基于FOA优化GRNN的船用柴油机涡轮增压系统故障诊断

Fault diagnosis for Turbocharging System of Marine Diesel Engine Based on GRNN Optimized by FOA

摘要:

为了及时有效的发现并排除船用柴油机涡轮增压系统的故障，文中采用果蝇优化算法（FOA）对广义回归神经网络（GRNN）的分布密度SPREAD进行优化选取，提出了一种果蝇优化算法和广义回归神经网络相结合的故障诊断新方法。收集某型号船用柴油机的样本集，采用相同的训练样本分别对FOA优化GRNN和RBF神经网络进行训练，并用相同的测试样本对以上两种模型进行验证。结果表明，与RBF神经网络故障诊断方法相比，FOA优化GRNN对柴油机涡轮增压系统故障模式的识别准确率更高。

Abstract:

In order to timely and effectively detect and eliminate the fault of marine diesel turbocharging system, the distribution density SPREAD of generalized regression neural network (GRNN) was optimized and selected by fruit fly optimization algorithm (FOA) in this paper.A new method of fault diagnosis based on fly optimization algorithm and generalized regression neural network is proposed. Sample sets of a marine diesel engine were Collected.The same training samples were used to train GRNN optimized by FOA and RBF neural network respectively.The above two models were verified by the same test samples.The results show that compared with the fault diagnosis method of RBF neural network , the method of GRNN optimized by FOA is more accurate to identify the failure modes of the diesel engine turbocharging system.