The Optimization Of Reactive Power Compensation For Distribution Network With Dgs

The exploitation of distributed generations not only can realize energyconservation, but also it can effectively improve the distribution network economicbenefits. Meanwhile, the optimal planning of reactive power compensation is ofgreat significance in improving voltage, decreasing the power loss and improving thequality of power. Consequently, it is more meaningful to study how the distributionnetwork with DGs can achieve optimal reactive power compensation allocation.According to actual conditions of the 10kV distribution network with DGs, amathematical model of reactive power compensation is established by regarding theDG as a PQ node in the paper. From the economics perspective, taking the timevalue of money into consideration, the total investment costs of reactive powercompensation devices are converted into the same investment expenditures in eachyear. As a result, the minimum sum of the annual power loss costs and the annualinvestment costs which have been transformed is selected as the objective function.In view of the distribution network with dynamic loads, three load models whichinclude the maximum load, the normal load and the minimum load are employed toapproximately simulate the loads change during the reactive power compensation. Inorder to solve the problem, the mathematical model above is divided into three submodels which include the maximum load sub-model, the normal load sub-model andthe minimum load sub-model. Furthermore, it is meaningful to combine the annualinvestment charges with the annual power loss costs to evaluate the finalcomprehensive benefit in the maximum load sub-model. Simultaneously, the annualpower loss costs are only considered in the normal load sub-model and the minimumload sub-model. Then three sub models are coordinate by the method of usingmaximum load loss time.Improved particle swarm optimization is applied to get the optimalcompensation position and back/forward sweep is employed to calculated theoptimal reactive compensation capacity.In the end, the simplified treatment of the models is made, by taking two reallines into consideration and presuming the networks with DGs. By calculating 12nodes and 20 nodes distribution network, the result illustrates that the method iseffective.