The Study On Artificial Intelligent Theory And Technology With The Application On Fault Diagnosis Of Large Centrifugal Compressors

Artificial intelligence(AI) is one of the frontiers of computer science. Expert systems(ES) and artificial neural network(ANN) are two main approaches of current AI research. Focusing on the fault diagnosis of a large centrifugal compressor, which is a widely used equipment in most of chemical and petrochemical plants, this paper is aimed to explore ways of solving some urgent theoretical problems in AI and to establish a knowledge—based fault diagnostic system for such equipment.In the first, a review of existing methods in AI research is presented. It concludes that AI is a promising and superior method to the solution of problems in the field of fault diagnosis.The two key aspects in AI research are knowledge acquisition and knowledge processing. Knowledge acquisition involves knowledge extration and knowledge representation, both of which are problem-dependant. Based on a System Hierarchy Method introduced in this paper and a full functional and structural analysis of large centrifugal compressors, the machanism, features and recognition methods for common faults are summarized. Knowledge representation plays an important role in building a knowledge base in a computer, in this paper, the production rules are selected and found to be best suited for representing fault diagnostic knowledge, most of which is logical.