1 February 2022 in IEEE Transactions on Cybernetics
Influenced by many complex factors, it is very difficult to obtain high-performance industrial power load forecasting. The industrial power load forecasting is deeply studied by fusing some machine-learning methods for industrial enterprise power consumers. As a result, a novel power load forecasting method is proposed by taking into account the variation of load characteristics in different regions, industries, and production patterns. First, through the improved K-means clustering analysis, the historical load data are classified as the production patterns to which they belong. Then, the prediction algorithm combining reinforcement learning with particle swarm optimization and the least-squares support vector machine is proposed. Finally, the improved algorithm in this article is used for short-term load forecasting separately by the load data in different patterns after the above processing. The forecasting method in this article is based on data driven with real datasets. The results of the simulation experiment show that the improved prediction algorithm can distinguish the changes in different production patterns and identify the load characteristics of different regions and industries with high prediction accuracy, which has practical application value.
20 January 2020 in IEEE Access
It is very difficult to predict the Maximum Power Demand (MPD) of customers in high performance because of various factors. In this paper, the problem of MPD prediction is studied by using fused machine learning algorithms. Firstly, an improved grey relation analysis method is adopted to analyze relevant influencing factors. Secondly, a modified prediction algorithm based on an adaptive cubature Kalman filter combined with Fbprophet is proposed according to the characteristics of customers’ MPD. Finally, the proposed algorithm of this paper is applied to predict MPD and cost is evaluated. Experiment results show that the improved MPD prediction algorithm can comprehensively consider the relevant factors, and has good performance in time series prediction.