Study Of Intelligent Decision Support System And Its Algorithm For Distribution Network Planning Based On GIS

Distribution network planning (DNP) work complexity manifests dominatingly on large-scales, multi-factors of uncertainty and inaccuracy, involving wide domains and so on. Only depending upon planning-workers' planning experience, it is unable to be satisfied for the requirement of modern DNP. So exploiting a set of intelligent decision system could not only relieve labor intensity of planning-workers improve scientific decision level of DNP, and it inevitably has important academic and realistic meaning. Introducing geographic information system (GIS) platform into DNP system could largely be convenient for planning-data collection and management, it would improve planning workers labor productivity and make planning process more visible and interactive. This paper has discussed the fundamental principle of intelligent decision support system (IDSS) for DNP based upon GIS and its main structure scheme, and emphatically studied GIS methods applying to DNP. In a word, this paper main work is as following:(1) Through analyzing the fundamental principle of decision support system, the universal structure scheme of IDSS for DNP based upon GIS is put forward. The system mainly includes sub-systems of history information analysis, load forecasting integrated analysis, substation planning, distribution network structure planning integrated analysis and so on. Established on GIS platform such system is achieved by utilizing object-oriented technology and covers database, model base, knowledge base and figure base. The system has the advantage of full visible information and powerful interactive function, and reveals the combination uniform of theory, computer tool and expert experience, and of scientificity and practicability.(2) Distribution network GIS database is the foundation of IDSS for DNP. Through the characteristic analysis of distribution network, the GIS data model of DNP is presented by vector model. Processing object is divided information figure-layer spatial by sub-layers and sub-processes, the data organization method and establishing flow is put forward for DNP GIS database. And the data origin and data warehousing of GIS for DNP are also introduced. (3) Distribution network load forecasting is a very important work for DNP IDSS. Using grey forecasting method to study load gross forecasting, aiming at the defect of grey forecasting model, the background modification method for grey model is put forward to improve the precision of load forecasting. Applying GIS to distribution network spatial load forecasting which makes the forecasting process operable. Utilizing GIS graphic platform to collect each type of spatial attribute information in forecasting region and revealing direct-view forecasting results in GIS platform. Meanwhile knowledge base reduction method in rough set theory is introduced to do a reduction for community attribute, the fuzzy inference regulation is established to regulate the support of land use and overcomes the shortcoming that experts design the fuzzy regulation, and could improve the accuracy of spatial load forecasting.(4) As substation addressing model and algorithm is a crucial link to IDSS for DNP based on GIS, the substation spatial layout model is set up and computed by multi-sources alternate iteration addressing method, and the standby address of substation is obtained. Using fuzzy analytic hierarch process (AHP) to blend the support-level data, and could choose the optimal substation address. In the process of choosing address, utilizing GIS data management and spatial analysis function to realize addressing needed multi-data visible, it will present more comprehensive guideline characteristic value for addressing work and largely enhance the efficiency of substation addressing.(5) For the method of distribution network structure optimizing planning is a critical component of IDSS for DNP based on GIS, the author studies the genetic algorithm and immune algorithm, and put forward a multiple population immune genetic algorithm (MPIGA) to distribution networ...