Power Plant Deaerator System Based On Granular Computing SDG Fault Diagnosis

The SDG model is an important branch of the qualitative model, can express complex causal relationship between the systems, contain mass underlying information, used for fault diagnosis, not only good maturity, strong adaptability, good robustness, and can provide fault propagation path and evolution of interpretation. In recent years, SDG fault diagnosis method has been widely concerned, scholars have developed rapidly, and has achieved good application effect in petrochemical industry. But ambiguous reasoning based on qualitative method inevitably has fault diagnosis method of SDG model in common cause the shortcoming of low resolution, and the problems of model and calculation of large system depends too much on the diagnosis speed slow, poor real-time performance.Granular computing research problem from different granular layer, is mainly used to deal with uncertain, fuzzy, incomplete information, is an effective tool for solving complex problems of fuzzy information processing, data mining, the. Knowledge reduction of granular computing based on guaranteeing the classification ability unchanged, remove redundant attributes, and derive the minimal rules, can the redundant nodes complex system reduction, improve the speed and real-time of fault diagnosis.In this paper, the theory of granular computing in rule reduction algorithm and reasoning algorithm is introduced to the fault diagnosis of SDG, solves the defects of SDG fault diagnosis, some researches have been carried out.(1) Using a new rule reduction algorithm based on GrC, used to compatible redundant attributes in decision table.(2) A rule based reasoning algorithm based on GrC, by computing the most similarity search the corresponding field information acquisition decision, this method can effectively reduce or even avoid the probability reasoning results conflict, resolution.(3) Using a SDG fault diagnosis method based on GrC, power plant deaerator system as an example, to build the SDG model, the decision table of fault diagnosis for SDG, generate the granule base after reduction, using GrC method of rule reasoning, fault source by computing the most similarity search, and in the Matlab simulation platform, the algorithm feasibility is verified. The simulation results show that, the method retains the advantages of SDG fault diagnosis, improve the diagnostic rate and resolution.