Reactive Power Optimization Of Distribution Network Considering The Output Fluctuation Of Distributed Power Generation

With the international community's attention to the resource development and protection and focus on energy transformation,clean energy power generation will be the main power source someday.It is an inevitable trend for a high percentage of sustainable energy to be connected to the power grid,thus,research on grid-generation is becoming increasingly important.After the distributed generation with high percentage of random output is connected to the electrified wire netting,the random variables in the system not only contain the load fluctuating,but also increase the uncertain factors such as the output of distributed generation,which makes the reactive power optimization of distribution network more complex.Therefore,it is of practical significance to study the reactive power optimization of power system under uncertain factors.In this paper,according to the grid connected characteristics of distributed generation,considering the operation economy,stability and reliability of distribution network,taking wind power and photovoltaic power generation as representatives,the reactive power optimization of distributed generation grid connected distribution network is studied from the perspective of probability.Firstly,the output characteristics of wind and solar power generation are studied,and the probability models of wind-solar power generation output and load output are selected to accurately describe the uncertainty of wind-solar power generation and load output.Three methods are used to establish the probability model of wind and solar power generation,and then the beta model and fitting process are determined.Photovoltaic power generation is set as beta distribution and load output is set as normal distribution.Secondly,the probabilistic power flow calculation in the process of reactive power optimization is studied.The three-point estimation method is used to sample the input random variables of the probabilistic power flow to determine the output power and probability of wind-solar power generation and load.Considering the output correlation of distributed generation in the same area,the three-point estimation method is improved.Principal component analysis is used to deal with the correlation of random variables,and the related random variables are transformed into independent spatial samples for sampling.The accuracy of the improved three-point estimation method is verified by an example.Then,considering the limit of the number of equipment actions in the whole day,the reactive power compensation is optimized under the condition that the system network topology is determined.The expectation of the output variable was obtained based on the improved three-point estimation power flow calculation,a mixed target dynamic reactive power flow optimization model is established,which includes the minimum expectation of active power loss,the voltage deviation and the L index of voltage stability.The model is transformed into a single objective to eliminate the dimension.Then the weighting factor and learning rate of PSO are ameliorated to solve the objective function better.The simulation results in the ameliorated IEEE33 distribution system verify the effectiveness of the proposed optimization method.The optimization method enables the power network to run within the constraints and ensures the security,stability and economy of the distribution network with distributed generation.Finally,considering the effect of reconfiguration on reducing network loss and improving voltage stability,the reactive power comprehensive optimization of distribution network with wind-solar power generation connected to the grid is carried out.The comprehensive optimization strategy of reactive power after reconfiguration is determined,and the mathematical model of reactive power comprehensive optimization is established.According to the standard deviation of output variables,the dynamic reconstruction period is divided to avoid frequent switch actions,and then the particle update strategy is improved to achieve the optimal objective function and output the reconstruction results and reactive power compensation capacity.The effectiveness of the proposed method is verified by an example.