The Fault Diagnosis Of Smart Substation Based On Rough Set And Evidence Theory

Fault diagnosis system plays an important role to enhance the security and economy of a power system in a substation. Along with the construction of smart substation and the rapid development of IEC61850, internal of substation can share information and data easily. A fault diagnosis method of smart substation based on rough set and information fusion technique is proposed for the uncertainty and multi-source factors of fault information in a substation.The present status of smart substation is introduced in this paper. In addition, the basic structure, main technical characteristics and the influence of new technology to the development of smart substation is also expounded. Communication mode of smart substation is studied in detail, including the technical characteristics of IEC61850 standard system, description language of IED configuration and the content and transmission characteristics of the GOOSE message and SV message.Combined with rough set method, this paper expounds the basic concepts of rough set theory, the information system, the approximation space and uncertainty expression. The discretization algorithm and reduction algorithm are studied in detail, and Naive Scaler discrete algorithm based on the correlation and reduction enumeration method based on Apriori algorithm are emphatically considered.In order to summarize and unify results based on analyzing the information. Information fusion technology of evidence theory is studied, including the application of probability distribution function, the trust function and likelihood function. Meanwhile, the uncertainty relation of the representation is expounded, especially the concept of Dempster synthesis principle and its application method.A 220/110 kV substation system is modeled and simulated, and fault alarm information is collected from GOOSE message and SV message. Fault section and fault device are described as decision attributes. By using the rough set and information fusion method of evidence theory, combined with the analytic hierarchy process to implement fault diagnosis, this method always has a good diagnosis effect in the case of wrong information transmission. Experimental results verify the correctness and effectiveness of the algorithm in this paper.