Using Measurement and Software to Optimize Part Setup for Machining

Aircraft manufacturing is international, with a complex supply chain dividing work among numerous companies around the world. Globalized manufacturing and outsourcing brings challenges and opportunities for suppliers.

Technical challenges include improving the quality of complex components with multistage manufacturing operations. Components must be highly accurate due to considerations of weight and assembly efficiency.

Competitive pressures require companies to be flexible with a fast response to customer needs, which often means increasing production but at reduced costs. Costs of material and machining—and the time required to remove excess material—are powerful incentives to minimize stock allowances for machining. Lead times are an important consideration in the overall costs.

This combination of challenges presents lucrative opportunities for suppliers that can meet the conflicting stringent requirements quickly and consistently.

Machining operations and part setting

Machining operations are carried out in a workpiece coordinate system. This is a temporary coordinate system created on the machine controller to create the finished part. The part-setting task involves setting the position of the workpiece—whether raw material or a part-finished component—to match the desired location of the finished part.

The complexity of this task varies considerably. Inaccurate part setting will result in scrap components, and slow part setting will have significant effects on the utilization of the machine.

Machining from a solid block of raw material is straightforward, especially if material allowances are generous. The material can be positioned on the machine, aligned with the axes, and the work coordinate system can be determined by “dead reckoning,” often by measurement from a corner. There can be a trade-off between material allowance and the accuracy of the positioning—a greater material allowance can compensate for less accurate positioning or alignment with the machine axes.

Part-finished components are more challenging, especially for thin-walled components or where stock allowances are minimal. The alignment must be maintained to tighter tolerances, reducing the margin for error. Traditionally this is achieved by physically positioning the workpiece to the required accuracy.

Precision fixtures can be used to hold the component in the desired position with a high degree of accuracy. This is the best method for high-volume, low-mix manufacturing. The fixture—and the development process for the fixture—are significant investments, but the investment pays off when producing a large number of components.

The case for fixtures becomes more difficult to justify for lower production volumes, or for a wider variety of parts. Designing and creating fixtures becomes a larger proportion of the overall manufacturing costs, and the maintenance and storage of the fixtures can become expensive. Short lead times for new parts may simply not allow enough time to produce suitable fixturing.

Fixtures become less effective if there is any variation in the part-finished components. High-quality manufacturing processes should have relatively consistent parts, but certain processes—such as casting, forging, or heat treatment—will result in some variation or distortion. In the extreme case, such as for one-off components, the traditional method for part setting involves a minimum of fixturing. The operator physically moves the pre-machined component until it is accurately located. This process requires a high degree of skill and is time-consuming and unpredictable, especially for larger, heavier parts. Accurate, physical location depends on having identifiable points or features that can be measured and adjusted until they are positioned correctly. Basic tools such as dial indicators can be used to align straight edges with machine axis directions, but it is difficult to align complex curved components without specific datum features.

Machine-tool probing

Machine-tool probing systems provide a well-established method for part setting. Probing cycles are combined with work coordinate system transformations to establish an accurate position for the part “as measured.”

Most machine-tool probing systems are supplied with routines for measuring simple features such as planes, edges, and bores. Typically these provide size and location information, and this is enough for simple part setting.

The operator sets an approximate work offset—one that is accurate enough to allow a measuring routine to measure the desired features while avoiding unwanted collisions—and runs a probing routine. The probing routine measures the positions of the parts in the current work offset, calculates the transformation between the measured part and the work offset, and uses this to match the finishing programs to the current position. Rather than relying on a physically accurate part setup, this approach adjusts the work offset to match the measured position.

Simple probing macros are not sufficient for locating more complex components, but the probing and alignment techniques from direct computer controlled coordinate measuring machine (DCC CMM) inspection software can be used to run similar measuring sequences directly on machine tools.

The alignment techniques common in inspection software can allow more sophisticated optimization. In particular, “best fitting” operations, which optimize the positioning of measurement results against CAD surfaces, can be used to optimize the position of complex free-form surfaces, either by minimizing the differences from nominal (least squares best fit) or using tolerances to take the amount of stock into consideration.

When more sophisticated software-based solutions are used, two different approaches can be used to update the machining programs to match the measured position:
Modifying the work offset, as with basic probing-based part setup. This is preferred for batch manufacturing because it allows existing, proven machining programs to be reused. A standard machining program is used every time, but the work offset is adjusted slightly to match the measurement.
Modifying the tool paths. This involves transferring the optimized alignment from the measuring software to the machining software. The work offset is not changed, but the tool paths are recalculated to match the required datum position. This is more versatile if significant alignment changes are required but requires verification of the new machining programs. This is the basis of “adaptive machining” techniques discussed later.

Case study: Eurofighter Typhoon’s foreplane

The Eurofighter Typhoon is a combat aircraft, and its wing platform is a delta wing (deltaplane) that uses a stabilator-type canard foreplane. BAE Systems manufactures the foreplane, a component of approximately 2.2 m × 1.1 m. The foreplane is diffusion-bonded from several titanium sheets and then super-plastic formed. The nature of the super-plastic forming process means that some distortion is inevitable, and deviations of up to 4 mm are expected.

After forming, the excess material from the super-plastic forming process must be milled so that the machined surface blends closely with the adjoining airfoil surfaces. The shallow, tapered shape adds to the difficulty of the blending process.

Original process
The original process involved locating the foreplane manually on a fixture to machine one side of the part. The operators then turned the component over and relocated it in a different fixture. This was repeated two more times, each time with a different fixture. The quality of the part depended entirely on the ability of the operators to adjust the foreplane and machining supports according to the fixture datums.

This four-stage setup process compensated successfully for the variability of the super-plastic forming process but was unsatisfactory in several respects. The quality was inconsistent due to the variability of the manual setting operations, so manual reworking was an integral part of the process. The process also took many hours, and the time required varied greatly, making scheduling difficult.

Revised process
The original fixtures have been replaced with a single vacuum fixture that holds the part centrally with significant overhang. This eliminates the need to turn the part over because the key areas can be machined in a single operation. The manual setup operations have been replaced by an automated, measurement-based part alignment process:
• The part is roughly positioned (to an accuracy of 4 mm) and secured on the vacuum fixture.
• The operator runs the software and enters part details, which are logged to ensure traceability.
• A five-axis probing program measures a series of key points on the part surface.
• The software uses the measurement results to calculate an optimized workpiece alignment and checks that the amount of movement required is within predefined tolerance limits.
• The software applies the optimized alignment to the machine by uploading a numerical controlled program that adjusts the work offset.
• The finish machining part programs run in the optimized alignment to create a perfect component.

Benefits
More than 700 production foreplanes have been produced, and the benefits have emerged as follows:
• Improvements in quality and consistency have almost eliminated the need for manual rework. (Previously almost every unit needed manual rework.)
• Manufacturing time has been cut by more than 60 percent. (The manufacturing process originally took 20–30 hours.)
• Manual setup operations have been dramatically reduced, and the part is now manufactured with a single fixture with one set of measuring operations.
• Setup times have become consistent, reducing variability and allowing better process planning. Combined with the reduction in machining time, this has allowed BAE Systems to bring subcontract work back in-house.
• Redundant fixtures are no longer required, saving handling, storage, and calibration. (Had the alignment process been used from the beginning, the fixtures would not have been required, saving the cost of their design and manufacture.)
• Operator variability has been reduced further with in-process measurement checks.
• The project has been rewarded internally with a BAE Systems’ Chairman’s Award.

Adaptive machining

Adaptive machining is a term used to describe more sophisticated measurement and machining combinations where the measurement is used to modify the shape of the finished component as well as the position. This requires closer integration between the measurement, design, and manufacturing processes.

The machining of composites is a typical adaptive machining application.

Many composite parts include drilled holes that are used for assembly purposes. Often the top of the rivet or bolt is kept flush with the surface by using a counter bore. Composite materials can vary in thickness for a variety of reasons, especially with open-mold processes, and this can cause significant problems with counter bores. If the counter bore is too deep, there may not be enough material for a durable joint. If the counter bore is not deep enough, the appearance and the aerodynamics will be compromised.

Adaptive machining can be used to overcome surface variation by measuring the part (either in part or as a whole) and mapping the actual surface. These measurement results are used to reshape or adapt the nominal shape of the part accordingly.

Composite materials may also be subtly distorted in other ways. The curing process will result in some variation, and the parts tend to be comparatively flexible. The parts also tend to “relax” as the fibers are cut during multistage machining operations. Again, adaptive machining can be used to subtly reshape the nominal model to reflect the actual shape of the part.

The remanufacturing of turbine blades is another common adaptive machining application. The high cost of components means that repair is a fas-growing industry. High-pressure turbine blades are typically repaired twice before they are replaced. The most common repair is to restore worn leading edges and tip shrouds. These are comparatively low-stress areas and can be repaired with weld overlay and remachining. The complication is that the high temperatures in service often cause subtle distortions to the blade, so the shape may have to be remodeled slightly.

In this case adaptive machining involves measuring the blade, “morphing” the CAD model to match the new, slightly distorted shape, adding a weld overlay, and machining back to the morphed shape.

Accuracy

It is important to manage the accuracy of machine-tool-based measurement, because machine tool probing is fundamentally different from measurement with a CMM with a charge-coupled device camera in a number of respects.

CMMs are maintained in a temperature-controlled environment, whereas machine tools actually generate a lot of heat during machining operations. This will change the accuracy, particularly with respect to characteristics such as squareness. It is important to take, monitor, and manage the measuring accuracy of the machine tool, especially when measurements are made during a manufacturing process.

Probing

In many ways a probe is an accurate mechanical switch: The probing system is triggered when the stylus makes contact with the component. There is a small amount of pre-travel before triggering, but this is very repeatable, at least for a fixed direction. CMM probing systems are not all equal, but they are intended for accurate 3-D measurement—the pre-travel response is quite even for all probing directions. Machine-tool probes fall into two categories: 3-D-measurement probes and setting probes. Three-dimensional probes have a consistent response in all probing directions, but setting probes do not. Simple part-setting operations involve probing along the X, Y, or Z machine axis. The response is different for each direction, but each direction is very repeatable. A simple calibration procedure for each individual axis direction allows sufficient accuracy for most part-setting operations as long as these are confined to the principal axis directions. But this is not suitable for complex 3-D probing operations, particularly with five-axis rotations.

It is worth bearing in mind that very tight accuracies may not always be necessary for part-setting operations, but it is important to manage the accuracy. This involves:
• Creating an initial benchmark of the overall measurement accuracy
• Regularly monitoring the performance to ensure that it remains within acceptable limits

Two approaches are in common use. Hardware-focused solutions involve measuring a known artifact and comparing the results with a benchmark established on a CMM. Software-focused solutions involve measuring simple known geometric shapes (typically a calibration sphere) in a variety of locations (e.g., using the rotation of a rotary table) and analyzing the results in a comparison against nominal data.

Portable measurement

Part-setting and software-fixturing processes work by modifying the work offset, and so require an accurate understanding of the initial work offset. This is one of the reasons why machine tool probing is so attractive for these operations (another being the automated measurement process).

Machine-tool probing is not suitable for all applications, however. If the component is larger than the machine, or if there are concerns over accuracy, it may not be possible to measure the part using the machine tool. Both of these problems are common with composite components and with robot machining operations, for example.

For these applications it is possible to use portable measurement to optimize the alignment. The part is measured and aligned as normal, but an extra step is required. The measuring machine must be “aligned” with the initial work offset of the machine tool. The measuring machine must measure some known datum features in the machine tool coordinates, and there is a wide variety of methods for doing this, including:
• Measuring physical datum features (for example, a fixture or calibration spheres in known coordinates)
• Comparing measurement results between the measuring machine and machine tool (if the machine tool is fitted with a probing system)
• Using the measuring machine to measure the position of the machine tool in various positions (for example, using a laser tracker to measure a reflector held by a robot)

The work offset of the machine tool is effectively measured by the measuring machine and becomes another measured alignment in the measurement software. The software calculates the transformation of the optimized part alignment relative to the current work offset, and creates a short numeric-controlled program to update the work offset to the desired, optimized location.

About The Author