S c i e n t i f i c a n d T e c h n i c a l C o n f e r e n c e
T I T L E : S c h e d u l i n g t h e P u m p O p e r a t i o n s i n W a t e r S u p p l y S y s t e m s U s i n g O p t i m i z a t i o n A l g o r i t h m s
Platinum Sponsor Silver Sponsor
W A T E R S E R V I C E S A N D T H E N E W E N E R G Y C H A L L E N G E S
AUTHOR: Alexandrescu Simona Adina (Miron), Luca Mihail, Alexandrescu Aurora, Mănescu Andreea
PUMP’S OPERATING COSTS
OPERATING
COSTS
ELECTRIC
ENERGY COSTS
MEINTENANC
E COSTS
OBJECTIVE FUNCTIONS
ELECTRIC
ENERGY COSTS
NUMBER OF
PUMP SWITCHES
OBJECTIVE
FUNCTIONS
TOTAL
VOLUME
DEFICIT = 0
PRESSURE FROM
DEMAND NODES
> 0
CONSTRAINTS
SPEA2 ALGORITHM DESCRIPTION
PARAMETER DEFINITION
CREATE THE INITIAL POPULATION
1. Random
2. Use a previous solution
EVALUATE THE
OBJECTIVE
FUNCTIONS
CALCULATE THE FITNESS
FUNCTIONS
1. Calculate strength value S(i)
2. Calculate raw fitness R(i)
3. Calculate the density
4. Calculate the fitness function
REPRODUCTION
1. Recombination
2. Mutation
SELECTION
from archive through
binary tournament method
ARCHIVE
the higher S(i) is, the
better solution i is
the higher R(i) is, the
worse solution i is
CONSTRAINT HANDLING
Invalid solutions are dominated by the valid ones;
When two solutions are invalid, the one with lower
pressure violations dominates the other;
For two solutions, the one with a lower total
volume deficit dominates the other;
If two solutions are feasible and have an equal total
volume deficit, the dominance criteria imposed by
the objective functions is applied.
Input:- text file which describes the regional water supply system; - initial population size;- maximum population size; - maximum archive size;- maximum number of generations varies from an execution to another;- number of parents; - mutation probability; - recombination probability;- individual mutation probability;- Individual recombination probability.
Matlab Algorithm:
EPANET_SPEA.m
SPEA2.m
F1.m
F2.m
Output:- Pump schedules presented as
binary values that correspond to
each time window.
Main elements of the algorithm that determines the
“near optimum” schedule of pumps
DESCRIPTION OF THE ALGORITHM
MODELING THE ANALYSED WATER SUPPLY SYSTEM
Timisesti – Iasi adductions modeled in EPANET
Pacurari and Aurora systems
No. of
execution Duration
Electric
energy
cost
(RON)
Number of
pump
switches
No. of
pressure
condition
breaking
EPANET
errors
Total
volume
deficit
1 5H -8M 3S 5570 60 0 0 -20
2 3H 11M 6S 5189 63 0 0 65
3 3H 11M 6S 5666 70 0 0 22
4 3H 58M 8S 5795 59 0 0 22
5 3H 58M 8S 6363 62 0 0 14
6 4H -25M -30S 5662 54 0 0 99
7 5H -17M 10S 5511 57 0 0 43
8 5H -8M 3S 5570 60 0 0 60
9 6H -12M -5S 5357 57 0 0 46
10 6H -8M 37S 5357 57 0 0 46
11 6H -8M 37S 5744 61 0 0 60
12 6H 11M 35S 6017 56 0 0 2
13 8H -27M 52S 5777 58 0 0 -30
14 8H -27M 52S 6146 68 0 0 1
15 8H -27M 52S 5589 57 0 0 1
16 6H -3M 21S 5605 56 0 0 0
17 6H -3M 21S 5533 54 0 0 3
18 6H -3M 21S 5565 55 0 0 -2
19 7H 1M -17S 5489 54 0 0 0
20 7H -30M -19S 5428 56 0 0 -7
21 7H 9M 6S 5489 54 0 0 0
22 8H -23M 21S 5421 55 0 0 0
23 8H -23M 21S 5420 54 0 0 1
24 8H -23M 21S 5404 54 0 0 -1
25 7H -48M 0S 5403 54 0 0 0
26 7H -48M 0S 5276 55 0 0 -1
27 6H 38M -15S 5253 55 0 0 -1
28 7H 8M 6S 5118 53 0 0 -1
29 7H 13M 5S 5118 54 0 0 0
Results obtained with SPEA2 program for Pacurari pumping system
RESULTS
EPANET results obtained with the pump controls
given by solution number 29
EPANET energy report resulted from functioning of the pumps
controlled by water level from Aurora tanks
Orarele de funcţionare a pompelor: roşu - pumpPac1-1,
albastru - pumpPac1-2, negru – pumpPac2-1, magenta –
pumpPac2-2, verde – pumpPac2-3
CONCLUSIONS
Time controlled pumps with schedules optimized by SPEA2 algorithm have an
operational cost lower than the one obtained by water level controlled pumps.
In this paper the best pump schedule given by SPEA2 algorithm generates an
operational cost lower with 34 % than the one generated by the water level
controlled pumps.
In order to find good solutions for complex problems like the one presented in
this paper, it is necessary to use either a large number of generations.
The method presented in this paper is a cheap and efficient way to lower the
operational costs.
An important disadvantage of the presented method is that pumps can be turned
on or off only at the start of each hour.
