中文摘要：

以船用高强钢的T 型接头为研究对象,在物理模拟试验的基础上,基于人工神经网络调用MATLAB软件的工具 箱建立船用高强钢T 型接头角变形B P神经网络模型,仿真焊接工艺参数、底板厚度、焊接顺序等因素对角变形的影响, 探索抑制船体钢结构焊接角变形的有效措施.结果表明该方法可以快速预测、预报船舶T 型接头焊接过程中产生的角 变形量.测试结果与仿真结果之间的偏差较小,采用分段退焊的角变形量最小, T 型接头焊接角变形随焊接电流的增大 而增大,随底板板厚增加而减小.

英文摘要：

bBased on the physical tests of marine T- joint,the back-propagation network was built by using Artificial Neural Network （ANN） tool box of MATLAB software during CO 2 welding. The influence of welding parameters, plate thickness and welding sequence on angular deformation was simulated. The valid path of controlling the weld-ing angular deformation for marine high-strength steel was attained. The results show that the method can predict and forecast the angular deformation of the T-joint in the ship. The deviation between the test result and the simula-tion result is small, the angular deformation of the segmented welding is the smallest, and the angular deformation increases with the increase of the welding current and decreases with the increase of the plate thickness.