基于人工神经网络的船舶高强钢焊接变形分析预测

以新型船舶高强钢为主要对象 ,在物理模拟试验和有限元分析基础上 ,通过人工神经网络建立船舶分段钢结构焊接变形数值模拟模型 ,仿真焊接工艺参数、板材厚度、焊接顺序等因素对船体钢结构变形的影响 ,探索抑制船舶高强钢焊接变形的有效途径。仿真结果表明 ,在各种因素中组焊顺序是影响焊接变形的主要因素 ,通过选择小电流同向分段退焊工艺将有效抑制船体分段钢结构焊接过程中产生的变形 ;采用上述方式建立的模型可以快速预测、预报船舶高强钢焊接过程中产生的变形量 ,为新型船舶研制、生产提供真正的理论指导 ,避免焊接试验的盲目性

The Prediction of Welding Deformation for the Marine High-Strength Steel Based on Artificial Neutral Network

This paper based on the physical simulation and finite element analysis (FEA) for the marine high strength steel,the numerical simulation of the welding deformation for the marine block steel structure was established by artificial neutral network (ANN).Through simulating the welding parameters,the thickness of armor plate and the welding sequence,the influencing factors to marine steel structure were studied.The valid path of controlling the welding deformation for marine high strength steel was attained.The simulation results showed that in all kinds of factors,the welding sequence was the main factor to influence the welding deformation.By choosing little electric current of the same direction backstep welding,the deformation in the welding process of the marine block steel structure was controlled.By above method,the deformation of marine high strength steel was predicted rapidly at the welding process.It really offered a theoretic direction for the marine development and manufacture,avoided the blindness of the welding experiment.