基于数据驱动的船用冷藏集装箱智能故障诊断方法研究  

Research on the Automated Fault Detection and Diagnosis in Marine Refrigerated Container Based on Data-driven

作　　者：郑超瑜[1,2] 俞文胜[1,2] 陈武[1,2] 

机构地区：[1]集美大学轮机工程学院,福建厦门361021 [2]福建省船舶与海洋工程重点实验室,福建厦门361021

出　　处：《船舶工程》2015年第12期60-64,共5页Ship Engineering

基　　金：福建省自然科学基金项目(2013J01201);福建省科技厅项目(201205180008);福建省教育厅科技项目(JA12187);厦门市科技计划项目(3502Z20123024)

摘　　要：采用数据驱动方法,避开船用冷藏集装箱复杂的控制策略,建立系统或部件模型用于智能故障诊断。根据正常运行状态的实验数据"训练"冷凝器模型,利用该模型预测冷凝器的运行状态,并与实际的故障数据进行比较。结果表明,数据驱动模型具有较高的预测精度,能够分辨故障状态,实现智能故障报警。The use of data-driven method can avoid complex control strategy of marine refrigerated container so as to establish system or component model for automated fault detection and diagnosis. According to the experimental data under normal operation condition, the condenser model of refrigerated container is trained to predict operation parameters, and the predicted data is compared with faulty data. The results show that the data-driven model has good prediction accuracy, which can identify fault state and realize automated fault alarm.

关 键 词：数据驱动 船用冷藏集装箱 智能故障诊断 

分 类 号：U664.87[交通运输工程—船舶及航道工程] TB61[交通运输工程—船舶与海洋工程]