A novel one-time unified configuration strategy with L-SHADE optimizer for enhanced power extraction in thermoelectric generator array

The thermoelectric generator (TEG) is a promising green energy source, converting wasted heat into electric power without relying on chemical reactions or emitting harmful substances. However, a significant challenge obstructing its practical deployment is the power losses due to nonuniform temperature distribution (NUTD) within its modules, particularly in applications like solar-TEGs and automotive systems. Studies in the literature have addressed this challenge by proposing dynamic reconfiguration approaches (DRAs) for TEG arrays to mitigate power losses under NUTD conditions. However, the main issue with these DRAs is their reliance on numerous switches and sensors, which increases the system's cost and complexity and necessitates regular maintenance. To provide a cost-effective solution, this paper introduces an innovative approach to designing a One-time Unified Configuration (OUC) for TEG arrays as an alternative to the standard series–parallel (SP) arrangement. This approach formulates an optimization problem to establish OUC as a unique layout for the TEG array under multiple NUTDs. The L-SHADE optimizer is tailored to determine the OUC layout that fits multiple NUTDs simultaneously. The yielded OUC ensures a consistent temperature distribution among the parallel strings of the TEG arrays, thereby maximizing the overall power output. The proposed OUC is rigorously implemented and tested on both symmetrical (9 × 9) and asymmetrical (10 × 15) TEG arrays and compared with DRA. Additionally, it is validated through a hardware setup involving a (5 × 5) TEG array under varying NUTDs. The results demonstrate a remarkable enhancement in TEG array output power with gains exceeding 10% compared to the standard SP configuration. Furthermore, the OUC eliminates the need for switches and sensors, resulting in cost savings of approximately 1498.5$ and 2775$ for the (9 × 9) and (10 × 15) TEG arrays, respectively, compared to DRA.