It’s possible to minimize unexpected tool failure by adhering to best practices, error proofing, and other automation strategies
When a tool breaks during a machining operation, the part being processed is often destroyed, and sometimes the machine is damaged. Aerospace parts are often complex shapes, manufactured from exotic materials that require prolonged machining cycle times. Therefore, a scrapped part is a significant loss in raw materials and value-added machining.
Because single-piece lot sizes are not uncommon in the aerospace industry, the loss of a single part is a real hit to production yield. An aircraft part failure can have catastrophic results, consequently compliance controls and risk mitigation makes reworking damaged parts more complicated than in other industries. The loss of a part or a machine due to tool breakage can have a significant impact on profitability and customer satisfaction.
Many of the specialized machine tools used in the industry perform mission-critical machining. Due to the cost and the long lead times for these machines, they are most likely bottleneck assets, and a crash can have a significant impact on production capacity.
There are many reasons for tool breakage during machining, and there is not one solution that can ensure 100% detection or protection. Purpose-built tool-breakage recovery cycles can also save parts and lost production. Given the cost of the machines, material, and value-added work-in-progress that are characteristic of the aerospace industry, several levels of preventive and detection strategies are justified to protect the company’s investment.
The value of the parts and types of materials machined in the aerospace industry demands that the highest quality tooling be used for most applications. But even the best tools will fail if the processing parameters in the part program are wrong for the particular tooling or application, or if the operator makes a mistake during setup or adjustment.
Aerospace parts are machined from forgings, castings, bars, and sheet stock, and from materials with generally poor machinability. Variations in material composition, surface conditions, and depths and widths of cut make it very difficult to specify optimum cutting parameters throughout the part program, and for every part produced.
Engine parts are manufactured from heat-resistant superalloys (HRSA), such as Inconel, Hastelloy, and Waspaloy. Titanium is also used for many aircraft parts. The machinability of these alloys is generally poor because of the very nature of the material structure that is required for the application. Cast or forged components typically have a rough, uneven surface.
High cutting forces and high temperatures are generated when these tough materials are processed. Carbide in the structures of HRSA materials is abrasive, and a tendency for work-surface hardening can cause tool notching. Other tool failure modes such as cratering, thermal cracking, chipping, edge buildup, and deformation can occur, as well as a machine crash, if the feeds, speeds, and depths of cut are not specified correctly for the application.
So the very nature of aerospace part machining is likely to cause uneven tool wear and high stresses, which are prescriptions for premature tool failure. These problems can be avoided, however, by optimizing process parameters.Even if the process parameters are perfect, tool-setup tasks and tool-wear offset adjustments are error-prone. Measurement, calculation, and data entry mistakes are common causes for tool breakage and machine damage.
Given the high machinery and work-in-process values typical in aerospace manufacturing, it makes sense to implement several levels of safeguards to protect these investments. Some potential solutions are well documented, such as sonic or vibration monitoring, and the use of inspection and tool-setting probes to error-proof the tool setup and adjustment processes. Data collection and Failure Mode and Effects Analysis (FMEA) techniques can provide valuable insight into the root causes of tooling failures and related machine crashes. This analysis can help select the most effective strategies for a particular business operation.
Abrasion wear is the ideal failure mode for a cutting tool, because it tends to be repeatable and predictable. When wear is even, indicators such as burrs, changes in surface finish, or the sound of a cut can alert the operator to take action well before catastrophic failure. Tool-management systems can be used to remind the operator to perform timely preventive-maintenance tasks.
However, many parts are produced with less-than-optimum cutting parameters. Perhaps the company’s CAM system incorporates a generic postprocessor, and the optimized data from the shop floor are not fed back to the system reliably. Perhaps the latest new wonder tool shows impressive results on one application so it is adopted for the shop, but the new process parameters are not edited in every part program.
Tooling companies typically offer free consulting services for the application of their products. They can suggest a variety of scenarios based on the particular tooling technology; for example, faster feeds and speeds with lighter depths of cut, or slower feeds and speeds with more aggressive depths of cut. The results are often impressive, reducing part-cycle times by a significant margin. But more importantly, the tools will be operating within their recommended cutting envelope.
Once all the cutting parameters are confirmed, an electronic change-management system can ensure that process is controlled; managing both the part programs and the tool-setup sheets.
Updating tool geometry and wear offsets are error-prone procedures that can cause a crash, damaging the part and perhaps the machine.
A typical offset update begins with the operator measuring a tool or a part. Then the operator must interpret the impact of the measurement, and calculate the offset value that must be typed into the CNC. Often signs get mixed up, and radii-versus-diameter problems can also complicate the calculation. Entering the value into the CNC offers additional opportunities for mistakes. Decimal point and sign errors are common, and sometimes the value is entered into the wrong offset location.
Inspection and tool setting probes are a traditional solution to error-proof tool geometry and wear offsets, and they are used extensively in the aerospace industry.
But most problems can also be error-proofed extensively by using the high-level macro programming available with most CNCs. The operator is prompted to make a measurement and enter the measured value into an easy-to-remember offset location in the CNC. This parameter can be an offset number that matches the tool pocket location, or it could be a constant offset number, such as 99 or 999.
The part-program macro language can perform a series of checks to error-proof the data entered. It can check that the expected offset location changed, confirming that the operator entered the measurement value into the right location. It can make sure the measurement value entered is within a reasonable range. Then it can calculate the correct offset value, considering the value entered, the dimension specification and tolerance, and the existing offset value. It can also put limits on the maximum offset that is reasonable considering the tool wear range specification. If all is well, the correct tool offset is updated automatically, eliminating all manual calculation and data entry errors.
When contouring complex shapes, there are many times when tools will make a partial cut. If tools are lightly loaded, it is difficult to maintain the proper chip formation, and rubbing may occur. Without the proper chip formation, heat is not evacuated from the tool and material, and tools may experience buildup or cratering tool-failure modes.
Though adaptive control is typically considered a productivity solution, there is a beneficial side effect to keeping the tool under constant load. If the adaptive system detects the load on the tool falling, it will speed up the feed rate to maintain the target load. If the load rises above the target load, the speed will be reduced accordingly. Keeping the tool under constant load improves chip formation and evacuation, and practically eliminates build-up, cratering, and chip re-cutting issues.
If machining process parameters are optimized for an application, tools will demonstrate the reliable and predictable abrasive-wear failure mode. In this case, a tool-management system can alert the operator to proactively replace tools before they cause a serious problem.
The CNC can monitor tool usage both in machining time and machining cycles. Limits can be established to warn the operator to replace inserts, and then stop the machine if the warnings are not acted upon. To extend the time between maintenance interventions, sister-tooling strategies can be implemented for high-usage tools, if extra pockets are available. If a machine has the ability to safely replace worn tooling while machining continues, then significant efficiency benefits can also be realized.
If tools are centrally managed in a toolcrib or by an outside vendor, it can be difficult to track usage parameters over the life of the tool. RFID tags can be used to load preset or measured tool geometry settings into the CNC, and to update the tool life tables. When the tool becomes inactive, the tool life and any updated wear offsets can be written back to the RFID chip embedded in the toolholder for the next machine to read. If the operator decides to override the CNC tool management system to command replacement before the expected tool life is reached, that information can also be flagged on the RFID chip.
Off-line systems can also monitor the RFID data stored with the tool. This information can be used for preventive maintenance in the toolcrib or to analyze any unusual failure patterns, and determine if root causes were related to a particular part, operator, shift, machine, or other factor.
When a tool breaks during machining, it may take several minutes before the operator can stop the process. If the break occurs when roughing, theoretically all that must be done to recover is to retract the tool, replace the insert and then resume the process from the last point that was machined cleanly. In reality, tool breakage recovery is a little more complicated.
One option is to return to a convenient restart block in the part program. This step might require several minutes or hours of machining air, a hit to productivity. Because it would be less worn than the tool it replaced, the new tool would likely re-cut the part. The lightly loaded tool could cause bad chip formation, which might lead to another tool failure.
A more-elegant solution uses purpose-built CNC features to manage the recovery. Activating the retract feature moves the tool rapidly away from the surface of the part. Machine axes can then be jogged to an inspection point for insert replacement. The tool-offset move-and-store feature can be used to update the active offset to clear any artifacts that were not machined correctly between the point in time that the tool broke and the time when the operator retracted the tool.
The tool retract-recover cycle can then be triggered to move the tool back to the retract point. Next, the CNC-retrace function can be used to reverse the part program to a point on the part with no artifacts. Finally, the tool-offset move-and-store feature is again used, preferably with a handwheel, to move the tool until it touches the surface of the material. The tool is now exactly at the right place, with a new and appropriate tool-offset value activated.
Broken tools may be especially catastrophic in certain conditions. (For example, if a machine attempts to tap a hole after the essential drilling operation failed.) A serious machine crash can occur with a twin-spindle lathe if the cut-off tool breaks prior to a spindle-to-spindle transfer. In these cases, it may be essential to inspect the tool after each operation to make sure it’s still intact.
Contact and noncontact tool measurement devices have been available for many years, and have been used for broken-tool detection. Noncontact devices offer faster tool-breakage detection, because the lack of mechanical contact means that the tool can pass through the laser beam at high speed. However, these systems have been difficult to mount on moving tables, and because the table on a HMC is typically a removable pallet, applications on such machines are even more complex.
A new tool-breakage sensor—the TRS1 from Renishaw (Hoffman Estates, IL)—offers many advantages for high-speed tool-breakage detection. This single-sided device can be mounted to any stable vertical surface. It’s ideally mounted so that it can monitor the tool in the vicinity of the tool change position, eliminating the need for moving the tool down to a table-mounted sensor. A sampling algorithm even differentiates between the tool and coolant drips for fast, reliable detection. It can even be used to check that parts have been separated by a cut-off operation in the case of the spindle transfer for a twin-spindle lathe.
No matter how much you error-proof a process, something unexpected will happen, and a tool will be commanded to move at high speed towards the part or a machine component. For expensive parts or mission critical machines, it makes sense to monitor the torque being generated by the servomotors for unusually high values.
Machine collisions and damaged cutting tools generate abnormally large load torques on the servo and spindle motors, as compared with normal rapid traverse or cutting feed. A maximum torque limit can be set in the part program so that excessive loads can be detected by the CNC in real-time. If the CNC detects an unexpectedly large torque-disturbance value, the CNC can stop the servomotor very quickly, or even reverse the motor by a specified value, minimizing possible damage to the part and machine.
For more comprehensive protection, fine torque-sensing learns a characteristic torque profile for the specific part being machined, and saves it in memory. During subsequent runs of the same part, fine torque-sensing detects any occasion when the monitored axis exceeds the learned profile by a specified tolerance.
Investment in work-in-progress and machinery is very high in the aerospace industry. Minimizing unexpected tool failure with best practices, error proofing and other automation strategies can be part of a comprehensive multilevel solution.
This article was first published in the March 2006 edition of Manufacturing Engineering magazine.