Trapezoidal Variable Weights Regression Tracing Control of MCSR in Cascading Failures

摘要

A trapezoidal variable weights regression control algorithm was proposed to proportionately calculate and change the weights of all variables in the rolling time domain window while tracking sensitivity and robustness were available. The operation principle of magnetically controlled shunt reactor magnetic circuit was analysed, and the weighted extensional model in time-varying parameters logarithmic regression function was established considering the dynamic relationship between the nonlinear component and the external power system according to theory of trend extrapolation. A control strategy of variable weights to reflect their importance in system identification was proposed to optimise voltage stability control. The algorithm has no iteration which can save computing time and memory and increase control flexibility. Advanced Digital Power System Simulator (ADPSS) which was developed by China electric power research institute contains the controller model by programmed. Simulation results in cascading failures used in practical operation parameters in projects confirmed the effectiveness of the control algorithm.