Advances in BIW welding – reduced RSW expulsions and improved 3D vision systems

1. A comprehensive and on-going approach to reducing expulsion rates from resistance spot welding.

2. Current capabilities of 3D laser vision systems for improving welding and brazing quality.

Takeaway #1

Reducing resistance spot welding expulsion rates through process analysis and smart data

The goal of an on-going study by BMW MINI UK and TWI Ltd (The Welding Institute, UK) is to reduce welding expulsion, an expensive problem caused when hot liquid metal from the spot weld lands on another part of the car. Also known as weld splash or liquid metal ejection, expulsions lead to:

• Burning of zinc coatings, reducing corrosion protection
• Visual defects on visible car parts

Identifying and reworking weld expulsion-damaged parts is very time-consuming.

Decreasing already low expulsion rates

The BMW Group’s MINI plant at Oxford already had a low 3.7% expulsion rate. But the goal of UK government’s WeldZero project is zero weld defects. With funding from WeldZero, MINI Plant Oxford and The Welding Institute systematically worked to bring the plant’s expulsion rate even lower for each of the 6000 resistance spot welds (RSWs) found on every MINI model.

MINI Plant Oxford uses state-of-the-art, high-end robots and welding guns, with integrated/adaptive controls for all spot welds — so it has no issues regarding poor quality, undersized, or low-strength welds. The only remaining issue is weld splash.

In the on-going study, data analysis is used to identify:

• Weld expulsion occurrences
• Their root causes
• Data patterns for each root cause

To inform production engineers of appropriate corrective actions.

Initial corrective measures for weld splash

• Reducing the weld gun air pressure, which was originally and intentionally “over-pressured” in the mistaken belief that overpressure would compensate for the differences in distances from the air supply to the weld gun. Lowering the air pressure, it turns out, better equalized the pressure rate for all weld guns, reducing weld expulsions while reducing air supply energy costs by 25%.
• Monitoring cooling water flow to the weld guns to discover blockages or drops in flow levels. Disruptions to the cooling water flow caused the welding electrodes to overheat and experience excessive electrode wear — leading to weld splash.
• Welding process data analysis of welding voltage, welding current, welding force, and measured resistance, which was then correlated to the welding robots having the highest occurrence of weld expulsions.
  
  

Additional weld expulsion factors found by data analysis

Then TWI and BMW performed a data analysis of the remaining weld expulsions to determine their leading factors:

• Panel mismatch: including poor part shape and springback, but also issues such as parts being pushed out of position by other parts.
• Effect of gaps: for higher strength and thicker components — or ones with three or four sheet stack-ups — gaps between panels can destabilize the welding process, leading to weld expulsion.
• Electrode edge distance: panel mismatch can cause the welding electrode to be too near to the edge of a panel, causing the welding zone to “break out” of the panel edge — with heavy expulsion as the result.
• Poor weld spot orientation: caused by panel mismatch or shape defect and resulting in the electrode being other than its ideal 90-degree orientation. This situation leads to weld expulsion, as well as electrode deterioration which, in turn, can cause more expulsions.
• Heavily-worn electrode tips: the amount of zinc alloying on the electrodes influences expulsion rates.
• Poorly-designed water cooling of weld guns: blockages or tight bends in the cooling channels restricted water flow, causing the electrode to overheat and wear out too quickly, leading to weld expulsions.

Determining the importance of each factor in weld splash

TWI then set up a robotic welding cell to simulate the BMW production processes in their own laboratory. They could then identify the process tolerances for each of the above-listed factors to determine the severity of each factor required to cause a weld expulsion.

They also identified weld process data signatures from the welder timers to diagnose the cause of expulsion for each case.

The “smoking gun”: electrode tip condition

The study determined the sensitivity of each expulsion factor was dependent on the electrode’s current condition. Electrode tip wear also influenced the expulsions’ data signatures.

Figure 2: Some of the process data analysis linking electrode wear condition to weld expulsion. Image courtesy of TWI Ltd and BMW MINI UK.

Next steps for the on-going study

“To achieve an in-process data analysis system that can diagnose the cause of [weld] expulsion, a model should be developed that can take into account the number of welds made with a set of [electrode] tips since the last dressing operation…A tool is under development to provide on-line, real-time identification of welders with unacceptable expulsion levels and diagnostics of the expulsion cause so that problems can be efficiently rectified.”

Takeaway #2

Current capabilities of 3D laser vision systems for improving welding and brazing quality

The penalty for having less-than-perfect welds on automotive parts with high safety requirements — such as EV battery enclosures — is significant. But as reported by Servo-Robot Corp., 3D laser vision camera systems can improve the productivity and quality of robotic laser welding, laser brazing, and arc welding for car components — including Body in White (BIW), chassis, and EV battery protective structures.

Cameras with rates of 2 kHz (2000 frames per second) or higher can be integrated into high-power (up to 30 kW) laser heads. These cameras — placed within 20 mm of the laser’s focus spot — can then track seams during high-speed welding, even on curved shapes, while also providing real-time process monitoring and post-weld inspection.

Figure 3: Tailer-welded blank welding with 3D laser camera seam tracking and inspection. Image courtesy of Servo-Robot Corp.

Large car parts + large welding robots = larger deviations from weld joint

Laser tailor-welded blanks (TWBs) for large parts — such as body side frames and door inner panels — require large welding robots. Because of the large scale, the robots may not be able to keep the laser spot within 100 microns of the weld joint. High-speed, laser-vision, seam-tracking cameras with zero backlash actuators can overcome this challenge, providing precision tracking at welding speeds of up to 15 meters/minute.

Combining cameras and software with a back-reflection sensor — part of a Laser Process Control System (LPCS) — provides detection of both surface and internal weld defects, with results immediately sent to the welding robot to prevent further bad welds.

The Laser Process Control System’s back-reflection sensor measures the thermal radiation emitted from the molten pool keyhole area. Changes to the absorption of the laser power by the weld joint and its molten pool indicate internal defects in the weld. For tailor-welded blanks, the defects caused by conditions such as gap variation, blank edge differences, contamination, or insufficient laser energy are detected in real time.

Laser and laser-hybrid welding of BIW

Similar 3D laser vision systems can be used for laser and laser-hybrid welding of BIW components, such as in the joining of the vehicle roof and body. They are also being used for laser and laser-hybrid welding of EV battery enclosures (aka cages, protective frames), including aluminum stitch welding and steel panel welding.

Checking weld bead geometry of laser-brazed roof seams

Laser-brazed seams, such as brazed joints for roofs, can also benefit from two-camera vision systems. The first camera finds and tracks the joint, locating its centerline. The second camera measures the bead’s geometry, finding surface defects to a resolution of 0.1 mm. Checking the internal soundness of brazed joints requires the use of a back-reflection sensor.

Getting weld wire where it’s needed

The biggest cause of quality issues and lost productivity in arc welding is not the process itself but instead the simple fact that the weld wire is not positioned correctly in the weld joint. The most common reasons for improper wire placement are detailed part variability, stack up of tolerances in an assembly, and distortion caused by the weld heat input which results in the joint not being where the robot is programmed to travel.

Seam finding with 3D vision systems locates the actual joint position. Then the wire position is modified to fit the real-world location of the joint to achieve optimum weld quality. In addition, if an unacceptable gap is present, an adaptive weld schedule can be used to open the process window of acceptability by changing the welding current, travel speed, or weaving pattern.

Figure 6: 3D cameras doing seam tracking for the welding of chassis siderails. Image courtesy of Servo-Robot Corp.

Automated visual arc weld inspection for Industry 4.0

Fully automated, robotic arc weld inspection is not only faster and more reliable than manual inspection — it also provides valuable data that can fulfill Industry 4.0 requirements. This data can help predict when an arc weld is trending towards becoming defective — as well as possibly indicate to what needs to be improved, such as part quality, fixture repeatability, or the welding process itself.

The inspection of arc welds is more difficult to perform than laser welds or brazing due to the larger variabilities in arc weld size, regularity, and surface smoothness.

For this reason, the most successful approach for arc weld inspection is to use a comparative approach: you establish the nominal weld quality on a “golden part” and then compare the actual production welds to it, looking for unacceptable differences. Excessive differences will indicate a welding operation that is not in control and thus more likely to produce defective welds.