機(jī)器人焊接的系統(tǒng)問題-外文文獻(xiàn)

《機(jī)器人焊接的系統(tǒng)問題-外文文獻(xiàn)》由會(huì)員分享，可在線閱讀，更多相關(guān)《機(jī)器人焊接的系統(tǒng)問題-外文文獻(xiàn)（5頁(yè)珍藏版）》請(qǐng)?jiān)谘b配圖網(wǎng)上搜索。

1Chapter 2 Robotic Welding: System Issues2．1 IntroductionRobotic welding research deals with the relevant technical and scientific aspects involved in the task of reproducing the work of the experienced and skilled human welder. Welding was for a long time a task performed only by humans, being a craft that combines skill with art and science. Automating welding is therefore a very difficult and demanding objective, because of the required adaptive behavior of the automatic system. It can be considered that any welding operation is constituted by three very different phases:1. Preparation phase: where the welding operator sets up the parts to be welded, the welding apparatus (power source, robot, robot program, etc.) and the welding parameters. The type of gas and the type of wire are also selected in this phase. If any CAD/CAM or other offline programming facility is used then a robot welding pre-program is available and should be placed on-line. This aspect is very important since currently most of the welding pieces are designed using CAD software. Consequently, that software should be used to generate robot programs that could work as starting points for the welding tasks, needing only minor tuning due to calibration. That may be done easily by the welding operator just by performing selected on-line simulations of the process calibrating in this way the robot program that should then be ready for production.2. Welding phase: considering a manual welding operation, the welder acts by adjusting the process variables just by continuously observing the welding operation and the correspondent results. If automatic equipment is used to perform the welding operation, then the same capabilities must be present, i.e., the system should be able to maintain the torch orientation while following the desired trajectory, perform seam tracking and change welding parameters in real-time. With those capabilities available the system should be capable of emulating the adaptive behavior showed by the manual welder.3. Analysis phase: this is normally a post-welding phase where the manual welder, or welding operator in the case of an automatic welding system, examines the obtained welds and decides if they are acceptable or if changes are needed in the two previous phases. In the particular case of an automatic system, this phase can be performed automatically, or by means of user input using specific software interfaces. When advance sensors are used, like laser 3D cameras, this phase can be executed on-line during the welding phase. This is particularly interesting since evaluation of welding 2quality on-line may influence the ongoing welding process. Consequently, when designing a fully automated robotic welding system all the above welding phases must be considered as a way to achieve a good performance and welding quality. The following sections detail some of the relevant problems, namely: modeling and control the welding process, system interfaces and programming environments.2．2 Modeling the Welding ProcessModeling the welding process is basically a theoretical problem (a physics problem mainly) and a technological problem, i.e., understanding the welding process requires theoretical studies but also extensive experimentation to obtain the governing models. Part of the current knowledge on welding is empirical and based on detailed experimentation,which focuses on technological aspects. Consequently, the strategy used in this book was to present the most interesting welding processes from a robotics and automation point of view, focusing on the technological characteristics and automatic system requirements. The physics of the process is briefly introduced and the reader referred to other technical publications, as a way to identify the process parameters relevant for each welding process.2．2．1 Definition and Detection of the Process ParametersTo design a welding robotic system the first step is to identify the process related parameters, i.e., the parameters that should be controlled in a way to obtain the desired quality, also defined by a set of accepted characteristics. The process related input parameters can be classified into three different categories:1. Primary inputs: variables that can be modified on-line during the welding process. Taking as example the GMAW process, the primary welding parameters are the voltage; the wire feed rate, and the torch speed. Technically, the voltage and the wire feed rate are analog signals commanded to the welding power source, and generated from the robot controller or process PLC. The torch speed is the desired speed commanded to the robot TCP for coordinated motion.2. Secondary inputs: variables defined when the process is selected and before any welding service. Using again as example the GMAW processThose parameters include the type or composition of the shielding gas, the flow of gas during the process, the torch angle, and the type and size of the wire to use.3. Fixed inputs: parameters that are fixed and cannot be changed by the user. These parameters are usually an imposition of the selected welding process, of the current welding procedure or of the physical setup. Parameters of this type include the 3joint geometry, plate thickness, physical properties of the plate metal, etc.All these parameters must be handled carefully, namely the correct preparation of the setup and the selection of the secondary inputs are fundamental to control the primary inputs efficiently. Another important set of parameters are the output parameters. Those parameters characterize the weld and are used to evaluate its quality. In a general way, there are two types of output parameters: geometrical and metallurgical.Geometrical parameters result from the process mass balance, and basically define the way how the transferred metal fills the welding joint. Consequently, the basic parameters used to classify an acceptable weld are the penetration, the bead width, the bead height and the cross-sectional area, if we consider a V-Groove weld, and the penetration and the length of both legs for a fillet weld. The penetration is a very important parameter to evaluate the quality of the weld, because it is related to the way the weld metal combined with the base metal during the welding process. Unfortunately, it is very difficult to control the penetration during the welding process since there is no way to measure it on-line. Sensors are used in robotic welding to detect and measure the process features and parameters, namely the joint geometry and the weld pool geometry and location, used for on-line control of the welding process. Nevertheless, sensors are also used to perform weld inspection and quality evaluation.The first basic thing to achieve with a fully automatic robotic welding system is the capability to follow precisely the joint to be welded. This is because the welding quality depends very much on the welding pool position apart from its geometry. A pre-programmed path cannot be obtained with the desired precision, since deviations from the programmed path are likely due to deficient path definition, but also due to material plate deficiencies and to the effect of heating the plates. Consequently, an on-line joint recognition and seam tracking system must be available. Several techniques have been used for joint detection and seam tracking, namely for welding robotic systems. Using the arc characteristics exploiting the proportional relationship between the welding current and the distance from the electrode to the work-piece, as proposed by Cook, was one of the first approaches. All technicians working in welding are familiar with the weaving techniques used to obtain the joint geometrical profile. In fact, the only thing that is needed is a current sensor and a comparison system: setting the reference as the current reading for a perfectly centered torch on the actual welding situation, the center of the welding joint can be obtained just be weaving the arc along the joint and reading the current signal. When the torch is centered the difference between the two signals is zero, and the signal can be used to position the robot carrying the torch.4This approach is slow, does not work well with all types of joints and requires weaving motions which may not be desirable or possible for certain welds. Nevertheless, this is a commercially available system. The utilization of vision systems permits more accurate results. Many researchers tried to use CCD cameras to obtain the joint image and compute the track to follow and even evaluate quality of the weld . Those approaches suffered from CCD saturation due to the light generated by the arc, and interference due to the electromagnetic field also generated by the electric arc, which did not make them ideal for on-line seam tracking. Nevertheless, they could be used for joint detection, for weld pool detection and for quality control, with the selection of an appropriate narrow band filter and/or optimizing the placement of the camera in such a way to avoid the arc light and electromagnetic interference. Since these solutions are not robust when several welding processes are considered, the most used approach with CCD cameras is to make a teaching pass before the welding process is actually initiated. Theoretically this is a good solution since a good reading of the welding seam can be recorded and used to guide the system during the welding process. The drawbacks of this approach are the reduction of arc-on time and the insensitivity to deviations, even if small, of the welding seam that can happen due to the extremely high temperatures characteristic of the welding process and deficiencies on the material of the plates to be welded.The laser based 3D cameras have been used successfully for joint detection, seam tracking and weld inspection. These cameras work in a very simple way, based on the principle of laser triangulation. A low power laser source is used to generate a laser beam that is projected onto the surface of the joint to weld. The reflected light is picked up by a lens that feeds the imaging system composed usually of a CCD or CMOS sensor. The laser reflected signals are extracted using filters and image processing software, which is a simple task since the laser signal has a very precise wavelength and power. In fact these laser cameras and related processing hardware and software, with some customization to the selected application, are very useful for evaluating most of the geometric parameters besides the already mentioned joint detection and seam tracking features. Since they are available with powerful APIs for general use, with standard interfaces for robot controllers and current computer hardware, this type of sensors constitute a powerful tool for robotic welding. Another very important and challenging parameter is the penetration. Basically, a good weld has constant penetration along the weld path, and consequently the welding system should be capable of keeping that goal despite possible variations in the joint geometry. If full penetration is achieved then there 5are some methods to observe the penetration on-line, but in cases of partial penetration there are no means to monitor its evolution. Several methods were designed to measure the penetration when full penetration is achieved, but most of them require backface bead measurements requiring access to the back of the work-piece, which isn’t always possible. Measuring front-face bead geometrical characteristics, along with the weld bead temperature provides the means to estimate the penetration. This means a good understanding of the welding process behavior, so that a precise model, correlating two dimensional measurements of the weld pools with the three dimensional shape of the same weld pool, can be written and used for on-line penetration control. Using ultrasonic techniques is also possible as an alternative, since a full model description is difficult to obtain and, although desirable, has not been achieved yet.Metallurgical aspects are also important for the welding quality, since they determine important mechanical characteristics like hardness, soundness, strength and residual stresses. Those very important mechanical parameters are not easy to measure on-line and are a consequence of several mechanisms. Nevertheless, they all result from the heat generated in the welding process. And since the welding process is basically based on heat, the following is needed to guarantee an acceptable weld:1. A certain peak temperature is needed to achieve a good metal fusion and penetration.2. A roughly uniform temperature distribution, centered in the weld joint, is required to achieve a constant weld.3. Acceptable cooling rates, compatible with the required metallurgical characteristics of the final work-piece, are also needed. All these requirements focus on the need to monitor all the thermal events of the welding process, which adds to the other geometrical measurements requirements that a successful robotic welding system must implement.