Risk-limiting Dispatch Of Power System With Renewable Under Extreme Weather

In extreme weather conditions,the active output of wind farms can drop steeply within a very short period.Such situations,when they occur,pose a great threat to the safe and stable operation of the power system,and pose new challenges for the scheduling of the power system.Taking the initiative to remove part of the load too early before the arrival of extreme weather will lead to an overly conservative strategy,which result a waste of wind power resources.Uncontrolled disconnection of turbines from the grid during extreme weather will cause an imbalance between power supply and demand,resulting in passive load shedding.Therefore,this paper focus on the problem of power system scheduling in extreme weather and applies risk-limiting dispatch to the system scheduling.The basic principle of risk-limiting dispatch is to constantly adjust the scheduling strategy according to the change of extreme weather-related information.Under the premise of meeting the risk constraints,the cost of using the quick response resources and emergency response resources on the power side of power system can be reduced to achieve a better scheduling strategy.Firstly,this paper introduces the basic principle of fixed time intervals risklimiting dispatch and propose a unit commitment-based fixed time intervals risklimiting dispatch model under extreme weather.The model is based on the basic framework of multi-stage stochastic optimization,where multiple decisions are made to continuously adjust the scheduling strategy and realize the cost reduction of quick response resources and emergency response resources in the system according to the changes in wind power forecast information before the actual operational stage comes.It is characterized by a fixed time interval between individual decisions.Secondly,this paper proposes the concept of data-driven adaptive time interval risk-limiting dispatch based on fixed time interval risklimiting dispatch and construct a data-driven adaptive time interval risk-limiting dispatch model The basic principle of data-driven adaptive time interval risklimiting is similar to that of fixed time interval risk-limiting dispatch,with the improvement that the data-driven adaptive time interval risk-limiting dispatch uses the information gain of additional data on wind power prediction results as the basis for the system to make a scheduling policy.This model makes fuller use of information about the uncertain variables in the system than does fixed time interval risk-limiting dispatch,further reducing the reduction of operation cost of power system.Finally,to reduce the computational burden of the risk-limiting dispatch model solution process,the stochastic unit combination model with three types of 0-1 variables and the generalized Benders decomposition algorithm are applied to the unit commitment-based risk-limiting dispatch to improve the computational efficiencyThe analysis of the case study shows that compared with the traditional twostage stochastic optimization,the fixed time interval risk-limiting dispatch can realize cost reduction of expensive quick response resources and emergency response resources in the system,and its scheduling solution is better than the two-stage stochastic optimization,and the lower the capacity of quick response resources in the system and the higher the scheduling cost,the better the fixed time interval risk-limiting dispatch presented.The data-driven adaptive intervalbased risk-limiting dispatch can further realize cost reduction of quick response resources and emergency response resources in the system compared with fixed time interval risk-limiting dispatch,and its scheduling results are better.In terms of solving the unit commitment-based risk-limiting dispatch model,the use of the unit commitment model with three types of 0-1 variables and the generalized Benders decomposition algorithm can effectively improve the model solving efficiency.