Method for generating a test profile for vibration testing of vehicle equipment based on data acquisition during route driving, system for performing the method, computer program, and computer readable medium

Abstract

The present invention relates to a method for generating a test profile for a vibration test of vehicle equipment based on data acquired during route travel, and includes a pseudo-damage spectrum calculation process (P1) for calculating a plurality of pseudo-damage spectra (2.1, 2.2, ..., 2.n) from a plurality of measurement signals (1.1, 1.2, ..., 1.n) using a spectrum calculation algorithm to generate a pseudo-damage spectrum for each measurement signal, where each measurement signal was recorded during route travel of a test vehicle before the start of the method, an extrapolation sub-process (P5), and a superposition sub-process (P6), in which, in the extrapolation sub-process (P5), each pseudo-damage spectrum is multiplied by a proportionality constant to generate a plurality of extrapolated pseudo-damage spectra (7.1, 7.2, ..., 7.n), and in the superposition sub-process (P6), the extrapolated pseudo-damage spectra are added together to generate a superposed pseudo-damage spectrum (3).

Description

[Technical Field]

[0001] The present invention relates to a method for generating a test profile according to the preamble of claim 1, a system for performing such a method according to claim 12, a computer program according to claim 1, and a computer readable medium according to claim 13. The present invention further relates to a system, a computer program, and a computer readable medium according to the independent claims. [Background technology]

[0002] Within the approval process of newly developed technical products in the automotive industry, the vibration resistance of such products is typically tested in particular. The products are typically any type of component, assembly, or device that is installed in vehicles and that is subjected to vibrations during the operation of these vehicles. The vibrations can be vibrations generated by the operation of the vehicle, or vibrations generated by the product itself (e.g., in an electric motor), or a combination of different types of vibrations.

[0003] In testing such products for their vibration resistance, the products are typically subjected to a specific vibration pattern on a special test stand (e.g. a so-called "shaker test stand") that simulates as closely as possible the vibrations that occur during operation of the vehicle in which the product is used. Such a vibration pattern is defined by a test profile.

[0004] Ideally, such a test profile should, on the one hand, reflect as closely as possible the vibration stresses during the service life of the respective product, but, on the other hand, it should also make it possible to keep the duration of the vibration test on the test stand as short as possible, e.g., in order to minimize the costs for the vibration test.

[0005] DE10 2020 114 973 A1 relates to a method for determining a test profile for a test or simulation on a component or vehicle to be tested. Data is measured or calculated as a function of time for different applications and stored. This data is then analyzed for damage content. The identified damage content for a specific application is made available to a user. The user can then select and combine specific time periods from the various use cases to create an individual test profile.

[0006] In the past, methods have already been proposed for generating test profiles that are determined based on data recorded during test runs, for example.

[0007] The drawbacks of known profiling methods are briefly described below: For better clarity, the methods have been grouped according to the features "model-based / non-model-based" and "damage-based / non-damage-based".

[0008] Model-based methods are based on describing the dynamic properties of the analyzed components by means of a physical system model, for example in the form of a FE model [10, 11] (references to the literature are given in square brackets at the end of the specification), or mathematically, by a set of differential equations [5-9]. A profile calculation algorithm is classified as a damage-based method if it is defined in such a way that the profiles created with it must satisfy the (computational) equivalence of the damage numbers from operational vibration measurements (e.g. on a vehicle on a route) with the damage numbers from a shaker test (the so-called damage equivalence principle).

[0009] The biggest drawbacks of this method are: Profiles calculated using non-damage-based methods (whether model-based or not) do not guarantee that the damage a product experiences during its life cycle in a vehicle (damage during operation) will be reproduced in the vibration test (damage during the test). Therefore, even if the test passes, it does not exclude that the damage inflicted on the product during the vibration test was too small compared to the actual damage in the vehicle (i.e., the test was too weak). On the other hand, if the test does not pass, there is a possibility that the damage during the test exceeds the damage during operation (a test that is too severe). In either case, this means that the test was not performed correctly. This drawback is shared by all non-damage-based methods, e.g., [2, 3, 4]. Model-based methods use a model of a linear, underdamped one-mass oscillator (Einmassenschwinger: EMS) to describe the vibration behavior of the test object. This is the case in methods [5-9]. However, the complex, nonlinear mechanical-dynamic behavior of real products (components) can only be represented inaccurately by relatively simple EMS models. Other methods [10, 11] use computer models created with the help of finite element tools for this purpose. However, setting up an FE model requires a lot of effort, which in turn leads to high costs for profile calculations.

[0010] Other drawbacks of profiling methods known in the art for vibration testing of vehicle equipment are: They are not universally applicable because they are either: a specific type of profile (e.g., sweep profiles only [2, 3], or noise profiles only [4, 9, 12, 13]), or A profile for vibrational excitation only (hereafter referred to as an excitation profile) and not for the vibrational response (e.g., a response profile that allows limiting the response amplitude of the test object when creating the excitation profile) [4-9], or A profile that allows damage equivalence testing, and that can be calculated only for certain combinations of calculation parameters (for example, only for Wöhler line gradient coefficient 4)

[13] , or They require data recorded on a vibration test stand, possibly in a pre-test

[13] , to trigger the profile calculation algorithm. In accordance with the present invention, it is recognized that such data can be generated in a simulation without the use of hardware (e.g., a test stand), thereby minimizing effort and costs, or They only work recursively [8, 9, 13], which is inefficient and requires repeatedly calculating the profile amplitude for each individual frequency in a loop, resulting in high computational costs. They do not take into account the need for extrapolation and superposition of load spectra (Lastkollektive) or damage numbers calculated from data recorded in road tests so that they are valid for all required service lives of components in the vehicle

[13] . Summary of the Invention [Problem to be solved by the invention]

[0011] SUMMARY OF THE INVENTION It is an object of the present invention to eliminate or at least reduce the disadvantages of the prior art. [Means for solving the problem]

[0012] This problem is solved by a method for generating a test profile as claimed in claim 1. The inventors have realised that this problem can be particularly well solved by means of a new, damage-based, non-model-based method for generating a test profile as claimed in claim 1.

[0013] The term "test profile" should be understood broadly. In particular, a test profile may include, for example, the following functional dependencies: The course of the amplitude of the acceleration signal (or another physical quantity) over frequency (especially in the case of a sweep profile), The progression of the power density spectrum (LDS) of the acceleration signal (or other physical quantity) over frequency (especially in the case of noise profiles).

[0014] The inventors have found that this type of method for generating test profiles can be more versatile and efficient than known methods. This method according to the invention, hereinafter referred to as the ASPEN method (short for "Automatisierte SchwingungsProfil‐Entwicklung"), or the ASPEN RoMi method, is used to generate test profiles from test runs on routes, which are understood to be special driving routes defined for various application profiles of a vehicle (e.g., highways, national roads, mountain passes, urban areas, etc.), on which realistic conditions regarding the vibration loads of the components under consideration are simulated. Routes can be combined in any proportion, which is where the name "RoMi" - "RouteMix" comes from. A known representative example of a route mix in the passenger car segment is CARLOS [1, 15]. Special tests must be distinguished from route runs. They are understood as tests of relatively short duration involving a uniform, continuous increase ("run-up") or decrease ("run-down") of one or more driving condition parameters. In most cases, this parameter is the rotational speed of the vehicle's drivetrain, e.g., the internal combustion engine or the electric drive (in electric vehicles). The dwell times of the driving condition parameters in these tests usually do not correspond to typical operational use. Therefore, they are not suitable for creating test profiles with the ASPEN RoMi method.

[0015] The data required for profiling are preferably recorded in a test vehicle during test runs on one or more routes (route mix), as described above. Alternatively, suitable functional and / or load test stands may also be used for this purpose, provided they allow for the simulation of the vibrations experienced by the product as they would be experienced in a vehicle on a route. In such tests, the mechanical-dynamic load and / or stress variables accompanying the vibration process (e.g. vibration acceleration, vibration velocity, dynamic vibration displacement, dynamic force, mechanical strain, etc.) are measured on the product and / or at its mounting points on the carrier using suitable sensors (e.g. acceleration sensors, velocity sensors, displacement sensors, force sensors, strain gauges, etc.) and using a suitable vibration recording system.

[0016] After the completion of a measurement campaign carried out in this way, the input data required for profiling with the ASPEN method are available as digitized vibration signals, which, after the usual signal pre-processing steps known to those skilled in the art (e.g. removal of measurement disturbances, cutting out relevant measurement intervals, filtering), can be directly fed to a computing unit or other system in which the ASPEN method is programmed for profile calculation.

[0017] To calculate the test profile, in the ASPEN method, in addition to the signals recorded on the route, advantageously a special time signal, called the reference signal, is processed.

[0018] In an advantageous embodiment, all test profiles calculated in the method computationally satisfy the principle of damage equivalence for the pseudo damage spectra calculated in the method.

[0019] In an advantageous embodiment, the reference signal processing process runs in parallel with the pseudo-damage spectrum calculation process and / or the extrapolation and convolution processes. The advantage of this is that both the convoluted pseudo-damage spectrum and the extrapolated reference signal pseudo-damage spectrum are substantially simultaneously available for the test profile generation process. In a typical embodiment, the reference signal pseudo-damage spectrum is calculated within the pseudo-damage spectrum calculation process. In a typical embodiment, the extrapolated reference signal pseudo-damage spectrum is calculated within the extrapolation and convolution processes, particularly within the extrapolation sub-process. However, it is alternatively possible to calculate the reference signal differently, for example, for other time points.

[0020] In this case, each measurement signal preferably depicts the time course of one and the same measurement value typically recorded on different routes of the route mix. This should be understood in particular to mean that a measurement value, e.g., acceleration, is defined at a specific position of a vehicle component in a specific measurement direction. This measurement value is then considered within the framework of multiple route runs of the test vehicle, and for each route run, the measurement signal for this measurement value is always recorded in the same direction. Thus, at the end of the multiple route runs, there are multiple measurement signals (in other words, sets of measurement signals) for one and the same measurement value - one measurement signal per route. The measurement signals have slightly different courses, since each route run causes different vibrations.

[0021] In a first, exemplary embodiment of the method, a set of measurement signals with only one measurement value is processed. The method then provides (as a primary result) only one test profile, with which the same stresses are applied to the component (at its selected measurement locations) in a vibration test of a defined duration as would be expected to occur on the root mix for all required operating periods (mileage). By predicted, we mean that the operating damage is computationally predicted by extrapolation and superposition processes. In this sense, the method provides a damage equivalent profile. The principle of damage equivalent underlies the definition of the method and is its most important feature.

[0022] In a second exemplary embodiment, the method generates multiple test profiles based on multiple measurement signal sets. Each measurement signal set is obtained by measuring signals of a specific measurement value along a route mix. In other words, the method considers multiple measurement values ​​instead of just one measurement value. For example, it is conceivable that acceleration for a specific vehicle component should be measured in the same direction at different points on the component. Each acceleration in one direction at each measurement position should be considered as a separate measurement value. Therefore, multiple measurement signal sets are generated for this component. Each measurement signal set corresponds to a specific measurement value. Therefore, instead of multiple measurement signals for one and the same measurement value, different measurement signal sets are present. Each measurement signal set consists of measurement signals whose number corresponds to the number of route runs performed by the test vehicle. (In all embodiments of the method, signals of the same measurement value are preferably recorded for each route.) In the embodiment with multiple measurement signal sets, instead of measurement signals for only a single measurement value, different measurement signal sets for multiple measurements are processed within the above process. The number of measurement signal sets corresponds to the number of measurements under consideration. In other words, the simplest case of the method according to the invention (see the first embodiment) corresponds to the case where only one measurement variable is considered. However, in the latter second exemplary embodiment, the method considers two, three, four or more measurement values, each of which is incorporated into the calculations of the aforementioned process. In this case, the method provides, after extrapolation and superposition processes, several profiles, each of which is damage-equivalent in the above sense for the position and orientation under consideration on the component (i.e. for each measurement value).

[0023] Furthermore, the methods are usually based on the calculation of so-called pseudo-damage spectra.

[0024] In a typical embodiment, within the framework of the method (depending on the problem, input data and setting parameters) several different types of test profiles or profiles with different characteristics are generated, in particular excitation profiles and response profiles, profiles for single-point and multi-point control, and profiles covering damage on individual routes or on the entire route mix.

[0025] In an advantageous embodiment, the pseudo damage spectrum calculation process comprises the following steps: a signal filtering step, within which each measurement signal is filtered with a plurality of bandpass filters, resulting in a plurality of filtered measurement signals; a classification step in the framework of which a load spectrum is formed from each filtered measurement signal by means of a classification, typically by dividing the entire amplitude range of each filtered measurement signal into classes, and in which the number of vibration cycles is determined for each class of amplitude, preferably by counting; a transformation step in which the possible mean amplitudes of each vibration cycle (which may have a mean value not equal to zero) are first transformed into damage-equivalent mean-free amplitudes, preferably using a Haigh diagram, and then the mean-free amplitudes are sorted in ascending order; a partial damage contribution calculation step in which the partial damage contribution is calculated for each damage equivalent number of vibration cycles at a mean-free amplitude, preferably using Wöhler lines; a total damage calculation step, within which for each filtered measurement signal the partial damage contributions are summed to a total damage, whereby each total damage is referred to as a pseudo-damage number for the respective filtered measurement signal; a pseudo damage spectrum formation step in which a pseudo damage spectrum is formed from the pseudo damage numbers by displaying the pseudo damage numbers as a function of the bandpass center frequency of the bandpass within the pseudo damage spectrum formation step;

[0026] The filtered measurement signals are preferably narrow-band filtered measurement signals. In a typical embodiment, the classification step comprises a rainflow counting step, within which a rainflow matrix is ​​generated for each filtered measurement signal with the aid of rainflow parameters. In the following, for the sake of simplicity, the term "pseudo damage spectrum" will be abbreviated in some places to "PSS".

[0027] In an advantageous embodiment, the noise profile and / or the sweep profile is calculated within the framework of the test profile generation process.

[0028] In an exemplary embodiment, the sweep profile is calculated according to the following formula:

number

[0029] In an exemplary embodiment, the noise profile is calculated according to the following formula:

number

[0030] In typical embodiments, the method is a computer-implemented method. In typical embodiments, the process runs at least partially automated.

[0031] The problem is further solved by a system for performing one of the aforementioned methods, the system being preferably suitable for at least partly performing and / or regulating and / or controlling a method for generating a test profile according to at least one of the aforementioned embodiments.

[0032] For this purpose, the system advantageously comprises suitable components, such as a pseudo-damage spectrum calculation component, and / or a spectrum calculation component, and / or an extrapolation and convolution component, and / or an extrapolation sub-component, and / or a convolution sub-component, and / or a test profile generation component, and / or a reference signal processing component, and / or a signal filter component, and / or a classification component, and / or a transformation component, and / or a partial damage contribution calculation component, and / or a total damage calculation component, and / or a pseudo-damage spectrum formation component, and / or a noise profile calculation component, and / or a sweep profile calculation component.

[0033] Advantageously, in the system, at least some of the aforementioned components are implemented using computer program code.

[0034] The problem is further solved by a computer program, which, when executed on a computer, causes the computer to carry out a method for generating a test profile according to at least one of the preceding embodiments.

[0035] In one embodiment of the present invention, the computer-readable medium comprises computer program code for performing one of the aforementioned methods. The term "computer-readable medium" includes, but is not limited to, hard disks and / or servers and / or memory sticks and / or flash memories and / or DVDs and / or Blu-rays and / or CDs. In addition, the term "computer-readable medium" can also be understood as a data stream, such as occurs when a computer program product is downloaded from the Internet. [Brief explanation of the drawings]

[0036] [Figure 1]1 is a schematic block diagram of a method according to the invention in a first embodiment; FIG. [Figure 2] FIG. 3 shows a schematic representation of the method according to the invention in a second embodiment as a block diagram. [Figure 3] FIG. 2 is a schematic block diagram of a pseudo damage spectrum calculation process typically used in the method according to the invention; [Figure 4] FIG. 3 is a schematic diagram of the method according to the invention in a third embodiment. [Figure 5] FIG. 2 is a schematic diagram of a pseudo damage spectrum calculation process typically used in a method according to the present invention; [Figure 6] FIG. 1 is a schematic diagram of the extrapolation and superposition process for multiple measurements. [Figure 7] FIG. 1 is a schematic diagram of a test profile generation process for multiple measurements. [Figure 8] FIG. 1 is an exemplary diagram of a load spectrum and a Wöhler line. [Figure 9] FIG. 5 is a schematic block diagram of the method according to the invention in a fourth embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0037] DESCRIPTION OF THE PREFERRED EMBODIMENT FIG. 1 shows a block diagram of a first embodiment of the method according to the present invention. In particular, FIG. 1 shows a pseudo-damage spectrum calculation process P1, an extrapolation and superposition process P2, a test profile generation process P3, and a reference signal processing process P4. The extrapolation and superposition process P2 includes an extrapolation subprocess P5 and a superposition subprocess P6. A plurality of measurement signals 1.1, 1.2, ..., 1.n are fed to the pseudo-damage spectrum calculation process. These measurement signals 1.1, 1.2, ..., 1.n are typically recorded during a test vehicle route (not shown) and each represent the time course of a specific vibration measurement value on the test vehicle. Each of the N measurement signals 1.1, 1.2, ..., 1.n is measured along a specific route of the test vehicle. As a result, there is one measurement signal per route for profile calculation. In the pseudo-damage spectrum calculation process P1, multiple pseudo-damage spectra 2.1, 2.2, ..., 2.n are calculated from these measurement signals 1.1, 1.2, ..., 1.n using a spectrum calculation algorithm. Thus, a pseudo-damage spectrum 2.1, 2.2, ..., 2.n is generated for each measurement signal 1.1, 1.2, ..., 1.n. ​​Details of the pseudo-damage spectrum calculation process P1 will be explained in more detail below. The pseudo-damage spectra 2.1, 2.2, ..., 2.n are then fed to an extrapolation and superposition process P2. In this process, each pseudo-damage spectrum 2.1, 2.2, ..., 2.n is first multiplied by a proportionality constant. As a result, multiple extrapolated pseudo-damage spectra are generated. For clarity, the extrapolated pseudo-damage spectra are not explicitly shown in FIG. 1. The extrapolated pseudo-damage spectra are then added in a superposition subprocess P6. As a result, a superposed pseudo-damage spectrum 3 is generated. In other words, from a number of measurement signals 1.1, 1.2, ..., 1.n recorded over different route runs of the test vehicle, a single superimposed pseudo damage spectrum 3 is formed. This superimposed pseudo damage spectrum 3 is then fed into a test profile generation process P3.In parallel with the pseudo-damage spectrum calculation process P1 and the extrapolation and superposition process P2, a reference signal processing process P4 also operates in the manner shown in FIG. 1 . In this reference signal processing process P4, a reference signal pseudo-damage spectrum is first calculated from the reference signal 5 using the spectrum calculation algorithm also applied in the pseudo-damage spectrum calculation process P1. For clarity, this reference signal pseudo-damage spectrum is not explicitly shown in FIG. 1 . The reference signal pseudo-damage spectrum is then multiplied by the proportionality constant already applied in the extrapolation subprocess P5. As a result, an extrapolated reference signal pseudo-damage spectrum 6 is generated. This extrapolated reference signal pseudo-damage spectrum 6 is also supplied to the test profile generation process P3. In the test profile generation process P3, a test profile 4 is then generated based on the superposed pseudo-damage spectrum 3 and the extrapolated reference signal pseudo-damage spectrum. The test profile 4 can then be supplied to a test object on a test stand. The test object on the test stand thereby experiences stresses corresponding to the stresses from all previous route runs under the influence of the test profile 4 calculated in this way.

[0038] The generation of the extrapolated reference signal pseudo-damage spectrum 6 is shown in FIG. 1 as being generated in a separate reference signal processing process P4. However, other variations of the generation of the extrapolated reference signal pseudo-damage spectrum 6 are also possible. For example, the reference signal can be generated directly within the pseudo-damage spectrum calculation process P1 and the extrapolation and superposition process P2. In other words, for example, the reference signal processing process P4 can be performed partly by the pseudo-damage spectrum calculation process P1 and partly by the extrapolation and superposition process P2.

[0039] FIG. 2 shows a block diagram of a schematic diagram of a method according to the present invention in a second embodiment. The method of FIG. 2 is very similar to the method of FIG. 1. However, unlike the method of FIG. 1, in the method of FIG. 2, in addition to the superimposed pseudo-damage spectrum 3, the extrapolation and superimposition process P2 outputs a number of extrapolated pseudo-damage spectra 7.1, 7.2, ..., 7n. These extrapolated pseudo-damage spectra 7.1, 7.2, ..., 7n are then also fed to the test profile generation process P3. The test profile generation process P3 can process them and output, at the output side of the test profile generation process P3, a number of test profiles 8.1, 8.2, ..., 8.n for individual routes in addition to the test profile 4 covering the damage on the complete route mix as described above. These test profiles 8.1, 8.2, ..., 8.n for individual routes are available as side results of the method in addition to the test profile 4 for the route mix (the main result of the method) and can also be used when testing the same product on a test bench. The particularity of test profiles 8.1, 8.2, ..., 8n is that they cover damage on each individual route (for the full driving period on this route within the framework of the route mix), but not on the route mix.

[0040] FIG. 3 shows a block diagram of a pseudo-damage spectrum calculation process P1, typically used in the method according to the present invention. It can be seen that the pseudo-damage spectrum calculation process in FIG. 3 includes a signal filtering step S1, a classification step S2, a conversion step S3, a partial damage contribution calculation step S4, a total damage calculation step S5, and a pseudo-damage spectrum formation step S6. In the signal filtering step S1, each measurement signal is filtered using multiple bandpass filters to generate multiple filtered measurement signals. In the classification step S2, a weight spectrum is formed from each filtered measurement signal using a classification. This is typically done by dividing the entire amplitude range of each filtered measurement signal into classes. For each class, the number of vibration cycles is preferably determined. Subsequently, in the conversion step S3, the amplitudes of each vibration cycle, which may have mean values ​​(vibration cycles may have mean values ​​that are not equal to zero), are first converted into damage-equivalent, mean-free amplitudes. This conversion step is optional in certain embodiments. The conversion in the conversion step S3 is preferably performed using a Haigh diagram. After generating the damage-equivalent mean-free amplitudes, these mean-free amplitudes are sorted in ascending order. In a partial damage contribution calculation step S4, a partial damage contribution is calculated for each damage-equivalent mean-free amplitude, preferably using a Wöhler line. In a total damage calculation step S5, the partial damage contributions are summed to obtain a total damage for each filtered measurement signal. Each total damage is then referred to as a pseudo-damage number for the respective filtered measurement signal. Finally, in a pseudo-damage spectrum formation step S6, a pseudo-damage spectrum is formed from the pseudo-damage numbers. This is typically done by displaying the pseudo-damage numbers as a function of the bandpass center frequency of the bandpass.

[0041] FIG. 4 shows a schematic diagram of the method according to the present invention in a third embodiment. In particular, FIG. 4 illustrates how the different computer-implemented components of the method according to the present invention are interconnected and how the profile calculation works in detail. In FIG. 4, a pseudo-damage spectrum calculation process P1, an extrapolation and superposition process P2, and a test profile generation process P3 are shown. Additionally, FIG. 4 illustrates a pseudo-damage spectrum calculation component 13, an extrapolation subcomponent 14, a superposition subcomponent 15, and a test profile generation component 16. In FIG. 4, these components 13, 14, and 16 are partially shown more than once. Therefore, the corresponding instances of each component are discussed below. In each instance of the same component 13, 14, and 16, the same calculation algorithm (FDDC, SPEX, or PRGN algorithm—described below) is implemented. The components 13, 14, and 16 are active at different points in the method according to the present invention, in particular when processing different signals and / or when further processing intermediate characteristic values ​​calculated in one of the previous steps. For the method of FIG. 4, there are n measurement data files 9.1, 9.2, ..., 9.n at the input side. These measurement data files 9.1, 9.2, ..., 9.n are typically generated by recording signals only on small, as representative as possible, sections of each of the n routes, and each contains measurement signals of the same measurement values ​​(not explicitly shown in FIG. 4). Each of the measurement data files 9.1, 9.2, ..., 9.n is fed to an instance of the pseudo-damage spectrum calculation component 13. The pseudo-damage spectrum calculation component 13 outputs a pseudo-damage spectrum at the output side. In FIG. 4, only the first pseudo-damage spectrum 2.1 is labeled to avoid overloading the diagram. The pseudo-damage spectra are then fed to the extrapolation and superposition process P2, where they are first processed by different instances of the extrapolation subcomponent 14 (one instance per pseudo-damage spectrum). Thus, there are a total of n instances.At the output side, multiple extrapolated pseudo-damage spectra 7.1, 7.2, ..., 7.n are available in the extrapolation subcomponent 14. Again, for clarity, only the first two extrapolated pseudo-damage spectra 7.1, 7.2 are labeled. The extrapolated pseudo-damage spectra 7.1, 7.2, ..., 7.n are supplied to the superposition subcomponent 15 on the one hand, and directly to n instances of the test profile generation component 16 on the other hand. The superposition subcomponent 15 superimposes the extrapolated pseudo-damage spectra 7.1, 7.2 (and all other available extrapolated pseudo-damage spectra not explicitly labeled, i.e., n spectra in total), so that the superimposed (and extrapolated) pseudo-damage spectrum 3 is available at the output side. This superimposed pseudo-damage spectrum 3 is also supplied to the test profile generation component 16. Therefore, it consists of n+1 instances of the test profile generation component 16: n instances for the extrapolated pseudo-damage spectra 7.1, 7.2, ..., 7.n, and one instance for the superimposed pseudo-damage spectrum 3. In parallel with processes P1, P2, and P3, an extrapolated reference signal pseudo-damage spectrum 6 is also generated in FIG. 4 . This is also provided to all n+1 instances of the test profile generation component 16. This extrapolated reference signal pseudo-damage spectrum 6 is generated by first processing the (separately generated) reference signal 5 in an instance of the pseudo-damage spectrum calculation component 13. As a result, a reference signal pseudo-damage spectrum 17 is generated. This reference signal pseudo-damage spectrum 17 is then provided to an instance of the extrapolation subcomponent 14. The extrapolation subcomponent 14 generates the extrapolated reference signal pseudo-damage spectrum 6. In FIG. 4, pseudo damage calculation parameters 10, extrapolation and convolution calculation parameters 11, test profile calculation parameters 12, and reference signal extrapolation parameters 18 are shown.The pseudo-damage calculation parameters 10 typically include one or more definitions for bandpass filters (e.g., their corner frequencies, filter order, coverage of individual filters, etc.), one or more parameters for the rainflow method, and / or a Wöhler line. The pseudo-damage calculation parameters 10 are made available to the pseudo-damage spectrum calculation component 13. The extrapolation and superposition calculation parameters 11 typically include information regarding the total driving duration on each route forming the route mix, preferably in hours. The extrapolation and superposition calculation parameters 11 are passed to the extrapolation and superposition process P2. The reference signal extrapolation parameters 18 typically include an indication of the intended duration of the vibration test of the product, preferably in hours. These are passed to the reference signal extrapolation subcomponent 14. The test profile calculation parameters 12 typically include a Wöhler line slope coefficient and / or a safety factor and / or a test factor and / or a safety factor and test factor. The test profile calculation parameters 12 are passed to a test profile generation component 16, in particular within the framework of a test profile generation process P3.

[0042] Figure 5 shows a schematic diagram of a pseudo-damage spectrum calculation process typically used in a method according to the present invention. In particular, Figure 5 shows how multiple damage numbers are determined from the measurement signal 1.1 that forms the pseudo-damage spectrum. This is illustrated in Figure 1 by the sequence of steps: signal filtering step S1, classification step S2, transformation step S3, partial damage contribution calculation step S4, total damage calculation step S5, and pseudo-damage spectrum formation step S6. Further details of Figure 5 are provided below.

[0043] Figure 6 shows a schematic diagram of the extrapolation and superposition process P2 for multiple measurements, details of which are described below.

[0044] Figure 7 shows a schematic diagram of the test profile generation process P3 for multiple measurements. Further details of Figure 7 are provided below.

[0045] An exemplary representation of the load spectrum and Wöhler line is shown in Figure 8. Details of Figure 8 are explained below.

[0046] FIG. 9 shows a block diagram of the method according to the present invention in its fourth, most complex embodiment. Unlike the methods shown in FIGS. 1, 2, and 4, the method shown in FIG. 9 processes multiple measurement signals measured on multiple routes, where the signals belong to different measurement points. In this embodiment, it is assumed that measurements were performed on a test vehicle (not shown) on a total of m driving routes, with n signals recorded on each route. This results in a total of m x n measurement signals MS1,1, MS1,2, ..., MS1,n, MS2,1, MS2,2, ..., MS2,n, ..., MSm,1, MSm,2, ..., MSm,n being supplied to the method according to the present invention. The signals measured (time-synchronously) on the routes were stored in the same (digitized) measurement data file. For example, in measurement data file 2 recorded during a drive on route 2, there are n measurement signals MS2,1, MS2,2, ..., MS2,n. Therefore, the number of measurement data files m is equal to the number of driving routes. Furthermore, in this embodiment of the method, it is assumed that signals with the same last index are generated by recording the same measurements on different routes, for example signals MS1,1, MS2,1, ..., MSm,1 by recording measurement 1, signals MS1,2, MS2,2, ..., MSm,2 by recording measurement 2, etc. This leads to the following particularities, which differ from the method embodiments in Figures 1, 2 and 4: In the pseudo damage spectrum calculation process P1, mxn pseudo damage spectra PS1,1, PS1,2, ..., PS1,n, PS2,1, PS2,2, ..., PS2,n, ..., PSm,1, PSm,2, ..., PSm,n are calculated from the measurement signals instead of n. In the extrapolation sub-process P5 of the extrapolation and superposition process P2, from these pseudo damage spectra, instead of n, likewise m x n extrapolated pseudo damage spectra ES1,1, ES1,2, ..., ES1,n, ES2,1, ES2,2, ..., ES2,n, ..., ESm,1, ESm,2, ..., ESm,n are calculated, each of which describes the damage for the measurement position and measurement direction in the component (product) under consideration on each individual route (but not on the entire route mix). In the test profile generation process P3, from these mxn extrapolated pseudo damage spectra, test profiles PP1,1, PP1,2, ..., PP1,n, PP2,1, PP2,2, ..., PP2,n, ..., PPm,1, PPm,2, ..., PPm,n are calculated, each of which covers damage for a measurement position and measurement direction on the component under consideration on each individual route (but not on the entire route mix) and is suitable for carrying out a vibration test of the component using profile control, introducing vibration at this position and in this direction. In such a test, an equivalent damage corresponding to the corresponding route is applied to the corresponding position and in the corresponding direction of the component. In the superposition sub-process P6 of the extrapolation and superposition process P2, instead of one, several superposed (and extrapolated) pseudo damage spectra SS1, SS2, ..., SSn are calculated, i.e., pseudo damage spectra for each measurement. Each such spectrum describes the damage of the component under consideration on the entire route mix for the corresponding measurement location (i.e., sensor placement location and orientation). Therefore, in the next step of the test profile generation process P3, n test profiles SPP1, SPP2, ..., SPPn are calculated, among others, from these n pseudo damage spectra SS1, SS2, ..., SSn, all of which cover the damage of the component over the entire route mix, i.e., each profile for its measurement position and measurement direction on the component, and are suitable for carrying out a vibration test of the component under consideration. This vibration test involves profile control at the corresponding position (sensor placement location) and introducing vibration according to the profile in the corresponding direction. In such a test, this is applied at a specific position on the component in the corresponding direction for this damage equivalent to the entire route mix.

[0047] The reference signal processing process P4 in this embodiment of the method remains the same as in the embodiment previously described in FIGS.

[0048] In the following, some embodiments of the present invention will be described in greater detail, in particular with the aid of Figures 4 to 9. Numbers in square brackets should be understood as references to documents at the end of the specification, as mentioned above. Numbers or lowercase letters in parentheses should be understood as references to the corresponding mathematical formulas indicated below by these numbers or lowercase letters in parentheses.

[0049] In a particular embodiment, the method according to the invention (as mentioned above, this particular method according to the invention is also called the ASPEN method or the ASPEN RoMi method) is based on the calculation and equivalence of damage numbers and therefore belongs to the group of damage-based methods. These calculations are carried out using the following procedure, which is known in operational stability: Formation of a load spectrum from the time values ​​under consideration (using classification or counting methods), and Conversion to damage numbers using the Wöhler material fatigue hypothesis (mathematically described by the Wöhler line) and the linear damage accumulation hypothesis (Palmgren-Miner law).

[0050] It should be noted here that the Wöhler lines used to calculate damage numbers are almost always based on assumptions and are not necessarily applicable to the measurements and / or stress states under consideration. That is, they are fictitious. Therefore, damage numbers calculated in this manner have no meaning with respect to the absolute time of failure or remaining life of the component under consideration. For this reason, they are referred to herein as pseudo-damage numbers (PSZ). However, PSZ can be meaningfully used for comparative damage-based calculations and analyses, such as the ASPEN method.

[0051] Based on the PSZ, a new fundamental parameter - the pseudo-damage spectrum (PSS) - is introduced in the ASPEN method, which describes the dependence of the PSZ of an oscillating time signal on the oscillation frequency. This is detailed in the subsection "Pseudo-Damage Spectrum" below.

[0052] The basic concept of the ASPEN RoMi method is as follows: The test profile generated for vibration testing of a component should computationally guarantee the equivalence of the two following PSSs at each vibration frequency f: ·PSS occurring after extrapolation and possible superposition of the required component life in the route mix traveled. The PSS that a component experiences in a vibration test on the first profile over time for a specified test period.

[0053] In this sense, the ASPEN RoMi method thus provides a damage-equivalent test profile.

[0054] Damage equivalence is applied individually to each location, for example on the component or its holder, where a signal is considered to be measured for profiling.

[0055] The ASPEN RoMi method consists of three computational modules, as shown in Figure 4: The FDDC (Frequency Dependent Damage Calculation) module, also known as the Pseudo Damage Spectrum Calculation Component 13, a SPEX (SuperPosition and EXtrapolation) module, also called the Extrapolation and Superposition component, including an Extrapolation subcomponent 14 and a Superposition subcomponent 15; and · PRGN (PProfile Generator) module, also called Test Profile Generation Component 16.

[0056] Here, on the one hand, acceleration signals are measured during a test run on a route. (Using the ASPEN method, a profile can be generated for each vibration measurement. For simplicity, acceleration is discussed here as the input value for the profile calculation.) On the other hand, a specially generated time signal - the reference signal - is evaluated. Therefore, for a profile calculation using this method, two of the three modules (FDDC and SPEX) are executed twice, see Figure 4. The evaluation of both the measurement signal and the reference signal from the route run is performed in the same way.

[0057] Typical calculation steps of the method are as follows: Calculation of PSS for each individual signal (FDDC module), Extrapolation and superposition of individual signals for PSS (SPEX module), · Test profile generation (PRGN module).

[0058] For simplicity, we first describe the profiling procedure with the ASPEN RoMi method from measurement data of only one measurement. This leads to some limitations, such as the impossibility of profiling for multi-point control, which will be addressed later in the individual descriptions of the SPEX and PRGN modules.

[0059] The following measurement technique terms are used in the method description: According to DIN 1319-1, the measured value is the time-varying physical value to be measured (e.g. acceleration). It is always related to a specific measurement position and measurement direction (since vibrations are direction-dependent). Furthermore, in the documentation it is denoted γ(t). Measurement location (measurement location) is a spatially limited local position on a component or its holder that is selected to capture a measurement value. For this purpose, a sensor (e.g., an acceleration sensor) is attached to this position. The measurement signal is the result of the measurement of the measurement values. It is generated during travel on a section of the route (measurement section) after a single trigger and stop of the measurement by the measurement system and is available in digital form. The measurement signal of the measurement value γ(t) captured during travel on the route i on such a measurement section is denoted as γ i It is called (t). A measurement data file is a file containing the measurement signals of all measurement values ​​that are time-synchronously captured by the measurement system during the journey on the measured section of the route, digitized and stored on a data carrier. The measurement data file has a name. The measurement signals are displayed here in digital form and saved under the name of the respective measurement value.

[0060] Suppose that for vibration approval of a component (intended for operation in a motor vehicle), a route mix consisting of n different design-related routes is defined. The travel time T RM is the individual root T Ri(i=1,...,n). The component under consideration will have its total life T LD This means that:

number

[0061] Furthermore, for data capture, a test was performed (e.g., in a test vehicle) on each section of each route, and a value γ (e.g., unidirectional speed) was measured at one location on the component being tested. This allows for a time period t along each route i (i=1, ..., n). γi For each measurement interval (measurement section), the acceleration signal γ i (t) exists, and it is assumed that each is saved in a separate measurement data file, so that there is a measurement data file with this signal for each route traveled. (This assumption applies to the description of the ASPEN RoMi method in all respects of this description, except for the "Additional Explanation" at the end of the description below, where the extrapolation and superposition of PSSs is addressed for the case of multiple measurement data files per route traveled.)

[0062] From these n signals (as input data), the damage-equivalent test profile is derived using the ASPEN RoMi method. In this case, the PSS is calculated for an assumed (fictitious) Wöhler line. Mathematically, this is expressed as follows: A , a A (Support point A), and k WL (slope coefficient), which is the same for all amplitudes a. WLSince it has a tensile strength of 1000 kJ / s, it passes without kinks and without any particular fatigue strength range. Furthermore, this is referred to herein as a simple Wöhler line. As is well known, the mathematical description of component damage using the Wöhler line applies only to fatigue [1, 15]. Therefore, it is required for profiling in the ASPEN method.

[0063] The test frequency range is f unt , …, f ob is.

[0064] The profiling for the considered case is described in the following steps a) to g), which are shown in Figure 4 as follows: calculation steps a) to c) are shown by thick branches with arrows in the left part of the diagram, d) to f) are shown by thin branches in the right part of the diagram, and for calculation step g), these branches are finally integrated into the PRGN module.

[0065] a) Calculating PSS for route running First, the defined test frequency range f = f unt , …, f ob In this case, each acceleration signal γ measured on each section of the route i=1, ..., n is i For (t), PSS D γ i (f) is calculated. For this purpose, a set of linear bandpass filters (e.g., Butterworth type) is defined, each with a narrow passband. The passbands of adjacent bandpass filters are adjacent to each other so that they do not overlap or have gaps between them, and they cover the entire test frequency range. For each input signal γ i (t) is filtered by them. A force spectrum is then formed from each bandpass filtered signal using a counting method. Classification methods that provide, in addition to the amplitude, also the mean value of the vibration cycle (the so-called two-parameter force spectrum) are preferred. In the ASPEN method, the rainflow method is used for classification. The rainflow method provides the rainflow force in the form of a rainflow matrix. Each such two-parameter weight spectrum is then converted into a damage-equivalent, mean-free, one-parameter amplitude weight with the aid of amplitude transformation according to Haigh

[15] (Haigh diagram), from which the PSZ is subsequently calculated with the aid of a simple Wöhler line and the linear damage accumulation assumption (e.g., in the "Miner elementary" form). Furthermore, each PSZ calculated in this way is assigned to the center frequency of the passband of the corresponding bandpass filter. The sequence of these PSZs in ascending order of filter center frequency is finally calculated as the signal γ i (t) PSS D γ i All these calculations are performed in the FDDC module, see the n calculation blocks FDDC (also called pseudo damage spectrum calculation component 13) in the left part of the diagram in Figure 4. The FDDC algorithm for calculating the PSS from the time signal is described in more detail below.

[0066] b) Extrapolation of PSS for route running These individual PSS D γ i (f) (i=1,...,n) is the signal measurement period t γ i To keep the testing and measurement effort during data capture as low as possible, γ i is usually the vehicle's required lifespan on this route (T LD actual run time T Ri is significantly smaller than t γ i < <T Ri The total travel time T on route i RiTo obtain a PSS that covers γ i (f) with coefficient k γ i Multiply by.

number

number

[0067] c) Superimposing PSS for route running The extrapolated PSS D calculated in the previous step according to (3.2) γ i,EX (f) (corresponding duration T Ri represents the damage of a component when traveling only on each individual route i (of n routes), where the component is designed to operate on the entire route mix. Therefore, we must determine the total damage experienced by the component in the complete mix of n different routes. The travel time on the route mix is ​​calculated by multiplying the travel time T Ri (see relation (3.1)), the damage measure of the root mix can also be calculated as the sum of the damage measures of the individual roots. Therefore, the individual extrapolated PSS D γ i、EX (f) is where the total PSS D γ、SPEX (f) can be added.

number

[0068] This procedure is called PSS superposition. PSS superposition is also based on the method of superposing the damage numbers known in the operational stability, or adding the PSZ of each operational condition (here, the root) to the total damage. Superimposed PSS D γ,SPEX (f) is the travel time T RM In this case, the running time T R1 , T R2 , …, T Rn is distributed among the n individual roots of this mix according to (3.1). T LD =T RM According to this assumption, the superimposed PSS D γ,SPEX (f) simultaneously for all its intended lifetime T LD The superposition of the PSS is also performed in the SPEX module, in particular in the superposition subcomponent 16 (see FIG. 4).

[0069] d) Reference signal generation Here, the question arises as to the determination of the damage equivalent height of the test profile. According to the ASPEN RoMi formulation, this means: VT In the vibration test, this height profile is RM This means that the same damage should be applied to the component as experienced in the root mix of the PSS D. This damage is calculated by the extrapolated and superimposed PSS D calculated in steps a) to c). γ,SPEX (f). The problem with converting PSS to damage-equivalent profile amplitudes is that the relationship between these two values ​​must be known. This depends, among other things, on the specified type of vibration test (sweep or noise). To solve this problem, a time signal r(t) suitable for the type of vibration test is independently generated and timed to the intended duration of the vibration test (T VT) to calculate the damage. r(t) is designated as the reference signal. This is nothing other than the control signal that the vibration test equipment generates when realizing a specific type of test profile (sweep profile or noise profile). For example, it can be generated as the actual vibration signal on a vibration test stand, equipped with an appropriate control system and recorded with a suitable measurement system. Many PC-based signal processing tools, such as Matlab®, Famos, LabView®, etc., also offer alternative possibilities. In these environments, the reference signal can be generated as a fictitious digital signal (for example, when creating a noise profile with the help of a random number generation algorithm). This is a time-saving and cost-effective possibility, since it does not require a vibration test stand to generate the vibrations or a measurement system to record the signal. The level of the reference signal can be chosen arbitrarily. However, it must be within the entire test frequency range f = f unt , …, f ob (following the concept of the ASPEN method) for r(t). If r(t) is the sweep signal (for sweep tests) or the LDS PSD Ref and if r(t) is a stochastic signal (for noise testing), the level is determined by the amplitude frequency curve (AFV) S Ref Therefore, S Ref =const, PSD Ref = const is required. In the latter case, this condition is satisfied when r(t) is within the test frequency range f=f unt , …, f ob This means that the signal is generated as white noise. The procedure for generating such a signal, together with the following steps e) and f), has been referred to as the reference signal concept and will be explained in more detail below.

[0070] e) Calculation of the PSS of the reference signal After the reference signal r(t) has been generated as described above, its PSS is calculated. For this, the FDDC algorithm is executed again (see calculation block 13 in the right part of Figure 4). For the ASPEN method to work correctly, the same calculation parameters must be used as for calculating the PSS of the signal from the route mix run (in step a). In particular, this concerns the configuration of the bandpass filter, as well as the configuration of the classification parameters and Wöhler line parameters. The calculated PSS is r The generated reference signal r(t) has a period t r It is effective against.

[0071] f) Extrapolation of the reference signal PSS A specified period T VT In order to calculate the damage caused by the vibration test, the reference signal D calculated in the previous step is used. r The PSS for (f) must be extrapolated to this period in the same way as the PSS for the route run (see step b), i.e., D r This is done by multiplying (f) by the corresponding extrapolation factor.

number

number

[0072] g) Calculation of test profile height Here, Root Mix D γ,SPEX (f) The superimposed PSS and the reference signal D r,EXTwo damage degrees are available for the extrapolated PSS in (f). In addition, component damage D r,EX (f) is the amplitude S Ref Vibration testing with sweep excitation by LDS PSD Ref In the vibration test using noise excitation by VT It is also known that this occurs when S Ref or PSD Ref with a different amplitude S U or another LDS value PSD U This can be converted into damage D γ,SPEX This damage equivalence conversion is defined by equation (3.17) (for sweep profiles) or equation (3.16) (for noise profiles), depending on the profile type. (In this, D Ref (f)=D r,EX (f) and D U (f)=D γ,SPEX (f).) The two relations (3.16) and (3.17) were developed independently and called the ASPEN transformation (AT). For this conversion of PSS to profile amplitudes, a PRGN module (also called test profile generation component 16) is implemented, see Figure 4. The conversion is then performed for a defined test frequency range f = f unt , …, f ob is performed separately for each frequency f. The sweep profile or noise profile calculated in this way forms the result of the method, which is suitable for damage-equivalent vibration testing of the component under consideration with profile control (single point control) at the locations where the acceleration values ​​x(t) are measured during the route travel.

[0073] Special features and additions to the above a) to g) procedures This is used separately to create a profile in each spatial axis. For this purpose, the signal γi(t) measured in the corresponding spatial direction is used. Therefore, this method is mainly suitable for vibration tests on test stands that have the possibility of generating vibrations in only one spatial axis (e.g., electrodynamic shaker test stands). If the vibration test equipment allows for the simultaneous generation of vibrations in three spatial directions (as is typically the case on servohydraulic test stands), profiles can first be created separately in each direction according to the procedure described above. Then, during the test, vibrations generated by the vibration control system corresponding to these profiles can be simultaneously introduced into the test object in mutually orthogonal directions. The reference signal can also be a multi-sweep (to generate a multi-sweep test profile) or a sine wave with a fixed frequency (to generate a profile that can perform dwell time testing). Following this uniform procedure, excitation and response profiles are generated. It only distinguishes between using signals of various measurement locations on the component itself (to generate the response profile) or on its holder (to generate the excitation profile). All these signals must be recorded synchronously in a route mix run. The excitation profile is usually used for test stand control, and the response profile is used to limit the amplitude of the component's response. If signals are present at multiple measurement locations and should be evaluated together to generate a profile, see the "Multiple Measurement Points" section immediately below. Except for the whole root component, this procedure also generates profiles to cover the damage of individual root components. For this purpose, in AT(3.16) and (3.17), the corresponding extrapolated PSS D γ i、EX (f) is D γ,SPEX (f) were used instead of (f). (These were calculated by default for each route in step b.) Such a profile is used to calculate the travel time T Ri Covers component stresses for

[0074] Multiple measurement locations So far, the method has been described for the simpler case when the test profile is derived only from the signal of one measurement (defined for one measurement position). In a slightly modified form, this procedure can also be used when measurement signals are recorded in a route run at several positions in the same direction (for example on the component and / or on its holder) and they are to be evaluated together for profiling. The differences in the method concern only the PRGN algorithm and only apply in the case of multi-point control. This case will be explained later.

[0075] Pseudo Damage Spectrum (PSS) The concept of pseudo-damage spectrum is introduced into the ASPEN method to represent the parameter dependence of PSZ on the center frequency of each narrow bandpass. The PSS is the fundamental and central parameter of the ASPEN method. When considering an oscillating time signal, the PSS describes the distribution of the PSZ of its individual harmonic components over the oscillation frequency. The dependence of a function on frequency is physically similar to the concept of a spectrum (as calculated from a time signal in the classical way by the Fourier transform

[14] ), so the term spectrum is used in the PSS concept.

[0076] According to its physical content, PSS is similar to FDS (Fatigue Damage Spectrum) parameters. FDS is widely used to analyze the fatigue behavior of components and to synthesize damage equivalent profiles using damage-based and model-based methods based on models of linear underdamped EMS [6-9]. Therefore, FDS absolutely requires a model of such EMS (with base point acceleration as input variable and relative displacement amplitude of EMS as vibration response). In contrast, the PSS calculation in the ASPEN method is not tied to a mathematical model of the component for which the test profile is to be generated. This is the difference between FDS and PSS.

[0077] Since the PSS consists of individual PSZs, the PSS is likewise fictitious, not real. Therefore, it is not suitable for absolute component life prediction. For example, a meaningful interpretation of the PSS can only be achieved by comparing two PSSs calculated for two different time signals. This means, in fact, that at certain vibration frequencies, loading of a component according to a time sequence is more damaging (more severe) than loading according to another. However, the calculation of the two PSSs must always be performed under identical conditions. This affects all PSS calculation parameters, especially the Wöhler line parameters and the width of the bandpass filter. If (rarely) a suitable Wöhler line is actually known for the component, material, measurements, and load case under consideration, it can be set and used in the ASPEN method. This means that the calculated damage numbers are no longer fictitious and can (possibly after extrapolation and superposition) provide information about the component's actual life. In this case, it is not the PSZs and PSSs but the actual damage numbers or damage spectrum.

[0078] Each PSS D(f) is Xis calculated from the time signal x(t) and is valid only for this period. It must be specified together with the PSS. Without this information, it is not possible to know for what time the (pseudo) damage has accumulated at each frequency.

[0079] PSS can be used not only for the synthesis of damage-equivalent test profiles, but also, for example, to compare the severity of two profiles of different types (e.g., a sweep profile with a noise profile).

[0080] PSS (FDDC algorithm) calculation

[0081] The FDDC module functions to calculate the PSS from the time signals measured during the route run and from the reference signal. X For each signal x(t) in the matrix, independent of its origin, one PSS D(f) is calculated according to the same algorithm described below. In the above notation, in this case x(t)=γ i (t),t X =t γ i (i=1, ..., n).

[0082] The FDDC algorithm includes the following: See Figure 5.

[0083] a) Signal Filtering In step S1, the signal x(t) is filtered with a set of narrow bandpass filters. First, the selected test frequency range f = f unt , …, f ob In , bandpass filters are set. They can be of any type, for example Butterworth, Bessel, Tschebyschew, etc. In order to achieve a sufficiently high resolution of the profile curve calculated over frequency, typically f unt = 10Hz to f obAt least 500...1000 bandpass filters are set in the frequency range defined for vehicle equipment vibration testing, up to 2 kHz. This results in a passband of only a few Hertz. Therefore, these bandpass filters can be called narrowband. Depending on the selection of filter synthesis parameters (varying corner and center frequencies), the bandpass filters can have equal or variable passband widths. Regardless, these parameters must be selected so that the passbands of adjacent filters do not overlap or have gaps between them. From this perspective, the filter order and filter type (conventional or "zero-phase" filtering) can be freely selected. Filtering an input signal x(t) with a set of m such bandpass filters provides m output signals. The j. bandpass output signal is defined as x(t). BP,j (t) (j=1, ..., m)

[0084] b) Classification of the bandpass filtered signals in classification step S2 Next, the output signal x of each filter BP,j From (t), a weight spectrum is formed using classification. For this, x BP,j The entire amplitude range of is divided into l classes, and the amplitude a i With the help of counting or classification methods, the number of cycles N OP,i For this purpose, various classification methods known from operational stability can be used. However, it is preferred to determine the amplitude a i In addition, for each cycle, the average value m i A classification method, such as the Rainflow method, is used that also allows the determination of the load spectrum N OP =N(a,m) is the time signal x under consideration BP,jIt represents the frequency of occurrence of cycles with a specific amplitude a and mean value m in (t). For the reasons mentioned above, in the ASPEN method, a rainflow classification is used to form the load spectrum. The rainflow classification provides rainflow loads in the form of a rainflow matrix. They are expressed in terms of the coordinate "mean value / amplitude". That is, each measured vibration cycle is characterized by its amplitude (half span) and mean value. (Thus, for example, if the ordinate is m / s 2 For an acceleration signal of m / s, the two abscissas of the rainflow matrix are also 2 (The ordinate contains the number of closed oscillation cycles.) Any residuals that may exist are counted into a matrix after the rainflow method is completed. In this way, each filtered signal x BP,j The rainflow matrix is ​​calculated for

[0085] c) Conversion to damage equivalent mean-free amplitude in conversion step S3 In the rainflow matrix, as mentioned above, each oscillation cycle has a corresponding amplitude a i and the average value m i However, with the help of the approach by Wöhler, known for its operational stability, only the amplitudes are used to determine the damage of the cycles (see next step d). As is known, this approach does not take into account the average value of the cycles. Since this approach can affect the damage of the vibration cycles in some circumstances, the amplitudes a can be affected by the average value of each vibration cycle in the rainflow matrix. i First, the damage equivalent mean amplitude (SMA) a SM,i This is done with the help of the Haigh method

[15] (called Haigh diagram), which is also known from operational stability. These transformed amplitudes are then sorted in ascending order. As a result, the number of oscillation cycles in the rainflow matrix (which remains unchanged during the Haigh transformation) is converted to SMA N OP =N(a SM )

[0086] d) Calculation of partial damage contribution in step S4 where each filtered signal x BP,j Each SMA a of the spectrum with l classes (i=1, ..., l) is formed from SM,i The damage level is calculated for

number

number

[0087] e) A total damage calculation step S5 for calculating the total damage for the output signal of each bandpass. According to (3.7), each amplitude a of each load spectrum SM,i Partial damage contribution D i After determining, the different amplitudes aSM,i The resulting damage D caused by the action of loads with (i=1, …, l) GS is calculated using the linear damage accumulation assumption (Palmgren-Miner rule) as follows:

number

[0088] f) Formation of PSS in pseudo damage spectrum formation step S6 Current test frequency range f=f unt , …, f ob The PSZ calculated for the output signal of all m bandpasses defined in BP The expression of PSZ depending on each filter center frequency is PSS D(f)(f=f BP ) results.

[0089] It should be mentioned again that each measurement signal in the FDDC module (also referred to as the Pseudo Damage Spectrum Calculation Component 13) is evaluated according to this same algorithm, without relying on signals at other measurement locations or on signals at the same location measured on other routes, so that, for example, even if multiple values ​​(signals at multiple measurement points) are recorded during the route run, this does not result in any change in the FDDC algorithm.

[0090] Extrapolation and superposition of PSS (SPEX algorithm)

[0091] time signal γ recorded during a run on route i i PSS D calculated from (t) using the FDDC algorithm γ i(f) is the measurement period t γ i This usually refers to the total travel time T Ri significantly shorter than t γ i <T Ri However, in vibration tests with a test profile, the full life of the component must be guaranteed. According to (3.1), this is the travel time T Ri Therefore, all PSSs calculated by the FDDC algorithm must be extrapolated and possibly convolved in the next step. For this purpose, the SPEX algorithm was developed. The extrapolation and superposition of PSS in the ASPEN RoMi method is based on a procedure known in operational stability, using which (real or fictitious) damage numbers are extrapolated and superposed. Extrapolation involves multiplying the damage numbers determined from tests on individual routes by the correspondingly calculated extrapolation coefficients. Superposition involves adding the extrapolated damage numbers together. Since each ordinate of the PSS is a PSZ, this procedure can be applied without modification to the extrapolation and superposition of PSS. In the following description of the SPEX algorithm, it is assumed that there is one measurement data file per route driven, as mentioned above. The extrapolation and superposition of the PSS for the case of multiple measurement files per route is described below.

[0092] To perform the extrapolation and superposition algorithm in the ASPEN RoMi method, the travel time T Ri Only must be specified.

[0093] In the block diagram of the ASPEN RoMi method in FIG. 4, the extrapolation of the PSS is represented by n calculation blocks 14 (left part of the diagram) and the superposition is represented by block 15.

[0094] Extrapolation of PSS for route running

[0095] The extrapolation of the PSS is described by equation (3.2), which is based on a constant k γ i PSS D γ i (f) is a pure multiplication procedure that calculates the total travel time on route i and the signal γ i (t) (individual) measurement time t γ i Relative to T Ri It is calculated as k γ i is the same for all frequencies f in the PSS, so the extrapolated PSS D γ i,EX (f) is always D γ i (f) is a scaled copy of (f).

[0096] Physically, this extrapolation method corresponds to the following idealized interpretation: the test vehicle drives on the measured section of route i with k γ i During each repeated run, the measured signal under consideration γ i For (t), the exact same PSS D γ i Then, at the end of all these repeated runs, a specified run time T Ri and the corresponding PSS D γ i,EX The damage,represented by (f), and,is reached.

[0097] The extrapolation algorithm of the ASPEN RoMi method is based on the measurement signal γ i (t) and the extrapolated PSS D for each measurement data phi γ i,EX (f) is calculated. A measurement data file exists for each route traveled, and T Ri is specified as the total travel duration on route i, the extrapolated PSS D γ i,EX (f) shows the pseudo-damage, depending on the frequency, which is measured at the position under consideration (of the component or its holder) (and only at this position) for the entire travel period T on the route under consideration. Ri The travel time on the route is calculated based on the required component life T LD(It is still assumed or required that components be designed for the entire route mix.) Therefore, T Ri <T LD The extrapolated PSS D γ i,EX (f) forms a side result of the SPEX algorithm.

[0098] Nevertheless, the extrapolated PSS D γ i,EX (f) is important Practical significance This allows for a comparison of the severity of different routes in the route mix. D calculated for the same measurement location from runs on different routes γ i,EX When (f) are superimposed on each other (e.g., graphically), conclusions can be drawn about which route is more damaging than the other route for the location under consideration, in which frequency range. Note, however, that the duration of travel on different routes may not be equal. That is, T R1 ≠T R2 ≠T R3 , etc.

[0099] PSS overlay for route running

[0100] The superposition of PSS in the ASPEN RoMi method is described by equation (3.4). This is done after extrapolation, and the signal γ for the same measurement γ i (t) multiple previously extrapolated PSS D γ i,EX (f) is a purely additive procedure, and the total PSS D γ,SPEX (f). The PSS is added separately for each frequency f. One conclusion from this is that the superimposed PSS D applied to the entire route mix γ,SPEX (f) is the extrapolated PSS D representing the simulated damage only on individual routes. γ i,EX (f) is more damaging than any of the other two, at all frequencies. RM The travel time above is for each individual route T Ri This is understandable as it is larger than the above travel time. See equation (3.1).

[0101] As mentioned above, the convolution algorithm of the ASPEN RoMi method generates one convolved PSS D for each measured value γ. γ,SPEX (f) Calculate this PSS D γ,SPEX (f) is the vibration frequency at which the component will vibrate over all pre-specified lifespans T LD Therefore, the simulated damage experienced at each measurement location is calculated for all the superimposed PSS D γ,SPEX (f) forms the main result of the calculation in the SPEX module. However, as a result, if only one measurement data file is fed to the profiling method, no superimposed PSS is formed. This may be the case, for example, if measurements are performed on only one route (no route mix).

[0102] Extrapolation of the PSS of the reference signal Reference signal D r The PSS of (f) (previously calculated in the FDDC algorithm) is extrapolated according to (3.5). For this purpose, the extrapolation coefficient k r is determined according to (3.6). This formula is used for vibration test T VT This period must be determined at the latest at this stage of profiling under the ASPEN RoMi Act. Since the reference signal is a single signal, no superposition is required for its extrapolated PSS.

[0103] Exemplary Embodiment

[0104] Here is given an example of the execution of the SPEX algorithm (also called the extrapolation and superposition process P2) on two measured values ​​x, y (γ=x, y) recorded in tests on n different routes of a route mix. x, y can be, for example, acceleration values ​​measured at two different positions on the component to be inspected or on its holder. These x and y signals are measured on sections of i routes (i=1, ..., n) and xi (t), y i (t), their measurement period t xi , t yi (Generally, t xi ≠t yi ). PSS D xi ,D yi has been previously calculated in the FDDC algorithm, which means that where t xi , t yi is supplied to the SPEX module together with

[0105] The sequence of extrapolation and superposition is shown in Figure 6. First, PSS D xi , D yi For this purpose, the total travel time T Ri The total life of the component (required) is specified. LD Ratio to T Ri is still determined by (3.1), and according to (3.3), the extrapolation coefficient k xi or k yi is calculated. PSS D xi , D yi According to equation (3.2), the extrapolation of xi or k yi This is done by multiplying them by The extrapolated PSS thus formed is shown in Figure 6 as D xi,EX (f), D yi,EX (f) (i = 1, ..., n), where they are superimposed. This means that, according to (3.4), all PSS D xi,EX (f) and all PSS D of channel y of each route yi,EX (f) and (f) are added together. This results in one superimposed PSS for each measurement γ, i.e., in the example under consideration, a total of two PSSs D x,SPEX (f), D x,SPEX (f). The results of the execution of the SPEX algorithm in the example under consideration (see Figure 6) are as follows: Multiple extrapolated PSS Dxi,EX (f), D yi,EX (f) (two PSSs for each route i, sub-results), and Extrapolated and superimposed PSS D x,SPEX (f), D y,SPEX (f) (Total of 2 PSS, main outcome). All of these can be fed into the PRGN algorithm to calculate different profiles.

[0106] Calculation of test profile (PRGN algorithm)

[0107] The PRGN algorithm works by determining the profile amplitude from the PSS, for which the ASPEN transform (AT) is used.

[0108] The PRGN algorithm is first detailed in "Description." In "Types of Resulting Profiles," the types of resulting profiles that the PRGN algorithm (and therefore the entire ASPEN RoMi method) computes are described. Then, in "ASPEN Transformations," the AT is described for two distinct cases: the generation of noise profiles and sweep profiles.

[0109] explanation

[0110] Input data The input data for the PRGN algorithm is as follows: i) Extrapolated PSS D for route travel γ i,EX (f) or extrapolated and superimposed PSS D γ,SPEX (f) (calculated in the SPEX algorithm) ii) the extrapolated PSS D of the reference signal r(t) r,EX (f) (also calculated in the SPEX algorithm) iii) the parameters of the reference signal r(t), i.e., ·Amplitude S Ref , if the reference signal r(t) is a sweep (to generate a sweep profile), or LDS PSD Ref , if the reference signal r(t) is generated as white noise (to generate a noise profile).

[0111] Notation Regardless of whether the PSS is extrapolated or extrapolated and convoluted, the profile calculation in the PRGN module is performed according to the same algorithm. What is important for the performance of the PRGN calculation is the measured signal γ i The only difference is the difference in (t), i.e., the profile is the extrapolated but not superimposed PSS D γ i,EX (f) or from the measurement γ (see further above for the definition of the term measurement), the profile is the extrapolated and superimposed PSS D γ,SPEX (f). This is only applicable if multi-point control is defined. For a clear description of the functionality of the PRGN module, the PSSs of a total of p measurement signals (or measurements) from any SPEX result data file should be evaluated with the PRGN algorithm, of which q (q≦p) are intended for multi-point control. The input PSSs of the PRGN algorithm (see the "Input Data" subsection above) calculated with the SPEX algorithm are renamed as follows compared to the definition above: D γ i,EX =D U,k - measurement signal γ of the measured values ​​γ measured on the travel route i The extrapolated PSS of (t). (The index i, which determines the number of routes i, is not meaningful for the description of the PRGN algorithm and is therefore omitted. Instead, the measurement signals measured on this route are simply numbered consecutively, and the extrapolated PSS of the measurement signals then obtain consecutive numbers k, k=1, …, p.) D γ,SPEX =D U,k - the superimposed PSS for the measurements γ calculated for all route runs on the route mix. These superimposed PSS for each measurement are now also simply given a consecutive index k (k=1, ..., p).

[0112] In other words, the extrapolated but non-overlapping PSSs need not be distinguished from the extrapolated and overlapping PSSs for the purposes of the PRGN algorithm description. Therefore, for the purposes of this discussion, these PSSs are uniformly U,k where k is the number of consecutive measurement signals or measurements in the input data file provided to the PRGN module.

[0113] PRGN algorithm description PSS D of all p measurement signals (or measurements) in the input data file (k=1, ..., p) U,k From the PRGN module, single point control PR V,k p individual profiles or profiles for multi-point control, or one profile PR MP is calculated. This is done in the following steps:

[0114] a) ASPEN Conversion (AT) First, single profile PR U,k The height of each extrapolated or extrapolated and superimposed PSS D U,k From the extrapolated PSS of the reference signal, r,EX (f) and its amplitude S Ref , or its LDS PSD Ref , and the slope coefficient k of the simple Wöhler line WL (described by equation (3.8)) is given and calculated. Depending on the type of test profile to be generated, this is done with the help of AT (3.16) or (3.17), i.e., When generating a sweep profile, AT is used in the form of (3.17), where the amplitude S of the previously generated sweep reference signal is Ref is specified. When generating a noise profile, AT is used in the form of (3.16), where (3.16) is the height of the previously generated white noise reference signal LDS PSD Ref is specified. In AT, the gradient coefficient k of the Wöhler line WL The same value of must be used in calculating the PSS from the signals recorded during the route run and from the reference signal as was used previously in the FDDC algorithm.

[0115] b) Profile calculation for multi-point control In the case under consideration, q out of the total p existing measurement locations are for multi-point control, where the individual profiles PR of these q measurement locations are U,k Combine these to create a profile PR U,MP In this case, a distinction can be made between average control and maximum control strategies. To obtain a profile suitable for averaging control, the individual profiles of all corresponding positions are arithmetically averaged at each frequency, i.e.

number

number

[0116] c) Weighting by safety and test factors Calculated Profile PR U,k and PR U,MPcan already be used in this form for vibration tests (with profile control at the corresponding points). However, the results of component tests using such profiles often do not provide the sufficient reliability required for series approval. This is because these profiles do not take into account the variations in stress and stress capacity of the test objects, nor do they take into account the uncertainty of the test results due to the limited sample size (i.e., the uncertainty resulting from testing a limited number of test objects). Therefore, in order to increase the reliability of the test results, the ASPEN RoMi method provides the possibility to increase the calculated profile amplitude along with the corresponding procedure for operational stability. For this purpose, the safety factor and test factor j are used. STF is introduced and applied to the previously calculated profile amplitude (j STF >1). That is, When generating a sweep profile, PR U,k and PR U,MP On each ordinate of j STF (k=1, ..., p) are multiplied. That is,

number

number

number

number

[0117] For practical applications, the safety and test factor j STF must be determined based on engineering considerations and / or obtained from known literature on operational stability, e.g.,

[15] .

[0118] As a result of the execution of steps a) to c), the single point control PR V,k p individual profiles or profiles for (k=1, ..., p) as well as profile PR for multi-point control MP Similar to the extrapolation and superposition in the PSS (SPEX algorithm), the calculation of the height of the test profile in the PRGN algorithm is performed for each frequency f independently, independent of other frequencies. Therefore, the symbol f is used in the above explanation, steps a) to c), to calculate the height of all PSS D U,k and omitted in the profile function PR.

[0119] Exemplary Embodiment

[0120] We will now illustrate the above PRGN algorithm in a concrete example. Assume that after performing route runs, FDDC calculations, and SPEX calculations, the extrapolated PSS or extrapolated and superimposed PSS are available (in one data file) for three measurement signals or measurements x(t), y(t), z(t). These three PSS are U,k(k=1, ..., 3). The first two values ​​x(t) and y(t) are for multi-point control. The profile to be generated must cover these PSSs.

[0121] The sequence of profile calculations using the PRGN algorithm, steps a) to c), for this case is shown in Figure 7. U,k In addition to the PSS D, a further input value is required for the algorithm, namely the extrapolated PSS D r,EX and the height A of the reference signal r(t) r is required. This means that if a sweep profile is to be generated, the constant amplitude S of the harmonic signal r(t) Ref and the constant amplitude PSD of the stochastic signal r(t) (white noise) if a noise profile is to be generated Ref Ar={SRef,PSD Ref}.

[0122] From these input values, we first calculate the slope coefficient k WL is specified, and three individual profiles PR U,k is calculated by AT(3.16) or (3.17) (k=1, ..., 3). The profile PR of the first two values ​​x(t) and y(t) U,1 , PR U,2 is then calculated using one of equations (3.10) or (3.11) to find the profile PR for multi-point control. U,MP In Figure 7, these equations are represented by the operator Φ. Depending on the type of profile to be generated (sweep or noise), the ordinates of all four profiles are finally integrated with the safety and test factors j STF or its square (j STF ) 2 is multiplied by

[0123] In the example under consideration, the results of the PRGN calculation are obtained by dividing the four profile functions PR V,k(k=1,...,3) and PRMP. Depending on the treatment of the reference signal and the use of AT equations (3.16) or (3.17), they can all be either sweep profiles or noise profiles. The first three PR V,k are the individual profiles. They are determined for vibration tests using profile control at the individual positions where signals with values ​​x(t), y(t), z(t) were recorded during the route run. The PRMP profile is implemented using multipoint control of the signals of two sensors placed as accurately as possible at the positions of recording of values ​​x(t) and y(t) during the route run. The strategy of profile control at these two positions (average control or maximum control) depends on whether equation (3.10) or (3.11) was used in calculating the amplitude of the profile PRMP.

[0124] Result Profile Types

[0125] Depending on the input data and parameterization, the PRGN algorithm (and thus potentially the entire ASPEN RoMi method in one run) will compute multiple different types of test profiles, or profiles with different characteristics. Below are descriptions of the different types of profiles, along with information on their execution in some cases.

[0126] a) Sweep profile and noise profile, profile for dwell time test The difference between these test profiles is the type of vibration generated using harmonic or stochastic vibration excitation. The first category includes tests using slip frequency excitation (sweeps) consisting of one tone (single sweep) or several tones (multi-sweep), as well as dwell time tests using fixed frequencies. All these test types are standardized by DIN EN 60068-2-6. A limitation of the ASPEN method is that in the case of multi-sweep profiles, the frequency bands of the individual sweep tones must not overlap. The second category includes broadband noise tests according to DIN EN 60068-2-64. Using the ASPEN method, profiles for combined excitation, in which one or more sweep tones are superimposed with noise, can also be created. This test type is standardized by DIN EN 60068-2-80. The prerequisite for this is that only one profile amplitude (e.g., the amplitude of a sweep profile or the LDS value of a noise profile) needs to be determined at one frequency.

[0127] b) Excitation and response profiles It makes sense to measure the vibration, load or stress values ​​for generating a profile using the ASPEN method both on the affected component and at its fixed or connection points on its holder (the holder can be, for example, the drive, such as the internal combustion or electric drive of a vehicle, the transmission of a vehicle, the body or the axle). Using the ASPEN method, excitation profiles are derived from data measured at support or fixed points of the component under consideration. These are used to generate the vibrations that are introduced into this component (test object) during the vibration test. In most cases, this excitation profile is also adjusted. If vibrations are recorded directly at the component under consideration (during a route run), the ASPEN method can also be used to calculate a response profile from this data. This describes the desired or required vibration amplitude that the test object should experience during the vibration test in response to the introduced excitation profile. In contrast to the excitation profile, the response profile of a resonating component is rarely used for shaker control, for various reasons. The only difference between excitation and response profiles is the selection of the measurement locations from which the profiles are derived and their handling during the vibration test, i.e., the control or (possibly with limitations) monitoring of the excitation profile based on the response profile. In the ASPEN method, the excitation profile and the response profile are generated according to the same algorithm.

[0128] c) Single-point and multi-point control profiles The difference between these profile types is whether the profile is controlled at one or more points (locations) in the vibration test. In single-point control, a vibration sensor is placed at one location in the test setup. Vibration is introduced into the structure at this location according to a pre-specified profile. In multi-point control, a pre-specified profile is adjusted according to the signals of multiple vibration sensors placed at different locations in the test setup. Profiles for single point control are generated by default in the ASPEN RoMi method, PRGN module, from all signals measured on the route or on the entire route mix, using the procedure described above. If signals are recorded (during a driving test) in several positions on a component or its holder and in different directions, the differences between these profiles are due only to the positions and measurement directions to which they are applied. Therefore, the test profile PR V,k If a profile is generated using the PRGN algorithm from the signal of a sensor with serial number k, the profile can be meaningfully used in a test only with profile control in the corresponding position and in the corresponding direction (sensor measurement direction) where this sensor was located in the route run. A profile for multipoint control can be generated from all signals measured on the route or on the entire route mix in any combination. To do this, it is only necessary to specify these signals. Extrapolated or extrapolated and superimposed PSS D for p measured signals U,k If p is available in the SPEX result data file, any q profiles of these signals can be specified in the PRGN module (2≦q≦p) for the calculation of the profile PRMP for multi-point control. Such a test profile PRMP is generated by jointly processing q individual profiles according to equations (3.10) and (3.11) and should also be controlled in the vibration test according to the signals of all these q sensors located at the corresponding positions. That is, this profile can only be correctly implemented by multi-point control of the signals at the corresponding sensor positions. The control strategy should also correspond to the algorithm for calculating a single profile. That is, if equation (3.10) is used for this purpose, average value control should be applied, while if equation (3.11) is used, maximum value control should be applied. These control strategies must be implemented by corresponding settings in the vibration control system of the test stand. If only one measurement data file (see above) is fed to the ASPEN RoMi method for profile generation, no test profile for multi-point control is calculated. This may be the case, for example, if measurements are performed only on one route (no route mix).

[0129] d) Profiles covering damage on individual routes or on the entire route mix When measurements are performed on roots associated with several different designs of root mix, the ASPEN RoMi method generates profiles that cover both the damage on each individual root and on the entire root mix. This occurs in a single run of the method. Individual routes Profile PR generated from extrapolated but not superimposed PSS V,k , PR MP is the travel period T only on the route i under consideration. Ri It does not take into account the running of other routes in the route mix. Also, it does not take into account the required component life T LD (It is still assumed or required that components be designed for the entire route mix.) In addition, Ri <T LD , such a profile is weaker than a profile that covers the entire route mix and therefore the required component life. Therefore, only on individual routes of the route mix is ​​the running period T Ri The profiles covering the range may only be of secondary importance for the actual vibration test. In this respect, these profiles are merely a by-product of the profiling of the ASPEN RoMi method. Nevertheless, the profile PR generated from the extrapolated but not superimposed PSS V,k , PR MPhas important practical significance. Above (see the last paragraph of the subsection "Extrapolation of PSS for route runs"), the possibility (and usefulness) of using this PSS to compare the severity of different routes in the route mix was already mentioned. With the help of profiles generated from these PSS, this task can now be solved in another way. PR calculated for the same measurement position or for the same measurement position from runs on different routes V、k or PR MP When the are superimposed on each other (e.g. in a graph), it is possible to draw a conclusion which route and which frequency range leads to a more severe profile in the vibration test than that of the other routes for the location(s) under consideration. A more severe profile means a greater (pseudo) damage. However, it should be noted that the running periods on the different routes are sometimes not equal, i.e. TR1 ≠ TR2 ≠ TR3, etc. Route Mix Profile PR generated from extrapolated and superimposed PSS V,k , PR MP Only the running period on the entire route mix and therefore all the pre-specified lifespans T of the components are LD They therefore form the main results of the profiling of the ASPEN RoMi method. However, for each profile, this result is not related to the corresponding specific position (a single profile PR V,k In the case of profile control at multiple corresponding positions (profile PR MP This only applies to profile control in the PR V,k , PR MP are the different positions (on the component or on its holder) that were used to generate

[0130] The assignment of profiles to groups a) to d) results from the definition of input data and calculation parameters.

[0131] The features from groups a) to d) are not contradictory. Each generated profile contains one feature from each group. For example, a test profile can be a sweep profile in terms of the type of vibration generated, an excitation profile in terms of the selection of locations for its control / application, and a profile for multi-point control in terms of the number of control locations. In this case, the running period covers only one route or the entire route mix at the same time. All these profiles can be used. The selection of one profile for the practical execution of the vibration test is made by the person in charge.

[0132] ASPEN Conversion

[0133] Here, equations are given to convert the PSS (i.e., the simulated damage experienced by a component in a running operation) calculated from a route run using the ASPEN method into damage-equivalent profile amplitudes. For this purpose, simple, non-iterative analytical equations (3.16) and (3.17) were developed for the ASPEN method. They are called the ASPEN transformation (AT) and were implemented in the PRGN algorithm (see above).

[0134] In both variations, AT applies for a simple Wöhler line, which is described by equation (3.8). As can be seen from equations (3.16) and (3.17), the desired profile height in this case is a function of the slope coefficient k WL It depends only on the abscissa or ordinate of the support point of the Wöhler line.

[0135] Which expression to use depends solely on the profile type to be generated. In this case, two profile types are distinguished:

[0136] a) Noise profile Such a profile requires stochastic vibration excitation and is described by an LDS. The AT for calculating the LDS height of a noise profile at frequency f is:

number

[0137] AT(3.16) is the LDS PSD of the stochastic oscillations at frequency f U Determine the magnitude of the damage at this frequency, D U D U From LDS PSD U to calculate the LDS PSD Ref Another stochastic oscillation (called the reference signal) with a damage value D Ref This relationship is used to generate the known LDS PSD Ref Generate a reference signal (here, a stochastic vibration signal) appropriate for the profile type using Ref (See the right part of the diagram in Figure 4.) This procedure is detailed above.

[0138] Calculated PSD in (3.16) U (f) is performed separately for each vibration frequency f, independent of the other frequencies. In the formula notation in the section describing the PRGN algorithm (see the "Description" section above), D U (f)=D U,k (f), PSD U (f)=PR U,k(f) (((3.16) is an arbitrary signal, so the index k is omitted), D Ref (f)=D r,EX (f).

[0139] Equation (3.16) is mathematically proven below, which is based on considering the relationship between the LDS and PSS of an ergodic normally distributed noise signal and its scaled copy.

[0140] b) Sweep profile Such a profile requires harmonic vibration excitation, which is described by the dependence of the amplitude S of such a harmonic vibration signal on its instantaneous frequency f, i.e., AFV S(f).

[0141] The AT for calculating the AFV of a sweep profile (consisting of a single sweep) at frequency f is:

number

[0142] AT(3.17) is the amplitude S of a simple harmonic oscillation at frequency f. U Determine the defined (required) damage value D at this frequency. U D U From AFV S U To calculate the amplitude SRef Another monoharmonic oscillation (called the reference signal) with a damage value D Ref This relationship is used to generate a known amplitude S Ref Generate a reference signal (here, a single sweep signal) that matches the profile type and calculate its damage value D at each frequency. Ref (see the right part of the diagrams in Figures 1 and 4), which will be explained further below. Calculation S in (3.17) U (f) is performed separately for each vibration frequency, independent of the other frequencies.

[0143] In the formula notation in the section where the PRGN algorithm is described (see the "Description" section above), D U (f)=D U,k (f), S U (f)=PR U,k (f) (((3.17) is an arbitrary signal, so the index k is omitted), D Ref (f)=D r,EX (f).

[0144] In addition to single sweep profiles, AT (3.17) is also applicable to the generation of the following profiles: i) Profiles for dwell time tests with monoharmonic excitation at a fixed frequency (fixed frequency sinusoidal). These can be considered as a special case of a single sweep profile where the sweep frequency remains constant. ii) Profiles with multi-sweep excitation.

[0145] In case ii), the restriction of the ASPEN method is that the frequency bands of the individual sweep tones of a multi-sweep profile must not overlap. This restriction means that even in the case of a multi-sweep profile, only one single sweep profile is always defined for each frequency. Therefore, equation (3.17) is valid in this case as well.

[0146] Equation (3.17) is further mathematically proven below. It is based on considering the relationship between the amplitude and PSS of a monoharmonic time signal with a fixed frequency and a scaled copy thereof.

[0147] Reference Signal Concept

[0148] As already mentioned above, all damage-based profiling methods encounter the problem of converting the damage numbers experienced by a component in running operation into damage-equivalent amplitudes in the test profile. The difficulty here is that the relationship between these two values ​​must be known in order to perform such a conversion non-recursively. In the ASPEN method, this problem is solved based on the concepts of reference signals and AT. AT, described by equations (3.16) and (3.17), was considered in the "ASPEN Transform" section. Here, the concept of reference signals is addressed.

[0149] Essentially, this method is based on the following: depending on the profile type, a specially defined time signal is generated independently, its PSS is calculated, and it is extrapolated to a specified period of the vibration test. See the right part of the diagram in Figure 4 (the generation of this signal can be internal or external to the method). The amplitude (or LDS) of this signal is predetermined, and its PSS is calculated. This procedure solves the above-mentioned problem, and the ratio between these two parameters, required for converting the PSS into a damage-equivalent profile amplitude, is therefore known at each frequency. The signal generated in this way is called a reference signal in the ASPEN method. Its type must necessarily correspond to the type of test profile to be created. For example, it must be generated as a sweep signal when calculating a sweep profile, or as a stochastic signal when calculating a noise profile. In detail, the concept of a reference signal for generating a test profile in the ASPEN RoMi method provides for the execution of the following steps:

[0150] a) Determining the profile type and parameters At the latest at this point in the profile generation, the type of test profile to be generated and the vibration test T VT (Note: For the FDDC evaluation of signals measured during a test run, the test frequency range f unt , …, f ob must already be selected first - see above). The types of profiles that can be generated by the ASPEN method are: Sweep profile (single sweep or multi-sweep) Profiles for dwell time tests (sinusoidal excitation at fixed frequency), and Noise profile (tests with stochastic vibration excitation).

[0151] b) Reference signal generation A reference signal is generated depending on the profile type and profile parameters selected above, i.e. Sweep Profile In a single sweep profile, the reference signal is applied over the entire test frequency range, f unt , …, f ob , as a single sweep of the selected type (linear, logarithmic), with amplitude S Ref , tuning rate R, and period t r It is generated with a period t r must contain an integer number of half-sweep cycles (this is the transition time from the lower frequency to the upper frequency of the swept sine wave signal). Ref can be chosen arbitrarily, but over the entire frequency range f unt , …, f ob must be constant over In a multi-sweep profile, the reference signal is Swp Individual sweep tones (n Swp Each sweep tone is defined in each frequency band and is divided into the following divisions [f Swp,1 ;f Swp,2], …, [f Swp,l-1 ;f Swp,l ], l=n Swp +1(f Swp,1 =f unt , f Swp,l =f ob ) over the entire test frequency range f unt , …, f ob In this case, the individual frequency bands must not overlap. Ref , and the tuning rate R must be chosen to be the same for each sweep tone (with respect to the tuning rate R, this condition ensures that all individual sweeps have the same sweep time for the individual frequency bands, i.e., are performed synchronously). Again, the period t r must be an integer number of half-sweep cycles. Amplitude S Ref is the entire frequency range f unt , …, f ob must be constant in Profile for residence time testing This is the simplest case, where the reference signal has a fixed frequency f Ref A selected period t r (preferably at least 1000 cycles of a sinusoidal vibration). This is the test frequency. The amplitude of the vibration, S Ref can be chosen arbitrarily, as for the sweep profile. Noise Profile The reference signal here has a constant power density PSD over the entire test frequency range. Ref The LDS PSD is generated as a stationary stochastic signal with Ref The value of can be chosen arbitrarily. The period of this oscillation t r , the actual LDS of the reference signal and the pre-specified PSD Ref In practice, the signal period t r =400s has been proven to be long enough. If such noise signals are generated using PC-based signal processing software (e.g., Matlab®, Famos, Labview (registered trademark), etc.), the random number generation function available therein can be used (e.g., in Famos, this function is called "random" and provides a digital white noise signal in the frequency band up to the Nyquist frequency).

[0152] c) PSS calculation After the reference signal r(t) is generated as described above, its PSS D r (f) is calculated. For this, the FDDC algorithm is again executed (see calculation block 13 in the right part of the diagram in Figure 4). The same calculation parameters must be used to calculate the PSS of the signal from the route mix run (see above). In particular, this concerns the configuration of the bandpass filter, as well as the classification and Wöhler line parameters.

[0153] d) PSS extrapolation PSS D calculated at the previous point r (f) is the period t of the generated reference signal r However, in order to calculate the damage equivalent profile in the ASPEN method, the period T VT The damage count (calculated) for all vibration tests of PSS D is required. r (f) is the period T VT This is done by a new execution of the SPEX algorithm, see calculation block 14 in the right part of the diagram in Figure 4, in which the extrapolation algorithm is depicted by equation (3.5). The extrapolation coefficient k r is determined according to (3.6). Extrapolated PSS D r,EX (f) is the specified duration T of the vibration test VTto form the result of all calculations for the reference signal. This PSS is further used in the PRGN algorithm for determining the profile amplitude (see section "Description", subsection "Description of the PRGN algorithm", a)).

[0154] Practical Tips Basically, the following possibilities are available for generating the reference signal: Generation as a real vibration signal on a vibration test table with an appropriate control system. The reference signal is only needed to calculate the relationship between its amplitude (or LDS) and the degree of damage for a specific type of test profile, so there is no need to set up a test object for this. The required vibration signal can be recorded on an empty shaker plate, i.e. without a test object. This applies regardless of whether the current test profile calculation is an excitation profile or a reaction profile. Generation as a fictitious digital signal in a PC-based signal processing tool (e.g. Matlab®, Famos, LabView®, etc.). This is an alternative, time-saving, cost-effective possibility, since it does not require a vibration test stand for vibration generation or a measurement system for signal recording. In this case, formally, the amplitude of the reference signal generated in this way can have any unit, or even no unit at all, since the signal is fictitious. However, for the ASPEN method to work correctly, it is absolutely necessary to assign to the ordinate of the generated reference signal the same unit as the ordinate of the signal from the route run for which the test profile is to be created. Correspondingly, the amplitude S of the digital reference signal to be generated Ref or LDS PSD Ref The units for must also be chosen, i.e. The amplitude S for generating the reference signal of the sweep type (when generating a single-sweep profile or a multi-sweep profile) or the type of harmonic oscillation with a fixed frequency (when generating a profile for a dwell time test) Ref The units of the signals from the route run for which the test profile is to be generated (e.g. acceleration values ​​- m / s 2 Units for S Ref etc.) LDS PSD for generating a white noise reference signal Ref The units of must be the same as the units of the noise profile to be generated (e.g., m / s 2 Unit PSD for acceleration values ​​measured in Ref -(m / s 2 ) 2 / Hz).

[0155] More Tips Care must be taken to ensure that the signal parameters used to generate the reference signal (e.g. sweep tuning rate, number of individual sweep tones and their frequency bands) are identical to those actually used in the subsequent execution of the test profile generated on the vibration test equipment for the vibration testing of the test object. The same reference signal, once generated, can be used to generate excitation and response profiles that are used together in the vibration test, referenced to the test object.

[0156] Further explanations / attachments

[0157] PSS extrapolation and superposition when there are multiple measurement data files for each driving route

[0158] Here we deal with the special case of PSS extrapolation and superposition in the ASPEN RoMi method, when there are multiple measurement data files per traveled route. This is the case, for example, when saving all measured sections on this route in one data file would result in too much data. Therefore, the measurement engineer may decide to trigger and stop measurements several times during the route travel. For the purposes of further explanation in this subsection, it is assumed that the data for each measurement interval is saved in a separate measurement data file. For example, the measurement data can be later split into several measurement data files in order to reduce the size of the measurement data files. The travel time to which the PSS of the time signals of this data file is extrapolated is specified separately in the SPEX module for each measurement file. Therefore, it is of course possible to calculate the total travel time T Ri Rather, the total running time T Ri must be divided among several sections of the route. As mentioned in the previous paragraph, there is then, by assumption, one measurement data file per route section. The sum of the travel times of all the subsections is the total travel time of the route T Ri The total running time should be as above. This explanation is illustrated by the following numerical example: Assume that the acceleration values ​​γ(t) are measured in two sections of a travel route i with durations of 1000 s and 1500 s and stored in two separate measurement data files. The signals in the two sections are expressed as γ i,1 (t) and γ i,2 (t) and the corresponding measurement period is t γi,1 and t γi,2 Therefore, t γi,1 =1000s,t γi,2 = 1500 s, and the total travel time on the route i under consideration within the route mix covering the entire lifespan is, for example, T Ri Assume =5000h. The operation of the ASPEN RoMi method using the FDDC and SPEX modules is as follows: First, two measurement files are fed to the FDDC module, and two signals are generated. γi,1 (t) and γi,2 (t) to two PSS D γi,1 (f) and D γi,2 (f) is calculated. These PSSs are also stored in individual data files, where they are fed to the SPEX module for extrapolation. Two PSSs D γi,1 (f) and D γi,2 (f) are again in two separate data files, so they are extrapolated separately. The extrapolation algorithm therefore Ri,1 , T Ri,2 These are the values ​​for each measurement signal γ i,1 (t) and γ i,2 (t) is the duration of the recorded trip on the section of route i. Their relationship to each other can be chosen arbitrarily, but it is only necessary to ensure that their sum is the total trip time on the route. That is, T Ri =T Ri,1 +T Ri,2 T Ri,1 , T Ri,2 A reasonable choice for is probably the same ratio as the measurement period, i.e., T Ri,1 / T Ri,2 =t γi,1 / t γi,2 If you select this, T Ri,1 =2000h, T Ri,2 = 3000h, and T Ri =T Ri,1 +T Ri,2 =5000h. In this regard, in the SPEX module, PSS D γi,1 is D γi,1,EX is extrapolated to D γi,2 is Dγ i,2,EX The extrapolated PSS D γi,1,EX is the running period T Ri,1 (on the first section of route i) and extrapolated PSS D γi,2,EX is the running period T Ri,2applies to the total travel time T on all routes i. Ri =5000h applied PSS D γi,EX is first calculated in the SPEX module by convolution, i.e., D γi,EX =D γi,1,EX +D γi,2,EX This is the main result of the SPEX calculation for the case study under consideration.

[0159] Proof of ASPEN transformation

[0160] The mathematical proof of AT, referring to equations (3.16) and (3.17), is based on the derivation of the relationship between the amplitude (or LDS) and PSS calculated for two scaled copies of an arbitrary time signal using the simple Wöhler line described by equation (3.8). This relationship is first established, followed by the proof of equations (3.16) and (3.17).

[0161] Relationship between the PSS of two scaled signals

[0162] An arbitrary time signal a(t) is considered. The PSS of this signal, determined by the FDDC algorithm of the ASPEN method for a simple Wöhler line (see above), is a (f). Multiplication of the signal amplitude by a fixed coefficient (constant) p leads to a new signal b(t) = a(t) p. The same Wöhler line can be used to calculate the PSS D of the signal b(t) using the FDDC algorithm. b When used to calculate (f), the PSS of two signals a(t) and b(t) are related to each other by:

number

[0163] Proof of formula (4.1) Assume that the signal a(t) is classified using the counting method, so that for each lasthorizon (i.e., amplitude) ai, there is an associated number of oscillation cycles N OP (ai) exists. i can also be the SMA if the counting method used determines the average value for each cycle in addition to the amplitude (e.g., the Rainflow method). i ;N OP (a i ) form the load spectrum. In Figure 8, this is shown schematically as a continuous dark brown curve. Each amplitude a i The damage is calculated according to (3.7).

number

number

number

number

number

number

number

[0164] Derivation of AT for noise profile cases

[0165] The mathematical derivation of equation (3.16) is based on the following consideration: the reference signal a(t) is the LDS PSD a (f) and PSS D a Assume that the noise is ergodic and normally distributed with (f). This PSS was determined using the FDDC algorithm described above. From a(t), a new (reference) signal b(t) is formed by multiplying the amplitude a(t) by a constant p, i.e., p: b(t) = a(t)·p. PSS D b (f) (default value) LDS PSD of the noise signal b(t) b (f) is required. The LDS of an ergodic signal a(t) of duration T is its Fourier transform FT a The magnitude of (f) can be calculated by squaring it as follows

[14] :

number

[14] . Therefore, b(t) also applies, i.e.,

number

number

number

number

number

[0166] Derivation of AT for the case of a swept profile

[0167] Due to the conditions of the ASPEN method mentioned above, it is sufficient to derive equation (3.17) only for the case of a single sweep profile. Such a profile determines a slip-frequency sinusoidal oscillation consisting of one tone. Furthermore, it can be assumed to be a monoharmonic oscillation with a fixed frequency for any sufficiently small selected time interval. This leads to a further simplification: AT (3.17) needs to be proved only for the case of a monoharmonic oscillation with a fixed frequency.

[0168] Such a monoharmonic (reference) signal a(t) with a fixed frequency f is considered here (in this case, the frequency f can be arbitrary). The AFV of a(t) is S a (f), PSS, D a This PSS is called (f). This PSS was determined using the FDDC algorithm described above. From a(t), a new (reference) signal b(t) is formed by multiplying the amplitude a(t) by a constant p, i.e., p: b(t) = a(t) p. PSS D b (f) (default value) of the AFV S of the signal b(t) b (f) is required. S b (f) and S a Since (f) is the amplitude of the two signals b(t) and a(t), it is clear from b(t)=a(t)·p that the following holds:

number

number

number

number

[0169] The invention is not limited to the described embodiments, the scope of protection is defined by the claims.

[0170] literature 1. Kohler, M.; Jenne, S.; Potter, K.; Zenner, H.: Loading and Load Assumptions in Operational Stability. Springer, 205S., 2012. 2. MBN 10438-2 "Fatigue strength of vibrating engine components - Safety measures, requirements and procedures for series production". Mercedes-Benz Factory Standard, January 2014, 22S ("Fatigue strength of vibrating engine components - Safety measures, requirements and procedures for series production"). 3.VW80200-1 Motoranbauteile“. Volkswagen AG Konzernnorm, Ausgabe 2009-03, 29S ("Engine Components" Volkswagen AG Konzern Standard) 4. DIN 30787-5 - Measurement and evaluation of mechanical-dynamic loads. Part 5: Derivation from test specifications. 09.2002, 20S ("Measurement and evaluation of mechanical-dynamic loads. Part 5: Derivation from test specifications") 5.Cornelis, B.; Dendas, B.; Carrella, A.: “Qualification testing of racecar equipment subject to engine-induced vibrations: how to derive a test profile using a mission synthesis procedure”, Siemens Industry Leuven. In: IMAC XXXV, Garden Grove, USA, 30.01.-02.02.2017 (URL: https: / / www.researchgate.net / publication / 313368832_A_Mission_Synthesis_procedure_for_Sine-on-Random_excitations_in_a_helicopter_application) (Qualification testing of racing car equipment subjected to engine-induced vibrations: A method for deriving test profiles using a mission synthesis procedure, Siemens Industries Leuven) 6. Halfpenny, A.: “Accelerated Vibration Testing Based on Fatigue Damage Spectra”. nCode International, 2009 (nCode publication, URL: https: / / www.ncode.com / images / GlyphWorks / Downloads / Whitepaper_nCode_AHP_AcceleratedVibrationTestingBasedonFatigueDamageSpectra_v2-Halfpenny.pdf) 7. Halfpenny, A.; Kihm, F.: “Mission Profiling and Test Synthesis Based on Fatigue Damage Spectra”. 9th International Fatigue Congress, Atlanta, USA, 2006. 8. Decker, M.; Kinscherf, S.; Hesse, R.; Dillinger, S.: “FatiResponse - Damage equivalence in vibration testing, test method for operational stability testing in the automotive industry”. DVM-Workshop “Test methods for damage equivalence in vibration testing”, Ottobrunn, 25-26.01.2017 (FatiResponse - Damage equivalence in vibration testing, test method for operational stability testing in the automotive industry). 9. Decker, M.: “Ableitung schadigungsaquivalenter Leistungsdichtespektren fur die Vibrationsprufung unter Verwendung von Schadigungs-Antwort-Spektren”. Dissertation, Heft 119, TU Darmstadt, Institut fur Stahlbau und Werkstoffmechanik, 133S, 2018 (Derivation of Damage Equivalent Power Density Spectra for Vibration Tests Using Damage Response Spectra) 10. Patent WO 98 / 14765 A “Method to Specify Random Vibration Tests for Product Durability Validation”, Ford Motor Company, 1998 (Method to specify random vibration tests for product durability validation) 11. Patent US 005565618A “Method to Specify Sinusoidal Vibration Tests for Product Durability Validation“. Ford Motor Company, 1996 12. Patent EP 3433593 B1 “Method and System for Accelerated Fatigue Damage Testing of an Object“. Siemens Industry Leuven, 2016 (Method and System for Accelerated Fatigue Damage Testing of an Object) 13. Patent DE 10236735 A1 "Method for generating equivalent noise profiles for vibration testing of vehicle components" BMW AG Munchen, 2002 (Method for generating noise profiles equivalent to driving damage for vibration testing of vehicle components) 14. Bendat, J; Piersol, A.: Random Data, Analysis and Measurements Procedures. John Wiley, 566S., 1986. 15. Haibach, E.: Betriebsfestigkeit. Verfahren und Daten zur Bauteilberechnung. Springer, 2. Auflage, 2002, 753S (Operational stability, methods and data for component calculations) [Explanation of symbols]

[0171] 1.1, 1.2, …, 1.n measurement signal 2.1, 2.2, …, 2.n Pseudo damage spectrum 3 Superimposed pseudo-damage spectrum 4 Test Profile (Route Mix) 5 Reference Signal 6. Extrapolated reference signal pseudo-damage spectrum 7.1, 7.2, …, 7.n Extrapolated pseudo-damage spectrum 8.1, 8.2, …, 8.n Test Profiles (Individual Routes) 9.1, 9.2, …, 9.n measurement data files 10 Pseudo Damage Calculation Parameters 11 Extrapolation and superposition calculation parameters 12 Test profile calculation parameters 13 Pseudo Damage Spectrum Calculation Component 14 Extrapolation Subcomponent 15 Superimposed Subcomponents 16 Test Profile Generation Component 17 Reference signal pseudo-damage spectrum 18 Reference signal extrapolation parameters Regarding the separate embodiment of the method in FIG. m Number of routes n the number of measurements recorded on each route MS i,j , i=1,…, m, j=1,…, n Measurement signal PS i,j , i=1,...,m, j=1,...,n pseudo damage spectrum ES i,j , i=1,...,m, j=1,...,n Extrapolated pseudo-damage spectrum SS j , j=1,...,n Extrapolated and superimposed pseudo-damage spectrum PP i,j , i=1,...,m, j=1,...,n Test profile (individual route) SPP j , j=1, ..., n Test profile (root mix) P1 Pseudo damage spectrum calculation process P2 Extrapolation and convolution processes P3 Test Profile Generation Process P4 Reference Signal Processing P5 Extrapolation Subprocess P6 Superimposed Sub-Process S1 Signal filtering step S2 Classification step S3 Transform Step S4 Partial damage contribution calculation step S5 All damage calculation step S6: Pseudo damage spectrum generation step

Claims

1. A method for generating a test profile (4), comprising: a pseudo damage spectrum calculation process (P1) in which a plurality of pseudo damage spectra (2.1, 2.2, ..., 2.n) are calculated from a plurality of measurement signals (1.1, 1.2, ..., 1.n) using a spectrum calculation algorithm within the pseudo damage spectrum calculation process (P1), generating a pseudo damage spectrum (2.1, 2.2, ..., 2.n) for each measurement signal (1.1, 1.2, ..., 1.n), each measurement signal (1.1, 1.2, ..., 1.n) having been recorded within the framework of a route travel of a test vehicle before the start of the method; an extrapolation process and superposition process (P2) including an extrapolation sub-process (P5) and a superposition sub-process (P6), wherein within the extrapolation sub-process (P5), each pseudo damage spectrum (2.1, 2.2, ..., 2.n) is multiplied by a proportionality constant to generate a plurality of extrapolated pseudo damage spectra (7.1, 7.2, ..., 7.n), and within the superposition sub-process (P6), the extrapolated pseudo damage spectra (7.1, 7.2, ..., 7.n) are added to generate a superposed pseudo damage spectrum (3); a test profile generation process (P3) which generates a test profile (P4) within the framework of said test profile generation process (P3); a reference signal processing process (P4) in which, within the framework of said reference signal processing process (P4), a reference signal pseudo-damage spectrum (17) is first calculated from a reference signal (5) using a spectrum calculation algorithm, and then said reference signal pseudo-damage spectrum (17) is multiplied by said proportionality constant to generate an extrapolated reference signal pseudo-damage spectrum (6), A method for generating a test profile (4), characterized in that the test profile (4) is generated based on the superimposed pseudo damage spectrum (3) and the extrapolated reference signal pseudo damage spectrum (6) within the framework of the test profile generation process (P3).

2. 2. The method of claim 1, wherein all test profiles calculated in the method computationally satisfy the principle of damage equivalence for the pseudo damage spectra (2.1, 2.2, ..., 2.n) calculated in the method.

3. 2. The method of claim 1, wherein the reference signal processing process (P4) runs in parallel with the pseudo damage spectrum calculation process (P1) and / or the extrapolation and convolution process (P2).

4. 4. The method according to claim 1, wherein each measurement signal (1.1, 1.2, ..., 1.n) depicts the time course of one and the same measurement value typically recorded on different routes (8.1, 8.2, ..., 8.n) of the route mix (4).

5. 5. The method according to claim 1, further comprising generating a plurality of test profiles (8.1, 8.2, ..., 8.n) based on a plurality of measurement signal sets within the scope of the method, wherein each measurement signal set depicts the time course of a particular measurement value and / or is obtained by measuring a signal of a particular measurement value on the route mix (4).

6. 6. The method according to any one of claims 1 to 5, characterized in that within the framework of the method (depending on the task, input data and setting parameters) a number of different types of test profiles (8.1, 8.2, ..., 8.n) or profiles with different characteristics are generated, in particular excitation profiles and response profiles, profiles for single-point and multi-point control, as well as profiles covering damage on individual routes or on the entire route mix (4).

7. The method according to any one of claims 1 to 6, wherein the pseudo damage spectrum calculation process (P1) comprises the following steps: a signal filtering step (S1), within which each measurement signal (1.1, 1.2, ..., 1.n) is filtered using a plurality of bandpass filters to generate a plurality of filtered measurement signals; a classification step (S2) in which a load spectrum is formed from each filtered measurement signal by means of a classification, typically by dividing the entire amplitude range of each filtered measurement signal into classes, and the number of vibration cycles is determined for each class of amplitude, preferably by a counting method; a transformation step (S3) in which, within the framework of said transformation step (S3), the possibly mean amplitudes of each vibration cycle are first transformed into damage-equivalent mean-free amplitudes, preferably using a Haigh diagram, and then said mean-free amplitudes are sorted in ascending order; a partial damage contribution calculation step (S4) in which the partial damage contribution is calculated for the number of vibration cycles at each damage equivalent mean value-free amplitude, preferably using Wöhler lines; a total damage calculation step (S5), in which, within the framework of said total damage calculation step (S5), for each filtered measurement signal, said partial damage contributions are summed to a total damage, whereby each total damage is referred to as a pseudo-damage number for the respective filtered measurement signal; and a pseudo damage spectrum forming step (S6) for forming a pseudo damage spectrum (2.1, 2.2, ..., 2.n) from the pseudo damage numbers by displaying the pseudo damage numbers as a function of the bandpass center frequency of a bandpass within the pseudo damage spectrum forming step (S6).

8. 8. The method according to claim 1, wherein within the framework of the test profile generation process (P3) a noise profile and / or a sweep profile is calculated.

9. 9. The method according to claim 1, characterized in that within the framework of the test profile generation process (P3), the conversion of damage to damage equivalent test amplitudes for generating sweep profiles is performed using a special transformation in equation (3.17), and the conversion of damage to damage equivalent LDS for generating noise profiles is performed using a special transformation in equation (3.16).

10. 10. The method according to any one of claims 1 to 9, characterized in that the entire structure and the entire calculation sequence of the method remain unchanged for the generation of a sweep profile or a noise profile, except for feeding respectively different types of reference signals to the input of the method and using different transformation formulas (3.17) or (3.16) in the test profile generation process (P3) of the method.

11. The method according to any one of claims 1 to 10, characterized in that the method is a computer-implemented method.

12. A system for carrying out the method according to any one of claims 1 to 11, said system being preferably suitable for at least partly carrying out and / or regulating and / or controlling the method for generating a test profile (4) according to any one of claims 1 to 11.

13. A computer program comprising, when executed on a computer, causing the computer to carry out the method according to any one of claims 1 to 11.

14. A computer readable medium comprising computer program code for carrying out the method according to any one of claims 1 to 11.

Images