Laser Scanner Corrects Aluminum Casting Distortions, Saving Two Months Time

The aluminum casting was left too long in an oven where a heat treating operation was being performed. The oven distorted the geometry to the point that operators for the subsequent machining operation felt that they would be unable to make a good part.

The distortion reached the level that they were not able to even determine an orientation that the part could be placed in that would eliminate the need to add material to achieve required dimensions. The part could have been recast, but that would cost $5,000, take two months, and delay the delivery of the finished product to the customer by nearly that long.

The solution turned out to involve an innovative use of NVision's laser scanner. Technicians used laser scanning to produce a point cloud with 10 million points that accurately described the geometry of the distorted casting.

Technicians used the laser scanner from NVision to produce a point cloud with 10 million points that accurately describes the geometry of the distorted casting. (Note: this image does not show the project discussed in this article.)

Then they imported the point cloud into software that made it possible to compare the actual geometry of the casting to the original solid model that expressed the design intent. The software automatically oriented the distorted part to minimize differences with the design intent while also highlighting the remaining differences. This made it easy to correct the distortion during the normal machining process and ship the product to the customer on time.

Laser Scanning

Laser scanning systems work by projecting a line of laser light onto surfaces while cameras continuously triangulate the changing distance and profile of the laser line as it sweeps along, enabling the object to be accurately replicated.

The laser probe computer translates the video image of the line into 3D coordinates, providing real-time data renderings that give the operator immediate feedback on areas that might have been missed.

Laser scanners are able to quickly measure large parts while generating far greater numbers of data points than probes, without the need for templates or fixtures. Since there is no contact tip on a laser scanner that must physically touch the object, the problems of depressing soft objects, measuring small details, and capturing complex free form surfaces are eliminated.

Instead of collecting points one by one, the laser scanner picks up tens of thousands of points every second. This means that reverse engineering of the most complicated parts can often be accomplished in an hour or two.

Laser scanning is able to reverse engineer parts which are so complex that they would be practically impossible to reverse engineer one point at a time. Finally, the software provided with the scanner greatly simplifies the process of moving from point cloud to CAD model, making it possible in minimal time to generate a CAD model of the scanned part that faithfully duplicates the original part.

Special software can be used to compare original design geometry to the actual physical part, generating an overall graduated color error plot that shows in a glance where, and by how much, surfaces deviate from the original design. This goes far beyond the dimensional checks that can be performed with touch probes.

The NVision laser scanner has been used on a wide array of aircraft projects.

Creating Spare Parts

Laser scanning makes it easier and less expensive to create spare parts that perfectly duplicate the originals. In the past, a major aerospace company used a CMM to capture points one-by-one by touching the probe to the surface of the part.

The primary limitation of a CMM is that the operator must manually move the steering system to track each point to be measured and the device captures points one at a time. But to accurately model the geometry of a complex 3D contour, such as is found on many military parts, you need millions of points, sometimes many millions, to get the geometry exactly right. Laser scanning makes it possible to create solid models from parts with complicated geometries in much less time and to a higher level of accuracy. These solid models can then be used as the basis for CNC programs that can be used to either produce the parts directly or to produce patterns or molds that will be used to produce the parts in subsequent casting operations.

A different challenge

The aerospace company originally purchased the NVision laser scanner to inspect incoming parts with complex geometries that it receives from suppliers.

The company scans the as-built geometry and compares it to the original solid model. The aluminum casting application represented a different type of challenge.

The casting cost about $5,000, so throwing it away would have involved a significant loss. Scheduling was an even more critical concern.

If the aerospace company had thrown the part away, it would have taken six to eight weeks to get a new one from the foundry.

Meanwhile, a product that was ready to deliver, except for that part, would have been sitting idle taking up space. That would have impacted the company's cash flow, plus the customer would not be happy to get the product late.

The part was supposed to be machined after undergoing the stress relieving operation that was the source of the problem. But the distortion induced by the stress relief operation was so great that if the company had tried to machine the part in its normal orientation, it would have been short on stock in several critical areas.

Engineers that looked at the casting were unable to determine a way to salvage the part, but they reasoned that laser scanning might help find a way. It took about two hours to scan the distorted part using the NVision laser scanner. The part was so big that an engineer split it up into sections and scanned each one separately.

He then exported the point clouds generated by the laser scanner in the STL format. The files were then converted into a surface model in Pro/ENGINEER.

Best-Fit Analysis

The engineer performed a best-fit analysis, which oriented the as-built model to match most closely with the design geometry. At this point, he performed a difference analysis, which showed with color codes the areas on the as-built part that were larger than or smaller than the final design.

The goal was to find an orientation in which every area on the as-built part was larger than the design intent. Larger areas can easily be removed during the machining process, while adding metal back to the part, though possible, is much more expensive and might affect its physical properties.

The best-fit analysis moved the part into an orientation that was close to solving the problem – only one area remained where there was insufficient stock to meet the design intent.

The engineer manually adjusted the orientation of the part on the screen until he had eliminated the small, undersize area. Next, he used the model to generate reference dimensions that were used to fixture the part. He provided a copy of the as-built model to CNC programmers, who checked their existing CNC program used to drive the machine tool that cuts the part to its final dimensions.

They determined that the existing program would work without any modifications. The CNC programmers also designed a new fixture to hold the distorted casting in the proper orientation. When the casting was run on a machining center with the new fixture and existing program, the result was a part that met all dimensional and structural requirements.

"In the past, we have frequently used laser scanning with great success to inspect or to re-create complex geometries," the engineer concludes. "But this application demonstrates a new potential application for this technology – repairing parts that have become damaged during the production process. The high accuracy of the NVision laser scanner made it possible to deliver a part that met all dimensional requirements. At the same time, the ability of laser scanning to quickly capture a large volume of data helped us meet the customer's delivery requirements and keep costs under control. Our expectation is that we will continue to find more uses for this powerful technology."