摘要
构建了并行测试系统(PATS)故障诊断效能评价指标体系,针对该指标体系在评价过程中存在的不确定因素,结合神经网络和模糊理论的优点,提出了基于神经网络的模糊综合评价方法 (FCEANN);详细给出了FCEANN的求解算法及步骤,结合实例进行仿真实验并与专家评价结果进行比较,对影响模型预测精度的因素进行了分析;仿真结果表明,基于BP神经网络和Elman神经网络的FCEANN均能够很好地模拟专家评价的全过程,能够准确地对PATS的故障诊断效能指标进行评价;并将两种神经网络的仿真结果进行了比较,结果表明,Elman神经网络得到的仿真结果精度更高。
Fault diagnosis efficiency evaluation index system for Parallel automatic test system was structured. Compounding the advantages of neural network and fuzzy logic, Fuzzy comprehensive evaluation approach based on neural network (FCEANN) was proposed for better evaluating the uncertain factors which exist in the fault diagnosis efficiency evaluation of parallel automatic test system. The solving algorithm and process of FCEANN was presented. The factors affecting the prediction accuracy were analyzed based on the comparison be tween the simulation result and the expert prediction result. The simulation results show that the FCEANN based BP neural network or Elman neural network can simulate the whole process of expert assessment efficiently, and can evaluate fault diagnosis efficiency index accurately, and the evaluation approach based on Elman neural network is more accurate.
出处
《计算机测量与控制》
CSCD
北大核心
2011年第12期2934-2938,共5页
Computer Measurement &Control
关键词
并行测试系统
神经网络
模糊综合评价方法
故障诊断效能
parallel automatic test system
neural network
fuzzy comprehensive evaluation approach
fault diagnosis efficiency