基于嵌入式ARM和Linux的电阻点焊监控器的研制

发布时间：2018-06-26 01:15

本文选题：电阻点焊 + 过程监控　；参考：《哈尔滨工业大学》2017年硕士论文

【摘要】：电阻点焊作为一种生产效率高、实现简单、成本低的连接方法,广泛应用于汽车生产中。但由于焊接工艺选择不合适、焊接参数波动、工人操作不当等原因,往往导致现场焊接质量不稳定。由于焊点质量检测往往需要抽样破坏性检验,此举既浪费成本又不具代表性,因此实现电阻点焊过程在线监测是工业现场的迫切需求。本文首先介绍了课题组针对实际现场需求,以STM32F303ZET6单片机及EasyARM-i.MX283开发板套件分别作为系统工作核心和计算交互核心,制作的一套能够稳定应用于电阻点焊生产现场的监控设备。该监控设备具有焊接过程多路模拟量信号采集功能,并能将采集信号进行文件存储并在系统的TFT屏上实时显示;具有焊接生产辅助功能,用户可以将当前工艺信息,如车辆数、焊点编号和组数等进行输入,每次焊接可以对工艺信息进行更新;可监测生产漏焊、多焊等情况,并可通过语音进行提示;另外,系统还具有以太网通信功能,可以通过上位PC机监测系统工作状态,为车间的网络化生产管理提供条件。之后介绍了点焊监控器在一汽红旗焊接现场的功能验证。经验证,系统在现场的交直流电阻点焊机均能稳定地完成信号采集、焊点信息与焊接参数曲线交互、焊点文件生成、以太网通信等功能。在焊点文件存储上,对直流电阻焊机,可以对每个逆变半波的焊接电流、电压、电阻、通流比等信息进行存储,也可以在80k Bit/s速度下完成焊接瞬时值的存储;对交流电阻焊机,可以对焊接电流、电网电压和原边电压以80k Bit/s速度下存储,并在文件末尾存储焊接过程的触发角和导通角数值。本文最后介绍了对中频直流电阻焊简化计算模型的推导和对现场采集数据的初步分析。在模型推导上,得出理想恒占空比、恒压、电阻电感恒定条件下的焊接电流计算公式。根据实验和经验数据,估算电流达到稳态时间约为10ms。推导了电感剔除式的动态电阻算法,通过数据仿真和实际数据验证,其电阻计算精度在电流未稳定时优于传统平均值算法,且该方法计算回路电感值在焊接过程基本保持稳定。通过计算不同焊接条件电感值,发现电感值对于焊接条件不敏感,不同电感相对中心值波动范围为-0.61%~1.52%。铁磁物伸入条件下平均电感值显著增加,较正常焊接条件平均电感值中心提升2.26%~3.76%。在数据初步分析上,比较了不同材料的动态电阻曲线特征,并选取电阻具有明显峰谷值特征的DP240材料进行多次试验,分析了焊接电流、工件层数和板厚对焊接过程的影响。介绍了基于动态电阻的飞溅识别和飞溅特征值计算方法,并通过提取动态电阻初始电阻、初始谷值时刻、初始谷值电阻等飞溅发生前的特征值,利用GRNN、PNN、SVM算法,对飞溅的产生进行预测。改变训练集和测试集分别运行500次,得到各算法判断正确率达87.5%以上的次数占总运行次数的比例分别为92.2%、95.0%和89.2%,很好的完成了焊接飞溅的预测工作。
[Abstract]:Resistance spot welding is widely used in automobile production as a connection method with high production efficiency, simple realization and low cost. However, due to improper selection of welding process, fluctuation of welding parameters and improper operation of workers, the quality of welding is often unstable. It is not only a waste of cost and non representativeness, so it is an urgent need to realize the on-line monitoring of resistance spot welding process. This paper first introduces a set of STM32F303ZET6 singlechip and EasyARM-i.MX283 development board suite as the core of the system and the core of the calculation interaction. The monitoring equipment is used in the field of resistance spot welding production. The monitoring equipment has multi analog signal acquisition function in welding process, and can store the collected signal and display it on the TFT screen in real time. With the auxiliary function of welding production, the user can make the current information of the work, such as the number of vehicles, the number of solder joint and the number of groups. In line input, the process information can be updated each time, the production leakage welding and multi welding can be monitored. The system can be prompted by voice. In addition, the system also has Ethernet communication function. The system can monitor the working state of the system through the upper PC machine and provide conditions for the network production management of the workshop. Then the spot welding monitor is introduced. It is verified that the AC and DC resistance spot welder in the field can accomplish the signal acquisition, the interaction of welding point information and the welding parameter curve, the welding point file generation, the Ethernet communication and so on. On the storage of solder joint, the welding current of each inverter half wave can be applied to the DC electric resistance welding machine. The voltage, resistance, flow ratio and other information are stored, and the welding instantaneous value can be stored at the speed of 80K Bit/s. For the AC resistance welding machine, the welding current, the grid voltage and the original side voltage can be stored at the speed of 80K Bit/s, and the trigger angle and the conduction angle value of the welding process are stored at the end of the file. The derivation of simplified calculation model of frequency DC resistance welding and the preliminary analysis of data collected on the spot. In the model derivation, the formula of welding current calculation under constant constant duty ratio, constant pressure and constant resistance inductance is obtained. According to the experimental and empirical data, the estimation of the dynamic resistance of inductor elimination is derived from the estimated current when the current reaches the steady state of 10ms.. The algorithm, through data simulation and actual data verification, its resistance calculation precision is better than the traditional average value algorithm when the current is not stable, and the method calculates the inductance value of the loop in the welding process basically stable. By calculating the inductance value of different welding conditions, it is found that the inductance value is insensitive to the welding condition and the different inductance is relative to the central value wave. The average inductance value of the -0.61%~1.52%. ferromagnetic field is significantly increased and the average inductance value center of the normal welding condition is raised by 2.26%~3.76%.. The dynamic resistance curve of different materials is compared with the average inductance value center of the normal welding condition, and the DP240 material with the characteristic of obvious peak and valley value is selected for multiple tests and the welding welding is analyzed. The influence of the current, the number of workpiece layers and the thickness of the plate on the welding process. The method of spatter identification based on dynamic resistance and the calculation method of spatter characteristic value are introduced, and the characteristics of the splash before the initial resistance of dynamic resistance, the initial valley value time, the initial valley value resistance and so on are extracted. The GRNN, PNN and SVM algorithms are used to predict the generation of spatter. The training set and the test set run 500 times respectively. The ratio of the number of times to the total operation times is 92.2%, 95% and 89.2% respectively, and the prediction of the welding spatter is completed.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TG438

【参考文献】