Gilbert, David Jonathan
(2017)
Laser process optimisation for in-situ repair of aero-engine components.
EngD thesis, University of Nottingham.
Abstract
Repairs on stress raising features (SRFs) generated on internal aero engine components such as compressor blades are sometimes necessary, dependent on size, to avoid further damage, such as crack propagation. Where possible, repairs are carried out whilst the engine remains on the wing (in-situ) to avoid costs incurred by engine removal/transportation, and fines from non-operation of the engine. A system comprising an SPI pulsed laser, a Surgical Innovations deployment probe, and OpTek beam guidance components, proposes to adapt the current method of SRF blending which uses a rigid grinding tool, inserted into the engine through the external inspection ports. The need for this update to the system arises from components in the new Trent XWB engine which are not compatible with the previous rigid grinding tool. Delivery is planned through a deployment probe with a set of laser beam guidance optics at the head.
Material removal was achieved using a nanosecond (1≤ τ ≤ 999 x 10-9s) pulsed Yb:YAG infra-red (1064nm) laser, which was used to ablate material in multiple small volumes (craters). Metallurgical studies were conducted to quantify the effect of the three key process variables (laser power, pulse frequency, scan speed) on secondary effects such as material recast layer thickness and oxidation. Through these experiments some optimal machining parameters were defined, which could produce repair geometries with a good compromise between minimising machining time and reducing component metallurgical damage. These parameters were defined as 13W of measured power output (input and output powers may vary dependent upon a number of atmospheric conditions such as ambient temperature and air pressure), 25-35kHz pulse frequency (dependant on laser system attenuation frequencies), and 400-500mm/s scan speed. Using these parameters repairs were generated with full ablation through the blade edge in 15 minutes.
A predictive model was made to simulate surface topography evolution throughout a laser ablation process with any given parameters, using some initial laser-material calibration experimentation. This model showed good predictive powers, where initial validation of the model showed errors under 9.39% for ablation on CVD diamond and under 1.3% for ablation in INCONEL® 718 alloy. The model had limited predictive capability for simulating material redeposition and as such further validation was carried out for more complex geometries. Micro-scale validation was carried out on diamond for complex features generated using varied laser scan track overlaps. Macro-scale validation was carried out on INCONEL® 718 for large features where a desired mass was to be removed through prediction of volume of
machined material using the model. Both further validation experiments had errors as low as -0.14% and 1.38% for diamond and INCONEL® 718 targets respectively.
High cycle fatigue testing was carried out on engine run compressor components from an XWB test engine. These components were machined using best-practice boreblending techniques employed in the field today, as well as laser machining developed throughout this project. Optimised machining parameters as discussed previously were used to generate laser repairs with minimal material damage. These laser repairs were performed using both laboratory laser equipment, and prototypes for tooling which is intended for use on in-situ repairs in the field. All of these machined samples were put under high cycle fatigue testing to simulate component fatigue life changes induced from each machining method, whilst all repair methods presented fatigue strengths that would surpass the fatigue performance of components which had sustained foreign object damage and had not been repaired. Repair geometries and locations were determined through finite element analysis performed by Rolls-Royce Compressor Stress teams, where a worst-case scenario for component loading could be simulated with these repairs. Each of the repair types demonstrated fatigue life above the minimum required as defined by Rolls-Royce for project advancement. Material microstructure remains largely unchanged when laser processing is used, when compared to mechanically repaired components. The only easily observable change is material recast, and this falls within acceptable limits as defined in Rolls-Royce laser machining standards in every case [1]. Fracture surface inspection confirms results seen in the metallurgical study. Average fatigue strength after 107 cycles was calculated for components repaired with each method, showing similar performance between laser and mechanical blending.
As the prototype tooling was in continuous development, the approach angle of the tool against the blade and hence the repair geometry changed, and as such repeated tests were carried out to validate fatigue performance on different repair geometries.
A number of figures and technical details have been redacted from this version of the project thesis to protect the intellectual property and commercial interests of Rolls-Royce PLC.
Actions (Archive Staff Only)
 |
Edit View |
Tools
Tools
![[thumbnail of DG Thesis Viva + RR Corrections 17.pdf]](https://eprints.nottingham.ac.uk/style/images/fileicons/application_pdf.png "DG Thesis Viva + RR Corrections 17.pdf")