Volume No. 6 Issue No.: 3A Page No.: 853-862 Jan-Mar 12

 

LEAST COST DESIGN OF WATER DISTRIBUTION NETWORK UNDER DEMAND UNCERTAINTY BY FUZZY - CROSS ENTROPY METHOD

 

Shibu A.* and M. Janga Reddy

Indian Institute of Technology Bombay, Mumbai (INDIA)

 

Received on : November 25, 2011

 

ABSTRACT

 

Water distribution system modeling problems are associated with large number of variables which are uncertain in nature. The uncertainties are due to (i) formula used (ii) coefficients in the formula and (iii) imprecise knowledge of the values of various parameters. Usually uncertainty exists in nodal demands and the uncertainty associated with nodal demands has to be considered for better design of water distribution networks. A two phase methodology for the least cost design of water distribution network based on fuzzy set theory and cross entropy method is proposed. The uncertain demands are considered as fuzzy sets, and the diameters for each pipe is selected from the commercially available diameters by cross entropy method. The model coded in MATLAB is linked to EPANET tool kit for hydraulic simulation. The proposed methodology was tested on Hanoi water distribution network, and the solutions obtained are compared with well known deterministic solutions from literature. The methodology is found to be effective in dealing with uncertainty in input parameters represented as fuzzy sets, and also the discrete diameters are very well handled by cross entropy method.

 

Keywords : Water distribution network, Fuzzy sets, Cross entropy, Optimization, Hydraulic Simulation

 

 

BACK