3-D Unbalanced Coordinate Transformation With Super-Twisting Control and Observation for Unbalance and Disturbance Load Voltage Mitigations in Stand Alone Four-Leg Inverters

This paper presents an enhanced dual-loop voltage vector control method for standalone power supply systems using a four-leg voltage source inverter. The proposed method aims to optimize the quality of load voltage and inverter output current waveforms under unbalanced three-phase loads, enhance load voltage disturbance rejection during unbalanced load changes, improve the system’s dynamic response during transients, and increase the robustness of both the inner four-leg voltage source inverter output current and outer load voltage control loops against system parameter uncertainties, as well as to reduce the number of sensors. In the inner current control loop, a super-twisting sliding mode control combined with a 3-D unbalanced transformation is proposed. This transformation is adopted herein for its ability to transform the three-phase unbalanced inverter currents into direct current quantities in the synchronous reference frame based on the dynamic estimation of their unbalanced phases and amplitudes. On the other hand, the second-order sliding mode control is adopted for its robustness and its ability to reduce the chattering problem. Moreover, a proportional-integral controller incorporated with a super-twisting observer for load disturbance observation is designed for the outer loop not only to get rid of sensors but also to ameliorate the load voltage dynamics, which enhances the load disturbance rejection and reliability, as well as decrease the system’s cost and size. In this, the super-twisting sliding mode observer is suggested not only for the load current estimation but also to build up the current feed-forward control scheme to estimate and reject the load disturbance quickly. The most significant advantages of the designed technique over existing strategies are that it can overcome the control issue of periodic currents caused by unbalanced loads, provides high robustness against parameter uncertainties and high load disturbance rejection ability, improves the transient responses, and reduces the ripples, as well as reduces the required number of sensors. Suggested method’s effectiveness is confirmed through various real-time simulation tests under different loading conditions using OPAL-RT-OP5700.