This paper addresses the problems involved in the automatic monitoring of the weld quality produced by robotized short-arc welding. A simple statistical change detection algorithm for the weld quality, the repeated Sequential Probability Ratio Test (SPRT), was used. The algorithm may similarly be viewed as a cumulative sum (CUSUM) type test, and is well-suited to detecting sudden minor changes in the monitored test statistic. The test statistic is based on the variance of the weld voltage, wherein it will be shown that the variance decreases when the welding process is not operating under optimal conditions. The performance of the algorithm is assessed through the use of experimental data. The results obtained from the algorithm show that it is possible to detect changes in weld quality automatically and on-line.

On-line quality monitoring in short-circuit gas metal arc welding

The gas metal arc welding (GMAW) process involves arc plasma, metal transfer, and weld pool phenomena. In addition, metal vapor is formed by evaporation from the high-temperature metal and mixes with the arc plasma. These phenomena interact with each other and are very complicated. A numerical approach that includes their interaction is therefore important for clarifying the GMAW phenomena. We have developed a unified model and used the model to investigate the influence of the shielding gas on the metal transfer. When argon shielding gas was used, for an arc current of less than 230 A, globular transfer occurred. For higher currents, spray transfer occurred. On the other hand, when Ar + 18 %CO2 gas was used, the transition from globular to spray transfer occurred at around 280 A. This difference was caused by changes in the driving force exerted on the molten metal by the arc plasma. The arc pressure that lifts up the molten metal and interferes with its detachment from the wire tip becomes stronger when the gas mixture is used.

Numerical simulation of GMAW process using Ar and an Ar–CO2 gas mixture

Controlled metal transfer implies controllable heat and mass in- puts and improved weld quality. Success of previous methods has been hindered by uncertainty in droplet detachment. In the proposed approach, an oscillation of the droplet is excited by switching the current from a peak level to a base level to initiate an active detachment period. The droplet oscillation is then monitored. When the droplet moves downward, the current is switched back to the peak level, which creates an ejection action and concludes the active detachment period. The droplet is detached by down- ward momentum and the increased cur- rent-induced electromagnetic force. Growth period of the succeeding cycle begins. The utilization of downward mo- mentum eliminates the need for a very high current to detach the droplet, which is conventionally used in pulsed gas metal arc welding (GMAW-P). This elim- ination guarantees that no unexpected detachment occurs during the growth period. Also, reliable high-frame rate vision monitoring of the oscillating drop- let guarantees that downward momen- tum will correspond with the ejection action. Hence, the metal transfer process can be precisely controlled. Experimen- tation has been conducted to demon- strate the effectiveness of the proposed approach in achieving the desired metal transfer process.

Active Metal Transfer Control by Monitoring Excited Droplet Oscillation

One of the requirements of a sensing system for feedback control of gas metal arc welding (GMAW) is the capability to determine the metal-transfer mode. A series of sensing experiments was performed during which the audio emissions, welding current fluctuations and welding voltage fluctuations ere recorded as a fuction of the transfer mode

Process control of gmaw : sensing of metal transfer mode

The gas metal arc welding (GMAW) process combines aspects of arc plasma, droplet transfer, and weld pool phenomena. In the GMAW process, an electrode wire is melted by heat from an arc plasma, and molten metal at the wire tip is deformed by various driving forces such as electromagnetic force, surface tension, and arc pressure. Subsequently, the molten droplet detaches from the tip of the wire and is transferred to the base metal. The arc plasma shape changes together with the metal transfer behavior, so the interaction between the arc plasma and the metal droplet changes from moment to moment. In this paper, we describe a unified arc model for GMAW, including metal transfer. In the model, we do not account for heat transfer in the metal, but the wire melting rate is determined by the arc current. The developed model can show transition from globular transfer at low currents to spray transfer at higher currents. It was found that electromagnetic force is the most important factor at high currents, but surface tension is more important than electromagnetic force at low currents in determining the transfer mode.

Numerical simulation of metal transfer in argon gas-shielded GMAW

This study reports the changes in droplet-trans fer mode and rate during gas metal arc welding as the voltage is varied at a series of current levels. The droplet-transfer rate was found to be maximum (approximately 100 s_1) for the voltage/current combi­ nations that are normally suggested by the electrode manufacturers and are con­ sidered optimum in the judgment of experienced welders. At voltages above or below the TV-wide optimum range, the transfer rate decreased by about 10 s_1 per V in the vicinity of the optimum condition. Furthermore, statistical analysis of the arc current and voltage data showed that during operation outside the optimum range, the welding arc was unstable and the current output was very irregular with varying cycle time between each droplet transfer. At the maximum droplet-transfer rate, the droplet-transfer cycle time was very consistent and revealed a narrow rate range, which correlated with the high stability and lower spatter at these optimum operating conditions. The possibility of using the concept of maximum droplet-transfer rate range with minimum rate fluctuation and corresponding arc current-voltage signals as a means of short-circuiting welding process control and automation is being considered. At voltages below the optimum range, high-speed video recording confirmed that the short- circuiting transfer was very unstable and the arc reignited explosively. Above the optimum voltage, the arc became longer and the droplets became visibly larger, with mixed globular and short-circuiting transfer. The droplets, however, were no longer directed uniformly to the weld pool, resulting in increased spatter. arc (FCA), submerged arc (SA) and gas tungsten arc (GTA) welding. With the exception of GTA welding, all these pro­ cesses require a consumable electrode, which has the dual function of carrying the current that heats the weld pool and providing filler metal to complete the weld joint. This dual function has long been a topic of research. Spraragen and Lengyel (Ref. 1) reviewed the basic princi­ ples of an electric arc and summarized the development of the field of welding arc physics. In particular, they concluded that in the area of liquid metal transfer from the electrode to the weld pool, the electromagnetic pinch force, gravity, shielding gas drag force and surface ten­ sion are the major forces that act on the electrode tip. Using high-speed cinemat­ ographic techniques, Muller, Greene, and Rothschild (Ref. 2) found that large spher­ ical liquid-metal droplets in a GMA arc decreased in size with increasing current. As the electrode feed rate was continu­ ously increased, however, a sudden decrease in droplet size occurred at what was termed the transition current. In addition, they determined that with inert gas shielding, the droplet composition remained constant during the metal trans­ fer. Lesnewich (Refs. 3-5) investigated the physics of arc welding using SMA and CMA welding. Particularly, he studied the effects of welding process parameters such as current, voltage, electrode polar-

T. A. Siewert

Papers

Metal Transfer in Gas Metal Arc Welding: Droplet Rate A study of droplet rates for various transfer modes verifies optimum operating parameters