基于PSO与BP神经网络的磁共振成像设备故障诊断研究  

Research on fault diagnosis of magnetic resonance imaging equipment based on PSO and BP neural network

作　　者：方佩玺 张姚昕 赵媛[1] Fang Peixi;Zhang Yaoxin;Zhao Yuan(Xinhua Hospital affiliated to School of Medicine,Shanghai Jiaotong University,Shanghai,200092,China)

机构地区：[1]上海交通大学医学院附属新华医院,上海200092

出　　处：《机械设计与制造工程》2025年第1期85-90,共6页Machine Design and Manufacturing Engineering

摘　　要：针对磁共振成像设备故障诊断准确性和效率低的问题,提出一种基于粒子群优化算法与反向传播神经网络结合邓普斯特-谢弗证据理论的故障诊断模型。该模型通过粒子群优化算法优化反向传播神经网络的参数,并结合邓普斯特-谢弗证据理论融合多传感器数据。实验结果表明,10种故障类型下所提模型的故障检测正确率为100%,对10种不同类型故障的平均检测准确率达96.2%,单样本检测耗时为17.5 ms。Aiming at the low accuracy and efficiency in fault diagnosis of magnetic resonance imaging equipment,a fault diagnosis method based on particle swarm optimization algorithm and back propagation neural network combined with Dempster-Schafer evidence theory is proposed.This method optimizes the parameters of back propagation neural network by particle swarm algorithm and fuses multi-sensor data by combining Dempster-Schafer evidence theory.Experimental results show that the average detection accuracy of the proposed model for 10 types of faults is 96.2%,the single sample detection time is 17.5 ms,and the accuracy rate reaches 100%in the detection of 10 types of faults.

关 键 词：粒子群优化算法 反向传播神经网络 磁共振成像设备 故障诊断 邓普斯特-谢弗证据理论 

分 类 号：TH772.2[机械工程—仪器科学与技术]