The Optimal Planning Of Urban Mid-Voltage Distribution Network

Urban mid-voltage (MV) distribution network planning is a complex NP-hardoptimal problem with characteristics such as multi-objects, uncertain, non-linear andmulti-stages. Geographical Information System (GIS), intelligent optimizationalgorithm, reliability worth theory, multi-objective and uncertain planning theory areadopted in this paper to do the research deeply as follows.Under the condition of different load density of power-supply areas and differentsubstation capacity, economy and reliability of some familiar connection modes fordistribution systems are calculated and analyzed to get their trends with load densityand substation capacity varied. The difference between these connection modes underthe same condition can also be known from the comparison result. And based onquantitative analysis and synthetic consideration, some useful schemes are proposed.Considering the specialties of distribution network spatial data structure, thispaper designs GIS spatial database of distribution network by applying the 'layer'concept of GIS and provides credible assistant design method for urban MVdistribution network planning.Shortest path method is used to produce preliminary radial network, and GeneticAlgorithm (GA) is used to optimize the preliminary network. As the different-lengthcharacter chromosome encoding strategy based on spatial GIS is adopted, distributionnetwork planning is really combined with geographic conditions. Through analyzingthe example, it is proved that the idea of combining network optimization problemwith graphical problem based on spatial GIS has succeeded in distribution networkplanning.To construct the urban MV distribution network from the global situation, thispaper puts forward a method of normal and reverse radial lines to realize automaticrouting of the link and loop network based on the automatic routing of radial network.That is, while routing the radial lines, consideration is also given to the tie lines at thesame time to better meet the demand of the overall planning. Then based on this, theautomatic routing of a three-sectioned and three-linked network is developed, withthe general planning for the automatic routing of the multi-sectioned and multi-linkednetwork proposed.Based on the spatial GIS of distribution network, the optimum number andlocation of switching stations are determined by adopting the method of substationlocating and capacity calculating. And the automatic routing of the incoming andoutgoing lines of switching stations is developed based on the automatic routing ofmulti-sectioned and multi-linked network.Tabu Search (TS) algorithm is also used to optimize the urban MV distributionnetwork in this paper. The advantage of adopting TS is that the feasible final networkis satisfied with the radial routing trait and the planning process is shortened.By comparing the different abilities when solving TSP by GA, TS and the hybridalgorithm of the two, this paper proves the advantages of improved GA (Genetic TabuHybrid Algorithm, GTHA) which adopting the TS mutation operation when solvingthe combination optimization problems. While optimizing the actual urban MVdistribution network using GA, TS and GTHA respectively, the comparison resultsshow that the search efficiency of GTHA is much better than the other two andGTHA proposed in this paper has a promising application.Then GTHA is used in optimal sectioned switches placements of radial lines indistribution network systems. The function of the model combines the annualinvestment and maintenance costs of sectioned switches with the customer outagecost which transforms reliability benefit to economy index in order to arrive at anoptimum decision. The algorithm described in this paper is testified by a simpleexample, and is suitable to be used in actual distribution systems.The fuzzy technology is used to make the economy and reliability indexes fuzzyin order that they have the same statuses. Based on this, the multi-objective model ofurban MV distribution network planning is designed. And the novel multi-objectiveGTHA is put forward with the selection operation applying the hybrid method, whichconsiders the traits of multi-objective optimal problem and includes paratacticselection method, elitist preservation strategy and sharing function method.Interval algorithm is adopted to express the uncertain load and modify theoriginal mathematic model and certain planning method. While the improved methodis compared with the certain planning method by applying in the same actual case, itis testified that the method described is suitable in the urban MV distribution networkplanning considering load uncertainty.With the general trends of electric power market innovation in China, theorientation, theory and evaluation rules of power system planning in the competitivemarket should be developed. The power generation planning theory of electricgeneration companies is suggested in this paper, also with the planning theory ofdistributed generations. And the uncertainties that the transmission-distributionnetwork planning faced and the new planning theory and methods under the marketare also put forward concretely.