A framework for integrated system of fault diagnosis in oil equipments based on neural networks

When the traditional expert system is used for the fault diagnosis in oil equipments, there are some problems, such as difficult knowledge acquisition, low inference efficiency, poor adaptability. Therefore, it is proposed that neural networks are combined with the expert system for fault diagnosis. This paper presents the development of a framework for integrated system of fault diagnosis in oil equipments based on neural networks. The framework employs a combination of technologies, including dynamic database, comprehensive knowledge base and neural networks. This paper describes how to represent fault diagnosis knowledge using the neural networks, and discusses design process of the inference engine based on fuzzy neural networks. The results demonstrate that the accuracy is higher using the proposed system for fault diagnosis in oil equipments, and it can meet real-time requirements of maintenance, so this system outperforms the traditional system.