Reactive Power Optimization Of Distribution Network Injected With Distributed Photovoltaic

Low-voltage(LV)distribution network is in the terminal of the power grid.With the characteristic of long lines and large impedance ratio and massive reactive demand,LV will often lead to LV qualification rate and affect consumers' voltage quality.Reasonable reactive power optimization schemes can reduce the system loss and improve voltage level effectively.With the development of new energy technology,the penetrance of distributed photovoltaic(PV)power supply in LV distribution is increasing,while providing renewable energy for power grid,it also impact the operation of the power grid.The output of distributed photovoltaic power supply is closely related to meteorological factors,and it has the characteristic of intermittent and randomness,witch will change the distribution network's power flow and increase the uncertain factors to the reactive power optimization of distribution network.So,when carry out reactive power optimization of distribution network,the effect of PV random output on optimal allocation should be considered,in order to improve the robustness of optimization scheme.The front-end centralized compensation method is widely used in distribution network presently,but it will often cause the front-end voltage higher than standard value and the back-end voltage lower than standard value,and the voltage qualification rate can't be improved effectively.In order to solve the problem effectively,the distributed reactive power compensation robust optimal allocation method is proposed.In the model,objective is to minimize total power loss of distribution network,with uncertain variables of load power,light intensity and temperature described as uncertain set.Its constraints include upper and lower limits of node voltage and reactive power compensation capacity,and limit of total number of installed compensation devices.A deformed Sigmoid function is used to approximate symbolic function to smooth non-differentiable function in the model.Bi-level optimization method is used to transform robust optimization model with uncertain variables to bi-level deterministic optimization model,which is solved alternately.Finally,interior point method is adopted to solve the deterministic model to obtain robust optimal allocation scheme of reactive power compensation.Analysis of a real LV distribution network demonstrates that the obtained robust optimal allocation scheme of reactive power compensation can ensure all node voltage qualified no matter how photovoltaic output and load power changes in their own fluctuation interval.Compared with deterministic reactive power optimization scheme,the method has stronger robustness.Aiming at the actual structure and operation of LV distribution network,the three-phase four-wire distributed reactive power compensation robust optimal allocation method is proposed.In this model,the asymmetric of loads and single-phase PV is considered,and the randomness of single-phase PV output and the fluctuation of loads is described as cassette uncertain sets.Three-phase simultaneous switching and single-phase independent switching reactive power compensation capacity is used as decision variables in this model.Simulation calculation of a real three-phase four-wire distribution network,the allocation scheme of reactive power compensation is obtained,which has stronger robustness,and ensure all node voltage qualified no matter how photovoltaic output and load power changes in their own fluctuation interval.Also,it reduce active power loss and unbalanced voltage of distribution network effectively,so the scheme is more applicable in engineering.