Performance Improvement of Low-Temperature Thermal Energy Conversion Systems via Physics-Based Control and Optimization

Low-Temperature Thermal Energy Conversion Systems (LT-TECS) are critical for recovering low-grade waste heat from geothermal and industrial sources however, their practical deployment is constrained by strong nonlinear system behaviour and the absence of adaptive, physics-informed control strategies. This study presents a physics-based, simulation-driven control and optimization framework aimed at improving the system-level performance of low-temperature thermal energy conversion systems operating below 200°C. A first-principles numerical model is developed to capture the coupled thermal electrical behaviour of the system over thermal source temperatures ranging from 50 to 200°C, heat fluxes between 0.5 and 5.0 kW m⁻², and electrical load values from 0.5 to 3.0. Performance prediction surfaces indicate electrical power outputs varying from approximately 300 to 2600 W and conversion efficiencies between 1% and 14%, highlighting strong sensitivity to operating conditions. A surrogate-assisted, physics-based optimization strategy is employed to construct control maps and implement a self-learning adaptive control loop. Time-domain simulations over training periods of 24 to 96 h demonstrate consistent performance improvements under optimized control. Average electrical power output increases by up to 2.2%, while conversion efficiency improves by approximately 0.7 percentage points compared to conventional control, accompanied by a 25 to 40% reduction in load and heat-flux fluctuations. The results confirm that physics-based adaptive control and optimization can deliver measurable and stable performance gains without hardware modification, addressing a key gap in low-temperature thermal energy conversion system operation. The proposed framework provides a transferable and cost-effective solution for improving the energy yield of existing LT-TECS infrastructure.