神经网络和证据理论融合的故障诊断方法研究

针对传统故障诊断方法存在的诊断准确率不高的问题，利用数据融合原理，将神经网络和证据理论进行有机结合，使两者优势互补，提出了神经网络和证据理论融合的故障诊断方法。通过简化网络结构提高了局部诊断网络的诊断能力，并使证据理论的基本可信度分配不再完全依赖专家进行的主观化赋值，从而实现了赋值的客观化。充分利用各种故障的冗余和互补信息，显著提高了故障诊断的准确率。在材料试验机伺服系统的液压泵上进行了模拟故障诊断试验，验证了该方法的有效性。

On Fault Diagnosis Method Fusing ANN and Evidence Theory

For the reasons of low fault diagnosis accuracy of traditional diagnosis methods, a fault diagnosis method fusing ANN and evidence theory was presented by means of data fusion theory. This method combined two data fusion methods-ANN and evidence theory-by using their superiority and avoiding their disadvantages. Through simplifying network structure, the diagnosis ability of the local diagnosis networks was advanced. Through making the basic reliability distribution of the evidence theory not completely depending on the expert subjective valuations, the impersonal valuations were realized. Through taking full advantages of various redundants and complementary fault informations,the accuracy of the fault diagnosis is improved evidently. This method is applied to fault diagnosis of axial piston pump of hydraulic servo system of material experimental machine. Through simulating the shoe-doffing fauls of this pump, the validity of this method is verified.