A cooperative control strategy of resistance spot welding process by combining the constant current control with the DRC method

The modeling control method based on the dynamic resistance characteristics of good nuggets, that is the DRC method, is an improvement on the dynamic resistance threshold method for the quality control of resistance spot welding. But there is still a control blind area in the initial four cycles. For this reason, the quality of every weld nugget could not be fully ensured. Thus a new fuzzy cooperative control method is put forward. It uses a multi-information time-control mechanism by combining the constant current control technology with the DRC method in a relay way. This whole-process control strategy has led to a good control effect and produced the dual-identical results in the weld nugget quality and the welding time.

Quality estimation of resistance spot welding of stainless steel based on BP neural network

The end value of the dynamic resistance curve of stainless steel was proved to have strong correlation with nugget size by experiments, so it was an important factor for estimation of weld quality. BP neural network was employed to estimate the weld quality, The end value of the dynamic resistance curve, welding current and welding time were selected as the input variables while the nugget diameter, which is closely related to weld quality, was selected as the output variable. Testing results shows that such network has fine fault tolerance and real-time quality estimation is possible.

An Electrothermal Model Based Adaptive Control of Resistance Spot Welding Process

Resistance Spot Welding (RSW) is a process commonly used for joining a stack of two or three metal sheets at desired spots. The weld is accomplished by holding the metallic workpieces together by applying pressure through the tips of a pair of electrodes and then passing a strong electric current for a short duration. Inconsistent weld and insufficient nugget size are some of the common problems associated with RSW. To overcome these problems, a new adaptive control scheme is proposed in this paper. It is based on an electrothermal dynamical model of the RSW process, and utilizes the principle of adaptive one-step-ahead control. It is basically a tracking controller that adjusts the weld current continuously to make sure that the temperature of the workpieces or the weld nugget tracks a desired reference temperature profile. The proposed control scheme is expected to reduce energy consumption by 5% or more per weld, which can result in significant energy savings for any application requiring a high volume of spot welds. The design steps are discussed in details. Also, results of some simulation studies are presented.

Neural network prediction of the shunt current in resistance spot welding

An error back propagation （BP） neural network prediction model was established for the shunt current compensation in series resistance spot welding. The input variables for the neural network consist of the resistivity of the material, the thickness of workpiece and the spot spacing, and the shunt rate is outputted. A simplified calculation for the shunt rate was presented based on the feature of the constant-current resistance spot welding and the variation of the resistance in resistance spot welding process, and then the data generated by simplified calculation were used to train and adjust the neural network model. The neural network model proposed was used to predict the shunt rate in the spot welding of 20# mlid steel （in Chinese classification） （in 2. 0 mm thickness） and 10# mild steel （in 1.5 mm and 1.0 mm thickness）. The maximum relative prediction errors are, respectively, 2. 83%, 1.77% and 3.67%. Shunt current compensation experiments were peoCormed based on the neural network prediction model proposed to check the diameter difference of nuggets. Experimental results show that maximum nugget diameter deviation is less than 4% for both 10# and 20# mlid steels with spot spacing of 30 mm and 50 mm.

Modeling and analysis of resistance spot welding based on neural network

A numerical study on the multi-parameter control method based on nonlinear auto-regressive with exogenous input neural network （NARX） is presented here. Welding current was set as the input parameter; electrode displacement and dynamic resistance were set us the output parameters. The NARX model using these parameters was set up to simulate the multi-parameter resistance spot welding process. By comparing actual experimental data and NARX model output data, it was validated that the results from the model reflect the relationship between input parameter and output parameters correctly under the influence of many affecting factors.

On-line evaluating on quality of mild steel joints in resistance spot welding

A method was developed to realize quality evaluation on every weld-spot in resistance spot welding based on information processing of artificial intelligent. Firstly, the signals of welding current and welding voltage, as information source, were synchronously collected. Input power and dynamic resistance were selected as monitoring waveforms. Eight characteristic parameters relating to weld quality were extracted from the monitoring waveforms. Secondly, tensile-shear strength of the spot-welded joint was employed as evaluating target of weld quality. Through correlation analysis between every two parameters of characteristic vector, five characteristic parameters were reasonably selected to found a mapping model of weld quality estimation. At last, the model was realized by means of the algorithms of Radial Basic Function neural network and sample matrixes. The results showed validations by a satisfaction in evaluating weld quality of mild steel joint on-line in spot welding process.

Quality estimation of the resistance spot welding based on pattern feature of the electrode displacement signal

The electrode displacement signal of the resistance spot welding process is monitored and mapped into a binary matrix. Some welded spots, from different welding current specifications, are classified into five classes according to the prototypes of the pattern matrices. A reliable quality classifier is developed based on Hopfield network when the tensile shear strength of the welded joint is measured as the quality indicator. The cross validation test results show that the method utilizing pattern matrix of the displacement signal to characterize nugget formation process is feasible and it can provide adequate quality information of the welded spot. At the same time, under small sample circumstance, the classifier presents good classification ability and it also can correctly estimate the weld quality in some abnormal welding process according to the pattern feature of the displacement signal.

Parameter selecting and quality predicting of spot welding based on artificial neural networks

This paper proposes a procedure for using artificial neural networks (ANN) in spot welding , and establishes spot welding parameter selecting ANN systems and spot welding joint quality predicting ANN systems . It has been proved that the ANN systems have high prediction precision , providing a new way of parameter selecting and quality predicting in spot welding .

Online estimation and characteristic analysis of double nugget diameters during aluminum/steel resistance spot welding process

Online estimation of the double nugget diameters was performed by means of a back propagation neural network.The double nugget diameters were obtained using actual welding experiment and numerical simulation,according to different characteristics of aluminum nugget and steel nugget.The input of the neural network was some key characteristic parameters extracted from dynamic power signal,which were peak point,knee point and their variation rate over time,as well as heat energy delivered into the welding system.The architecture of the neural network was confirmed by confirming the number of neurons in hidden layer through a series of calculations.The key parameters of the neural network were obtained by means of training 81 arrays of data set.Then,the neural network was used to test the remaining 20 arrays of verifying data set,and the results showed that both of the mean errors for the two nugget diameters were below 3%.In addition,corresponding analyses showed that the accuracy of two nugget diameters was higher than that of tensile-shear strength.

基于改进麻雀搜索算法优化BPNN的电阻点焊质量预测

电阻点焊技术由于具有高效、自动化程度高等焊接特点,被广泛应用于汽车、航空航天和公共交通等制造领域,由于焊点在封闭状态下进行,焊接过程存在诸多影响因素且无法直接检测,因此,准确预测电阻点焊质量是生产过程中必不可少的环节.本文以2219/5A06铝合金为研究对象,在3种不同的装配条件(包括间隙和间距)下进行电阻点焊工艺信号的分析,并进行人工智能建模.为了提高电阻点焊质量评价的性能和效率,本文采用Logistic-Tent(LT)复合映射改进麻雀搜索算法(SSA)对反向传播神经网络(LT-SSA-BPNN)模型进行优化,模型的输入和输出分别为多信号融合后的变量和熔核直径.实验结果表明,与传统的标准反向传播神经网络(BPNN)模型相比,经过LT-SSA-BP模型优化后,预测结果的平均绝对误差(MAE)、均方误差(MSE)和均方根误差(RMSE)分别降低了36.17%、17.55%和51.75%.同时,LT-SSA-BP神经网络在添加了不同间隙和间距条件作为训练集后,其预测稳定性明显提高,可以成功预测电阻点焊质量.

基于CAN总线的电阻焊网络控制系统

介绍了CAN总线特点,分析了电阻焊网络控制系统的工作原理和硬件结构,制定了适合电阻焊生产的CAN 通信协议,设计了底层电阻焊网络控制器的控制软件、上／下位机通信软件和上层管理软件。试验结果表明,设计的基于CAN总线的电阻焊网络控制器具有控制精度高、工作稳定可靠、通信速度快和可扩展性强等特点,能够实现电阻点焊的在线网络控制与数据管理,可以应用于焊接单元制造信息网络智能集成。

基于相关性分析和SSA-BP神经网络的铝合金电阻点焊质量预测

基于电阻点焊过程中工艺信号特征,在不同间距、不同间隙和不同间距与间隙3种条件下,引入相关性分析方法分析工艺信号与熔核直径之间的相关性,并建立基于麻雀搜索算法-BP神经网络(sparrow search algorithmback propagation neural network,SSA-BP)的电阻点焊质量预测模型,将功率、焊接电流、焊接电压和动态电阻作为预测模型输入特征.结果表明,经麻雀搜索算法优化后的BP神经网络在测试集上的决定系数R2、均方误差(meansquare error,MSE)、均方根误差(root mean square error,RMSE)和平均绝对误差(mean absolute error,MAE)分别为0.95,1.55,1.24和0.90,均优于BP模型.获得了功率、焊接电流、焊接电压和动态电阻与熔核直径的映射关系,可为焊接的工艺参数设计提供依据.

基于电阻点焊过程与形貌的飞溅成因分析及控制方法

【目的】在汽车白车身制造生产过程中,电阻点焊是板件连接的主要焊接工艺,但电阻点焊焊接过程中焊渣飞溅对白车身焊接质量影响较大,还会导致设备故障等问题,进而增加返修成本并降低生产效率。因此,该文对降低电阻点焊白车身的飞溅成因分析与控制方法展开研究。【方法】提出了一种飞溅成因分析与控制方法的操作流程,从多维度分析焊接过程与焊接形貌,进而针对性地实施措施,并在实际生产中验证飞溅成因分析与控制方法的操作流程。【结果】在验证过程中。通过对焊装车间飞溅率较高的工位基于分析结果进行优化,使得车间整体飞溅率由14%降低至6%,飞溅率低于优化前50%,有效减少预打磨工位的工作量和飞溅引起的故障率。【结论】飞溅成因分析与控制方法的理论体系保证了白车身焊接质量,提高了生产效率,并为后续工作提供思路指导与经验借鉴。焊接飞溅率是焊装车间主要控制指标之一,在生产过程中还需要不断监测、调整和保持,从而制定更有效的标准流程,减少飞溅对白车身质量的影响。

CAN总线在电阻焊网络控制中的应用

本文简要介绍了CAN总线的特点,分析了CAN总线在电阻焊网络控制中的应用,制定了系统节点ID和数据帧,设计了系统通信硬件电路和软件模块。电阻点焊实验结果表明,基于CAN总线的电阻焊网络控制系统通信速度快,工作稳定可靠,可扩展性强,能够应用于焊接单元制造信息网络智能集成。

电阻点焊质量Hopfield神经网络在线评估方法

目的电阻点焊广泛应用于汽车、家电等领域,但目前少有准确的无损质量评价方法。为此,研究一种基于低碳钢板焊接功率信号的焊接质量在线评估方法,并探索利用该信号来评价电阻点焊焊点质量的可能性。方法对焊接电流、电压信号进行测量和分析,研究功率信号表征焊接质量的可靠性,提出一种有效的模式特征提取方法,将动态功率信号转换为二值图像并用二值矩阵表征,该方法避免特征提取和选择,且尽可能保留焊点质量信息。通过拉剪试验将焊接样本分为6种不同的焊接等级,利用Hopfield关联记忆神经网络建立焊接质量分类器,将具有不同焊接质量水平的焊接样本模式特征矩阵记忆为稳定状态。结果将焊接样本的模式特征矩阵输入分类器,通过Hopfield网络关联记忆将其收敛到最相似的稳定状态,最终锁定了稳定状态对应的焊接质量。60个测试样本中59个样本都可以被正确分类,该分类器的分类准确率达到98%。结论分类性能试验结果表明,所提出的模式特征提取方法快速、有效,并能可靠地在线评估低碳钢板的焊接质量。

车身制造中电阻点焊飞溅产生的原因及控制方法

目前国内外白车身焊接广泛采用电阻点焊工艺进行焊接,但电阻点焊过程中受到焊接打点位置、角度、母材板间隙、电极和焊接参数等因素影响,不可避免地会产生大量焊接飞溅,尤其近年来车身镀锌板、高强度钢板的使用,导致飞溅问题尤为严重。通过对点焊飞溅产生的原因进行分析,找出白车身焊接过程中产生飞溅的影响因素,并结合生产实际情况提出相应的控制方法,从而降低了飞溅的发生概率,提高了车身焊接品质。

Improving RSW nugget diameter prediction method:unleashing the power of multi-fidelity neural networks and transfer learning

This research presents an innovative approach to accurately predict the nugget diameter in resistance spot welding(RSW)by leveraging machine learning and transfer learning methods.Initially,low-fidelity(LF)data were obtained through finite element numerical simulation and design of experiments(DOEs)to train the LF machine learning model.Subsequently,high-fidelity(HF)data were collected from RSW process experiments and used to fine-tune the LF model by transfer learning techniques.The accuracy and generalization performance of the models were thoroughly validated.The results demonstrated that combining different fidelity datasets and employing transfer learning could significantly improve the prediction accuracy while minimize the costs associated with experimental trials,and provide an effective and valuable method for predicting critical process parameters in RSW.

基于DSP的电阻点焊检测与智控技术研究

研制了一种基于DSP的电阻点焊检测与控制系统 .在软、硬件设计中分别采用了抗干扰技术 ,保证了系统的可靠性 .误差分析和工艺试验表明 。

基于声发射信号的铝合金点焊裂纹神经网络监测

铝合金热加工过程的冶金行为比较复杂,在电阻点焊快速加热和冷却条件下,极易产生裂纹缺陷。基于虚拟仪器技术,以Lab VIEW为软件平台,结合Matlab数值分析软件,构建了电阻点焊过程声发射信号采集分析及铝合金点焊裂纹监测系统。以2A12铝合金电阻点焊熔核冷却结晶过程,即点焊焊接循环维持阶段的声发射信号为研究对象,提取与声发射信号强度相关的振铃计数、能量、有效电压及5层小波分解125～250 k Hz频带能量系数4个特征参数作为输入矢量,裂纹作为输出矢量,建立3层BP神经网络铝合金点焊裂纹的监测模型,并利用测试样本对该模型进行验证。结果表明,裂纹监测的正确率达到89.1%,为监测铝合金电阻点焊裂纹提供了一种有效的方法。

人工神经网络技术在点焊质量控制中的应用研究

利用人工神经网络技术对交流电阻点焊的多个动态电参数进行融合处理,建立起以交流点焊过程中动态电参数作为输入空间;以熔核尺寸为输出空间,可用于实时在线检测和预测的低碳钢点焊质量监测系统。所建监测系统的熔核直径的平均预测误差小于5%,熔核高度的平均预测误差小于8%,完全可以满足工程实际的需要。