16 January 2026 in Measurement Science and Technology
To address the high-precision thermal control requirements of semiconductor lasers in laser soldering applications, we propose a thermal control method based on long short-term memory model predictive control (LSTM-MPCs). First, we constructed precise mathematical models for the thermoelectric cooler and components of the thermal control system using non-equilibrium thermodynamics and other methods. The LSTM-MPC thermal control system was derived by introducing the diophantine equation, and an experimental platform for the thermal control of semiconductor lasers in laser soldering was established. The experimental results from random current input variations showed that the average temperature overshoot caused by changes in the laser input current remained within 0.86%. The stability of the laser’s central wavelength improved, with the proposed method reducing the central wavelength drift by 88%. In laser soldering performance experiments, this control method effectively controlled the output power of the semiconductor laser, enhancing the linear fit of the weld point temperature curve to the ideal temperature curve by 21.05%. These findings provide theoretical and engineering guidance for the development of high-performance, reliable laser soldering systems.