Study On Load Model Identification And Load Characteristics Classification And Synthesis

Power load model has a great impact to power system analysis, computing, planning, operating and control, but because of the random time-variation, composition diversity, geographic dispersion and severe non-linear characteristics of the composite load, the load modeling is still a recognized problem of electric power system both at home and abroad. In this paper, a number of issues of load modeling are studied.Adequate data sources are the basis for load modeling. In this paper, several ways to obtaining the required data of load modeling are compared, including Supervisory Control and Data Acquisition (SCADA), Wide Area Measurement System (WAMS), a special load characteristic measuring device, fault record and monitor system and so on. With the development of fault recording technology, it has many advantages for load model identification using fault record data, such as low investment, the full use of a large number of fault recorder data, and so on.The selection and evaluation of load model is very difficult. In this paper, several common load models are identified and compared, including synthesis load model, linear differential equation model,50% constant impedance+50% motor model,4-6 models and 3-3-4 models. And through the EPRI-36 bus system, the result is illustrated the advantages of the synthesis load model and the problems of current load model.A new method of "anti-perturbation method" is used to analyzing the parameters sensitivity of the synthesis load model. The principle of this method is that a parameter is fixed for a typical value and other parameters are identified, and then according to the fitting results of the model response and measured response to determine the sensitivity of this parameter. Finally six parameters of Xs, Rr, Kpm, Mlf,PV, QV come to have greater sensitivity.Classification and synthesis for load characteristics is the main ideas to solve the random time-variation and geographic dispersion characteristics problem of the composite load. In this paper, a new method is presented to solve these problems, this is Classification and Synthesis of load characteristics based on an Improved Fuzzy C-means clustering algorithm. First of all, the fuzzy C-means clustering algorithm has been improved to make it applicable to time-series data clustering problem, and then it is used to cluster with the measured response as the feature vector. Finally, the general load model of this Category is obtained by the direct identification to the cluster center. At last this method is verified through the EPRI-36 bus system, the simulation results show that this method is convenient, fast, accurate, practical and effective.