Integral Sliding Mode Observer based Finite Set Model Free Predictive Current Control of Plug-in Fuel Cell HEV with Common Mode Voltage Mitigation

Abstract

Rapid increase in prices, decrease in resources of oil and natural gas, increasing global warming and pollution have shifted focus of the community towards plug-in hybrid electric vehicles. Plug-in fuel cell hybrid electric vehicle utilizes a fuel cell as the main power source while battery and ultra-capacitor as auxiliary power sources. Plug-in mains are used to recharge the batteries. The robustness of the model predictive control of induction machine can be increased by using the ultra-local model as it expresses system output in term of its inputs without requiring any specific information about the plant. Observer-based controllers utilize available sensors to estimate state variables, eliminating the need for all sensors. Adaptive backstepping sliding mode control technique is employed to generate control laws for DC-DC converters for fuel cell, ultra-capacitor, battery and plug-in mains. A finite set model predictive current control of induction machine has been proposed as it has a fast response, less computation and has a relatively simpler controller design. Neutral point clamped inverter (three level inverter) has been proposed to mitigate common mode voltage generated at the motor bearing. An integral sliding mode observer has also been proposed to estimate the stator current and unknown vector variable of the ultra-local model. Lyapunov’s theory is employed to ensure the asymptotic stability of the proposed system. The control scheme has been validated in MATLAB/Simulink (R2022b) using the ODE-45 solver which shows off the comparison of the proposed controller with conventional model predictive control, two-level inverter, backstepping controller and its variants. The results indicate that the proposed controller outperforms the rest of its counterparts in tracking, fast convergence and reachability and exhibits better dynamical performance.