基于鲸鱼优化算法的焊缝尺寸预测

姚屏; 李文强; 陈威; 何日恒; 张佩美; 张广潮

doi:10.12073/j.hjxb.20240701001

基于鲸鱼优化算法的焊缝尺寸预测

姚屏^{1, 2,},
李文强¹,
陈威¹,
何日恒¹,
张佩美¹,
张广潮¹

1.
广东技术师范大学自动化学院, 广州, 510630
2.
广州理工学院智能制造与电气工程学院, 广州, 510540

基金项目: 国家自然科学基金资助项目(51805099)，广东省教育厅重大专项资助项目(2020ZDZX2019)，广东省高等教育学会实验室管理专业委员会基金项目(GDJ2022032)

详细信息

作者简介:
姚屏，博士，教授，主要从事智能制造，制造过程智能化检测与控制方面的研究. Email: ypsunny@163.com

中图分类号: TP 391
计量
- 文章访问数: 112
- HTML全文浏览量: 11
- PDF下载量: 37
出版历程
- 收稿日期: 2024-06-30
- 网络出版日期: 2024-11-14
- 刊出日期: 2024-11-24

Prediction of weld size prediction based on Whale Optimization Algorithm

1.
College of Automation, Guangdong Polytechnic Normal University, Guangzhou 510630, China
2.
College of Intelligent Manufacturing and Electrical Engineering, Guangzhou Institute of Science and Technology, Guangzhou 510540, China

摘要

摘要:
在机器人电弧焊过程中，准确预测焊缝尺寸对于控制焊接成形质量具有重要意义. 文中提出了一种融合鲸鱼优化算法(whale optimization algorithm, WOA)与深度信念网络(deep belief network, DBN)的预测模型，简称WOA-DBN. 该模型以电流、频率、占空比和焊接速度作为输入参数，进行了四因素十水平正交试验，构建了针对电弧焊焊缝尺寸的预测模型. 为提升算法的搜索效率、增强收敛性能并避免陷入局部最优解，引入了混沌反向学习初始化种群、非线性收敛因子以及模拟退火操作和自适应变异扰动等策略，建立了一种混沌鲸鱼优化算法优化的深度信念网络模型，即AAMCWOA- DBN. 通过试验对比，AAMCWOA-DBN模型在预测精度和性能指标方面均优于传统的WOA-DBN模型，熔宽预测的MAPE仅有1.85%，余高预测的MAPE仅有0.47%. 文中利用人工智能算法对电弧焊的焊缝尺寸进行预测，为焊接成形控制和焊接质量的智能化提供了新的研究视角和方法，有望在相关领域得到应用.
- 人工智能 /
- 机器人电弧焊 /
- 焊缝尺寸预测 /
- WOA-DBN /
- AAMCWOA-DBN
Abstract:
In the robotic arc welding manufacturing process, accurate prediction of weld seam size is important for controlling the quality of weld formation. In this study, a prediction model fusing whale optimization algorithm (WOA) and deep belief network (DBN), referred to as WOA-DBN, is proposed, which constructs a prediction model of weld size for arc welding manufacturing, using current, frequency, duty cycle, and welding speed as the input parameters. Welding seam size prediction model. In order to improve the search efficiency of the algorithm, enhance the convergence performance and avoid falling into the local optimal solution, this paper introduces the strategies of chaotic inverse learning initialization population, nonlinear convergence factor, as well as the simulated annealing operation and adaptive mutation perturbation, etc., and establishes a deep belief network model optimized by the chaotic whale optimization algorithm, i.e., AAMCWOA-DBN. Through experimental comparison, the AAMCWOA-DBN model outperforms the traditional WOA-DBN model in terms of prediction accuracy and performance metrics, the MAPE of the fused wide forecast is only 1.85% and the MAPE of the remaining high forecast is only 0.47%. This study utilizes artificial intelligence algorithms to predict the weld seam size of arc welding manufacturing, which provides new research perspectives and methods for the intelligent control of weld shaping and weld quality, and is expected to be applied in related fields.
- artificial intelligence /
- robotic arc welding /
- weld seam size prediction /
- WOA-DBN /
- AAMCWOA-DBN

HTML全文

图 1 焊缝示意图

Figure 1. Schematic diagram of weld. (a) three measurements of the weld position; (b) schematic diagram of weld cross-section

下载: 全尺寸图片幻灯片

图 2 隐藏层层数和隐元个数对预测精度的影响

Figure 2. Influence of the number of hidden layers and neurons on prediction accuracy

下载: 全尺寸图片幻灯片

图 3 焊缝尺寸预测流程

Figure 3. The prediction process of Weld Bead Size

下载: 全尺寸图片幻灯片

图 4 不同模型结果对比

Figure 4. Comparative Analysis of Results. (a) melting width prediction; (b) excess height prediction

下载: 全尺寸图片幻灯片

图 5 误差分布图

Figure 5. Error Distribution Chart. (a) melting width error; (b) excess height error

下载: 全尺寸图片幻灯片

图 6 模型性能检验

Figure 6. Model performance tests

下载: 全尺寸图片幻灯片

表 1 试验因素水平表

Table 1 Experimental factors and levels table

因素	焊接电流 I/A	频率 f/Hz	占空比 D(%)	焊接速度 v/(cm·min⁻¹)
水平	53	0.5	28	25
	60	1	30	30
	68	1.5	32	35
	75	2	35	40
	83	2.5	38	45
	90	3	40	50
	98	3.5	42	55
	105	4	45	60
	113	4.5	48	65
	120	5	50	70

下载: 导出CSV

表 2 WOA-DBN的超参数

Table 2 Hyperparameters of WOA-DBN

隐层层数 K/层	隐层神经元数				学习速率	预训练最大迭代次数 T₁/次	反向微调最大迭代次数 T₂/次
隐层层数 K/层	n₁	n₂	n₃	n₄	$ \eta $	预训练最大迭代次数 T₁/次	反向微调最大迭代次数 T₂/次
4	10	25	20	15	0.03	5	32

下载: 导出CSV

表 3 性能指标对比分析

Table 3 Performance Metrics Comparative Analysis

预测模型	指标	性能指标
预测模型	指标	时间t/s	平均绝对百分比误差ε_MAPE(%)	平均绝对误差ε_MAE	平均绝对误差值θ
AAMCWOA-DBN	熔宽	1.62	1.85	0.13	0.1255
WOA-DBN	熔宽	2.93	3.96	0.47	0.3335
AAMCWOA-DBN	余高	1.145	0.47	0.08	0.0765
WOA-DBN	余高	1.985	1.28	0.17	0.2025

下载: 导出CSV

表 4 预测精度分析

Table 4 Prediction Accuracy Analysis

指标	预测模型	最低精度ε(%)	最高精度ε(%)	分布占比δ(%)
指标	预测模型	最低精度ε(%)	最高精度ε(%)	≥95%	≥98%
熔宽	WOA-DBN	92.1	99.4	65	5
熔宽	AAMCWOA-DBN	96.4	99.6	100	70
余高	WOA-DBN	87.6	99.1	20	10
余高	AAMCWOA-DBN	94.8	99.6	100	20

下载: 导出CSV

参考文献(10)

[1]	石玗, 朱珍文, 张刚, 等. 金属电弧增材成形控制关键技术及研究现状[J]. 材料导报, 2022, 36(12): 135 − 142. doi: 10.11896/cldb.20090337 Shi Yu, Zhu Zhenwen, Zhang Gang, et al. Key technology and status of metal arc additive morphology control[J]. Materials Reports, 2022, 36(12): 135 − 142. doi: 10.11896/cldb.20090337
[2]	柏久阳, 王计辉, 林三宝, 等. 铝合金电弧增材制造焊道宽度尺寸预测[J]. 焊接学报, 2015, 36(9): 87 − 90. Bai Jiuyang, Wang Jihui, Lin Sanbao, et al. Width prediction of aluminium alloy weld additively manufactured by TIG arc[J]. Transaction of the China Welding Institution, 2015, 36(9): 87 − 90.
[3]	Xiong Jun, Yin Ziqiu. Forming appearance control of arc striking and extinguishing area in multi-layer single-pass GM AW-based additive manufacturing[J]. The International Journal of Advanced Manufacturing Technology, 2016, 87(1): 579 − 586.
[4]	周春东, 章晓勇, 彭勇. 等. CMT Cycle Step 工艺参数对焊缝表面特征纹路及成形尺寸的影响[J]. 焊接学报, 2023, 44(5): 95 − 101. doi: 10.12073/j.hjxb.20220126001 Zhou Chundong, Zhang Xiaoyong, Peng Yong, et al. Influence of CMT cycle step process parameters on weld surface feature patterns and forming[J]. Transactions of the China Welding Institution, 2023, 44(5): 95 − 101. doi: 10.12073/j.hjxb.20220126001
[5]	Hu Z Q, Qin X P, Li Y F, et al. Multi-bead overlapping model with varying cross-section profile for robotic GMAW-based additive manufacturing[J]. Journal of Intelligent Manufacturing, 2020, 31: 1133 − 1147. doi: 10.1007/s10845-019-01501-z
[6]	赵鹏, 吕彦明, 周文军, 等. 钨极惰性气体保护焊电弧增材制造单焊道尺寸预测[J]. 机械工程材料, 2020, 44(11): 78 − 82,91. doi: 10.11973/jxgccl202011014 Zhao Peng, Lu Yanming, Zhou Wenjun, et al. Prediction of Single Weld Based Size of TIG Welding Arc Additive Manufacturing[J]. Materials for Mechanical Engineering, 2020, 44(11): 78 − 82,91. doi: 10.11973/jxgccl202011014
[7]	孙家豪, 张超勇, 吴剑钊, 等. 基于神经网络的316L不锈钢激光焊焊缝形貌预测[J]. 焊接学报, 2021, 42(12): 40 − 47. doi: 10.12073/j.hjxb.20210526003 Sun Jiahao, Zhang Chaoyong, Wu Jianzhao, et al. Neural network-based prediction of weld morphology for laser welding of 316L stain less steel[J]. Transactions of the China Welding Institution, 2021, 42(12): 40 − 47. doi: 10.12073/j.hjxb.20210526003
[8]	Yao P, Zhou K, Tang H. Effects of operational parameters on the characteristics of ripples in double-pulsed GMAW process[J]. Materials, 2019, 12(17): 2767. doi: 10.3390/ma12172767
[9]	Yao P, Zhou K, Huang S. Process and parameter optimization of the double-pulsed GMAW process[J]. Metals, 2019, 9(9): 1009. doi: 10.3390/met9091009
[10]	Wang Xiaoming, Wei L, Yajie L, et al. Prediction model of refined gasoline blending formula based on PSO-DBN[J]. China Petroleum Processing & Petrochemical Technology, 2022, 24(3): 128 − 138.

施引文献(10)

期刊类型引用(3)

1.	陶旭阳，何建萍，徐磊. 超薄板脉冲微束等离子弧焊熔池振荡频率特征. 轻工机械. 2020(03): 24-27+32 . 百度学术
2.	陶旭阳，何建萍，徐磊. 基于LabView脉冲微束等离子弧焊熔池振荡信息提取. 智能计算机与应用. 2020(03): 371-374 . 百度学术
3.	顾玉芬，席保龙，李春凯，石玗，代悦，丁彬. 基于熔池振荡的GTAW熔透实时传感与控制. 电焊机. 2020(12): 5-8+108 . 百度学术