Course Name: Power System Operations, Control and

Market Mechanism
Course Code: EEE 465
Credit: 3.00
 Optimal Power Flow
8.1 INTRODUCTION
❑Economic dispatch (ED) is the calculation that finds the lowest-cost generation
dispatch for a set of generators that is constrained within the individual generator
limits and results in a total generation that equals the total load plus losses.

❑ED calculation ignores the specific details of the network that the generators are
connected to and lumps all the network effects into the losses and total load
demand.

❑ED calculation ignores the effect that the dispatch of generation has on the loading
of transmission branches or the effect it has on bus voltages.

❑The dispatch of generation does, in fact, have an important effect on transmission

flows.
 Optimal Power Flow
8.1 INTRODUCTION
❑ OPF, as the name implies, couples the ED calculation with a power flow calculation
so that the ED and the power flow are solved simultaneously.

❑The total losses of the power system are simply part of the power flow calculation
and are reflected in the loading of the generation on the reference bus—thus, there
is no need to specifically calculate the losses since they are inherently a part of the
power flow.

❑More importantly, the ED can be constrained to meet transmission system limits

such as MW or MVA flow limits on lines or transformers or voltage limits on buses.

❑The result is the generation dispatch representing the minimum $/h total generation
cost and that also solves the power flow at that optimum.
 8.2 THE ECONOMIC DISPATCH FORMULATION (Revisited)
In the ED calculation, we solve the following problem:
1. Objective function = total generation cost in $/h.
2. Each generator is within its minimum and maximum limits.
3. Sum of all generator outputs = total load + losses.
 8.2 THE ECONOMIC DISPATCH FORMULATION (Revisited)
We can write the ED problem as –

(the u vector appearing in the f (Pgen , u) now contains the parameters of the generator cost
functions) and is subject to –
8.3 THE OPTIMAL POWER FLOW CALCULATION COMBINING
ECONOMIC DISPATCH AND THE POWER FLOW
In the OPF calculation, we solve the following more complex problem:
❖ Objective function = total generation cost in $/h (same as ED)

❖ Generator limit inequality constraints (same as ED)

❖ For each bus,

❑ Generation minus load at the bus = the net power flow at the bus. This is done using
the Y matrix power flow equation for the bus.

❑ This introduces the voltage magnitude and phase angle of each bus, except for the
reference bus, as unknowns into the problem.

❑ This also adds 𝟐 𝑵𝒃𝒖𝒔 − 𝟏 new equality constraints into the problem representing the
𝑁𝑏𝑢𝑠 − 1 complex Y Matrix equations for all buses except the reference bus.
➢The power flow itself results in a set of generator outputs that equals
the load plus losses as required.

➢Do not have to explicitly calculate the losses or the generator incremental
losses as this is all incorporated in the AC power flow equations.

➢As in a power flow, must force the reference-bus voltage magnitude

and phase angle to be fixed quantities.

subject to
❑Where, the vector of parameters u now contains –

1. the generator cost function parameters,

2. the generator real and reactive power limits, and
3. all the fixed parameters for the power system’s transmission system.

❑The Y matrix includes –

1. the bus fixed voltages at generators that are regulating,

2. the reference-bus fixed voltage magnitude and phase angle.
❑The equation 𝝎 𝑷𝒈𝒆𝒏 , 𝒖 = 𝟎 is now actually 𝟐𝑵𝒃𝒖𝒔 equality constraints representing
the power flow Y matrix equations, and

❑𝒈 𝑷𝒈𝒆𝒏 , 𝒖 ≤ 𝟎 is a set of inequality constraints representing every generator real

power limits and reactive power limits.

❑The great advantage to this formulation is the fact that in addition to a very accurate
calculation of losses inside the OPF, other transmission system constraints can be
added to the set inequality constraints.
❑For example,
❑For each transmission line or transformer,
1. MW flow or MVA flow is less than or equal to the branch’s MW flow or MVA flow
limit:

or,

The result is a dispatch of generation that does not overload the transmission line. In
addition, each bus voltage magnitude in the power system can be forced to be within
limits:
2. For each bus in the network,
8.4 OPTIMAL POWER FLOW USING THE DC POWER FLOW

❑Formulate the OPF with the transmission system constraints represented using a DC
power flow.
❑The DC power flow itself, it gives a good approximation to the AC power flow and is
much faster and easier to set up and solve.

❑Once again we start with the objective function:

Formulate the DCOPF problem using the Lagrange multiplier approach of
ELD. We can write the basic formulation of the problem as –

▪ This formulation includes a term at the right-hand end that is there as an equality
constraint to force the reference-bus phase angle to 0 (without this
constraint the OPF will not solve properly) .
▪ The generator inequality constraints can then be added so that we obtain –
8.5 EXAMPLE 8A: SOLUTION OF THE DC POWER FLOW OPF
The expression for the Lagrangian with the Power Flow Equations written out becomes

where we have now assumed the reference bus to be bus 1.

❖Solve the Lagrangian by taking the derivatives of L with respect to each
independent variable in the problem.
❖The independent variables are –
8.6 EXAMPLE 8B: DCOPF WITH TRANSMISSION LINE LIMIT IMPOSED
✓ Add a constraint to the system representing the flow limit on line 1–2. The
limit will be set to 150MW.

✓ Since the flow on line 1–2 was at 158.1776MW when we dispatched it

earlier, imposing the 150MW limit will result in the line being held at the
limit. Thus, we can simply add another equality constraint to the Lagrangian.

✓ Note that normally we shall use an inequality constraint for line flow limits,
but in this introduction we will use the equality constraint.
✓ The added constraint is –
The set of linear equations is shown as follows –
 The lambda values are

Because the line flow on line 1–2 is at its limit, the lambda values are different, the LMP
values for the buses are equal to these lambda values. Finally, the bus phase angles and power
flows are -
8.9 SOLUTION OF THE ACOPF

To solve the ACOPF, the bus power equations require a real and an
imaginary part to take care of the real power and reactive power.
3.3 Characteristic features of OPF
❖ The main features of an OPF is its ability to include the detailed
network configuration and bus-wise demand balance for both active
and reactive power, within the optimization framework which, thereby,
enables an exact representation of incremental losses.

❖ The OPF can include many operating constraints and model other issues
also. Specifically- limits on reactive power generation in addition to real
power generation, power flow limits (in either MW or MVA) on the
transmission lines and limits on the bus voltages ensure that the system
is operated in a secure manner.
3.3.1 OPF Applications
One of the main features of the OPF is that it is a flexible analytical tool, which allows
the use of different objective functions to solve different problems. The objective
functions commonly used for operations and planning studies are as follows –

❑ Minimize cost of operation

❑ Minimize the deviation or minimize control shift
❑ Loss minimization
❑ Minimize the cost of load curtailment
❑ Minimize the cost of installation of new capacitors and reactors and/or cost of
MVAR supplied
❑ Minimize the total emissions, and so on.
 Unit Commitment
Ref Text Book of K. Bhattacharya Ref Text Book of Allen J. Wood
• Chapter 2: Power System • Chapter 4: Unit Commitment
Economic Operation Overview • Art. 4.1, 4.1.1, Example (4A-4C)
• Art. 4,4.1,4.2
4. UNIT COMMITMENT (UC)
❑ UC problem is a somewhat longer term scheduling problem usually covering a
time range from 24 hours (1 day) to 168 hours (1 week) ahead, and is handled by
the operator in the pre-dispatch stage.

❑ The operator needs to take decisions on how to commit or de-commit (i.e. keep
running or shut down) its available units over the week, or over the next day.

❑ The input to the operator is the demand forecast for the next week or next day, as
the case may be, aggregated for the whole system.

❑ In the same way, that the ELD problem is formulated, the operator seeks to
minimize its system costs over the planning horizon in a UC problem, while
meeting the forecast demand to decide upon the unit up/down status for every
hour.
4. UNIT COMMITMENT (UC)
❑Planning activity is essential as the system load varies over a day or even
over a weekly period and hence it is not economical to keep all the units
on-line for the entire duration.

❑A proper schedule for starting up, or shutting down the units can save
costs significantly.

❑UC problems are much more complex to solve, compared to the ELD
problem discussed earlier, due to the presence of binary decision variables
on unit status (on/off).

❑Depending upon the need of the system and computational facilities, the
utilities choose to use UC models that suit their requirements.
4.1.1 Economic Dispatch vs. Unit Commitment
❑ The essential difference between the UC and ED problem.

Economic Dispatch (ED) Unit Commitment (UC)

• The ED problem assumes that there • The UC problem, on the other hand, is
are 𝑁𝑔𝑒𝑛 units already connected to more complex.
the system. • We may assume that we have 𝑁𝑔𝑒𝑛
• The purpose of the economic dispatch units available to us and that we have a
problem is to find the optimum forecast of the demand to be served.
operating policy for these 𝑁𝑔𝑒𝑛 units.

The question that is asked in the UC problem area is approximately as follows:

4.1.1 Economic Dispatch vs. Unit Commitment
❑ The UC problem may be extended over some period of time, such as the 24 h of a day or
the 168h of a week. The UC problem is a much more difficult problem to solve.

❑ The solution procedures involve the economic dispatch problem as a

subproblem. That is, for each of the subsets of the total number of units that are to be
tested, for any given set of them connected to the load, the particular subset should be
operated in optimum economic fashion.

❑ This will permit finding the minimum operating cost for that subset, but it does not
establish which of the subsets is in fact the one that will give minimum cost over a
period of time.
If we are to supply a load of 550 MW, what unit or combination of units should be used to
supply this load most economically?
 Example 4A
❑ To solve this problem, simply try all combinations of the three (03) units.
❑ Some combinations will be infeasible if the sum of all maximum MW for the units
committed is less than the load or if the sum of all minimum MW for the units
committed is greater than the load.
❑ For each feasible combination, the units will be dispatched using the techniques of ED.
The results are presented in Table 4.1.
 Example 4A

❑Note that the least expensive way to supply the generation is not with all
three units running or even any combination involving two units.

❑ Rather, the optimum commitment is to only run unit 1, the most

economic unit. By only running the most economic unit, the load can be
supplied by that unit operating closer to its best efficiency.

❑If another unit is committed, both unit 1 and the other unit will be loaded
further from their best efficiency points such that the net cost is greater
than unit 1 alone.
 Example 4B
❑ Suppose the load follows a simple “peak-valley” pattern as shown in Figure 4.1.
❑ If the operation of the system is to be optimized, units must be shut down as the load
goes down and then recommitted as it goes back up.
❑ We would like to know which units to drop and when. One approach to this solution is
demonstrated in Example 4B, where a simple priority-list scheme is developed.
 Example 4B
Question: Suppose we wish to know which units to drop as a function of system load. Let
the units and fuel costs be the same as in Example 4A, with the load varying from a peak of
1200 MW to a valley of 500 MW.
Solution:
❖ To obtain a “shutdown rule,” simply use a brute-force technique wherein all
combinations of units will be tried (as in Example 4A) for each load value taken in steps
of 50 MW from 1200 to 500.

❖ The results of applying this brute-force technique are given in Table 4.2. Our shutdown
rule is quite simple.
4.1 DC: The Basic Mode
4.1 DC: The Basic Mode

▪ The operator's objective function, while solving the UC problem still remains the
same, i.e., to minimize the system costs.
▪ However, due to the longer time-scale of the problem, the total system cost will be
affected by the start-up and shut down decisions of generating units, effected within
the planning period. These are explained below –
1) Fuel Cost: There have been two different approaches to represent fuel
costs in UC models. The first and the most common approach has been
to use a cost characteristic derived from the heat-rate characteristics and
represented by a polynomial function, which is usually quadratic, and
can be written as follows,

The other approach has been to represent the generator cost as a constant, which
is derived from the generator's average full load cost.
2) Start Up Cost: This component appears in the UC objective function to take into
account the cost incurred during a generator start-up operation. This is often modeled as
a function of the time for which the unit was off-line.
3) Shut Down Cost: Usually this component of cost is not considered in UC models
since it is not very significant compared to other costs. However, a constant cost
representation has been used in, and is included when the unit undergoes a shut down.

The composite objective function for the UC problem can be constructed using the above
as follows:

❑ UST, USD and W are integer decision variables denoting the status of the unit at
hour K.
❑ W denotes the unit status (1 = running, 0 = off);
❑ UST denotes the unit start-up state (1 = start-up, 0 = no start-up) and
❑ USD denotes the unit shut down state (1 = shut down, 0 = no shut down).
 4.1.2 Constraints in Unit Commitment
➢ This constraint ensures that the operator has scheduled enough capacity for a particular hour so that the
demand at that hour is met.

➢ This can also include any pre-decided import or export contracts with other utilities, and a certain amount
of reserve capacity. A typical demand-supply balance constraint is given as follows –

➢ The term RESV denotes the spinning reserve in the system, which is a reserve available to the
operator from among its spinning units, i.e. from the generators already running.

➢ Therefore, this reserve is available almost instantaneously to the operator in case of need.

➢ The operator has the very important responsibility of maintaining adequate spinning reserves
in the system, not only on a total-MW basis, but he also needs to take care of the location aspect
of this reserve, taking into account transmission capacities available in the system.
 4.1.2 Constraints in Unit Commitment
➢ The operator generally uses his experience or certain rules for determining this reserve to
be maintained in his system.

➢ The RESV component could typically comprise a base component, a fraction of the load
requirement and a fraction of the high operating limit of the largest on-line unit.

4.1.2.2 Minimum Up and Down Time Constraints on Thermal Units

✓ These constraints ensure the minimum number of hours a unit must be on, before it can
be shut down (minimum up-time) or,

✓ The minimum number of hours a unit must be off-line before it can be brought on-line
again (minimum down-time).

✓ These constraints are usually applicable to large thermal units.

4.1.2.2 Minimum Up and Down Time Constraints on Thermal Units

❑ MUT is the minimum time in hours the unit should be running before shut down;

❑ MDT is the minimum time in hours the unit should remain shut down before start-up,

4.1.2.3 Generation Limit

This constraint describes the allowable range of generation available for scheduling, as
defined by the maximum and minimum limits of the unit.
4.1.2.4 Must-Run Units
❖ Some units such as large coal-based units or nuclear units cannot be start up or shut down
on a day-to-day basis following the daily load variations because they involve very high
start-up costs and other technical constraints.
❖ Such units need to be assigned a must-run status.

4.1.2.5 Crew Constraints

These constraints pertain to the number of units that can be started at the same time in a
particular plant.
4.2 UC: Additional Issues

4.2.1 Multi-Area UC

UC has been adapted in the context of multi-area systems where individual utilities are
interconnected by tie lines and the dispatch is carried out jointly.

4.2.2 Fuel Constrained UC

❖ Fuel supply planning is an important issue , which when properly coordinated with
generation scheduling can reduce system costs significantly.

❖ The fuel co-ordination problem has been considered over a time horizon of one-month, for
three types of units (a) oil-fired (b) fuel constrained LNG/LPG fired and (c) pumped
storage hydro units.
4.2 UC: Additional Issues

4.2.3 Ramp Rate Constraint on Thermal Units

This constraint limits the inter-hour generation change in a unit, and is particularly applicable
to coal-based thermal units.

RUP and RON are the ramp-up and ramp-down constants of a unit.

4.2.4 Transmission Constraints

4.2.5 Environmental Issues

Example 4C: Suppose a power system consisted of two isolated regions: a western region
and an eastern region.
✓ Five units, as shown in Figure 4.3, have been committed to supply 3090 MW.

✓ The two regions are separated by transmission tie lines that can together transfer a
maximum of 550 MW in either direction.

✓ What can we say about the allocation of spinning reserve in this system?
With the exception of unit 4, the loss of any unit on this system can be covered by the spinning reserve on
the remaining units. Unit 4 presents a problem, however. If unit 4 were to be lost and unit 5 were to be run
to its maximum of 600MW, the eastern region would still need 590MW to cover the load in that region.
The 590MW would have to be transmitted over the tie lines from the western region, which can easily
supply 590MW from its reserves. However, the tie capacity of only 550MW limits the transfer. Therefore,
the loss of unit 4 cannot be covered even though the entire system has ample reserves. The only solution
to this problem is to commit more units to operate in the eastern region.