PEEN FORMING SIMULATION

CROSS-REFERENCE TO RELATED APPLICATION

[0001] The present application claims the priority of United States Provisional Patent Application No. 62/559,936, filed on September 18, 2017.

TECHNICAL FIELD

[0002] The present application pertains to peen forming equipment and related process.

BACKGROUND OF THE ART

[0003] Peening processes are primarily used as cold-work surface treatments to improve material properties of a parts. There exist a variety of peening processes such as shot peening, laser peening, needle peening and ultrasonic peening.

[0004] All peening processes aim at plastically deforming a thin layer of material near the surface of the part. The strain mismatch between this layer and the rest of the part - left unaffected by the treatment - induces a compressive residual stress field near the surface. Compressive residual stresses can inhibit or delay the initiation or propagation of cracks, thus improving the fatigue life of the part. The tools used to plastically deform the material may differ from one process to the other. For example, laser peening relies on shock waves generated by the expansion of a plasma generated near the surface of the part by a laser pulse, whereas needle peening relies on a mechanical striker propelled at high velocity. Shot peen forming is a particular peening process that consists in bombarding metal parts with small shots at high velocities.

[0005] The strain mismatch between peened and unpeened material also results in a distortion of the part. This effect is negligible when the intensity of the treatment is low and the part massive; it can be significant when thin-walled structures are peened with high intensity treatments.

[0006] Peen forming consists in the application of a peening process to a part with the intent of inducing such distortions. Peen forming can be used to form parts. For example, aircraft lower wing panels are usually peen formed from an initially flat machined aluminum panel. Fig. 1 shows schematically the application of the shot- peening process on a part, as well as typical peening-induced plastic strains and the associated residual stress field. It can also be used to correct small amplitude unwanted distortions. This latter application of the process is called distortion correction or straightening.

[0007] When peened uniformly, plates of uniform thickness may deform into hemispherical, elliptical or cylindrical shapes, as shown in Fig. 2. To obtain more complex shapes, more advanced peening strategies may be used. Varying the intensity of the treatment and/or the density of impacts (a.k.a. coverage) over the surface is one possible strategy. Another technique, called stress-peen forming, consists in elastically pre-straining the part. This results in larger plastic strains in the direction in which the part is being stretched. A combination of these techniques is usually required to form complex industrial parts.

[0008] Although peen forming has been used for numerous years, in particular in the aerospace industry to form wing and fuselage panels, process parameters are still determined most of the time either by a costly trial and error approach, or by analogy with existing peening recipes on similar parts. Moreover, peening processes may be operated in free-hand mode by operators. The introduction of simulation tools is desirable as it has the potential to shorten the time necessary to design new peening strategies and/or enable exploring new designs for which the entity performing the peening has limited experience, without the costly trial and error approach.

SUMMARY

[0009] It is an aim of the present disclosure to provide a method and system to simulate a peen forming process.

[0010] It is a further aim of the present disclosure to provide a method and system to output a peening process model (peening parameters and pattern) to peen form a part based on a simulation.

[0011] It is a further aim of the present disclosure to provide a method and system for operating peening equipment with a peening process model to peen form a part based on a simulation. [0012] Therefore, in accordance with a first embodiment of the present disclosure, there is provided a system for producing a peening process model comprising: a processing unit; and a non-transitory computer-readable memory communicatively coupled to the processing unit and comprising computer-readable program instructions executable by the processing unit for: obtaining models of peening induced loads as a function of a plurality of peening treatments, obtaining models of a current shape and of a target shape of a part, using the models of the part, identifying a plurality of combinations of peening treatment and peening pattern to reach the target shape of the part, simulating a peening of the part with the plurality of combinations of peening treatment and peening pattern using the models of peening induced loads and the model of the current shape of the part, selecting one of the plurality of combinations of peening treatment and peening pattern from the simulating, and outputting the selected one of the combinations as a peening process model adapted to drive peening equipment to peen form the part from the current shape to the target shape.

[0013] Further in accordance with the first embodiment, obtaining models of peening induced loads includes for instance generating the peening induced loads as a function of the plurality of peening treatments.

[0014] Still further in accordance with the first embodiment, generating the peening induced loads includes for instance characterizing peening induced plastic strains using an inverse reconstruction procedure to compute plastic strains from curvature measurements and/or residual stresses.

[0015] Still further in accordance with the first embodiment, the compute of plastic strains from curvature measurements and/or residual stresses includes for instance obtaining the curvature measurements and/or residual stresses with X-ray diffraction.

[0016] Still further in accordance with the first embodiment, generating the peening induced loads includes for instance interpolating one of the peening induced loads from a database of characterized peening induced loads.

[0017] Still further in accordance with the first embodiment, identifying a plurality of combinations of peening treatment and peening pattern includes for instance identifying one of the plurality of combinations with a maximum intensity for the peening treatment without any constraint on a peened area of the part.

[0018] Still further in accordance with the first embodiment, simulating the peening of the part includes for instance beginning the simulating with the one of the plurality of combinations with a higher intensity for the peening treatment.

[0019] Still further in accordance with the first embodiment, selecting one of the plurality of combinations of peening treatment and peening pattern from the simulating includes for instance discarding the target shape as unfeasible if the target shape is not achieved through the simulating with the one of the plurality of combinations with the higher intensity for the peening treatment.

[0020] Still further in accordance with the first embodiment, simulating the peening of the part includes for instance simulating the peening of the part with a progressive decrease of an intensity of the peening treatment relative to the higher intensity to reach a simulated shape within tolerances of the target shape of the part for a lower intensity of the peening treatment.

[0021] Still further in accordance with the first embodiment, simulating the peening of the part includes for instance simulating the peening of the part with a progressive decrease of an intensity of the peening treatment from the plurality of the combinations to reach a simulated shape within tolerances of the target shape of the part for a lower intensity of the peening treatment.

[0022] Still further in accordance with the first embodiment, simulating the peening of the part includes for instance identifying a checkerboard pattern, and setting an upper bound constraint on a peened area for the simulating of peen only areas that have a more pronounced influence over the target shape.

[0023] Still further in accordance with the first embodiment, the system is for producing for instance a peening process model for shot peening, and wherein the peening treatment includes for instance parameters including shot speed, type of shots and coverage.

[0024] Still further in accordance with the first embodiment, obtaining the models of the current shape of the part includes for instance obtaining a model of a three- dimensional scan of the part. [0025] Still further in accordance with the first embodiment, outputting the peening process model includes for instance outputting the peening process model as human readable instructions for manually performing the peen form of the part.

[0026] Still further in accordance with the first embodiment, outputting the peening process model as human readable instructions includes for instance maps of the part with intensity and/or coverage.

[0027] Still further in accordance with the first embodiment, outputting the selected one of the combinations as a peening process model includes for instance driving the peening equipment to peen form the part from the current shape to the target shape.

[0028] Still further in accordance with the first embodiment, obtaining the models of the peening induced loads includes for instance obtaining the models of the peening induced loads as eigenstrains as a function of the plurality of peening treatments.

[0029] Still further in accordance with the first embodiment, an automated peening apparatus comprises for instance the system described above.

[0030] In accordance with a second embodiment of the present disclosure, there is provided a system for producing a peening process model comprising: a peening load modeling module for obtaining models of peening induced loads as a function of a plurality of peening treatments; and a peening optimization module for obtaining models of a current shape and of a target shape of a part, identifying a plurality of combinations of peening treatment and peening pattern to reach the target shape of the part using the models of the part, simulating a peening of the part with the plurality of combinations of peening treatment and peening pattern using the models of peening induced loads and the model of the current shape of the part, selecting one of the plurality of combinations of peening treatment and peening pattern from the simulating, and whereby the system outputs the selected one of the combinations as a peening process model adapted to drive peening equipment to peen form the part from the current shape to the target shape.

[0031] Further in accordance with the second embodiment, the peening load modeling module generates for instance the peening induced loads as a function of the plurality of peening treatments. [0032] Still further in accordance with the second embodiment, the peening load modeling module characterizes for instance peening induced plastic strains using an inverse reconstruction procedure to compute plastic strains from curvature measurements and/or residual stresses to generates the peening induced loads.

[0033] Still further in accordance with the second embodiment, the peening load modeling module computes for instance the plastic strains from curvature measurements and/or residual stresses by obtaining the curvature measurements and/or residual stresses with X-ray diffraction.

[0034] Still further in accordance with the second embodiment, the peening load modeling module interpolates for instance at least one of the peening induced loads from a database of characterized peening induced loads to generate the peening induced loads.

[0035] Still further in accordance with the second embodiment, the peening optimization module identifies for instance a plurality of combinations of peening treatment and peening pattern by identifying one of the plurality of combinations with a higher intensity for the peening treatment without any constraint on a peened area of the part.

[0036] Still further in accordance with the second embodiment, the peening optimization module begins for instance the simulating with the one of the plurality of combinations with the higher intensity for the peening treatment.

[0037] Still further in accordance with the second embodiment, the peening optimization module discards for instance the target shape as unfeasible if the target shape is not achieved through the simulating with the one of the plurality of combinations with the higher intensity for the peening treatment.

[0038] Still further in accordance with the second embodiment, the peening optimization module simulates for instance the peening of the part with a progressive decrease of an intensity of the peening treatment from the plurality of the combinations to reach a simulated shape within tolerances of the target shape of the part for a lower intensity of the peening treatment.

[0039] Still further in accordance with the second embodiment, the peening optimization module simulates for instance the peening of the part with a progressive decrease of an intensity of the peening treatment from the plurality of the combinations to reach a simulated shape within tolerances of the target shape of the part for a lower intensity of the peening treatment.

[0040] Still further in accordance with the second embodiment, the peening optimization module identifies for instance a checkerboard pattern while simulating the peening of the part, and sets an upper bound constraint on a peened area for the simulating of peen only areas that have a more pronounced influence over the target shape.

[0041] Still further in accordance with the second embodiment, the system outputs for instance a peening process model for shot peening, and wherein the peening treatment includes for instance parameters including shot speed, type of shots and coverage.

[0042] Still further in accordance with the second embodiment, the peening load modeling module obtains for instance the models of the current shape of the part as a three-dimensional scan of the part.

[0043] Still further in accordance with the second embodiment, the system outputs for instance the peening process model as human readable instructions for manually performing the peen form of the part.

[0044] Still further in accordance with the second embodiment, the peening process model has for instance human readable instructions includes for instance maps of the part with intensity and/or coverage.

[0045] Still further in accordance with the second embodiment, the peening equipment is for instance driven to peen form the part from the current shape to the target shape.

[0046] Still further in accordance with the second embodiment, the peening equipment peen forms for instance the part from the current shape to the target shape.

[0047] Still further in accordance with the second embodiment, the peening load modeling module obtains for instance the models of the peening induced loads as eigenstrains as a function of the plurality of peening treatments. DESCRIPTION OF THE DRAWINGS

[0048] Fig. 1 is a schematic view showing (a) shot peen forming of a part and (b) typical peening-induced plastic strains and the associated residual stress field;

[0049] Fig. 2 is a schematic illustration of experimental deformed shapes for 1 x 1 m AA2024-T3 panels of varying thicknesses peened uniformly with the same treatment;

[0050] Fig. 3 is a block diagram of a system for simulating and operating a peening process in accordance with the present disclosure; and

[0051] Fig. 4 is a schematic view of (a) a model of an initially flat panel divided into a number of subdomains that can either be unpeened, peened on one side, or peened on both sides and (b) the section properties used for any domain.

DETAILED DESCRIPTION

[0052] The present disclosure pertains to a method and system for simulating and operating a peen forming process, the system being generally illustrated at 10 in Fig. 3. The method and system determine a relationship between a part, a desired shape, a.k.a., a target shape for this part, and the peening parameters of a peening process, such as the type of shots, characteristics of the shot stream, peening trajectory, etc. The method and system 10 of the present disclosure may be used to set peening equipment parameters and perform a peening process to obtain the target shape. The method and system of the present disclosure may automatically compute optimal peening patterns that fit a given target shape of a part, and subsequently output a peening process model to peen the part into the target shape, or operate the peening equipment 30 to produce the part.

[0053] The system 10 is of the type including a processing unit (a.k.a., a controller), with one or more processors, generally shown at 20 that is devised to simulate a peening process on a part. The processing unit 20 may be specifically dedicated to simulating the peening process, and may also be used to drive the peening equipment 30 to produce parts. Part data is generally shown at A1 , A2 and is provided as an input in the processing unit 20. As an example, A1 and A2 may be three-dimensional digital models of a part, with A1 being the current shape of the part (prior to peening) and A2 being the target shape of the part. The current shape of the part A1 may result from technical drawings, from a 3D scan of the part A1 , or from other representation methods, for the model to reflect the current shape of the part A1. The method and system 10 are consequently used either for distortion correction (e.g., when deflections are small), or for peen forming (e.g., when deflections are moderate or large) in simulating the peening of the part. The target shape of the part A2 may be by design. The 3D models of the parts A1 and A2 may be in any appropriate digital format.

[0054] The processing unit 20 may include a non-transitory computer-readable memory communicatively coupled to the processing unit 20 and comprising computer-readable program instructions executable by the processing unit 20 for performing various sequences of steps and/or functions by way of modules. For example, the processing unit 20 may have various modules by which various functions may be performed to achieve the simulation, and to operate peening equipment 30, if desired. Accordingly, the processing unit 20 may output a peening process model M that may be used by peening equipment, such as that shown at 30, or other equipment. The peening process model M may be part specific, i.e., for each part A1 a unique peening process model M is defined. For example, slight variations between the initial state of several instances of the same part caused by variations during prior manufacturing processes may require different peening models M to obtain the same end-result, e.g., parts with the same target shape. It is also possible to use the same peening process model M for a plurality of parts if these parts can have the same current shape, or if current shape variations result in a target shape of various parts within tolerances.

[0055] The peening equipment 30 has the capacity of operating with given process parameters to perform given process patterns, for instance by integrating the processing unit 20. It can be a fully automated peening equipment that controls peening intensity, coverage, relative position of the tools with respect to the part, etc. For example, the nozzle shown in Fig. 1 is part of the peening equipment 30. A trained machine operator could also use the given process parameters to perform the peening process. In this latter case, the peening process model M is output in the form of human readable instructions such as listings of peening parameters, peening intensity maps, coverage maps, etc.

[0056] The processing unit 20 may have a peening load modeling module 21. The module 21 is used to link process parameters, such as shot speed, type of shots, coverage, to peening induced loads that will develop inside the part during peening. The peening induced loads may be characterized in terms of plastic strains (also referred to as eigenstrains) or any derived quantity such as so-called induced residual stresses (i.e., residual stresses that would exist inside a part if it had not been allowed to deform after peening). Plastic strains are a convenient way to model peening induced loads as they can usually be considered independent of the geometry of the part in typical applications of peen forming processes, i.e., applying the same peening treatment to two specimens made of the same material but whose geometries differ results in the same near-surface plastic strains. Conversely, once peening induced loads are characterized for a given treatment, they can be used as an input in a model of the peen forming process to compute the deformed shape of random parts. This property is known in the art not to hold for very high intensity treatments (such as high intensity laser peening) or in the vicinity of geometric features such as holes and free edges. Another example may be the characterization in the form of metric tensors to define the length of and angle between tangent vectors. Metric tensors are known to be functions taking as input a pair of tangent vectors at a point of a surface to produce a real number scalar to generalize many of the familiar properties of the dot product of vectors.

[0057] The module 21 can include a database of peening treatments, i.e. treatments for which the peening loads induced by the treatments in a given material were previously characterized. Interpolation in the database enables to approximate peening induced loads for new treatments. More data-points in the database may usually imply more accurate predictions. Data-points can be generated experimentally, or by any other method. For example, data-points can be generated by performing simulations by the module 21 , the results of which then populate the database for subsequent use. The experimental characterization of peening induced plastic strains implies the use of an inverse reconstruction procedure to compute plastic strains from other measurable quantities, such as curvature measurements or residual stresses obtained by X-ray diffraction. The method and system 10 may rely on the modeling of peening induced loads as eigenstrains, to enable the casting of the shape optimization of peen formed plates.

[0058] The module 21 can also consist of an analytical, semi-analytical, numerical, or empirical model of the peening process. For example, a finite-element simulation of the impact of shots on a representative volume of material can be used to obtain the peening load modeling of module 21 . In any case, these models of the peening process are process-specific, i.e. different models are required for different peening processes, whereas the same database structure can be used to store and interpolate peening induced loads for various peening treatments.

[0059] The processing unit 20 may also have a peening optimization module 22. The peening optimization module 22 computes peening treatments liable to form part P1 into part P2. The peening module 22 may consists of two interacting sub- modules: a peen forming model sub-module 22A and an optimization algorithm sub- module 22B. The optimization algorithm sub-module 22B produces candidate peening treatments, the outcome of which (i.e. the deformed shape of the part peened with this specific treatment) is then simulated via peen forming model sub- module 22A and compared against the target shape. If the agreement between the computed deformed shape and the target shape A2 is unsatisfactory, another iteration is performed. The process is repeated until convergence, or until a stopping criterion is met. The latter case may correspond to a failure of the process.

[0060] The peen forming model in the sub-module 22A may consist in a structural model of part A1 which takes a peening treatment as an input, the sub-module 22A computing the deformed shape of part A1 for this specific treatment. In an embodiment, it can be a finite element model. In anticipation of the optimization process, the model is partitioned in a number of domains. If a finite element module is used, each domain can be meshed with one or several elements. The partition in domains is shown in (a) of Fig. 4. In the illustrated embodiment, each domain is assigned tri-layer laminate section properties as in (b) of Fig. 4. Peening induced loads are input in the upper and lower layers of the laminate in the form of an in- plane expansion (i.e. eigenstrain). This can be done via a thermal analogy commonly used in the art, by prescribing a through thickness distribution of non-zero thermal expansion coefficients, and by applying a unit increment of temperature. The central layer does not expand. In Fig. 4, the part is shown as being a thin-walled structure. However, the system 10 may also perform simulations on massive structures by using appropriate 3D meshes.

[0061] To cast the optimization problem into a form that can be solved by the optimization algorithm sub-module 22B, each free surface of each domain that is allowed to be peened is linked to an optimization variable A i ^■ - [0, 1] , where i is an integer used to index the faces. The amplitude of the eigenstrain profile input in the model near the surface of domain i is equal to X i, where is an upper bound on the expansion set by the optimization module 22, as described below. The case }'■: = 1 corresponds to a fully peened area whereas J!':- = 0 corresponds to an unpeened area. A peening pattern is completely characterized by the vector χ with . V representing the number of faces) that collects all optimization variables.

[0062] The peening optimization module 22 performs an optimization procedure to find a combination of optimization variables that best fit the given target shape A2. The optimization problem can be cast in the form

where z and z _ta r _g et are vectors of coordinates of a subset of nodes, respectively in the current deformed configuration and in the target shape configuration, A is a vector that contains the area of the faces associated with each optimization variable, and Ί«^ _.'Η · is an upper bound on the peened area that can either be user-defined, or set by the optimization module 22. The constraint A χ < Λ ,, ^, is introduced as a remedy to pathological checkerboard patterns that typically arise with the unconstrained formulation. Depending on the problem at hand, it may be necessary to consider alternative expressions for the cost function and/or the constraints.

[0063] The model of the sub-module 22A used to compute the deformed shape A2 can be either geometrically linear (e.g., small deflections), or geometrically nonlinear (e.g., moderate to large deflections and rotations). If the model of the sub- module 22A is linear, the dependence of on X can be made explicit by using the superposition principle. If D is a matrix whose ^L column contains the nodes coordinates computed for Λ

recast in the equivalent form

This form corresponds to a norm approximation problem with constraints, which is convex. Convexity guarantees that if the optimization algorithm sub-module 22B finds a solution to the optimization problem, this solution is a global optimum, i.e. a better solution cannot be found. If alternative expressions for the cost function and/or the constraints are considered, the convexity might not hold anymore.

[0064] To solve the optimization problem, the optimization algorithm sub-module 22B can address the first form of the problem with generic gradient based algorithms. These algorithms exhibit fast convergence rate; usually, the overall shape of the pattern becomes apparent after a small number of iterations. Other optimization strategies known in the art might also be relevant, depending on the optimization algorithms available. If a linear peen forming model is used in sub- module 22A, the optimization problem is convex and can be addressed with more specific convex optimization algorithm. The second form can also be addressed with a constrained least square algorithm. This forms requires the assembly of the D matrix, which is time consuming for large models.

[0065] According to an embodiment, the peening optimization module 22 conducts a sequence of steps - an optimization loop - to compute peening patterns, based on the peening treatment and with a view to reaching the target shape A2. The following sequence may be used by the peening optimization module 22 to iteratively identify suitable peening patterns and peening treatments from scratch, based on the optimization strategy presented above.

(i) An evaluation is made to ensure that the target shape is 'feasible', i.e., that there exist at least one combination of peening treatment and peening pattern that can approximate the target shape A2 with sufficient accuracy. A brute force approach for such evaluation consists in running a simulation without any constraint on the peened area, with the most intense peening treatment admissible for the current application (i.e. with set to its upper limit). If the agreement is unsatisfactory at the end of the analysis, i.e., if the converged peening treatment computed by the optimization algorithm sub-module 22B does not approximate the target shape with appropriate tolerances, the target shape may most likely not be obtained, whereby the method and system 10 may discard the target shape as a feasible shape. Alternatively, a peening treatment of higher intensity is identified, i.e., one that is evaluated to exceed the minimum intensity of peening treatment to produce the target piece. Another expression for higher intensity is higher threshold intensity, upper threshold intensity, etc.

(ii) If the evaluation indicates that the target shape A2 is feasible, the peening optimization module 22 may re-run the analysis by progressively decreasing the intensity of the treatment (relative to the higher intensity) until the agreement becomes unacceptable, i.e., the simulated shape goes beyond the tolerances of the target shape A2. This gives a lower bound on the intensity of the treatment. A lower intensity treatment may result in lesser damage to the part.

(iii) Based on the results of (ii), an appropriate peening treatment is identified and selected for yielding a suitable expansion, by using the load modeling module 21. The load modeling module 21 can identify an existing treatment, or suggest process parameters likely to yield the desired expansion, for example by interpolation in a database of treatments.

(iv) Another optimization loop is performed with the treatment identified in (iii). If the optimal peening pattern obtained at the end of this loop presents numerical artifacts such as checkerboard patterns, the upper bound constraint on the peened area may be enforced by the peening optimization module 22. This constraint forces the code to peen only areas that have the most pronounced influence over the final target shape. Progressively decreasing the upper bound A steers the optimal pattern towards 0-1 designs, i.e., consisting of only fully peened and unpeened areas, and makes unwanted checkerboard patterns disappear. Some applications might require smooth transitions between peened and unpeened area. Such gradients can usually be obtained for moderate values of ^.- _: - _Λ or with additional smoothing constraints known in the art. Hence, other constraints may be added in the simulation.

The sequence (i) to (iv) can be performed automatically to minimize human intervention. [0066] To speed up the optimization process, one can first go through the whole optimization procedure by using a geometric-linear peen forming model in sub- module 22A, which is fast to evaluate, and for which the optimization problem can be shown to be convex, with the associated advantages. For a linear-geometric model, most quantities used during the optimization process can be computed once at the beginning of the analysis, outside of the optimization loop. The stiffness matrix used by finite element models is an example of such quantities. This, together with the reordering and pre-factorization of all matrices in anticipation of solving large linear systems considerably speeds up the solution process. The optimal pattern obtained at the end of the optimization process can then be used as an initial guess for a second optimization phase, this time with a geometrically non-linear model.

[0067] If in-plane deformations are of secondary importance (for example at the beginning of the design process of a new part), it is possible to reduce the number of optimization variables by half by allowing only one of the layers of each element to expand at a time, instead of both at the same time. This comes from the fact that increasing the expansion in both layers by the same amount results in an in-plane expansion only (no additional bending). Halving the number of optimization variables in this manner reduces the computing cost of the optimization loop.

[0068] At the outset of the optimization loop, the optimization peening module 22 produces a peening process model M that includes the peening pattern to control the position of the nozzle relative to the part, and the peening treatment of the shot or like peening medium, for the part P2 to be peen formed from the shape A1 to the shape A2. An equipment driver module 23 may be integral to the processing unit 20 and may use the peening process model M to peen form the part P2, by driving the peening equipment 30.

[0069] The method and system 10 of the present disclosure are able to compute 0-1 patterns which can be peened by masking the appropriate areas. It is also more flexible as it is not restricted to least square problems form, i.e., the procedure can handle various forms of the cost function and constraints.

[0070] Consequently, the method and system may be used to compute peening patterns for distortion correction (i.e. small amplitude deformations), for instance on thin walled structures such as integrally stiffened panels. The method and system may also be used to compute optimal peening patterns for peen forming applications when deflections are small to moderate, or when the parts are firmly held into place during peening (e.g., by some supporting device or clamps) and released afterwards.

[0071] The method and system 10 are occasionally described above as being used with shot peening. However, since the physical source of distortions (i.e., the expansion of subsurface plastically deformed layers of material) is the same for all peening processes, the method and system 10 of the present disclosure can also be applied to these processes. The system 10 consequently produces a peening process model by obtaining models of peening induced loads as a function of a plurality of peening treatments, obtaining models of a current shape and of a target shape of a part, using the models of the part, identifying a plurality of combinations of peening treatment and peening pattern to reach the target shape of the part, simulating a peening of the part with the plurality of combinations of peening treatment and peening pattern using the models of peening induced loads and the model of the current shape of the part, selecting one of the plurality of combinations of peening treatment and peening pattern from the simulating, and outputting the selected one of the combinations as a peening process model adapted to drive peening equipment to peen form the part from the current shape to the target shape.