The Research On Maintenance And Decision Support System For Electric Power Plant Equipments Based On Data Mining

Modern Electric Power Company is a capital and technology intensive company, which mainly reflected in the equipment management. The high efficiency and standardization of equipment operation, maintenance and other production activities is the core of the whole company healthy operation. Good operation of equipment is the basic condition for company survival and company creation of economic profit, but also safety protection. Maintenance System for Power Plant Equipment running status of equipment to complete accurate diagnosis and prediction of the safe operation of equipment and the national economic development is of great significance.Classic data mining algorithm used in power plant equipment identification and diagnosis include decision tree algorithm, neural networks and support vector machines. On the basis of analyzing the advantages and disadvantages of various classic decision algorithms, this thesis makes an improved support vector machine method used for the diagnosis and prediction of power plant equipment operating status. Improved support vector machine method regards support vector machines as the core algorithm, and utilizes quantum genetic algorithm to dynamically search the optimal training parameters. Experimental results show that this model has higher accuracy on diagnosis and prediction of equipment fault.One the basis of asset management projects of power plant enterprise, this thesis combines data mining methods and deeply digs mass of equipment operation data to establish power plant equipment fault diagnosis and prediction model and apply it to decision support system of power plant equipment. The model provides supports from equipment failure list shown by the system, curve comparison graph of equipment operation condition, trend analysis graph of equipment operation, and equipment defect analysis graph and as well as others for all leaders and electric power operating personnel focus on surveillance, roving surveillance or maintenance and replacement decisions.The system is based on B/S structure, and uses MVC pattern to implement, including five modules:user management, equipment information management, equipment defect management, equipment operation management and equipment central repository. The achievement of power plant equipment maintenance decision support system based on data digging has great significance for saving equipment maintenance costs and improving equipment utilization.