Reactive Power Reserve And Optimization Based On Network Partitioning

Reactive power/voltage control plays an important role in maintainingpower system operating in economic and stable manner. At present theinterconnection of the regional power grids is gradually formed with thedevelopment of the Ultra High Voltage Grid. The growing system scaleand more complex control characteristics have been brought about newchallenges to optimal reactive power dispatch. Thus, the systemmanagement by dividing it into a series of decoupled voltage control area,based on the regional character of reactive power, will be a feasibleapproach.Network partitioning method for voltage control is investigated in thispaper. The concept and definition of reactive power reserve based onvoltage control area (VCA-RPR) is developed, and VCA-RPRoptimization model is formulated.In network partitioning, definition of electrical distance to measurethe coupling degree of reactive sources and a multi-objective modularityindex to measure the partitioning quality are proposed, and then amulti-level network partitioning method is developed. Numerical resultson several large scale systems demonstrate that the proposed method caneffectively handle the high dimensional clustering problem and reduce thecomputational cost by the coarsening-partitioning-refining phases,leading to partition schemes with high quality. Partitioning schemesobtained by the method are of high intra-electrical coupling and lowinter-coupling, and the connectivity of each partitioned voltage controlarea is guaranteed.Based on the investigation on bus voltage vulnerability index and weighting factor of reactive power reserve (RPR), definition of VCA-RPRis proposed as the weighted sum of RPR of each reactive source. Thecontrol characteristics of generator terminal voltage is analyzed and provedfor a more proper control variable selection. And then the optimizationmodel is built, which takes VCA-RPR as objective and conventionalphysical limit as constraints, such as and bus voltage limit and generationlimit. A successive regional optimization algorithm is developed, whichselect the key areas and optimize VCA-RPR of these areas respectively.The optimization model and algorithm are validated by numerical resultson IEEE118system. After optimization, VCA-RPR of key areas isincreased and voltage stability margin is thus greatly enhanced, the voltageprofile is also improved. Only a few decoupled sub-problems with lowdimension are needed to be solved in the proposed regional optimization,which greatly reduce the problem scale and complexity, and thus themethod has high computational efficiency and fault tolerance. Numericalresults on a real2123-bus system show that the proposed method can beapplicable to large scale systems.