Modeling of Medium and Long-term Transaction Behavior of Generators Based on Bayesian Network

With the advance of the new electricity reform, the annual planned electricity quantity of China's electric power system is gradually replaced by medium and long term transaction electricity quantity. Regardless of the electricity market participants of the medium and long term trading the imbalance of the uncertainty of market behavior, capacity, or trading results randomness factors such as power grid operation mode of the unexpected, results in uneven distribution of tidal current and power grid operation safety margin decline, increased the complexity of the system operation arrangement, may lead to serious power system security. Therefore, based on the Dynamic Simulation Platform for Macro-energy Systems, this paper develops the application of electricity market transaction Simulation based on the hybrid Simulation method of combining computer agent model and experimental economics. Combined with statistical analysis, causality analysis and behavior analysis three research paradigms to analyze the medium and long term trading behavior of power producers. Firstly, the key driving factors of power producers' trading behaviors in the electricity market are explored by combining expert knowledge, prior knowledge and relevant literature. Based on this, a proxy model of trading behaviors with the same distribution characteristics as real participants' trading behaviors is established by using Bayesian network model, and its validity is verified.