059.pdf

Welding Journal | January 2014

WELDING JOURNAL 59 indirect quality assurance by measuring some related parameters. Some measure electrode force dynamics and electrode displacement, others study welding current and voltage changes to predict weld quality based on models of the welding process. Such methods allow one to only predict the weld quality, not measure it, due to the indirect nature of these methods. Ultrasound methods use ultrasonic waves that easily penetrate the metal sheets and bring back information about the internal structure of the spot weld. For this reason, ultrasonic testing was traditionally used for offline weld inspection. From the 1960s, different research groups attempted to develop real-time ultrasonic testing technology (Refs. 2–7). With the relatively recent use of robots for spot welding, along with the introduction of servo motors and tip dressers, spot welding has become a much more stable process to implement real-time ultrasonic inspection inline with production. Today, our research group advanced to the level of installation of half a dozen prototype inline ultrasound units at several assembly plants around the world. The biggest progress has been achieved with our long-term partner Chrysler Corp. at one of its plants in Windsor, Ont., Canada. This aricle describes the current level of technology along with the particular problems and advancements of the system installed at an industrial facility. Inline Ultrasonic Inspection of Spot Welding In resistance spot welding, the metal sheets are joined by means of melting the base metal with high electric current. The current is delivered by two electrodes, which squeeze the sheets together thus developing Joule heat. As the welding gun makes hundreds of welds, the electrodes experience deformation and foreign material pick up. This leads to gradual degradation of the original welding conditions and eventually to the production of unacceptable welds. Periodic current stepping, tip dressing, and electrode cap replacements are routinely used in production. Still, with the introduction of new materials, it becomes harder to predict the tip conditions and to implement timely adjustments. Some means of real-time control become a necessity. In the current inline ultrasound setup, the piezoelectric transducer is installed in the cooling water stream inside the welding electrode — Fig. 1. Sound waves propagate through the cooling water and copper electrode to reach the welded plates. Cooling water is used as a couplant to deliver sound from the transducer to the copper electrode cap. The dry contact between the electrode and metal sheets allows ultrasonic waves to penetrate further due to the high pressure exerted by the electrodes. The sound experiences partial reflections at every boundary, including the solid-liquid boundary of the molten nugget. In mild steel around the melting temperature, the solid metal has an acoustic impedance of 32.7 MRayl, while its liquid state shows 26.5 MRayl. Calculations and experiments show that this impedance mismatch at the liquid nugget’s boundary reflects enough sound energy to be reliably detected. The transducer, which works as both emitter and receiver, collects information from every reflecting boundary on the wave path and forms an A-scan. Proper time gating of the A-scan allows software to analyze the information from the area of interest; the metal sheets and spot weld itself — Figs. 2, 3. Figure 4 shows a schematic view of the ultrasonic signature of the spot welding process, represented as an M-scan. Such an M-scan is composed of multiple Ascans of the same point on the weld captured successively in time. Every A-scan is simply a time-voltage graph acquired by the receiving transducer. The first Ascans begin to shoot before welding starts. In this case, the system works as a simple ultrasonic thickness gauge, receiving reflections from every sheet. One can see interfaces 1, 2, and 3 appearing horizontally and parallel to each other since these are stationary in time. These are reflections off of the copper-steel, steelsteel, and steel-copper boundaries. Scanning continues throughout the welding process (Fig. 4) and some time after it with a time interval of 3 ms between A-scans. When current is turned on, the metal sheets’ temperature increases, which leads to sound velocity reduction. Thus, the back wall reflections from both sheets begin to arrive later in time as temperature increases. When the base metal begins to melt, the steel-steel boundary disappears and so does sound reflection off of it. The liquid metal nugget then grows from the steel-steel boundary into both sheets. Impedance mismatch between solid and liquid steel allows the sound waves to reflect off the top and the bottom of the nugget and thus make them visible on the A-scan and correspondingly on the M-scan (lines 4 and 5 in Fig. 4). The two reflections continue to move apart as the nugget grows. When welding current is shut off, the system begins to cool and the process reverses. Figure 5 presents real ultrasonic scans of underwelded and properly welded spot Fig. 2 — Two electrodes squeeze steel plates before welding (left). Gated A-scan (right). Fig. 3 — Two electrodes squeeze steel plates during welding (left). Gated A-scan (right). Fig. 4 — Schematic M-scan of the resistance spot welding process.


Welding Journal | January 2014
To see the actual publication please follow the link above