Power Producers In The Electricity Market Bidding Behavior Analysis

As the most important market member in the electric market, generator has very important influence to the whole electricity market. Under certain market condition, generator probably utilizes market power by physics or economic withholding, result in the electricity price to be higher than the normal competitive price, even the electricity price spikes occurred. For electricity market to run steadily and safely, and to protect the healthy development of market, it is very important to analyze and control the generator bidding.Firstly, this paper defines CPI (Capacity Price Index), and applies CPI to analyze the generator bidding of Zhejiang electricity market. The analysis results show: CPI can reflect generator abnormal high bidding sensitively and detect generator's withholding behavior effectively. It not only can reflect the short-term change of the generator bidding, but also can reflect the long-term change accurately. It is valuable to analyze and control the generator bidding.Secondly, an average electricity price difference-integration model is proposed in this paper. This model can transform unit's bidding curve of power producer in market into a one-dimensional feature vector, so can implement classification of unit's bidding using classical clustering method. Through clustering calculation of bidding curve from Zhejiang electricity market, it is shown that bidding curve can be classified accurately and efficiently using 20-segment model and square sum of deviations in the average electricity price difference-integration model. Moreover, CPI(Capacity Price Index), HHI index can be used to analyze the bidding clustering results further.Electricity price is the basis of decision making for each participant in electricity market. This paper uses the method of CMAC neural network to build day-ahead electricity price short-term forecasting models. By this method, different models are designed for the different settlement periods respectively. Then using the data of California electricity market, apply CMAC and BP neural network to forecast short-term electricity price. The comparison between two results shows that CMAC neural network can work more steadily and faster in electricity price short-term forecasting than BP.