ANALYSIS AND CONTROL OF MODES OF POWER SYSTEM FOR ENHANCING OPERATIONAL EFFECTIVENESS OF POWER MARKET

MINISTRY OF EDUCATION AND TRAINING
THE UNIVERSITY OF DANANG

TRAN PHUONG NAM

ANALYSIS AND CONTROL OF MODES
OF POWER SYSTEM FOR ENHANCING
OPERATIONAL EFFECTIVENESS OF
POWER MARKET

Major: Power System & Network
Code: 62.52.50.05

SUMMARY OF DOCTORAL DISSERTATION

Danang - 2016


This dissertation has been finished at:
The University of Danang

The supervisors 1: Ass. Prof. Dinh Thanh Viet
The supervisors 2: Prof. La Van Ut

Reviewers 1: Prof. Tran Dinh Long
Reviewers 2: Ass. Prof. Vo Ngoc Dieu
Reviewers 3: Prof. Nguyen Hong Anh

The dissertation is protected before the council meeting marked
doctoral dissertation at the University of Danang in December 26,
2015 08:30am.

Dissertation can be found at:
- National library of Viet Nam, Hanoi.
- Learning & Information Resource Center, the University of Danang.


1

INRODUCTION
1. Rationale
Nowadays, the operation of power systems in the power market
model has been researched and applied by many countries. In
Vietnam, starting on July 1st, 2012 competitive source power market
has been operation. Problem analysis and control of power system
mode to enhance the operational efficiency of the power market, it has
an important role and great significance. In particular, assessement
and control of congestion as well as voltage stability power system
are not only safe operation of power system but also economic
operation of power market, it should be considered study.
2. Aims of study
- Study and analysis of the power system congestion in power
market operation.
- Study and analysis of voltage stability of power system in
power market operation.
- Study of FACTS devices for congestion control, voltage
stability control and power market welfare.
- Design website model for congestion as well as voltage
stability monitoring in spot market operation.
- Study of design congestion as well as voltage stability
monitoring model in spot market operation.

3. Subjects and scope of study
- Subjects: LUF factor, dV/dQ, LMP and social welfare; neural
network; SVC and TCSC; information technology infrastructure of
power market.
- Scope of study: congestion and voltage stability in power
market operation of IEEE 39 bus and Vietnam power system in 2016.


2

4. Methodology
From document, experiment study and mathematical model.
CHAPTER 1: OVERVIEW OF POWER MARKET
AND POWER SYSTEM OPERATION
IN POWER MARKET MECHANISM
1.1. Deregulate electricity sector and towal competitive power
market
Electricity sector restructuring has been researched and applied
by many countries. Depending on the characteristics of the power
system, political mode, the restructure purpose of each country is
different so that electricity sector restructuring is different.
1.2. Introduction to competitive power markets
Power market models has been introduted as: competitive power
generation market with a single buyer, competitive wholesale power
market, competitive retail power market.
1.3. Introduction to power market model in the world and
Vietnam
Power market models has been introduted as: Asia power market,
Australia power market, United states of america power market,
Europe power market, Vietnam power market.

Generation market
2005

2009

Testing

Wholesale market
2014

Full

2016

Testing

Retail market
2022

Full

2024

Testing

Full

Fig 1.5: The levels of development of the Vietnam power market

3

1.4. Power market operation
1.4.1. Day ahead market and spot market
SMO

GenCo/
DisCo

● Check quotations
● mobilization schedule
of day ahead market

● Mobilization schedule of
hour ahead market
● Dispatch

Day ahead market
(Day D-1)

Spot market
(Day D)

● Quotations of day D

● Announces the availability power
level (only in case of problems)
● Operates generators following
dispatch

● Calculate
price market
and payment
Day D+1

Fig 1.7: Day ahead market and spot market
1.4.2. Bid, auction and mobilization schedue in power market
The auction is arranged according to the capacity ranges from
low to high price for supply curves and vice versa for demand curves.
Auction in the power market is operated by SMO, which is based on
the quotations of the buyer and the seller.
1.4.3. Optimal power flow in the power market
The objective function is a minimum of social welfare in total:
nG

nD

i 1

j 1

Min( CGi ( PGi )   BDj ( PDj ))

(1.3)

1.4.4. Some economic criteria in the power market
Locational marginal price - LMP at Busi is three components
included in the marginal price at reference bus, marginal loss cost
from reference bus to Busi and marginal congestion price from
reference bus to Busi:

LMPi  i  ref  lossi  congestioni ($ / h)

Revenue of GenCo and payment or DisCo payment:

(1.11)


4

RGenCoi ( DisCoj )  LMPi  Pi

(1.12)

($ / h)

Profit of GenCo and DisCo is determined as:
GenCoi  RGenCoi  CGenCoi ($ / h)

(1.13)

DisCoj  BDisCoj  RDisCoj ($ / h)

(1.14)

Hourly social welfare can be determined as:
nG

nD

i 1

j 1

 hSocial   GenCoi    DisCoj ($ / h)

(1.15)

Moreover, annual electricity losses can be determined as:

A  Pmax  (MWh / year )

(1.16)

Thus, cost for annual electricity losses can be determined as:

Y  A  c ($ / year )

(1.17)

1.5. Analysis and control modes of power system in power market
operation
1.5.1. Analysis and control congestion
- Congestion analysis methods: Congestion analysis on line
utilization factor - LUF, congestion analysis on locational marginal
price - LMP.
- Methods of congestion control and management: using
economics method such as locational marginal price - LMP, Financial
Transmission Right - FTR; using techniques such as building new
lines, upgrading and overload protection for the old line; adding
compensation devices to increase power system control (fixed

compensation and FACTS devices…)


5

1.5.2. Analysis and control voltage stability
- Voltage stability analysis methods: Continuation power flow CPF, V-Q sensitivity.
- Voltage control methods: Excitation current adjusting of the
generator, output voltage adjusting of transformer by voltage fixel
divider or on load tap changer - OLTC, using compensation device for
change voltage loss on the line as synchronous compensator, fixed
capacitor, FACTS devices…
1.5.3. The impact of power system modes control on power market’s
criterias
Power system modes control will change power market’s
criterias in formulas 1.11, 1.12 and 1.17.
CHƯƠNG 2: ANALYSIS AND ASSESSMENT
VOLTAGE STABILITY OF POWER SYSTEM IN
POWER MARKET OPERATION
2.1. Background
With analysis and control of power system’s modes in power
market condition, analysis and assessement voltage stability through
dV/dQ sensitivity will be implemented the lager number of calculation
to implement system of equations. Power system in power market
condition is heavily influenced from the power market elements (bid,
auction, mobilization schedule…). Need to find a new analysis and
assessement voltage stability for power market.
2.2. Voltage stability of power system

6

Voltage stability related to the ability to capacity balance of the
load buses, particularly reactive power balance.
2.3. Analysis and assessement voltage stability of power system in
spot market
2.3.1. The impact of quotations element on analysis and assessment
of power system
The objective function relating to quotations:
ng

ng

i 1

i 1

Min (CGi ( PGi )) D1 Min (ai  bi PGi  ci PGi2 )

(2.4)

The system of equations is determined from Newton -Raphson:
J PV    
J QV   V 

 P   J P
 Q    J
   Q

(2.5)

J R1 matrix is determined as:
J R1 

V
Q

(2.8)

Diagonal elements of J R1 is V-Q sensitivity at buse.
2.3.2. Artificial neural network for technical analysis and
assessement
Neural network

Input

p

w

n



f

a

b
1

Fig 2.5: Neural network with one input


7

Types of error of neural network assessement between target
value ti and assessement value ai can be determined as:
Actual error:

ei  ti  ai

(2.10)

Absolute error:

ei  ti  ai

(2.11)

MAE 

Mean Absolute Error:

1 n
 ei
n i 1

Mean Absolute Percentage Error: MAPE 

(2.12)

1 n ei
 100 (2.13)
n i 1 ti

1 n
(2.14)
 (ei )2
n i 1
Where: ti is target value, ai is assessement value, n is the
MSE 

Mean Squared Error:

number of text sample.
2.3.3. Design neural network for application of analysis and
assessement voltage stability of power system in spot market
- For basic neural network structure:
Input

Hidden

Output

PgT
SCADA system

T

Qg

.
.
.

.
.
.
(dV/dQ)T

Pl T
QlT

.
.
.

.
.
.

VT

Fig 2.6: MLP neural network model with basic structure
- For neural network structure of proposal dissertation:


8

Input

Hidden

Output

PgT
SCADA system
QgT

.
.
.

.
.
.

Pl T
QlT
Biding system
T

V

(dV/dQ)T
.
.
.

.
.
.

CgT

Fig 2.7: MLP neural network model with structure of proposal
dissertation (VSA-PM)
2.3.4. Algorithm flowchart of neural network training in voltage
stability assessement of power system in spot market
Algorithm flowchart is impletemented as: create neural
network; create learning text sample; create weight, bias, the number
of iterations; training by Levenberg - Marquardt algorithm; checking
stop condition; saving neural network structure with new weight and
bias; text neural network.
2.3.5. Application of neural network in analysis and assessement
voltage stability of IEEE 39 bus power system in spot market
- Analysis neural network training perpormance:

Fig 2.11: Error MSE: (a) basic structure, (b) proposal structure


9

- Text and analysis neural network with examined data:
(a) Distribution of actual error

(b) Distribution of absolute error, MAE=0.0005
Absolute error
MAE

(c) Distribution of absolute percentage error, MAPE=0.038%
Absolute percentage error
MAPE

Fig 2.14: Distribution of actual error, absolute error and MAE
with basic structure
(a) Distribution of actual error

(b) Distribution of absolute error, MAE=0.0005
Absolute error
MAE

(c) Distribution of absolute percentage error, MAPE=0.038%
Absolute percentage error
MAPE

Fig 2.17: Distribution of actual error, absolute error and MAE
with proposal structure


10

- Text and analysis neural network with mutations data:

Bus

Bus

(a)

(b)

Fig 2.20: Basic structure: (a) MAE, (b) MAPE

Bus

Bus

(a)

(b)

Fig 22: Proposal structure: (a) MAE, (b) MAPE
Therefore,

proposal

neural

network

structure

shows

assessement performance through error of MSE, MAE, MAPE are
better than the basic neural network structrure for IEEE 39 bus power
system.
2.3.6. Application of neural network for analysis and assessement

voltage stability of southern Vietnam power system in 2016 in spot
market
- Analysis neural network training perpormance:


11

Training
Best

Epochs
(b)

Fig 2.23: Error MSE: (a) basic structure, (b) proposal structure
- Text and analysis neural network with examined data:
(a) Distribution of absolute error, MAE=4.13e-07
Absolute error
MAE

(b) Distribution of absolute percentage error, MAPE=0.5887%
Absolute percentage error
MAPE

Fig 2.24: Distribution of absolute error, absolute percentage error
with basic structure
(a) Distribution of absolute error, MAE=2.89e-07
Absolute error
MAE

(b) Distribution of absolute percentage error, MAPE=0.4370%

Absolute percentage error
MAPE

Fig 2.25: Distribution of absolute error, absolute percentage error
with proposal structure


12

- Text and analysis neural network with mutations data:

Sample value (PowerWorld)
ANN value (Matlab)

Bus

Fig 2.28: Assessement dV/dQ, error MAE and error MAPE of
basic structure with disconnected branch 500KV
Tan Dinh - Song May

Sample value (PowerWorld)
ANN value (Matlab)

Bus

Fig 2.29: Assessement dV/dQ, error MAE and error MAPE of
proposal structure with disconnected branch 500KV
Tan Dinh - Song May

13

Therefore,

proposal

neural

network

structure

shows

assessement performance through error of MSE, MAE, MAPE are
better than the basic neural network structrure for Southern Vietnam
bus power system in 2016.
CHAPTER 3: STUDY AND APPLICATION OF FACTS
DEVICE CONTROL OF MODES OF POWER SYSTEM FOR
ENHANCING EFFECTIVENESS OF POWER MARKET
3.1. Background
With premium features and benefits of FACTS devices,
research and installation of FACTS will be expected the effective
operation of the power system and power market.
3.2. Overview of FACTS device
With specifications, SVC, STATCOM, UPFC devices can be
utilized to enhance voltage stability as well as TCSC, UPFC devices
can be utilized to control power flow.
With economic criteria, cost of SVC and TCSC following
KVAr are cheaper than other FACTS devices.

Dissertation choose SVC and TCSC to calculate and connect
operational power system in conditional power market.
3.3. Application of choice placement of FACTS device for modes
control of power system in power market operation
3.3.1. The impact of FACTS device to modes control of power system
to power market’s criterias
Application of FACTS device will change power flow in power
system, so criterias of GenCos, DisCos and market will change as
formulas from 1.13 to 1.17.


14

3.3.2. Methodology of choice placement of FACTS device into power
system in power market operation
With methodology of choice placement of TCSC as well as SVC
into power system in power market operation, it will be analyzed the
technology first, and then it will analyze the economic after.
3.3.3. Assessment FACTS device efficiency to enhance benefit of
electricity losses
Dissertation proposes choice of placement of FACTS device
following criteria for benefit of electricity losses fall:
(3.20)
BLFACTS  D(A )  c ($)
Criteria for benefit of electricity losses fall with scenarios:
n

scenarioi
BLFACTS   BLFACTS
($)

_ scenarioi  t

(3.21)

i 1

3.4. Analysis and choice of placement of FACTS device in IEEE
39 bus power market
3.4.1. Analysis and choice of placement of SVC
SVC is located alternate weak buses of voltage stability, and P-V
is analyzed next, effectiveness achieves positive effects. The slope of
the P-V curves are reduce, voltage Vmin of curves are improved with
the connection cases of SVC.
Tab 3.3: Change criterias of P-V cuver with connection cases of SVC
Without SVC

Bus 26

Bus 27

Bus 28

Bus 29

Ptotal-max (pu)

4,725

4,725

4,725

4,725

4,725

Vmin (pu)

1,015

1,016

1,016

1,028

1,053

With two SVC effective cases, case of SVC connection at bus
SVC

29 shows the highest economic efficiency through BL

value.


15

Control value
Average value

Fig 3.14: Annual BLSVC criteria in connection cases of SVC with
control V=1,02pu and V=1,03pu
3.4.2. Analysis and choice of placement of TCSC
With two TCSC effective cases, case of TCSC connection at
TCSC

branch 4-14 shows the highest economic efficiency through BL
value.
Control value
Average value

TCSC

Fig 3.19: Annual BL
criteria in connection cases of TCSC with
control XTCSC= 0,5XL and XTCSC= 0,7XL
3.5. Analysis and choice of placement of FACTS device in
Vietnam power market in 2016
3.5.1. Analysis and choice of placement of SVC


16

With three SVC effective cases, case of SVC connection at bus
Tan Dinh shows the highest economic efficiency through BLSVC value.
Control value
Average value

Fig 3.26: Annual BLSVC criteria in connection cases of SVC with

control V=1,02pu and V=1,03pu
3.5.2. Analysis and choice of placement of TCSC
With two TCSC effective cases, case of TCSC connection at
branch Di Linh - Tan Dinh shows the highest economic efficiency
through B/C value.
Control value
Average value

TCSC

Fig 3.27: Annual BL
criteria in connection cases of TCSC with
control XTCSC= 0,7XL and XTCSC= 0,75XL


17

CHAPTER 4: STUDY AND DESIGN OF MONITORING AND
CONTROL MODES OF POWER SYSTEM IN
POWER MARKET OPERATION
4.1. Background
Nowadays, the information technology infrastructure of
electrical sector in many countries around the world are upgrading
development. However, when the power market operation, the
completion of hardware and software of the information technology
infrastructure is an important issue of power market.
4.2. Overview of SCADA/EMS system
SCADA/EMS is supervisory control and data acquisition/Energy
Management System. This system is the optimal control center
solution to enable a secure and efficient operation of the electric

power system.
4.3. Introduction of SCADA/EMS system of Vietnam power
system
According to overall assessment, SCADA/EMS should be further
upgraded to meet the high operation requirements as well as
mechanism power system.
4.4. Connection of SCADA/EMS with operational system of power
market
In mechanism power market, SCADA/EMS system is connected
directly to power market operational system. SCADA/EMS system
focus to collect, monitor and control power system, operational
system of power market focus to operate transaction of the members
participating in the power market.


18

4.5. Website model design for monitoring of voltage stability and
congestion in spot market (WEB-CVM-PM)

Fig 4.7: Page of congestion collection data

Fig 4.8: Page of congestion assessement


19

Fig 4.9: Page of voltage stability collection data

Fig 4.10: Page of voltage stability assessement

ASP.NET language is used to design web. The web structure
includes one the homepage Index and six main page for descriptions,
user manual, statistical data collection and congestion assessement,
statistical data collection and voltage stability assessement.


20

4.6. Proposal of model for monitoring and control of congestion of
power system in spot market
The model combine between the voltage stability assessement
method with LUF factor for monitoring, assessement and control
congestion in spot market is proposed as Fig.4.11. In this proposed
model, the parameters signal of transmission capacity on the branchs
are sent from RTU and SCADA through SCADA/EMS system of
electrical power system, the parameters of capacity limits on the
branch are sent from data center as proposed in chapter 2. Website
implement the publication and monitoring of congestion. In addition,
TCSC with suitable location has been selected and adjustment
(mentioned in Chapter 3) will be used to control congestion of power
system to enhance social welfare.
Power system

Signal from RTU/IED

SCADA/EMS system
MVAij
System data

MVAijmax

Assessement LUF on-line

WEB-CVM-PM
monitoring

No congestion

Congestion
TCSC
Congestion control

Fig 4.11: Model of monitoring and control of congestion of power
system in spot market


21

4.7. Proposal of model for monitoring and control of voltage
stability of power system in spot market
In this proposed model, the parameters signal of buses are sent
from RTU and SCADA through SCADA/EMS system of electrical
power system, bids of spot market hour ahead market are sent GenCos
through bid system of power market. MLP neural network is
applicated to design voltage stability assessement as proposed in
chapter 2. Website implement the publication and monitoring of
voltage stability. In addition, SVC with suitable location has been
selected and adjustment (discussed in Chapter 3) will be used to
control voltage stability of power system to enhance social welfare.
Power system

Quotation from GenCos

Calculate dV/dQ
off-line

System data

Training MLP
off-line

MLP data

Auction
system

Signal from RTU/IED

SCADA/EMS
system

VSA-PM
asesses dV/dQ on-line

WEB-CVM-PM
monitoring

Stability

No stability

Off-line training
Monitoring and control on-line

SVC
voltage control

Fig 4.12: Model of monitoring and control of voltage stability of
power system in spot market


22

CONCLUSION AND FUTURE WORK
1. Conclusion
- The dissertation has proposed a new method (VSA-PM) to
analysis and assess voltage stability of power system in power market
operation. The model combines between the MLP neural network with
dV/dQ sensitivity. MLP neural network structure is added a vecto C gT
that are related to information from GenCos’s quotations.
- The dissertation has applicated VSA-PM to analysis and
assess error of MSE, MAE and MAPE in IEEE 39 bus and southern
Vietnam in 2016. The new method VSA-PM has achieved positive
results in voltage stability in spot market. This new method has been
more effective than the structure of the previous study (reduction of
error MSE, MAE, MAPE).
- The dissertation has proposed method of choice of placement
of FACTS device (SVC, TCSC) for power system in power market
operation with techno-economic criterias. Technical criteria is
boundary conditions, economic criteria is prerequisite. Technical
criteria is LUF factor for TCSC and dV/dPtotal (curve slope P-V) for

SVC; the considered economic criteria is benefits in electricity losses
fall of FACTS device.
- With new proposal method, dissertation applied to calculate
and analyze in IEEE 39 bus power system and southern Vietnam
power system in 2016. In the case of IEEE 39 power system, the most
efficient location to install SVC is the connection bus 29, the location
for TCSC is at branch 4-14. In the case of south Vietnam power
system in 2016, the most efficient location to install SVC is at bus Tan
FACTS

Dinh, TCSC at branch Di Linh - Tan Dinh. Moreover, via the BL


23

criteria, the dissertation has compared the efficiency between SVC
and TCSC, and the results have shown that TCSC brings about more
economic values than SVC.
- The dissertation has proposed a website model for monitoring
of voltage stability and congestion in spot market (WEB-CVM-PM).
- With new proposal website, dissertation has applied a
monitoring and control of congestion as well as voltage stability in
spot market. This model combines between proposal method of
congestion and voltage stability, WEB-CVM-PM and SVC/TCSC.
2. Future work
- In technology development of integrated circuit technology, it
should research to design a microchip controller for voltage stability
assessement of power system in power market through VSA-PM
method.
- In the ultra high voltage power system with great length as

Vietnam power system, Northern’s operational problems had little
effect on the South and vice versa. Maybe study to find the optimal
placement of FACTS devices for each region. Therefore, the proposed
algorithm and methodology of mixed FACTS devices (the same type
or different type) through technology and economic criteria will be
studied for the future.