The establishment of power system operation mode plays an important role in the safety, stability, economy and high quality operation of power grid. In order to enable the power grid operation mode staff can more quickly judge the advantages and disadvantages of the power flow of the operation mode according to the mode change and improve the flexibility and efficiency of the mode compilation, a fast evaluation method of the power flow based on the deep belief network is proposed. Firstly, considering the safety and stability guidelines of power system, the method establishes the evaluation index system of power flow state from three aspects of safety, stability and economy. Then the comprehensive evaluation method based on entropy-weight TOPSIS is adopted to comprehensively evaluate the evaluation indexes of multiple operation modes. Then, the index values of the above power flow evaluation system and the comprehensive evaluation results of entropy-weight TOPSIS are respectively used as the input and output data of the rapid evaluation model of power flow based on DBN. DBN algorithm based on RBM is used to extract deep features to complete unsupervised learning process, and then the supervised BP neural network is used as the conventional fitting layer to obtain the evaluation results. The flexibility and practicability of comprehensive assessment is effectively improved. In this paper, an example of IEEE39 nodes system is used to verify the effectiveness of the proposed model and algorithm.
Keywords operation mode, assessment Index, comprehensive evaluation of power flow, deep belief network, rapid evaluation