temporal-spatial coordination of distributed energy ... and microgrid-0613.pdf · "a...

Temporal-Spatial

Coordination of

Distributed Energy

Resources (DERs)

in Microgrids

Dr Yan Xu | Nanyang Assistant ProfessorSchool of Electrical & Electronic EngineeringNanyang Technological University (NTU)Email: [email protected]: https://eexuyan.github.io/soda/index.html

© 2020 Yan Xu All Rights Reserved

Xu Yan (NTU) Copyright reserved



mailto:[email protected]

https://eexuyan.github.io/soda/index.html

1

2

3

REIDS Project

DER Control

DER Operation

2

4 Hierarchy Coordination

Islanded microgrid

Grid-connected microgrid

5 DER Planning

Energy dispatch

Volt/Var regulation

Distributed generation units

Energy storage systems

0. Outline

1. REIDS Project

2. Control1) Islanded mode2) Grid-tied mode

3. Operation1) Energy dispatch2) Volt/Var regulation

4. Hierarchycoordination

5. Planning1) DG planning2) ESS planning3) PRO algorithm

Volt/Var control

Active power balancing

Timescale

ms ~ seconds

mins ~ hours

ms ~ hours

years ~ decades




▪ Renewable Energy Integration Demonstrator – Singapore (REIDS)

3

0. Outline

1. REIDS Project





http://erian.ntu.edu.sg/REIDS/Pages/AboutREIDS.aspx





4

0. Outline

1. REIDS Project





64,400 m2




Solar Renewables

Wind Renewables

Marine Renewables- Solar PV- Solar Thermal Electric

- Small toIntermediate- Onshore & Offshore

- In-stream Tidal

Renewable Energy Generation

Energy Storage

- Batteries – Li-ion, RedoxFlow

- Super-capacitor- Flywheel- Power-to-gas

Fuel Cell

-H2

Loads

- On-island loads(residential,commercial, SME’s)

- Desalination- Aquaculture- Agro processing- Clean mobility- …- Connection to

neighboring islands

- Remote Monitoring& Control

- AC/DC

HybridMicrogrid #1

Diesel Gen

-Diesel

Hybrid Microgrid

#N

Energy Sources

LoadsStorage

Hybrid Microgrid

#M

Energy Sources

Storage Loads

Grid Connection

▪ REIDS Roadmap and Framework

5

0. Outline

1. REIDS Project





Phase I – 4 independent MGs (500kW-1MW each)

Phase II – 4 MGs in a cluster configuration (100kW-250kW each)




0. Outline

1. REIDS Project





▪ Onboard Industry Collaborators






▪ REIDS Electrical Structure

– Interlock (IL1, IL2 &IL3)

– Feeders

– Transformer

– Power line

– Communication line (feedback/command signals)

ESS –Energy Storage System

RH –REIDS Hub

SAS –Shared Assets Set

LVMG –Low Voltage Micro-Grid

MG – Micro-Grid

SAS-1

LVMG Cluster-1

MG2

MG3

MG0

MG1

MG4

MG5

MG7

MG6

Loads

Operating MG4 and /or MG5 with RH & shared assets set -1

LVMG Cluster-1

SAS-2

Operating MG6 and /or MG7 with RH & shared assets set -2

LVMG Cluster-2

Operating Clusters 1 & 2 with RH & shared assets set 1 & 2

LVMG Cluster-3

6.6 kV

Central

Link Bus

LVMG Cluster-3

LVMG Cluster-2

IL1 IL2

IL3

RHEMS

400V Switchgear

Diesel Engine

Solar Arrays

Wind Turbine

ESS Load Solar Arrays

Wind Turbine

ESSLoadDiesel Engine

7

0. Outline

1. REIDS Project








▪ Onsite pictures

MG0 Test & Commissioning - March 2017

400kW PV

8

0. Outline

1. REIDS Project








▪ Onsite pictures

9

0. Outline

1. REIDS Project








10

0. Outline

1. REIDS Project





▪ Our research Framework: system-level coordination of DERs

Distributed generator (DG)

RES (Solar, Wind)

Micro-turbine, CCHP

Energy storage (ES)

Battery, Supercapacitor

Thermal ES

Hierarchy coordination

Flexible Load

Demand response

HVAC, Smart building

Primary & Secondary control

· Voltage control

· Energy management

· Volt/Var regulation

Operation (Tertiary control)

· Optimal sizing

· Optimal siting

DER allocation/planning

· Frequency control

· Centralized control

· Distributed control

· Decentralized control· Robust optimization

· Distributed optimization

· Stochastic optimization

· Robust optimization

· Probability-robust optimization

· Stochastic optimization

ms ~ seconds mins ~ hours years ~ decades

Multi-stage coordination

Time frame

Distributed

Energy

Resources

(DERs)

Me

tho

do

log

ies

Controller design Optimal dispatch Investment decision-making

Distribution network Distribution network

Coordinated optimization model Coordinated optimization model

· Hardware-in-the-loop test

Pro

ble

ms




11

▪ Control of DERs in Microgrids

2. Grid-connected mode:▪ DER for f support▪ DER for V support

1. Islanded mode:▪ Distributed control

(event-triggered, finite-time)

▪ Hardware-in-the-Loop(Hil) validation

0. Outline

1. REIDS Project





...DG-1

Power System

(a) Centralized control

DG-2 DG-n

Central controller

...DG-i

...DG-1

Power System

DG-2 DG-n...DG-i

...Agent-1 Agent-2 Agent-n...Agent-i

(c) Distributed control

...DG-1

Power System

DG-2 DG-n...DG-i

...Local

Control-1...

(b) Decentralized control

Local

Control-2

Local

Control-i

Local

Control-n

...DG-1

Power System


DG-2 DG-n

Central controller

...DG-i

...DG-1

Power System

DG-2 DG-n...DG-i



...DG-1

Power System

DG-2 DG-n...DG-i

...Local

Control-1...


Local

Control-2

Local

Control-i

Local

Control-n

...DG-1

Power System


DG-2 DG-n

Central controller

...DG-i

...DG-1

Power System

DG-2 DG-n...DG-i



...DG-1

Power System

DG-2 DG-n...DG-i

...Local

Control-1...


Local

Control-2

Local

Control-i

Local

Control-n




12

▪ Hierarchical control of an islanded microgrid

Hierarchical control framework of islanded microgrids

0. Outline

1. REIDS Project





opened

➢ Tertiary control (centralizedor distributed)

▪ Economic dispatch,optimal power flow.

➢ Secondary control(centralized or distributed)

▪ V/f restoration andaccurate power balancing

➢ Primary control(decentralized)

▪ Inner control loops anddroop control

▪ Local V/f regulation andpower sharing




13

▪ Distributed Control – Spatial Coordination of DERs

✓ No need for a central controller ✓ One node only communicates with neighbouring nodes ✓ Share communication and computation burden among nodes ✓ Higher resilience, plug-and-play, scalability, data privacy

( ) ( )( ( ) ( ))i

i ij j i

j N

x t a t x t x t

= − 0

1

( ) ( )( ( ) ( )) ( ( ) ( )).n

i ij j i i i

j

x t a t x t x t g x t x t=

= − + −

0lim ( ) ( ) 0it

x t x t→

− =lim ( ) ( ) 0i jt

x t x t→

− =

12 14

21 23

32 34

41 43

0 0

0 0

0 0

0 0

a a

a a

a a

a a

=

A

Example of communication graph Adjacent matrix of the graph

a) Average consensus control b) Leader-follower consensus control

0. Outline

1. REIDS Project

2. Control 1) Islanded mode2) Grid-tied mode


4. Hierarchy coordination

5. Planning 1) DG planning2) ESS planning 3) PRO algorithm




14

▪ Secondary Controller Design – Principle 0. Outline

1. REIDS Project





nom P

i i i im P = −

nom Q

i i i iV V m Q= −

Droop control

nom P

i i i im P = −

nom Q

i i i iV V m Q= −

nom ( ) ( )P P

i i i i im P dt u u dt = + = +

Taking Derivative

Problem formulation

ss

Primary Control

Secondary Control

P1P 2P

nom

Example: Frequency

Regulation in Islanded

Microgrids with Two DGs

1

( ) ( )N

ref

i ij j i i i

j

u a g =

= − + −

1

( ) ( )N

V ref

i ij j i i i

j

u a V V g V V=

= − + −

1

( )N

p P P

i ij j j i i

j

u a m P m P=

= −

1

( )N

Q Q Q

i ij j j i i

j

u a m Q m Q=

= −

Apply consensus control rule

nom ( ) ( )Q V Q

i i i i iV V m Q dt u u dt= + = +




▪ Cross-national hardware-in-the-loop (HiL) testbed

15

Jointly developed by NTU (Singapore), University of Strathclyde (UK), and G2E Lab (France)

• Microgrids system with OPAL-RT in Singapore.

• Distributed controllers in Raspberry Pi in UK and France.

• Software environment based on gRPC and data exchange via Redis cloud server.

Y. Wang, T. L. Nguyen, M. H. Syed, Y. Xu*. "A Distributed Control Scheme of Microgrids in Energy

Internet and Its Multi-Site Implementation." IEEE Transactions on Industrial Informatics, 2020.

0. Outline

1. REIDS Project








▪ HiL Validation Results – Controller performance

16Y. Wang, T. L. Nguyen, M. H. Syed, Y. Xu*. "A Distributed Control Scheme of Microgrids in Energy Internet and Its

Multi-Site Implementation." IEEE Transactions on Industrial Informatics, 2020.

0. Outline

1. REIDS Project





gRPC

server i

gRPC client 1

Agent i (python based program)

Process consensus

algorithm, transfer data

gRPC client 2

gRPC client j

Raspberry i (IP address i)

...gRPC

server j

gRPC client 1

Agent j (python based program)

Process consensus

algorithm, transfer data

gRPC client 2

gRPC client i

Raspberry j (IP address j)

......

TCP/IPTCP/IP

Local Area Network

emulated by NS3

Cloud

Server

UDP UDP

Structure of each agent based on gRPC

Test system: 10-DG with two controller in UK and France

(Each controller for 5 DGs)

a) step load change case b) Real PV and load profile case

Real

Po

wer (

kW

) 4

3

0

Time (s)0 120 240 360 720480 600

2

1

Volt

ag

e M

ag

nit

ud

e (

V)

326

328

324

3200 120 240 360 720480 600

322

Fre

qu

en

cy (

Hz)

50.15

50

50.3

49.85

49.70 120 240 360 720480 600

(b)

(a)

(c)

Real

Pow

er (

kW

) 5

4

6

3

0

Time (s)0 30 60 90 180120 150

2

1

Volt

age M

agnit

ude (V

)

330

325

335

320

3100 30 60 90 180120 150

315

Fre

quency (H

z)

50.2

50

50.4

49.8

49.60 30 60 90 180120 150

(a)

(b)

(c)




17Y. Wang, T. L. Nguyen, M. H. Syed, Y. Xu*. "A Distributed Control Scheme of Microgrids in Energy Internet and Its

Multi-Site Implementation." IEEE Transactions on Industrial Informatics, 2020.

Communication delay emulated by NS3 simulation tools.

Real

Po

wer (

kW

)

8

6

10

4

-2Time (s)

0 30 60 90 180120 150

2

0

12

Vo

ltag

e M

ag

nit

ud

e (V

)

335

325

345

315

0 30 60 90 180120 150305

Fre

qu

en

cy (H

z)

50.2

50

50.4

49.8

49.60 30 60 90 180120 150

(a)

(b)

(c)

System oscillation under large delay, which

can be mitigated by tuning the control gain.

0. Outline

1. REIDS Project





▪ HiL Validation Results – Communication delay

500ms Delay

Dela

y (m

s)

800

600

1000

400

Time (s)0 30 60 90 180120 150

0

200

150ms Delay

Test system: 5-DG MG with one MAS in UK

✓ Larger control gain -> converge faster-> withstand smaller delay.

✓ Smaller control gain -> convergeslower -> withstand larger delay




18Y. Wang, T. L. Nguyen, M. H. Syed, Y. Xu*. "A Distributed Control Scheme of Microgrids in Energy

Internet and Its Multi-Site Implementation." IEEE Transactions on Industrial Informatics, 2020.

0. Outline

1. REIDS Project





Communication failures

▪ HiL Validation Results – Communication failures

Real

Pow

er (

kW

)

6

4

8

10

-2Time (s)

0 30 60 90 180120 150

2

0

Volt

age M

agnit

ude (V

)

330

325

335

315

0 30 60 90 180120 150310

320

Fre

quency (H

z)

50.2

50

50.4

49.8

49.60 30 60 90 180120 150

(a)

(b)

(c)

Test system: 5-DG MG with one controller in UK

Inaccurate power

sharing

Secondary control start

✓ Failure of communication will affect the convergence speed

✓ Loss of communication will lead to inaccurate power sharing

Plug-and-playXu Yan (NTU) Copyright reserved



▪ Event-Triggered Distributed Control of Islanded Microgrids

19Y. Wang, T. L. Nguyen, Y. Xu*, et al, “Cyber-Physical Design and Implementation of Distributed

Event-Triggered Secondary Control in Islanded Microgrids ,” IEEE Trans. Industry Application, 2019.

Primary

Controller

…DG #1Real Power

Sharing Control

Eq. (14)

Frequency

Restoration

Eq. (33)

Microgrid

Event-Triggered Secondary Controller#i

DG #2 DG #i

Primary

Controller Primary

Controller

Communication

Network

1

2

Node i

(DG i)

i

Event-Triggered

Secondary

Controller#1

iu

V

iu

ETC#1

Eq.(15)- (18)

ETC#2

Eq.(29)-(32)

Event-Triggered

Secondary

Controller#2

{ , ,

, }

P

i i i

Q

i i i

K P

V K Q

{ , ,

, }

P

i i i

Q

i i i

K P

V K Q

{ , ,

, }

P

j j j

Q

j j j

K P

V K Q

Reactive Power

Sharing Control

Eq. (28)

Voltage

Restoration

Eq. (34), (36)

P

j jK P

P

i iK P

Q

j jK Q

Q

i iK Q

P

iu

Q

iu

Eq. (8)

Eq. (9)

nom

i

nom

iV

{ , }nom nom

i iV

2 2(1 )( ) ( ) ( )

P P P

P P Pi i

i i i

i

df t e t t

d

−= −

1inf{ | ( ) 0}Pi Pi P

k k it t t f t−= =

Event-Trigger Condition for f and P:

1inf{ | ( ) 0}Qi Qi Q

k k it t t f t−= =

2 2(1 )( ) ( ) ( )

Q Q Q

Q Q Qi i

i i i

i

df t e t t

d

−= −

Time

Tri

gg

eri

ng

Co

nd

itio

n

2(1 )( )

P P P

Pi i i i

i

i

dt

d

−

2

( )P

ie t

Effects of ETC

Reduce real-time communication and calculation burden

0. Outline

1. REIDS Project





Event-Trigger Condition for V and Q:




▪ Controller Hardware-in-the-Loop (CHil) Test

20Y. Wang, T. L. Nguyen, Y. Xu*, et al, “Cyber-Physical Design and Implementation of Distributed Event-

Triggered Secondary Control in Islanded Microgrids ,” IEEE Trans. Industry Application, 2019.

HiL testbed with Raspberry Pi and OPAL-RT

Microgrid topology with four DGs

DG-1 DG-2

DG-3DG-4

12Z

23Z

34Z

1 2

34

Communication

Network

Load-1 Load-2

Time (s)0 10 20 30 60

Real

Pow

er (k

W) 10

12

6.0

4

DG-1

DG-2

DG-3

DG-4

50

8

6

2

0

Load-2 ONLoad-1

OFF

40

3.0

10.0

5.0

4.0

2.0

Fre

qu

ency (H

z)

50.4

50

50.8

49.6

49.2

Load-2 ON

Load-1

OFF

DG-1

DG-2

DG-3

DG-4

Time (s)0 10 20 30 605040

Time (s)0 10 20 30 60

Agen

t N

um

ber

3

4

50

2

1

40

0. Outline

1. REIDS Project





Communication requirement




▪ Finite-Time Distributed Control of Energy Storage Systems

21

Y. Wang, T. L. Nguyen, Y. Xu*, D. Shi, “Distributed control of heterogeneous energy storage systems in islanded microgrids:

Finite-time approach and cyber-physical implementation,” Int. J. Electrical Power & Energy Systems, 2020.

Control diagram of one ESS unit

AC Bus

Voltage

Control

Loop

Current

Control

Loop

PWM

ESS Dynamics

Eq. (3)-(6)

iP i

iV

f

iL

L

ii

o

iL

o

iio

iV

Power

Calculation

f

iC

+ +

iQ

+

+

+

+

P

iK Q

iK cos( )i iV t nom

i

nom

iV

Distributed Finite-Time Secondary Control

Battery

DC/AC

Inverter LCL Filter

i i

E

dc

iV

dc

iI

+

Nominal Values

Calculation

Eq. (9),(10)

V

iu iu E

iu

Finite-time consensus control law

0. Outline

1. REIDS Project





Control objectivesXu Yan (NTU) Copyright reserved



▪ Finite-Time Distributed Control of Energy Storage Systems

22

Y. Wang, T. L. Nguyen, Y. Xu*, D. Shi, “Distributed control of heterogeneous energy storage systems in islanded

microgrids: Finite-time approach and cyber-physical implementation,” Int. J. Electrical Power & Energy Systems, 2020.

Linear consensus control Finite-time consensus control

Under the same power overshoot, the proposed controller converges

much faster (74s vs 120s)

0. Outline

1. REIDS Project





Time (s)0 40 80 120 200

Rea

l P

ow

er (k

W) 8

12

-4

ESS-1 ESS-2 ESS-3 ESS-4

160

4

0

Sta

te-o

f-C

har

ge (

%)

48

52

44

40

36

32

Fre

quen

cy (H

z) 50.3

50.6

49.4

49.7

50

330

328

326

324

322

Volt

age

Mag

nit

ude (

V)

Time (s)0 40 80 120 200160

Time (s)0 40 80 120 200160

Time (s)0 40 80 120 200160

11

-2.2

Converged

t 120s

Power overshoot

Load-2

ONLoad-1

OFF

Converged

t 120s

Load-1

OFF

Load-2

ON

Converged

t 120s

Load-2

ON

Load-1

OFF

Converged

t 120s

Load-1

OFFLoad-2

ON

50.02

49.66259.5

6259.5

326

323

(a)

(b)

(c)

(d)

Time (s)0 40 80 120 200

Real

Po

wer (k

W) 8

12

-4

ESS-1 ESS-2 ESS-3 ESS-4

160

4

0

Sta

te-o

f-C

harg

e (%

)

48

52

44

40

36

32

Fre

qu

ency (H

z) 50.2

50.4

49.6

49.8

50

Time (s)0 40 80 120 200160

Time (s)0 40 80 120 200160

Time (s)0 40 80 120 200160

11

-2.2

Converged

t 74s

Power overshoot

Load-2

ONLoad-1

OFF

Converged

t 74s Load-1

OFF

Load-2

ON

326

326.5

325

324.5

324

Vo

ltag

e M

agn

itu

de

(V)

325.5

Converged

t 74s

Load-1

OFF

Load-2

ON

Load-1

OFF

(a)

(b)

(c)

(d)

6159.549.7

50.02

Load-2

ON

61.559.5324.5

325.5

Converged

t 74s




Voltage control support:

mitigate voltage deviation

(seconds to minutes)

Frequency and voltage are

dominated by the main grid

through point of coupling

connection (PCC).

Frequency control support:

mitigate frequency variation

(ms to seconds)

▪ Grid-connected mode of Microgrids (DER support)

23

0. Outline

1. REIDS Project





PCCXu Yan (NTU) Copyright reserved



▪ Frequency Support from Aggregated Energy Storage

24Y. Wang, Y. Xu*, Y. Tang, et al “Aggregated Energy Storage for Power System Frequency Control: A Finite-Time

Consensus Approach,” IEEE Trans. Smart Grid, May 2018,

Primary Control

Secondary Control (AGC)

Energy Storage Aggregator

System

Disturbance

Observer

Band-

Pass

Filter

ESS 1

ESS 2

ESS n

...

Co

mm

unic

atio

n N

etw

ork

Proposed Disturbance Observer

Proposed load frequency control (LFC) framework

( ), , , ,

( ) ( )2

1( ) ( ) ( ) ( ) ( )

2

ii i

i

mi L i RES i tie i ESA ii

Df t f t

H

P t P t P t P t P tH

= −

+ − + − +

1 1( ) ( ) ( ) ( )ai

mi mi gi gibi bi bi

TP t P t P t P t

T T T = − + +

1 1 1( ) ( ) ( ) ( )gi gi ci i

gi gi i gi

P t P t P t f tT T R T

= − + −

,

1,

( ) 2 ( ( ) ( ))

M

tie i ij i j

j j i

P t T f t f t=

= −

,( ) ( ) ( )i i i tie iACE t B f t P t= +

( ) ( ) ( )ci P i I iP t K ACE t K ACE t = − −

LFC with primary control

Tie-line power flow

Secondary control

Lead acid

batteryLithium ion battery Vanadium redox battery

…

0. Outline

1. REIDS Project








▪ Frequency Support from Aggregated Energy Storage

25

Y. Wang, Y. Xu*, Y. Tang, et al “Aggregated Energy Storage for Power System Frequency Control: A

Finite-Time Consensus Approach,” IEEE Trans. Smart Grid, May 2018,

(a)

1 2 3 4

8 7 6 5

1 2 3 4

8 7 6 5

(b)

00

( ) ( ), 1,2..., .

( ) ( )

i ESS i

i i

e t K p ti N

p t u t

==

=

*

0 ma

0

x

0

( )(

( ) ( )

) ESA

ESA

ESSe t K

P tp t

P

p t

=

=

1, if ,=

0, otherwise

( )

.

i j

ija

v v E

01, if ,=

0, otherw

( )

ise.

i

ig

v v

0

1

0

1

( ) ( ( ( ) ( )) ) ( ( ( ) ( )) )

( ( ( ) ( )) ) ( ( ( ) ( )) )

N

i ij i j i i

j

N

ij i j i i

j

u t a sig e t e t g sig e t e t

a sig p t p t g sig p t p t

=

=

= − − −

− − − −

0 00 0lim ( ) ( ) 0, lim ( ) ( ) 0i i

t T t Te t e t p t p t

→ →− = − =

0 0 0( ) ( ), ( ) ( ), , 1, 2,... .i ie t e t p t p t t T i N= = =

A = [aij]

G = diag{gi}

Matrices to

describe graph

Leader model

ESS model (follower)

Adjacent Matrix

Pinning Matrix

Proposed Control LawControl Objectives: achieve LFC

power reference with consensus SOC

Communication topologies of ESSs

0. Outline

1. REIDS Project





Consensus SOC

LFC power reference

LFC power reference




▪ Simulation Results

26Y. Wang, Y. Xu*, Y. Tang, et al “Aggregated Energy Storage for Power System Frequency Control: A Finite-Time

Consensus Approach,” IEEE Trans. Smart Grid, May 2018.

Time (s)0 600 1200 1800 3600

Sta

te-o

f-ch

arg

e (%

)

45

47

49

51

2400 3000

53

2

1

3

4

6

5

7

8

55

600

Converged

0. Outline

1. REIDS Project





Time (s)0 600 1200 1800 3600

Pow

er (

MW

)

-0.32

0

0.32

0.64

2400 3000-0.64

21

34

65

78

0.33

Time (s)0 600 1200 1800 3600

f

(Hz)

-0.4

-0.2

0

0.2

0.4

2400 3000-0.4

-0.2

0

0.2

0.4

f

(Hz)

(a) Conventional LFC scheme

(c) Proposed control scheme

f

(Hz)

-0.4

-0.2

0

0.2

0.4(b) LFC with ESA w/o observer

Consensus SOC




▪ Thermostatically Controlled Loads (TCLs) for frequency support

27

Temperature dynamics of TCL:

Heat exchange with

the ambient

( ) ( ) ( )( ), .i a i

th ith

dT t T t T tC P t i

dt R

−= − G

Thermal energy

from VFAC

( ) 0 0 01

( )2( ) 0

i

i

i

ii a s

iith thth th th

Bx

WAx

d t

dtu T t T TT P

d tC RC R C

dt

= + + − + − −

Assume power state αi is a continuous

variable from 0 to 1.

( )( ) ,

2

i si

T t T Tt i

T

− + =

G

Comfort state βi is an index from 0 to 1.

Comfort zone of TCL:

State-space model of TCL:

TOUT

TIN

TOUT

Heating system

HCT

TOUT

HCIN

HCER

0. Outline

1. REIDS Project








▪ Thermostatically Controlled Loads (TCLs) for frequency support

28

Y. Wang, Y. Xu, and Y. Tang, “Distributed Aggregation

Control of Grid-Interactive Smart Buildings for Power

System Frequency Support,” Applied Energy, 2019. Leader-follower consensus controller

...

kf

Multi-Area Power

System Frequency

Model

Eq. (1)-(5)

Leader Control

with Two Modes

Eq. (16)-(19)

DSMC of Multiple

GISBs Eq. (20)-(26)

Thermal-Electrical Model

of Multiple GISBs

Eq. (6)-(11)

Communication

Network of GISBs

Eq. (12)-(14)

aggP

Frequency of

kth Area

0 0,

,j j

1u

Nu ...

,i i

From/to

adjacent GISBsControl signal and system

states of corresponding GISBs

Total power from the

aggregation of GISBs

References

...

Time (s)0 300 600 900 1800

0.4

-0.4

0

0.8

1200 1500

(a) Area 1

(c) Area 3

(b) Area 2 -0.8

0.4

-0.4

0

0.8

-0.8

0.4

-0.4

0

0.8

-0.8

f

(Hz)

f

(Hz)

f

(Hz)

Proposed method

LFC w/o GISBs

Time (s)0 300 600 900 1200 1800

Pow

er (

MW

)

4

3

2

1

GISB aggregator power

Aggregator baseline power

Energy absorption

Energy injection

1500

0. Outline

1. REIDS Project





Leader control mode: f support mode and

comfort recover mode

Time

Base-line

Power

Comfort Recovery ModeFrequency Support Mode

Fre

quen

cy

Fre

quen

cy

Power

Power in

FSMPower range

in CRM




▪ Ancillary Service Support from Smart Building Community

29Y. Wang, Y. Tang, Y. Xu*, et al, “A Distributed Control Scheme of Thermostatically Controlled Loads

for Building-Microgrid Community,” IEEE Trans. Sustainable Energy, 2019.

Smart building community:

TCLs+PVs

Initial value updating scheme

1

To other groups

(communication rate: Δtc)

2

3

m

Power dispatch signal

(t=0)

PV and load regulation

(sampling rate: Δts)

Initial value update

Po

wer

(k

W)

3

1

0

5

2

4

Time (h)11:30 11:40 11:50 12:20 12:3012:1012:00

Com

fort

Sta

te (

p.u

.)

0.7

0.5

0.4

0.9

0.6

0.8

Time (h)11:30 11:40 11:50 12:20 12:3012:1012:00

Pow

er (

kW

)

200

300

Time (h)11:30 11:40 11:50 12:20 12:3012:1012:00

Total power consumption of TCLs

Power injection of the community

Total power injection of PV and load

400

Pow

er

(kW

)

300

350

400

Pow

er (

kW

)

0

100

200

-100

0

st st …st

…

dt

ct

One dispatch period (research scope)

Initial value update

t

ct

Distributed control

0. Outline

1. REIDS Project








▪ Real-time Voltage/Var Control (VVC) Support from DERs

➢ Existing Challenges: High PV penetration level, massive EV charging.

➢ Voltage quality issues: Voltage rise, drop and fast fluctuations.

➢ Potential solutions: inverter-assisted voltage/var support

30

0. Outline

1. REIDS Project





ES




▪ Real-Time Coordinated Voltage/Var Control Controller

31

Controller design:

Y. Wang, M. H. Syed, E. Guillo-Sansano, Y. Xu*, and G. Burt “Inverter-Based Voltage Control of Distribution Networks: A Three-Level

Coordinated Method and Power Hardware-in-the-Loop Validation,” IEEE Transactions on Sustainable Energy, 2019.

( )

( ) [ ( ) ]( ) ( )

t

i

j tI Ii i i

V j

u t K V tT t T t

= −= −

− −

Level I: Ramp-rate Control -> smooth voltage fluctuation

Level II: Droop Control -> immediate voltage support

( ( ) ), ( )

( ) 0, ( )

( ( ) ), ( )

IIi i i

IIi i

IIi i i

K V t V V t V

u t V V t V

K V t V V t V

−

=

−

Level III: Distributed Control -> voltage regulation to

acceptable range

1

( ) [ ( ( ) ( ))] ( )

NIII III III IIIi i ij j i

i

u t G a u t u t e t

=

= − +

( ( ) ),

( ) 0,

( ( ) ),

IIIi i i

i

IIIi i i

K V t V V V

e t V V V

K V t V V V

−

=

−

Netw

ork

Volt

age

II-Dro

op

Co

ntro

l

I-Ram

p-ra

te C

on

trol

III-Dis

tribu

ted

Co

ntro

l

Time0

VVC operational range

0. Outline

1. REIDS Project




5. Planning 1) DG planning2) ESS planning 3) PRO algorithm Dynamic

consensus




▪ Simulation Tests

32

Effectiveness of ramp-rate control

Real-time voltage/var control from inverters0. Outline

1. REIDS Project








▪ Power Hardware-in-the-Loop (PHiL) Test

Distributed Control

Communications

15 kVA

Converter

10 kW

7.5 kVAR

Load bank

Grid

PV #1

Bus 1

Distribution

Network

Bus 0

PV #2 PV #3 PV #4 PV #5

PV #6

Bus 6

PV #7

Bus 7

Bus 2 Bus 3 Bus 4Bus 5

90 kVA

Power

Interface

Beckhoff

EL3702

GTNET

GTNET

Analog signal

Digital signal

33

0. Outline

1. REIDS Project










▪ Power HiL Results and Eigenvalues

34

Real Axis-3 -2 0.5

Imag

inar

y A

xis

-1

1

2

-1 0-2

0

unstablestable

*

** 70i

IIIG =

20iIIIG =

90iIIIG =

Voltage profiles under step load changes

Voltage profiles under real PV and load data

Trace of eigenvalues under different control gains

0. Outline

1. REIDS Project










▪ Operation of DER - Energy Dispatch & Volt/Var Regulation in Microgrid

%

%

%

%%

Wind Turbine Photovoltaics

Electric Vehicle Battery

Diesel Generator Fuel Cell

Flexible load

Flexible load

Flexible load

Flexible load

Main Grid

Main Grid

Micro Grid

Power line

Communication Line

Microgrid Controller

Grid requests

• Control variables:1) Micro-turbine 2) Energy storage3) Demand response4) Capacitor banks5) On-load tap changers6) PV inverters

• Parameters: 1) Load demand2) Wind and PV output 3) Electricity price4) Network parameters (R,X,B)

• State variables:1) Bus voltage2) Branch power flow3) Power exchange with main grid

Uncertain

35

Active

power

resource

Reactive

power

resource

• Network model: 1) Linearized Dist-Flow 2) Second-order cone model

0. Outline

1. REIDS Project








36

Principle: coordinate different DERs in different timescales against uncertainty.

➢First-stage: slower-responding DER in longer timescale.

➢Second-stage: faster-responding or more flexible DER in shorter timescale.

▪ Two-stage coordinated operation – Temporal Coordination of DERs 0. Outline

1. REIDS Project





➢ Frist-stage decisions are implemented before uncertainty realizes and will be fixed in the second-stage.

➢ Second-stage decisions will be re-optimized and implemented after uncertainty realizes, therefore it is a recourse action to the first-stage decision.

st st…

st

dt

st st…

stTime

Second-stage: faster-responding

or more flexible DERs

hours~day

dt

15mins~hour

First-stage: slower-responding DERs

hours~day

15mins~hour

Recourse

action

Two-stage coordinated

optimization modelXu Yan (NTU) Copyright reserved



▪ Optimization Methods

37

Method Stochastic Programming Robust Optimization (RO)

UncertaintyModeling

Probabilistic scenarios based on probability distribution function (PDF)

Uncertainty setwith bounds and budgets

Inputs Point prediction Interval prediction

Model Optimize under expectation Optimize under worst case

Advantages • Simpler formulation and solution process

• No need for PDF• Fully robust within the

uncertainty sets

Disadvantages • Need for PDF• Probabilistic robustness

• Complex formulation and solution process

• May be conservative

0. Outline

1. REIDS Project








▪ Robustly Coordinated Energy Management Day-ahead Price-based Demand Response & Hourly-ahead Microturbine

38

Level Price Rate (%) Load Rate (%)

1 70 107.9

2 80 104.8

3 90 102.3

4 100 100.0

5 110 98.0

6 120 96.2

7 130 94.6

8 140 93.1

9 150 91.8

10 160 90.5

Price-based Demand Response (PBDR)

𝑷𝒕𝑫 = 𝐀 𝑷𝒓𝒕

𝜺 where 𝜀 is price elasticity of electric demand, and A is a constant value

modeling the relationship between the price and load demand. E.g., the

price elasticity of load is -0.38 for Australian power systems.

0. Outline

1. REIDS Project









C. Zhang, Y. Xu, Z. Y. Dong, "Robust Coordination of Distributed Generation and Price-Based Demand Response

in Microgrids," IEEE Trans. Smart Grid, 2018. Web-of-Science Highly Cited Paper 39

Two-Stage Operation Framework

0. Outline

1. REIDS Project





Two-Stage Robust Optimization (TSRO) model

Uncertainty modeling –uncertainty set

Objective function







Modelling for Price-based DR

0. Outline

1. REIDS Project





• Considering the characteristics of the uncertain

load demands, in (9), the revenue from the

demands is split into two parts i.e. the predicted

revenue based on the predicted load demands and

the uncertain revenue difference from the

predicted one.

• Constraints (10) and (11) support the calculation

functions of these two revenue items respectively.

• Constraint (12) denotes the decision variable for

each PBDR level is binary.

• Constraint (13) guarantees that only one PBDR

level decision can be carried out for each hour.

• Constraint (14) and (15) guarantees the bills for

the customers cannot increase and the energy

which the customers can use cannot decrease.

These mean that the proposed PBDR does not

reduce the customers’ economic benefits.







Day-ahead Interval Prediction Hourly Microturbine Dispatch

Day-ahead PBDR Decision

0. Outline

1. REIDS Project








▪ Robustly Coordinated Energy Management Hourly-ahead energy storage & 15min-ahead direct load control (DLC)

0. Outline

1. REIDS Project





42C. Zhang, Y. Xu, Z. Y. Dong, "Robust Operation of Microgrids via Two-Stage Coordinated Energy

Storage and Direct Load Control," IEEE Trans. Power Syst., 2017.

Two-stage robust optimization model

ESS economic model

ESS operation model

Two-stage coordination framework




▪ Robustly Coordinated Energy Management Hourly-ahead energy storage & 15min-ahead direct load control (DLC)

0. Outline

1. REIDS Project





43C. Zhang, Y. Xu, Z. Y. Dong, "Robust Operation of Microgrids via Two-Stage Coordinated Energy

Storage and Direct Load Control," IEEE Trans. Power Syst., 2017.




▪ Two-Stage Dispatch of Hybrid Energy Storage considering battery health

Relationship between the number of life cycles and the DOD of Ni-Cd batteries

C. Ju, P. Wang, L. Goel, and Y. Xu, “A two-layer energy management system for microgrids

with hybrid energy storage considering degradation costs,” IEEE Trans. Smart Grid, 201744

0. Outline

1. REIDS Project








▪ Two-Stage Dispatch of Hybrid Energy Storage considering battery health

45

0. Outline

1. REIDS Project





✓ First-stage: battery dispatch with SOH

degradation cost

✓ Second-stage: supercapacitor dispatch




▪ Multi-Energy Microgrid

46

Battery StoragePhotovoltaic CellWind Turbine Fuel Cell

Upstream Grid

Heat Storage Tank Heat Recovery Unit Micor-Turbine Absorption Chiller Ice Storage Tank

Electric Boiler Electric Chiller

CCHP Plant

Power Heat Cooling

Power Demands

Heat Demands

Cooling Demands

0. Outline

1. REIDS Project





Z. Li and Y. Xu*, “Optimal coordinated energy dispatch for a multi-energy microgrid in grid-connected and islanded modes,”

Applied Energy, 2017. Web-of-Science Highly Cited Paper




▪ Multi-Energy Microgrid – Modeling of thermal part

47Z. Li and Y. Xu*, “Optimal coordinated energy dispatch for a multi-energy microgrid in grid-

connected and islanded modes,” Applied Energy, 2017. Web-of-Science Highly Cited Paper

Starting point Ending point

Indoor temperature

Electrical network

Heat network

Ambient

temperature

18

Point of

Connection

69-bus distribution network

27-pipe heat network

Heat loadHeat node

19 20 21 22 23 24 25 26 27

EB5

EB2

EB6

EB4

EB1

EB3

𝑄𝐶𝐻𝑃 𝑚𝑠,𝑘 ,𝑡 𝑄𝑠,𝑘 ,𝑡−

𝑄𝑟 ,𝑘 ,𝑡+

Primary supply pipes

Primary return pipes

Secondary supply pipes

Secondary return pipes

Heat sourceHeat exchanger Heat load

Heat exchanger

Heat load

𝑄𝑟 ,𝑘 ,𝑡−

𝑄𝑠,𝑘 ,𝑡+

𝑄𝑙𝑜𝑎𝑑

𝑚𝑟 ,𝑘 ,𝑡

𝑄𝑙𝑜𝑎𝑑

Y. Chen, Y. Xu*, Z. Li, “Optimally Coordinated Dispatch of Combined-

Heat-and-Electrical Network,” IET Gen. Trans. & Dist., 2019.

District Heat Network

Vertical section of a pipe

Coupled electric-thermal network

Thermal conduction of a building

0. Outline

1. REIDS Project








4848

▪ Multi-Energy Dispatch – Two-Stage Coordinated Operation

First stage: Day-ahead Dispatch for 24 Hours of Next Day

Hour 1 Hour 3 Hour 22 Hour 23 Hour 24……Hour 2

Second stage: Intra-day Online Dispatch within Each Hour

t1 t2 t4 t5 t6 t7 t8 t9 t10 t11 t12t3

Day-ahead

Forecasting

Short-lead-time

Forecasting

Power

Loads

RES

Generation

Transaction

Prices

Two-stage

Stochastic

Optimization

Temporally-coordinated Stochastic Operation Framework

1 2 n, , ...min ( ) ( )

S

n nz y y y

n N

F z L y

+

. .

( , ),

A

n n

s t z F

y z n

Two-Stage Sstochastic Optimization model

G FC OM EX ST SD HRMIN F C C C C C C= + + + + −

, ,( / )T M

t i t i

FC G MT MT

t N i N

C P t

= ,1 ,1( )

T

t t

EX B BUY S SELL

t N

C P P t

= − , , ,[ +... ( + )]

= + T W H

t i t i t i

OM WT WT TST TSTC TSTD

t N i N i N

C P P P t

, ,

( )

{0 }T M E

t i t -1 i U

ST CG CG CG

t N i N N

C max ,U -U C

= , ,

( )

{0 }T M E

t -1 i t i D

SD CG CG CG

t N i N N

C max ,U -U C

= ,

T M

t i

HR HR L

t N i N

C H t

=

, , , , ,t i min i t i t i max i

CG CG CG CG CGU P P U P

, , , ,down i t i t -1 i up i

CG CG CG CGR t P P R t −

, 1 , ,0,b 1 , ,... , ( ), ,+ += − − + − t b t b t t i t i

PF PF PF L PTP P P P P b Br i i t

, 1 , ,0, 1 , , ( ), ,t b t b t b t i

PF PF PF LQ Q Q Q b Br i i t+ += − −

, 1 , , ,

0( ) / ,b ( , 1), ,t i t i b t b b t b

BUS BUS PF PFV V R P X Q V Br i i i t+ = − + +

max , max1 1+t i

BUS BUS BUSV V V−

, , , , ,t i t i t i t i t i

L MT PT TSTD TSTCH H H P P= + + −

, , ,min i i t i max i i

ES ES ES ES ESCap E Cap

Z. Li and Y. Xu*, “Temporally-Coordinated Optimal Operation of a Multi-energy Microgrid under Diverse Uncertainties,”

Applied Energy, 2019.

0. Outline

1. REIDS Project








Z. Li and Y. Xu*, “Temporally-Coordinated Optimal Operation of a Multi-

energy Microgrid under Diverse Uncertainties,” Applied Energy, 2019. 49


Day-ahead dispatch results

Intra-day dispatch results

Input data for modelling

Get simulation results

Predicted wind,

solar output

Forecasting

load demands

Coordinated system-wide dispatch modelling

Parameters of

system, units

Solve the Second stage model

GAMS software, CPLEX solver

CHP

Electric boiler

Power outputs of

controllable units

Charging/discharging

power

On-off statues of

controllable units

CHP

Electric boilerEES

Solve the First-stage model

GAMS software, CPLEX solver

Real-time

electricity price

Other real-time

dataRES output

Input data

First-stage

(Day-ahead

stage)

Second-

stage

(Intra-day

stage)

-4000

-2000

0

2000

4000

6000

8000

1 5 9 13 17 21

Pow

er

(kW

)

Time (h)

MT WT PV PTC BS EX Power

-0.4

0

0.4

0.8

1.2

-600

-200

200

600

1000

1400

1 3 5 7 9 11 13 15 17 19 21 23

Sta

te o

f C

har

ge

Po

wer

(kW

)

Time (h)

MT PTC TST Heat SOC

-0.4

0

0.4

0.8

1.2

-800

-400

0

400

800

1200

1600

1 3 5 7 9 11 13 15 17 19 21 23

Sta

te o

f C

har

ge

Po

wer

(kW

)

Time (h)

MT PTC TST Heat SOC

-0.4

0

0.4

0.8

1.2

-800

-400

0

400

800

1200

1600

1 3 5 7 9 11 13 15 17 19 21 23

Sta

te o

f C

har

ge

Po

wer

(kW

)

Time (h)

MT PTC TST Heat SOC

0. Outline

1. REIDS Project








Z. Li and Y. Xu*, “Temporally-Coordinated Optimal Operation of a Multi-energy Microgrid under Diverse Uncertainties,”

Applied Energy, 2019. 50


Method #1: Single-stage deterministic operation

Method #2: Single-stage stochastic operation

Method#3: Two-stage deterministic optimization

Cost

Security

Item Method #1 Method #2 Method #3 Our Method

Uncertainty level 1 (Lower Uncertainty)

Average cost ($)

Average voltage violation (%)

2183.46 2149.65 2468.20 2440.22

30.40 16.50 0 0

Uncertainty level 2 (Medium Uncertainty)

Average Cost ($)

Average voltage violation (%)

2218.89 2188.97 2483.19 2450.78

74.70 49.80 0 0

Uncertainty level 3 (High Uncertainty)

Average Cost ($)

Voltage violation (%)

2341.64 2282.66 2556.04 2508.65

97.20 77.90 0 0

0. Outline

1. REIDS Project








▪ Multi-Energy Demand Response

51

indoor temperature control (thermal load) and price-based DR (electric load)

to counteract uncertain renewable power generation, load, and ambient temperature

0. Outline

1. REIDS Project





C. Zhang, Y. Xu*, Z.Y. Dong, “Robustly Coordinated Operation of A Multi-Energy Microgrid with Flexible Electric and

Thermal Loads,” IEEE Trans. Smart Grid, 2018.

Day-ahead robust optimization model

Intra-day optimization model




▪ Multi-energy demand response

52C. Zhang, Y. Xu*, Z.Y. Dong, “Robustly Coordinated Operation of A Multi-Energy Microgrid with Flexible Electric and

Thermal Loads,” IEEE Trans. Smart Grid, 2018.

0. Outline

1. REIDS Project





Day-ahead optimization results - PDR Day-ahead optimization results – thermal ESS

Intra-day results – CCHP and indoor temperatureIntra-day optimization results – thermal load




▪ Robustness VS Conservativeness

53C. Zhang, Y. Xu*, et.al, "Robustly Coordinated Operation of A Multi-Energy Micro-Grid in Grid-Connected

and Islanded Modes under Uncertainties," IEEE Trans. Sustain. Energy, 2020.

Robustness under Different Uncertainty Budgets

Robustness:

Possibility of a feasible solution (or no operating constraint violation) whatever uncertainties realize (Advantage)

• Full Robustness: Always a feasible solution

Conservativeness:

Compromise in optimization process when considering uncertainties (Drawback)

Design of Uncertainty Budgets

• Larger Budgets

-> Higher Robustness

-> Higher Conservativeness

• Uncertainty Degree Analysis

0. Outline

1. REIDS Project








▪ Multi-Energy Dispatch GUI Prototype

54

0. Outline

1. REIDS Project








▪ Robustly Coordinated Energy Management Distributed robust optimization for Networked-Hybrid AC/DC Microgrids

55Q. Xu, T Zhao, Y. Xu*, et al, " A Distributed and Robust Energy Management System for Networked

Hybrid AC/DC Microgrids," IEEE Transactions on Smart Grid, 2020.

1. Affinely adjustable robust optimization modeling

2. Model convexification

3. Distributed solution based on ADMM

0. Outline

1. REIDS Project





Individual MG

Networked-MG




56

Scenario I: centralized deterministic;Scenario II: centralized stochastic; (100)Scenario III: proposed

A 3 networked microgrid system in an IEEE 4 bus system

A 30 networked microgrid system in a revised IEEE 123 bus system

▪ Simulation results 0. Outline

1. REIDS Project








▪ Two-stage Coordinated Volt/Var Regulation under uncertaintyHourly dispatch of CB and OLTC & 15min dispatch of PV inverters

57

Volt/Var Control Model

Two VVC timescale

Stochastic Programming Model

Two decision stage

CB and OLTC hourly dispatch here-and-now variables

Inverters 15min dispatch wait-and-see variables

Random load and RES Stochastic variables

Network and utility code Model state variables

Load and RES forecasting Scenario construction

VVC signals Solution results

15-min dispatch of RES inverters

Hourly dispatch of CBs and OLTCs

Two-stage

stochastic

programming

1st-stage decision

2nd-stage decision

time

t1

t2 t3 t4

➢ First-Stage: Slow switching devices such as OLTCs and CBs are scheduled one day ahead.

➢ Second-Stage: Fast responding devices such as PV-associated inverters are updated to operate in short time-window.

Y. Xu*, Z.Y. Dong, et al, “Multi-timescale coordinated

voltage/var control of high renewable-penetrated distribution

networks,” IEEE Trans. Power Syst., 2018.

0. Outline

1. REIDS Project





smaller granularity Xu Yan (NTU) Copyright reserved



▪ Mathematical modeling

58

0. Outline

1. REIDS Project









59

0. Outline

1. REIDS Project





Y. Xu*, Z.Y. Dong, et al, “Multi-timescale coordinated voltage/var control of high renewable-penetrated

distribution networks,” IEEE Trans. Power Syst., 2018.




▪ Multi-Objective Adaptive Robust Voltage/VAR Regulation

60C. Zhang, Y. Xu*, Z.Y. Dong, "Multi-Objective Adaptive Robust Voltage/VAR Control for High-PV Penetrated

Distribution Networks," IEEE Trans. Smart Grid, 2020.

•Minimizing voltage deviation conflicts with

minimizing network power loss.

•Multi-objective “min-max-min” problem

Adaptive Weighted Sum (AWS)

Normal Boundary Intersection (NBI)

Key point:

1) Voltage deviation index: load-weighted voltage

deviation index (LVDI)

2) Which MOP algorithm is more efficient to generate

accurate Pareto front and get a fair trade-off?

a) Classic Weighted-Sum (CWS)

b) Classic ε-Constrained (CeC)

c) Adaptive Weighted-Sum (AWS)

d) Normal Boundary Intersection (NBI)

0. Outline

1. REIDS Project








61C. Zhang, Y. Xu*, Z.Y. Dong, "Multi-Objective Adaptive Robust Voltage/VAR Control for High-PV Penetrated

Distribution Networks," IEEE Trans. Smart Grid, 2020.

The AWS and NBI algorithms are suggested

depending on different optimization

requirements.

✓ If a relatively accurate Pareto front with

high computation efficiency is required, the

AWS algorithm is preferred.

✓ If a more accurate Pareto front with evenly

distributed solutions or the “knee” solution

is required, the NBI algorithm is preferred.

(a) CWS; (b) CeC; (c) AWS; (d) NBI

▪ Multi-Objective Adaptive Robust Voltage/VAR Regulation0. Outline

1. REIDS Project








62

✓Operational optimization and real-time control are traditionally decoupled.

✓Existing two-stage coordination methods are all for operational timeframe (e.g., day-ahead & hourly-ahead or hourly-ahead & 15mins-ahead).

✓Need to coordinate the operation level and control level for enhanced system performance, i.e., optimizing the operation decisions considering the real-time controllers’ effects, or simultaneously optimizing operational variables and controller parameters .

▪ Hierarchically coordinated operation and control of DERs0. Outline

1. REIDS Project





st st…

st

dt

st st…

stTime

Real-time control

mins~hours

dt

ms~s

Online dispatch

mins~hours

ms~s ms~s ms~s




RO: optimize 𝛼𝑡 , 𝛽𝑡considering the worst

case and the inverter

output dispatch

Obj.: loss reduction

Optimize 𝑄𝑙𝑟

considering the reactive

power reserve for local

droop voltage control

Obj.: loss reduction

Control 𝑄𝑖𝑛𝑣 with the

real-time local voltage

measurement

Obj.: voltage deviation

reduction

Third Stage: Inverter Droop Voltage Control (real time in first short period)

Second Stage: Inverter Output Dispatch (4 short periods in 1 hour)

First Stage: CB and OLTC Scheduling in a Rolling Horizon (4 hours)

Implemented DiscardedHour-ahead interval prediction

15-minute-ahead

point prediction

Real-time data

15-minute dispatch

reference

63

▪ Three-Stage Robust Volt/Var Control (TRI-VVC) 0. Outline

1. REIDS Project





C. Zhang, Y. Xu*, Z.Y. Dong, et al “Three-Stage Robust Inverter-Based Voltage/Var Control for Distribution

Networks with High PV,” IEEE Trans. Smart Grid, 2018. Web-of-Science Highly Cited Paper

Piecewise droop

controller for

inverters




1 12 24 36 48 60 72 84 96

Time Interval (15 min)

-500

-250

0

250

500

Inv

ert

er

Outp

ut

(kV

ar)

PV 9 PV 11 PV 21 PV 27

1 2 3 4

Time (hour)

20

40

60

80

100

PV

Outp

ut

(%

)

1 2 3 4

Time (hour)

50

60

70

80

90

Lo

ad

Dem

an

d (

%)

1/0.25 12/3 24/6 36/9 48/12 60/15 72/18 84/21 96/24

Time Interval/Time (15-min/Hour)

0

20

40

60

80

100

Pre

cen

tag

e o

f C

ap

acit

y (

%)

PPV

PD

1 6 12 18 24

Time (hour)

0

100

200

300

CB

Ou

tpu

t (k

Var)

-5

0

5

10

OL

TC

Po

siti

on

CB 3 CB 6 CB 23 OLTC


64

0. Outline

1. REIDS Project





C. Zhang, Y. Xu*, Z.Y. Dong, et al “Three-Stage Robust Inverter-Based Voltage/Var Control for Distribution

Networks with High PV,” IEEE Trans. Smart Grid, 2018. Web-of-Science Highly Cited Paper




▪ Hierarchically-Coordinated Voltage/VAR Control (HC-VVC)

65C. Zhang and Y. Xu*, "Hierarchically-Coordinated Voltage/VAR Control of Distribution Networks Using PV

Inverters," IEEE Trans. Smart Grid, 2020.

✓ Central VVC considers the

network level information

(power flow)

✓ Local VVC focuses on the real-

time variation (bus voltage)

Inverter Droop Control Model

• The central VVC hierarchy implements the

base reactive power output setpoint of each

inverter, i.e. 𝑄𝑖𝑏𝑎𝑠𝑒 under the expected

operating condition.

• The local VVC hierarchy implements the

local droop control by adjusting the

reactive power output responding to the

local voltage deviation. ∆𝑄 = 𝑓(∆𝑉)

0. Outline

1. REIDS Project





linear droop controller for inverters




66C. Zhang, Y. Xu*, "Hierarchically-Coordinated Voltage/VAR Control of Distribution Networks Using PV

Inverters," IEEE Trans. Smart Grid, 2020.

Voltage control results:

In response to the local bus voltage variation, the

inverter reactive power output moves along the

droop control curve.

The mean bus voltage magnitude with the HC-

VVC is very close to 1 p.u.

Comparison with other VVC methods

HC-VVC: least voltage violation rate; least

voltage magnitude deviation; second least

average power loss; second average voltage

close to 1 p.u.

▪ Hierarchically-Coordinated Voltage/VAR Control (HC-VVC)0. Outline

1. REIDS Project








▪ Fully Distributed Two-Level Volt/Var Control

67

Y. Wang, T. Zhao, C. Ju, Y. Xu*, P. Wang “Two-Level Distributed Voltage/Var Control of Aggregated PV Inverters

in Distribution Networks,” IEEE Trans. Power Delivery, 2019.

st st…

st

dt

Real-time distributed control within aggregators (Ts )

dt

15-min distributed dispatch among aggregators (Td )

st st…

stTimeD

ispat

ch

pow

er

Reactiv

e po

wer

constrain

ts

Two-level VVC with time scale coordinationDistributed dispatch by ADMM

VV

Time-

varying

term

dkq

VV

aggkq

d resk kq q+

aggkq

Time-

varying

term

Dispatched

Base Value

d resk kq q−

( )aggkq t

( )kV t

Distributed real-time voltage control

……

MV

Netw

ork

( )B = N, E

Main Grid Bus-1 Bus-2 Bus-N

1 1 1( )G = M , D

LV

System

with

PV

Aggreg

ato

rs

2 2 2( )G = M , D ( )agg agg aggN N NG = M , D

PCC PCC PCC

…

Upper-Level

Communication Link

Lower-Level

Communication

Link

0. Outline

1. REIDS Project









68Y. Wang, T. Zhao, C. Ju, Y. Xu*, P. Wang “Two-Level Distributed Voltage/Var Control of Aggregated PV Inverters

in Distribution Networks,” IEEE Trans. Power Delivery, 2019.

Time (Hour)

0 4

PV

/Lo

ad (

%)

0

20

40

60

80

100

12 16 248 20

Timescale of

one-hour

simulation

PV

Load

Time (Hour)

0 4

Net

work

Loss

(k

W)

0

50

100

200

12 16 248 20

150

250 w/o proposed method

with proposed method

only droop method

Time (Hour)0 6

Vo

ltag

e (p

.u.)

0.9

0.95

1

1.05

1.1

18 2412Time (Hour)

0 6 18 2412

Vo

ltag

e (p

.u.)

0.9

0.95

1

1.05

1.1(a) (b)

Time (H:M)10:00 10:15

PV

/Load

(%

)

0

20

40

60

80

100

10:30 10:45 11:00

PV

Load

Time (H:M)10:00 10:15

Vo

ltag

e (p

.u.)

0.96

1

1.04

1.08

10:30 10:45 11:00

Time (H:M)10:00 10:15

Reac

tiv

e P

ow

er (

kV

ar)

-200

0

200

400

10:30 10:45 11:00-400

Agg

Bus No.

24-hour simulation with 15 minutes sampling One-hour simulation with 1 second sampling

0. Outline

1. REIDS Project





Zoom in

Zoom in




▪ Hierarchically Coordinated Operation and Control for DC microgrid clusters

69Q. Xu, Y. Xu*, et al, " A Hierarchically Coordinated Operation and Control Scheme for DC Microgrid

Clusters under Uncertaint," IEEE Transactions on Sustainable Energy, 2020.

0. Outline

1. REIDS Project








▪ Hierarchically Coordinated Operation and Control for DC microgrid clusters

70

Dispatch results

Real-time control results

Simulation results when local

controller responds to the scheduling

results from operation level of MG2

at 9h, 10h and 11h (which is at 10s,

20s and 30s in the simulation)Simulation results of MG2 during 9h-10h with PV and loadfluctuations in Matlab/Simulink.

Q. Xu, Y. Xu*, et al, " A Hierarchically Coordinated Operation and Control Scheme

for DC Microgrid Clusters under Uncertaint," IEEE Transactions on Sustainable

Energy, 2020.

0. Outline

1. REIDS Project








71

▪ Optimal Planning of DERs in Microgrid

Objective: Minimize total investment costs Constraints: operational limits

network constraintscomponent constraints, etc.

Variables: size, site, type, installation year, etc.

0. Outline

1. REIDS Project





Network

model

Uncertain

load

Uncertain

renewablesDER model

Distributed generators Solar photovoltaic, Wind

turbine, Micro-turbine, CCHP

Energy storage system Electrical storage, Thermal

storage, Hydrogen storage

Stochastic programmingRobust optimization Probability-weighted robust optimization…




72

▪ Optimal Placement of Heterogeneous Distributed Generators

Proposed two-stage DG placement method

Z. Li and Y. Xu*, “Optimal Placement of Heterogeneous Distributed Generators in a Grid-Connected

Multi-Energy Microgrid under Uncertainties,” IET Renewable Power Generation, 2019.

,{ ( ) ( ) }

z x

sz CF x CG

Operation StageInvestment Stage

NPV Max F z G x

= − +

in ( | ) in{ ( | ) [ ( | , )]}

. . |

( | , )= in ( | )

( , )

w

w

y

M G x c M S w c E Q w c

s t w CS z

Q w c M L y c

y CL w

= +

System multi-stage operation model

Sub-stages for system operation

D1 D5D4D3D2 D363 D364 D365D362D361

Investment Stage

H4H3H2H1 H23 H24H22H21

RES outputs

Power demands

Hot water needs

Outdoor temperature

Operation Stage

Y1 Y2 Y4 YN-3 YN-2 YN-1 YNY3 Y5 Y6 YN-4YN-5

P1 P2 Pk PTPT-1

P: Investment Phase Y: Year D: Day H: Hour N: Project Lifetime T: Total Phase Number

H4H3H2H1 H23 H24H22H21

Day-ahead Dispatch for 24 hours of Next day

RES outputs

Power demands

Hot water needs

Outdoor temperature

Operation Stage

Intro-day Online Dispatch (Within Each Hour )

Q1 Q2 Q3 Q4 Q96Q95Q94Q93

Day-ahead

forecasting

Short-lead-time

forecasting H: Hour Q: A quarter of hour

0. Outline

1. REIDS Project





Year/Bus 3 6 9 12 18 22 25 27 30 33CCHP unit：Cap-65

1-4 65 195 65 130 0 0 130 65 0 655-8 65 195 65 195 0 65 130 65 65 65

9-12 65 195 65 195 65 65 130 65 65 130Electric boiler: A

1-4 123.2 96 90.7 0 146.8 41.5 167.8 92.9 0 05-8 123.2 96 128.1 0 146.8 88.5 214.8 92.9 0 0

9-12 138.5 96 129.6 0 146.8 120.4 216.2 92.9 0 0Electric boiler: B

9-12 0 0 0.0 57.1 0 66.7 74.1 0 0 0Electric chiller:A

1-4 200.2 0 68.4 0 105.8 0.0 141.3 0 0 48.25Electric chiller: B

1-4 0 0 25.0 0 0 50.7 74.3 0 0 05-8 0 0 50.3 0 0 103.8 115.2 0 0 0

9-12 0 0 90.9 0 0 143.0 115.2 0 0 0Photovaltics: B

1-12 137.6 137.6 136.2 122.2 118.9 137.7 181.6 181.6 177.8 169Wind turbine: A

1-12 80 0 0 0 0 80 80 0 80 0Wind turbine: B

1-12 0 240 120 0 240 0 240 0 120 0




73

▪ Optimal Deployment of Heterogeneous Energy Storage

Z. Li and Y. Xu*, “Optimal Stochastic Deployment of Heterogeneous Energy Storage in a Residential Multi-Energy Microgrid

with Demand-Side Management,” IEEE Transactions on Industrial Informatics, 2020.

TOUT

TIN

TOUT

Heating system

HCT

TOUT

HCIN

HCER

Typical structure of a room in a residential building

Structure of the thermal storage

Risk-averse objective function

1 2, , ...,( ) [ ( ) ( )]

qQ

q qx w y y y

q N

MinG x Min S w c L y

= +

. . |

( , ),

w

q q

s t w CS z

y CL w q

Proposed multi-stage stochastic deployment model

0. Outline

1. REIDS Project








74

▪ Planning results

Deployment Results For Battery Storage (kWh)

Year/Bus 3 6 18 22 25 27 30 331-3 1500 0 0 1500 1500 0 0 04-6 1500 466.0 101.7 1500 1500 0 0 473.27-9 1500 466.0 101.7 1500 1500 378.0 81.19 473.2

Year 1-9 Group 1 Group2 Group3Cooling storage tank 0 0 0

Heat storage tank 1800 1800 1800

Deployment Results For Thermal Storage (kWh)

21

26 27

192021

28

3 4 6 75 8 9 10 11 12 13 14 1715

23 24 25

16 18

22

29 30 31 32 33

Thermal Group #3

Thermal Group # 1

Utility Grid

RES

RES

CCHP

EB

Substation

RES

CCHP

EB

RESRES

RES RESRES

CCHP

EB

Thermal Group #2

Z. Li and Y. Xu*, “Optimal Stochastic Deployment of Heterogeneous Energy Storage in a Residential Multi-Energy Microgrid

with Demand-Side Management,” IEEE Transactions on Industrial Informatics, 2020.

0. Outline

1. REIDS Project








▪ Probability-Weighted Robust Optimization (PRO) for DG Planning

75C. Zhang, Y. Xu*, Z.Y. Dong, "Probability-Weighted Robust Optimization for

Distributed Generation Planning in Microgrids," IEEE Trans. Power Syst., 2018.

Probability-Weighted Uncertainty Sets

Solution Algorithm

PRO Formulation

0. Outline

1. REIDS Project





Problems identification: Robust optimization only

considers the worst case under a single day profile, while

stochastic programming cannot cover full spectrum of

uncertainties and thus full operational robustness.

Our aims: to ensure a full robustness for the short-term

operation under the uncertainties over the long-term

planning horizon.




▪ Probability-Weighted Robust Optimization (PRO) for DG Planning

76

Main

Grid

Substation

Microgrid

Candidate Buses

12

3

45

67

89

10

1112

13

14

15

16

17

18

19

2021

22

2324

25

2627

28

2930

31

32

33

C. Zhang, Y. Xu*, Z.Y. Dong, "Probability-Weighted Robust Optimization for Distributed Generation Planning

in Microgrids," IEEE Trans. Power Syst., 2018.

0. Outline

1. REIDS Project








1. C. Zhang and Y. Xu*, “Hierarchically-Coordinated Voltage/VAR Control of Distribution Networks using PV

Inverters,” IEEE Trans. Smart Grid, 2020.

2. C. Zhang, Y. Xu*, Z.Y. Dong, and R. Zhang, “Multi-Objective Adaptive Robust Voltage/VAR Control for

High-PV Penetrated Distribution Networks,” IEEE Trans. Smart Grid, 2020.

3. Z. Li, Y. Xu*, et al, “Optimal Deployment of Heterogeneous Energy Storage System in a Residential Multi-

Energy Microgrid with Demand Side Management,” IEEE Trans. Industrial Informatics, 2020.

4. Q. Xu, Y. Xu*, Z. Xu, L. Xie, F. Blaabjerg, “A Hierarchically Coordinated Operation and Control Scheme for

DC Microgrid Clusters under Uncertainty,” IEEE Trans. Sustainable Energy, 2020.

5. Q. Xu, T. Zhao, Y. Xu*, Z. Xu, P. Wang, and F. Blaabjerg, “A Distributed and Robust Energy Management

System for Networked Hybrid AC/DC Microgrids,” IEEE Trans. Smart Grid, 2020.

6. C. Zhang, Y. Xu*, and Z.Y. Dong, “Robustly Coordinated Operation of a Multi-Energy Micro-Grid in Grid-

Connected and Islanded Modes under Uncertainties,” IEEE Trans. Sustainable Energy, 2020.

7. Y. Wang, T.L. Nguyen, M. Syed, Y. Xu*, E. Guillo-Sansano, V.H. Nguye, G. Burt, and Q.T. Tran, “A

Distributed Control Scheme of Microgrids in Energy Internet Paradigm and Its Multisite Implementation,”

IEEE Trans. Industrial Informatics, 2020.

8. Y. Wang, M.H. Syed, E. Guillo-Sansano, Y. Xu*, and G.M. Burt, “Inverter-based Voltage Control of

Distribution Networks: A Three-Level Coordinated Method and Power Hardware-in-the-Loop Validation,”

IEEE Trans. Sustainable Energy, 2020.

9. Y. Wang, T. L. Nguyen, Y. Xu*, and D. Shi, “Distributed Control of Heterogeneous Energy Storage Systems

in Islanded Microgrids: Finite-Time Approach and Cyber-Physical Implementation,” Int. J. Electrical Power

and Energy Systems, 2020.

10.Q. Xu, Y. Xu, C. Zhang, and P. Wang, “A Robust Droop-based Autonomous Controller for Decentralized

Power Sharing in DC Microgrid Considering Large Signal Stability,” IEEE Trans. Industrial Informatics,

2020.

▪ Publication list in DER and Microgrid

77




11. Z. Li and Y. Xu*, “Temporally-Coordinated Optimal Operation of a Multi-energy Microgrid under Diverse

Uncertainties,” Applied Energy, 2019.

12.C. Zhang, Y. Xu*, Z.Y. Dong, “Robustly Coordinated Operation of a Multi-Energy Micro-Grid in Grid-

Connected and Islanded Modes under Uncertainties,” IEEE Trans. Sustainable Energy, 2019.

13.Y. Wang, T. Zhao, C. Ju, Y. Xu*, P. Wang “Two-Level Distributed Voltage/Var Control of Aggregated PV

Inverters in Distribution Networks,” IEEE Trans. Power Delivery, 2019.

14.Y. Wang, Y. Xu, and Y. Tang, “Distributed Aggregation Control of Grid-Interactive Smart Buildings for

Power System Frequency Support,” Applied Energy, 2019.

15.S. Sharma, Y. Xu*, A. Verma, et al, “Time-Coordinated Multi-Energy Management of Smart Buildings

under Uncertainties,” IEEE Trans. Industrial Informatics, 2019.

16.Y. Wang, T.L. Nguyen, Y. Xu*, et al, “Cyber-Physical Design and Implementation of Distributed Event-

Triggered Secondary Control in Islanded Microgrids,” IEEE Trans. Industry Applications, 2019.

17.Z. Li, Y. Xu*, et al, “Optimal Placement of Heterogeneous Distributed Generators in a Grid-Connected

Multi-Energy Microgrid under Uncertainties,” IET Renewable Power Generation, 2019.

18.Y. Wang, Y. Tang, Y. Xu*, et al, “A Distributed Control Scheme of Thermostatically Controlled Loads for

Building-Microgrid Community,” IEEE Trans. Sust. Energy, 2019. – Web-of-Science Highly Cited Paper

19.Y. Chen, Y. Xu*, et al, “Optimally Coordinated Dispatch of Combined-Heat-and-Electrical Network,” IET

Gen. Trans. & Dist., 2019.

20.R. Xu, C. Zhang, Y. Xu*, Z.Y. Dong, “Rolling Horizon Based Multi-Objective Robust Voltage/VAR

Regulation with Conservation Voltage Reduction in High PV-Penetrated Distribution Networks,” IET Gen.

Trans. & Dist., 2019.

21.Y. Xu*, Z.Y. Dong, et al, “Multi-timescale coordinated voltage/var control of high renewable-penetrated

distribution networks,” IEEE Trans. Power Syst., 2018.

22.Z. Li and Y. Xu*, “Optimal coordinated energy dispatch for a multi-energy microgrid in grid-connected and

islanded modes,” Applied Energy, 2018. – Web-of-Science Highly Cited Paper


78




23. Y. Wang, Y. Xu, Y. Tang, et al “Aggregated Energy Storage for Power System Frequency Control: A Finite-

Time Consensus Approach,” IEEE Trans. Smart Grid, 2018.

24. Y. Wang, Y. Xu*, Y. Tang, et al, “Decentralized-Distributed Hybrid Voltage Regulation of Power Distribution

Networks Based on Power Inverters,” IET Gen. Trans. & Dist., 2018.

25. C. Zhang, Y. Xu*, Z.Y. Dong, et al, “Probability-Weighted Robust Optimization for Distributed Generation

Planning in Microgrids,” IEEE Trans. Power Syst., 2018.

26. C. Zhang, Y. Xu*, Z.Y. Dong, et al, “Robustly Coordinated Operation of A Multi-Energy Microgrid with

Flexible Electric and Thermal Loads,” IEEE Trans. Smart Grid, 2018.

27. C. Zhang, Y. Xu*, Z.Y. Dong, “Three-Stage Robust Inverter-Based Voltage/Var Control for Distribution

Networks with High PV,” IEEE Trans. Smart Grid, 2017. – Web-of-Science Highly Cited Paper

28. C. Zhang, Y. Xu*, Z.Y. Dong , et al, “Robust operation of Microgrids via two-stage coordinated energy storage

and direct load control,” IEEE Trans. Power Syst., 2017.

29. C. Zhang, Y. Xu, Z.Y. Dong, “Robust Coordination of Distributed Generation and Price-Based Demand

Response in Microgrids,” IEEE Trans. Smart Grid, 2017. – Web-of-Science Highly Cited Paper

30. W. Zhang, Y. Xu*, Z.Y. Dong, “Robust SCOPF using multiple microgrids for corrective control under

uncertainty,” IEEE Trans. Industrial Informatics, 2017.

31. C. Ju, P. Wang, L. Goel, and Y. Xu*, “A two-layer energy management system for microgrids with hybrid

energy storage considering degradation costs,” IEEE Trans. Smart Grid, 2017.

32. Y. Xu, Z.Y. Dong, K.P. Wong, et al, “Optimal capacitor placement to distribution transformers for power loss

reduction in radial distribution systems,” IEEE Trans. Power Systems, 2013.

*Corresponding author

To appear in 2021: Y. Xu, Y. Wang, C. Zhang, and Z. Li, “Coordination and optimization of distributed energy

resources in microgirds,” IET Book Press.79





temporal-spatial coordination of distributed energy ... and microgrid-0613.pdf · "a...

Documents