Research On Voltage Predictive Control Algorithm Of Regional Power Grid Based On Reduced Model

Voltage stability problems of regional power grid are important researchcontents to ensure the security and stability of power system. Generally, layeredcontrol mode is widley used in our country, which includes primary voltage control(local control), secondary voltage control (reginal control) and tertiary voltagecontrol (global control). In the study of regional power grid voltage control, thecontrol methods are usually based on the constraint conditions of power flowcaculations, which can not consider the dynamic behavior of the power system. Inthis paper, model predictive control method is applied to the regional power gridlayered voltage control in order to enhance the global dynamic stability performanceof regional power grid. Thus, the paper contained the following research:Firstly, establishing the layered nonlinear power system mathematical model ofthe regional grid is the groundwork, which included real power system model(primary voltage control layer) and voltage predictive control model (secondaryvoltage control layer). Secondly, applying the voltage predictive control model andpredictive control theory, we formed the regional power grid voltage multi-steppredictive control optimization problem via the sample variables of real power gridmodel using WAMS. It took minimizing the deviation between the output of thesystem and its reference as optimization objectives, nonlinear prediction equations ofsystem as constraint conditions and system state variables and control input ranges asun-equivalent constraint conditions in order to optimize the reference voltages oflocal system controllers. Thirdly, using the reference voltages from the secondaryvoltage control layer, we completed the layered regional power grid voltagepredictive control. We took advantage of a4-machine power system to verify theeffectiveness of the predictive control methods compared with traditional control.Finally, for high dimensions of nonlinear dynamic model problem, we studied thevoltage predictive control algorithm of regional power grid based on balancedreducted model. This algorithm realized multi-step prediction, considered theconstrains of states and inputs and shorten the optimization calculating time bybalanced reduction for high dimensional nonlinear dynamic model.This paper used the online interior point numerical method to solve the aboveoptimization problem. We programmed on MATLAB platform and took IEEE testsystem model to verify the effectiveness of the predictive control methods.