Automated optimization of weld seams

WWW.THEFABRICATOR.COM APRIL 2009

April 28, 2009

By:

Automated vision inspection and optimization technology can help streamline manufacturing operations and improve part quality. This technology detects defects in welding seams that often are missed by manual inspection and provides documentation that can be used in researching and resolving product issues.

Automate Inspection System

Figure 1Automate Inspection System

Emerging technologies for automated welding inspection have reached new levels of sophistication in recent years. While automated technology systems for use in weld seam inspection have been in practice in Europe for more than two decades, automated optimization followinginspection is an operation many manufacturers and fabricators worldwide are taking a closer look at.

Next-generation Inspection and Optimization

Most manufacturers and fabricators agree that streamlining weld seam inspection and optimization on production lines is a top operations priority. Human inspection is inherently subjective and can be error-prone, even when conducted by the most skilled inspectors. But how efficient are automated technologies, and how do they operate?

New-generation automated weld inspection and optimization technologies are based on a laser triangulation method.

A semiconductor laser, housed in a compact sensor, projects a line across the welded seam. A high-speed camera, also housed in the sensor, captures the line as an elevation profile. A 3-D image of the welded seam surface then is created from synchronizing the relative motion of the sensor and the object. Any relevant deviations of a compliant welded joint can be identified with accuracy and objectivity. Figure 1depicts a laser line stripe across a weld.

This image acquisition method permits entire classes of defects to be identified by sophisticated pattern-matching software algorithms. Similar to the process of optical character recognition (OCR), in which software translates an acquired envelope image's address into data, modern weld seam inspection (WSI) systems also capture images. From the acquired image, the software identifies defects that otherwise might be missed using manual inspection methods. With new inspection and optimization systems, these defects are inspected and, if needed, optimized then inspected again, a process that can help manufacturers avoid defects and product damages.

WSI systems can be characterized by major components with diverse functions. The first component is the image acquisition unit. Image acquisition is collected from sensors that include both imaging hardware, such as a charged couple device (CCD), and a laser.

The image acquisition hardware is in turn hosted by a computing platform. This platform includes necessary hardware to communicate with the sensor, the network (if any), and the operator.

The last major component is the inspection software employed when images from welds are collected. The software is used to identify errors objectively. This is important in environments in which detailed specifications are used to discern goodwelds from badwelds.

Specifications typically cover some or all of these important weld characteristics:

  • Porosity/Holes
  • Weld Length
  • Weld Volume
  • Throat Thickness
  • Weld Width
  • Misalignment
  • Weld Height (A-height)
  • Excessive Convexity
  • Incomplete Filling
  • Grooves
  • Undercuts
  • Excessive Spatter
  • Burn-through

Indications of errors can be difficult to spot with the human eye. For example, the weld on the coupon in Figure 2fails because of weld height.

Figure 3shows how the gray-scale image of this weld appears to the human eye. The elevated portions are visible, but still difficult to see. If this same scan is submitted to the software, the height module creates an image that is intuitive to an inspector and clearly shows where the errors appear (Figure 4).

In this particular tool, the white and blue line maps a location value so that the location of the defects can be seen more clearly. This information helps reduce rework time because no effort is expended to locate the errors.

Automotive Industry Moves to Automated Weld Seam Optimization

The automated and optical welded seam inspection process has been used for simple inspection tasks in industrial production for years. But in vehicle production, manufacturers often look to go one step further, putting into practice an automated optimizing line cycle process.

Many of the leading auto manufacturers have committed themselves to breaking new ground with innovative engineering and quality standards. For example, production lines house components in several different assembly stages with each line of operation using similar production methods. Components with varying numbers of seams are welded together in a separate welding booth, and the welded assembly is inspected automatically during the next stage of production. If a defect is detected, the results are analyzed for corrections.

One European engineering plant that implemented automated inspection and optimization of welded seams reported that millions of welded seams have been successfully inspected and, if required, repaired. This type of approach brings automation to nearly 100 percent accuracy, because of application of the information collected.

Engineering experts agree that automation in weld seam inspection gives auto manufacturers added safety measures during the assembly process by enabling them to archive all relevant weld data for each automobile that moves off the assembly line. This information can be used as an audit trail for quality control and continuous improvement.

According to industry experts, advances in automated welding technology continue to make significant contributions in terms of cost savings to the U.S. automotive industry and improve vehicle craftsmanship.



Robert S. King III

Contributing Writer
Vitronic Machine Vision LTD.
Phone: 502-266-2699

Related Companies

More in Testing and Measuring from TheFabricator.com

Published In...

www.thefabricator.com

Read more from this issue

comments powered by Disqus