Bayesian Optimization-Random Forest Model and Northern Goshawk Optimization Algorithm Predict and Optimize the Timing of Switches

In order to improve the acceleration efficiency and energy utilization efficiency of the energy recovery multistage reluctance launcher, the trigger time of each coil is set reasonably. In this article, a method is proposed to predict and optimize the <sc xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">on-off</small> timing of insulated gate bipolar transistor (IGBT) switches in the loops of the energy recovery reluctance launcher by using the Bayesian optimization-random forest model (BO-RF) and the northern goshawk optimization (NGO) algorithm. Through the orthogonal experiment method, a finite element simulation dataset of <sc xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">on-off</small> timing of all switching levels was constructed. A regression prediction model of <sc xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">on-off</small> timing and muzzle speed was established based on the Bayesian optimization-Random Forest model. Then, the <sc xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">on-off</small> timing of all switching stages was optimized according to the established model combined with the NGO algorithm. The proposed optimization algorithm was validated on the established three-stage energy-recovery reluctance launcher experimental platform. After optimization, the armature muzzle speed increased from 20.08 to 33.52 m/s, and the launcher efficiency improved from 2.40% to 6.69%. The results demonstrate that the method is well suited for optimizing IGBT switching timing and provides a valuable reference for the intelligent design of reluctance launchers, offering significant engineering application value and potential for broader adoption.