Neuro-genetic energy management for hybrid fuel cell power train

This paper deals a new strategy developed for optimizing the energy flow by using evolutionary algorithms implemented on a hybrid fuel cell vehicle (HFCV) to reach the best performance, fuel economy, emission and acceptable operation of this hybrid structure. This paper investigates the applicability of the Evolutionary based algorithms for hybrid system optimization problems. Artificial neural networks (ANN) are a computational paradigm modeled on the human brain that has become popular in recent years, especially in engineering problems. Genetic algorithms (GA) are a class of search algorithms modeled on the process of natural evolution. In this paper the modeling phase is done using ANN. The sub models of system including battery and fuel cell has been replaced fully or partially with proper ANN. The GA is used for optimization phase and also for ANN weight and threshold selection. With respect to dynamic behavior of this optimization problem, the system is simulated to demonstrate the validity and the convenience of evolutionary approach. Hence an object oriented programming (OOP) tool is developed for simulation of this hybrid structure. It prepares a good environment for supervisory control of stack as a major part of HFCV. The simulation results confirm the feasibility and encourage more research towards an actual application