The contents in this report have been published in 5th CIRP CSI 2020 under the title- ‘Data Driven Optimization of Vibropeening’ © 2020 The Authors. Published by Elsevier B.V.
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Inconel (IN) 718 is one of the nickel-based superalloys that is widely adopted in aerospace manufacturing of gas turbine engine; for instance, the high-pressure compressor rotors in Rolls-Royce Trent XWB engine (Feldmann et al. ). Aerospace components are subjected to high levels of alternating vibrational stress and extreme environmental conditions, thus high cycle fatigue (HCF) failure is one of the key challenges. As such, improving the fatigue life through a suitable surface treatment process becomes very important. Shot peening is a surface treatment process which can generate high compressive residual stress (RS) at the surface and work-hardened layer with the impingement of metallic shot at high velocities. This enhances the fatigue resistance but produces high surface waviness due to the dimples created by the shot. The high surface roughness acts as stress raisers and crack initiation sites, which lead to reduction in fatigue life (Dowling et al. ). Hence, vibropolishing is required as an additional step to finish the surface roughness (Ra) to below 0.25µm to prevent any reduction in fatigue life improvement introduced by shot peening and to meet the aerodynamic requirements (Feldmann et al. ). An alternative mass surface treatment process known as vibropeening (VP) has been identified as it combines the beneficial impact of the shot peening process to impart compressive RS as well as achieving identical surface finish as vibropolishing in one single step, simultaneously resulting in an equivalence in HCF strength as per current method of manufacturing (Feldmann et al. ).
Studies by Ahluwalia et al. introduced the concept of double vibropolishing by introducing an external vibration on Ti-6Al-4V workpieces attached onto a vibrator; consequently, the vibropolishing cycle time significantly increased more than 50%. This could be attributed to the high relative velocity and impact force between the media and vibrating workpieces in high frequency (Ahluwalia et al. ). In the production route of vibratory processing, the fixture that holds the workpiece is normally stationary for ease of bulk production; however, in vibratory processing, the adoption of vibrating fixture onto the trough so that they are vibrating in the same frequency and amplitude can greatly improve the process cycle time as mentioned by Sangid et al. . In this study, there are two optimizations conducted on the conventional VP process. Firstly, a smart sensor setup is introduced to develop more accurate process monitoring and optimize workpiece setup to improve peening capability. Next, an external actuator is explored to optimize the fixture setup and compare the residual stress and microstructural analyses between stationary fixture, vibrating fixture, and actuator.
HCF high cycle fatigue
RS residual stress
SEM Scanning electron microscope
RPM revolutions per minute (unit of motor frequency)
The experiments were conducted on a vibratory trough manufactured by Walther Trowal, TFM58/32VP with trough size of 580 mm x 320 mm x 360 mm. The trough is equipped with a PU lining container and motor that can drive the trough to vibrate at controlled frequency and amplitude. The setup configuration is illustrated in Fig. 1 with immersion depths of test piece labelled. In stationary fixture mode, the fixture holding the workpieces is attached on an external support, whereas in vibrating fixture mode, the fixture is attached onto the sides of the trough. An additional actuator is also positioned at a distance in front of the workpieces.
Almen deflection of Almen strips were collected at 10, 20, 40 and 80 minutes and measured using an Almen gauge to generate a saturation curve at each setting using a saturation curve solver from Electronics Inc. The RS profile of flat coupons was measured using XSTRESS Robotic X-Ray Diffraction (XRD) measurement system with layer removal using electropolishing steps. The microstructure of the test coupons is investigated using the Carl Zeiss EVO Scanning Electron Microscope (SEM) equipped with Energy Dispersive X-Ray (EDX) Spectrometer, which scans a sample with a focused electron beam of resolution 1.9nm at 30kV SE, and 8nm at 1kV SE, with the magnification range from less than 5x to 1000kx. The flat coupons are cut to view the cross sections and hot-mounted with conductive resin at 190ºC. The samples are then grinded, polished and chemical-etched using Waterless Kalling’s No.2 reagent for 70 seconds.
Two different test pieces are used in this investigation, i.e. Almen strips and IN718 flat coupons. The Almen strips used are of type N from Electronics Inc., with flatness of ±0.0010 and hardness of 44-50 HRC. The flat coupons are wire-cut from the heat-treated IN718 disc bore with dimensions of 30mm x 40mm x 4.5mm. The as-received condition of the coupons is similar as high pressure compressor blisks condition, which is degreased, heat-treated and anodized with Rolls-Royce proprietary heat treatment steps. The experimental parameters are shown in Table 1.
3. Results and Discussion
Due to word count restriction, in this section I have included the residual stress and material results alone from the original paper.
3.1. Residual stress profiles
In this section, the impact of fixture vibration was investigated in terms of RS and microstructure. Optimum treatment intensity of VP process is obtained at the lower-middle section, i.e. immersion depth of 130mm below the media surface level for the same trough (Chan et al. ). Thus, Figs. 2 and 3 show the RS profiles at different sub-surface depths of vibropeened IN718 flat coupons at immersion depth of -130mm with the external actuator compared to conventional processes in longitudinal and transverse direction respectively.
In VP process with stationary fixture, the as-received high surface tensile stress of ≈ +400MPa can be reduced to compressive stress state with maximum compressive stress at approximately -200MPa in both directions. When the fixture is vibrating together with the trough simultaneously, the RS profile has greatly improved with surface compressive stresses of ≈ -380MPa in longitudinal and ≈ -360MPa in transverse directions, and the maximum compressive stresses are able to achieve ≈ -750MPa in both directions. The depth of influence can also increase up to 100μm. This proved that fixturing plays an essential role in the efficiency of vibratory process as demonstrated by Sangid et al. . Furthermore, when a vibrating external actuator at frequency of 50Hz is inserted into the vibratory trough in front of the test piece, the RS profile has not shown much improvement compared to a vibrating fixture even though a slight increase in Almen intensity was observed. The higher media velocity near the vibrator has been captured by high-speed camera in previous work by Mediratta et al. .
The fundamental mechanism of RS imparting in VP can be deduced from this. In shot peening, the energy of the high velocity media is transferred to the component as impact thereby resulting in plastic deformation. In VP, similar plastic deformations are able to be introduced through a relative impact between component and media. Continuous relative vibration impacts component mass (usually in tens of kilos) with the media playing a crucial role in determining the stress imparted onto the surface. It can be concluded that for effective VP, the component must hit the media and not only the media hitting a stationary component. Key process variables for impact hence include relative vibration amplitude and frequency of component mass to media at given hardness ratio between component to media. This is low for a stationary fixture, hence less impact force and thereby lesser plastic deformation as demonstrated by Fig. 4. Further optimization can be done by designing fixtures that can increase the relative amplitude by keeping the trough amplitude and frequency constant.
The SEM images of IN718 sample with as-received condition and treated using stationary and vibrating fixtures are shown in Fig. 4(a), fine δ precipitates can be found in the Nickel matrix, which is normally found in heat-treated Ni-based alloys (Wang et al. ). This is expected as the samples used are from heat-treated IN718 disc-forging. The δ phase particles precipitated and intragranular δ precipitation can be observed as well. (Kumar et al. ). At the near-surface section, some granular structure can be seen and is revealed to be Niobium Carbides (NbC) after EDX analysis, which remain undissolved after heat treatment processing.
From Fig. 4(b), some grain boundaries are formed after the VP with stationary fixture, which could be due to plastic deformation. A discontinuous white layer can be seen after the VP is compared to as-received condition. The formation of this layer could be associated with the plastic deformation on the material, or due to rapid heating and cooling (Ahmad et al. ). Since these samples are from the same raw parts that undergo the same heat treatment, the major contribution of this layer would be related to severe plastic deformation next to the grain structure; however, by stationary fixture, the layer is discontinuous and explains the low compressive RS at the surface. Under 1000x magnification in Fig. 4(b), the white layer is very thin, approximately 3.5 to 5.4 μm wide. Needle-like δ precipitates have increased in number as compared to the as-received condition (Deng et al. ).
From Fig. 4(c), the SEM images of VP-treated IN718 samples using vibrating fixture showed that more grain boundaries are formed and more obvious compared to VP with stationary fixture, which suggests the higher compressive RS and greater depth of influence of VP on the samples using vibrating fixture. A thin continuous white layer can be observed in Fig. 4(c), and shows more even surface compared to previous SEM of stationary fixture. Similar to stationary fixture, more needle-like δ precipitates can also be found, unlike more spherical δ precipitates found in as-received samples (Wang et al. ). Besides, Laves phases can also been found in this sample, which are usually formed during the heat treatment process (Deng et al. ). In addition, dislocation and slip bands can be found in the grains and concentrated at surface and sub-surface layers up to approximately 50μm. These slips bands are related to the deformation and could be attributed to the higher compressive RS as achieved by vibrating fixture (Aezeden et al. ).
4. Conclusion and Future Work
The authors have demonstrated that fixture vibration is necessary for increased residual stress compared to a non-vibrating fixture, which coincides with the research of Sangid et al. . Residual stress and SEM microstructural analysis have been then used to identify the effect of the vibrating fixture on the IN 718 test pieces. SEM study showed the difference of grain structure and dislocation density between vibrating and stationary fixture modalities and supported the difference in RS as observed. More in-depth microstructure study can be done in the future, such as transmission electron microscopy (TEM), electron backscatter diffraction (EBSD) mapping, and micro-hardness test to understand the microstructural change, even with the setup with external actuator. Furthermore, future studies can also use some representative coupons, such as aerofoils to have more accurate results, as thickness and curvature might have differences in microstructure and RS generation.
Authors thank the Advanced Remanufacturing Technology Centre (ARTC) and Rolls-Royce for providing resources, facilities and technical support, as well as Drs. Thomas Haubold and. Goetz Feldmann from Rolls-Royce Deutschland for technical discussions and guidance.
 G. Feldmann, C. C. Wong, W. Wang and T. Haubold, "Application of Vibropeening on Aero-Engine Component," in 2nd CIRP Conference on Surface Integrity (CSI), Nottingham, 2014.
 N. E. Dowling, Mechanical Behavior of Materials: Engineering Methods for Deformation, Fracture, and Fatigue, Pearson, 2013.
 K. Ahluwalia, R. Mediratta and S. H. Yeo, "A novel approach to vibratory finishing: Double vibro-polishing," Materials and Manufacturing Processes, vol. 32, no. 9, pp. 998-1003, 2017.
 M. Sangid, J. Stori and P. Ferriera, "Process characterization of vibrostrengthening and application to fatigue enhancement of aluminum aerospace components-part I. Experimental study of process parameters," International Journal of Advanced Manufacturing Technology, vol. 53, pp. 545-560, 2011.
 W. L. Chan, K. Ahluwalia and A. Gopinath, "Parametric Study of Fixtured Vibropeening," Metals, vol. 9, no. 910, 2019.
 Mediratta R., K. Ahluwalia and Y. S. H. , "Surface motion analysis of double vibro-polishing of Ti-6Al-4 V," International Journal of Advanced Manufacturing Technology, vol. 97, pp. 1113-1122, 2018.
 Z. Wang, D. Zhou, Q. Deng, G. Chen and W. Xie, "The Microstructure and Mechanical Properties of Inconel 718 Fine Grain Ring Forging," The Minerals, Metals and Materials Society, 2010.
 D. Kumar, S. Idapalapati, W. Wei and N. Srikanth, "Effect of Surface Mechanical Treatments on the Microstructure-Property-Performance of Engineering Alloys," Materials, vol. 12, 2019.
 C. Ahmad, M. Hossein, R. Alex, J. Mohammad, B. Ali and M. Eric, "Microstructure and mechanical properties of surface and subsurface layers in broached and shot-peened Inconel-718 gas turbine disc firtrees," Materials Characterization, vol. 132, pp. 53-68, 2017.
 D. Deng, R. L. Peng, B. Håkan and M. Johan, "Microstructure and mechanical properties of Inconel 718 produced by selective laser melting: Sample orientation dependence and effects of post heat treatments," Materials Science & Engineering A, vol. 713, pp. 294-306, 2018.
 M. Aezeden, "Cyclic Deformation of Hastelloy and Inconel Alloys and Slip Bands Formation," International Journal of Modern Engineering Research (IJMER), vol. 3, no. 2, pp. 1253-1255, 2013.
Rolls-Royce Singapore, Central Technology Group, 1 Seletar Aerospace Crescent,