Lessons learnt from thermo-mechanical feasibility assessment of pavement solar collectors using a FE-ANN approach

This study presents a feasibility assessment of harvesting heat from asphalt pavement and a structural performance evaluation of a Pavement Solar Collector (PSC) in Utrecht, The Netherlands. The potential for interface failure was evaluated through experimental tests to determine the yield shear stress/strength ratio (YR) parameter. Subsequently, pavement responses were computed for two scenarios (with and without PSC) to evaluate damage factors using a bottom-up fatigue cracking approach. This study introduces a hybrid approach, combining Finite Element (FE) analysis with Artificial Neural Networks (ANN) to significantly reduce the computational time required for thermal FE simulations. This FE-ANN prediction model is employed to conduct parametric studies on three key input parameters of PSC systems: pipe embedment depth, inlet supply temperature, and pipe section length. The structural response analysis showed that the YR remained below 0.25 for temperatures below 32.5°C, peaking at 0.68 for 45°C, indicating a minimal risk of interface shear failure. Cumulative damage factor calculations for the two scenarios showed that operating the PSC system between May and September could yield an 18% reduction in overall damage. It was also established that pipe section length exerted the most profound influence on the thermal responses of the PSC among the three selected parameters. The investigation revealed that a PSC system's total annual heat harvesting capacity could achieve a noteworthy 1.17 GJ/m², accompanied by an average outlet water temperature of 15°C. Moreover, an increase in pipe section length from 50 to 300 meters resulted in a significant decrease of up to 50% in the total annual heat harvest capacity of the PSC system while simultaneously causing the outlet temperature to rise to 30°C.