Fuzzy-granulation based controller for low distortion in PWM inverter with reduced load power

Inverters with pulse width modulation (PWM) for batteries are used in many applications including uninterruptible power supply (UPS) and energy storage system (ESS) for microgrids. According to the recommendations for energy suppliers, the delivered AC voltage must meet some requirements such as suitable voltage levels (between 92% and 105% of the nominal value according to national standard), and total harmonic distortion (THD) that has to be low (less than 10% according to the standard). Since the loads connected to the energy supplier change over time, the performance of a classical controller (designed for the nominal load power) is compromised, which is observed in simulations of the output voltage and THD of the inverter under disconnection of 90% and 98% of the nominal load. Therefore, a Mamdani-type fuzzy controller is designed, which is likely to be robust under those events. Rule base of the fuzzy controller is built by using granulation of information produced by simulation of a classical resonant-PI (R-PI) controller. It is shown that the designed fuzzy controller is able to reduce significantly THD in output voltage of the PWM inverter. Although the root mean square (RMS) value of the output voltage is slightly reduced, it still complains with the PRODIST Module-8 standard, as it is verified through simulations in MATLAB/Simulink.