Unit Commitment Problem and Solution

Reference Book: J. J. Grainger, W. D. Stevenson, “Power System Analysis”, McGraw Hill

Problem: Schedule fossil fuel-fired thermal units by minimizing the aggregate cost (stat-up cost,
operating fuel costs, shut-down costs).

Assumption:
 Neglect transmission line losses.

total available generation

total system load

 System load on each sub-interval of the load cycle is constant.

 Start-up cost and shit-down costs are fixed.

Problem formulation:

( ) combination of interval
( ) minimum production cost of combination ( )
( ) sum of economic dispatch costs of individual units.
( ) cost of transition from combination ( ) to combination ( ) between intervals and

( ) cost associated with stage

( ) ( ) ( )

 Assume that the system load levels at the beginning and end of the day are same. Thus, the state
of the system is same at the beginning and at the end of the day.
( )
 The minimum unit commitment cost is found by minimizing the sum of the current single-stage
cost plus the minimum cumulative cost over the later stages of the study.
 The computational procedure known as “dynamic programming” can be used to solve the unit
commitment problem.
 Recursive formula to find out the feasible combinations using the dynamic programming
approach is;
( ) { ( ) ( ) ( )}
{ ( )}

best combination for stage

( ) minimum cumulative cost at stage

Page 1 of 7
Example

The system load in the figure below is to be supplied by combinations of the four generating units given
in the table.

Generating Loading Limits Fuel Cost Parameters

Unit Number Min (MW) Max (MW)
1 100 625 0.0080 8.0 500
2 100 625 0.0096 6.4 400
3 75 600 0.0100 7.9 600
4 75 500 0.0110 7.5 400

Load (MW)

2000
1800

1600
1600 1400
1400

1200 1100 1100

800

400
0
Time, (h)
4 8 12 16 20 24
Assuming the start-up cost of each thermal generating unit is $3000, and the shut-down cost is $1500,
determine the optimal unit commitment policy to serve the system load. Only must-run units 1 and 2 are
to operate at the first and final stages of the load cycle.

Answer

There are 4 discrete load levels in the daily load curve and 6stages (duration of each stage is 4 hours).
There are 4 units to meet the load.
Total number of combinations

Unit
1 1 1 1 1 0 0 1 0 1 1 0 1 0 0 0
2 1 1 1 0 1 1 0 0 1 0 1 0 1 0 0
3 1 1 0 1 1 0 0 1 0 1 1 0 0 1 0
4 1 0 1 1 1 1 1 1 0 0 0 0 0 0 1

Since, unit 1 and unit 2 must run, the possible combinations are

Page 2 of 7
 First, it is required to find out the output of each unit and the associated production cost of each
combination to serve the daily load curve.
 Perform the economic dispatch analysis as illustrated below:

Average fuel cost of plant ( ⁄ )

Incremental fuel cost ( ⁄ )

When two units are on economic dispatch:

( ) ( )

(∑ ) ( ) (∑ ) (∑ )

General scenario with number of generators to meet the load:

(∑ ) ∑ (∑ ) (∑ ) (∑ )

 Example calculation:
Load level
Combination [ ]

(∑ )

(∑ ) (∑ )

Page 3 of 7
 Average fuel cost of each unit:

( ⁄ )

( ) ( ) ⁄

( ) ( ) ⁄

( ) ( ) ⁄

( ) ( ) ⁄

Total production cost during the stage ( )

 Some combinations are infeasible. For example, the 9th combination; [ ] is infeasible to
meet the load demand of 1400 MW since this demand cannot be served even after loading unit 1 and
unit 2 to their maximum capacities. Such combinations are ignored.
 Some units may violate their capacity limits during the economic dispatch solution. Thus, the output
of those units must be set at the violating limit. For example, consider the 2 nd combination;
[ ] in meeting the load demand of 1800 MW.

(∑ )

(∑ ) (∑ )

Unit 2 output violates the upper limit. Therefore, set the unit 2 output as and
distribute the remaining load of ( ) between unit 1 and unit 3.

Page 4 of 7
 (∑ ) ( )

(∑ ) (∑ ) ( )

Unit 1 output violates the upper limit. Therefore, set the unit 1 output as .
Unit 3 output

Perform economic dispatch analysis for each loading level as shown in the table below:

( )
( ) ( ) ( ) ( )
( ⁄ ) ( ⁄ ) ( ⁄ ) ( ⁄ ) ( ⁄ )

10.0898 261 384 219 235 2861 3565 2570 2466 45848
10.8051 351 459 290 - 3801 4349 3312 - 45848
10.7740 347 456 - 298 3758 4317 - 3123 44792
12.0727 509 591 - - 5608 5859 - - 45868

10.804 351 459 290 300 3801 4349 3312 3145 58425
11.7165 464 554 382 - 5073 5419 4347 - 59356
11.7112 464 553 - 383 5073 5407 - 4079 58236
Infeasible Combination

11.2801 410 508 338 344 4452 4890 3841 3631 67259
12.3241 541 617 442 - 5999 6176 5069 - 68976
12.3361 542 618 - 440 6011 6188 - 4765 67856
Infeasible Combination

11.7562 469 558 386 387 5132 5466 4394 4126 76472
12.9316 625 625 550 - 7063 6275 6458 - 79184
Infeasible Combination
Infeasible Combination

Page 5 of 7
 [$314,360] [$183,576]
[$314,812] [$258,316] [$184,696]
[$312,912] [$254,484] [$183,768] [$107,296]
[1 1 1 1] x1 x1 [$0] x1 [$0] x1 [$0] x1 x1
$45,848 $58,428 $70,908 $76,472 $58,428 $45,848
[$3000] [$3000]
[$1500] [$1500]
[$318,288] [$189,288]
[$314,240] [$1500] [$254,884] [$3000] [$185,908] [$1500]
[$316,840] [$255,552] [$189,480] [$106,724]
[1 1 1 0] x2 x2 [$0] x2 [$0] x2 [$0] x2 x2
$45,848 $59,356 $68,976 $79,184 $59,356 $45,848
[$4500] [$1500]
[$6000] [$4500] [$4500] [$4500] [$3000]
[$312,668] [$3000]
[$317,620] [$3000] [$258,264]
[$315,720] [$254,432] [$105,604] [$1500]
[1 1 0 1] x3 [$3000] x3 [$0] x3 x3 x3
$44,792 $58,236 $67,856 $58,236 $44,792

[$361,536]
[$363,108] [$3000] [$1500]
[$364,780]
[1 1 0 0] X
x49 X
x49
$45,868 $45,868

Stage 1 Stage 2 Stage 3 Stage 4 Stage 5 Stage 6

The least overall cumulative cost from stage 1 to stage 6

Page 6 of 7
The optimal unit commitment schedule is given below:

Stage Load Level (MW) Combination Units

1 1100 Unit 1, Unit 2
2 1400 Unit 1, Unit 2, Unit 4
3 1600 Unit 1, Unit 2, Unit 4
4 1800 Unit 1, Unit 2, Unit 3, Unit 4
5 1400 Unit 1, Unit 2, Unit 4
6 1100 Unit 1, Unit 2

Page 7 of 7