Power Grid Resource Data Homologous Maintenance Based on Complex Representation Disambiguation Model Component

Power grid resource planning is the key link of stable power supply and the important engine of smart grid system construction. The same source maintenance of power grid resources is a heavy systematic work, including line loss calculation, load rate analysis, geographical survey, grid optimization and a large number of plates. The project volume is large, and data processing is difficult. Aiming at the problem that there may be a lack of data information in the multi-modal data of the same source of power grid resources, an algorithm for disambiguation model is proposed. Based on the idea of multi-instance multi-label learning, the method transforms the multi-modal data of complex objects into multi-modal multi-instance form so that each modal data of an object can be regarded as a package, which is composed of multiple instances. The consistency between different modal data packages of the same object is met so as to eliminate the influence caused by the inconsistency between modal examples. According to the multi-marker prediction task of power grid resources, the M3DN is used to propose a method using the correlation between markers based on the optimal transmission theory to improve the accuracy of prediction and mine more hidden relationships between markers. In the experimental analysis, a variety of algorithms are used to compare with the content of this paper. The experimental results show that the performance of this algorithm is 2.5% ahead of other methods in the benchmark data set. The density of data mining is better than other algorithms, and the effectiveness of interaction matrix is 5% higher than other methods. This study has a certain theoretical and practical engineering application value. It is of great significance to improve the efficiency of power data homologous maintenance, operation and maintenance. It enhances the level of power grid resource integration, and better serves the smart grid business.