In-Process Measurement Makes Sense
When doors and other aircraft assemblies arrive at the fuselage, assembly mechanics at most aerospace companies are not surprised when they do not fit properly. Doors that are not quite right will often pass inspection because the company’s quality control checks are inadequate and subject to human error. So, the mechanics just adjust the hinges and door stops, sanding and pounding the skin around the opening.
They can spend up to four days fitting each door to a commercial jetliner.
Work-around techniques such as these are still commonplace throughout all manufacturing industries, even though most companies actively embrace and implement quality control procedures. Rather than rejecting incorrectly fitting parts and subassemblies and re-engineering the process to make them right, manufacturers continue to use loose go-with-work/no-go tests and work-around procedures to cover out-of-control processes.
Manufacturing engineers and technicians know that building in-line rework into the assembly process only lengthens overall cycle time and drives up costs. The problem is that many of them do not have measurement and analysis tools for troubleshooting common fit problems quickly and economically. The lack of immediate feedback on where the mating parts interfere with one another, and by how much, frustrates their attempts to pinpoint the problem and communicate exactly what went wrong to both external vendors and internal operations.
Unfortunately, the tools for gathering, tabulating and interpreting such hard data have been time consuming and difficult to use for troubleshooting fit problems. Moreover, the data- collection tools available to manufacturing engineers deliver reams of data days to months after the fact, requiring them to sort through the numbers to find the relevant trends and clues. As a consequence, the consensus between manufacturing and management is that there is more profit in working around problems to ship product now rather than hold up production for quality control.
One such set of tools is portable coordinate measurement machines (CMMs) used in tandem with 3-D measurement and analysis software. When parts are not mating correctly, a manufacturing technician can bring the portable CMM to the part, rather than the other way around. These tools are as mobile as rulers, calipers, micrometers and other manual gages, but differ in that they do not rely on the operator’s skill to get accurate, repeatable measurements and that they have the ability to analyze the results.
Portable CMMs create a 3-D blueprint of objects and assemblies. Some are laser-tracking devices, where the operator guides a mirrored target over the object’s entire surface. The tracker emits a laser beam that is reflected back from the target-up to 230 feet away-allowing the system’s laptop computer to simultaneously draw and record all of its 3-D measurements.
Articulated-arm portable CMMs allow the operator to guide the arm’s tip over the surface of the object to achieve the same result, but are more accurate and are ideal for within a 4 to 12-foot working volume.
Engineers and designers use the data to reverse-engineer, improve or create new parts and products-or improve how they are made. Quality control personnel use it to verify that their parts are made correctly, and, if not, where they need to be corrected.
Computer-Assisted TroubleshootingWhen one automaker gave its assembly line such a tool, its manufacturing engineers and technicians eventually eradicated the longstanding go-with-work routine for fitting van doors. Using a portable measurement device and analysis software, they were able to measure incorrectly mating pieces and compare the actual parts to nominal data stored in the company’s CAD files right on the shop floor. Because technicians could pinpoint and quantify errors accurately, manufacturing engineers could then correct upstream processes so that subsequent doors would fit without adjustments.
The initial measurements identified a specific upstream door assembly operation as the problem. Further analysis revealed that assembly workers would introduce forces on the door that sometimes caused enough deformation to throw the assembly out of tolerance, even if the components were made to nominal dimensions.
To find out where the process exerted the unwanted force, technicians just broke the process down into steps and checked the relevant portions of the assembly for deviation from nominal. They started by measuring the assembly tool and the first part and then each part of subassembly added afterward.
There was no need to stop production of the door and send various parts and subassemblies to a temperature- controlled room for checking on a CMM. Although measuring during assembly slowed production of the one door, perhaps taking twice as long to build, it was finished afterward. Sending it to a CMM would have taken longer, and the door would still require assembly when it returned to the line.
The unwillingness to stop production to measure sub-assemblies also was a significant factor in preventing an aerospace giant from analyzing why the skin of a popular large jetliner bubbled when manufacturing joined two sections of the forward fuselage.
A fixed CMM was the only tool the company had for quantifying the problem at the time, but no one wanted to go through the trouble of measuring the piece on it because just getting it to a temperature-controlled room could take up to four weeks. Then the measurement process could take a month or two, and engineering’s interpretation of the reams of numbers would take another several days. So manufacturing engineering continued with piecemeal solutions.
Over time, the consensus was that the problem was due to a faulty skin-stretching process. Before building a new stretching tool, however, the obstacle to measuring the sections before and after joining vanished when a portable CMM became available. By quantifying how the skin and frames were not integrating, the company’s manufacturing engineers found that the poor fit was due to the frames, not the skin, and traced the cause to the tooling that made the set of frames. Had they acted on their original guess, the new skin stretching would have made the problem worse.
Making No-Go Situations GoThis capacity to solve problems quickly and on the spot can save manufacturers millions of dollars, not only by eliminating work-around procedures, such as the rework for fitting airplane doors, but also by eliminating situations in which in-line finessing cannot make a part fit. These no-go situations require either stopping production until the problem is resolved or scrapping the part altogether. In these cases, engineers, technicians and sometimes vendors argue about what is right and wrong until someone can eventually produce hard data describing the fit and function of the part.
A no-go situation was causing exactly these problems when an automaker attached locking fuel doors to car bodies on one of its assembly lines. For two years, engineers there chased a fit problem that caused a 17% door- rejection rate. Some of the fuel doors would not lock because somehow they were shifting slightly during assembly. The hole in the latch would not mate with the door-release pin.
As manufacturing engineers and technicians tackle such no-go situations, their enthusiasm for the troubleshooting technique typically builds. Seeing dimensionally correct parts going together faster, easier and better only creates an incentive for troubleshooting and correcting the go-with-work situations.
The process of fixing dimensional problems will eliminate built-in production line delays, decrease overhead and increase sales because of better quality and lower prices. Q
Quality OnlineFor more information on in-process measurement, visit www.qualitymag.com to read these articles:
- “In-line Probing for Process Improvement,” By David Bozich
- “CMM Programming Saves Time and Money,” By Larry Adams
- Case Study: “Life, Liberty and the Pursuit of Quality”