Prediction and optimization of multi-layer and multi-pass welding process parameters based on GA-BP neural network
-
摘要:
针对目前多层多道焊工艺参数的选择问题，利用遗传算法(genetic algorithm, GA)对BP神经网络(back propagation neural network, BPNN)进行优化，提出多层多道焊成形预测及焊接工艺参数优化方法，旨在为工艺参数选取提供有效指导，提高焊接生产效率及焊接质量. 首先通过分析多层多道焊图像，提出采用三次样条插值法与自适应分段法进行特征点识别，然后根据焊接顺序、焊道工艺建立焊接过程各焊道横截面积形状预测模型，运用解析法进行焊接工艺参数预测，进一步结合不同焊道工艺参数优选原则，采用改进神经网络进行焊接工艺参数优化，从而建立具有实时性的焊接工艺参数与焊缝轮廓关系模型.结果表明，该方法对多层多道焊中各焊道焊接工艺参数提供有效预测，试验结果满足实际需求，对提高焊接产品质量、简化焊接工艺参数选取具有实际意义.
Abstract:To solve the selection of process parameters for multi-layer and multi-pass welding, a strategy based on GA-BP neural network is proposed to predict the weld forming and optimize welding process parameters in the multi-layer and multi-pass welding. Firstly, by analyzing welding images, a cubic spline interpolation and adaptive segmentation methods is proposed to identify the feature points in multi-layer and multi-pass welding. Then, a prediction model for the cross-sectional shape of each weld bead during the welding process is established, and an analytic method is used to predict the welding process parameters. Further combining the principles of optimizing different welding process parameters, an improved neural network is used to optimize multi-layer and multi-pass welding process parameters, the real-time model for the relationship between welding process parameters and weld formation is established. Finally, the feasibility of the multi-layer and multi-pass welding process parameter selection strategy proposed in this paper was verified through experiments. The experiment shows that this method provides effective prediction of welding process parameters for each pass in multi-layer and multi-pass welding, and the experimental results meet practical needs.
-
-
![]()
图 1 试验平台及Q235低碳钢板示意图
Figure 1. Test platform and test plate. (a) experimental platform; (b) schematic diagram of low carbon steel plate with material Q235
![]()
图 2 试验焊道排序示意图
Figure 2. Schematic diagram of welding sequence for experimental welding
![]()
图 3 图像处理
Figure 3. Image processing. (a) original drawing; (b) traditional image processing; (c) cubic spline interpolation method; (d) adaptive segmented slope method
![]()
图 4 遗传算法优化神经网络流程图
Figure 4. Neural network flow chart optimization by genetic algorithm
![]()
图 9 相关性示意图
Figure 9. Correlation diagram. (a) relationship between welding speed and weld width; (b) relationship between welding speed and weld height
![]()
图 10 焊道面积与焊接速度和送丝速度的关系
Figure 10. Relationship between pass area and welding speed and wire feeding speed
![]()
图 11 相关示意图
Figure 11. Correlation diagram. (a) relationship between welding speed and cross section area; (b) relationship between wire feeding speed and pass cross section area
表 1 试验工艺参数
Table 1 Test process parameters
焊接速度
vh/(mm·s−1)送丝速度
vs/(m·min−1)焊接电流
I/A电弧电压
U/V3 ~ 7 3 ~ 7 89 ~ 173 19.7 ~ 22.3 
下载: 导出CSV
表 2 部分处理试验数据
Table 2 Partial processing of test data
组号 焊接工艺参数 焊缝形貌 送丝速度
vs/(m·min−1)焊接速度
vh/(mm·s−1)焊缝宽度
W/mm焊缝高度
H/mm1 3.00 3.00 7.80 2.96 2 3.00 4.00 7.61 2.50 3 3.00 6.50 6.50 2.36 4 3.40 4.50 8.267 2.36 5 3.40 6.00 7.48 2.173 6 3.80 3.50 9.10 2.70 7 3.80 5.50 8.206 2.24 8 4.20 6.00 7.167 2.30 9 4.20 6.50 6.907 2.253 10 5.00 3.00 11.933 3.44
下载: 导出CSV
表 3 试验结果对比
Table 3 Comparison of test results
送丝速度
vs/(m·min−1)焊接速度
vh/(mm·s−1)平板成形
面积Ss/mm2打底焊
面积Sp/mm2填充焊第1道
面积Sa/mm2填充焊第2道
面积Sb/mm2盖面焊第1道
面积S1/mm2盖面焊第2道
面积S2/mm2盖面焊第3道
面积S3/mm23.0 5 2.91 2.776 2.785 2.811 2.841 2.849 2.918 3.4 4 3.46 3.297 3.205 3.311 3.312 3.358 3.4 3.4 5 2.99 2.8 2.849 2.819 2.831 2.902 3.032 3.4 6 2.71 2.673 2.639 2.672 2.56 2.683 2.734 3.8 5 3.19 3.015 3.101 3.006 3.042 3.086 3.17
下载: 导出CSV
-
[1] 胡丹, 吕波, 王静静, 等. 焊缝表面缺陷激光视觉传感 HOG-SVM 的检测方法[J]. 焊接学报, 2023, 44(1): 56 − 62. Hu Dan, Lü Bo, Wang Jingjing, et al. Study on HOG-SVM detection method of weld surface de fects using laser visual sensing[J]. Transactions of the China Welding Institution, 2023, 44(1): 56 − 62.
[2] 李志强. 不同焊接工艺参数对焊缝组织的影响[J]. 焊接技术, 2019, 48(12): 20 − 23. Li Zhiqiang. Effects of different welding process parameters on weld microstructure[J]. Welding Technology, 2019, 48(12): 20 − 23.
[3] 贾剑平, 徐坤刚, 李志刚. 改进型BP网络在优化焊接工艺参数中的应用[J]. 热加工工艺, 2008(21): 98 − 100. doi: 10.3969/j.issn.1001-3814.2008.21.029 Jia Jianping, Xu Kungang, Li Zhigang. Application of improved neural network in optimizing welding parameter[J]. Hot Working Technology, 2008(21): 98 − 100. doi: 10.3969/j.issn.1001-3814.2008.21.029
[4] 张恩慧, 苟建军. 基于Matlab/GUI的SMAW焊接工艺参数优化系统[J]. 热加工工艺, 2015, 44(15): 195 − 197,200. Zhang Enhui, Gou Jianjun. Process parameter optimization system of SMAW based on Matlab/Gul[J]. Hot Working Technology, 2015, 44(15): 195 − 197,200.
[5] Chen F F, Xiang J, Thomas D G, et al. Model-based parameter optimization for arc welding process simulation[J]. Applied Mathematical Modelling, 2020, 81(5): 386 − 400.
[6] Zhang Ke, Chen Yixin, Zheng Jian, et al. Adaptive filling modeling of butt joints using genetic algorithm and neural network for laser welding with filler wire[J]. Journal of Manufacturing Processes, 2017, 30(12): 553 − 561.
[7] Shanmugasundar G, Karthikeyan B, Santhosh Ponvell P, et al. Optimization of process parameters in TIG welded joints of AISI 304L-Austenitic stainless steel using taguchi’s experimental design method[J]. Materials Today: Proceedings, 2019, 16: 1188 − 1195.
[8] 张恒铭, 石玗, 李春凯, 等. 工艺参数对自保护药芯焊丝焊接烟尘的影响[J]. 焊接学报, 2020, 41(11): 31 − 37. doi: 10.12073/j.hjxb.20200108001 Zhang Hengming, Shi Yu, Li Chunkai, et al. Effect of process parameters on welding fume of selfshielded flux cored wire[J]. Transactions of the China Welding Institution, 2020, 41(11): 31 − 37. doi: 10.12073/j.hjxb.20200108001
[9] 郭磊, 李思豪, 郭利霞, 等. 基于改进MULTIMOORA方法的PCCP焊接工艺参数优选[J]. 焊接学报, 2022, 43(3): 74 − 79. doi: 10.12073/j.hjxb.20211018001 Guo Lei, Li Sihao, Guo Lixia, et al. Optimization of PCCP welding process parameters based on improved MULTIMOORA method[J]. Transactions of the China Welding Institution, 2022, 43(3): 74 − 79. doi: 10.12073/j.hjxb.20211018001
[10] 张华军, 张广军, 蔡春波, 等. 厚板弧焊机器人自定义型焊道编排策略[J]. 焊接学报, 2009, 30(3): 61 − 64. doi: 10.3321/j.issn:0253-360X.2009.03.016 Zhang Huajun, Zhang Guangjun, Cai Chunbo, et al. Self-defining path layout strategy for thick plate arc welding robot[J]. Transactions of the China Welding Institution, 2009, 30(3): 61 − 64. doi: 10.3321/j.issn:0253-360X.2009.03.016
[11] 郑银湖, 宋永胜, 邓静. 基于simufact. welding的中厚板多层多道焊数值模拟分析[J]. 电子世界, 2021, 611(5): 95 − 97. Zheng Yinhu, Song Yongsheng, Deng Jing. Numerical simulation analysis of multi-layer and multi-pass welding of medium thick plate based on simufact welding[J]. Electronics World, 2021, 611(5): 95 − 97.
[12] 喻宁娜, 莫胜撼, 戴建树. 基于激光视觉传感的焊缝图像阈值分割法研究[J]. 焊接, 2015(5): 21 − 24, 69. doi: 10.3969/j.issn.1001-1382.2015.05.006 Yu Ningna, Mo Shenghan, Dai Jianshu. Threshold segmentation methods of weld image based on laser vision[J]. Welding & Joining, 2015(5): 21 − 24, 69. doi: 10.3969/j.issn.1001-1382.2015.05.006
[13] Hsieh, Hou, Andrews, et al. Cubic splines for image interpolation and digital filtering[J]. Acoustics, Speech and Signal Processing, IEEE Transactions on, 1978, 26(6): 508 − 517. doi: 10.1109/TASSP.1978.1163154
[14] Sagi Filin, Norbor Pfeifer. Segmentation of airborne laser scanning data using a slope adaptive neighborhood[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2006, 60(2): 71 − 80.
[15] 周方明, 孙红辉, 宋辉, 等. CMT窄间隙多层多道焊接工艺参数对焊缝成形的影响[J]. 江苏科技大学学报(自然科学版), 2011, 25(2): 126 − 130. Zhou Fangming, Sun Honghui, Song Hui, et al. Effect of CMT multi-layer narrow gap welding parameters on weld formation[J]. Journal of Jiangsu University of Science and Technology(Natural Science Edition), 2011, 25(2): 126 − 130.
[16] 陈振款, 何建萍, 李芳, 等. 基于BP神经网络薄板P-PAW搭接的间隙自适应工艺参数优化[J]. 材料科学与工艺, 2024, 32(1): 18 − 24. Chen Zhenkuan, He Jianping, Li Fang, et al. Optimization of adaptive process parameters for P-PAW lap welding gap of sheet metal based on BP neural network[J]. Materials Science and Technology, 2024, 32(1): 18 − 24.
[17] 徐健宁, 张华, 胡瑢华, 等. 熔焊快速成型中焊接工艺参数与焊缝几何尺寸的关系[J]. 焊接技术, 2008(4): 10 − 13. doi: 10.3969/j.issn.1002-025X.2008.04.004 Xu Jianning, Zhang Hua, Hu Ronghua, et al. Relationship between the parameters of rapid prototying welding and welding seam sizes[J]. Welding Technology, 2008(4): 10 − 13. doi: 10.3969/j.issn.1002-025X.2008.04.004
[18] Xiao Xinyuan, Shi Yonghua, Wang Guorong, et al. Study of image processing for V-shape groove and robotic weld seam tracking based on laser vision[J]. China Welding, 2008, 17(4): 68 − 73.
计量
- 文章访问数: 328
- HTML全文浏览量: 112
- PDF下载量: 89
