Welding Inspection Automated Vision Systems Improve Quality

Automated vision systems for in-process welding and for quality inspection helps to improve the bottom line as well as the quality of the welding process. Vision systems manufacturers agree that the big driver for automated vision systems is the automotive industry as carmakers push their suppliers to tighten controls on their processes and quality.

Vision systems have to be considered separately for welding processes and for inspection applications, Dean Elkins, senior general manager for the Midwest and eastern regions for Motoman, said.

"If we look at vision systems as they relate to the welding process, interest in that has grown from problems in variations in piece parts that customers are getting from their suppliers," Dean Elkins said. "It’s a way to make up for inconsistencies in fixture design, as a means of finding a part and tracking the joint. Most of the time, these systems are structured as a light or a laser-based system."

Laser-based seam tracking offers the benefits of faster tracking of medium to heavy deposition applications. Many times a laser-based process will be tied to a secondary product known as 'adaptive fill'. Dean Elkins said the adaptive fill products are variable, and have the laser looking at weld joint geometry. They adjust welding parameters when needed.

"The vision system looks ahead of the arc so the robot knows what’s coming in terms of joint profile. Then the robot can speed up or slow down, decrease or increase wire feed speed, and increase or decrease deposition of metal to provide more of a lights out approach to welding," Dean Elkins said.

Adaptive fill technology is growing fast in automotive applications such as catalytic converters and suspension components, areas in which we are seeing increased levels of vision based inspection systems. Dean Elkins said vision systems for those parts are increasing as the major automobile manufacturers hold their suppliers to higher and higher process capability standards.

Motoman uses thirdparty providers of vision inspection systems.

One of those is Servo-Robot Inc. which offers two primary systems, ArcScan and LasScan.

ArcScan system

The ArcScan system is a visual weld inspection and measurement system designed for arc welded components. It has been used in applications that range from in-process real time weld quality measurement to offline auditing applications, the company said. In its on-line information, Servo-Robot said objective consistent measurements made with its ArcScan systems are superior to inspection done manually by people, and provide more information than periodic weld cross sectioning. Inspection speeds are in the range of 2 m/ min to 5 m/min (6.5 ft/min to 16.4 ft/min).

Benefits of the ArcScan include ensuring that welds meet quality standards and design requirements; identifying variations in the process; validating the process during prototype and test samples; and ability to measure undercuts, entry angle and toe radius.

LasScan/Bead Scan systems

LasScan/Bead Scan systems are Servo-Robot’s visual weld inspection systems designed for laser brazing processes.

These systems deal with detecting the very small pinholes that can develop in the brazing process to dimensions as large as a car body. For both large and small dimensions, the systems maintain accuracy of position. LasScan systems feature dual sensors that allow for simultaneous geometry detection and small defect detection. Seam finding is integrated into the system to detect and track the locations of both parts and joints.

Servo-Robot said LasScan operates at inspection speeds of to 10 m/min.(32.8 ft/min), and can detect pinholes as small as 0.1 mm (0.003937 in.), while having the ability to inspect small fillet welds and parts that have equal thicknesses.

Jeff Noruk, president Servo-Robot Corp., a U.S. subsidiary of the Canadian-based company, said his company has seen a trend toward more automated vision systems for welding.

"Obviously, automotive was the first industry in which we placed those systems, but the interest is up even outside of automotive," Jeff Noruk said.

However, he added that, in most cases, automated vision inspection for small companies isn’t feasible.

"Shops maybe have one or two robots to handle welding jobs, but they don’t think about the inspection. It’s tougher, realistically, for small shops to implement automated vision. I can’t think of any small shops that have our system," Jeff Noruk said.

It is not that small shops aren’t interested in automated welding inspection, but there’s a misconception that if they adopt automated vision inspection a shop can eliminate its quality person.

"It’s tougher because small shops aren’t as far up the curve. I have to tell them ‘you still need a person in quality who knows welding to understand what the vision system is telling them,’ " Jeff Noruk said. Typically, it is the shop that has 200 or more people and a minimum of five robots that represents the optimum for putting in robotic vision systems, unless a shop has a very critical part and the customer is willing to pay for it, Jeff Noruk added.