基于遗传算法优化的电弧增材再制造焊道尺寸预测模型

Prediction model of weld pass size for arc additive remanufacturing based on genetic algorithm optimization

电弧增材再制造过程中,成形件通过层层累积的方式进行制造。每道焊缝的几何尺寸对构件的成形精度有着重要的影响,是分层切片和路径规划的重要依据之一。为了研究焊接工艺参数与焊道尺寸之间的关系,文中采用正交试验法进行单道焊缝成形试验,建立了基于遗传算法的神经网络模型,可以预测不同焊接工艺参数(焊接电流、焊接速度、送丝速度)下单道成形的熔宽和余高。结果表明,焊缝宽度的预测平均误差为2.05%,焊缝高度的预测平均误差为5.09%,该模型实现了对焊道尺寸的较高精度的预测。结合多层多道试验对模型进行了验证,结果显示成形件成形良好,成形精度较高。文中建立的神经网络模型为电弧增材制造过程的工艺优化提供了依据。

In the arc additive remanufacturing process, the formed parts are manufactured by layer-by-layer accumulation. The geometric size of each weld has an important influence on the forming accuracy of components, and is one of the important bases for layered slicing and path planning. In order to study the relationship between welding process parameters and weld bead size, orthogonal experiments were used to conduct single-pass weld forming experiments in this paper, and a neural network model based on genetic algorithms was established, which could predict weld width and reinforcement of single pass forming under different welding parameters such as welding current, welding speed and wire feeding speed. The results showed that the average prediction error of weld width was 2.05%, and the average prediction error of reinforcement was 5.09%. The model achieved a high-precision prediction of the weld pass size. The model was verified by multi-layer and multi-pass experiments, and the results showed that the formed parts were well formed and the forming accuracy was high. The neural network model established in this paper provided a basis for the process optimization of the arc additive manufacturing process.