Research Of Hydroelectric Engineering Investment Control Based-on Power Price Forecast And Realization Of Maximum Of Efficiency

Usually hydroelectric engineering investment control is carried out in design,bid and construct stage. In the essay, the control of the investment of hydroelectric station and the realization of the most profit are approached, at the same time, some research of Grey Model and Artificial Neural Network is applied in forecast of system marginal price. Currently, competitive bidding is gradually carried out in power plant market in our country: compensate cost appropriately, confirm gains rationally, calculate tax by law, stick load fairly, and carry out "same net same quality same price". So in order to reduce the investment of the hydroelectric station, it is necessary to reduce the network power price. For the hydropower project of the same, the cost lower and the benefit higher, on the contrary, the investment higher and the benefit lower. So the investor must consider whether invest and how to invest to gain the most earnings according the network power price, when he invest on the hydroelectric station. Short-term and long-term accurate forecast is the key for the investment on hydropower station and achievement of the most profit. Accurate forecast of system marginal price next day will be helpful to hold market chance and constitute the best policy of quantity of electricity and power price to realize the most profit. And several sorts of models for 96 spot price forecast are proposed by search. Based on historical price materials, form GM(1,1) and forecast 96 spot price next day by seeking alteration regularity which is hidden in price materials. With the thought of the interaction between load and price, form the GM(1,2) and forecast 96 spot price next day. Considering limitation of GM(1,1), apply the modified method of GM(1,1) and forecast 96 spot price next day. Based on historical price materials, form Backpropagation Neural Network(BP) and forecast 96 spot price next day by making full use of fitting goodness of BP. With the thought of fitting goodness of BP and that GM can reduce random disturbance, connect BP with GM in series and forecast 96 spot price next day. According to influencing factors and vary regularity, connect GM with ANN based on weather sectors in parallel and forecast 96 spot price next day. Through analogue forecast, and giving considerations to the prediction accuracy and stability, author thinks that GM(1,1) and modified method of GM(1,1) and connection BP with GM are better prediction models in forecast 96 spot price next day.