两种优化训练的FNN在船舶柴油机故障诊断中的性能研究

本文针对船舶柴油机故障诊断系统,基于遗传算法（genetic algorithm,GA）和蚁群优化算法（ant colony optimization algorithm,ACOA）构造了2种优化训练的模糊神经网络（fuzzy neural network,FNN）智能故障诊断模式,给出了该模糊神经网络智能故障诊断系统的结构及其参数选取方法,通过对船舶柴油机燃烧子系统的FNN模型结构权值和阈值优化训练的故障诊断仿真研究,对两种方式的性能进行对比研究,仿真测试结果表明,基于ACOA的诊断模型具有更好的故障诊断知识表达准确性和较快的收敛速度等特点,具有较好的应用前景。

Performance Research of Fault Diagnosis Using FNN Based on Two Kinds Optimized Methods for Marine Diesel Engine

Two kinds of intelligent fault diagnosis based on fuzzy neural network optimized and trained by the genetic algorithm （GA） and ant colony optimization algorithm （ACOA） were proposed and constructed in this paper. The structure and the parameters of intelligent fault diagnosis system for the two kinds of fuzzy neural network were introduced. The fuzzy neural network systems optimized and trained by GA and ACOA were applied in the fault diagnosis of the marine diesel engine＇s combustion subsystem. By simulation that has been carried out to evaluate the performance of proposed method and to compare their performance with conventional FNN fault diagnosis method, the simulation results the knowledge expression and the precision of fault diagnosis can be improved and have good quick convergence performance using fault diagnosis base on ACOA. This intelligent fault diagnosis method has the good application prospects in other similar system.