遗传神经网络故障诊断自适应Petri网模型

摘要

研究故障诊断优化问题.针对传统Petri网难以精确地描述故障现象和故障原因之间的复杂关系,造成故障诊断难以精确,提出了将遗传算法、神经网络和传统Petri网模型结合,形成了一种改进的自适应的加权Petri网模型以及模型的构造算法,同时在此基础上,采用改进的遗传算法对神经网络模型的权值进行优化训练,并给出了采用构造的自适应模糊Petri网模型对故障进行诊断的具体步骤.仿真实例验证了算法的有效性,对柔性制造系统实例的故障进行诊断,验证了此自适应的加权模糊Petri网模型结合了Petri网和遗传算法的优点,具有很强的故障推理能力以及自适应能力,能有效地对故障进行诊断.%Fault diagnosis optimization problems were studied. Traditional Petri nets are difficult to accurately de scribe the symptoms and the complex relationship between the cause of the malfunction, resulting in that accurate di agnosis is difficult. The paper put forward an improved adaptive weighted Mohu Petri net model which combined ge netic algorithm, fuzzy logic and traditional Petri net model. On this basis, an improved genetic algorithm model was used to optimize the weight training, and the specific steps of fuzzy fault diagnosis was given based on the structural a daptive Petri net model. Simulation results verified the effectiveness of the algorithm, and proved that the adaptive weighted fuzzy Petri net model combines the advantages of network and genetic algorithm, and has a strong fault rea soning ability and adaptive ability.