人工神经网络电阻点焊质量监测模型的研究

利用 BP网络对电阻点焊的动态电参数进行融合处理 ,建立了以电阻点焊动态电参数为输入空间、点焊熔核尺寸为输出空间的电阻点焊质量智能监测模型。检验结果表明 :该模型的熔核尺寸预测精度完全满足工程实际需要 ;有很大的实用价值。在网络训练中采用综合改进的 BP算法 ,使得网络的训练效率大大提高     

基于电极位移信号特征分析的电阻点焊质量监测

提出一种基于点焊过程信息采集和处理的焊点质量在线监测方法。通过对电极位移、动态电阻信号的实时采集和分析,利用电阻信号动态特征刻画熔核形成的不同阶段,从同步位移信号中提取12个与焊点质量相关的动态特征参量。通过对提取的特征参量与作为焊点质量评价指标的抗剪强度之间的相关性分析,选取相关性显著的特征参量作为输入和抗剪强度为输出,建立线性回归、非线性回归及径向基函数神经网络焊点质量监测模型。监测模型的有效性检验结果表明,建立的三种监测模型都可实现对焊点质量的在线监测。径向基函数神经网络模型的监测准确率高于其他两种模型,其平均验证误差为2.28%,最大验证误差低于10%。     

连续顺序电阻点焊分流率的反向传播神经网络预测

建立以材料电阻率、板厚、焊点间距为输入空间,分流率为输出空间的连续顺序电阻点焊分流率的3层误差反向传播(Back propagation,BP)神经网络预测模型.依据电阻点焊恒流控制的特点和点焊过程的电阻变化规律建立分流率的理论计算模型,由该模型所得数据作为样本对网络进行训练和检验.对2.0 mm厚度的20钢及1.5 ...     

基于神经网络的激光熔覆高度预测

激光成形过程中,对熔覆高度进行实时检测,从而实现熔覆高度闭环控制是成形高质量零件的保证.激光成形过程是一个多参数耦合的非线性过程,大量激光参数对成形熔覆表面质量具有重要影响.在分析激光参数对熔覆高度影响的基础上,建立利用激光工艺参数预测熔覆高度的误差反向传播(Back propagation,BP)神经网络模型,完成了网络算法设计.通过激光成形试验采集样本,利用训练样本对所建立的网络进行训练,完成网络输入输出高度映射关系,并利用测试样本对所训练的网络进行检验.仿真试验表明,神经网络熔覆高度预测模型具有很高的精度,验证了该预测模型在理论和实践上的可行性与有效性.神经网络熔覆高度预测模型为实现激光加工过程熔覆高度实时预测与闭环控制打下基础,对提高成形产品质量具有重要意义.     

Ni基-WC合金粉末激光熔覆层形貌的预测

基于BP神经网络,建立网络模型对激光熔覆层形貌尺寸进行预测,研究激光熔覆特征信号(蓝紫光信号、红外辐射信号、可听声信号)和激光熔覆形貌(熔覆层高、宽)之间的对应关系。结果表明,该网络模型预测平均误差小,检验精度高,具有较好的预测能力。     

基于卷积神经网络的GTAW熔透预测

焊接熔池信息能反映熔透状态,但建立熔池与熔透状态的关系十分困难。针对该问题,提出一种基于卷积神经网络(Convolution neural networks,CNN)的GTAW背面熔透预测模型。通过基于被动视觉的熔池二维图像采集,建立了用于CNN网络训练和测试的训练集和测试集。其次建立了CNN背面熔透预测模型,优化了学习率、batch-size、迭代次数等网络参数。研究发现:第一层卷积核尺寸为9×9,最后一层含有64个卷积核使模型在预测准确率和训练时间上综合表现达到最佳。使用训练集数据训练模型,将训练完成的模型在测试集上对背面熔透进行预测,取得了高于96%的预测准确率。通过对预测模型的特征映射进行可视化分析,模型是通过熔池边缘,反光点位置和熔池尾部等特征来预测背面熔透情况。     

基于BP神经网络的焊接工时定额模型研究

为了提高焊接工时定额的计算效率和精确度,消除传统定额方法过度依赖线性公式进行计算而忽略工时定额与其加工工艺参数之间的非线性关系的不足,将BP神经网络技术应用到焊接工时定额中,建立多输入单输出的神经网络模型,并运用实际数据对建立的神经网络进行训练,通过测试样本验证了该网络模型输出的仿真值与实际值之间的误差在5%以内。验证结果表明,用BP神经网络方法建立的焊接工时定额模型能够对焊接工时进行准确的预测。     

点焊熔核尺寸与工艺参数关系的模型化处理

针对生产中常用的 1Cr18Ni9Ti板材 ,采用正交试验设计法研究了其交流点焊接头的熔核尺寸 (熔核直径、焊透率 )受主要焊接工艺参数影响的规律性。对实验数据进行了多元线性回归 ,建立了熔核尺寸与焊接电流、焊接时间、电极压力、电极端面尺寸、工件厚度之间关系的数学模型 ,并采用该模型对熔核尺寸进行了预测     

CO2气体保护焊的焊缝形貌建模及虚拟化仿真系统开发

建立了CO₂气体保护焊工艺参数与焊缝几何尺寸(熔宽、熔深)之间的多层感知机神经网络预测模型,并基于焊接试验数据训练模型,确定了模型的数学解析式;通过分析焊缝截面和表面形貌特征,建立焊缝形貌的虚拟化仿真模型;通过python编程开发了焊缝形貌预测与虚拟化仿真系统。结果表明：所建立的多层感知机神经网络预测模型对熔宽预测的最大偏差为0.097 mm,模型拟合优度为0.999 269,对熔深预测的最大偏差为0.051 mm,模型拟合优度为0.999 567;建立了以焊缝熔深和熔宽为输入变量的焊缝截面形貌数学模型和以焊缝熔宽为输入变量的表面形貌数学模型。     

基于级联森林的电阻点焊熔核直径预测算法

异质材料电阻点焊是非线性的复杂物理过程，焊接工艺参数与熔核直径的数学映射模型难以建立。鉴于此，提出结合自适应差分进化算法与级联森林进行电阻点焊熔核直径预测。考虑两层不等厚异质板材电阻点焊实验数据的特征，以材料的物理特性和化学成分表征材料信息，并将工艺和材料视为质量影响因素作为模型的输入特征；分别选择随机森林、XGBoost和LightGBM作为元学习器级联到传统级联森林中，利用自适应差分进化算法搜寻最优超参数确定级联结构，引入SHAP理论分析影响熔核直径的因素及其作用。实验结果表明，经超参数优化的级联森林模型在测试集上的R²、MSE分别为0.9480、0.0315,优于基线模型，同时确定了熔核直径的重要影响因素，为新材料焊接的工艺参数设计提供依据。     

Numerical modeling of heat transfer and fluid flow in hybrid laser–TIG welding of aluminum alloy AA6082

This paper describes a three-dimensional numerical model based on finite volume method to simulate heat transfer and fluid flow in laser–tungsten inert gas (TIG) hybrid welding process. To simplify the model and reduce the calculation time, keyhole dynamics are not considered; instead, a new modified volumetric heat source model is presented for the laser source to take into account the effect of the keyhole on the heat transfer into the workpiece. Due to the presence of arc current, an appropriate electromagnetic model based on the Maxwell equations are also solved to calculate electromagnetic forces in the weld pool. The results of computer simulation, including temperature, current density, electromagnetic, and melted material velocity field, are presented here. Furthermore, several dimensionless numbers are employed to recognize the importance of fluid flow driving forces in the weld pool. It is deduced that the fluid flow has an important effect on the weld pool shape. It is also founded that among the driving forces, Marangoni force is dominant fluid force in the weld pool. Besides, calculated results of hybrid welding process are compared with those of TIG and laser welding processes. The weld pool depth is relatively the same, but the width of the weld pool is highly larger in hybrid welding than lone laser welding. Eventually, the presented model is validated by comparison between calculated and experimental weld pool shape. It is founded that there is a good agreement as the capability of this model can be proved.     

An improved artificial neural network for laser welding parameter selection and prediction

In the laser welding production, the selection and prediction of welding parameters is essentially important to guarantee weld quality. Artificial neural networks (ANN), which perform a nonlinear mapping between inputs and outputs, are an alternative approach for developing welding parameter forecasting model. In this paper, in order to speed up the convergence and avoid local minimum of the conditional ANN, genetic algorithm simulated annealing (GASA) based on the random global optimization is inducted into the network training. By means of GASA method, weights and threshold of neural networks can be globally optimized with short training time. Meanwhile, the gray correlation model (GCM) is used as a pre-processing tool to simplify the original networks based on obtaining the main influence elements of network inputs. The GCM–GASA–ANN method combines the complementary features of three computational intelligence techniques and owns very good applicability. Through the simulation and analysis of an orthogonal experiment, the proposed method can be proved to have higher accuracy and to perform better than the traditional ANN to forecast the laser welding parameters.     

激光焊匙孔特征的近红外与X射线传感分析

由于多传感匙孔特征参数可以有效地反映大功率激光焊接质量状态,本文研究了匙孔特征信息的提取方法并建立了焊缝成形预测模型。以大功率盘形激光焊接304不锈钢为试验对象,应用近红外高速摄像机和X射线视觉成像系统同时提取了焊接过程中的熔池动态图像,并分割出匙孔区域。针对近红外图像,应用矩方法导出匙孔的不变矩特征,同时定义并提取匙孔面积和最前端点纵坐标两个特征;针对X射线图像则提取匙孔深度和熵两个特征。在不同激光功率条件下得到匙孔特征并进行特征融合分析,然后建立了3个BP神经网络焊缝成形预测模型。探索了匙孔形态、焊接条件和焊接状态三者之间的联系,实现了对焊接过程的在线监测。试验结果表明,将两个传感器获取的匙孔特征信息融合并进行主成分分析变换后,熔宽和熔深的预测绝对误差平均值分别为0.18mm和0.57mm,比基于单个传感器获取匙孔特征建立的BP神经网络分别减小了0.03mm和0.31mm,显示提出的方法能够有效在线监测大功率盘形激光焊接状态。     

低功率激光诱导电弧复合焊接钛合金薄板工艺研究

采用低功率脉冲YAG激光诱导非熔化极惰性气体保护(Tungsten inert gas,TIG)焊电弧复合热源实现了1 mm厚TC4钛合金薄板的优质焊接,研究激光诱导电弧复合焊接过程中热源能量匹配、热源间角度、对接间隙对焊缝成形的影响规律。结果表明,钛合金薄板低功率脉冲YAG激光诱导TIG电弧复合热源焊接过程中,激光能量与电弧能量之间的相互匹配将显著影响焊缝的表面成形。相对于电弧功率的变化,焊缝成形对激光功率变化的敏感度更高。随着热源间角度减小,激光诱导电弧复合热源传热能力增强;由于复合焊接速度快、热输入小、焊接试板横向变形小,当对接间隙为0~0.5 mm范围内时均能获得良好的焊缝成形。为了使焊缝成形均匀连续,焊接过程中需要对焊缝背面采用氩气进行保护,当保护气体流量为5~8L/min时获得最佳焊接接头。     

Process Characterisation of Pulsed Nd:YAG Laser Seam Welding

A wide range of pulsed laser welding parameters was identified. These include average peak power density (APPD), peak power, mean laser power, traverse speed, pulse repetition rate, duty cycle, pulse energy, spot size, and pulse duration. The type of laser beam temporal pulse shape studied was a rectangular power pulse. The effects of pulsed laser welding parameters on heat flow, weld dimension, and weldability are investigated. The study shows that weld quality is principally affected by APPD, mean power, and traverse speed, of which APPD is the most critical process parameter. A processing map containing the APPD effects is constructed as a guide for producing good welds.     

Adaptive volumetric heat source models for laser beam and laser + pulsed GMAW hybrid welding processes

Laser + pulsed gas metal arc welding (GMAW) hybrid welding process is an attractive joining technology in industry due to its synergy of the two processes. It is of great significance to conduct fundamental investigations involving mathematical modeling and understanding of the hybrid welding process. In this study, an adaptive heat source model is first developed for laser beam welding. Through combining the ray-tracing method with the keyhole profile determination technique based on the local energy balance, the keyhole shape and size are calculated and correlated to the distribution parameters of the volumetric heat source model. Then, thermal action characteristics in laser + pulsed GMAW hybrid welding are considered from viewpoint of macro-heat transfer, and a combined volumetric heat source model for hybrid welding is developed to take consideration of heat input from laser, pulsed gas metal arc, and overheated droplets. Numerical analysis of thermal conduction in hybrid welding is conducted. The shape and size of fusion zone and weld dimension in the quasi-steady state are calculated for various hybrid welding conditions, which have a fair agreement with the experimental results.     

Understanding the process parameter interactions in multiple-pass ultra-narrow-gap laser welding of thick-section stainless steels

In laser welding, typical welding penetration depths are in the order of 1–2 mm/kW laser power. The multipass laser welding technique, based on the narrow-gap approach, is an emerging welding technology that can be applied to thick-section welds by using relatively low laser power, but the process is more complicated since it is necessary to introduce filler wire to narrow-gap weld configurations. The aim of this work was to understand significant process parameters and their interactions in order to control the weld quality in ultra-narrow-gap (1.5 mm gap width) laser welding of AISI grade 316L stainless steel. A 1-kW IPG single-mode fiber laser was used for welding plates that were 5 to 20 mm in thickness using the multiple-pass narrow-gap approach. Design of experiments and statistical modelling techniques were employed to understand and optimise the processing parameters. The effects of laser power, wire feed rate, and welding speed on the weld homogeneity, integrity, bead shape, gap bridgability and surface oxidation were studied. The results were evaluated under different optimising constraints. The results show that the models developed in this work can effectively predict the responses within the factors domain.     

Face milling for hypoid internal bevel gear using new type of tilt milling machine

An hourglass-like heat source model is developed to simulate the geometry of weld beads of laser beam welding. The welding experiments with different welding parameters were carried out to find whether the welds are full-penetrated or not. The laser powers ranged between 1.0 and 5.0 kW with steps of 1.0 kW. Five levels of the welding speed were used: 1.0, 2.0, 3.0, 4.0, and 5.0 m/min. The numerical analysis based on finite element method was used to predict the shape of weld pools during welding. To give a better representation of weld bead, an hourglass-like heat source model is developed to describe the unique shape of 1060 steel weld bead for laser beam welding without filler metal instead of the commonly used conical heat source. Comparing the predicted fusion profile with the experimental cross section of weld, it is found that hourglass-like heat source model is more consistent with the experimental results than the conical heat source for 1060 steel. In addition, the best heat source parameters are obtained by an error analysis. Compared to the experimental results, the simulated results for the welds using Power = 4000 W and Speed = 2.0 m/s and 5000 W and 2.0 m/s match the experimental results well.     

Analysis of weld pool dynamic during stationary laser–MIG hybrid welding

A mathematical model was established to simulate the weld pool development and dynamic process in stationary laser–metal inert gas (MIG) hybrid welding. Surface tension and buoyancy were considered to calculate liquid metal flow pattern; moreover, typical phenomena of MIG welding such as filler droplets impinging weld pool, electromagnetic force in the weld pool, and typical phenomena of laser beam welding such as recoil pressure, inverse Bremsstrahlung absorption, and Fresnel absorption were all considered in the model. The laser beam and arc couple effect was introduced into this model by the plasma width during hybrid welding. Transient weld pool shape and complicated liquid metal velocity distribution from two kinds of weld pool to a unified weld pool were calculated. Furthermore, the simulated weld bead geometries were in good agreement with experimental measurement.     

Nd:YAG激光+脉冲MAG电弧复合热源焊接参数对焊缝熔深的影响

试验研究Nd:YAG激光 脉冲MAG电弧复合热源焊接过程中焊接参数对焊缝熔深的影响.研究结果表明,复合热源焊缝熔深随电弧功率和激光功率的增大而增大,随焊接速度的增大而减小,并且在相同参数下,复合热源焊缝熔深稍大于激光焊缝熔深而显著大于脉冲MAG焊缝熔深.对于不同焊接电流,光丝间距在0～3 mm内复合热源焊缝取得最大熔深,且取得最大熔深的光丝间距与焊接电流大小有关;复合热源焊缝熔深在离焦量为2 mm时取得最大值.试验结果分析表明,在激光 电弧复合热源焊接过程中激光功率不仅决定复合热源焊缝熔深,而且可以极大地提高焊接速度:MAG电弧也可提高Nd:YAG激光焊的热效率.