10-03-2012, 04:34 PM
Economic dispatch using fuzzy logic
Abstract
A Fuzzy logic based solution for emission constraint economic dispatch is demonstrated in this
work. The aim of this work is to minimize the fuel cost of operation of an interconnected power system
with emission constraints. This non-linear constrained multi-object optimization solution has been
solved by other investigator using AI computation methods. These solutions though provided a global
optimal value, are not suitable for computer implementation, due to computational complexity and large
computation time.
Fuzzy logic has served a black box modeling tool for engineers in application where no proper
mathematical models are available. In the first phase of the work various fuzzy based models are
investigated to see how they can model the nonlinearities in the generator cost equation. On intensive
testing it was found that the fuzzy based model with two inputs the error and change in error was able to
attain the optimal solution with highly reduced computational time, Compared to methods like simulated
annealing. Five membership functions were defined for error and change in error which resulted in a
rule a rule base with twenty five rules. The error and change in error were represented with various
possible membership function among them the triangular membership was chosen by trial and error
method. The defuzzification method chosen was centriod method.
Results for a three generator test system have been presented to validate the proposed method. Hence
this method may be used in place where a combined ECEDP is necessary.
Abstract
A Fuzzy logic based solution for emission constraint economic dispatch is demonstrated in this
work. The aim of this work is to minimize the fuel cost of operation of an interconnected power system
with emission constraints. This non-linear constrained multi-object optimization solution has been
solved by other investigator using AI computation methods. These solutions though provided a global
optimal value, are not suitable for computer implementation, due to computational complexity and large
computation time.
Fuzzy logic has served a black box modeling tool for engineers in application where no proper
mathematical models are available. In the first phase of the work various fuzzy based models are
investigated to see how they can model the nonlinearities in the generator cost equation. On intensive
testing it was found that the fuzzy based model with two inputs the error and change in error was able to
attain the optimal solution with highly reduced computational time, Compared to methods like simulated
annealing. Five membership functions were defined for error and change in error which resulted in a
rule a rule base with twenty five rules. The error and change in error were represented with various
possible membership function among them the triangular membership was chosen by trial and error
method. The defuzzification method chosen was centriod method.
Results for a three generator test system have been presented to validate the proposed method. Hence
this method may be used in place where a combined ECEDP is necessary.