基于深度学习的高压油泵可靠性试验故障分析方法研究

介绍了船用柴油机高压油泵可靠性试验平台的系统架构,开展了对试验数据的清洗处理和基于机器学习的试验典型故障的研究。利用TensorFlow平台对试验中两种典型故障进行原因分析和模式识别,建立高压油泵可靠性试验故障诊断模型。结合柴油机高压油泵可靠性试验中大量反复出现的故障数据进行训练,大大提高了故障诊断模型的准确度,为高压油泵生产研制部门提供应急故障处理方案提供了依据,为实现船舶动力行业智能诊断提供了新的思路。

Research on Fault Analysis Method of High Pressure Oil Pump Reliability Test Based on Deep Learning

This paper introduces the system architecture of the reliability test platform for high-pressure oil pump,and the cleaning and processing of test data and the study of typical faults based on machine learning.The cause analysis and pattern recognition of two typical faults in the test are carried out by using TensorFlow platform,and the fault diagnosis model of high-pressure oil pump reliability test is established.The accuracy of fault diagnosis model is greatly improved by combining a large number of repeated fault data from reliability test of high-pressure oil pump of diesel engine,which provides a basis for emergency fault treatment plan for high-pressure oil pump production and development department,and lays a foundation for intelligent diagnosis of ship industry.