基于概率神经网络的船用柴油机共轨系统限流阀故障诊断研究  

Fault Diagnosis for Fuel Limiting Valve of Marine Diesel Engine with Common Rail System Based on Probabilistic Neural Network

作　　者：金江善[1,2] 董晓露 秦慈伟 桑晓琳 JIN Jiangshan;DONG Xiaolu;QIN Ciwei;SANG Xiaolin(Shanghai Marine Diesel Engine Research Institute,Shanghai 201108, China;National Engineering Laboratory for Marine and Ocean Engineering Power System,Shanghai 201108,China)

机构地区：[1]中国船舶重工集团公司第七一一研究所,上海201108 [2]船舶与海洋工程动力系统国家工程实验室,上海201108

出　　处：《内燃机工程》2018年第3期81-87,共7页Chinese Internal Combustion Engine Engineering

摘　　要：基于船用柴油机高压共轨系统工作原理分析限流阀失效模式,利用AMESim液力仿真平台搭建共轨系统模型,通过仿真分析得出限流阀在3种典型故障模式下的故障特征。建立概率神经网络对限流阀故障进行诊断,结合试验与仿真数据完成网络训练和验证。结果表明:采用概率神经网络可有效地实现限流阀故障诊断。The failure mode of the fuel limiting valve was analyzed based on the working principle of the highpressure common rail system（HPCRS）of marine diesel engine.A simulation model of HPCRS was built with the AMESim hydraulic simulation platform,and the characteristics of three typical fault modes of the fuel limiting valve was obtained by simulation.A probabilistic neural network was established to diagnose failure of the valve,and test and simulation data were used to complete the training and verification of the network.The results show that the probabilistic neural network can realize the fault diagnosis of the fuel limiting valve effectively.

关 键 词：船用柴油机 共轨系统 限流阀 故障诊断 概率神经网络 

分 类 号：TK421[动力工程及工程热物理—动力机械及工程]