Research On Multi-level Combined Fuzzy Comprehensive Evaluation For Alternative Transmission Network Planning

Although there are numerous alternative transmission network planning schemes, experts and decision-makers may lack an objective standard to make a subjective evaluation, and the traditional evaluation which maximizes the technical and economic index is not comprehensive, so it's necessary to study the comprehensive evaluation which can deal with multiple constraints effectively. Besides, there are many defects in traditional evaluation such as unscientific index system, randomness of the choice of evaluation method, unreasonable index weight and much randomness of the process and so on. This thesis carries out the research on comprehensive evaluation aimed at solving the problems above.Firstly, based on the existing index system from “Power Grid Planning Index System Reasearch Report of Guangdong Grid Co”, a set of comprehensive evaluation index system is established to consider characteristics of the grid planning based on the background of the project. The index system can meet various requirements of the transmission network planning. It is divided into three levels in total, including 6 first grade indexes, 16 second grade indexes and 60 third grade indexes. At last, a detailed analysis of the significance?calculation formula?original data and evaluation criterion of each index is presented in this paper.Secondly, an optimization model which combines expertise with data information is set up to give the indexes reasonable weights. Firstly, the advantages and disadvantages of subjective weighting method and objective weighting method are analyzed in this paper. Secondly, based on the theory of mathematical statistics and linear algebra, the optimization model which is aimed at minimizing the deviation between the combination weights and subjective/objective weights is established in hyperspace and it can give the indexes reasonable weights. Finally, in order to check the rationality of the index weight, the average closeness is used in the paper.Thirdly, considering the fuzziness and hierarchy in transmission network planning, Multi-level combined fuzzy comprehensive evaluation is put forward in this paper: firstly, for the sake of giving the indexes reasonable weights, a transformation rule converting the fuzzy relation matrix into index data matrix is built in this paper. Then, in order to quantify the fuzziness in grid planning reasonably, fuzzy relation matrix is obtained by establishing an optimization model which is aimed at minimizing the deviation between the combination membership vector and single membership vector. At last, considering the validity of the maximum subordination principle, the weighted average fuzzy operator is used to evaluate the projects of transmission network planning scientifically.Fourthly, in order to give indexes weights and quantify the fuzziness reasonably, an improved particle swarm optimization algorithm(IM_PSO) is developed to solve the optimization mathematical model. And Lots of methods are put forward to improve PSO's convergence performance, and the principal aspects are as follows: the strategy of self-adaptive inertia weight; the nonlinear strategy of acceleration coefficient; the consideration of the information of the closest particle; the use of chaotic sequences to initialize the evolutionary population; the introduction of gauss mutation operator and so on. The simulation results show that the IM_PSO has better behaviors in convergence performance, computational stability and skipping from Local optimization.At last, an empirical study on three transmission network planning schemes of Guangdong Power Grid from “the 12 th five-year power system plan” is carried out and the evaluation results verify the validity and reasonability of the evaluation model.