New AI tool speeds up design of efficient heat-to-electricity generators (www.techradar.com)

Researchers in Japan have developed an innovative AI tool, TEGNet, which dramatically accelerates the design process of thermoelectric generators (TEGs) by reducing simulation times from thousands of seconds to mere fractions of a second. This breakthrough enables predictions of generator performance with over 99% accuracy while using only 0.01% of the computational resources required by traditional finite-element solvers. TEGs, which convert temperature differences directly into electricity without moving parts, have the potential to efficiently utilize waste heat from various sources like factories and vehicles but have been limited by high costs and moderate performance. The significance of TEGNet extends beyond speed—it has the potential to reduce manufacturing costs and open new avenues for thermoelectric applications. By identifying designs that require simpler fabrication and reducing reliance on expensive materials like bismuth telluride, TEGNet could make waste-heat harvesters and advanced heat pumps more affordable and accessible. While there are inherent thermodynamic limitations to heat-to-electricity conversion, this AI-driven approach signifies a notable advance, promising to bring TEGs closer to industrial competitiveness for the first time in history.