Research Of Models On Short Time Electric Marginal Price Forecasting

System marginal price is the uniform price reflecting short-term supply-demand relationship of electric commodities in the electric power market. It contacts customers, market supervisors and individual energy enterprises. It has economic relations with supply side and demand side. At present in most electric power market the electric price bases on market clear price or system marginal price, for instance the New England electric power market in U.S.A. and several provincial electric markets which our country realizes tentatively, etc. System marginal price is the product price of an individual energy enterprise whose profit relies on successful bidding policy, and bidding policy is generally based on accurate mastery of short-term market tendency whose key is to carry out forecast of system marginal price. Therefore, research on forecast of system marginal price is an urgent task for all individual energy enterprises in competitive circumstances. Meanwhile, from the point of customers, system marginal price forms their unit cost of using energy and forecast of system marginal price makes dynamic cost control possible. From the angle of market supervisors, forecast of system marginal price promotes the healthy, stable, and orderly development of energy market and supplies foundations for establishment of all energy price policies. Accordingly, forecast of system marginal price is significant for three kinds of participators in power market. The thesis mainly researches the prediction of short-term system marginal price. The main influencing factor and the variational characteristic of the electric price are deeply discussed at first. Reaffirming the regularity of electric price and the indispensable factor, which play an effecting role. The short-term electric price is influenced by a great deal of factors, especially some human factors and the factors of the market power, etc. These factors have influenced on electric price seriously, have made the prediction of the electric price difficult. The thesis analyzes the impact on electric price of market power, imports of supply and demand ratio, which weighs market power in the forecasting model. The effectiveness of the methods is testified by simulating examples. Both power system load and price are time series. Theoretically speaking, methods applied to forecast load could be used to forecast price, such as Time Series Analysis, Artificial Neural Network and Wavelet Transform etc. But electricity price is more difficult to forecast than load because of its inherent characteristics such as levity and many influencing factors. Comparing several methods, RBF neural network model is established to forecast system marginal price. Aiming at the characteristic of RBF neural network, a hierarchical genetic algorithm for RBF neural network is proposed to train network parameters and configuration simultaneity. The feasibility of model is testified by simulation examples.