| With the substantial of energy consumption in the modern society, the traditional sources of energy is less and less unable to meet the growing demand for human consumption, so the Fuel Cell as a new energy which has unique advantages of energy saving and environmental protection is researched by more and more countries.PEMFC is made of many sections which are connected in series, each piece of battery will influence the performance and security of the fuel cel stack, at the same time, detection of fuel cell stack work status and find the best working conditions to prolong its service life is particularly important. Therefore the need for each monomer fuel cell output voltage to be monitored in time. This paper takes the pieces voltage measure of fuel cell stack as research object. Develop the design of the voltage monitor system based on the optical relay of PEMFC pieces.The main research contents as follows:According to the function needs of the system and the special applied environment, this paper has designed a structure of monitor system based on the optical relay, using the optical relay to achieve different battery switch, to solve the traditional voltage acquisition of accumulative electric potential problem.At the same time the design of potential lifting unit effectively solves the problem of the negative voltage acquisition.On the basis of analysis of PEMFC polarization curve and the performance index of the system, this paper designs the master-slave single chip voltage monitor system hardware, including the gating unit, high precision signal conditioning unit, negative voltage measuring unit and the USB hardware interface circuit. This paper also analyses their structures and characters detailed.This paper designs the next-bit machine detection unit software and the main control unit software. According to the inspection system for vehicle application, this paper makes a strict standardized communication application layer protocol, developes CAN and USB application layer protocol in detail, and gives the corresponding protocol format. This paper gives a method based on support vector machine to diagnosis fault single cell. It uses SVM classifier function on fuel cell fault recognition, based on the maximum distance principle optimization of hierarchical SVM model, so that each node of the support vector machine has the maximal margin classification, optimizes the multilevel binary tree structure of SVM, realizes fault diagnosis and recognition of single cell.The whole monitor system has master-slave mode of detection, with high reliability, scalability, has realized the fuel cell stack voltage real time high precision detection. This system has applied in New Technology State Key Laboratory of50KW fuel battery test platform in the Wuhan University of Technology, and it has got a good performance. |