Load Forecasting Based On Multi-Variable LS-SVM And Fuzzy Recursive Inference System

The development trend of future power grid is the smart grid, with the industry increases dependence on electricity, the requirement of power supply reliability and power quality is increasing, with the characteristics of efficient, clean, safe, reliable and interaction, the realization of smart grid can't depart from the support of the accurate load forecasting technology. Meanwhile china is actively promoting the reformation of electric power system, the key tasks include orderly promoting price reform, straighten out the price formation mechanism; promoting the reform of electric power trading system, improving the trading mechanism to establish the power market; establish an independent trading mechanism, form an fair and standard market transaction platform. The development of electric power Demand Response (DR) brings great change to the traditional power utilization mode. Combined with real-time electricity price, consumers can adjust their power utilization mode by their energy demand. This makes load forecasting more complicated. At present some research results of the domestic load forecasting under the environment of Demand Response have been obtained, with the opening gradually of power market, it's important to study the load forecasting method under the complex environment of Demand Response.This subject regards the load forecasting method under the complex environment of the smart grid's Demand Response as the investigated subject, the relationship between the power market's retail side and demand side was discussed to explore the higher accuracy of load forecasting method. The primary coverage as follows:First of all, selected sample data by the calculation of load series shape similarity. Considering the inherent law of the prediction moment's load and the load some time before, and the load curve's differences between the overall similarity and local similarity, introduce the load curve coefficient to the choice of forecasting samples, can reduce the training data, greatly improve the speed and forecasting performanceNext, the multi-input and two-output least squares support Vector Machine (LS-SVM) was proposed to preliminarily predict the load and price at the same time.Then, considering the interaction between the real-time electricity price and load, the fuzzy recursive inference system based on data mining technology was adopted to imitate the process of thought, by mining the association rules between electricity price variation and load variation to simulate the game process between the electricity price forecasting and the load forecasting.Finally, the preliminary forecast results of multi-variable LS-SVM prediction algorithm were recursively corrected until the forecasting results were tending towards stability. Multi-variable LS-SVM can avoid running into local optima and has an excellent capacity of generalization, the improved association rules mining algorithm and loop predictive control algorithm have good completeness and robustness, and can correct the forecasting result approximately in every real situation. Simulation results of the actual power system show that the proposed method has better application effects.