Research on Load Forecasting Model for Distribution Network Based on Virtual Definition of Internet of Things

In the process of smart grid construction, the increasing demand for electricity in power-consuming areas often leads to the operation of distribution network equipment at full load or even overload, which puts great pressure on the protection of the power system. At present, equipment load forecasting models are mainly used to predict and analyze various data-related factors of equipment, which are divided into temporal factors and nonsequential factors, to reduce the complexity of equipment management on embedded platforms. According to the original intention of ubiquitous power Internet of Things (IoT) construction, this paper uses the embedded platform of distribution network equipment to build a fog network. It also adopts a virtual defined network to manage the data communication of each node, and effectively utilizes edge computing to complete data processing tasks. A lightweight load forecasting scheme with both time series and nontime characteristics is proposed, and a neural network feature model is pre-trained. For the sensing parameters extracted from embedded systems of power equipment and load prediction, the unidirectional continuous cyclic long short-term memory (LSTM) connection is regarded as a time series neural network modeling. We then consider the determining factors of historical data on equipment operation capability to predict the effectiveness of load capacity, and extract load data feature vectors to compare potential relationships. Simulation tests have shown that the research method proposed in this paper is more effective in predicting load capacity and analyzing different temporal patterns of loads, facilitating the decomposition of data in the initial stage. Software-defined virtual networks use lightweight structures to reduce network operational pressure and minimize data space occupation. The research results can achieve the important function of early warning of abnormal operation of the power grid system, reduce the operational risks of the power grid system, and have significant implications for improving the economic and social benefits of the power system.