CURENT’s Power Electronics Based

Reconfigurable Grid Emulator

Leon M. TolbertMin H. Kao Professor

October 18, 2017

• Overview of CURENT – NSF/DOE Research Center established in

2011

• Grid Models - Large Scale Testbed

• Reconfigurable Grid Emulator – Hardware Testbed

U.S. Wind and Solar Resources

3

Best wind and solar sources are far from

load centers.

Distance provides diversity of sources.

Transmission networks must play

a central role in integration.

http://www.eia.doe.gov/cneaf/solar.renewables/ilands/fig12.html

Wind

Solar

Population

CURENT Vision• A nation-wide transmission grid that is fully monitored and dynamically controlled for high efficiency, high

reliability, low cost, better accommodation of renewable sources, full utilization of storage, and responsive load.

• A new generation of electric power and energy systems engineering leaders with a global perspective coming

from diverse backgrounds.

Multi-terminal HVDC

Monitoring and sensing

Communication

Control and Actuation

Computation

4

Power Grid

Measurement

&Monitoring

Communication

Actuation

PMU0PMU

PMUPMU0

PMUFDR

WAMS

Communication

PSS

Generator

Storage

HVDC

Wind Farm

FACTS

Solar Farm

Responsive Load

Wide Area Control of

Power Grid

What is CURENT?

5

Day Hour Minute Second Cycle

Device

Substation

Region

Balancing

Authority

Wide Area

Ultra-wide Area

AGC

LTC

AVR

UFLS

SVC

Fixed Comp.

RAS

Schemes

Unit

CommitmentEconomic

Dispatch

PSS

HVDC

Device

Protection

Today’s Operations

Some Wide Area and Some Fast but not Both

Limited communication

Minimal sensing

Traditional uncoordinated controlsDistributed coordinated actuation with

extensive measurements

6

CURENT Research Thrusts

Enab

ling

Te

chn

olo

gie

sEn

gin

ee

red

Sys

tem

sFu

nd

ame

nta

l K

no

wle

dge

ControlControl Actuation Actuation

Control Architecture

Actuator & Transmission Architecture

System-level Actuation Functions

Communication& Cyber-security

Estimation

Economics & Social Impact

Barriers· System complexity· Model validity· Multi-scale· Inter-operability

Barriers· Poor measurement design· Cyber security· Actuation & control limitation

· Barriers· Lack of wide-area control

schemes· Measurement latency· Inflexible transmission systems

MonitoringMonitoring ModelingModeling

Situational Awareness & Visualization

Wide-area Measurements

Modeling Methodology

Hardware Testbed

Large Scale Testbed

Testbeds

Control Design & Implementation

7

• Overview of CURENT – NSF/DOE Research Center established in

2011 – headquartered at The University of Tennessee

• Grid Models - Large Scale Testbed

• Reconfigurable Grid Emulator – Hardware Testbed

Engineered System: CURENT Concept

Multiple functionalities:

1. Driving requirements for the four research thrusts.

2. Testing technologies developed from the four thrusts.

3. Demonstrating an engineered system that can

• Accommodate high penetration of renewables at reasonable costs

• Fully utilize grid capacity with new system security paradigm

Systems will have elements of scaled physical

transmission network and actuation, monitoring,

control, and communication.

9

Engineered System Testbed Objectives

LTB/H

TB

Demonstrate CURENT-developed controls,

wide-area responsive load,

and wide-area renewable generation.

Provide research platforms

for testing thrust technologies,

especially modeling and control thrusts.

Study ways to increase the transmission

capability, presently constrained due to

network security considerations.

Test different power electronics

technologies and system

architectures

for improving power flow and

reliability.

Develop scenarios to evaluate

impact for high penetration of

renewable energy sources,

responsive loads, and energy

storage on the future grid.

Include real-time communication

networks, real-time control,

protection, cyber security, and actuation.

10

Engineered Systems – CURENT Concept

• High penetration of

renewable energy sources

• Flexible DC and AC

transmission

• Accommodate load and

source variability,

responsive load

• Improved situational

awareness, ultra-wide-area

control

Multi-terminal HVDC

11

MISO

PJM

ERCOT

• Large Scale Testbed (LTB):

Virtual Grid Simulator with an

Energy Management and

Control System (Matlab based

and Commercial-tool based)

• Regional and National Power

Grid Models

• Hardware Testbed (HTB):

Grid Emulator Development

and Real-time Scenario

Demonstration

CURENT Testbed Projects

12

Year 1~3

Generation I

Regional grid models

with > 20% penetration

of renewables and HVDC

connections

Model development for primary

and secondary frequency and

voltage controls in regional grids

Scaled down system models

suitable for testing in RTDS and

HTB

Scenario development to include

diverse system operating

conditions

Year 4~6

Generation II

Reduced North American system

model with > 50% penetration of

renewables and HVDC connections

Extension of frequency and voltage

control models to North American

grid and for damping control and

transient stability control

Communication system modeling

including cyber attacks

Scenario development for North

American grid

Year 7~10

Generation III

Large model of North American system

with >80% renewables and HVDC

connections

Fully integrated system model of real

time communication, coordinated

control, actuators, monitoring and load

response

Detailed scenarios for contingencies

and cyber attacks sufficient to

demonstrate resilience

LTB Demonstration Plan

13

LTB - US Grid Model Development

• Objectives:

o Develop several reduced-order models of the US interconnected power grid for use in demonstrating wide-area interconnections to transmit large amount of energy from renewable resources.

o Evaluate disturbance scenarios for the design and verification of wide-area control methodologies.

• Goals

o Developing a reduced model from a very large data set, such as the EI model

o Improving and updating the existing reduced models - WECC model

o Identify HVDC link opportunities

Multi-Terminal DC system and overlay HVDC system in NPCC system

Overlap HVDC grid for WECC system

Interconnection of WECC, EI and ERCOT via B2B HVDC

Implementation of HVDC Overlay with renewables

14

• Overview of CURENT – NSF/DOE Research Center established in

2011

• Grid Modes - Large Scale Testbed

• Reconfigurable Grid Emulator – Hardware Testbed

HTB Top-level Goals and Needs

16

• System-level outcome

Interconnected (reduced bus model) EI/WECC/ERCOT with 80% renewable, featuring HVDC overlay and

regional MTDC, fully-monitored transmission & some monitored loads, fully integrated closed-loop control on

frequency, voltage, damping, and adaptive RES for improved transfer limit and reduced reserves

• HTB

Modeling/building: reduced models of each of the 3 interconnections with variable RES levels up to 80%

Control architecture: 3-layer traditional control with central control, regional control, and local control (also

internal converter control).

Protection architecture: Local level protection.

Communication architecture: Ability to emulate power system communication

Event capability: black start with renewables, restart, normal operation, fault, scheduled/unscheduled change

of loads/sources/lines

Operation: interactive, scenario setting, visualization

Needs from: 1) actuation – real-time capability/modes, Var sources, inertia source, HVDC transmission & flow

control; 2) monitoring; 3) modeling/estimation, and 4) control

Hardware Grid Emulation System Testbed (HTB)

• Emulate various grid scenarios with interconnected clusters of scaled-down generators,

loads, and energy storage.

• Demonstrate tools developed by research thrusts.

WT II

DC

cab

le 1

DC

cab

le 2

DC cable 4

DC cable 3

VSC 4

1G1 5 6

G2 2

3

G3/WT III

1110

G44

97 8

L7 L9C7 C9

10 km25 km 25 km10 km

110 km 110 km

VSC 3

VS

C 2

VS

C 1

12 13

L12

L13

G14

14

110 k

m

66 k

m

33 km

WT I

Wind FarmWind Farm

Three-Area

System

Area 1Area 2

Area 3

Multi-Terminal

HVDC

17

Hardware Testbed Architecture and Accomplishments

Reworked hardware cabinets and software to make entire system more modular. Multiple simultaneous control functions and software hierarchy established.

Added new generation/load cabinet and transmission line cabinets so larger system can be emulated.

Battery emulator for Lead-acid and Li-ion batteries and interconnecting two-stage converter.

Virtual SG control of wind turbine converters.

Transmission line fault emulation and power system protection function emulation.

Analysis of PMU cyberattacks and multiple event detection.

Realized different WECC scenarios with high penetration renewables and HVDC overlay: MBVSA, power system separation RAS, safety supervisor based emergency control.

Generator I

Output

Inductors

Generator II

Load I

RectifierBuilding

Power

DC Bus Short Distance

Transmission

Line Emulator

Cluster 1

Cluster 2

Cluster n

Cluster n+1

Long Distance Transmission

Line Emulator

HVDC

Cluster n+2

Cluster m

Monitoring

Hardware Room

Visualization and Control Room

CTs, PTs FDR, PMU

ControlCAN Bus

7-18

HTB Hardware Expansion and Reconfiguration

Back row

• 2 cabinets for Multi-

terminal HVDC system

• 1 wind farm cabinet

19

Middle rowFront row

• Areas 1 and 4

• Transmission line cabinet 1

• Physical inductor cabinet

• PMU cabinet

• Areas 2 and 3

• Transmission line

cabinet 2

HTB Reconfiguration

• Installed a physical

inductor cabinet, allowing

quick hardware switching

between the new 4-Area

System and the previous

3-Area System

• Added input/output buses

to all cabinets,

simplifying inter-cabinet

connections

Hardware Testbed (HTB) Background

20

Power circulates within a

single area

Communication, Control, and Visualization（4 area）

Computer 1

(Area 1)

Central

Controller

Computer 2

(Area 2)

Computer 4

(Area 3)

Computer 6

(Area 4)

Station 1 Station 2 Station 3

Computer 3

(RTDS)

Visualization Room Layout Control Center functions

Area control center:

• Control local area

• Independent from each other

• Dispatch transmission lines

• Implemented with AGC, local

state estimation, voltage

monitoring, etc.

Central controller:

• Only for automatic scenario

sequencing and demonstration

• Future system level testing

Visualization computer:

• Only for display of system

information on the video wall

7-21

Station 4

Computer 5

(HVDC)

Hardware Test-bed Advantages

Broad time scales in one system - microseconds for power electronics to miliseconds and seconds for power system event.

Integrate real-time communication, protection, control, and power (and cyber security).

Multiple power electronic converters (for wind and solar and energy storage) with separate controls.

Capable of testing actual communication and measurements.

A useful bridge from pure simulation to real power system application.

22

Modular Emulators in HTB

23

Emulators

Generator Emulator Synchronous generator

Load Emulator

Induction machine

Constant impedance, constant current, and constant power load (ZIP)

Wind EmulatorWind turbine with permanent magnetic synchronous generator (PMSG)

Wind turbine with doubly-fed induction generator (DFIG)

Solar Emulator Solar panel with two-stage PV inverter

Transmission Line Emulator Back-to-back converter to emulate AC transmission lines

Energy Storage Emulator Batteries (Li-Ion, Pb-Acid, and flow), supercapacitors, and flywheels

RT Simulator Interface Integrate RTDS and LTB with HTB

HVDC Emulator Multi-terminal HVDC overlay

Combined Model Emulator Emulate combined model in single emulator

Fault Emulator Emulate short circuit faults – demonstrate system relay protections

Voltage Type

Current Type

2017

Development

Multiple Simultaneous Control Functions

24

Local control level

Regional control level

State EstimationMeasurements

Pi, Qi, Pf, Qf, V, I, θij State

Estimator

V, θ

Topology

Processor

Different

Control Block

Network

Observability

Check

Ctrl.

∆ω

∆P12

B1

ACE

s

KI

ACE Based AGC

Local Area

Frequency

DeviationPSS

To Excitation

PSS

Frequency

Difference

Between AreasWADC

To Excitation

WADC

Margin Not

Enough

Transfer Active

Power Limit

Monitoring

Reactive

Power

Support

Voltage Limit Monitoring and Control

Dispatch, Irradiance/Wind speed

Variable irradiance level, wind speed, and load

power consumption can be sent to the emulators.

Central control level

Renewable Energy Mode

Selection

Different operating modes can be

selected: MPPT, inertia emulation,

voltage regulation mode, etc.

Scheduling

Droop

1/R∆ω To Governor

To Governor

Protection

Under-voltage protection

Under-frequency protection

Over-frequency protection

Over-voltage protection

Overcurrent protection

Energy

Storage

Renewable

Energy

Generator

Visualization

Visualization

Current Controlled Emulator• Power stage description

o Three phase converter using IGBTs

o Three phase inductor as filter

o Three phase voltage controlled by paralleled inverter

o DC bus stiff and stable regulated by regenerative active regulator

o Testing structure of load emulator performance

25

Regenerative three phase current controlled emulator

test schematic

+

-

gridLoad emulator

regulating iabc

Inverter

regulating vabc

Regenerative

active rectifier

regulating Vdc

Current Controlled Emulator (cont’d)

• Controller design description

o Single closed loop current controller

o Proportional-integrator (PI) controller in dq domain

o Current references from emulated model, calculated with real-time data

o Converter dq0 frame frequency extracted from PLL

26

AUTib

ic

abc/dq

ia

ia ib ic

id iq

kp+ki /s

kp+ki /s

dd dq

da db dc

abc/dq

PLL

θ

vab

vbcvca

abc/dq

vd vq

id_ref

iq_ref

Emulated

model

+

+

-

--

+

+

+

Vdc

3ωL

Vdc

3ωL

Control structure of current controlled emulator PLL structure

vab vbc vca

Line voltage

à

Phase voltage

abc/dq

va vb vc

vq

kp+ki /s

ω1/s

θ

PI controller limit lowpass

filter

integral

+

-0

Current Controlled Emulator (cont’d)

• Emulated model description

o Input with voltage and frequency

o Output with current references

o Mathematical model fed by real time data, output calculated within IGBT

switching cycle

o Power system model dynamics are slow compared with power electronics

controller

27

vd

id_ref

iq_refEmulated

model

vq ω v0

i0_ref

External

command

Emulated model input and

output structureExample of induction motor dynamics and

detailed current controller performance

Mitigation of Unstable Harmonic Resonances in HTB

7-28

Zov1

*

1 1clvG V

+

Zov2

*

2 2clvG V

+

…

*

1 1clcG I

Yoc1

*

2 2clcG I

Yoc2

…

Connection Network

Transmission line

Voltage-controlled

converters

• Generators

• Renewables in voltage-

control mode

Non-ideal: non-passive

output impedances

Current-controlled converters

• Loads

• Renewables in current-

control mode

• HVDC

• Energy storage

Non-ideal: non-passive

output admittances

HTB: converter-based power system

Unstable harmonic resonances

Unstable harmonic resonances

Approach to Mitigation

Adjusting converter controller parameters:

• Trade-off

Increase the passivity of the converter

impedances or admittances

Reduce control bandwidth of converters

(still sufficient for emulation of power system)

G5: iG5 [10 A/div]

[time: 20 ms/div]

FFT(iG5)[Freq.: 250 Hz/div]605 Hz60 Hz

Synchronous Generator Emulator

Electrical Model

Two-axis Model:

Droop

AGC

Governor Turbine

Pm

Pe

DMs

1 ω

ωs

s1 δ

ωg

Mechanical Model

PSS

Utref

Ut

1sT

K

e

A

Efmax

Efmin

Efω

Excitation Model

Iabc

Generator

ModelVabc

Machine Terminal Voltage Control

Gate Signals

Uabc

~a

~b

~c

29

Generator Emulator Synchronization

G1

X2G2

I1

I2Load

X1

X3

G1 starts up with the load.

Transition 1

G2 connect into the system open loop by using a PLL to lock the

system frequency.

Transition 2

Enable voltage closed loop control

of G2 with PLL.

Transition 3

Alternate frequency reference from PLL output to mechanical

model output.

I2 I1ILoadTransition 1

Transition 2

Transition 3

Two generator emulators synchronized

T1 T2 T3

Current output of G1, G2, and load

during synchronization

Synchronization Process

Generator

Electrical

Model

Generator

Mechanical

Model

I2

Voltage

Controller

and PLL

V2

δr

δPLL

Vt

Pe

Vtref

δ

Frequency of G1 and G2 during

synchronization

Active power of G1 and G2 during

synchronization

30

Wind Turbine Emulator

Vabc

Wind turbine

model

Control

strategies

Gate Signals

Wind Speed

Wind power model and pitch model

Electrical and mechanical models of PMSG

Average model of the two converters

Physical Models

MPPT and reserved power control

Droop and inertia emulation

Reactive power control

Control strategies

Wind Speed

9m/s

Wind Speed

12m/s

Wind Speed

10m/s

Wind Speed

13m/s

Grid

Voltage

WT

Current1 p. u.

Detailed

Waveforms

Generator Side Converter

Grid Side Converter

Output Filter

Variable wind speed experiment

Full converter with permanent magnetic

synchronous generator

31 31

Virtual Synchronous Generator (VSG) Control of

Type-4 Wind Turbine

32

Comparison between VSG MPPT

and traditional MPPT

Machine Side Converter

controls wind turbine speed

• MPPT

• Follow grid

demands

Grid Side Converter

emulates the generator behavior

• Dispatch-able

• Variable emulated inertia

Storage Side Converter

maintains DC Voltage

Minute Level

Energy Storage

• Provide energy

buffer

Objective: Let renewable energy sources behave like the

synchronous generators in power system

VSG can track maximum power point,

and provide inertial response

Carry out dispatch command under

variable wind speed

0

2

0 200 400 600 800 1000 12000

10

20

0

0.5

1

0

0.5

1Storage SoC (p.u.)

Wind Speed (m/s)

Time (s)

Turbine Speed (p.u.)

Power (p.u.)

Transition from VSG MPPT to VSG Normal Operation

1

Generator Load Wind

Solar Power Emulator

MPPT under irradiance (S: W/m2) change

Two-stage PV inverter

Vabc

Solar power

model

Control

strategies

Gate Signals

Irradiance

i1: Generator current

i2: PV current

vg: Load voltage

PV panel model considering the irradiance and

temperature

Boost converter model

Inverter model including the LCL filter

Physical Models

MPPT and reserved power control

Droop and inertia emulation

Reactive power control

Low voltage ride through (LVRT)

Control strategies

33

ZIP Load Emulator

• ZIP load (constant impedance,

constant current, and constant

power)

Coefficients a + b + c = 1, d + e + f =1.

Ia & Ib

Input voltage

+

-

Vdc

Vabc

Load

Model

Iabc

Voltage Control

Gate Signals

Iref

Coefficients and the output current amplitude in the

ZIP load model change, which enables the load to

switch from constant power to constant

impedance/current during low voltage fault condition.

34

Induction Motor Load Emulator

Mechanical torque

Electrical model

among V, I and

flux

Mechanical

model

Torque Rotor speed

OUT:

Current

references

IN:

Converter

voltages

Torque

demand

PLL

Phase A starting up current

• Induction motor load

35

Induction Motor Emulator Startup

36

Induction Motor Emulator

ib

ic

abc/dq

ia

ia ib ic

dd dq

da db dc

abc/dq

PLL

θ

vab

vbcvca

abc/dq

Emulated

model

+

+

-

-Controller

Same input conditions

& torque command

• The comparison between simulation and experimental references results verifies the correct calculations of

emulated model in DSP.

iq_ref

id_ref

id

iq

Tem

ωr

Tem

ωr

Energy Storage Emulators

Energy Storage

Compressed AirBatteries and

Ultra-capacitorsFlywheels

37

BESS Emulator Overview

38

Technologies: Lithium Ion, Lead Acid, and Vanadium redox flow

Models: battery and two-stage power electronics interface

Control: command-based active and reactive power control

Applications: frequency control and voltage support

Point of Emulation

Internal Battery Characteristics

39

Polarization Curve

Li-Ion

Li-Ion Cell

VRB Cell

Pb-

AcidVRB

Load Loss

With IE

Without IE

Battery Energy Storage Emulator

Emulator Attributes

• Two modes of operation: Command and Inertia Emulation

• Choice between Lithium Ion, Lead Acid, Flow batteries

• Constant current – constant voltage charging algorithm

• Independent active and reactive power control

• Operates on its own HTB inverter

7-40

Inertia Emulation Control Test

Point of Emulation

Flywheel Energy Storage Emulator

Flywheel working states triggered by frequency regulation

demands:

Acceleration: area frequency high/stable, absorbing

power from grid

Standby: reserve energy in spinning kinetic form

Deceleration: area frequency sags/need power

support, supplying load

Flywheel current Generator 1 current

1G1

5 6

Flywheel

2

3 G31110

G4

4

97 8

L7 L9C7 C9

10 km25 km 25 km10 km110 km 110 km

Load 7 current

Acceleration

Standby with load

increase

Deceleration

Re-acceleration

0.9916

1.0083

With flywheel

Without flywheel

Flywheel reaches ωmin

Constant torque

acceleration Standby

Constant power

accelerationDeceleration

Deceleration mode

index

Acceleration mode index

41

Combined Model Emulator Implementation

42

ea

eb

ecn

Ia,b,c

VSG

VSG+LOAD

Load

L R ia ea

ib

ic

eb

ec

n

Ia,b,c

Generator Model

ZIP Load Model

Voltage Control

Va,b,c-ref

IZIP-ref

Gate

Signal

ud

uqid

iq

uq

udid

iq

Standalone Load Simulation

Combined Load and SG Simulation

ud

uqid

iq

uq

udid

iq

Standalone Load Simulation

Combined Load and SG SimulationStandalone load and combined load and SG emulation results comparison

(a) Standalone load emulation and (b) Combined load and SG emulation.

(a)

(b)

Bus Short-Circuit Fault Emulation

EUT

+-

P

Fault Emulator

-PI

Va*

Va

Za

Ia da

-PI

Vb*

Vb

Zb

Ib db

dc

block

Control block diagram of

double-line-to-ground fault

Realize short-circuit faults by controlling

the corresponding phase of emulator

terminal voltages to zeroVoltages of three-phase short circuit Line-to-line short circuit

Currents of three-phase short circuit Single-line-to-ground short circuit

with 1 Ω grounding resistance

43

Validation of the Developed SG Emulator

under Three-Phase Fault

0 0.02 0.04 0.06 0.08 0.1 0.12

-30

-20

-10

0

10

20

Time (s)

Phase A

Curr

ent

(A)

0 0.02 0.04 0.06 0.08 0.1 0.12

-20

-10

0

10

20

30

40

Time (s)

Phase B

Curr

ent

(A)

Simulation

Experiment

0 0.02 0.04 0.06 0.08 0.1 0.12

-30

-20

-10

0

10

20

Time (s)

Phase C

Curr

ent

(A)

0 0.05 0.1 0.15

0.999

0.9992

0.9994

0.9996

0.9998

1

Time (s)

Roto

r S

peed (

p.u

)

Simulation

Experiment

Rotor Speed

Phase A Current Phase B Current Phase C Current

Amplitude and phase

error are less than 5%

when using second

norm error for analyzing

dynamic response over

a frequency range

of 0 to 200 Hz.

6th-order synchronous generator

model with saturation effect on d-

axis

Fault through shorted lines

44

AC or DC Transmission Line Emulator

fLVDC

208V/480V

G

208V

ai

bici

480V

G

fL

P1

Q1

P2

Q2

ACI

RectifierInverter

1V 2 0V X

1G1

5 6

G2

2

3 G31110

G4

4

97 8

L7 L9C7 C9

10 km25 km 25 km10 km110 km 110 km

Area 1 Area 2

Back-to-back structure with two terminals

Line 7-9Comparison between the emulator and simulation with line

impedance change (line drop)

G1 frequencySimulation in MATLAB/Simulink

Emulator in HTB

Emulator in HTB with DC offset control

Transmission Line Emulator Attributes

• Vary the line length (impedance) for different scenarios

• Short circuit or open line faults

• Reclosing emulation

• Emulate multiple parallel lines

• Emulate FACTS applications such as CVSR

45

Transmission Line Fault Emulation

7-46

Transmission lineLline/2 Rline/2 Lline/2 Rline/2Lsource1Rsource1

Lsource2Rsource2Vsource1Vsource2

Simulation Verification

0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 0.55 0.6-100

0

100

Time(s)

Cu

rre

nt (A

)

Phase A current on source 1 side

Original

Emulation

0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 0.55 0.6-100

0

100

Time(s)

Cu

rre

nt (A

)

Phase B currenton on source 1 side

Original

Emulation

0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 0.55 0.6-100

0

100

Time(s)

Cu

rre

nt (A

)

Phase C current on source 1 side

Original

Emulation

System model

Vab1

Vbc1

Vca1

L-L to

Phase

Va1

Vb1

Vc1

Vab2

Vbc2

Vca2

L-L to

Phase

Va2

Vb2

Vc2

(Vx1-Vx2)/2Va(LR/2)

Vb(LR/2)

Vc(LR/2)

Normaln

n

n

Va1(LR/2)

Vb1(LR/2)

Vc1(LR/2)

f

f

f

Va2(LR/2)

Vb2(LR/2)

Vc2(LR/2)

f

f

f

Fault

Fault

Fault

Fault

Normal

Va2(LR/2)

Vb2(LR/2)

Vc2(LR/2)

Va1(LR/2)

Vb1(LR/2)

Vc1(LR/2)

×-1

(V-IR/2)/(L/2)

Ia1

Ib1

Ic1

Ia2

Ib2

Ic2

1/s

I

Transmission line emulation model (normal and fault condition)Mode

Selection

Transmission line emulatorLsource1Rsource1

Lsource2Rsource2Vsource1Vsource2Ia1

Ib1

Ic1

Ia2

Ib2

Ic2

Vab1

Vbc1

Vca1

Vab2

Vbc2

Vca2

Fault Recover Fault Recover

Voltage Source Converter (VSC)-Based

Multi-Terminal HVDC Overlay

7-47

DC cable 1

DC cable 3

DC cable 2

VSC 1 VSC 2

VSC 3

Summer Power Flow

DC cable 1

DC cable 3

DC cable 2

VSC 1 VSC 2

VSC 3

Winter Power Flow

• DC overlaying an AC system is suitable for transferring remote renewable energy to load centers

• Less converter numbers and potential cost benefit of multi-terminal configuration compared to

building multiple point-to-point transmission lines

• Easy power flow reverse, smaller footprint by using voltage source converter (VSC)

• DC power flow controller and DC fault protection need to be addressed

System Protection

• Inverter DSP’s have been programmed to have their own

protection based on system model parameters

• Parameters were chosen by WECC protection standards

• Automatic actions such as load shedding and generator

trips can be taken

• Loads have under-voltage and under-frequency protection

• Generators have over-current, under-frequency, and over-

frequency protection

7-48

0 5 10 15 200

0.2

0.4

0.6

0.8

1

0

t x( )

200 x

Generator Overcurrent Protection Curve

t x( )0.0515

7 x0.02

1

0.114

Under-

frequency

Over-

frequency

Tripping time

59.4 60.6 3 minutes

58.4 61.6 30 seconds

57.8 - 7.5 seconds

57.3 - 0.75 seconds

57 61.7 instantaneous

Generator under and over frequency protectionLoad under-frequency

Frequenc

y Set-

Point, Hz

Tripping

time,

Cycle

Load

Dropped,

%

59.1 14 5.3

58.9 14 5.9

58.7 14 6.5

58.5 14 6.7

58.3 14 6.7

Load under-voltage

Voltage

Set-Point,

p.u

Tripping

time, s

Load

Dropped,

%

0.9 3.5 5

0.92 5 5

0.92 8 5

Vulnerability Exploitation

The communication protocol used by

PMUs and PDCs: IEEE C37.118

Lack of user and message

authentication à Packet injection attack

Security recommendations and best

practices to be evaluated

DNP3, TLS, WPA2

Vulnerability Assessment of Phasor Networks

Routers

Ethernet

RoutersPMUs FDR UGA

Substation without PDC Routers

Ethernet

RoutersPMUs FDR UGA

Substation with PDC PDC

Super PDC

Ethernet

RoutersStorage

Control Center

Router

SuperPDCIT

Center

Ethernet

RoutersStorage

Log recorder

Router

Relay Protection

Office

Communication Backbone Network

Attaker

Phasor Communication Network

Multiple Event Analysis for HTB

Background: Existing event detection algorithms can handle single event case by exploring their abnormality,

e.g., frequency increment over a threshold indicates occurrence of load shedding. However, existing work fails

to decompose and identify multiple events or cascading events which tangle the signals together.

Objective: We propose a multiple-event analysis based on sparse coding to detect and recognize multiple,

involved single events (i.e., generator trip, line trip, and load shedding) in an online fashion.

Tasks:

1) Multiple events detection when signals from all buses are reliable.

2) Multiple events detection when there is 10% data hacked (missing signals).

Results: We conducted the experiment based on a three-area HTB system under two testing

scenarios.

3-50

Accuracy Bus1 Bus2 Bus3 Bus4 Bus5Majoriy

Voting

S1C 100 100 90 100 90 100

M2C 80 80 70 80 80 80

Table 1: Performance of testing scenario 1

Accuracy -Bus1 -Bus2 -Bus3 -Bus2, 3 -Bus4, 5 -Bus1, 2,

5

S1C 100 100 100 100 100 100

M2C 80 80 80 80 80 80

Table 2: Performance of testing scenario 2 (missing data)

Note: S1C denotes single event, M2C denotes double events.

We can collect frequency signal from 5 buses.

Majority Voting is applied to boost the performance.

There are 10 S1C cases, and 5 M2C case.

Note: where “-” denotes the missing buses, e.g., “”-Bus1” indicates

that the signal from bus 1 is missing, and “-Bus1, 2, 5” denotes that

buses 1, 2, and 5 are all missing.

Voltage Collapse Scenario

Collapsing

point

Collapsing

point

Stability

margin goes

zero

Stability margin

goes zero

1G1

5 6

G2

2

3 G31110

G4

4

97 8

L7 L9C7 C9

10 km25 km 25 km10 km110 km 110 km

Ramp the active power of the load on

bus 9 (20% const. impedance, 20%

const. current, 60% const. power) to

1.65 p.u. (1485 MW)

531 MW

Without Control

The system enters a state of voltage instability (or voltage

collapse) when the increase in

load demand causes a

progressive and uncontrollable

decline in voltage.

Load increase

Voltage Collapse Scenario with STATCOM

Reactive power

injected

Stability margin goes

under threshold

Reactive power

injected

With Control

When the margin is close to a

defined threshold, e.g. 15%, it will

trigger the control action, such

as reactive power

compensation.

1G1

5 6

G2

2

3 G31110

G4

4

97 8

L7 L9C7 C9

10 km25 km 25 km10 km110 km 110 km

Ramp the active power of the load on

bus 9 (20% const. impedance, 20%

const. current, 60% const. power) to

1.65 p.u. (1485 MW)

531 MW

RTDS Interface with HTB

G1

Area 1

G2 LD7

1 6

2

7HTB

G3

G4

310

4

9

Transmission

Line LD9

Area 2

Emulator 1

Emulator 2 Emulator 3

Emulator 4

Emulator 5Emulator 6

2.45 mH

1.2 mH

0.7 mH

2.5 mH

0.7 mH

0.7 mH10.7 mH

RTDS

Po

wer

Inte

rface

+ -

12

Dig

ital

Inte

rface

Po

wer

Inte

rface

+ -

12

Dig

ital

Inte

rface

12 133 mH

10 mH 6 mH

LD12 LD13

G5

152.5 mH

HVDC1 HVDC2

abc/dq0

Voltage

ReferenceFilter

Vabc Vdq0

abc/dq0

Current

ReferenceFilter

Iabc Idq0

Power

ConverterDigital

InterfacePLL

Digital Side

Voltage

Closed-Loop

Control

Closed-Loop

Control

abc/dq0

abc/dq0

ΔθI Current Delay

Correction

ΔθV Voltage Delay

Correction

FilterVdq0fb

FilterIdq0fb

Feedback

Voltage

Feedback

Current

dq0/abc

dq0/abc

Voltage

ModulatePWM

θ

θ 2

1θ=ωt

Simulator Hardware

VS

Zs

ZLVL VM =VL+ε

IL

IM =IL

VS

Zs

ZLVLVM =VL

Is

IM =Is+ε

Simulator Hardware

VL

Interface Algorithms:

Voltage Type

Current Type

Area 3 in

RTDS

Two RTDS

interfaces

with the HTB

Area 1 and 2 in HTB

Combination of Voltage and Current TypeRTDS Interface Attributes

• Expand HTB to more than 40 buses

• Unique system that has both control and

power hardware interface

53

3-Area CURENT HTB Structure

WT II

DC

cab

le 1

DC

cab

le 2

DC cable 4

DC cable 3

VSC 4

1G1 5 6

G2 2

3

G3/WT III

1110

G44

97 8

L7 L9C7 C9

10 km25 km 25 km10 km

110 km 110 km

VSC 3

VS

C 2

VS

C 1

12 13

L12

L13

G14

14

110 k

m

66 k

m

33 km

WT I

Area 1 Area 2

Area 3

Multi-terminal

HVDC

Wind FarmWind Farm

Three-area system including multi-terminal HVDC

and wind farm

Three-area system:

• Three areas interconnected by three long

transmission lines (220 km, 220 km, and 132

km respectively)

• Area 3 connected to a multi-terminal HVDC

(MTDC) system with DC lines (two 100 km,

one 60 km, and one 70 km)54

Scenario - High Renewable Penetration – Inertia Emulation

Scenario

• Replace G2 by an onshore wind farm.

Together with offshore wind provided by HVDC, system renewable penetration can reach 80%.Event triggered by a HVDC converter failure.

Solution

• Frequency and voltage support from

onshore wind farm and the HVDC converters

• Curtailment and voltage mode control when necessary

• Integration of energy storage to further enable grid

support controls

Expected outcome

• System frequency and voltage

within acceptable range

0 2 4 6 8 10

0.65

0.7

0.75

0.8

0.85

0.9

0.95

0 2 4 6 8 1059.2

59.4

59.6

59.8

60

60.2

Wind Turbine Active Power (p.u.)

Area Frequency (Hz)

Time (s)

MPPT with inertia emulation

MPPT

Base case with generator

Voltage mode with storage

Voltage mode

HTB scenario test results

55

• Scenario

o 50% renewable penetration

o One poorly damped inter-area oscillation mode between Area 1 and Area 2

• Wide-area Damping Controller (WADC)

o Based on the measurement-driven model, which is identified by using ring-down data and ambient data

o Controller input: frequency difference between Area 1 and Area 2

o Controller output: added to exciter voltage reference of Generator 1

• System disturbance

o Load increase (Bus 9, 0.7 p.u. to 1.1 p.u.)

Scenario – Wide-area Damping Controller

0 5 10 15 20 25 30 35 40-3

-2

-1

0

1

2x 10

-3

t(s)

f 1 -

f4(p

.u.)

without control

with control

Frequency deviation Tie-line power, Bus 7-Bus 9

0 5 10 15 20 25 30 35 400.18

0.2

0.22

0.24

0.26

0.28

0.3

0.32

t(s)

PB

us7-B

us9 (

p.u

.)

without control

with control

0 2 4 6 8 10 12 14

0

5

10

time (s)

Dam

pin

g R

atio (

%)

without control

with control

Estimated damping ratio using Matrix Pencil56

Rebuild a N+1 buses equivalent (For 3-area HTB, N = 2) using boundary

buses measurements.

Calculate the transfer limits for each tie line.

Monitor the stability margins.

If margin < threshold, take remedial action.

• Collapse scenario: load increase on Bus 13

until system collapses.

• Stable scenario: with the same load increase,

the remedial action (Q support from MTDC) is

taken against the voltage instability triggered by

the stability margin monitoring.

Scenario – Measurement-Based Voltage Stability

Assessment and Control for Load Area (3-Area)

40 60 80 100 120 140 1600.3

0.4

0.5

0.6

0.7

0.8

0.9

Time (s)A

ctiv

e P

ow

er (

p.u

.)

P9to13

P9to13Limit

P9to13(after Q support)

P9to13Limit(after Q support)

Stable

40 60 80 100 120 140 1600.3

0.4

0.5

0.6

0.7

Time (s)

Act

ive

Pow

er (

p.u

.)

P7to12

P7to12Limit

P7to12(after Q support)

P7to12Limit(after Q support)

Collapse

Stable

Stability Margin

Stability Margin

57

New MBVSA algorithm:

Equivalent circuit model

HTB 4-Area Structure

58

Four-Area

System

Four-area system including multi-terminal HVDC overlay, large wind farms,

solar energy, and energy storage (battery-based)

DC Line 2-3

DC Line 2-6

139

L2

L9

Area 2 Area 3

Area 1

Multi-terminal

HVDC

G1

11

c

L11PV7

PV5

L7

5

Area 4

G2

2

L8

7

PV6L6

6

L10

WT 3

WT 4

10

4

c

VSC 1VSC 2

VSC 3

DC Line 3-6

T 11-6

T 8-2 T3-4

550 km

580 km355 km

969 km

1674 km1625 km

260 km

S8

S9

S10

S11

Rebuild a N+1 buses equivalent (For 4-area HTB, N = 2) using boundary

buses measurements.

Calculate the transfer limits for each tie line.

Monitor the stability margins.

If margin < threshold, take remedial action.

• Collapse scenario: load increase on Bus 11 until system

collapses.

• Stable scenario: with the same load increase, the

remedial action (reactive power support from HVDC at

bus 6 and energy storage at bus 11) is taken against the

voltage instability triggered by the stability margin

monitoring.

Scenario – Measurement-Based Voltage Stability Assessment and Control

(4-area)

7-59

MBVSA algorithm:

Equivalent circuit model

Without

reactive

power

support

With

reactive

power

support

• Collapse scenario: three-phase short circuit at the

end of line 3-4, line 3-4 is tripped by the overcurrent

protection, the system becomes unstable

• Stable scenario: when overcurrent protection is

triggered it sends the signal to trip line 2-8, this action

separates the system into two stable islands.

Scenario – Power system separation remedial action scheme

7-60

Without the

separation

With the

separation

Scenario Development

61

• Interactive voltage control scenario: To

demonstrate the application of energy storage in

voltage control. Users can increase load at one or

multiple load buses (8,9,10 or 11), the energy

storage injects reactive power to the corresponding

bus when the active power reaches 90% of the limit,

saving the system from voltage collapse.

• State predictor scenario: To demonstrate the

application of state predictor in transient instability

prevention using a semi-analytical approach. After a

generator trip, the system will show real time and

10-second future predicted rotor angle of

generators. All energy storage in the system injects

active power to prevent transient instability when

predictor detects instability.

HTB Future Work

• Continued HTB Construction and Expansion

Provide more flexible control and visualization options – easier reconfigurability

Expand the HTB structure to the national grid with RTDS or LTB.

Continue integration of multi-terminal HVDC into HTB with MMC structure.

• Functionality enhancement Add more capabilities to the communication architecture including the ability to test cybersecurity on

HTB and robustness to failures in communication, sensor failures, and actuator failures.

Enhance the HTB control function under the condition of the loss of monitoring or measurement.

Enhance the HTB control function with load identification function, cross regional AGC control.

Comprehensive integration of protection schemes for power electronics connected sources (wind,

PV, etc.)

Use testbed to develop microgrid controller to be demonstrated on EPB’s system in Chattanooga

62

HTB Demonstration Plan

Year 1~3

Generation I

Hardware implementation of

the power electronics based

emulators, including large wind /

solar / storage farm emulation.

Integrate PMU/FNET data into

HTB.

Multiple load and scenario

demonstrations

(multi-terminal HVDC, hybrid

AC/DC, multi-area oscillation

and control, high renewable

energy penetration).

Year 4~6

Generation II

Implementation of sensing,

monitoring, actuation, and protection

in real-time.

Integrate with real-time simulation.

Scenario demonstrations (multiple

HVDC links between wide areas,

major tie line and wind farm outage

dynamic effects, coordinated power

flow control over large distances,

demonstrate system resilience to

attacks, energy storage impact).

Year 7~10

Generation III

Coordinated high penetration

renewable control

demonstration.

Automatic real time

reconfiguration for selected

outage scenarios.

Ultra-wide-area coordinated

real-time communication and

control on a system hardened

against coordinated cyber

attack.

7-63

Acknowledgements

This work was supported primarily by the ERC Program of the National

Science Foundation and DOE under NSF Award Number EEC-1041877

and the CURENT Industry Partnership Program.

Other US government and industrial sponsors of CURENT research

are also gratefully acknowledged.

64