Multiple welding seam video transmission and monitoring of welding robot in local area network

A method of transmission and monitoring in local area network (LAN) for the welding seam video information of binocular vision welding robot is proposed. This paper uses MPEG-4 video CODEC functions of DivX, video for Windows (VFW) video capture technology, multi-threaded network communication idea and WinSock Socket in Visual C++ 6.0. In the end, client/server (C/S) communication model based on TCP/IP protocol has been developed. Experimental results show that the method can achieve the multiple welding seam video??s real-time transmission in LAN. Moreover, the Canny operator is used to realize the edge feature extraction of the seam image which is captured by the client, so that the edge of the seam extracted is continuous and clear.     

Research on voids deformation welding condition for manufacturing of heavy forgings

Crushing and diffusion welding are two critical healing stages of interior void defects in heavy forgings. The healing result depends on many factors during the forging process, such as stress, temperature, deformation and type of material, while the void diffusion welding condition is still not well known at the present. This paper is concerned with the deformation welding condition of the closed void interface in heavy ingot during the hot forging process. A void crushing experiment is carried out to recognize the microstructure of the closed void interface. According to the healing mechanism at high temperature, a new physical simulation model is setup to study the deformation welding process of the closed void interface based on the theory of atom diffusion and the interface contact mechanics prototype. Compared with the experimental results, the influence of deformation degree, forming temperature and holding time on the welding quality is discussed, and then the deformation welding condition of closed void interface is presented. The proposed condition helps to improve forging technology and product quality.     

Modeling, identification and simulation of DC resistance spot welding process for aluminum alloy 5182

Currently the aluminum alloy resistance spot welding(AA-RSW) has been extensively used for light weight automotive body-in-white manufacturing.However the aluminum alloys such as AA5182 have inferior weldability for forming the joint due to their high reflectiveness to heat and light.Therefore it is necessary to further develop the high performance control strategy and the set-up of a new welding schedule.The welding process identification is the essential issue where the difficulty arises from the fact that the AA-RSW is a nonlinear time-varying uncertain process which couples the thermal,electrical,mechanical and metallurgical dynamics.To understand this complicated physical phenomenon a novel dual-phase M-series pseudo-random electrical pattern is adopted to excite the AA-RSW electrical-thermal process and the thermal response is recorded according to the welding power outputs.Based on the experimental information,the transfer function of an AA-RSW electrical- thermal mechanism is identified,and the optimum model order and parameters are determined.Subsequently a control-oriented DC AA-RSW model is established to explore the welding power control algorithm.The simulated results of the control model show agreement with the experimental data,which validates its feasibility for the corresponding welding control.     

Self-organizing map based quality assessment for resistance spot welding with featured electrode displacement signals

To classify the quality of the resistance spot welding process, a relationship between the welder electrode displacement curve characteristics and the weld shear force has been explored. Eleven statistical features of the displacement signals are extracted to represent the welding quality. Self-organizing map (SOM) neural networks have been employed to discover their quantitative relationship. In order to identify the influence of various displacement curve features, all of the available combinations have been used as inputs for SOM neural networks. Further we analyze the impact of each feature on the classification results, yielding the best quality-indicative combination of characteristics. There is no determinant relationship between the welding quality and the level of expulsion rate. The quality of welding is most impacted by the maximum electrode displacement, the span of welding process and the centroid of the electrode displacement curve. The experiments show that SOM is feasible to assess the welding quality and can render the visualized intuitive evaluation results.     

Optimal posture searching algorithm on mobile welding robot

An independently developed mobile welding robot system is introduced. Kinematics model is the base of robot planning. With the Denavit-Hartenberg method, robot kinematics equation can be gotten. The robot welding torch inclination angle consists of the working and moving angles. According to the kinematics model, both the two angles and torch position can be calculated. Under special moving angle, a two-step algorithm is used to obtain the optimal manipnlator posture. The calculation result shows that the Mgorithm has high precision.     

Application of image morphology in detecting and extracting the initial welding position

A method of image morphology in detecting and extracting the initial welding position during the autonomous welding process is described. During the process, firstly visual sensing technology is used to capture the straight seam image, and secondly the image edges are detected by morphological corrosion edge detection algorithm, with which can retain the critical information while filter other interferences effectively at the same time. Then morphological processing algorithm is used to conduct the direction of filter by selecting the multidirectional linear structuring elements and finally get the initial weld position point coordinates with the Hough transform. The algorithm is simple, rapid, self-adaptability with high accuracy for interferences except long lines so as to accomplish the entire process of detecting the initial welding position. It can meet the practical demands of automatic guidance for robotic welding.     

驾驶人手机通话行为中基于图像特征决策融合的手势识别方法

为鲁棒检测自然环境中驾驶人的通话行为, 提出了一种驾驶人手机通话手势的识别方法。运用Adaboost算法检测驾驶人面部区域, 在YC_gC_r色彩空间中分别对面部肤色亮度分量和色度分量进行稀疏网格间隔采样, 由此建立了肤色的高斯分布模型; 针对驾驶室光照强度的不均匀性, 提出了肤色分量的漂移补偿算法, 建立了适应光照变化的在线肤色模型, 以准确分割左右手部肤色区域; 运用HOG算法获取手部肤色区域的2 376维HOG特征向量, 运用PCA方法将HOG特征降至400维; 同时提取手部肤色区域的PZMs特征, 并采用Relief算法筛选出权重最大的8个PZMs特征向量, 建立了融合PCA-HOG特征和Relief-PZMs特征的通话手势支持向量机分类决策。试验结果表明: 基于PCA-HOG特征的手势识别率为93.1%, 对光照变化的鲁棒性较好, 但易受到手部与头部转动的干扰; 基于Relief-PZMs特征的手势识别率为91.9%, 对于头部与手部姿态的耐受度较好, 但光照鲁棒性较差; 基于PCA-HOG和Relief-PZMs多元特征融合方法的手势识别率达到94.5%, 对光照波动、手部与头部转动等干扰条件具有较好的适应性。      

High temperature indentation test to improve constitutive model for welding simulation

The high temperature gradients experienced during fusion welding lead to a local dilatation and metallurgical transformations that generate inhomogeneous plastic deformation, residual stresses and distortions in the welded parts which can affect the service life of the structures. To predict such residual stresses and distortions, finite element analysis is nowadays widely used. This work reports a 3D finite element model for welding simulation. The proposed model is based on semi coupled thermo-mechanical analysis using a double ellipsoidal model of heat source. Firstly, a disk heated in its central zone is investigated. Secondly the validation concerns the simulation of a fully 3D two pass butt weld. The idea in this work concerns the calibration of the heat source parameters with the help of the inverse analysis to improve the heat flow predictions. For the stress analysis, an annealing temperature is introduced to force the material to lose its hardening memory above a given temperature. The predicted residual stresses as well as the predicted distortions are found to be sensitive to the annealing temperature.     

Analytical Solution for Model-Based Dynamic Power Factor Measurement in AC Resistance Spot Welding

On the basis of welding transformer circuit model, a new measuring method was proposed. This method measures the peak angle of the welding current, and then calculates the dynamic power factor in each half-wave. An artificial neural network is trained and used to generate simulation data for the analytical solution, i.e. a high-order binary polynomial, which can be easily adopted to calculate the power factor online. The tailored sensing and computing system ensures that the method possesses a real-time computational capacity and satisfying accuracy. A DSP-based resistance spot welding monitoring system was developed to perform ANN computation. The experimental results suggest that this measuring method is feasible.     

基于车载视觉的行人检测与跟踪方法

为提高城市交通环境下车辆主动安全性,保障行人安全,提出了基于车载视觉传感器的行人保护方法. 利用Adaboost算法实现行人的快速检测,结合Kalman滤波原理跟踪行人,以获取其运行轨迹.该方法利用离散 Adaboost算法训练样本类Haar特征,得到识别行人的级联分类器,遍历车载视觉采集的图像,以获取行人目 标;结合Kalman滤波原理,对检测到的行人目标进行跟踪,建立检测行人的动态感兴趣区域,利用跟踪结果分 析行人的运行轨迹.试验表明:该方法平均耗时约80ms/帧,检测率达到88%;结合Kalman滤波原理跟踪后,平 均耗时降到55ms/帧,实时性较好.      

Spatter rate estimation of GMAW-S based on partial least square regression

This paper analyzes the drop transfer process in gas metal arc welding in short-circuit transfer mode （GMAW-S） in order to develop an optimized spatter rate model that can be used on line. According to thermodynamic characters and practical behavior, a complete arcing process is divided into three sub-processes： arc re-ignition, energy output and shorting preparation. Shorting process is then divided as drop spread, bridge sustention and bridge destabilization. Nine process variables and their distribution are analyzed based on welding experiments with high-speed photos and synchronous current and voltage signals. Method of variation coefficient is used to reflect process consistency and to design characteristic parameters. Partial least square regression （PLSR） is utilized to set up spatter rate model because of severe correlativity among the above characteristic parameters. PLSR is a new multivariate statistical analysis method, in which regression modeling, data simplification and relativity analysis are included in a single algorithm. Experiment results show that the regression equation based on PLSR is effective for on-line predicting spatter rate of its corresponding welding condition.     

基于Marc和遗传算法的焊接工艺参数优化

焊接变形的大小主要受焊接速度、焊接电流、电压、送丝速度、焊接顺序等工艺参数的影响.根据焊接热加工特点和实际需要,针对设计变量数目繁多、类型多样的焊接工艺,采用遗传算法与有限元分析相结合的方式,对重点参数进行了优化.以二维平板对接焊的模型为例,以工件最小焊接变形量为优化目标,经过遗传算法进化后,结合有限元方法进行每组参数组合的焊接过程热力耦合分析,找出焊接能量和焊接速度的最优组合,为焊接工艺的优化和制定提供科学的评价和预测.     

A Generalized Rough Set Modeling Method for Welding Process

Modeling is essential, significant and difficult for the quality and shaping control of arc welding process. A generalized rough set based modeling method was brought forward and a dynamic predictive model for pulsed gas tungsten arc welding (GTAW) was obtained by this modeling method. The results show that this modeling method can well acquire knowledge in welding and satisfy the real life application. In addition, the results of comparison between classic rough set model and back-propagation neural network model respectively are also satisfying.     

基于最大信噪比算法在缸盖振动信号中的分离研究

对柴油机缸盖表面振动信号进行分析处理,可以判定内部零部件的状态以及柴油机的工作状态,但实际上由于各路信号时域和频域有混叠,很难准确分离．针对现有内燃机监测诊断信号分析方法的不足,研究了基于最大信噪比的盲源分离算法．该算法将求优过程转化为广义特征值问题求解,整个过程无须迭代．通过对仿真振动信号的分离表明其有效性,从试验对象某四缸柴油机上拾取实测缸盖振动信号处理,结果显示了该算法在柴油机故障监测诊断上有良好的应用前景．     

Modified gray level difference-based thresholding segmentation and its application in X-Ray welding image

Thresholding is a popular image segmentation method that often requires as a preliminary and indispensable stage in the computer aided image process, particularly in the analysis of X-ray welding images. In this paper, a modified gray level difference-based transition region extraction and thresholding algorithm is presented for segmentation of the images that have been corrupted by intensity inhomogeneities or noise. Classical gray level difference algorithm is improved by selective output of the result of the maximum or the minimum of the gray level with the pixels in the surrounding, and multi-structuring of neighborhood window is used to represent the essence of transition region. The proposed algorithm could robustly measure the gray level changes, and accurately extract transition region of an image. Comparisons with other approaches demonstrate the superior performance of the proposed algorithm.     

列车轮对状态的融合监测系统

为了提高列车轮对故障诊断准确率和改善现有列车轮对状态在线监测方法的不确定性,结合多传感器信息融合原理,设计了列车轮对融合监测系统,采用特征层融合自适应加权算法进行了轮对状态融合监测,以自适应的方式寻求最优加权因子,使状态测量值总均方误差最小,比较了特征层融合自适应加权算法、模糊数据关联算法、变结构多模的状态估计算法和BP神经网络算法的计算结果。比较结果表明:当轮对两端轴承均出现故障后,两传感器输出的测量值分别为22.0470和21.0250,而此融合算法计算出的估计值为4.2642,融合值最接近真值,因此,列车轮对融合监测系统可靠性高,抗干扰性强。     

动车组铝合金车体焊缝质量等级评价的应力因数计算方法

建立了某型动车组铝合金焊接车体的有限元分析模型，对焊缝部位进行简化建模，焊缝与实际存在的差异在等效结构应力计算中进行修正；基于标准BS EN 12663—1:2010分析了车体承受的载荷，采用Box-Behnken正交矩阵设计确定了车体的9个疲劳载荷工况；对车体有限元模型施加多轴载荷，分析了车体侧墙上的4条长焊缝部位的应力分布，确定了6个应力因数计算的关注点；采用名义应力法和等效结构应力法计算车体侧墙焊缝的应力因数，对比分析了2种应力分析方法。分析结果表明:2种应力分析方法在循环次数为1.0×107的许用应力范围不同，名义应力为16.40 MPa，等效结构应力为26.61 MPa；6个关注点的名义应力范围均小于其等效结构应力范围，得到的车体焊缝6个关注点的名义应力和等效结构应力的应力因数分别为0.33、0.25、0.50、0.49、0.76、0.62和0.23、0.24、0.39、0.45、0.61、0.48；针对同一焊缝的关注点，名义应力法计算的应力因数大于采用等效结构应力法计算的应力因数；名义应力法存在很大的分散性，导致应力因数偏大，而等效结构应力法物理意义更明确，计算的应力因数更为合理。      

基于手机传感器的车辆轨迹实时在线压缩方法

随着具有定位功能的各类便携式移动设备的普及,产生了大量的移动目标时空轨迹数据,庞大的数据规模对轨迹数据管理和分析带来了严峻的挑战.?车辆时空轨迹数据压缩算法,通过监测分析车辆在不同运动行为模式下智能手机内置线性加速度传感器和方向传感器的数据变化规律,识别车辆的转向行为和变速行为,并根据识别结果请求GPS传感器定位,记录...     

Multi-Step Model Predictive Control Based on Online Support Vector Regression Optimized by Multi-Agent Particle Swarm Optimization Algorithm

As optimization of parameters affects prediction accuracy and generalization ability of support vector regression (SVR) greatly and the predictive model often mismatches nonlinear system model predictive control, a multi-step model predictive control based on online SVR (OSVR) optimized by multi-agent particle swarm optimization algorithm (MAPSO) is put forward. By integrating the online learning ability of OSVR, the predictive model can self-correct and adapt to the dynamic changes in nonlinear process well.     

Branch-Activated Multi-Domain Convolutional Neural Network for Visual Tracking

Convolutional neural networks (CNNs) have been applied in state-of-the-art visual tracking tasks to represent the target. However, most existing algorithms treat visual tracking as an object-specific task. Therefore, the model needs to be retrained for different test video sequences. We propose a branch-activated multi-domain convolutional neural network (BAMDCNN). In contrast to most existing trackers based on CNNs which require frequent online training, BAMDCNN only needs offline training and online fine-tuning. Specifically, BAMDCNN exploits category-specific features that are more robust against variations. To allow for learning category-specific information, we introduce a group algorithm and a branch activation method. Experimental results on challenging benchmark show that the proposed algorithm outperforms other state-of-the-art methods. What’s more, compared with CNN based trackers, BAMDCNN increases tracking speed.