中文摘要：

传统的模型预测控制（model predictive control．MPC）采用基于脉冲响应的非参数模型作为预测控制模型，计算复杂，很难直接应用到快速的实时控制系统领域。该文提出了一种改进的MPC，并将此改进的MPC技术成功地引入到电流型PWM整流器（current source PWM rectifiers．PWM—CSR）的功率因数校正系统中。改进的方法从PWM—CSR控制量与被控制量的传递函数入手，得出整流器的一阶差分方程作为预测控制模型，同时省略传统模型预测控制（MPC）中参考轨迹这一不确定因素，因而克服了传统MPC计算量大，实现困难的缺点，并保留了传统MPC反馈校正、滚动优化的优点。在用DSP TMS320LF2407A为开发平台的实验样机结果表明：改进的MPC用于PWM—CSR的功率因数校正系统中，具有优良的稳定控制性能和快速的动态响应性能。

英文摘要：

In traditional model predictive control （MPC）, non-parameter model based on pulse response is usually adopted as predictive control model, which results in complex computation, so this method is difficult to be applied to real-time control system directly. An improved MPC method was presented and applied to power factor correction system of current source PWM rectifiers （PWM-CSR）. In the proposed MPC, the first-order differential equations were derived from the transfer functions of controlling variables and controlled variables in PWM-CSR, and the equations were used as predictive control model, meanwhile, the uncertain factor reference trajectory was left out. Consequently, considerable computation and inconvenient realization of traditional MPC were avoided and the merits of feedback correction and dynamic optimization of traditional MPC were reserved in the new MPC. Experimental results based on DSP TMS320LF2407A show the satisfied steady performance and fast dynamics characteristics are obtained in the PWM-CSR power factor correction system with the proposed MPC.