Adaptive machining shows its mettle in manufacture and repair5 April 2002
The dimensional variability and complex shapes found in gas turbine components pose challenges for conventional automated machining. One solution is adaptive machining, where a toolpath unique to each component is generated using digitised data. Paul Walton, TTL, Aylesbury, UK
Surprisingly, the manufacture and repair of gas turbine blades is still largely being carried out in the traditional way, by people wearing leather gloves and hand-held grinding, finishing and polishing devices.
The problem lies with the fact that individual blades and vanes that make up full engine-sets are not geometrically identical and therefore are not completely interchangeable. While the more crucial dimensions are tied down to very close manufacturing tolerances and are tightly controlled, others are more open-tolerance and are hence less well suited to automated machining operations.
This is where adaptive machining comes into play, bringing with it levels of accuracy and repeatability once deemed impossible. But what exactly is meant by adaptive machining?
Engraving an egg
An interesting and unusual demonstration that helps understand the technology is one developed recently by UK-based company TTL, using raw eggs. Like the multitude of blades, vanes and combustion components that make up a modern gas turbine engine, no two chicken eggs are identical. To demonstrate the principles of adaptive machining, a name or company logo is engraved into the shell of an egg using a standard 5-axis machine tool. The process is achieved by first digitising the surface of the egg, then transferring the captured information to a computer system which processes the data and is subsequently able to understand mathematically, the exact shape and size of the egg. A toolpath is generated that is based upon the digitised data and is hence absolutely unique to each individual egg. The resulting numerically controlled code is used to control the machine and allows cutting evenly at a pre-defined depth of 0.05mm (0.002 in) into the eggshell regardless of the surface variations.
The problems encountered when engraving an egg are technically equivalent to machining many features of new cast or forged components, or removing the coatings or excess weld from repaired ones.
Of course, there is nothing new in digitising. Nor is there anything new in CNC machining. The trick in developing successful adaptive machining applications is to deliver not only technical excellence, but economic viability also. This requires the putting together into one neat, efficient solution, of a number of different technologies, the end result being a cost-effective, productive and efficient alternative to more traditional methods of manufacture. Certainly, with ever-increasing competition and reduced turnaround times, technology for technology's sake will never win the day. Decisions to invest in adaptive machining technology are usually driven by a combination of technical, quality, economic and health and safety requirements.
But why do gas turbine components show this geometric variability? The answer to that lies in how they are manufactured. Turbine blades, nozzle guide vanes and compressors are either cast or forged, both being relatively imprecise processes. Castings, for example, can be prone to shrinkage variations resulting from material and temperature fluctuations, and die wear can influence the resulting shape of both cast and forged parts. Subsequent heat treatment can play a part also, where the component is relieved of its internal stresses and modified metallurgically.
The problems of variable geometry in new parts are further compounded in repair applications, because here, in addition to the original manufacturing tolerances, the repairer must deal with the issue of wear and distortion that have occurred during operation. Considering the very high ambient temperatures and corrosive conditions under which some components operate, it is little wonder that the resulting variations in shape and size can be quite considerable.
Figure 3 (not available) shows a "new part" - a titanium forged compressor rotor - which has a snubber or mid-span support on the aerofoil. For the reasons already stated, the blending of the snubber onto the aerofoil is an operation that is traditionally performed by hand, and which has tended to be labour intensive, expensive and time consuming. Moreover, original equipment manufacturers (OEMs) have claimed that as much as 30 per cent non-conformance has been experienced as a result of the manually-based process giving rise to significant wasted time end effort, along with unnecessarily large inventory costs. Systems using adaptive machining have revolutionised quality throughput and productivity. One such system is claimed by the customer to have reached payback within eight months of installation. In this particular example, the "snubber" was CNC machined to a precise position on the component, and "adaptively" blended in the same operation onto the aerofoil surface. Depending upon the blade size, up to 800 per cent savings in cycle time have been achieved compared with the previous manual methods.
Repair, refurb and overhaul
The maintenance, repair and overhaul sector of the gas turbine market, in particular, makes extensive use of hand labour. Many of the weld removal and blending operations in this sector are still carried out predominantly by hand, but with the intrinsic value of components being so high - particularly in the industrial sector - there are very valid reasons to utilise technology that not only increases throughput, but also greatly increases accuracy, quality and reliability.
Figure 4 shows a typical example of adaptive machining applied in the gas turbine maintenance, repair and overhaul field.
In this example, adaptive technology is being used to accurately machine the weld-bead at the tip of an unshrouded turbine blade. Such software-based machining technology has been shown to cut cycle times by a factor of three whilst greatly enhancing the quality of the finished product. Furthermore, the lack of dependence upon operator input makes the process significantly more consistent.
A further application of adaptive machining is in the refurbishment of blades of all kinds. Figure 5 shows a small compressor rotor set-up for single-hit automated weld removal to both the tip and the leading edge. Special fixtures are invariably required in order to gain access to the relevant areas and the application software is either purpose-written or is at least customised to the particular job on an "as required" basis. In this application the business case changes slightly, since the intrinsic value of the components is relatively low (as compared with an industrial gas turbine "bucket", for example), but the volumes repaired are very much higher and can with some repairers reach the hundreds of thousands per year. Here once again, cycle times are reduced by up to 100 per cent and a typical adaptive machining cycle can digitise, compute and precisely restore the profile of a tip-welded compressor in well under two minutes. Accuracies can be to within +0.00mm/+0.01mm of the parent material, which represents a huge reduction in subsequent hand-polishing work.
The nozzle guide vane presents its own problems. The geometry is often very difficult to deal with - being effectively a single, double or triple aerofoil constrained at both ends by platforms. Steep walls, tight spaces and awkward fixtures frequently hinder access for the machine tool and unlike most compressor and turbine repairs, the nozzle guide vane suffers seemingly random weld and braze patches. These factors make the automated repair of nozzle guide vanes very difficult, but solutions have been developed.
For example, using vision-based weld identification, it is possible, with sophisticated software, to vastly simplify the process of differentiation between welded and non-welded areas. In this way, the original parent material can be digitised and used for reverse engineering, whilst the welded areas are used to form the boundaries inside which the machining process takes place.
Choosing the right equipment
In the world of 3-axis CNC machines, within any given price-range, one machine tends to be quite similar to another. Each will have a spindle, a toolchanger, a CNC control, etc, and each will be capable of moving all three axes simultaneously, and therefore be able to describe complex 3-dimensional shapes. Of course, there will be differences in quality and performance, but fundamentally they all do roughly the same job.
This is not the case with five-axis machines, however. Here, the physical performance of the machine is only one factor. At least as important are the configuration of the machine and the suitability of that configuration to the application.
So the selection process for many 5-axis applications needs a much closer look as not only does the machine need to meet the performance criteria, but it must also be of a suitable configuration.
Figure 7 (not available) shows a typical adaptive machining installation. The architecture of the manufacturing cell will vary with the application. But, wherever possible, standard machine tools are used. This minimises costs and lead-times, and greatly simplifies longer-term maintenance issues. The installation shown in Figure 7 (not available) is a Hermle 5-axis machining centre.
Labour rates, work volumes and the nature of the intended application itself all play their part in determining whether adaptive machining is appropriate in a particular application. Recent experience has shown it to be a valuable approach in the complex business of gas turbine component manufacture and repair. Its significance has now ben recognised with the award to TTL of a Queen's Award for Enterprise.