Theory And Its Application For Cost Management And Bidding Optimization Of Power Generation Enterprise In Electricity Market

The marketing environment for China's power companies is continuously changing, which leads to unprecedented pressure of competition. On one hand, the competition among power plants is gradually shifting from asset scales to cost management, and the cost priority strategy becomes the consensus. On the other hand, the analysis and forecast of electricity price are greatly emphasized, under the market reform aimed at "price bidding". Based on the analysis and summaries towards the marketing environment for domestic power generation companies, this paper identifies and filters the main risk elements, which then compose the risk factors sets. The structure model is explained here, and the hierarchical directed graph is given, which is used to obtain the conclusion that both cost and price are key and relatively controllable risk factors for power companies. The further research of this paper focuses on the mode of cost management and the analysis and prediction for the market price.The paper sums up the traditional cost management mode, and discusses the practical application situation of activity-based costing management mode in generation plants. The design mainly includes four parts. Firstly, establish the resource database, which is the base of the further apportion in activity and cost objects. Secondly, set up power activity database that is with clear levels and in accordance with industrial standards. The work refers to the relevant business processes in power enterprises and international standards. Thirdly, determining the core costing objects of generation equipment systems and units, then device-dependent maintenance cost, as well as all the unit-dependent cost. Fourthly, the design for cost driver model. Aimed at above various links, this paper goes deep into discussion, and analyzes the relevant factors which cause the important influence during the design for the activity-based costing management model. In this way, the design guideline can be made, and the specific model design process in every links.There are many price analysis and prediction technologies, and the main methods are usually boiled down to two kinds. One is based on statistics, while the other one is based on artificial intelligence. The generalized auto-regressive conditional hetero-skedasticity model is widely used, due to its higher adaptation to the hetero-scedasticity nature of price. However, the parametric assumptions conditions required by the model restrict its application. Based on the time series method, the paper studies the nonparametric estimation in the conditional variance function, and applies it to the day-ahead electricity price forecast. Compared to the traditional generalized autoregressive conditional hetero-skedasticity and other models, practical calculation examples show that the method proposed in this paper forecasts the price spikes, which is often observed in the electricity market, more effectively.Using a single model to predict electricity price will have its unique adaptability due to its characteristics, while using multi-model combination enables integrate useful information provided by different methods to improve the prediction accuracy. When dealing with the weight of the results from various models, a method making use of environment information is put up to evaluate the reliability at the predicting time point, then take comprehensive treatment towards the single model's credibility through data fusion principle in Dempster-Shafer's evidence theory, and eventually the combined weights are determined. Practical examples prove that multi-model combination based on the Dempster-Shafer's evidence theory is effective in volatile price forecast, whose forecasting accuracy is obviously higher than other methods.On the basis of the above theoretical research, this paper probes into the optimization coordination mechanism and system design of market bidding for power generation companies, to lay down the basis for practical application.