Study Of Multivariate Statistical Analysis Application For External Insulation Selection

Hydroelectric resources are abundant in Western China. With the execution of Development of West China and West-to-East Power Transmission, massive power is required to be delivered from western region to eastern industry center. Transmission lines will run through some complex circumstance regions. In these regions, meteorologic condition is quite different from that of plain regions. Accumulation of ice and snow, low atmospheric pressure, natural and industrial contamination has obvious effect on the external insulation of transmission lines. At present, most transmission lines in service in home and abroad run trough low altitude and light icing regions. So there are few design and service experiences for external insulation design in high altitude, icing and polluted regions. As a result, the selection of external insulation in complex circumstance regions is an urgent problem to be solved. The dissertation studies the application of Multivariate Statistical Analysis on the selection of external insulation in complex circumstance regions. The purpose of the dissertation is to explore a new way of selecting external insulation. The paper, for the first time, applies two methods of Multivariate Statistical Analysis---Principal Component Analysis and Support vector Machines to select external insulation in complex circumstance regions. Based on Principal Component Analysis, the characteristic indexes of multi influencing factors that influence the performance of insulators are analyzed. The comprehensive evaluating mathematic model of insulator is built. The performance of several types of insulators under multi influencing factors is comprehensively evaluated. And the performance of insulators in complex circumstance regions is comprehensively evaluated by the mathematic model. Based on SVM, using the characteristic indexes of atmospheric pressure, contamination and icing as inputs, the mathematic model that evaluating the performance of external insulation is built. And the performance of insulators in complex circumstance regions is evaluated by the mathematic model. Based on SVM, using the characteristic indexes of atmospheric pressure, contamination and icing as inputs, the mathematic model that forecasting the flashover voltage of insulators is built. And the flashover voltage of insulators in complex circumstance regions is forecasting by the mathematic model.