| With the rapid development of Smart Substation, the run of power grid is getting big changes in technology. Meanwhile, a series of problems are born, for example the problem that the backward mode of maintenance can’t fit the need of highly automation of Smart Substation. Thus the development of the condition-based maintenance of Smart Substation has already become very urgent. Currently, the State Grid has established the guides for condition evaluation of some transmission equipments, such as SF6 high-voltage circuit breaker, oil-immersed power transformers. Meanwhile, the study on the condition-based maintenance of primary equipment has also acquired many achievements at home and abroad. Based on the background, the paper focuses on the research on the condition evaluation methods of secondary system of Smart Substation which has not been made. The specific work related is as follows:(1) The establishment of the condition evaluation system of the secondary system of Smart Substation. Firstly, the evaluation scope of secondary system of Smart Substation is defined, according to related concepts in the standards, to collected its state information conveniently. Secondly, the principle of the evaluation system is set based on the experience of the related study, to provide standard for selecting the evaluation indexes. Thirdly, the indexes are divided to different layers from three aspects:running environment, intelligent devices, communication system. Finally, a complete condition evaluation system of secondary system of Smart Substation is established scientifically.(2) The build of the evaluation algorithm models. First, the evaluation indexes are divided to six classes from two angles:the trend of data change and the priority levels. Then, the different models of membership function are set up to correspond the six classes, and some examples are given to illustrating how to set parameters according the standard files. Finally, all the basic indexes can be quantitatively evaluated by the six algorithm models.(3) The determination of the weights of evaluation indexes. To begin with, the hierarchical structure of the different BP neural networks is built according the hierarchical structure of evaluation system. Next, the trainings of the different BP neural networks are used to fitting the evaluation process of the secondary system of Smart Substation. Among those, the inputs of the networks are the real-time evaluation results of the basic indexes, the expect outputs of the networks are the real-time evaluation results of the basic indexes, and the special nodes in engineering practice are selected as the training samples. When the actual outputs keep with the expect outputs, the trainings succeed. Finally, the weights can be calculated by using the presented formulas.(4) One example of a smart substation is given to hackling above three parts content. Ultimately, the condition evaluation method in this paper is proved feasible in engineering practice. |
