Research On Electric Power Load Forecasting And Power Optimization For Commercial Building Based On Data Mining

With the development of economy, the power consumption of society grows rapidly, the power grid capacity is insufficient, the energy utilization rate is not high. At the same time, the problem of environment and resources is also facing great challenges. Therefore, how to improve the level of intelligent electricity is important for improving the efficiency of social electricity and achieving energy conservation and emissions reduction. Nowadays, the energy consumption of commercial buildings is increasing. So it’s of great significance to make use of electricity intelligent management. New commercial buildings installing electrical energy consumption monitoring system provide favorable data for the research.Data mining technology can excavate valuable information from the vast data. The dissertation aimed at commercial building load forecasting and electric optimization problems by using data mining technology were studied.Firstly, based on the analysis of the characteristics of commercial power load, the Rough set was used to reduce the important factors of load forecasting. The model of Wavelet Support Vector Machine was built to forecast the commercial power load by combining the Wavelet theory with the Support Vector Machine. Example shows that the accuracy of load forecasting of the mutational load sequence is better than the model of SVM and the model of Artificial Neural Network.Secondly, based on the characteristics of commercial power load, the commercial load was divided to fixed load and controllable load. The power consumption optimization model of the air conditioning load of the controllable load was established by minimizing the cost of electricity and optimizing thermal comfort as the goal of power control. The load scheduling optimization model for the transfer load of controllable load was established. The non-dominated Genetic Algorithm and the Adaptive Genetic Algorithm were used to find the optimal solution respectively. The result shows that the method helps users reduce the cost of electricity on the premise of not affecting the electricity comfort.Finally, the software of business intelligence management was developed by C# and MATLAB hybrid programming technology, realizing the function of commercial power load prediction and optimization of commercial electricity.