You are on page 1of 4

2004 IEEE Intemational Conference on Electric Utility Deregulation, Restructuring and Power Technologies (DRPT2004) April 2004 Hong


Risk Based Spinning Reserve Allocation in

Deregulated Power Systems
Z Song, L Goel, Senior Member, IEEE, and P Wang,

Abstract-. In a deregulated power system, spinning reserve participants. Unit Commitment (UC) risk was once used
can be allocated to different ancillary service providers at to evaluate the SR problem in a traditional power system
different locations based on the supplier bids and customer [3]. It is defined as the probability of the system load
choices on the reliability levels required. Different inadequacy.
allocations of spinning reserve are expected to create
However, in a deregulated power market SR units can
different impacts on the reliability of the physical power
system. When transmission failures occur in a physical be allocated to different locations of a physical system.
power system, the system constraints may be violated and This may have much impact on the UC risk due to the
spinning reserve may not be immediately committed to the occurrence of line overloads when transmission lines fail.
system to serve the customers. This paper investigates the Risk Based SR Allocation Method (RBSRAM) is
impact of transmission system failures on spinning reserve proposed in this paper. It uses the contingency
allocation. A technique is proposed to determine the enumeration technique under each possible SR allocation
locations of spinning reserve based on the minimum unit schedule to find the minimum UC risk and the
commitment risk. corresponding SR allocation. The SR market model,
which is combined,with this method, is also illustrated in
this paper. The details of this method are described in
Key words-spinning reserve allocation, risk taking, unit Section II of this paper. The SR market model involved
commitment risk. Nomenclature
with this allocation method is investigated in Section 111.
Selected system studies using this method are shown in
Section IV, and the conclusions are presented in Section
Spinning Reserve (SR) management has evolved since the
onset of deregulation of power systems. SR is purchased
through a bilateral contract or in a SR market, which is 11. RISKBASEDSPMNlNG RESERVE ALLOCATION
independent of the energy market. In a real time SR METHOD(RBSRAM)
market, a new method to determine the SR market
Figure 1 shows the flowchart of this new method.
clearing price and dispatch schedule is generally based on
Taking the unit commitment risk to evaluate the SR
a proposed model with a certain optimization objective,
such as to maximize market benefit [I]. However, under allocation enables the physical power system to be more
certain unexpected situations such as transmission line reliable under system failures. In the proposed approach,
failures, system congestions will occur. Consequently the the reliability required SR capacity would be allocated
cleared SR would not be able to get committed to the among the selected SR providers based on the minimum
system. This will result in undesirable reliability risk taken. The cost for each generating unit is assumed to
problems. be equal in the current version of this method.
Hence new techniques should he developed to solve A. Contingency enumeration
this reliability problem in a deregulated power market.
The contingency enumeration method checks all the
Reference [2] points out that it is extremely helpful to
possible states of a physical system. The analytical
think of reliability primarily as a risk taking and
technique [3], a probabilistic measure of the system
management process. Actually in most practical power
reliability, is used in the contingency enumeration within
markets the responsibilities for risk taking have been
RBSRAM. Up to second-level contingencies were
clearly defined through bilateral contracts among market
investigated in RBSRAM. These include up to two units'
outage and two transmission lines' outage. Assuming that
2 Song i s with Powcr Market Rcsearch Group. School of Electrical all the units and transmission lines can exist in two
& Electronic Engineenng., Nanyangm Technological Univcrsiry,
Singapore (e-mail: songzitong@pmail.nN.~d".~g). (up or down), the probability of a unit or a transmission
L Gael is with Power Markct Research Group, School of Elcetrieal& line on outage at time T can be shown by (1) as follows:
Electronic Engineering, Nanyang Technological University, Singapore
(E-mail: pd = (1)
P wang i s with power Market Research croup, schoolof Eicctncal Considering a system with n generating units and m
& Electronic Engincenng, Nanyang Technological Univcrsiry, Singapore transmission lines, the probability of all components being
available is

49 1
2004 IEEE International Conference on Electric Utility Deregulation, Restructuring and Power Technologies (DRPT2004) April 2004 Hong Kong

m The line power flow e, between node iand j can be

'4 =n(l-
P d @ , n ( l - PdU)
(') calculated in (6)

When the failure of each component in this system is e, =-Bi-6,

independent, the probability of 7 components outage can
Where is the resistance between node i and j
he expressed in (3)
C. Transmission line constraints
Once the transmission congestion problem occurs, the SR
(3) will have a risk that cannot be committed to the system.
Under system failure conditions the power flow has a
probability that violates the transmission limitation, which
can be expressed by (7)

Q Start

Electricity and SR biddings

Pfl 5 5-
The power flow P+, on line / cannot exceed the

transmission limitation F,"'" on line I .

SR allocation i
Generating contingency j Spinning reserve market is a market that is independent
of the energy market. Like the energy market it can he
gcneration hlfilled through real-time bidding or through bilateral

k;. ..
- line contingency k
contracts. Bilateral contracts are arranged between market
participants. IS0 only acts as the manager of bilateral
contracts. However, in real-time bidding the IS0 offers its
own auctions for spinning reserve services, and the
DC flow analysis Calculatc UC reserve will he allocated by the ISO. Sequential dispatch
risk for statc j and simultaneous (joint) dispatch are two current
altematives for dispatching energy and reserve [ 11.
Simultaneous approach is based on formulating the
Calculate U C risk for 5 dispatching problem in the context of energy and ancillary
1 / 1 1 optimization coordination. In a sequential dispatch, the
energy market is first cleared followed by the settlement
of reserve market. The optimization of energy and
spinning reserve is done separately. This paper
investigates the risk of spinning reserve allocation in a
real-time bidding market with sequential dispatching. An
hourly bidding model is used, whose time frame can be
shown in Figure 2 [ 5 ] .
Determine SR allocation

Figure 1 Flowchart of RBSRAM methad

B. DCloadflow -ww=Yofrr
The risk evaluation method is based on the DC load
flow. This function can be shown in (4)
PIP = BOB (4)
where psp is the of injected power into each Figure 2 Time frame of the functioning o f the SR scwicc market
node, B~ is the symmetric flow Jacobian matrix and B is
the vector Of power angle Of each node. The can be ' After the energy market settles down the IS0 offers its
own auctions for SR, and SR providers submit their bids
solved in (5) for hour "k" at hour "k-I". The cleared SR will be
e=Bc'p'p (5) dispatched between hour k to k+l. This is repeated for the

2004 lEEE International Conference on Electric Utility Deregulation, Restructuring and Power Technologies (DRPT2004) April 2004 Hong Kong

next hour and so on. This paper applies the RBSRAM

method to the SR market. The model for this market
structure is shown in Figure 3. The program to simulate
the effects of the unit commitment risk (UCR) on the SR
allocation is discussed in the following sections in detail.


I for cvaiuaiion I

DMinumum risk and corresponding

SR allocation

Figure 3 SR market madel with RBSRAM

The proposed technique was applied to the IEEE-
Reliability Test System (RTS) [6]. The single-line
diagram of the 24-bus IEEE-RTS is shown in Figure 4.
The system has 10 generating buses, I O load buses, 33
transmission lines, 5 transformers and 32 generating units. Figure 4 Single-line diagram of the IEEE-RTS
The total installed generating capacity for this system is
3405 mW. The peak load in one day was assumed to be TABLE I
2850 mW. By using the reliability equivalent technique
[7], the generation system is restructured into three
Gencos. Gencol owns the generation providers at buses 1,
2 and 7, which consists of I 1 generating units. Eight
generating units connected to buses 18, 2 I and 22 belong
to Genco3. A total of 13 generating units connected to
Buses 13, 15, 16 and 23, constitute Genco2. We take the
peak load hour as the consideration point within 24 hours.
The objective of RBSRAM is to find the minimum
commitment risk and the corresponding SR allocation.
After the energy market is cleared, there are 12 units left
in this system, which can be bid to the SR market. The
units left are shown in Appendix A.
System reliability required SR capacity is about 10%
of load (the deviation is less than 3 mW). Considering all
the 12 units left for SR bidding, there are 4 7 unit
combinations whose total capacity satisfing the SR
capacity requirement. In addition, there are 7 SR
allocation schedules after the unit combinations are
categorized by Genbus. The UC risk and the
corresponding SR allocation can be seen in Table 1.
It can be seen from Table 1 that in last SR schedule the
UC risk has the minimum value of 0.00044601. Based on
this SR allocation the IS0 can purchase 216 mW SR from
Genco 1, 24 mW from Genco 2 and 50 mW from Genco

2004 IEEE Intemational Conference o n Electric Utility Deregulation, Restructuring a n d P o w e r Technologies (DRPT2004) April 2004 H o n g K o n g

V. CONCLUSIONS [7] P. Wang, “Impact of Gcncration Reserve Management on

Reliability in Electric Power Market”, Proc. IPQC 2002, Singapore,
This paper has proposed a new risk evaluation method Octobcr 2002.
named RBSRAM and illustrated a SR market model in
conjunction with this method. It has been shown that
under certain system failure conditions SR allocation can VIII. BIOGRAPHY
significantly affect the UC risk. In such situations, the
cleared SR has a high possibility of not being dispatched Zitang Sang was bom in Changchun in China. on August 9, 1977. She
obtained the B.Eng degree in Shanghai Jiao Tong University, China,
to the system. The risk evaluation method becomes more 2000. Her employment cxpericncc includcd thc Changchun Powcr
feasible in such reliability assessment of deregulated Company. Her research interest i s spinning ~ C S C W C mangament in the
power systems. Risk taking should also be considered by deregulated power markct.Now she is pursuing the Ph.D. degrec in
Nanyang Tcchnological University, Singapore.
the IS0 in determining the SR allocation schedule. The
RBSRAM method can help the ISOs to determine SR
allocations among all the SR bidders based on the
minimum unit commitment risk.


II I I 20 0.0354610 0.3546100 I
21 2 I 20 0.0909091 0.9090910 I
22 2 I 20 0.0909091 0.9090910 I

24 2 I 76 0.0909091 0.9090910 I
231 23 2 155 0.0080429 0.0504286 I


[I] Kwok W. Chcung. Payman Shamroblahi. David Sun, James
Milligan, and Mike Potishnak, “Energy and Ancillary Service
Dispatch for Ule Interim IS0 New England Electricity Market”,
IEEE Transactions on Power Systems, Vol. 15, No. 3, August
[2] M. I l k , I. A m , Y. Yoan. and E. Fumagali, “Assessing Reliability
as the Electric Power Industry Re~rmChlres’’,The Electnc Power
Engineering Joumal, March 2001.

[3] Roy Billinton, Ronald N. Allan, “Reliability Evaluation of Power
Systems”. Second Edition, 1996, Plcnum Prcss, USA.
[4] Mar@ Ilic, Francisco Galiana, Lester Fink, “Powcr System
Rcstruch~ring: Engineering and Economics”, 1998, Kluwer
Acadcmic Publishers, USA.

Papers Presented ai Conferences (Unpublished)

[5] Jin Zhong, Kankar Bhattacharya, “Design of Competitive Markets
for Spinning Reserve Service”, Powcr Engineering Swiety
Summer Mceting, 2002 IEEE, Vol 3, 21-25 July 2002, Chicago,
[6] C. Grigg, “The IEEE Reliability Test Systcm: 1996”. Paper 96
whl 326-9 PWRS, IEEE PES Winter Power Meeting 1996.

Papersfrom Conference Proceedings (Published):