Control strategies for energy harvesting systems

This thesis presents the development and evaluation of advanced control strategies for energy harvesting systems, focusing on vehicle regenerative suspensions and wave energy converters. The main goal is to design controllers that maximize energy collected from external disturbances, like road profiles or ocean waves, while adhering to physical and operational constraints. A central challenge addressed is the design of effective causal approximations for optimal control solutions, which are often anti-causal because they depend on the future of external signals. To achieve these objectives, the thesis develops several novel control frameworks. These include H∞/H2 output feedback controllers that use the Maximum Induced Power Control (MIPC) strategy; a new multi-objective Linear Quadratic Regulator (LQR) that incorporates an H∞ formulation to enforce a bound on the Ride Index (RI); and Economic Model Predictive Control (EMPC). The EMPC framework is explored in two forms: a predictive controller for wave energy converters that relies on wave forecasts, and a stochastic version for regenerative suspensions that operates without needing a road preview. The investigation of these strategies leads to several key findings. The H2 MIPC controller proved more effctive than its H∞ counterpart, and the Inerter Pendulum Vibration Absorber (IPVA) was identified as the best actuator for regenerative suspensions. The new LQR strategy significantly increased harvested energy, showing improvements of 37.3% and 27.8% over the regenerative damper and H2 MIPC state feedback methods, respectively. This LQR approach also has lower computational overhead, making it well-suited for real-time implementation. Furthermore, EMPC strategies achieved the highest levels of energy harvesting and superior constraint satisfaction for both regenerative suspensions without road preview and wave energy converters using wave prediction. All developed control strategies were rigorously validated through extensive numerical simulations and comparative analyses.