Robust Control Strategy Of Dynamic Voltage Restorer

During power quality problems, voltage sag is the main reason for not working properly of voltage-sensitive load. Dynamic Voltage Restorer (DVR) can compensate voltage sag to normal value in milliseconds, and ensure power quality of sensitive loads. So control strategy of DVR determines the voltage compensation effect. In practical systems, due to the measurement error of power parameters (transformer noise), aging and disturbance signals, compensation devices should have certain robustness to meet the actual requirements. However, conventional DVR controller is only designed according to accuracy and system response time, and it doesn't take into account the effects on the DVR of the transformer noise in high-frequency, the perturbation of the system model (typically in the high-frequency), the load disturbance and the actuator saturation.For single-phase DVR with no transformer, dual closed-loop feedback controller is designed. The outer feedback controller is designed to enhance the system robustness, while the inner loop proportional feedback controller is designed to improve the system response time. The state equations of nominal object are formulated based on the mathematical system model. Three weighting functions required in the mixed-sensitivity optimization design are selected according to the requirements for system performance to get the outer feedback controller. Then the proportion controller is designed using capacitor current as the inner feedback. The resulting dual closed-loop feedback controller not only reduces the steady-state error, but also enhances the robustness of the system. Simulation results verify the effectiveness of the controller. In addition, loop-shaping method is used to enhance robustness of the DVR with the transformer. This method can guarantee the stability of system, and obtain trade-off between system performance and robustness. First select the pre-and post-compensation to make the singular values of the nominal controlled object become the desired open-loop gain, and then the robust stabilization controller of the normalized coprime-type perturbed system is used as the final controller. This controller enhances the robustness of the system. For the project implementation of the controller, PI controller parameters are obtained according to the controller obtained above using chaotic particle swarm optimization algorithm. The PI controller is used in DVR which meets system robustness requirement. The method provides a new approach to the solution of PI parameters in the project.