Оптимизация параметров лазерного легирования стали 30ХГСН2А с использованием генетического алгоритма и нейросетевого моделирования

The process of laser alloying of structural steel with chromium has been optimized using a genetic algorithm and neural network modeling. Finite element modeling of the process of laser alloying of 30ХГСН2A steel was performed in the ANSYS Workbench software environment taking into account the temperature dependence of the thermophysical properties of materials. A face-centered version of the central compositional design of the experiment was used to construct a surrogate model of laser alloying. Time intervals corresponding to the durations of three laser pulse fronts and the peak power densities of these fronts were selected as experimental factors. Maximum temperatures in the processing zone were analyzed as responses. The process was optimized taking into account the limiting values of the maximum temperatures in the processing zone for three moments in time corresponding to the ends of the laser pulse fronts, at two points of the finite element model. The parameters obtained as a result of optimization were compared with the parameters calculated based on the finite element modeling data. The maximum relative error of temperature values obtained using the genetic algorithm was no more than 4.0 %.