模糊神经网络的动力电池故障诊断系统

Power battery fault diagnosis system based on fuzzy neural network

动力电池的故障症状和故障原因之间存在一定的不确定性和模糊性,设计了一种基于模糊神经网络的动力电池故障诊断系统。该系统结合了模糊逻辑和神经网络的优点,模糊逻辑能对故障症状与故障原因之间的模糊关系进行准确描述,对模糊性信息具有较强的表达能力,而神经网络则具有强大的学习和获取知识的能力,能够避免模糊逻辑依赖于模糊规则的局限性,能够对信息进行并行推理,将两者结合,可以提供一个更加精确、有效的电池故障诊断系统。为了对设计的故障诊断系统进行验证,通过MATLAB自带电池模型的参数设置进行电池故障的模拟,然后,利用设计的故障诊断系统进行仿真测试,结果表明,该系统能正确诊断出电池故障的原因并给出各种电池故障的严重程度,因此,本系统可以应用于动力电池管理系统进行电池故障的准确诊断和故障原因的分析与报警,提高动力电池使用的安全性。

There exists uncertainty and fuzziness between the fault symptomand fault cause of the power battery. The battery fault diagnosis systembased on fuzzy neural network was designed. The system combines the advantages of fuzzy logic and neural networks, as fuzzy logic can accurately describe the fuzzy relation between fault symptoms and fault cause and has strong ability on fuzzy information expression. Neural network has great ability to learn and acquire knowledge, avoiding the fuzzy logic limitation that dependent on the fuzzy rules, and can carry on the parallel reasoning to the information. Combination of these can provide a more accurate and efficient battery fault diagnosis system. To verify the fault diagnosis system of the design, the battery fault simulation was carried out by using the parameters of the battery model in MATLAB. Then, the fault diagnosis system was used for the simulation test. Simulation test results show that the system can correctly diagnose the cause of battery fault and the severity of all kinds of battery faults. Therefore, the system can be applied to the accurate battery fault diagnosis and analysis and alarmof the battery fault cause in power battery management system, so as to improve the safety of power battery.