Research On Generation Self-Scheduling Based On Distributionally Robust Optimization Method

In a deregulated electricity market,the power generation companies need to compete to maximize their own profits,thus they need to bid based on the bidding curve.In the POOL trading market,the power generation companies and the consumer propose the supply and demand,and then the ISO will establish a suitable market clearing rules in accordance with the grid transmission capacity and network constraints.The power generation companies can only provide the corresponding power generation plan to the ISO through the prediction of the LMPs.In the fully competitive electricity market will produce more uncertainty,with the market-oriented operation of the system bidding.Therefore,the challenge for power generation companies is how to establish an effective self-scheduling of power generation per the electricity price forecasted by the electricity market.self-scheduling of power generation model is a kind of programming problem with uncertain random variables.The model has some uncertainties,such as load and LMPs.The difficulty of the solution will increase with the increase of the system size,the constraint condition and the uncertainty parameter.For the programming problem with uncertain random variables,the modeling method is stochastic programming method and robust optimization method.The former introduces random variables to describe the uncertainty,but this method needs to know the complete distribution of random variables,and will produce a large number of discrete sample points,resulting in the calculation scale is too large;the latter does not require the probability distribution of parameters,It is only necessary to assume that uncertain parameters belong to an uncertain set,but this method may be too conservative and does not use some probability statistics that can be obtained.In this paper,we consider that the LMPs have a certain range and a certain probability distribution.Therefore,a method of combining stochastic programming with robust optimization can be used to solve the problem of generating self-scheduling with uncertain LMPs.Although the probability distribution of LMPs is uncertain,the moment parameters such as expectation and covariance of LMPs can be obtained by long-term statistics.The short-term probability distribution of LMPs is similar to the probability distribution of long-term statistics,but its moment is generally different,that is,the moment is uncertain.Therefore,solving this kind of problem by using the method of distributionally robust optimization under moment uncertainty can deal with LMPs.Firstly,a min-max problem is set up,which is based on the distribution set with expectation and the covariance of the LMPs.Then the Lagrange duality principle is used to solve the original problem by converting it into a SDP.The feasibility of the proposed method is validated by the analysis of IEEE 30Bus system.