Multi Energy Resource Combined Unit Commitment Under High Wind Power Penetration Level System

With the diversification of energy types,the proportion of renewable energy in the grid is increasing rapidly.Among the numerous renewable energy sources,wind power is a kind of reliable and high quality clean energy.In order to achieve the goal of renewable energy accounts for more than 60%of primary energy consumption in 2050,Wind power as a pillar of future renewable energy has received widespread attention.However,on the development of wind power in China,the high penetration of wind power output brings great impact to the safe and stable operation of power system,which leads to large scale wind farms output cannot be fully cosumed.This problem,in addition to restricted by the limited transmission capacity,the different geographical location of source and load and others hardware factors,inaccurate prediction methods and modeling methods,low accuracy of power distribution scheduling are also the reasons.Therefore,it is necessary to integrate the existing resources and accept the new scheduling method under the condition of high penetration wind power in order to explore an more economic,reliable unit commitment and economic dispatch plan.According to the existing problems of wind power generation and its application in unit commitment,the dissertation analyzes wind power scenario generation and reduction method,modeling and solving the probabilistic unit commitment under high wind power penetration level system,and adding mutiple controllable load into the model,solveing and analyzing the unit commitment based on source-load coordination.Firstly,for the wind power scenarios generation and reduction,Latin hypercube inverse sampling(LHS)and improved K-medoid clustering based on particle swarm optimization are applied to the analysis of wind power forecast error,scenarios generation and scenarios reduction.The wind power output scenarios are reconstructed by the combination of wind power forecast and its empirical error distribution.Based on the statistical data of wind power output,kernel density estimation is applied to the calculation of the probability distribution of wind power forecasting error,the LHS is used to generate the scenarios.In order to apply the generated scenarios to the calculation of unit commitment,we propose an improved K-medoid clustering based on particle swarm optimization for scenarios reduction.Aiming at the problem of slow convergence speed and variable clustering result in traditional K-mediods clustering,particle swarm optimization(PSO)algorithm is used to accelerate the convergence speed of clustering and optimize the clustering quality.The BPA wind power forecast error data in first half year of 2016 is used for scenarios analysis,a 24 hour wind power output model is built,Latin hypercube inverse sampling and improved K-medoid clustering are uesd for scenarios generation and reduction.Then,in view of the problem of wind curtailment and load loss caused by high wind power penetration,this dissertation presents a probabilistic unit commitment model with high wind power penetration level.In the modeling,wind curtailment and load loss are added into the probabilistic unit commitment model,wind reserve is considered and part of the total output of the units are used as an emergency reserve.In this dissertation,the problem is transformed into a mixed integer quadratic programming problem,and the treatment methods for the start-up and shutdown costs and the minimum start-up time of the thermal power units are given.A wind farm is added into the example of 10 unit cases.In order to analyze the problem under the high proportion of wind power output integration,a relatively extreme wind power output scenarios is established.The shortage of wind power output,the excess of wind power output and the shortage of thermal units are analyzed in this model,and the flexibility and effectiveness of the proposed probabilistic unit commitment model is verified.Finally,in order to solve the problem of coordination of multiple types of controllable load in power grid,this dissertation presents a source-load coordination based probabilistic unit commitment model under high penetration of wind power.In the modelling,this controllable loads are divided into four types according to the difference of response characteristics and response methods:interruptible load,energy storage load,energy translational load,demand response load.The model takes the minimum expected cost of the unit operation as the goal,combined with the real time electricity price,analyzes and discusses the operation and the compensation cost of various controllable loads.In the example,the four kind of controllable loads are added into probabilistic unit commitment model.The consumptive ability of various controllable load on wind power are analyzed and compared.The example verified the proposed source-load coordinated probabilistic unit commitment model can better optimize the source-load start-stop mode,reduce expected operating cost,decrease the load loss and wind curtaiment under unchange thermal units capacity.