神经网络与D-S证据理论分层融合的柴油机综合故障诊断方法研究

Diesel Engine Synthesized Fault Diagnosis Technique Based on Hierarchically Fusing Neural Network and D-S Evidence Theory

针对柴油机传统故障诊断方法处理数据量大、故障类型复杂多变的问题时存在诊断准确率不高的现状,利用数据融合原理,将神经网络和证据理论进行有机的结合,提出了神经网络和证据理论分层融合的柴油机故障综合诊断方法.该方法通过并行神经网络的结构提高局部诊断网络的诊断能力,并给出了基本可信度分配的客观化方法,充分利用各种故障的冗余和互补信息,可显著提高故障诊断的准确率.诊断实例表明,该方法能显著提高柴油机故障诊断系统的效率.

For the reasons of low diagnosis accuracy of traditional diesel engine fault diagnosis methods in handling diagnostic problems such as lots of data and various complex faults,a diesel engine synthesized fault diagnosis technique fusing neural network and evidence theory is presented by means of data fusion theory.In this technique,the diagnosis ability of the local diagnosis networks is advanced through parallel neural network structure,and an impersonal means obtaining basic reliability distribution of evidence theory is given,and then the accuracy of the fault diagnosis is improved obviously by taking full advantages of various redundant and complementary fault information.Finally an example is applied for fault diagnosis of ship diesel engine,and diagnostic results indicate that the technique is available,which can improve the efficiency of diesel engine fault diagnosis system evidently.