Research on Accurate Redundancy Perception Systems and Technology to Improve the Supply Guarantee Capacity of Industrial Parks

With the high proportion of decentralized resources such as distributed new energy, energy storage systems, and electric vehicles, higher demands are placed for the acquisition, perception, processing and integration of multisource data in the industrial park, so as to improve the power supply capacity and quality of the industrial park. However, the traditional perception processing technology of distribution networks makes it difficult to deal with the problems of data type diversity, inconsistent time scale and unequal spatial distribution in industrial parks. Therefore, this paper proposes a precise redundant perception technology and system research method to improve the supply guarantee capacity of industrial parks. Through virtual modeling and real-time simulation, the method combines the physical distribution network with the virtual system to accurately capture the spatiotemporal characteristics and internal correlation of multi-source data. Specific techniques include pseudo-measurement generation, data completion and correction, spatio-temporal interpolation and feature extraction. Pseudo-measurement generation uses historical data and state estimation to provide high precision estimation for unmeasured nodes; data completion and correction use KNN and interpolation method to handle missing values and abnormal values; and realize data synchronous interpolation based on space-time graph neural network. The outcomes of the experiments indicate that this approach is capable of remarkably enhancing the data - handling capability and supply assurance capability within industrial parks. Moreover, it offers dependable technical backing for the stability, precise perception, economic efficiency, and dispatching optimization of intelligent distribution networks.