Download as pdf or txt
Download as pdf or txt
You are on page 1of 6

1

2006 IEEE PES Transmission and Distribution Conference and Exposition Latin America, Venezuela

Availability & Reliability Evaluation of Dokan


Hydro Power Station
A. R. Majeed and N. M. Sadiq

Scheduled outage: “an outage that results when unit is


deliberately taken out of service at selected time, usually for
purpose of construction, preventive maintenance, repair, or
reserve” [1].

II. INTRODUCTION

D OKAN Hydro Power Station (DHPS) has an installed


capacity of 400 MW. It consists of 5 identical
independent units of 80 MW capacity per each. DHPS
is lying on Lesser Zab River. It is affiliated to Dokan town in
Sulaimany governorate and it is about 290 Km northeast to
Baghdad capital of Iraq. Dokan dam was built in 1954-1957,
and power station building was built in 1975-1977. The power
station equipment has been installed by the Russian Company
Index Terms--Hydro Power Plant, Markov Modeling, Steady (TECHNOPROMEXPOET).
State Probability, Failure and repair rate. Each unit of DHPS units consist of several sub-units such as
Penstock, Butter fly valve, Spiral case, Turbine, Generator,
Excitation system, Speed governor,…etc. We will focus on
NOTATION
those sub-units that cause the unit failure. The unit failure will
A availability affect the Availability and Reliability of the unit and the
R reliability power plant.
U unavailability Evaluating Availability and Reliability will
ENG energy not generated A. Play an important role of knowing performance,
Ȝi failure rate of state i ability, and weakness of each unit.
μi repair rate of state i B. Help planning and deciding periodical maintenance,
fi frequency of state i replacing or repairing when failure occurs.
Pi probability of state i The following is review of related papers and books.
i unit number i is down Methodology of finding the steady-state probability of
Subscript S denotes to scheduled continuous Markov model is presented in [1]. The elements of
Subscript F denotes to force Markov reliability analysis are presented in [2]. An evaluation
of generator Markov models used to obtain indices of unit
I. NOMENCLATURE reliability is represented in [3]. Reliability of generation
system containing multiple hydro plant using simulation
Availability: “the proportion of time, in the long run, that is in, techniques is evaluated. Hydro plant is modeled considering
or ready for, service” [1]. the effects that system operating and water management
Reliability: “the probability of a device or system performing policies have on the reliability indices [4]. An animated power
its function adequately for the period of time intended under generating system reliability performance evaluation model is
the operating conditions intended” [1]. represented. The model has real-time animation capabilities
Energy not generated: the amount of energy not generated due which allow the user to view graphically the interaction
to failure occurrence. among various system components [5]. The reliability of
Forced outage: “an outage result from emergency conditions hydro power equipment is evaluated which establishes an
directly associated with component or unit requiring that unit engineering basis for rehabilitation investment decision [6].
be taken out of service immediately, either automatically or as Reliability of individual generating station is evaluated that
soon as switching operations can be preformed” [1]. concerned with individual generating station adequacy and
security [7].
N. M. Sadiq is M.Sc. student, University of Sulaimany, Iraq (e-mail:
nawzad_977@yahoo.com).
Dr. A. R. Majeed is the Headmaster of Department of Electrical Engineering,
Sulaimany University, Iraq (e-mail: drassomajeed2000@yahoo.com).

1-4244-0288-3/06/$20.00 ©2006 IEEE


2

III. MATHEMATICAL BACKGROUND Up-state


To model a unit or a system we use Markov model, which is a
state-space representation. The matrix form of this
μ
representation is [1]:
λ
P ( t ) = P ( t ) A (1)
Down-state

Where P (t) is a row-vector of derivative of the state Fig. 1. Two-state model


probability consists of dP1 (t)/dt, dP2 (t)/dt,………; P(t) is a
row-vector consist of the elements of state probabilities; and The hydro-unit is transit from up-state to down-state, either
A is the transitional rate matrix with elements aij=Ȝij for ij, due to forced or scheduled outages.
and aij=-™Ȝij for i=j. The elements of each row of matrix A To derive the Markov model of a Hydro-unit we assume:
always add up to 0. The steady-state probability can be 1. The failure and repair rates are exponentially
determine by much simpler task of solving the set of linear distributed.
equations 2. There is no transition between the scheduled and
force outages. The unit after repairing is immediately
returning to up-state.
PA=0 (2)
From the above definition a developed Markov model is
solution of (2) requires an additional equation, which is driven as follows:
obtained from the fact that the probability of states must
always add up to 1; that is, μ S
Up-state
μ F

=1 (3) UpĹ
¦P
i
i
DownĻ
substituting 3 in one equation of the set (2) and solve will λ S
λ F
gives steady state probabilities.
The frequency of encountering each state equals the Scheduled outage Force outage
probability of been in the state times rate of departure from Fig. 2. Three-state model
that state or rate of entry to that state. According to the
definition We can classify events of Hydro-unit and it’s down states
fi = Pi*Rate of departure of state Pi (4) into:
1. Reserve, Preventive maintenance, and
overhaul.
IV. METHODOLOGY 2. Generator.
Since the power plant was overhauled in 1999 by 3. Turbine (inlet gate, penstock, spiral case, butter
TECHNOPROMEXPOET, this paper consider the operational fly valve, turbine bearing, and runner).
data of the period 2001-2005. The data of each year and for 4. Excitation system (thyristor, cooling system,
each unit is time scheduled. After tabulating all the data, we equipped transformer, and etc…).
classified for each unit the different types of failures, taking 5. Governor system (servo motors, wicket gates,
into account the various sub-units and systems. According to speed governor, and etc…).
that classification we defined Markov states. Failure rate and 6. Main Unit Transformer.
repair rate of all states are found from the classified data. The 7. Main Unit Circuit Breaker.
determination of availability and reliability from their 8. External Effects.
definition is completed. We introduce an over all model for More developed model is driven as follows:
the power plant considering all the five units together.

μ 1
0
Up-state
V. MODELING μ 2 μ μ 8
3

UpĹ
A. UNIT MODELING DownĻ
To model a hydro-unit generally according to its mode of
operation. It can be divided into up-state and down-state. The λ 1 λ 2 λ 3
λ 8
state-space diagram is as follows: 1
Scheduled
2
Generator 3
Turbine
8
External Effect

Scheduled outage Force outage


ĸ ĺ

Fig. 3. Developed hydro-unit model


3

The state transition matrix of Fig. 3 is as follows: The state space diagram of the five units is as follows:

ª− (λ1 + λ2 +""+ λ8) λ1 λ2 λ3 λ4 λ5 λ6 λ7 λ8 º


« »
« μ1 − μ1 0 0 0 0 0 0 0 »
«
« μ2 0 −μ
2
0 0 0 0 0 0 »
»
«
«
μ 3
0 0 −μ
3
0 0 0 0 0 »
»
« μ4 0 0 0 −μ
4
0 0 0 0 »
«
« μ5 0 0 0 0 −μ
5
0 0 0 »»
«
«
μ6 0 0 0 0 0 −μ
6
0 0 »
»
« μ7 0 0 0 0 0 0 −μ
7
0 »
« »
¬« μ8 0 0 0 0 0 0 0 −μ »

The state probabilities are as follows:

TABLE 1
STATE PROBABILITY VALUE

State
State Probability
Number
0 μμμμμμμμ
1 2 3 4 5 6 7 8 D d D 0

1 λμ μμ μμ μ μ
1 2 3 4 5 6 7 8 D d D 1

2 μλ μμ μμ μ μ
1 2 3 4 5 6 7 8 D d D 2
Fig. 4. State –space modeling of five units

3 μμλμμμμμ
1 2 3 4 5 6 7 8 D d D 3

4 μμ μλ μμ μ μ D
1 2 3 4 5 6 7 8 d D 4 The transition rate matrix of Fig. 4 is determined by the same
5 μμμμλμμμ
1 2 3 4 5 6 7 8 D d 5 D way as the unit transition rate matrix. The state probabilities
are determined by the same way as for unit modeling. The
6 μμ μμ μλ μμ D d 6 D
1 2 3 4 5 6 7 8
probability of state 1 is the probability that the five unites are
7 μμ μ μ μ μλ μ
1 2 3 4 5 6 7 8 D d 7 D up
8 μμ μμ μμμλ
1 2 3 4 5 6 7 8 D d 8 D
μ μ μ μ μ ∏ (μ + λ
5
D=d0+d1+d2+d3+d4+d5+d6+d7+d8 p = i =1 i
) (5)
1 1 2 3 4 5 i

The frequencies of encountering states are as follows: Probability of state 32 is the is the probability that all the units
are down
TABLE 2
FREQUENCY OF ENCOUNTERING STATES
λλ λ λ λ ∏ (μ + λ i)
5
p = 1 2 3 4 5 i =1
(6)
32 i
State
Rate of departure Frequency of state
Number
0 λ +λ
1 2
+ " + λ8 (λ + λ + " + λ ) d D
2 8 0
1
The frequency of encountering state 1 is
1 μ 1
μ d1 D 1

2 μ 2
μ d D 2 2
f = (λ 1 + λ 2 + λ 3 + λ 4 + λ 5 ) p (7)
1 1
3 μ 3
μ3d D 3

4 μ 4 μ d4 D 4
The frequency of encountering state 32 is
5 μ μ d D
= (μ + μ μ μ μ)p
5 5 5

f + + + (8)
6 μ 6
μ d6 D 6
32 1 2 3 4 5 1

7 μ 7
μ d7 D 7
By the same way the other state probabilities and frequencies
8 μ 8
μ d D
8 8
are found.

VI. RESULTS
B. Plant Modeling
A. Unit Modeling
To Model DHPS the five units should be studied together. The
number of failure rates and repair rates of a unit for one year In this section we will apply the general unit model derived in
and for all the units are taken to determine the plant section 5.A. According to the operational behavior of the unit
availability and reliability. through a year its model is derived. For example we will take
the Unit-1 for 2001 as a case study.
4

Case Study
The units Availability and Reliability curves for 2001-2005
Unit-1 transit from up-state to stand-by state 149 times with
Unit-1 Unit-2
average remaining time in that state 60 hr. the total time spent
1
in that state was 6566 hr. The unit transit from up-state to 1

0.8 0.8
state-3 (Turbine over speed 1st stage %115) 3 times for 52 hr

Percent Value
percent Value
0.6 Availability 0.6 Availability
with average 1 hr repair and to state-4 (loss of excitation 0.4 Reliability 0.4 Reliability

protection) 2 times for 388 hr with average 1 hr repair. 0.2 0.2

0 0
The one hour repair is considered from the stoppage of the 2001 2002 2003 2004 2005 2001 2004 2005

unit and inspection of the system by the concerned engineer Year Year

then reporting to the operation section. The failure and repair


rates for these states and their probabilities are shown bellow: Unit-3 Unit-4

1 1
TABLE 3 0.8 0.8

Percent Value

Percent Value
THE FAILURE AND REPAIR RATES OF UINT-1 FOR 2001 0.6 Availability 0.6 Availability
0.4 Reliability 0.4 Reliability

State Failure rate Repair rate State Frequency of 0.2 0.2

Number Failure/year Repair/year probabilities state 0 0


2001 2002 2003 2004 2005 2001 2002 2003 2004 2005
0 --- --- 0.413668877 63.70500706 Year Year

1 149 110 0.560333297 83.48966126


3 3 52 0.023865512 0.071596536 Unit-5
4 2 388 0.002132314 0.004264628 1

The Availability and Reliability of Uint-1 Percent Value 0.8

0.6 Availability
p 0.4 Reliability

A= p +p+p 0
= 0.940869 0.2

0
0 3 4
2001 2002 2003 2004 2005
According to the definition of reliability the reliability is Year

considered as the probability of the unit with out failure.


States 0 and 1 are the two states that are with out failure thus: Fig. 6. Annual unit Availability & Reliability
R = p + p = 0.97400217
0 1

In the same way for the other units availability and reliability B. Plant Modeling
are evaluated as shown in Fig. 5
In this section we will apply the general system model derived
2001 2002
in section 5.B. According to the behavior of the system along
1
the year its model is derived. For example we will take the
1
0.9
0.8
0.9
0.8 year 2001 as a case study.
Percent Value

Percent Value

0.7 0.7
0.6 Availability 0.6 Availability
0.5 0.5
0.4 Reliability 0.4 Reliability
0.3
0.2
0.3
0.2
Case Study
0.1 0.1
0 0
1 2 3 4 5 1 3 4 5
Unit Numbe r Unit Number The 5 unit’s behavior for 2001 is as follows:

2003 2004
TABLE 4
UNIT’S FAILURE AND REPAIR RATES FOR 2001
1 1
0.9
0.8 0.8
Unit Failure Rate Repair Rate
Percent Value

0.7
Percent Value

0.6 0.6
Availability Availability
0.5
0.4 Reliability 0.4 Reliabiity
Number Failure/year Repair/year
0.3
0.2
0.1
0.2
Unit-1 154 550
0 0
1 2 3 4 5
1 3 4 5 Unit-2 85 54
Unit Number Unit Number

Unit-3 131 66
2005
Unit-4 174 137
1
0.9
0.8
Unit-5 130 96
Percent Vlaue

0.7
0.6 Availability
0.5
0.4 Reliabiity
0.3
0.2
Availability of DHPS is given bellow, since all the five units
0.1
0 are operated in parallel. 2-out-5, 3-out-5, and 4-out-5
1 2 3 4 5
Unit Number availability are evaluated also for determination of reliability.
The average generation for 2001-2005 was adequate and
Fig. 5. Unit Availability & Reliability histogram for 2001-2005 sufficient of demand, such that 2 or more units are operating.
5

According to the definition of reliability the reliability of TABLE 6


SYSTEM AVAILABILITY, RELIABILITY AND ENS OF 2001-2005
system will be probability 2-out-5.
ENG
TABLE 5 Year Availability Reliability
MWh
DHPS STATE PROBABILITY AND AVAILABILITY DETERMINATION FOR 2001
2001 0.9713727 0.7928423 361918
State State Rate of Frequency of 2002 0.9675902 0.7863049 299375
Number Probability Departure State
2003 0.9929837 0.9202063 158407
1 0.019026972 674 12.8241789
2 0.005327552 1070 5.700480709 2004 0.9968919 0.9629095 234419
3 0.029949863 643 19.25776178 2005 0.9746583 0.8337461 4230928
4 0.037765656 609 22.99928443
5 0.024165643 637 15.39351449 The Availability and reliability of 2001-2005
6 0.025765691 640 16.49004211
7 0.008385962 1039 8.713014085
8 0.010574384 1005 10.62725556
9 0.00676638 1033 6.989670535 Availability and Reliability Curves
10 0.007214393 1036 7.474111584
11 0.05944594 578 34.3597532
1
12 0.038038512 606 23.0513382
13 0.040557106 609 24.69927748 0.9

P ercen t V alu e
14 0.04796514 572 27.43605984
15 0.051140992 575 29.40607059 0.8 Availability
16 0.032724308 603 19.73275773
17 0.016644863 974 16.2120967 0.7 Reliability
18 0.010650783 1002 10.67208489
0.6
19 0.01135599 1005 11.41276959
20 0.013430239 968 13.00047143 0.5
21 0.014319478 971 13.90421299
22 0.009162806 999 9.153643437
2001 2002 2003 2004 2005
23 0.075500683 541 40.84586932 Year
24 0.08049971 544 43.79184232
25 0.051510485 572 29.46399732
26 0.064952793 538 34.94460273 Fig. 7. Plant Availability & Reliability of 2001-2005
27 0.021140191 937 19.8083591
28 0.022539919 940 21.18752371
29 0.014422936 968 13.96140181 Energy Not Generated
30 0.018186782 934 16.98645447
31 0.102240508 507 51.83593744
32 0.028627342 903 25.85048999 450000
ENG in Mega watt hour

Probability Probability Probability 400000


Availability
2 out of 5 3 out of 5 4 out of 5 350000
300000
0.971372658 0.792842322 0.444814492 0.116235685
250000
200000
The intended condition is the probability of 2 out of 5 because 150000
under this condition the generation is sufficient and the power 100000
plant is reliable. That is why the reliability equals: 50000
0
R=0.792842322 2001 2002 2003 2004 2005
The energy not generated is another important index of Yaer
reliability measure for generation power plants. ENG is one of
the adequacy indices that show the plant energy not generated Fig. 8. Plant ENG 2001-2005
because the power plant units were failed. The ENG is can be
calculated:
The mean values of failure and repair rates are determined.
Unavailability=1-A The average failure Ȝ & repair μ rates for the five years and
for each unit are calculated as shown in Table 7. These values
ENG=U*annual average energy (in mega watt hour) help to find the system reliability indices.
6

TABLE 7 • Unit-4 at 2004 its left servomotor was defected two times
MEAN FAILURE AND REPAIR RATES
and at 2004 its thrust bearing segments were damaged and
Unit Mean Failure Mean repair rate replaced.
Number rate failure/year repair/year • Unit-5 reliability level was approximately constant through
Unit-1 154 307 out that period.
Unit-2 233 291 The availability of units was also affected by the mentioned
reasons.
Unit-3 212 349
(4) In Fig. 7 we note that at 2001 and 2002 there are poor
Unit-4 179 290 point reliability. The specified definition of reliability
Unit-5 179 299 used for plant reliability determination is the most
important factor, because at those two years units were
The predict values of system reliability indices down for long period due to reasons mentioned above.
System failure rate=192 failures/year (5) We realize from our study of the plant availability and
System repair rate=307 repairs/year reliability that the maintenance program and skill of
System Availability= 0.991977259 engineers and technicians play an important role for
System Reliability= 0.926357958 improving the performance of the units and increasing the
availability and reliability of the units and the power
plant.
VII. DISCUSSION AND CONCLUSION
VIII. REFERENCES
(1) We did this study because of importance of DHPS and the
necessity of the region to that power plant. 60% of
[1] J. Endrenyi, Reliability Modeling in Electric Power systems, A Wiley&
regional demand is covered by this generation plant. The Sons, 1980.
continuity of operation of units for mentioned reason is [2] Ioannis A. Ppazoglou, “ Elements of Markovian Reliability Analysis”,
the priority. Relaibility Engineering, Proceeding of the ISPRA-Course Held at the
(2) The weak points that cause poor point availability and Escuela Technica Superior de Ingnieers Navales, Madrid, Spain, pp-
171-203, September 1986.
reliability are given in Table. 8: [3] John R. Soeth, and A. D. Patton, “A Comparison of Alternative
generating unit reliability models”, IEEE Trans. Power System, Vol. 4,
TABLE 8 No. 1, pp. 108-114, February 1989.
UNIT MAJOR FAULTS THAT AFFECT THE RELIABILITY INDICES [4] R.N. Allan and J. Roman, “Reliability Assessment of Generation System
Containing Multiple Hydro Power plant Using Simulation Techniques”,
Unit Down time in IEEE Trans. Power System, Vol. 4, No. 3, pp. 1074-1080, August 1989.
Year Cause of trip
Number Hr due to trip [5] Hussein Salehfar,and Siswa Trihadi, “ Animated Power-Generation
Thrust bearing over heating and Ssytem Reliability Evaluation”, Proceeding Annual Reliability and
Unit-2 2001 3552 Maintainability Symposium, pp. 248-254, 1994.
bearing segments were damaged
[6] Department of Army, U.S. Army Crops of Engineer, “Reliability
Defect of left hand serve motor Analysis of Hydro Power Equipment”, Technical Letter No. 1110-2-
Unit-1 2002 7000
and operating ring of turbine 550, 30 May 1997.
Turbine bearing over heating and [7] R. Billinton, Hau Chen, and Jiaqi Zhou, “Individual Generating Station
Unit-4 2004 3346 Reliability Assessment”, IEEE Trans. Power System, Vol. 14, No. 4, pp.
segments were damaged
1238-1244, November 1999.
(3) Unit-2 for years 2002 and 2003 was completely shut
down, because the unit circuit breaker (CB) was damaged
IX. BIOGRAPHIES
and there were no spare circuit breaker to be replaced, due
to sanction on Iraq at that time and the complexity of
Asso R. Majeed was born in Sulaimani, on 1957. He
providing the spares from the manufacturing company. received his M.Sc. & Ph.D. in electrical engineering
From Fig. 6 we conclude that the reliability and availability of from Baghdad University/ Iraq in 1981and 1997
some units of the plant is decreasing a lot in some years. respectively. Recently he is head of electrical
engineering Dept. in Sulaimany University/Iraq. His
• Unit-1 reliability is decreased form 0.97 at 2001 to 0.3 at area is power system reliability and system planning.
2002 due to left hand servomotor defect. Its reliability is (E-mail: drassomajeed2000@yahoo.com)
increasing, but not reaching the level of 2001. At 2003 the
reliability was affected by turbine and generator shaft
alignment, while at 2004 by speed governor defect and at Nawzad M. Sadiq was born in Sulaimani, on June 7,
2005 by run-out of turbine shaft and turbine abnormal 1977. He received his B.Sc. in electrical engineering
vibration. from Salahadden university-Erbil/Iraq in 1999. Now
• Unit-2 thrust bearing segments were damaged at 2001, he is M.Sc. student in electrical engineering Dept.
Sulaimani University.
which caused poor reliability. (E-mail: nawzad_977@yahoo.com)
• Unit-3 at 2001 has the lowest value of reliability comparing
with other years, because of thrust bearing oil-cooler
dismantled.

You might also like