Granular Computing Reduction Method Of SDG Non-accommodating Fault Decision Table And Its Application

Because the system scale and system complexity growing day-by-day, All kinds of unexpected changes often occur in the system interior. Therefore, we need to correctly detect and diagnose these changes, and take corresponding measures to re-configure the system to maintain complicated system's performance at a high level. Besides, in order to protect the system's reliability and the security, we need to rapid and timely diagnosis the failure symptoms which monitored Momentarily. Therefore, The failure diagnosis technology already became a research hot spot in the automatic control domain.SDG-based fault diagnosis technology has the ability to contain massive information, We can use the storage knowledge information stored in the SDG model to reveal system variable's causal relation and the influence, to search for the disturbance of fault source, So as to identify the reasons of the system disturbance effectively, and diagnosis the system fault timely.The SDG model is one kind of process model, is the integration with the artificial intelligence domain's method. Granular Computing is an important branch of artificial intelligence methods. Granular Computing was introduced into the SDG fault diagnosis technology to carry on the compression or the reduction of the diagnosis characteristic value and to exclude unwanted characteristic variables. This can simplify the complexity of conventional diagnostic techniques and statistical workload greatly, so as to improve the efficiency of fault diagnosis technology.In this paper, using granular matrix reduction algorithm to simplify SDG fault decision table in order to transform the attribute reduction into the relevant computation of the binary granular matrix. The study has developed the research and the application scope in the fault diagnosis domain with emerging intelligence computational method-granular computing.This article will combine the Granular Computing and SDG fault diagnosis method together, Granule is the unity of data structure and relationship, which can effectively describe and express the topology structure of SDG, binary granule matrix can effectively describe the nodes and edges information of SDG model. Meanwhile, the granular computing can simplify the redundant knowledge and the structure of SDG model, which realize the purpose of simplifying system structure and knowledge rules. The main work are as follows:1,The representation of binary information granules based on the SDG model,carry on the knowledge granulation to the SDG model;2,The establishment of condition attribute and decision attribute's binary granule in SDG model;3,The definition of binary granule matrix,The knowledge granulation of SDG model;4,SDG model's simplification,carry on the knowledge reduction of SDG model;5,Simulation and compared confirmation.