A New Model for Load Tracker for an Electrochemical Power System with DC-DC Converter based on Artificial Neural Networks and Fuzzy Logic

Abstract

    This paper presents a new methodology to improve the performance efficiency of an electrochemical power system type of a Proton Exchange Membrane Fuel Cell (PEMFC), using a new tracer model based on the use of Artificial Neural Networks (ANN), and Fuzzy logic (FL), to track load changes. The use of the proposed Neuro-Fuzzy Tracker aims to achieve the optimal operation of the power system, by providing the load with the required electrical energy with high accuracy and at a high speed that is compatible with the speed of changes in the load, while achieving the optimal consumption of fuel used in the electrochemical energy conversion processes, in addition to improving the value of the electrochemical energy conversion efficiency factor. In this context, the proposed tracker model in the research consists of two models, the first neuron and the other Fuzzy logic model, working in sequence. The proposed neural model estimates the optimum operating current to be extracted from the PEMFC system to match the instantaneous load requirements. The proposed neural model estimates the optimum operating current to be extracted from the PEMFC system to match the instantaneous load requirements. While the Proposed Fuzzy logic Controller (PFLC) is working, it determines the appropriate variations of the operating ratio used to adjust the duty cycle of the DC-DC voltage  Converter o track the load. The simulation results for PEMFC load tracking performed in Matlab/Simulink environment showed that the proposed Neuro-Fuzzy Tracker (ANN-PFLC) achieves optimal performance of the system, compared with using other reference models and compared with the direct coupling condition without tracker and without a voltage converter.