225 rpm

Utility4.16 kVbus

2.5 MW

MT

Propulsion motor-drive or generator system under test

=

~

=

~

7.5MVA

12.47 kV

50 MVA 115 kV systemCity of Tallahassee

2.5MW

PHIL…Power-Hardware-in-Loop

5 MWAC-DC-ACConverter

“Amplifier”

Reproduces simulatedvoltage waveforms:4.16 kV nominal+20%/-100% 40-65 HzHarmonics up to 1.5 kHz

=

~

~

=

=

~

~

=

Experimental4.16 kV AC bus

DC bus0-1.2 kV

Equ

ipm

ent m

anuf

actu

red,

com

plet

e C

omm

issi

onin

g ex

pect

ed e

arly

2Q

200

7

Simulate Electric System Response

Simulate DynamicTorque Loads

FeedbackMeasured Data

Real TimeSimulator

Simulates electric power systems & controls in real time14 racks → 756 electrical nodes > 6000 control componentsTimer step (typical):Δt = 1.5…50μs400 digital and 200 analog I/Oenable “real-world”feedback


5 WM PEBB-based “signal amplifier”Key Specifications – AC source

• Continuous power rating 5 MW / 6.25 MVA• Full 4-quadrant operation • Output voltage range

– 0…120% (4.16 kV nominal setting), – 0…100% (8.2 kV setting, 50% power)

• Y output w/ accessible neutral• Current overload capability

– 130% 10s every 10 min– 165% 1s every min– 193% 87ms every min

• Base frequency range 40 – 65 Hz– 400 Hz possible for short time

• No-load and full load voltage THD ≤ 1%• Experimental side switching frequency

– IGBT devices 5 – 8 kHz– Twin configuration effective 10 – 16 kHz

• Filter cut-off frequency 1800 Hz

4.16 kV utility bus

4.16 / 8.2 kV experimental bus

AC Voltage reference

from RTDS

AC Current feedback to

RTDS


5 WM PEBB-based “signal amplifier”Key Specifications – DC source

4.16 kV utility bus

4.16 / 8.2 kV experimental bus

1.2 kV experimental DC bus

DC Voltage reference

from RTDS

DC Current feedback to

RTDS

• Continuous current rating 2.5 kA• Full 4-quadrant operation• Output voltage range

– 0.6…1.15 kV (1.5 – 2.8 MW)– 0…0.6 (reduced linearity)

• Bi-polar DC ungrounded• Full-load voltage regulation ≤ 1%• Closed loop voltage control bandwidth

limit ≥ 1 kHz (for 200 V step request)• Facilitates combined AC/DC operation

– 2.5 MW/4.16 kVAC and 2.5 MW/1 kVDC


16 kW PHIL of a Gas Turbine with Motor Drive

0 0.5 1 1.5 259

60

61

62

freq

uenc

y (H

z)

Generator frequency during motor load reverse

0 0.5 1 1.5 20

0.02

0.04

0.06

time (s)

rms

curr

ent (

kA) Load current (RMS) during motor load reverse

Experimental results from 50 kW PEBB system

Utilitybus208 V

15 kW @ 900 rpm

480 V

M1 M2

Experimental Bus-1@ 208 V

Load bank0-50kW

VSD1 VSD2

ABB PEBB/PECfs = 4…10 kHz, 50 kVA

RTDSRTDS

SimulatedBus-1

GGIX

IX VX

VX

SimulatedSystem


16 kW PHIL of a Motor Drive Trip after Pulse Load

Experimental results from 50 kW PEBB system

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 256

58

60

62

Freq

uenc

y (H

z)

Generator frequency response during source side pulse load

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2-0.05

00.050.1

0.15

Time (s)

Cur

rent

(kA

)

Hardware current (IA) response during source side pulse load

0.55 0.6 0.65 0.7 0.75 0.8 0.85-0.06-0.04

-0.020

0.020.040.060.08

Out of sync trippedDiode rectifier modeActive front

mode

Important: PHIL simulation remained stable resulting in correct Drive-System interaction throughout the event!


5 MW HTS Propulsion Motor Power Hardware-in-Loop Setup at CAPS

M3 Spee

d fe

edba

ck

Torq

ue

refe

renc

e

Speed reference

Torq

ue fe

edba

ck

4.16kV/60Hz utility feed

VSD

2

VSD

1

VSD

3M1M2

Gaseous He refrigeration

Exciter

Fiel

d cu

rren

t mea

sure

men

t

Real TimeSimulator

M. Steurer, S. Woodruff, T. Baldwin, H. Boenig, F. Bogdan, T. Fikse, M. Sloderbeck, and G. Snitchler, ”Hardware-in-the-Loop Investigation of Rotor Heating in a 5 MW HTS Propulsion Motor”, presented at the Applied Superconductivity Conference 2006, Seattle, WA, USA, and accepted for publication in the IEEE Trans. Applied Superconductivity


Sea-State Hydrodynamic Model

Motor (propeller)

speed (rpm)

Propeller hydrodynamic

model

Empirical torque coefficients

Empirical thrust coefficients

Ship hydrodynamic model

Ship velocity

Wave model

Random wave height

Propeller torque output

Wave velocity

Propeller velocity with

respect to water

Dynamometer torque

reference

+

+

Sea-state


127

128 SS4, Motor speed = 50%

Field Current Waveforms

126

127

130

131

SS4, Motor speed = 75%

SS4, speed = 90%

130

131SS5, Motor speed = 50%

0 20 40 60 80 100 120

130

131SS5, Motor speed = 90%

t [s]

Fiel

d C

urre

nt [A

]

Axis limits: Field current: ±1A


12:00 16:00 20:0030

30.5

31

31.5

32

32.5

33

Time [h]

SS4 , Motorspeed = 35%…75%

SS4, Motor speed = 90%

SS5

SS2

Rotor pole Temperature TP [K]

Heating from Sea-State Multi Frequency Torque Oscillations

Final value from 24 h heat run


Testing 10 MVA STATCOM for Wind Farm

01/01/04 03/01/04 04/30/04 06/29/04 08/28/04 10/27/04 12/26/0490

95

100

105

110

Vol

tage

at C

ondo

n W

ind

[%] 5 min Condon Wind SCADA Data for 365 days starting 01-Jan-2004

• Investigate feasibility and performance of NCSU’s ETO STATCOM on an existing wind farm with voltage fluctuation problems– Provide guidance to BPA & NCSU on

STACOM sizing and wind farm behavior– Test new controller in CHIL setup– Test power converter ion PHIL setup

• Investigate high-fidelity wind farm aggregation


40 turbines fully modeled, scaled by 2.075 to account for the 83 units actually in the field

40-turbine RTDS model with generic STATCOM

Fossil

DeMoss

Maupin

Condon

C1 C2

C3

C4

C5

C6

C7

C8

FSIG1Turbine 1

ModelWind

speed 1

CaT

1

CbT

1

34.5 kV

SS1

Turbine N Model

Wind speed N

CaT

N

CbT

N

SSn

N units connected

through underground

cabling

0.6 kV

TT1

TTN

ETOSTATCOM

TCW

WF

Turbine i Model

Wind speed i

CaT

i

CbT

i

SSiTTi

FSIGi

FSIGN


BPA model validation with SCADA data

0 10 20 30 40 50 60 70 80 90-20

-15

-10

-5

0

5

Vol

tage

dev

iatio

n [%

], P

ower

[MW

or M

VA

r]

Time [min]

2s Condon Wind SCADA Data on 3-5-2006

-Pcw/3Qcw(Vcw-1.0)*100

FSU received this data on 07-17-2006



0 10 20 30 40 50 60 70 80 900.92

0.94

0.96

0.98

1

1.02

1.04

1.06

Time (min)

Vol

tage

(pu)

Voltage profiles from 2s SCADA PQ injections at Condon

Vcw SCADAVcw Sim



0 10 20 30 40 50 60 70 80 900.92

0.94

0.96

0.98

1

1.02

1.04

1.06

Time [min]

Vol

tage

[pu]

Voltage profiles from 2s SCADA PQ injections at Condon

Vcw SCADAVcw Sim

Includes C-bank switching at Fossil


Model validation against 2 kHz data records

Sea West control panel(photo Steurer, Aug 2006)

0 1000 2000 30003

5

4

5

5

0 1000 2000 3000103

103.5

104

104.5

105Voltage at wind farm [pu]

time [ms]

From transient recorder (time resolution 0.5ms)

From RTDS model (time resolution 0.05ms)

C8 turns on, C1, C2, C6, C7 already onC2 turns on time [ms]

160 min

1.050

1.045

1.035

1.040

1.030

When cap switching control at Condon Wind works properly, the voltage stays within BPA criteria as FSU studies predicted- As confirmed by BPA (11/22/2006)


15 20 25 30 35 40 45 500.8

0.85

0.9

0.95

1

1.05

1.1

Power (MW)

Vol

tage

at C

W S

ubst

atio

n (p

u)

Wind Farm P-V Relationship With DeMoss Connection Open (RTDS)

0 MVAR Caps3 MVAR Caps6 MVAR Caps10 MVAR Caps12.75 MVAR Caps15.5 MVAR CapsManually Switch CapacitorsManually Switch CapacitorsWith 10 MVAR STATCOM

Impact of STATCOM with DeMoss Connection Open

…trying to maintain 1.05 pu…trying to maintain 1.00 pu…with 1.05 pu set point

Generic STATCOM model

V-control loop not optimized

Observed steady-state performance only

Caps switched manually to keep STATCOM VAR supply close to zero

Limit of operation w/ STATCOM (operating at 10MVAr)


Possible Testing of a 13.75 MVA STACOM in RTDS-PHIL Facility at CAPS

• Full power capability for 10 MVA rating

• Current regulation

• Limited system and wind farm dynamics

I0*) Requires additional line reactor for filtering and current sharing

2.5 MW =

~

~

=

STATCOM under test

=

~

=

~

~

=

Experimental4.16 kV AC busUtility 4.16 kV bus

2.5 MWSimulate Electric System Response

FeedbackMeasured Data

Real TimeSimulator

T17.5MVA

12.47 kV 50 MVA 115 kV systemCity of Tallahassee

Up to ±13.75 MVAr

± 6.25 MVAr± 7.5 MVAr

IR ≈ 0

Utility voltage reference for controls

T10.1

T10.2

Variable Voltage Source Converter - VVSI1 I2V0


Imperfect Interface Causes Simulation Errors

Vs (1V, 60Hz)

R1 (20ohm) R2 (0.1ohm)

R3 (20ohm)VO 0.055 0.06 0.065 0.07 0.075 0.08 0.085-0.2

0

0.2

0.4

0.6

time (s)

volta

ge (V

)

interface input (Vo)interface output (Vamp)

0 0.05 0.1 0.15 0.2-1

-0.5

0

0.5

1

time (s)

volta

ge (V

)

Vo of original circuitVo of PHIL circuit

Original Circuit

PHIL Implementation

Highly precise amplification

Vs VO

R1 R2

+-

R2

+-

VAMPR3V2

VFB=V2

Large error in the PHIL simulation result

Interface uses relaxation method where a common component is implanted both in hardware and in software

Simulated Hardware

W. Ren, M. Steurer, T. L. Baldwin, “Improve the Stability of Power Hardware-in-the-Loop Simulation by Selecting Appropriate Interface Algorithm”, in Proc. of the ICPS 2007 to be held in Edmonton, ALB, Canada, May 6-10 2007


• Increasing integration of power electronics and other advanced technologies into utility power systems and renewable energy systems requires improved modeling and simulation methods, tools, and expertise

• CHIL, and especially PHIL, can reveal hidden issues not well modeled in off-line simulations

• PHIL requires further research to improve overall simulation accuracy compromised by the inherent PHIL interface characteristics

• Large computational power of CAPS RTDS setup can help with model development and validation otherwise impossible with PC based simulations

Concluding Remarks


0.1 0.105 0.11 0.115

-0.1

0

0.1

0.2

0.3

time(s)

volts

Simulation result comparison

OriginalTLMDIM

0.09 0.1 0.11 0.12-1

-0.8

-0.6

-0.4

-0.2

0

0.2

time(s)

volts

Simulation result comparison

PHIL Test Case – Nonlinear Load

RL=2.0ΩVS

RS=2.0Ω LLK=0.001H

V1 C=0.01F( 1.0pu 60HZ+ 0.1pu 300Hz)

Simulator Hardware

Simulation time step: 100μs

Interface time delay: 500μs

ITM, TFA, PCD becomes unstable

LLK = 0.1mH


Lead Compensation

-150

-100

-50

0

50

100

Mag

nitu

de (d

B)

102

104

106

-180

-90

0

90

180

Phas

e (d

eg)

Bode Diagram

Frequency (rad/sec)

originalcompensatedcompensator

0.15 0.16 0.17 0.18 0.19-0.4

-0.3

-0.2

-0.1

0

0.1

0.2

0.3

0.4

time(s)

volta

ge(V

)

v1 response (linear case, RL=0.1ohm)

OriginalITMITM__comp

controller and power hardware-in-loop methods for accelerating

Documents