Shot Peening And Deep Rolling; Development Of A Comparison Model
Abstract: This article describes the method and the results for the development of a direct comparability between shot peening and deep rolling. Therefore the method of generating saturation curves by shot peening of Almen strips was used to evaluate the effect of deep rolling. In analogy to shot peening, saturation curves were created for deep rolling by variation of rolling pressure, path clearance and rolling direction. Parameter sets targeting an intensity range representative for the application on aero engine parts were generated. Finally the output of both techniques on the surface roughness was evaluated.
Companies’ daily business is dictated by producing economically and safely. The focus is therefore also on the manufacturing processes, as well as on their capabilities, accuracy, reproducibility and economic efficiency. The economic efficiency is determined, beside the cost of the process, by the impact on the manufactured part. To enhance the high cycle fatigue (HCF) resistance of aerofoils on compressor disks, a couple of mechanical surface treatments are state of the art technology. Besides laser shock peening, cavitation peening and ultrasonic shot peening, the two most used methods are shot peening and deep rolling. 
In almost every rotating aero engine, part work hardening techniques are in use due to a positive impact on the fatigue properties. Shot peening is used for high complex shaped parts like blisks (blade integrated disks), whereas the main use of deep rolling lies on rotational symmetrical components. Through plastic deformation of the surface layer, an improvement of the mechanical strength is generated. Beside the increase of the dislocation density, caused by the plastic deformation that elevates the surface hardness, residual compressive stresses are placed in the surface and subsurface layers. The hardness increase generates a higher resistance against crack initiation; the residual compressive stresses against crack propagation. Both have a significant effect on the HCF performance. 
The surface roughness has an effect on the HCF strength as well. The unsystematic impacts of the spherical shots during peening increase the surface roughness against what the deep rolling treatment decreases the roughness, by rolling parallel and overlapping tracks. Shot peening has the advantage to be able to peen high complex geometries by leading the shots through hose and nozzles directed to the surface area. Deep rolling is limited by the accessibility, the tooling design (transmission of load) and the treatment time. One mayor advantage of deep rolling on the other hand is the capability of inducing residual compressive stresses in deeper material layers than shot peening.
The two mechanical surface treatment technologies (shot peening and deep rolling) are used for a similar field of application. Until now, no method to compare the induced energy through shot peening with the force at deep rolling is known. In many reports, the intensity and coverage gained by shot peening is compared with the pressure and the path clearance of deep rolling. There is no direct comparability.
The development of parameter sets that produce similar residual stresses in the surface layer is only possible through extensive experimental series and cost intensive diffractometric examinations with X-ray or synchrotron validation.
3.1 Shot peening
The shot peening method [3;4] as it is used in industry today, has existed since the 1920s. At the end of the 1940s, J.O. Almen invented the Almen strip method, which enables the comparison of different peening applications. Geometrically defined spring steel strips are used to produce a saturation curve . For that purpose, a couple of strips are peened with identical parameters except the feed rate. The different feed rates generate different deflections of the Almen strips whereat the decrease of the feed rate creates an increase of the deflection. By reducing the feed rate, the deflection reaches saturation so that an additional reduction does not have an effect on the deflection. It is possible to calculate the saturation point, at which a doubling of the treatment time produces a deflection increase of less than 10% by reading the saturation curve (see diagram 1). This point identifies the intensity of the peening application.
The intensity value, which can be easily obtained, allows a quantitative statement about the energy transported in the accelerated stream. In first approximation, this value allows information about the residual compressive stresses induced in the peened component. For this consideration, the material specific deformation characteristics of the peened material define the peening effect.
A second important value is the coverage.  The coverage identifies the proportion between the shot indentation area and the total area to be peened. For aviation applications, the minimal required coverage is 100% caused by the demand for uniform surface topography and prevention of stress peaks. The coverage is, contrary to the intensity, not detected on Almen strips but direct on the component surface. This is caused by the different elastic and plastic deformation characteristics of the different raw materials.
3.2 Deep rolling
Neither the method nor the term are clearly defined or classified in a standard. Besides the name deep rolling, the terms deep rolling, roller burnishing or low plasticity burnishing are commonly used. Deep rolling combines three main physical effects in cold work hardening of the surface. Through a spring or a hydraulically pumping system, a pressure is transferred to a rolling element, which passes the pressure onto the component surface. The rolling element produces an elastic and plastic deformation of the surface layer, which is controlled by the rolling force and results in cold work hardening and residual compressive stresses. Furthermore, the treatment has a smoothing effect on the surface. This depends strongly on the previous surface quality of the component and the finishing quality of the rolling body. All of the three affects: the cold work hardening of the surface, the introducing of residual compressive stresses and the reduction of the roughness, have a positive effect on the fatigue strength of components. In addition, the effect of decreasing the roughness is a clear advantage compared to shot peening which increases the roughness, especially concerning aerodynamical surfaces. With deep rolling, it is also possible to roll aligned paths, which open the opportunity to create structured surfaces (riblet structures).
The contact force is the main parameter, controlling the different residual compressive stresses induced through deep rolling. The pressure and the lining of the tooling control this force. Currently, there is no standardized control method like the intensity detection in shot peening. One method is not standardized yet, but a newly developed process control monitors and checks the relevant parameters for deep rolling force.
In contrast to the intensity, the deep rolling coverage is very easy to obtain. 100% coverage is defined as the distance between two rolling paths that is half the width of one path. These values are determined similar to the shot peening coverage directly on the component. The required coverage for deep rolling is 100% as well, to avoid stress steps or sharp rolling edges between the paths, which can support a crack initiation.
4. Experimental series
Deep rolling of Almen strips is used to gain the typical intensity range of 0,2 – 0,3 mmN for shot peening of blades. Two experimental series were carried out, one with the deep rolling tool HG3 and one with HG6. Each of this series included a variation of pressure, path distance and rolling direction. The investigation about the impact of different rolling element material, the influence of different lubricants and the feed rate, showed no significant effect.
After producing the saturation curves and the identification of the intensities, deep rolling parameters that create intensities comparable to shot peening were evaluated. These were operated on original engine material specimens to develop the coverage and to check the roughness.
4.1 Variation of pressure
The variation of the hydrostatic pressure during deep rolling can be compared with the variation of the air pressure to transport the media during shot peening. With increasing the pressure, the available energy to plastically deform the surface layer increases the analog of the indentation depth. With both methods, the deflection of the Almen strip increases and therewith the intensity. The shot peening treatment affects as well an increase of the roughness.
4.2 Variation of coverage
Shot peening coverage is produced by media hitting the surface randomly. Deep rolling in contrast generates parallel paths with defined distances. The variation of the path distances produces (similarly to the decrease of the feed rate during shot peening) an increase of the deflection of the Almen strips. This is caused by the increase of the force effect that is dependent on the surface area. With decreasing the path distance, the increase of the deflection reduces until a saturation value is reached. Beyond this value, a further decrease of the path distance causes no further increase of the Almen strip deflection.
4.3 Rolling direction
The variations described above were each carried out for one rolling direction. In each case, variation of pressure and coverage were operated for lengthwise (length) and crosswise (width) direction. The different induced residual stress fields of the different rolling direction cause this. The amount of residual stresses measured crosswise is typically higher than lengthwise. That means that a treatment with identical parameters will generate a higher deflection (intensity) after crosswise treatment. Such a direction dependency was not observable at shot peening within an angle of impact of 90° ± 15°.
5.1 Intensity curves
The intensity curves were created for a specific pressure and a certain rolling direction. For each of the curves, a set of Almen strips was rolled with different path distances. With the deflection of the Almen strips dependent on the processing time, the saturation curves can be created. Target was the adjustment of the pressure in a way where the intensity gained by rolling lengthwise as well as the intensity gained by rolling crosswise is within the required intensity range of 0,2 – 0,3 mmN. In all diagrams, the shot peened saturation curves for 0,2 mmN and 0,3 mmN are shown.
Diagram 2 shows the saturation curves for both tools (HG3 & HG6) at a pressure of 80 bar. Noticeable on the one hand is the difference between the saturation curves of the lengthwise (lw) and crosswise (cw) treatments and on the other hand, the difference between the different ball diameters of the tools.
Diagram 3 shows the saturation curves for the HG6 tool at 60 bar. With this pressure, it is possible to obtain the specified intensity range with both lengthwise and crosswise treatments. The lengthwise treatment induces an intensity of 0,23 mmN; the crosswise of 0,31 mmN.
Diagram 4 shows, analog to diagram 3, the saturation curve for the pressure of 100 bar of the HG3 tool. Both saturation curves lie within the required range; the lengthwise treatment produces an intensity of 0,26 mmN; the crosswise of 0,31 mmN.
The experimental series show that the intensity difference between lengthwise and crosswise treatments are approx. 0,05 mmN for the HG3 tool and 0,08 mmN for the HG6 tool.
For each set of parameters, two faces of original engine material specimen were shot peened and deep rolled. Subsequently, three positions on each face were measured with a tactile process using a Perthometer (made by Mahr) each lengthwise and crosswise to the rolling direction.
Diagram 5 shows the input roughness, which was Ra 0,25 µm on average. The shot peening treatment with the low intensity of 0,2 mmN increases the roughness up to Ra 0,55 µm. High intensity of 0,3 mmN increases the roughness even up to Ra 0,85 µm. The difference between 100% and 200% coverage was within measurement inaccuracy. Both peening treatments increase the roughness significantly, which makes an additional process step necessary to reduce the roughness.
After the treatment, the deep rolled surfaces show a uniform surface structure and are below Ra 0,15 µm for all pressures and both tools. There was no difference in roughness between lengthwise- and crosswise-treated surfaces.
6. Summary and outlook
The examination has demonstrated that a transfer of the intensity determination from shot peening to deep rolling is possible. It was shown that the variations of rolling pressures, path distances and rolling direction generate saturation curves. Regarding rolling balls with different diameters, it was presented that identical parameters result in different intensities. Furthermore the influence of rolling element materials, lubricants and feed rates were investigated, but did not show an effect.
In addition two sets of parameters for deep rolling were developed, which leads to saturation curves and intensities similar to shot peening values used for today’s aero engine components. A clear interaction between intensity and rolling direction was identified. This made a parameter adjustment necessary concerning the different intensities of lengthwise- and crosswise-treated specimens. The parameters were adapted in a way where both intensities reached the target range. The pressure adjustment has to be set very precisely so that the reproducibility of the saturation curves and the intensities are given.
Subsequently, specimens made from original engine component materials were deep-rolled with the generated parameter sets to check the roughness evolution and to compare the results with the results from shot peening. Both shot peening treatments show a clear increase of roughness, opposite to all deep rolling treatments, which show a significant decreasing effect. With focus on the aerodynamic requirements on aero engine components, deep rolling is beneficial as it makes the polishing process step obsolete.
For the analysis of the treatment effects regarding the Almen strips and the intensity, the specimens are scanned via X-ray diffraction to get the residual stress distributions. This examination will show if the intensities and the residual stresses are related to each other. Furthermore, a comparison of the high cycle fatigue on treated original engine components would be desirable, but was not part of this investigation.
The authors gratefully thank the energetic support of the staff of ECOROLL AG and Rolls-Royce Deutschland Ltd. & Co KG. Part of this work was funded by the "Bundesministrium für Wirtschaft" (BMWi) through the "Luffahrtforschungsprogramm" (support code 20T0908C).
 SCHULZE, V.: Modern Mechanical Surface Treatment: States, Stability, Effects. Germany: Wiley-VCH Verlag GmbH & Co. KGaA, 2006
 WOHLFAHRT, H.: The Influence of Peening Conditions on the Resulting Distribution of Residual Stress. 1984 (Conference Proceeding ICSP2). – 316–331 S.
 HOROWITZ, I.: Oberflächenbehandlung mittels Strahlmittel. Bd. 2. Essen: Vulkan-Verlag, 1976
 WAGNER, L.: Shot Peening. Germany: Wiley-VCH Verlag GmbH & Co. KGaA, 2003
 AUTOMOTIVE ENGINEERING, Society of (Hrsg.): SAE J443 Procedure for Using Standard Shot Peening Test Strip. United States: Surface Enhancement Committee, 2003
 AUTOMOTIVE ENGINEERING, Society of (Hrsg.): SAE J2277 Shot Peening Coverage Determination. United States: Surface Enhancement Committee, 2009
Dipl.-Ing. Goetz G.Feldmann
Dr.-Ing. Thomas Haubold
Rolls-Royce Deutschland Ltd.&Co.KG
Dipl.-Ing. Sirko Fricke
Ecoroll AG, Celle, Germany