A New FNN Model and Its Application in Hydraulic Pump Trouble Diagnosis

摘要

The troubles in a hydraulic system are usually caused by certain source(s). When a trouble source is initiated, related symptoms spread through the possible paths. There are two trouble diagnosis methods in hydraulic systems: one is to locate the troubles (source) per symptoms of hydraulic systems, and another one is to detect the troubles per trouble sources which lead to hydraulics systems damage. In this paper, a new FNN diagnosis model for hydraulics troubles has been established. In the new model, at first the previous two diagnosis methods are conducted, and then the further analysis is performed over the conclusions obtained from these methods. This new diagnosis method is concise, reliable, and easy to be used in workshop. Meanwhile this FNN model has self-learning function, and can display the knowledge structure used for diagnosis process. The advantages of this method have been proved in the trouble diagnosis case of PV250DF axial piston pump in a steel company.