In this study, according to the experimental results related to the dissimilar laser welding of brass-stainless steel 308, a performance approximation method called artificial neural network (ANN) was used. Welding speed, focal length, peak power, pulse width, and frequency were taken as input parameters, and temperature and melting ratio were considered as target characteristics. The ANN results were compared with the experimental results and the error percentage between them was obtained. Maximum relative errors were 9.63%, 10.55%, and 6.13% for brass alloy temperature, stainless steel, and melt ratio, respectively. Based on this comparison, the percentage of error between the experimental data and the ANN was at a reasonable level; so, this numerical method could be used with low time and cost. Also, by considering seven and five neurons in the hidden layer, the lowest mean squared error was obtained for temperature and melting ratio, respectively.

中文翻译：

---

基于人工神经网络模型的黄铜和 SS 308 异种激光焊接中激光参数对熔化率和温度分布的影响

在本研究中，根据与黄铜-不锈钢 308 异种激光焊接相关的实验结果，使用了一种称为人工神经网络 (ANN) 的性能近似方法。焊接速度、焦距、峰值功率、脉冲宽度和频率作为输入参数，温度和熔化比作为目标特性。将人工神经网络的结果与实验结果进行比较，得到它们之间的误差百分比。黄铜合金温度、不锈钢和熔体比的最大相对误差分别为 9.63%、10.55% 和 6.13%。基于这种比较，实验数据与人工神经网络之间的误差百分比处于合理水平；因此，这种数值方法可以以较低的时间和成本使用。还，