Power System Operation

Chapter 3

Power system Optimization

1
• Introduction

• Optimization

• Power system optimization

Introduction to optimization in power system
What is optimization ?

What would you like to optimize in power system ?

This session will introduce different types of optimization problems,

with different levels of accuracy or system aspects included
Optimization
Optimization is the mathematical discipline which is concerned with finding the maxima and
minima of functions, possibly subject to constraints.

In multivariable function to find the maxima and minima the most commonly adopted method.
Lagrange multipliers.
Optimization characterization
X can be a vector with several variables

Minimization of f(x) is same as maximization

of –f(x)
Optimization using Lagrange multipliers
Assume an objective function f (x), and a set of inequality constraints h(x) and
equality constraints g (x)
We define the Lagrangian as:
𝐿(𝑥, 𝜆, 𝜇ሻ = 𝑓(𝑥ሻ + 𝜇𝑇 . ℎ(𝑥ሻ + 𝜆𝑇 . 𝑔(𝑥 ሻ

This is an augmented cost (objective) function, which has transformed our

constrained problem in x into an unconstrained problem in x, And λ or

𝑧 = [𝑥, 𝜆, 𝜇ሿ𝑇
General formulation of an optimization
problem
An optimization problem is generally written as follows:
Objective: min or max f (x)
s.t.:
Equality constraints: 𝑔(𝑥ሻ = 0
Inequality constraints: h 𝑥 >0

x εℜ𝑛 (of dimension n) x= [x1,x2, . . . ,xn]T ).

Assume that there are m equality constraints and p inequality constraints
Problem can be constrained or unconstrained
The optimal solution is X*
The initial guess is denoted as 𝑥0
New objective: minz L(z)
μ and λ are often just written as λ and are called the Lagrange
multipliers
In order for an optimum to be found, the following first order
conditions are necessary:
Optimization in power systems
Many power system problems can be formulated as an optimization problem
 Minimum losses in the system
 Searching for the optimal operating point within security bounds
 Maximum transfer capacity
 Best commitment of generators
 Lowest CO2 emissions
Multiple calculations are possible (brute force)
Optimization using mathematical methods
Meta-heuristic optimization through the use of genetic algorithms (e.g. simulated annealing, partical swarm
optimization,. . . )
In practice, we use a (mathematical or GA) solver instead of solving the equations
Constrained optimization
In power system operations we usually have a constrained optimization problem.
There are two type of constraints :
 Equality constraints
• In Electrical power system , the basic load flow equations are the equality
constraints
 Inequality constraints
• Voltage limits
• Generator limits (active and reactive power)
• Transmission line limits
• Reserve requirements
• Transformer tap settings
• Network Security constraints
Solution methods
 ECONOMIC OPERATION OF POWER SYSTEMS
 In EPS, the first step is to properly assess the load requirement of a given
area where electrical power is to be supplied.

 This power is to be supplied using the available units such as thermal,

hydroelectric, nuclear, etc.
 Generation costs can vary widely ,with different technology balances.
 The capital cost necessary to build the generator
 The costs to actually necessary to produce electric power (fuel +maintenance)
 For example ,nuclear and some hydro have high capital cost and low operating cost
and natural gas has low capital cost and higher operating costs
 Cont’d
 A power system has several power plants and each power plants has several generation
unit.
 At point of time ,the total load in the system is met by the different generating unit in the
power plant.
 Fuel cost is the major portion of the operating cost and it can be controlled

 Optimum economic efficiency is achieved when all the generators which are running in
parallel are loaded in such a way that the fuel cost of their power generation is the
minimum.
 One that deals with minimum cost of production is called “economic dispatch”
 In any EPS, the cost has two components as under:
 The Fixed Costs: Capital investment, interest charged on the money borrowed, tax paid, labor, salary, etc. which are
independent of the load variations.

 The Variable Costs: which are dependent on the load on the generating units, the losses, daily load requirements,
purchase or sale of power, etc.

 Factor affecting the operating cost of a power system are :

 Generator efficiency, transmission losses, fuel cost, etc.

 Of these, the fuel cost is the most important factor.

 As a given power system is a mix of various types of generating units, such as hydro, thermal,
nuclear, hydro-thermal, wind, etc., each type of unit contributes its share for the total operating cost.

 Since fuel cost is a predominating factor in thermal (coal fired) plants

 Economic load dispatch (ELD) is considered usually for a given set of thermal plants in the foregoing discussion.
Cost of producing electrical energy
The cost of producing electrical energy from a fuel depends on:
 Fuel type
 Efficiency of the power plant (age)
 Specific constraints (e.g. local temperature for cooling,. . . )
 Location of the plant

In case start-up, shut-down and operating are included in the optimization, the problem is called unit
commitment

Both Economic Dispatch and Unit Commitment can be performed with and without taking the power
system into account
 PERFORMANCE CURVES:

 The Performance Curves useful for economic load dispatch studies include
many different types of input-output curves as under:
 1. Input Output Curve: A plot of fuel input in Btu/Hr. as a function of the
MW output of the unit.

Boiler turbine unit

Input-output curve of thermal power plant
 The output is normally the net electrical output of the unit

 Heat Rate Curve: A plot of heat rate in Btu/kWH, as a function of the

MW output of the unit.
 Thus, it is the slope of the I-O curve at any point. The reciprocal of heat
rate is termed as the ‘Fuel Efficiency’.
 Incremental Fuel Rate Curve: A plot of incremental fuel rate (IFC) in
Btu/kWH as a function of the MW output of the unit, where,
 IFC= ∆input/∆output = Incremental change in fuel input/ Incremental
change in power output
Incremental Fuel Cost Curve: A plot of incremental fuel cost (IFC) in Rs./kWH as a
function of the MW output of the unit, where, IFC in Rs./kWH = (Incremental fuel rate
in Btu/kWH) (Fuel cost in Rs./Btu)

The IFC is a measure of how costlier it will be to produce an increment of power

output by that unit.
For reference
1 Btu (British thermal unit) = 1054 J
1 MBtu = 1x106 Btu
1 MBtu = 0.29 MWh
 Economic Dispatch: Formulation

• The objective of the ED is to calculate the instantaneous output power of the

generating units such that all the load demands are satisfied with minimum
costs subject to a constraints that the total
• Generation = total load + losses
• The cost of operating generator i , is a function of the generation of
• generator i : C(PGi )
m
Minimize CT   Ci ( PGi )
i 1
Such that
m
 PGi  PD  PLosses
i 1
Economic Dispatch Lagrangian
For the economic dispatch we have a minimization
constrained with a single equality constraint
m m
L(PG ,  )   Ci ( PGi )   ( PD   PGi ) (no losses)
i 1 i 1
The necessary conditions for a minimum are
L dCi
(PG ,  )  ( PGi )    0 (for i  1 to m)
PGi dPGi
m
PD   PGi  0
i 1
Cost functions of generators
Generators cost curves are not usually smooth. However the curves can be
adequately approximated using piece-wise smooth, functions.
Marginal cost or incremental cost is the derivative of the cost function
Polynomial (mostly quadratic or cubic) functions

Piece-wise linear function

What is economic dispatch for a two generator
system PD  PG1  PG 2  500 MW and
C1 ( PG1 )  1000 20 PG1  0.01PG21 $/h
C2 ( PG 2 )  400 15PG 2  0.03PG22 $/h
Using the Lagrange multiplier method we know:
dC1
( PG1 )    20  0.02 PG1   0
dPG1
dC2
( PG 2 )    15  0.06 PG 2   0
dPG 2
500  PG1  PG 2  0
We therefore need to solve three linear equations
20  0.02 PG1   0
15  0.06 PG 2   0
From the solution ,both generator have same marginal
500  PG1  PG 2  0 cost ,and this common marginal cost is Lamda.
0.02 0 1  PG1   20 
 0 0.06 1  PG 2    15 
Intuitively:
     If marginal costs of generators were different, then by
 1 1 0      500  decreasing production at higher marginal cost generator,
and increasing production at lower marginal cost of the
 PG1   312.5 MW 
 P    187.5 MW  generator we could lower overall costs.
 G2   
    26.2 $/MWh 
 CONSTRAINTS IN ECONOMIC OPERATION OF POWER SYSTEMS:
• Various constraints are imposed on the problem of economic operation of power
systems as listed below:
 Primary constraints (equality constraints):
• Power balance equations:
• Pi - PDi - PL=0; Qi - QDi - QL =0; i=buses of the system

i = 1,2,….n, are the power flow to the neighboring system.

• The above constraints arise due to the need for the system to balance the
generation and load demand of the system.
 Secondary constraints (inequality constraints):

These arise due to physical and operational limitations of the units and components.
𝑃𝑖𝑀𝑖𝑛 <𝑃𝑖 < 𝑃𝑖𝑀𝑎𝑥 , Active power

𝑄𝑖𝑀𝑖𝑛 <𝑄𝑖 < 𝑄𝑖𝑀𝑎𝑥 reactive power ,i = 1, 2,….n, the number of generating units in the system.

 Spare Capacity Constraints:

These are used to account for the errors in load prediction, any sudden or fast change in load demand, inadvertent
loss of scheduled generation, etc.
Here, the total generation available at any time should be in excess of the total anticipated load demand and any
system loss by an amount not less than a specified minimum spare capacity, PSP (called the Spinning Reserve) given
by:

PlG (Generation) ≥ Σ PL (Losses) + PSP + PDj (Load)

 Thermal Constraints:
• For transmission lines of the given system:

• 𝑆𝑖𝑀𝑖𝑛 < 𝑆𝑖𝑏 < 𝑆𝑖𝑀𝑎𝑥, i = 1,2,….n , the number of branches, where, 𝑆𝑖𝑏 the branch
transfer MVA. bi
 Bus voltage and Bus angle Constraints:
 Bus voltage and Bus angle Constraints are needed to maintain a flat bus
voltage profile and to limit the overloading respectively.
 𝑉𝑖𝑀𝑖𝑛 < 𝑉𝑖𝑏 < 𝑉𝑖𝑀𝑎𝑥 , i = 1, 2,….n
 δ𝑖𝑀𝑖𝑛 < δ𝑖𝑏 < δ𝑖𝑀𝑎𝑥 i = 1,2,….n; j = 1,2,….m
 where, n is the number of nodes and m is the number of nodes neighboring
each node with interconnecting branches.
• Other Constraints:
 In case of transformer taps, during optimization, it is required to satisfy
the constraint:
 𝑇𝑖𝑀𝑖𝑛 < 𝑇𝑖 < 𝑇𝑖𝑀𝑎𝑥
 where 𝑇𝑖 is the percentage tap setting of the tap changing transformer
used.
 In case of phase shifting transformers, it is required to satisfy the
constraint:
 𝞅𝑖𝑀𝑖𝑛 < 𝞅𝑖 < 𝞅𝑇𝑖𝑀𝑎𝑥
 where 𝞅 i is the phase shift obtained from the phase shifting transformer
used.
 ELD with capacity constraints:

Considering N generating unit supplying a Load PD MW, and taking the MW limit and
neglecting the transmission loss.
 Suppose the fuel cost of unit ‘i’ is given by:
2
 𝐶𝑖 (𝑃𝑖 ሻ = 𝑎𝑖 +𝑏𝑖 *𝑃𝑖 +𝑐𝑖 *𝑃𝑖 $/Hr ,so that the total cost, 𝐶𝑇 = Σ 𝐶𝑖 (𝑃𝑖 ) i= 1,2,… N
 The ELD problem can now be stated mathematically as follows:
 Minimize 𝐶𝑇 = Σ 𝐶𝑖 (𝑃𝑖 ) ,i = 1,2,… N Such that Σ Pi = PD and, 𝑃𝑖𝑀𝑖𝑛 <𝑃𝑖 < 𝑃𝑖𝑀𝑎𝑥

𝑑𝐶𝑖
=λ for 𝑃𝑖,𝑚𝑖𝑛 < 𝑃𝑖, < 𝑃𝑖,𝑚𝑎𝑥
𝑑𝑃𝑖
𝑑𝐶𝑖
λ for 𝑃𝑖, =𝑃𝑖,𝑚𝑎𝑥
𝑑𝑃𝑖
𝑑𝐶𝑖
λ for 𝑃𝑖, =𝑃𝑖,𝑚𝑖𝑛
𝑑𝑃𝑖
 From the above equations, if the outputs of the unit, according to optimality
rule, is:
 Less than its minimum value, then it is set to 𝑃𝑖,𝑚𝑖𝑛 , the corresponding IC
will be greater than the system λ,
 More than its maximum value, then it is set to 𝑃𝑖,𝑚𝑎𝑥 , the corresponding
IC will be less than the system λ,
 With in its maximum and minimum values, then the corresponding IC
will be equal to the system λ.
 ELD with Transmission losses
In determining the economic distribution of load between plants, we encounter the need to consider
losses in the transmission lines.
 For system with N generating units ,the total cost of all the fuel for the entire system is the sum of
the fuel cost of the individual units
 𝐶𝑇 = σ𝑁 𝑖 𝐶𝑖 (𝑃𝑖 ),i = 1,2,… N
The total Megawat power input to the network from all the unit is the sum
 𝑃𝑇 = σ𝑁𝑖 𝑃𝑖 ,i = 1,2,… N , Where 𝑃𝑖 , the ith individual output of the units injected to the
network
the constraints equation on the minimum value of the cost function is given by
 𝑃𝐿 +𝑃𝐷 - σ𝑁 𝑖 𝑃𝑖 =0, where 𝑃𝐷 is the total power delivered to loads , 𝑃𝐿 is the transmission loss of
the system
Using the Lagrange multipliers, the new cost function formed by combining the total fuel cost and
the equality constraints
 𝐿 = 𝐶𝑇 +λ*(𝑃𝐿 +𝑃𝐷 - σ𝑁 𝑖 𝑃𝑖 )
Example :
• Incremental fuel cost in dollars per Mhr for a plant consisting of two units are given by:
• λ1 = 0.008𝑃1 +8;
• λ2 = 0.0096𝑃2 +6.4;
• Assume that both units are operating a t all times, that total load varies from 250 to1250 MW, and
that maximum and minimum loads on each unit are to be 625 and100 MW, respectively Find the
incremental fuel cost of the plant and the allocation of load between units for the minimum cost of
various total loads.
• SOLUTION: at light load generator 1 will have high incremental fuel cost and operates at its lower
limits of 100MW for which the λ1 is is $8.8/MWh. When the output of generator 2 is also 1 00
MW, λ2 is $7.36/MWh.
• Therefore, as plant output increases, the additional load should come
from generator 2 until equals $8.8/MWh . The plant λ and outputs of
each unit for various values of total output 𝑃𝑇
• Therefore, for minimum cost the derivative of L with respect to each Pi equals
to zero ,
• Since PD is fixed and the fuel cost of any one unit varies only if the power
output of the unit is varied
𝜕𝐿 𝜕𝐶𝑖 𝑃𝐿
• = + λ( − 1)=0
𝜕𝑃𝑖 𝜕𝑃𝑖 𝜕𝑃𝑖
1 𝜕𝐶𝑖
•λ = 𝑃𝐿 ∗
(1−𝜕𝑃 ሻ 𝜕𝑃𝑖
𝑖

• The equation is sometimes written in the form

𝜕𝐶𝑖
• λ =𝐿𝑖 ∗
𝜕𝑃𝑖
• where 𝐿𝑖 ∗ is called the penalty factor of the plant and given by
1
𝐿𝑖 =
𝑃𝐿
(1 − ሻ
𝜕𝑃𝑖
• The minimum fuel cost is obtained when the incremental fuel cost of each unit
multiplied by its penalty factor is same for all generating units in the system
𝜕𝑃
• The penalty factor 𝐿𝑖 depends on 𝐿 , which is a measure of the sensitivity of the
𝜕𝑃𝑖
transmission system losses to changes in 𝑃𝑖 alone
• To include the effect of the transmission loss ,the total loss is expressed as a
quadratic function of the generator
• SPINNING RESERVE
 Spinning reserve (SR) is the term used to describe the total amount of generation available from all the
synchronized (spinning) units of the system minus the present load plus the losses being supplied. i.e.,

 Sp.Res., PSP = {Total generation, ΣPlG} - { Σ PDj(load) + Σ Pl (losses)}

 The SR must be made available in the system so that the loss of one or more units does not cause a large
drop in system frequency.

 SR must be allocated to different units based on typical Council rules. One such rule is as follows:

 ‘SR must be capable of making up for the loss of the most heavily loaded unit in the system

 ’ Reserves must be spread around the system to avoid the problem of ‘bottling of reserves’ and to
allow for the various parts of the system to run as ‘islands’, whenever they become electrically
disconnected.
• Q1. The I-O characteristics of two generating units are as follows:
2
𝐹1 = 0.0008P𝐺1 +0.2P𝐺1 +5𝐵𝑡 𝑢Τℎ

2
𝐹2 = 0.0005P𝐺2 +0.3P𝐺2 +4𝐵𝑡 𝑢Τℎ

Determine the economic operation of those two units when

delivering a total 500MW power demand
Solution: The incremental fuel rate of the two generating units can
be obtained as follows:

At equal incremental rate, the economic operation is optimal,

thus at 𝜆1 = 𝜆2

Given a system Load of 500MW ,then :

Q2
• Suppose the I-O characteristics of three generating units are as follows:

We wish to determine the economic operation point for these units when delivering a total of
500MW and 800MW power demand ,respectively, considering the inequality constraints of the
three units which are given as below.
• Solution :

When Delivering a total demand of 500MW ,the dispatch is given as following :

By checking the inequality constraints of units, we know that all the power outputs of the units are
within the limits. Thus they are the optimum results, and there is no violation of the inequality
constraints.
When delivering a total demand of 800MW power demand , the dispatch is as follows:

By checking the inequality constraints of units, we can see that the power output of unit 1
violated its upper limit. Therefore , for economic operating point , we set unit to the
maximum limit which is 250MW ,and finally we get new power balance for the economic
operating point by subtracting unit 1 and we get new power balance

Applying the principle of equal incremental rate for units 2 and 3, we have
 Then we can get the following two equations

Solving the above the equations, the power dispatch becomes

 Exercise
Q3. the I-O characteristics of two steam plant can be expressed analytically as under
with 𝑃𝐺1 and 𝑃𝐺2 in MW

𝐶𝑇1 = (2.3 ∗ 𝑃𝐺1 + 0.0062 ∗ 𝑃𝐺1 2 +25ሻ106 𝑘𝑐𝑎𝑙 𝑠Τ𝐻 𝑟

2
𝐶𝑇2 = (1.5 ∗ 𝑃𝐺2 + 0.01 ∗ 𝑃𝐺2 + 35ሻ106 𝑘𝑐𝑎𝑙 𝑠Τ𝐻 𝑟

The calorific value of coal at plant#1 and plant#2 are respectively equal to 4000
kCals/kg. and 5000 kCals/kg. The corresponding cost of coal is $.55/- and $.65/-per
Ton. Find the following: (i) Incremental Fuel Rate in kCals/MWHr (ii)Incremental
Fuel Cost in $./MWHr
• Q4 The incremental costs of a two unit system are given by:
• λ1 = 0.008𝑃1 +8;
• λ2 = 0.0096𝑃2 +6.4;

• Find the incremental cost and the distribution of loads between the
two units for optimal operation for a total load of 1000 MW. What is
this value if the same total load is equally shared among the two
units?
• Q5. The fuel costs in $ ./Hr. for a plant of three units are given by:

2
𝐶𝑇1 = (40 ∗ 𝑃𝐺1 + 0.1 ∗ 𝑃𝐺1 +100ሻ $Τ𝐻 𝑟

2
𝐶𝑇2 = (0.125 ∗ 𝑃𝐺2 + 30 ∗ 𝑃𝐺2 +80ሻ $Τ𝐻 𝑟

2
𝐶𝑇3 = (20𝑃𝐺3 + 0.15 ∗ 𝑃𝐺3 +150ሻ $Τ𝐻 𝑟

Find the incremental cost and the distribution of loads between the three units

for optimal operation for a total load of 400 MW, given that the max. and min.

capacity limits for each of the units as 150 MW and 20 MW respectively.

Q6 𝐺𝑖𝑣𝑒𝑛 𝑡ℎ𝑎𝑡 ∶
𝐶𝑇1 = (0.2𝑃𝐺1 + 40ሻ $Τ𝐻 𝑟
𝐶𝑇2 = (0.25𝑃𝐺1 + 30ሻ $Τ𝐻 𝑟

Calculate and tabulate the load shared by each unit for optimal operation
if the total load varies from 50 to 250MW, in steps of 50MW, given that
max.MW is 125 and min. MW is 20 for both the units.
Q7. Given that :

𝐶𝑇1 = (30𝑃𝐺1 + 0.09 𝑃𝐺1 2 + 110ሻ $Τ𝐻 𝑟 ,12 ≤ 𝑃𝐺1 ≤ 125 MW

2
𝐶𝑇2 = 12𝑃𝐺1 + 0.1 𝑃𝐺1 + 135 $Τ𝐻 𝑟 , 25 ≤𝑃𝐺1 ≤ 125 MW

Calculate and tabulate the load shared by each unit for optimal sharing of the
total load in the range 50-250 MW in steps of 100 MW. Also find λ system in
each case
Q 8. Three plants of total capacity 500 MW are scheduled for
operation to supply a total load of 310 MW.
Find the optimal load schedule if the IC curves and limitations are:

• λ1 = 30+ 0.12𝑃𝐺1 30 ≤ 𝑃𝐺1 ≤ 150 MW

• λ2 = 40+ 0.20𝑃𝐺2 20 ≤ 𝑃𝐺2 ≤ 100 MW
• λ3 = 10+ 0.16𝑃𝐺3 50 ≤ 𝑃𝐺3 ≤ 250 MW