Study On Unit Commitment Of Power System Integrated With Large-scale Wind And Photovoltaic Power

With the continuous improvement of the grid-connected capacity of new energy such as wind power and photovoltaic,showing a large-scale integration situation in China,the randomness and volatility of its output bring severe challenges to the dispatching and operation of the power grid.Due to the prediction technology of wind power and photovoltaic energy is not mature enough,the modeling of new energy output becomes more and more important.For large-scale wind power and photovoltaic energy connected power systems,the existing unit commitment model can only maintain the grid stability by improving the spare capacity of the unit,discarding excess energy or even removing the load.These measures have greatly sacrificed economics.Therefore,this dissertation studies the unit commitment of large-scale new energy connected power system scheduling.The main work is as follows:Firstly,a new energy output model considering the correlation of wind power and photovoltaic energy output is proposed.The predicted value of the wind and photovoltaic energy output is divided into several intervals,and the kernel density estimation method is used to fit the measured value of the energy output of each interval,so as to obtain the output prediction error distribution of the wind and PV energy.The wind and PV power output distribution were fitted to the wind and PV combined output curves through five common copula functions.The optimal combined output distribution were obtained by comparing the results with the rank correlation coefficient and the Euclidean distance method.Secondly,a scenario generation method based on Latin hypercube sampling is proposed.Combined with the wind-PV correlation model proposed in this dissertation,thousands of wind and PV combined output scenarios are extracted in each period,forming the initial timing scenarios.The cholesky decomposition method is used to process the initial scenarios to reduce the correlation between scenarios.The k-means clustering method is used to reduce initial scenarios,and the K-DBI index is used to evaluate the effectiveness of clustering.The clustering number is optimized to obtain the optimal scenario set describing the randomness of wind power and photovoltaic energy output.Taking the historical data of a neighboring wind farm and photovoltaic power station as an example to verify the effectiveness of the proposed scenario analysis method.Thirdly,the unit commitment model based on scenario analysis is found.The fuel cost,start-up cost,wind and PV curtailment and load-cutting penalty cost are set in the objective function.The active power balance,unit output constraint,unit climbing constraint,unit minimum on-off time constraint and line security constraint are set in the constraint.In order to reduce the computational burden of the model,the higher order term in the objective function and the constraint is linearized,and the model is solved by Yalmip+Gurobi.Finally,aiming at the situation of large-scale wind and photovoltaic power integration,a stochastic unit commitment model considering frequency security constraint is proposed.The dynamic frequency model is studied when the system is disturbed,and the frequency security constraint is derived,which is added to the constraints of the unit commitment model;Meanwhile,the framework of the unit combination model is improved by setting decision variables of unit output for each scenario to strengthen the role of the random output scenario.In the IEEE-118 node system,the simulation analysis is carried out.Calculate and analyze the same group of data using the traditional unit commitment model and the proposed model in this dissertation respectively,and the results are compared to verify the superiority of the unit commitment model proposed in this dissertation.