Lecture No.

CHAPTER Economic Dispatch of Thermal Units

THREE

3.1

INTRODUCTION The power system engineer is faced with the challenging task of planning and

successfully operating one of the most complex systems of today's civilization. The efficient planning and optimum economic operation of power system has always occupied an important position in the electric power industry. A typical electric power system comprises of three main elements. Firstly, there are consumers, whose requirements of electrical energy have to be served by electrical power system. Secondly, there has to be means by which these requirements are served i.e., the generating plants. Finally, there is transmission and distribution network, which transports the power from producers to the consumers. The prime objective of a power system is to transfer electrical energy from the generating stations to the consumers with; Maximum safety of personal and equipment. Maximum continuity (security reliability and stability) Maximum quality (frequency and voltage with in limits) Minimum cost (optimum utilization of resources). Electrical energy can not be stored economically. However, it can be stored as potential energy in hydro systems and pumped storage schemes, but this represents a small fraction of the installed capacity of most industrialized nations. The larger portion of generating plant is thermal. For cold turbo alternator set 6 to 8 hours are

37

Power system Operational Planning Economic Dispatch

Lecture No. 3

required for preparation and readiness for synchronization. A hot unit can be synchronized within 15 minutes and fully loaded in 30-60 minutes. It is therefore essential that the demand must be met as and when it occurs. The present day load demand handling is a challenging task for Power System Operation Engineer. In this context system operation necessitates the following for power system: it must be highly interconnected it must be automated it must have Operational Planning In the early days the power system consisted of isolated stations and their individual loads. But at present the power systems are highly interconnected in which several generating stations run in parallel and feed a high voltage network which then supplies a set of consuming centers. Operating an automated electric power system is an extremely complex task. The objective of power system control is to provide a secure supply at a minimum cost. Figure 3.1 illustrates the operation and data flow in a modern power system on the assumption of a fully automated power system based on real-time digital control. Although such an extreme degree of automation has not yet been implemented, the activities in the boxes are performed by most utilities. In some cases computation is performed off-line, in others on-line, the degree of human supervision or intervention, varying considerably from utility to utility. There are three stages in system control, namely generator scheduling or unit commitment, security analysis and economic dispatch. Generator scheduling involves the hour-by- hour ordering of generator units on off the system to match the anticipated load and to allow a safety margin. With a given power system topology and number of generators on the bars, security analysis assesses the system response to a set of contingencies and provides a set of constraints that should not be violated

38

Power system Operational Planning Economic Dispatch

Lecture No. 3

if the system is to remain in secure state. Economic dispatch orders the minute-to-minute loading of the connected generating plant so that the cost of generation is a minimum with due respect to the satisfaction of the security and other engineering constraints. These three control functions require reliable knowledge of the system configuration, i.e., circuit breaker and isolator position and system actual P and Q flows. This data is collected from thousands of metering devices and transmitted to control centers, usually over hired telephone lines. It is statistically inevitable that owing to the numbers of devices involved, interference over communication lines unreliable data will be present. State estimation is a mathematical algorithm that provides reliable database out of an unreliable set of information. In traditional power control centers where all activities are channeled through human operator, the experienced control engineer looking at wall mimic diagram of power system takes in multiplicity of data. He mentally assesses their compatibility with a degree of confidence and he can pinpoint grossly corrupted piece of information. Human beings are good state estimators.

39

Power system Operational Planning Economic Dispatch

Lecture No. 3

Figure 3.1 Power System Control Activities

3.2

OPERATION PLANNING OF POWER SYSTEM The operational planning of the power system involves the best utilization of

available energy resources subjected to various constraints to transfer electrical energy from generating stations to the consumers with maximum safety of personal/equipment without interruption of supply at minimum cost. In modern complex and highly interconnected power systems, the operational planning involves the steps

40

Power system Operational Planning Economic Dispatch

Lecture No. 3

o Load Forecasting, o Unit Commitment, o Economic Dispatch o Hydrothermal Coordination o Voltage Control o Frequency Control

TIMES SCALES INVOLVED FOR ACTIVITIES IN PLANNING OPERATION AND CONTROL Various activities that are combined under the broader area of power system operation and control do not have the same time scale. For example at one extreme is the time taken to build a hydropower station (up to 10 years), while on the other extreme is the time interval between detection and interruption of a fault on power system (up to 80 m sec). Time scales involved for various activities in power system planning, operation and control are summarized as follows: YEARS System expansion planning, construction, maintenance scheduling and planned outages. MONTHS DAYS HOURS MINUTES Preliminary load forecasting, generation estimation and contingency planning Short term load forecasting, reserve assessment and generation scheduling. Unit commitment, preliminary economic dispatch and contingency analysis. Economic dispatch, power interchanges, frequency control and security assessment. SECONDS

41

Power system Operational Planning Economic Dispatch

Lecture No. 3

Protection and C.B operation, automatic voltage and frequency control. Time scale particular to system operation point of view may be stated as: Unit Commitment hours to days to week Economic Dispatch minutes to hours Security Analysis every few minutes and on demand System equipment milliseconds to seconds

(Automatic voltage control i.e., tap changers and excitation control and generator set governor control)

SPINNING RESERVE The total generating capacity required to be available on the bars is always larger than the anticipated load. The difference between these quantities is called system spare or spinning reserve. The spinning reserve (s. r.) for a given generating unit can be defined as the extra amount of the active power that can be obtained from that unit with in a specified interval of time (a few minutes) by loading it at its maximum rate through governor action. The spinning reserve of the power system will be available to makeup the outage of any generating unit or to meet unexpected increase in demand. The spinning reserve should be at least equal to the rating of the largest unit on the bars. The system characteristics will determine the percentage interval after which the spinning reserve must be available if excessive drop in frequency is to be averted. The specified post outage time and maximum loading rate of a generating unit will fix its ceiling of spinning reserve. The maximum loading rate of a turbo alternator unit is determined empirically and is dictated thermal considerations. Typical maximum loading rate of Turbo alternator Gas turbine : : 2-5 MW /min (Ramping) 30 MW /min

First response reserve capacity can be provided by hydro or pump storage and by gas turbine. Such plant can be started up automatically when frequency falls below critical value.

42

Power system Operational Planning Economic Dispatch

Lecture No. 3

The decision of how much reserve capacity the system should carry depends on diverse factors such as type of generating plant, unit sizes, and degree of desirable reliability and security. It usually amounts to less than 10% of the load. Typical spinning reserve on CEGB Day time s. r of 1000 MW Drops at. Night to 680 MW Of 1000 MW, 640 MW is provided by partly loaded sets on the bars capable of supplying the demands within 5 minutes and sustaining this output. The remaining 360 MW is available from the F testing pump storage scheme there is also a standing reserve of 500 MW provided by gas turbines not in synchronization with the system but able to supply the demand within 5 minutes. Finally there is a standby reserve of 800 MWof gas turbine plant not included in standby reserve capacity capable of achieving the demand within hours. 3.3 ECONOMIC DISPATCH Having solved the unit commitment problem and having ensured through security analysis that present system is in a secure state then the efforts are made to adjust the loading on the individual generators to achieve minimum production cost on minute-to-minute basis. This loading of generators subjected to minimum cost is in essence the economic dispatch problem and can be defined as a computational process of allocating generation levels to the generating units in the mix so that the system load may be supplied entirely and most economically. Load dispatching is essentially an online activity and is normally associated with an online forecasting / prediction system. The economic dispatch calculations are performed every few minutes, which must ensure that all the committed units, sharing in the economic dispatch calculations, are operating in such a way that the overall system operation cost is minimum and the recognized system constraints are satisfied.

43

Power system Operational Planning Economic Dispatch

Lecture No. 3

3.3.1

POWER SYSTEM VARIABLES FOR ECONOMIC DISPATCH On a bus bar there are four variables namely P, Q, V and . Out of the four

two are specified and two can be determined from load flow analysis. In economic dispatch problem, further degrees of freedom are necessary, in other words some elbow room is available within which variables can be adjusted for the minimization of the operating cost. For the economic dispatch analysis, we can classify the variables in the power system into three categories: 1. Control or decision variables U These are the variables over which we have complete control, within specified limits. Typical control variables are the active power injection and voltage magnitude at generation buses. 2. State or dependent variables X We have no direct control over these, as their value is not known until the completion of load flow study. Typical dependent variables may be voltage magnitude and angles at load buses. 3. Output variables Z These are function of other variables at least one of which is state variable. The output variables are determined after the completion of load flow study. Typical output variables are the P and Q injected at slack bus, the complex power flows over the transmission lines, the cost of power generation, the network losses etc. 3.3.2 VARIOUS CONSTRAINTS IN POWER SYSTEM The present modern power system has to operate under various operational and network constraints. Broadly speaking there are two types of constraints: Equality Constraints Inequality Constraints

EQUALITY CONSTRAINTS
44 Power system Operational Planning Economic Dispatch

Lecture No. 3

Equality or network constraints are basic load flow equations under steady state condition. There is a balance between active and reactive power, therefore
n

P
i =1 n i =1

g,i

(Pd + Pl ) = 0 (Q d + Q l ) = 0

(3.1)

g,i

Where Pl and Ql are network losses. In terms of and state variables mathematically can be written in short hand form g(X,U)=0 INEQUALITY CONSTRAINTS Inequality constraints are of two types Hard type Soft type Hard type are those which are definite and specific like the tapping range of on-load tap changing transformer, where as soft type are those which have some flexibility associated with them like the nodal voltages and phase angles between nodal voltages. Generator Constraints The KVA loading on a generator is given by Pi2 + Qi2 Ci2 The maximum active power generation of a source is limited by thermal consideration and minimum power generation is limited by flame instability of the boiler. If the power output of a generator for optimum operation of the system is less than a prespecified value Pmin, the unit is not put on the bar because it is not possible to generate that low value of the power from that unit. Hence generator power Pi can not be
45 Power system Operational Planning Economic Dispatch

(3.2)

P 2 + Q 2 and this should not exceed a

pre-specified value Cp because of temperature rise conditions i.e.,

Lecture No. 3

outside the range stated by the inequality Pg, min. Pg, i Pg, max. Voltage Constraints Voltage magnitude and phase angle at various nodes should vary within certain limits i.e., Vi, min. Vi Vi, max. i, min. i i, max Running Spare Capacity Constraints These constraints are required to meet Forced outages of one or more generators on the system Un expected load on the system The total generation should be such that in addition to meeting the load and losses, a minimum spare capacity should be available i.e., G Pi + Pso Where G is total generation and Pso is some prespecified power. A well planned system is one in which spare capacity Pso is minimum. Transformer Tap Settings In case of auto transformer we have 0 t < 1.0 For two winding transformer, if tapings are provided on secondary side, then we have 0tn Where n is transformation ratio Phase shift limits of phase shifting transformer i,min. i i, max. Transmission Line Constraints
46 Power system Operational Planning Economic Dispatch

Lecture No. 3

P and Q flows through the line are limited by the thermal capability of the circuit and is expressed as: C i Ci, max. Where Ci, max. is the maximum loading capacity of the ith line. The mathematical short hand notation that encompasses all the inequality constraints can be written as:
h (U, X ) 0

(3.3)

3.3.3

Economic Dispatch Mathematical Formulation In general the economic dispatch problem can be formulated in mathematical

terms as: Minimize a scalar objective function f(X,U) Subject to Where X = state (dependent) variables U = control (independent) variables gi( X , U ) = power flow equations ( non- linear ) hk( X , U ) = inequalities ( limits ) VARIABLES The system state or dependent variables X are generally gi ( X , U ) = 0 hk ( X , U ) 0

V, nodal voltage on PQ nodes X = , voltage angle on PQ and PV nodes

The control or independent variables U are generally

V, on slack nodes P,Q on PQ node U= P, V on PV nodes t transformer taps

47 Power system Operational Planning Economic Dispatch

Lecture No. 3

Equality Constraints

The equality constraints gi ( X , U ) = 0 are the nodal load flow equations i.e., YV = I Where I = S*/V* For analysis and computation purposes it is more convenient to express the problem by a single relation and to separate the complex quantities into their coordinate components. Usually the problem can be decomposed into rectangular (real and imaginary) or polar (magnitude and angle) co-ordinates. Simple arithmetic operations are faster in rectangular system than in polar form which involves trigonometrically expressions. The polar form however uses V and explicitly which offers opportunity to exploit the natural physical operation of the system into P- and Q-V dependency. The power flow equations in polar form are usually expressed in terms of power mismatch at each node i.e.,
Pi = Pgi Pdi Vi Vk (G ik Cos ik + B ik Sin ik ) = 0
k

Q i = Q gi Q di Vi Vk (G ik Sin ik B ik Cos ik ) = 0
k

Where ik = i k g = generation d = demand G = Network Conductance B = Network Susceptance K = node number directly connected to node i usually G << B, ik << 20 and Vi = 1 10% Load flow problem is non-linear and sparse and can be solved by Newton or approximate Newton methods. In this method, Jacobin is formed in each iteration and

48

Power system Operational Planning Economic Dispatch

Lecture No. 3

equations solved to correct the unknown V and . Mathematically we can write as:

P = H Q J
3.3.4

N L V

Solution Methods

Economic dispatch is a constrained optimization problem. The solution of optimization problems is province of a branch of mathematics called mathematical programming. The label is some what confusing as the work programming has nothing to do with computer programming , although of course , computer are invariably used for the solution of such problems. A wide variety of optimization techniques has been applied for solving Economic Dispatch problems. The techniques can be classified as
The approaches used for ED problem may be listed as: Merit Order Approach (old method) Equal Incremental Cost Criterion (widely used) Linear Programming (Easy Constraint Handling) Dynamic Programming Non-Linear Programming (Exact Methods) a) Ist Order Gradient Based Techniques (Dommel & Tinney Method) b) Second Order Method (Optimal Power Flow )
1 2 Non-linear Programming (NLP) Quadratic Programming (QP) Objective function and constraints are non-linear. Special form of non-linear programming Objective function-quadratic Constraints are linear 3 NewtonBased Solution Necessary conditions of optimality commonly referred to as Khun-Tucker conditions are obtained In general these are non-linear equations requiring iterative methods of solution. The

49

Power system Operational Planning Economic Dispatch

Lecture No. 3

Newton method is favoured for its quadratic characteristics. 4 Linear Programming Linear programming treats problems with constraints and objective function formulated in linear forms with non-negative variables. Simplex and Revise Simplex methods are known to be quite effective for solving LP problems.

3.5 3.5.1

ECONOMIC DISPATCH OF THERMAL UNITS-------METHOD OF SOLUTIONS Merit Order Approach

In this method it is assumed that incremental cost of all generators is constant over the full range or over successive discrete portions within the range. The economic way of loading the machine is in the order of highest incremental efficiency. This method therefore needs forming of a table which could be looked into for any load condition and does not need any complicated calculations. Assuming linear cost function, the optimal economic dispatch can be described as: Minimize Subject to And
N

FT = Fi (Pi )
i =1

= PD Pi
i =1

Pim Pi 0 Pi PiM 0

The Augmented cost function is given by

Facost = Fi + (PD Pi ) + i (Pi Pi ) + i (Pi Pi )

m m M M i =1 i =1 i =1 i =1

1. If for some i the inequality constraints are inactive then

F =
50 Power system Operational Planning Economic Dispatch

Lecture No. 3

2. If for some i then 3. If some i then

Pi = PiM
F

Pi = Pim
F

Thus it can be concluded that if ith generator is between its limits, the remaining generators are either at its upper or lower limit depending on whether their incremental cost F is larger or smaller than F. This analysis leads to the following strategy for optimum dispatch. A table is made up of the available generating units with their corresponding values of the incremental cost in ascending order. For an example Unit 2 5 7 13 10 1 dF/dt $/Mwh dF2/dt dF5/dt dF7/dt dF13/dt dF10/dt dF1/dt

For a given load if unit 13 is operating between its limits, units 2,5 and 7 are at their upper limit and units 10,1 at their lower limit. In practice it is arranged that not all units below 13 are at their lower limit, but perhaps only 10 and 1, so that spinning reserve is ensured. The above table is known as the merit order list of the units, the method is known as Order of Merit method.
3.5.2 Equal Incremental Cost Criterion (Neglecting Transmission Loss)

Let there be a system of thermal generating units connected to a single bus-bar serving a received electrical load Pload. It is required to run the machines so that cost of

51

Power system Operational Planning Economic Dispatch

Lecture No. 3

generation is minimum subject to total generation are equal to total demand when transmission losses have been neglected. This problem may be solved by using Equal Incremental Cost Criterion. Working philosophy of their criterion is as: When the incremental costs of all the machines are equal, and then cost of generation would be minimum subject to equality constraints. The economic dispatch problem mathematically may be defined as: Minimize Subject to Where FT = F1 + F2 + FN is the total fuel input to the system Fi = Fuel input to ith unit Pi = The generation of ith unit This is the constrained optimization problem with equality constraints only. Three methods can be applied for locating the optimum of the objective function. These methods are: Direct substitution Solution by constraint variation The method of Lagrange multipliers However, in this analysis method of Lagrange multiplier will be used. The working philosophy of this method is that constrained problem can be converted into an unconstrained problem by forming the Lagrange, or augmented function. Optimum is obtained by using necessary conditions.
Case-1 No Generation Limits i.e. Pi (min.) Pi Pi (max.) Neglected.
N

FT = Fi (Pi )
i =1 N

(3.5)
n

PD = Pi or = Pload Pi = 0
i =1 i =1

(3.6)

The augmented unconstrained cost function is given by

52

Power system Operational Planning Economic Dispatch

Lecture No. 3

L = FT + .
N = FT + . PD Pi i =1

(3.7)

The necessary conditions for constrained local minima of L are the following:

L =0 Pi
L =0
C-I

(3.8)

(3.9)

First condition gives

L FT + .(0 1) = 0 = Pi Pi
or

FT F =0 T = Pi Pi

Q FT = F1 + F2 + FN
then

FT dFT = = Pi dPi
and therefore the condition for optimum dispatch is

dFT = dp i
or
b i + 2c i Pi =

(3.10)

(3.10)

where
FT = a i + b i Pi + c i Pi2

53

Power system Operational Planning Economic Dispatch

Lecture No. 3

C-II

Second condition results in

N L = PD Pi = 0 i =1

or

P
i =1

= PD

(3.12)

In summary,

When losses are neglected with no generator limits, for most economical operation, all plants must operate at equal incremental production cost while satisfying the equality constraint given by equation (3.12 ).

Pi =

. b i 2c i

(3.13)

The relations given by equation (3.13) are known as the co-ordination equations. They are function of . An analytical solution for is given by substituting the value of Pi in equation (3.12), i.e.,
N

i =1 N

. b i = PD 2c i
N b 1 i = PD 2c i i =1 2c i
N

(3.14)

.
i =1

PD +
i =1 N

bi 2c i

1 2c i =1 i

(3.15)

Optimal schedule of generation is obtained by substituting the value of from eq. (3.15) into eq. (3.13).

54

Power system Operational Planning Economic Dispatch

Lecture No. 3

Solution of Equation (3.13) Iteratively

In an iterative search technique, starting with two values of , a better value of is obtained by extrapolation, and process is continued until Pi is within a specified accuracy. However rapid solution is obtained by the use of gradient method. Equation (3.14) can be written as
f (.) = PD

Expanding left hand side of the above equation in Taylors series about an operating point (k) and neglecting higher order terms results in
f (.)
(k )

df (.) (k ) + . = PD d
k

.(k ) =

PD f (.) df (.) d P (k ) df (.) d

(k )

(k )

.(k ) =

(k )

P (k ) 1 2c i

(3.16)

and therefore
.(k +1) = .(k ) + .(k )

(3.17)

where

P (k ) = PD Pi(k )
i =1
P

(3.18)

The process is continued until P(k) is less than a pre-specified accuracy.

Case II Generation limits i.e., Pi(min) Pi Pi(max.) Included

The power output of any generator should not exceed its rating nor should it be below that necessary for stable boiler operation. Thus generators are restricted to be within given maximum and minimum limits. The problem is to find the real power generation for each unit such that objective function (i.e., total production cost) is minimum, subject to equality constraints and the inequality constraints i.e.,
55 Power system Operational Planning Economic Dispatch

Lecture No. 3

Minimize: Subject to:

FT = Fi (Pi )
i =1

= PD Pi
i =1

And:

Pi Pi (max.) 0 Pi (min.) Pi 0

for i =1,2,,N

This is constrained optimization problem both with equality and inequality constraints. The optimization problem with equality and inequality constraints is handled well by Lagrange multiplier method. However, Khun-Tucker conditions complement the Lagrange conditions to include inequality constraints as additional terms. For establishing optimal solution for this problem, we consider the case of two machines and would be generalized for N machines. The problem may be stated as: Minimize: Subject to:
FT = F1 (P1 ) + F2 (P2 ) (P1 , P2 ) = PD P1 P2

g P = P P+ 0 1 1 1 1 P P1 P g 2 P1 = P1 P1 0
1 + 1

g P = P P+ 0 P P2 P 3 2 1 1 g 4 P2 = P2 P2 0
2 + 2

Lagrange function becomes

L(P, , ) = FT + .(P1 , P2 ) + 1g 1 (P1 ) + 2 g 2 (P1 ) + 3 g 3 (P2 ) + 4 g 4 (P2 ) = F1 (P1 ) + F2 (P2 ) + .(PD P1 P2 ) + .1 P1 P1+ + .2 P1 P1 + .3 P2 P2+ + .4 P2 P2
Condition 1

The necessary conditions for an optimum for the point Po, o, o are:

56

Power system Operational Planning Economic Dispatch

Lecture No. 3

L (P, , ) = 0 Pi
F1(P1 ) .1 .2 = 0 F2 (P2 ) .3 .4 = 0

Condition 2
i (P ) = 0 PD P1 P2 = 0

Condition 3
g i (P) 0 P1 P1+ 0 P1 P1 0 P2 P2+ 0 P2 P2 0

Condition 4
.g i (P) = 0
.i 0

( . (P . (P . (P
2

.1 P1 P1+ = 0
1
1

2
2

) . 0 P ) = 0 . 0 P ) = 0 . 0 P ) = 0 . 0
1 2

+ 2

Case-I

If optimum solution occurs at values for P1 and P2 that are not at either upper limit or lower limit then all values are equal to zero and
F1(P1 ) = F2 (P2 ) =

57

Power system Operational Planning Economic Dispatch

Lecture No. 3

That is, the incremental cost associated with each variable is equal and this value is exactly the .
Case- II

Now suppose that the optimum solution requires that P1 be at its upper limit

(i.e., P

P1+ = 0 and that P2 is not at its upper or lower limit. Then

.1 0 and .2 = .3 = .4 = 0

From condition-1
F1(P1 ) = .1 F1(P1 ) F2 (P2 ) =

Therefore, the incremental cost associated with the variable that is at is upper limit will always be less than or equal to , whereas the variable that is not at limit will exactly equal to .
Case- III

Now suppose that optimum solution requires that P1 be at its lower limit (i.e., P1 P1 = 0 ) and that P2 is not at its upper and lower limit. Then
.2 0 and 1 = .3 = .4 = 0

From Condition I, we get:

F1(P1 ) = + .2 F1(P1 ) 0 F2 (P2 ) =

Therefore, the cost associated with a variable at its lower limit will be greater than or equal to , whereas, again incremental cost associated with variable that is not at limit will exactly equal to .
Case- IV

If the optimum solution requires that both P1, P2 are at limit and equality constraint can be met, then and non-zero values are indeterminate. Let
58 Power system Operational Planning Economic Dispatch

Lecture No. 3

P1 P1+ = 0 P2 P2+ = 0

Then

.1 0

.3 0

.2 = .4 = 0

Condition-1 would give

F1(P1 ) = .1 F2 (P2 ) = .2

Specific values for , 1 and 2 would be undetermined. For the general problem of N variables Minimize: Subject to: And:
FT (Pi ) = F1 (Pi ) + F2 (P2 ) + ... + FN (PN ) L P1 P2 ... PN = 0

P P+ 0 i i for 1 = 1,2,..., N Pi Pi 0

Let the optimum lie at Pi = Pi0, i= 1,2,,N and assume that at least one xi is not at limit. Then, if Pi < Pi+ and Pi > Pi then Fi(Pi ) = if Pi = Pi+ then Fi(Pi ) if Pi = Pi then Fi(Pi )
3.5.3 ITERATION METHOD

The incremental production cost of a given plant over a limited range is given by

dFi = a i + b i Pi dPi

(1)

Where bi = slope of incremental production cost curve

59

Power system Operational Planning Economic Dispatch

Lecture No. 3

ai = y-intercept of incremental production cost curve The necessary condition for optimum schedule is as:

dFi = dPi

(2)

Subject to equality and inequality constraints i.e.,

= PD Pi
i =1

(3)

Pi Pi (max.) 0 fori = 1,2,3,..., N Pi (min.) Pi 0

For optimal schedule equations (2) and (3) can be solved simultaneously. As inequality constraints have also to be taken into account, the following iterative method known as iteration method may be used for solution.

1. Assume suitable value o This value should be more than the largest intercept of the incremental cost characteristic of various generators.

60

Power system Operational Planning Economic Dispatch

Lecture No. 3

2. Compute the individual generations P1, P2,,PN corresponding to the incremental cost of production from equation 2. In case generation at any bus is violated during that iteration and the remaining load is distributed among the remaining generators. 3. Check if the equality
N

P
i =1

= PD is satisfied.

4. If not, make a second guess and repeat the above steps.

3.6 PROBLEM

PROBLEM #1
This problem has been taken as standard test system for testing the algorithms in IEEE Transaction papers.

The fuel costs functions for three thermal plants in $/h are given by F1 (P1) = 500 + 5.3 P1 + 0.004 P12 F2 (P2) = 400 + 5.5 P2 + 0.006 P22 F3 (P3) = 200 + 5.8 P3 + 0.009 P32 If the total load to be supplied PD = 800 MW, then find the optimal dispatch and total cost by the following methods: 1. Analytical method 2. Graphical demonstration 3. Iterative technique using gradient method Given that line losses and generation limit have been neglected. Solution:

61

Power system Operational Planning Economic Dispatch

Lecture No. 3

N b PD + i i=1 2ci = N 1 2c i=1 2 b b b PD + 1 + 2 + 3 5.3 5.5 5.8 2c1 2c2 2c3 800+ 0.008 + 0.012 + 0.018 = = 1 + 1 + 1 1 + 1 + 1 0.008 0.012 0.018 2c1 2c2 2c3

800 + 1443.0555 = 8.5$/MWh 263.8889

The co-ordination equation is given by

P2 =

b2 2c 2

Substituting the values of , we have optimal dispatch,

P10 =

b1 8.5 5.3 = = 400.00 2c1 2(0.004)

0 b 2 8.5 5.6 P2 = = = 250.00 2(0.006) 2c 2

0 b 3 8.5 5.8 P3 = = = 150.00 2c 3 2(0.009)

P10 + P20 + P30 = 400+250+150 = 800 MW = PD

2. Graphical Demonstration

The necessary conditions for optimal dispatch are.

dF1 = 5.3 + 0.008P1 = dP1 dF2 = 5.5 + 0.012P2 = dP2

62

Power system Operational Planning Economic Dispatch

Lecture No. 3

dF3 = 5.8 + 0.018P3 = dP3

Subject to P1 + P2 + P3 = PD To demonstrate the concept of equal incremental cost for optimal dispatch, we plot the equal incremental cost of each plant on the same graph as shown in figure3. To obtain a solution various values of could be tried until one is found which produces Pi = PD for each , if Pi < PD, we increase otherwise if Pi > PD we reduce . Thus the horizontal dashed line move up or down until at optimum point 0, Pi = PD For 0 = 8.5 Satisfying P10 = 400 Pi = PD P20 = 250 P30 = 150

Figure 3.2 Plot of Incremental Costs

63

Power system Operational Planning Economic Dispatch

Lecture No. 3

-ITERATION METHOD

64

Power system Operational Planning Economic Dispatch

Lecture No. 3

3. Iteration using gradient method

For the solution using iterative method, we assume starting value of (1) = 6.0 from the co-ordination equations.

Pi =

bi 2c i

We find Pi and then check Pi = PD if this condition is not satisfied, we

65

Power system Operational Planning Economic Dispatch

Lecture No. 3

update by

Where

(k +1)

(k)
=

+
P

(k)

(k) 1

2c 2

ITERATION 1
(1) = 6.0

P1(1) =
P2(1) = P3(1) =

6.0 5.3 = 87.50 2(0.004)

6.0 5.5 = 41.67 2(0.006) 6.0 5.8 = 11.11 2(0.009)

Pi(1) = 87.50 + 41.67 + 11.11 = 140.28 PD = 800 MW P(1) = 800-140.2777 = 659.72

P

Pi(1) PD we calculate
(1) = P (1) 659.7222 = = 2.5 1 1 1 1 1 1 + + + + 2c1 2c 2 2c3 0008 0.12 0.019 Thus new value of is (2) = (1) +(1) = 6.0+2.5=8.5 ITERATION 2 P1(2) = 8.5 5.3 = 400.00 2(0.004)

66

Power system Operational Planning Economic Dispatch

Lecture No. 3

P2(2) =
P3(2) =

8.5 5.5 = 250.00 2(0.006)

8.5 5.8 = 150.00 2(0.009)

Pi(2) = PD - Pi(2) = 800 (400+250+150) = 800 800 = 0 Pi(2) = 0, the equality constraint is met in two iteration.
Optimal Dispatch

P10 = 400 MW P20 = 250 MW P30 = 150 MW 0 = 8.5 Total fuel cost is given by Ft = F1 (P1) + F2 (P2) + F3 (P3) = {500 + 5.3(400) + 0.004(400)2} + {400+5.5(250) + 0.006(250)} + {200+5.8(150)+0.009 (150)2} Ft = 6, 682.5 $/h.
PROBLEM 2

Let there be three units with following data: Unit 1: Coal fired steam units Input - output curve:

MBtu 2 H1 = 510 + 7.2 P1 + 0.00142 P1 150 P1 600 h

Unit 2: coal fired steam unit Input - output curve:

67

Power system Operational Planning Economic Dispatch

Lecture No. 3

MBtu 2 H2 = 310 + 7.85P2 + 0.00194 P2 100 P1 400 h

Unit 3: coal fired steam unit Input - output curve:

MBtu 2 H3 = 78 + 7.97 P3 + 0.00482 P3 50 P1 200MW h

It is required to determine the economic operating point for these three units when delivering a total of 850 MW with reference to the following cost data:
Case-I

Unit 1: Unit 2: Unit 3:

Case-II

fuel cost = 1.1 $/MBtu fuel cost = 1.0 $/MBtu fuel cost = 1.0 $/Mbtu

The fuel cost of Unit 1 decreased to 0.9 $/MBtu

Solution. Case-I

Fuel costs are given by: F1 (P1) = H1 (P1) x 1.1 = 561+7.92P1 + 0.001562 P12 $/h F2 (P2) = H2 (P2) x 1.0 = 310+7.85P2 + 0.00194 P22 F3 (P3) = H3 (P3) x 1.0 = 78+7.97P3 + 0.00482 P32 $/h $/h

If is the incremental cost, then condition for optimal dispatch are:

dF1 = 7.92 + 0.003124P = dP1

(1)

68

Power system Operational Planning Economic Dispatch

Lecture No. 3

dF2 = 7.85 + 0.00388 = dP2

dF3 = 7.97 + 0.00964 = dP3
and P1+P2+P3 = 850 (4)

P1 =

7.92
0.003124

P2 = P3 =

7.85
0.00388

7.97
0.00964

Putting the values of P1, P2, & P3 in = (4), we have

7.92
0.003124

7.85
0.00388

7.97
0.00964

= 850

= 9.148 $/MWh Now P10 = 393.2 Mw P20 = 334.6 Mw P30 = 122.2 Mw 150 P1 600 100 P2 400 50 P3 250

Pi0 = 393.2 + 334.6+122.2=850 All the generations are within their limits and equality constraint is also satisfied.
Case-II

Cost of Coal for Unit 1 decreases to = 0.9 $/MBtu The fuel cost of coal-fired plats is given by

69

Power system Operational Planning Economic Dispatch

Lecture No. 3

F1 (P) = H1 (P) x 0.9 = ( 510 + 7.2 P1 + 0.00142 P12) x 0.9 = 459 + 6.48 P1 + 0.0128 P12 F2 (P2) = 310 + 7.85 P2 + 0.00194 P22 F3 (P3) = 78 + 7.97 P3 + 0.00482 P32

dF1 = 6.48 + 0.0256P1 dP1 dF2 = 7.85 + 0.00388P2 dP2 dF3 = 7.97 + 0.00964P3 dP3
The necessary conditions for an optimum dispatch are:

dF1 = 6.48 + 0.0256P1 = dP1 dF2 = 7.85 + 0.00388P2 = dP2 dF3 = 7.97 + 0.00964P3 = dP3
And P1 + P2 + P3 = 850 MW From equations (1), (2) & (3), we have. P1 = P2 = P3 = 6.48 0.00256 7.85 0.00388 7.97 o.oo964

(1) (2) (3)

(4)

70

Power system Operational Planning Economic Dispatch

Lecture No. 3

Substituting the value in equation (4), we get = 8.284 $ / Mwh Substituting the value of , we have optimal schedule as: P1 = 704.6 MW P2 = 111.8 MW P3 = 32.6 MW Pi = P1 + P2 + P3 = 849.0 Mw 850 Mw
This schedule satisfies the equality constraint, but unit 1 and unit 3 are not within the limits. Unit 1 violate upper limit & unit 3 violate lower limit. To solve this problem, we clamp the violated generations to their corresponding limits and optimal set is obtained by satisfying necessary conditions.

Unit 1 is fixed to Max limit i.e., P1 = 600 Mw Unit 3 is fixed to Min limit i.e., P3 = 50 Mw
Q P1 + P2 + P3 = 850

P2 = 850- (P1+P3) = 850- (600+50) = 200 Mw Now the schedule is as: P1 = 600 Mw P2 = 200 Mw P3 = 50 Mw The necessary conditions for this schedule are as:

C1 C2 C3

dF1 Q P1 = P1 (Max) = 600 MW (Clamped) dP1 dF2 = Since P2 is within inequality 200 P2 400 = 200Mw dP2 dF3 Q P3 = P3 (Min) = 50MW (Clamped) dP3

71

Power system Operational Planning Economic Dispatch

Lecture No. 3

dF1 dP1 dF2 dP2 dF3 dP3 dF1 dP1

600

= 6.48 + 0.0256(600) = 8.016$/MWh = 7.85 + 0.00388(200) = 8.626$/MWh = 7.97 + 0.00964(50) = 8.626$/MWh

200

50

Condition 1
600

= 8.016 < 8.626

True

Condition 2 = 8.626 $/MWh Condition 3 True

dF3 dP3

50

= 8.452 < 8.626

False

Condition 3 is violated, so for the optimal schedule, we keep unit 1 at its max limit, allow unit 2 and unit 3 incremental costs equal to . P1 = 600 Mw

dF2 = 7.85 + 0.00388P2 = dP2 dF3 = 7.97 + 0.00964P3 = dP3

P2 + P3 = 850 P1 = 250 MW From equations (1) & (2), we have
7.85 7.97 = 250 + 0.00388 0.00964 = 8.576 P2 + P3 =

(1) (2) (3)

Now the economic schedule for unit 2 and 3 corresponding to value of = 8.576 $/MWh
72 Power system Operational Planning Economic Dispatch

Lecture No. 3

P2 = 187.1 Mw P3 = 62.90 Mw

dF1 dP1

600

= 8.016 < 8.576

True True

dF2 dF3 = = 8.576. dP2 dP3

The figure represents the graph of the economic schedule.

PROBLEM # 3

Unit 1: Unit 2: Where

H1 = 80 + 8 P1 + 0.024 P12 H2 = 120 + 6 P2 + 0.04 P22

20 P1 100 20 P2 100

Hi = Fuel input to unit i in Mbtu/hour Pi = Unit output in Mw. 1. Plot input/output characteristic for each unit. 2. Calculate net heat rate in Btu/Kwh and plot against output in Mw 3. Assume Fuel cost 1.5 $/Mbtu. Calculate the incremental production cost in $/Mwh of each unit and plot against output in Mw.
Solution 1. Plot input/output characteristic for each unit. UNIT 1 P1 (Mw) H1 (Mbtu/h) P2 (Mw) UNIT 2 H2 (Mbtu/h)

20 30 40 50 60

249.6 341.6 438.4 540.0 646.4

20 30 40 50 60

256.0 336.0 424.0 520.0 624.0

73

Power system Operational Planning Economic Dispatch

Lecture No. 3

70 80 90 100

757.6 873.6 994.4 1120.0

70 80 90 100

736.0 856.0 984.0 1120.0

1200 1000 800 600 400 200 0 20 30 40

20 P1 100 H1=8 P1+0.024 P12+80

H1 BTU/hr

50

60
P1 (MW)

70

80

90

100

74

Power system Operational Planning Economic Dispatch

Lecture No. 3

1200 1000
H2 BTU/hr

20 P1 100 H2=6 P2+0.04 P22+120

800 600 400 200 0 20 30 40 50 60 70 80 90 100

P2 (MW)

b) Calculate net heat rate in Btu/Kwh and plot against output in MW

UNIT 1 P1 H1/P1 P2 UNIT 2 H2/P2

20 30 40 50 60 70 80 90 100

249.6/20 = 12.48 341.6/30 = 11.38 438.4/40 = 10.96 540/50 = 10.8 646.4/60 = 10.77 757.6/70 = 10.82 873.6/80 = 10.92 994.4/90 = 11.04 1120/100 = 11.20

20 30 40 50 60 70 80 90 100

256/20 = 12.80 336/30 = 11.20 424/40 = 10.60 520/50 = 10.40 624/60 = 10.40 736/70 = 10.50 856/80 = 10.70 984/90 = 10.93 1120/100 = 11.20

75

Power system Operational Planning Economic Dispatch

Lecture No. 3

H1/P1 vs P1
15 H1/P1 10 5 0 0 20 40 60 P1 80 100 120

H2/P2 vs P2
15 H2/P2 10 5 0 0 20 40 60 P2 80 100 120

2. Assume Fuel cost 1.5 $/Mbtu. Calculate the incremental production cost in $/Mwh of each unit ad plot against output in Mw. F1 (P1) = H1 x 1.5 = 120 + 12P1 + 0.036 P12 F1 (P2) = H2 x 1.5 = 180 + a P2 + 0.06 P22

dF1 = 12 + 0.072 P1 dP1 dF2 = 9 + 0.12 P2 dP2

76

Power system Operational Planning Economic Dispatch

Lecture No. 3

P1& P2 (Mw) 20 30 40 50 60 70 80 90 100

PROBLEM # 4 Unit 1: Unit 2: Coal Fired Steam Coal Fired Steam

dF1/dP1 ($/Mwh) 13.44 14.16 14.88 15.6 16.32 17.04 17.76 18.48 19.20

dF2/dP2 ($/Mwh) 11.4 12.6 13.0 15.0 16.2 17.4 18.6 19.8 21.0

H1 (P1) = 510 + 7.2 P1 + 0.00142 P12 150 P1 600 Mw Fuel Cost = 1.1 Rs/Mbtu H2 (P2) = 310 + 7.85 P2 + 0.00194 P22 100 P2 400 Mw Fuel cost = 1.0 Rs/Mbtu 1. Determine Economic Schedule for load demand of 728 Mw. 2. Determine total fuel cost/hour. 3. Determine the specific cost at Economic Scheduled generation. 4. Determine the average specific cost of unit 1 & unit 2. 5. For the following capacity charges determine the total cost per unit for unit 1 & Unit # 2 Capacity Charges for Unit 1 = 6 times the av. sp. cost Capacity Charges for Unit 2 = 4 times the av. sp. cost Solution F1 (P1) = H1 (P1) x 1.1 = 561 + 7.92 P1 + 0.001562 P12 F2 (P2) = H2 (P2) x 1.0 = 310 + 7.85 P2 + 0.00194 P12

77

Power system Operational Planning Economic Dispatch

Lecture No. 3

dF1 = 7.92 + 0.003124 P1 = L (1) dP1 dF2 = 7.85 + 0.00388 P2 = L (2) dP2 P1 + P2 = 728L (3)

For economic schedule the necessary condition is:

dF1 dF2 = = dP1 dP2

From =S (1) & (2), we have, P1 = 7.92 0.003124 P2 = 7.85 0.00388

Substituting in equation (3), we get, = 9.1486 Rs/Mwh Corresponding to this value of , the optimum values of P1 and P2 (satisfying equality constraints are : P1 = 393 MW P2 = 335 MW 2. Fuel cost per hour

Ft = F1 + F2 = F1

393

+ F2

335

= 3914.8 + 3157.5 Rs/h = 7072.3 Rs/h

3. The specific cost at Economic Scheduled generation.

Specific Cost of Unit 1. Sp Cost1 = 3914.8/393 Rs./Mwh = 9.96134 Rs./Mwh = 0.996134 Paisa/Kwh Specific Cost of Unit 2 Sp Cost2 = 3157.5/335 = 9.4253 Rs./Mwh

78

Power system Operational Planning Economic Dispatch

Lecture No. 3

= 0.9425 Paisa/Kwh Specific Cost of System = 9.801466 Rs./Mwh = 0.97146 Paisa/Kwh. 4. Specific Cost of Unit 1
% Load 0 25% 50% 60% 70% 80% 90% 100% Fuel Cost Per Hour (Rs./h) F1 P1=0 F1 P1=0 = Various Load = 561 Speicfic Cost (Rs./Mwh) F1 ( P1 ) / P1 561/0 = 178.145/150 = 11.8943 3077.58/300 = 10.2586 3614.64/360 = 10.0406 4162.94/420 = 9.911754 4722.48/480 = 9.83851 5293.28/540 = 9.80236 5875.32/600 = 9.7922

F1 P1=150 = 1784.145 F1 P1=300 = 3077.58 F1 P1=360 = 3614.64 F1 P1= 420 = 4162.94 F1 P1= 480 = 4722.48 F1 P1=540 = 5293.28 F1 P1=600 = 5875.32

Average Sp Cost of Unit 1 = 61.74618/6 = 10.29103 Rs./Mwh = 10.29103 Rs./Mwh (Average of Six readings 25% to 90%) Average Specific Cost of Unit 2. Average Sp Cost of Unit 2 = 58.88669/6 = 9.814778 Rs./Mwh = 0.9814778 Rs./Kwh (Average of Six readings 25% to 90%)
Capacity Charges Evaluation Unit 1

Rate of Capacity Charges = 6 x Av. Sp Cost = 6 x 10.29103 = 61.74618 Rs./Mwh = 6.174618 Ps./Kwh Capacity Charges/h for the declared capacity of 600 Mw = 61.74618 x 600 = 37047.708 Rs./h
Unit 2

Rate of Capacity Charges = 4 x Av. Sp. Cost

79 Power system Operational Planning Economic Dispatch

Lecture No. 3

= 4 x 9.814778 = 39.259 Rs./h Capacity Charges /h for the declared capacity of 400 Mw = 39.259 x 400 = 15703.6 Rs./hour. Total Cost per Unit of Energy = Capacity (fixed) Charges + Av. Sp Cost Unit 1 = 61.7461 + 10.29103 = 72.03721 Rs./Mwh Unit 2 = 39.259 + 9.81477 =49.07389 Rs./Mwh
3.9 REFERENCE

[1]. A.J.Wood & Bruce F. Wollenberg, Power Generation , Operation & Control, John Wiley & Sons, 1996.

80

Power system Operational Planning Economic Dispatch