Based On Neural Network Fault Diagnosis Of High-pressure Ammonia Pump System

This paper first introduced the purpose of fault diagnosis and significance, as well as the need for fault diagnosis. Research on fault diagnosis expert system of meaning and purpose. Introduced the current domestic and international fault diagnosis expert system development. Maintenance system and fault diagnosis of technological development. Expert system introduces the basic knowledge, including the definition of expert systems, general structure, basic features, structural characteristics, fault classification. Introduced the basic knowledge of neural networks, development in the law to avoid the application of fault diagnosis.In this paper, in-depth exploration into the characteristics of high-pressure ammonia pump failure on the basis of studies of the neural network diagnostic expert system in the high-pressure ammonia pump fault diagnosis system applications. A competitive neural network with a combination of BP network, neural network and expert system combining ideas. Based on this idea, the basic network layer corresponding to different signs of the formation of neural networks, basic network connectivity stratified high-pressure ammonia pump fault diagnosis expert system for the initial model.In this paper, a neural network architecture based on a neural network to explore further in the knowledge that the object-oriented thinking. Combination of neural networks in the text to try to resolve this problem. Proved through experiments, based on the idea of this article can only be achieved to some extent, this idea, in future trials is expected to further improve the formation of neural networks. For example, you can try to vary joint neural network (BAM) to show that the realization of the transfer of property space. The establishment of inter-attribute reasoning mechanism.