Methods for predicting the service life of components

Abstract

Method for computer-aided determination of the damage effect of defects in a component, comprising the steps of detection of a defect in a volumetric data set of the component, calculation of force-induced local deformations and / or stresses on a 3D data set of the component using structural simulation, calculation of a damage parameter to characterize the damage effect (including service life, strength, performance) of the defect based on a combination of the information from the structural simulation and the detected defect and / or other detected defects in a neighborhood of the defect.

Description

[0001] The invention relates to a computer-aided method and an associated system with which the load-bearing capacity or stress limits and thus the service life and performance of a component can be predicted non-destructively on the basis of volumetric calculations of the damaging effect of actual existing defects - also called inhomogeneities or incompleteness, sometimes also imperfections or defects - in a component.

[0002] Metallic components in particular often exhibit inhomogeneities in the form of external or internal defects, with internal defects including gas porosity, shrinkage cavities, cracks, bonding defects and foreign material inclusions, which lead to a reduction in component strength and thus the expected service life, especially under cyclic stress and / or dynamic / alternating loads, where the terms stress and load are to be understood synonymously - especially from a mechanical point of view.

[0003] In the current state of the art, a series of metallographic cross-sections are first prepared for the visual inspection of metallic components with regard to defects, particularly porosity. This is an invasive, component-destructive approach. Based on these cross-sections, a defect size distribution, a defect position distribution, or, more generally, a distribution with respect to various defect parameters can then be generated. For each defect, parameters such as size, position, shape parameters, and cross-sectional area can be determined. In addition to the porosity percentage, parameters relating to pore clusters are also determined and evaluated. Determining the maximum cross-section of an inhomogeneity or defect is not possible, as each defect is randomly sectioned and its spatial extent is evaluated only in a purely random section plane (see also BDG Guideline P 202).Computed tomography (CT) and subsequent defect analysis can also be used to generate histograms based on three-dimensional defect parameters. In general, computed tomography allows for the non-destructive determination and evaluation of most important defect parameters, including morphology, in all three dimensions, comprehensively and with sufficient accuracy, and—compared to the time-consuming preparation and visual evaluation of cross-sections—in a very short time. Due to the typically high number of defects found in components—in the range of 1000 or more—each individual defect is not currently evaluated together or superimposed with a local stress / deformation field. This also applies to the estimation of an individual service life based on a defect. Instead, statistical statements about the overall defect distribution, such as histograms or average defect size, are used., educated.

[0004] The parameters for the indicated inhomogeneities (defects) are then used to estimate the service life based on material models. Only defect characteristics detectable by CT or light microscopy can be evaluated. Furthermore, service life estimation is currently based solely on distributions or histograms, and thus averaged for the entire component. Alternatively, in addition to estimating component service life based on histograms, finite element analyses (FE analyses) are performed on CAD datasets of the component. An evaluation in the form of a purely geometric defect analysis of local deformations based on global component stresses can also be incorporated into the considered material models.

[0005] A defect-related assessment of the expected component service life based on all detected defects is not performed in the prior art, as an evaluation in conjunction with a deformation analysis is not available for this purpose. Therefore, typically only an estimation of the load-bearing capacity or stress limits, and thus the service life or performance of a component or structure, is made based on the overall distribution of defects. Crack paths or local crack initiation behavior cannot be directly determined in this way, as this requires a local analysis of individual defects in conjunction with a deformation analysis.

[0006] The methods known to date thus only use statistical probability statements for the components under investigation, applying a determined, mostly averaged, distribution of inhomogeneities (defects). Based on this, the object of the invention is to predict the damaging effect of inhomogeneities, such as defects in the form of porosity, in a component with higher accuracy and less time.

[0007] This problem is solved by a method according to claim 1, an associated system according to claim 14, and an associated computer program according to claim 15. Preferred embodiments are found in the dependent claims.

[0008] The invention comprises a method for the computer-aided determination of the damaging effect of defects in a component, comprising the steps of an indication of at least one defect in a volumetric data set of the component, a calculation of force-induced local deformations and / or stresses on a 3D data set of the component by means of structural simulation, and a calculation of a damage parameter to characterize the damaging effect of the defect based on a combination of the information from the structural simulation and the indicated defect and / or further defects in the vicinity of the indicated defect that influence the characterization of the damaging effect.

[0009] In the computer-aided method for determining the damaging effects of defects in a component, the results of defect detection, which generally indicates inhomogeneities, and structural simulation are superimposed and jointly evaluated. This enables a separate, local analysis of the defects. Particularly when combined with specific material models, this allows for an estimation of the damaging effect of each individual defect on structural integrity, e.g., on performance and service life. The calculated key performance indicators can be based on any combination of information from the structural analysis and the gray-scale-based defect analysis.

[0010] The method thus encompasses, in particular, the superimposition of geometric information, especially the defect position and / or shape, with simulation information in the form of a locally resolved stress field within the framework of deformation analysis in a single method or software program, and the resulting precise fusion of these two pieces of information. For this purpose, the relevant CT and CAD datasets are typically first aligned in a common coordinate system, preferably by means of a fixed alignment, which can also be referred to as rigid alignment. An additional elastic alignment or registration is used, in particular, if the datasets to be superimposed exhibit shape differences, as frequently occurs when superimposing actual and CAD geometries due to manufacturing variations.

[0011] The terms deformation and stress are usually directly linked. Deformation of a body causes mechanical stress within the body. A structural simulation (or structural analysis), or more precisely a structural mechanics simulation (or structural mechanics analysis), and its results therefore typically include both a deformation and a stress analysis. Normally, both the simulated stresses and the simulated deformations are important.

[0012] The detection of inhomogeneities or defects is preferably carried out using special algorithms, in particular by utilizing component / defect models as well as locally adaptive subvoxel-accurate surface determination. The detection of inhomogeneities or defects can generally also be automated and / or optimized using machine learning or artificial intelligence.

[0013] The term defect or inhomogeneity encompasses all deviations from the ideal state in the material of the component under investigation, including defects such as cracks, pores, bonding defects, cavities, dislocations, inclusions, as well as notches and variations in the microstructure, composition, or mixing ratio of different atoms, compounds, or substances. As in materials physics, the terms defect, inhomogeneity, and flaw are initially used synonymously here. Strictly speaking, an inhomogeneity in a broader sense can be described as a defect in the component's microstructure, but it is not necessarily a flaw.

[0014] The term volumetric dataset typically refers to a dataset in which defects can be detected, such as by a computed tomography (CT) scan, and which generally represents a collection of values ​​in three-dimensional space. In contrast, a 3D dataset usually represents a dataset that can be used for simulation, which is particularly relevant for CAD or CT data.

[0015] Structural simulation encompasses, in particular, the calculation of physical parameters of a three-dimensional structure or a representation of a component. The finite element method (FE method) is preferably used as the simulation method. Specific variations of the simulation model preferably employ linear-elastic or plastic approaches. Multiple load cases can be considered within the structural simulation, particularly iteratively. Special simulation technologies can also be used, such as fictitious-domain methods, mesh-free FEM methods, immersed-boundary methods, voxel-based methods, and finite-difference methods.

[0016] The invention utilizes, in particular, a precise three-dimensional representation of a component with its defects to predict, for example, the service life of the actual component with high accuracy based on the damaging effects of defects. In contrast to previously applied methods, the invention employs a defect-related approach. According to the prior art, methods using defect-related evaluation must perform the defect detection and simulation steps separately; that is, in a first step, the defect analysis, which corresponds to the indication of inhomogeneities (analysis of defects), would have to be output, and in a second step, this would have to be read into an FE program.

[0017] Currently, only separate programs exist for defect detection (program A) and simulation (program B). This requires, for example, determining the cross-sectional area of ​​a defect in the stress direction. First, the 3D volume of the defect must be output from program A, and then the complete stress state (a tensor-valued quantity), or at least the first principal direction of the stress state, must be read from program B at the corresponding points. Both pieces of information would then have to be read into a hypothetical program C and combined to calculate a so-called projected area according to the principal direction. Due to the limited output interfaces of defect detection and simulation software, and the lack of a dedicated program C, this is not yet possible according to the current state of the art, or it is far too complex to perform for 1000 or more defects.This can be practically achieved, in particular, by adequately combining or merging software A and B into a single software.

[0018] In a particular embodiment of the invention, the 3D dataset comprises a CAD model and / or a CT dataset. CAD stands for Computer Aided Design. A CAD model can contain numerous metadata elements in addition to an ensemble of point coordinates or values ​​in a coordinate grid, which increase the accuracy of the representation of a component.

[0019] If a 3D dataset is available in the form of a CAD model for structural simulation, it is free of internal defects, meaning the CAD model is free of inhomogeneities and thus without any internal flaws. Consequently, a simple approach can initially be taken to consider far-field approximation or averaging with respect to flaws. If a 3D dataset is available in the form of a CT dataset for structural simulation, then actual flaws and microstructures can also be taken into account. Since a CT dataset typically comprises a large amount of data, in this case, a selection of critical areas, especially selected flaws, is preferably carried out via submodeling, starting from a CAD simulation, which can lead to significant time savings.The aforementioned selection of critical areas is preferably based on heuristics relating to parameters such as defect size and / or defect clusters, in particular, for example, pore size and / or pore clusters, and / or increased far-field stress and / or stress intensity factor. The aforementioned or similar alternative approaches typically incorporate information from both defect detection and FE simulation.

[0020] In a preferred embodiment of the invention, the volumetric dataset comprises a measurement dataset and / or a computed tomography (CT) dataset and / or a series of micrographs. This enables the acquisition of, in particular, the structural properties of a component and their associated spatial distribution, including all defects or flaws of the actual component. The measurement data are preferably in grayscale form. In the case of micrographs, 2D grayscale images of a respective micrograph plane are typically available. For example, a relatively regular 3D dataset can be obtained by using a grid of parallel micrograph planes. In the case of a CT measurement, the result is typically a voxel grid with grayscale values ​​at each grid point.

[0021] In a particularly preferred embodiment of the invention, the defect is indicated and / or detected by means of grayscale analysis and / or a self-learning algorithm, wherein the self-learning algorithm is configured to characterize the shape of the defect, in particular to indicate it; that is, grayscale indications are captured, characterized, selected, and, if applicable, identified as inhomogeneities. Grayscale analysis is a typical evaluation method for images and, by extension, is also applicable to any type of measurement data exhibiting areal or volumetric features. Preferably, an analysis for grayscale indications and / or defects is performed using a self-learning algorithm.A self-learning algorithm primarily utilizes artificial intelligence, enabling the analysis algorithms to optimize their own parameters or methods—usually iteratively—according to previously specified target parameters. In a preferred embodiment, the self-learning algorithm can not only detect the position of a defect / fault but also characterize its shape. This characterization is typically achieved using parameters such as size and sphericity. Specifically, a 3D model of the defect can be created, for example, in the form of a surface model. Such detailed knowledge of the defect—especially in 3D—is essential for all subsequent steps, such as calculating the cross-sectional area in various projections and / or incorporating the defect into the simulation.

[0022] It should be noted that defect detection should not only identify defects, i.e., determine their presence and / or location, but preferably also characterize them geometrically. This is also necessary for many variations of steps 3 and 4 explained below. Characterization typically involves analyzing the 3D structure of the defects. In the simplest case, each defect can be approximated as a sphere or ellipse with specific dimensions. In a preferred variant, however, the surface of each defect is precisely segmented in order to determine, for example, the exact cross-sectional area for arbitrary sections and projections, to calculate local surface characterizations such as curvature values, and to perform morphological modifications to the shape, such as creating a convex hull or identifying the boundary region.Furthermore, precise knowledge of the 3D structure of a defect is required in order to include the influence of the defect in the structural mechanics simulation as accurately as possible.

[0023] In another preferred embodiment, the damage index is determined for a surrounding area adapted to the defect. Many defects encompass larger areas or are particularly influenced in their effect by neighboring defects. Typically, only a small proportion of defects are point defects. Consequently, a suitably defined surrounding area is preferably included in the calculation of the (local) damage index. This can be achieved, in particular, by considering an abstracted defect or flaw geometry, e.g., a convex hull instead of, for example, a C-shaped defect or flaw cross-section. The convex hull of the flaw is thus used instead of the actual flaw shape to represent the effect of the flaw more realistically.Preferably, two defects are considered as a single defect if the distance between the two defects is less than the square root of the combined area of ​​both defects. The defect area under consideration can, for example, also include a band with a defined thickness around the defect, where, in particular, the thickness is proportional to the defect size.

[0024] In a further embodiment of the invention, the prediction and / or calculation of a crack initiation location on the component and / or an associated crack initiation lifetime and / or an associated crack propagation lifetime and / or a damage tolerance is based on the damage characteristic value of a defect and / or the damage characteristics in the vicinity of a defect. In addition to the aforementioned damage tolerance based on the damage characteristic value of a defect, a combination of damage characteristics of specific defects is also possible. In particular, the determination of the crack initiation location is made possible by combining geometric information and simulation. This can be achieved, for example, by searching for the highest stress values ​​on the defect surface and / or by searching for sharp edges, e.g., by surface curvature analysis taking the stress state into account.This is not possible with existing methods that work with averaged distributions of inhomogeneities (defects).

[0025] A crack initiation site represents a defect where the material or solid structure typically exhibits a volumetric (e.g., pores, cavities) or planar void (e.g., bonding defects, cracks). This void, particularly under alternating load / stress, acts like an internal notch and serves as a starting point for defect propagation. A crack can also originate from the surface of the material without any cavities being present. For estimating the service life of a component, the crack initiation lifetime and / or crack propagation lifetime are of particular importance, as a crack initiation—more precisely, a notch—typically develops into a pronounced crack and thus a component failure under alternating load and / or cyclic stress. The respective lifetime can be calculated, in particular, by simulating defect growth.

[0026] Furthermore, based on various material mechanics design concepts, potential crack initiation sites can be directly identified by assigning a model-based damage parameter to each defect. This enables the classification of defects, particularly with regard to the prioritization of critical defects, their severity, and / or their impact on service life, or other criteria or weightings. This allows for the identification of particularly critical defects. In addition to the crack initiation site, the crack initiation lifetime and crack propagation lifetime can be estimated based on the damage parameters, especially when using specific models or modeling techniques, which are explained below.

[0027] A crack initiation-relevant parameter, also known as crack initiation, typically includes the stress concentration in the so-called critical plane, preferably also incorporating the crack initiation direction and location. Crack identification is performed in particular based on the so-called Theory of Critical Distance and / or Theory of Critical Area and / or Theory of Critical Volume and / or on hotspot methods and / or gradient methods. Damage-relevant material parameters are specifically considered when calculating the crack initiation lifetime. An assessment of the severity of a defect is usually an output value of the inventive approach. In a few cases, however, an assessment can be made based solely on defect detection, i.e., without the simulation and combination steps, if the location is particularly clear.This is the case, for example, when the defect is so small that an impact can be ruled out a priori based on heuristics. This is where this approach comes in. In a further embodiment of the invention, less precise simulations can be performed in areas that are considered not particularly critical from the outset, e.g., with coarser network discretization, in order to save time and / or computing resources.

[0028] In a further preferred embodiment of the method, this also includes the calculation of a crack growth-relevant parameter such as a defect area perpendicular to the loading direction and / or a stress and / or a stress intensity factor and / or a J-integral and / or an energy release rate and / or a defect / surface distance based on the damage characteristic of a defect and / or the damage characteristics in the vicinity of a defect. In the preferred embodiment, a defect area is calculated from the combination of the defect shape and the principal stress direction. An intensity factor is calculated from the defect area, the stress, the distance to the surface, and material-specific parameters. Alternatively, the intensity factor is calculated directly from the defect shape and the stress field via a J-integral or via the energy release rate.

[0029] The lifetime is calculated from the intensity factor using material models and material properties (e.g., Paris line).

[0030] In a particularly preferred embodiment of the method, the crack propagation lifetime is calculated based on the determination of a stress intensity factor and / or a J-integral and / or mean stress effects and / or hydrostatic stress effects and / or multiaxiality / triaxiality and / or crack closure and opening effects and / or anisotropy. The J-integral, in particular, denotes a continuum-mechanics-based parameter determined using an energy balance in the form of a line integral around the crack tip. Local plastic deformations are taken into account. Triaxiality, in particular, denotes a scalar quantity composed of the ratio of the mean normal stress to the von Mises equivalent stress.

[0031] In a further preferred embodiment, the calculation of a defect / damage tolerance is based on defect coagulation and / or defect / crack coagulation and / or on crack-stopping effects. Coagulation in materials refers in particular to the merging or rearrangement of defects, such as smaller cracks, into new, usually larger cracks.

[0032] In a specific embodiment of the method, non-effective defects, such as small size, large distance from the component edge, and / or low ambient stress, are not considered when calculating damage parameters. In particular, defects or inhomogeneities that are not heuristically relevant are not calculated, preferably to reduce computational demands and thus accelerate the calculations by focusing on relevant areas or locations. The method excludes non-effective defects primarily due to statistical considerations, as these have no impact on the damage. A corresponding selection of inhomogeneities or defects, determining when they should be considered, can be specified, for example, using predefined limits or thresholds.

[0033] In a particular embodiment, averaged stress values ​​from the structural simulation are preferably used to calculate a damage characteristic value for a defect. Averaging can be performed in the vicinity of the defect of a specific thickness. Averaging can also be performed by including the stress concentrations at neighboring defects. In particular, a damage characteristic value for a defect can also be calculated by including the object surface and thus the distance from the defect to the surface. The calculation of a general defect lifetime is typically performed from the sum of the initiation and propagation parameters.

[0034] In a further preferred embodiment of the method, this includes calculating a component characteristic value from the damage characteristics of several defects. This makes it possible, in particular, to determine a component-specific characteristic value that can represent an overall result. Specifically, it is conceivable to calculate the lifetime of the defect with the highest probability of failure, and / or to average the lifetimes of those defects with the highest probabilities of failure, and / or to determine the most critical defect, i.e., the defect that would likely be responsible for the component's destruction, especially with prediction of the crack location.

[0035] In a preferred embodiment, the calculation of a component characteristic number is based on the use of models according to Murakami and / or Shiozawa and / or Paris and / or damage-relevant material characteristics e.g. according to Smith-Watson-Topper and / or damage accumulation.

[0036] In a further preferred embodiment, the component identification number indicates the location of a first crack initiation and / or a weak point of the component and / or the crack propagation lifetime and / or a load-bearing capacity under static and / or dynamic / alternating / cyclic loads / stresses. The latter can, in particular, be based on component S-N curves specifying time- and fatigue-resistant load or stress limits.

[0037] In another preferred embodiment, the real defect is directly coupled or correlated with external / internal casting simulations, including with regard to mold filling, solidification and cooling behavior, as well as with reverse engineering, including with regard to casting optimization and / or structural topology optimization.

[0038] The invention further comprises a system with a device adapted to perform the method according to the invention. The system can, in particular, be a computer system that provides the volumetric or measurement data. Specifically, the computer system can include a measurement system that acquires the measurement data for component characterization, such as a computed tomography scanner or a microscope.

[0039] The invention also includes a computer program on a data carrier locally in or on the computer system or externally on servers (LAN / WLAN) or in a data cloud for carrying out the method according to the invention.

[0040] The invention will now be explained in more detail with reference to the drawings and a preferred embodiment.

[0041] The drawing shows Fig. 1 An illustration of the process steps of an embodiment according to the invention.

[0042] The Fig. Figure 1 schematically illustrates the sequence of steps in an embodiment of the method according to the invention. The abbreviations S1 to S5 shown in the figure mean: S1: Measurement / recording of volumetric component data (CT, microscopy, etc.) S2: Data analysis: Detection of defects / inhomogeneities S3: Structural simulation: Calculation of force-induced local deformations and / or stresses on a 3D CT dataset of the component S4: Calculation of a damage index for each defect using a combination of structural simulation and detected / indicated defect S5: Calculation of a component characteristic value to characterize the damage potential

[0043] In a first step S1, the component to be examined is measured using computed tomography or microscopy in the case of evaluating polished sections.

[0044] In step S2, a data analysis is then carried out, which in particular includes an indication of the inhomogeneities (defects) in the measurement data set as part of a gray value analysis.

[0045] In step S3, the structural simulation is performed based on the 3D CT dataset of the component. This is done, for example, using finite element methods. In this process, forces acting on the component, particularly in the form of load cycles, are simulated, and the resulting stresses and deformations in the component are calculated.

[0046] Step 3 can also include importing a previously created CAD structural simulation. It is crucial that the entire displacement / stress field be imported in its entirety, not just the displacement / stress values ​​at individual points, as is usually the case with a standard export from an FEM program. The importing program should therefore provide at least partial FEM functionality, specifically the ability to interpret displacement / stress fields defined on the nodes of an FEM mesh. Current state-of-the-art defect analysis software has not been able to achieve this.

[0047] In step S4, the damage parameters are calculated to characterize the damaging effect of defects, based on a combination of information from step S2 (indication of homogeneities (defects) in the 3D CT dataset) and from step S3 (structural simulation based on the 3D CT dataset). The datasets from steps 2 and 3 are typically automatically aligned with each other, as the structural simulation is directly based on the geometric information from the 3D CT dataset. Separate alignment of datasets is necessary when datasets from different sources are combined. The combination of datasets can be performed in various ways, particularly based on both theoretically and heuristically determined information.

[0048] In a fifth step S5, the ensemble of damage parameters assigned to the defects is converted into a component parameter to characterize the damage potential and to predict / estimate the load-bearing capacity or stress limits and thus the performance / strength / service life for the individual component.

[0049] The following section describes a more detailed example of determining key performance indicators for characterizing the service life of a component. First, inhomogeneities (defects) are identified using a grayscale analysis of an additively manufactured aluminum component. In a further step, the stress far field per defect is calculated as the mean value of the stress tensor in the vicinity of the defect. The vicinity is defined as a defect dilated by the effective defect radius.

[0050] From the first principal component of the stress and the defect cross-sectional area, projected into the first principal direction, a stress intensity factor (SIF) according to Murakami is calculated. Here, internal defects are taken into account with a geometry factor of 0.50, and surface defects with a geometry factor of 0.65.

[0051] Crack growth simulation is performed using linear elastic fracture mechanics (LEFM) with the Paris line, where the user specifies material parameters C and m. The initial crack area is calculated from the projected pore area, which is derived from defect detection and the first principal stress direction. The component is considered destroyed when the equivalent crack radius reaches a predefined limit. In the defects investigated in this example, the crack initiation phases are very short and not significant. In addition to the calculated crack initiation phases, the notch stresses and the Smith-Watson-Topper and Rainbow crack initiation lifetimes can be calculated for each defect from the 3D CT simulation according to the invention.

Claims

[1] Method for computer-aided determination of the damage effect of defects in a component, comprising the steps of indication of a defect in a volumetric data set of the component, calculation of force-induced local deformations and / or stresses on a 3D data set of the component by means of structural simulation, calculation of a damage parameter to characterize the damage effect of the defect based on a combination of the information from the structural simulation and the indicated defect and / or other defects in the vicinity of the indicated defect that influence the characterization of the damage effect. [2] Method according to claim 1, wherein the 3D data set comprises a CAD data set and / or a CT data set. [3] Method according to any of the preceding claims, wherein the volumetric data set comprises a measurement data set and / or a computed tomography data set and / or a series of micrographs. [4] Method according to any of the preceding claims, wherein the defect is indexed by means of a grey value analysis and / or a self-learning algorithm, wherein the self-learning algorithm is configured to characterize the shape of the defect. [5] Method according to one of the preceding claims, wherein the damage index is determined for an environment adapted to the defect. [6] Method according to any of the preceding claims, further comprising the calculation of a crack initiation site on the component and / or an associated crack initiation lifetime and / or an associated crack propagation lifetime and / or a damage tolerance based on the damage characteristic of a defect and / or the damage characteristic values ​​in a neighborhood of a defect. [7] Method according to one of the preceding claims, wherein, when calculating the damage parameters, defects that do not have a heuristic effect, such as a small size and / or a small distance to the edge and / or a low ambient stress, are not taken into account. [8] Method according to any of the preceding claims, further comprising the calculation of a crack growth-relevant parameter such as a defect area perpendicular to the loading direction and / or a stress and / or a stress intensity factor and / or a J-integral and / or an energy release rate and / or a defect / surface distance based on the damage parameter of a defect and / or the damage parameters in a neighborhood of a defect. [9] Method according to the preceding claim, wherein the calculation of the crack propagation lifetime is based on the determination of a stress intensity factor and / or a J-integral and / or mean stress effects and / or hydrostatic stress effects and / or multiaxiality / triaxiality effects and / or crack closing and opening effects and / or an anisotropy. [10] Method according to any of the preceding claims, further comprising the calculation of a defect / damage tolerance based on coagulation of a defect and / or crack stopping effects. [11] Method according to any of the preceding claims, further comprising the calculation of a component characteristic number from the damage characteristics of several defects. [12] Method according to the preceding claim, wherein the component identifier is based on the use of models according to Murakami and / or Shiozawa and / or Paris and / or damage-relevant material identifiers according to Smith-Watson-Topper and / or damage accumulation. [13] Method according to the two preceding claims, wherein the component identifier indicates a location of a first crack and / or a weak point of the component and / or the crack propagation lifetime and / or a load-bearing capacity under static and / or dynamic / alternating loads. [14] System comprising a device adapted to perform the method according to claim 1. [15] Computer program on a data carrier locally in or on the computer system or externally on servers (LAN / WLAN) or in a data cloud for carrying out the method according to claim 1.

Images