Study Of The Competitive Bidding For Power Plant

Electric industry is one of the most important industries in our national basic industries. In order to rectify the malpractice of lacking competition in a long-term monopolistic operation mechanism, the reform that taking "the separation from the electric power network and the bidding for pool purchase" policy has been gradually conducted. It is the necessity for the transition of our nation's economy from planned economy into market economy. How to know, analyze and control real-time cost to win competition in electricity market is the most important problems for independent power suppliers.The assistant decision-making system for generation bidding of power supplier is studied in this dissertation. It just adapt under the electric reform situation which power plant is separated from electric power plant and the bidding for pool purchase. This paper explores a system of bidding strategies by which can offer multi-assistant bidding advices for decision-makers. This dissertation studies the important factors which affects assistant bidding system for generation of power supplier market factors: running cost analysis subsystem, profit subsystem and market forecast subsystem. To forecast load, this dissertation mainly not only introduces statistics and neural Network methods but also uses two neural methods to continue training. It also illustrates the influence of input layer, midst layer, output layer's amount and training precision. To forecast market clearing price mainly uses statistics and neural methods. To quote price is based on cost analysis. This dissertation gives model based on cost adding and marginal cost. Efficiency is based on integration of forecast and fact.The feasibility of these mathematics models is verified by comparing the calculation results with the data collected in the plant. Therefore, it will hold an instructional significance for assistant decisions making system for generation bidding of power supplier.