Research Of The System Reliability Varying With Load Levels And Reliability Parameters In Composite Generation And Transmission Systems

This thesis was supported by the National Natural Science Foundation of China ("Research on the theory and application of Reliability Tacing in Power System", No.50577072).The system demands and reliability parameter of components have significant impacts on the reliability of power systems. The bus and system reliability indices are obviously functions of multiple variables with the random characteristics, such as the system load levels, the reliability parameter of components and the structhure of network. It's very difficult to accurately evaluate the system reliability using a single load level and constant reliability parameter of components. In many situations, such as the reliability optimization, the proposed schemes of a power system are usually related to several repeated reliability evaluation processes due to different reliability and electrical parameters. Hence, it's necessary to discuss the rules of the system reliability varying with load levels and reliability parameters, which can be used to directly evaluate the system reliability without a repeated calculation when the parameters are changed. The thesis presents regression models of the system reliability varying with load levels and reliability parameters on the base of a comprehensive analysis of variations of the system reliability with these parameters. The main contents are as follows:â‘ The model of the generation system reliability varying with different load levels and generator reliability parameters is proposed.The model for the generation system reliability varying with different load levels is formulated as a quadratic regression model with one variable using the variable transformation technique, which is helpful to show an approximate linear ralationship between the reliability and load levels. The original model for the system reliability varying with load levels can be obtained using the inverse transformation of variables. The proposed regression models have been verified to be highly significant based on the significance tests on some cases. Similarly, a nonlinear regression model with multiple variables for the generation system reliability varying with generator reliability parameters is also presented. The two models not only indicate the variation rules of the generation system reliability with load levels and generator parameters, but also can be used to forecast the reliability of generation system.â‘¡The model of the system reliability varying with load levels in composite generation and transmission systems is presented.Whereas the complex influence of load levels to system flow, the curves of reliability indices varying with load levels at single bus in composite generation and transmission system are analysed. And they indicate that the reliability level desend like a ladder with increace of bus load. Based on this case, the piecewise regression model of system reliability indices about single bus load is proposed. The analysis of break points indicate that their primary causes of formation are the discontiuity of system available capacity and constraint of line capacity. When load exceeds certain available capacity level by the square, system state is transferred to energy insufficient from energy adequate. This kind of step change is more obvious when there are many generators with the same capacity. At last, the piecewise regression model about all bus loads is conducted based on the model about single bus load. The test of significance indicates the regression equations are very siginificant.â‘¢The model of the system reliability varying with the reliability parameters in composite generation and transmission system is presented based on the stepwise regression technique and orthogonal design theory.If the model is built derectly by multiple linear regression analysis, the regression equations will include too much variables. And those variables aren't significant to system reliability not only increase the complexity of regression equations, but also influence the stability of regression equations. On the other hand, the amount of sample increases is exponential function of component amount. So the regression analysis is difficult to carry out. Whereas, the model of composite generation and transmission system reliability varying with component reliability parameters is proposed, which integrates stepwise regression and orthogonal experiment design. The stepwise regression could eliminate variables which are not significant to system reliability. The orthogonal experiment design could reduce amout of sample largely. The caculation result to RBTS system indicates that the model is still accurate while the amount of sample is reduced largely. The model could be applied to sensitivity analysis of components.