Open-loop and / or closed-loop control unit for a test bench for carrying out a fatigue test, and method

Abstract

This disclosure describes an open-loop and / or closed-loop control unit (100) for a test bench for carrying out a fatigue test on a test body. The open-loop and / or closed-loop control unit (100) has the following components: an input interface (110), which is designed to detect a load signal (102), which indicates a load acting on the test body, and a deformation signal (104), which indicates a deformation of the test body; a computing unit (120), which is designed to determine damage energy (122) using the load signal (102) and the deformation signal (104); a parameter determination unit (130), which is designed to determine an operating parameter (132) of an actuator for applying a load to the test body depending on the damage energy (122); and an output interface (140), which is designed to provide an actuator signal (142) depending on the determined operating parameter (132).

Description

[0001] Fraunhofer Society...eV

[0002] P149145PC00

[0003] Control and / or regulation unit for a test bench for performing a fatigue test

[0004] The disclosure relates to a control and / or regulation unit for a test rig for performing a fatigue test on a test specimen, a test rig comprising the control and / or regulation unit, a method for performing the fatigue test on the test specimen, and a computer-implemented method for performing a numerical fatigue test on a test specimen. A standard procedure for investigating the fatigue strength and service life of a component under mechanical stress is fatigue testing. In this process, the component is cyclically loaded and unloaded to determine, for example, an acceptable service life, i.e., the number of load cycles at which mechanical failure of the component occurs under cyclic loading and unloading. A load cycle is defined as a complete period of loading and unloading.

[0005] Since fatigue testing is complex, the service life of intricate components is often determined using numerical simulation. Numerical simulation of components allows for obtaining information about a component's service life more cost-effectively and quickly than would be possible with a physical fatigue test. Input parameters for numerical simulation include the component's geometry, a time-dependent load function, and knowledge of the material properties of the materials from which the component is to be manufactured. However, determining these material properties often necessitates the investigation of these materials using a physical fatigue test. Once the material properties are determined, it is assumed that the service lives of components with different geometries can be determined purely numerically.

[0006] The so-called S-N curve is typically used as input for numerical simulation. This curve establishes a relationship between the material's tolerable number of load cycles and a given constant load amplitude. The S-N curve is also known as the fatigue life curve. Furthermore, a load-deformation relationship is usually used as input, indicating the material's deformation under a given load. Both relationships can be determined by fatigue testing a specimen containing the material under investigation at a constant load amplitude. When evaluating fatigue test results, a distinction must be made between two types of material behavior: materials with linear-elastic load-deformation behavior and materials with elastic-plastic load-deformation behavior.Linear-elastic load-deformation behavior means that the material behavior under cyclic loading and unloading is completely reversible, and that the correlation between load and deformation is determined by a constant factor over the entire service life. The required service life curves are generated using load-controlled S-N tests with constant load amplitudes and mean loads.

[0007] Elastic-plastic load-deformation behavior means that the material's behavior is not constant over its lifetime, i.e., from initial loading to failure, but rather the material undergoes work hardening or softening. Work hardening means that a greater load (e.g., force) is required to achieve the same deformation. Conversely, with softening material behavior, only a smaller load (e.g., force) is needed to achieve the same deformation. Due to these so-called transient processes, the shape of the load-deformation hysteresis usually changes continuously.

[0008] Design concepts that consider elastic-plastic material behavior predominantly assume a cyclically stabilized load-deformation behavior (stress-strain behavior) at half the material's service life. This means that the load-deformation relationship is assumed to be constant over the material's service life and is equated with the load-deformation relationship at half the service life, i.e., at half the tolerable number of cycles. It is assumed that differences in the load-deformation behavior in the first and second halves of the material's service life balance each other out. To determine the S-N curve, the tolerable number of cycles at a constant load or deformation amplitude is calculated for a multitude of different load or deformation amplitudes.

[0009] However, when comparing computational and physical lifetime determination, sometimes significant discrepancies have been observed between numerical prediction and actual measurement. Various approaches have been developed to reduce this discrepancy. So far, however, none of these approaches have consistently succeeded. The aim of this disclosure is therefore to present a new approach for determining material properties using stress testing and for determining lifetimes using numerical simulation.

[0010] According to a first aspect of this disclosure, this approach is realized in a control and / or regulation unit for a test bench for performing a fatigue test on a test specimen according to claim 1.

[0011] According to a second aspect, this approach is realized by a method for performing a fatigue test on a test specimen according to claim 10.

[0012] Furthermore, the approach according to a third aspect is realized by a computer-implemented method for performing a numerical fatigue test on a test specimen according to claim 14.

[0013] The devices and methods of the various aspects of this disclosure are described in detail below. First, the focus will be on the first aspect of this disclosure, namely the control and / or regulating unit.

[0014] The control and / or regulation unit according to the first aspect of this disclosure comprises an input interface, a computing unit, a parameter determination unit and an output interface.

[0015] The input interface is configured to acquire a load signal, indicating a load acting on the test specimen, and a deformation signal, indicating a deformation of the test specimen. The processing unit is configured to determine a damage energy using the load and deformation signals. Furthermore, the parameter determination unit is configured to determine an operating parameter for an actuator used to apply a load to the test specimen, depending on the damage energy. Finally, the output interface is configured to provide an actuator signal based on the determined operating parameter.The described control unit is based on the understanding that existing design concepts for fatigue testing assume that stress cycles on a given load or deformation horizon contribute the same amount to damage over the service life and are therefore damage-equivalent. This assumption cannot be fulfilled by either load- or deformation-controlled tests due to the stress-dependent changes in a material's response, i.e., its cyclically transient material behavior, and the underlying control variable. Consequently, significant discrepancies between experimentally and computationally determined service lives occur during damage assessment within the framework of fatigue strength verification.

[0016] The control unit described above therefore follows a novel approach: instead of load- or deformation-controlled fatigue testing, it performs damage-energy-controlled fatigue testing. Damage energy is used as a measure of the damage to the test specimen caused by loading and / or unloading. This makes it possible, for example, to keep the damage per cycle constant and thus determine an acceptable number of cycles as well as the relationship between, for example, stress and strain in a material as a function of accumulated damage.

[0017] The following describes embodiments of the control and / or regulation unit with additional, optional features.

[0018] In one embodiment of the control unit, the processing unit can be configured to acquire the load acting on the test specimen and its deformation as a time-correlated pair of values. Furthermore, the processing unit can be configured to determine the damage energy by integrating over an area bounded by a load-deformation curve formed by these value pairs. "Time-correlated" here means that a value for the load acting on a test specimen is combined with a corresponding value for the deformation present at the same time as a single value pair. The resulting curve thus reflects the temporal evolution of the load and deformation. The damage energy can be determined by integrating over an area bounded by this curve.

[0019] The literature presents various approaches to calculating damage energy. According to one definition, damage energy can be obtained by calculating the area bounded by a load-deformation hysteresis loop. In one embodiment of the control unit, the computing unit can therefore be configured in a first calculation mode to determine the damage energy by calculating the area bounded by a closed hysteresis curve. The calculated damage energy can thus be the damage energy per cycle.

[0020] According to another definition, the damage energy can be obtained by calculating an area spanned between a load-deformation coordinate system and the load-deformation curve between two load reversal points. Therefore, in one embodiment, the computing unit can be configured in a second calculation mode to determine the damage energy by calculating the area bounded by a segment of the load-deformation curve between two successive load reversal points and coordinate axes of a load-deformation coordinate system. In another embodiment, the origin of the load-deformation coordinate system can additionally be located at one of the load reversal points, preferably the earlier one.Furthermore, the coordinate system can be oriented such that a deformation coordinate axis of the load-deformation coordinate system points in the direction of a time-dependent progression of the load-deformation curve. The calculated damage energy can therefore be the damage energy per half cycle.

[0021] To determine the damage energy by calculating the area, in one embodiment of the control and / or regulation unit the computing unit can be additionally designed to calculate the area using the upper, middle, and lower sum approach, the trapezoidal approach and / or Simpson's rule.

[0022] In one embodiment of the control unit, the parameter determination unit can be configured to determine a reference value deviation as a function of the damage energy and a predetermined damage energy reference value. Furthermore, the parameter determination unit can be configured to determine the operating parameter as a function of the reference value deviation. This allows for a control loop in which the load exerted on the test specimen by the actuator is adjusted so that damage energy corresponding to the damage energy reference value is introduced through loading and / or unloading. Therefore, in one embodiment, the parameter determination unit can be configured to adjust the operating parameter as a function of the reference value deviation in such a way as to reduce the reference value deviation.The damage energy reference value can be constant or time-varying. It can also refer to a full or half cycle of vibration.

[0023] The term "load" is defined differently in the literature. Within the scope of this disclosure, the term "load" can be understood both as a stress on the test specimen by means of forces and moments (bending moment, torsional moment) and as a strain, i.e., a stress on the test specimen. Stresses can be, for example, in the form of a nominal stress, a technical stress, and / or a true stress of the test specimen. Therefore, in one embodiment of the control and / or regulation unit, the load indicated by the load signal can be in the form of a force acting on the test specimen, a moment acting on the test specimen, a nominal stress, a technical stress, and / or a true stress of the test specimen.

[0024] Furthermore, in one embodiment of the control and / or regulation unit, the deformation of the test body indicated by the deformation signal can be in the form of an elongation, a shear, a twist angle and / or a vibration displacement of the test body.

[0025] Furthermore, the operating parameter can include, for example, a power and / or a frequency specification.

[0026] Furthermore, the damage energy can include the energy absorbed by the test specimen through elastic deformation. Additionally, the damage energy can also include the energy absorbed by the specimen through plastic deformation.

[0027] The control unit according to one of the embodiments described above can be used in a test rig for performing a fatigue test on a test specimen. Such a test rig will be described below.

[0028] The test rig comprises a holding device designed to accommodate the test specimen, and at least one actuator that can be mechanically connected to the test specimen via the holding device and is designed to apply a load to the test specimen. The test rig also includes a load sensor and a deformation sensor. The load sensor is designed to provide a load signal indicating a load applied to the test specimen. The deformation sensor, on the other hand, is designed to provide a deformation signal indicating a deformation of the test specimen.Furthermore, the test bench comprises a control and / or regulation unit according to one of the described embodiments, whose input interface is connected to the load sensor and the deformation sensor in order to detect the load signal and the deformation signal, and whose output interface is connected to the at least one actuator in order to provide the actuator signal to the at least one actuator.

[0029] The following describes embodiments of the test bench with additional, optional features.

[0030] In one embodiment of the test rig, the at least one actuator can be configured to apply a load to the test specimen by means of forced excitation. The actuator can be, for example, a hydraulic, pneumatic, piezo-ceramic, and / or electrodynamic actuator. Furthermore, the actuator can comprise a spindle drive, a linear motor, and / or an eccentric.

[0031] In another embodiment, the at least one actuator can be designed as a resonance pulser, which is configured to subject the test body to a load via a mechanical resonant circuit with the test body as a spring with imbalance and / or electromagnetic excitation and / or ultrasonic excitation.

[0032] Furthermore, in one embodiment of the test bench, the load sensor can be a force sensor and / or a torque sensor. The load sensor can also include a force transducer or load cell, a scale, a load cell, a pressure sensor, a force-measuring bolt, a bending beam, and / or an acceleration sensor.

[0033] In another embodiment, the deformation sensor can be a strain sensor, preferably a strain gauge and / or an extensometer. Alternatively or additionally, the deformation sensor can also be an optical sensor, preferably a fiber Bragg grating-based sensor, a Brillouin optical time domain-based sensor, an optical frequency range reflectomerism-based sensor, and / or an image correlation sensor.

[0034] In general, the test specimen can be any body to be examined. This includes, for example, material specimens specifically used to investigate material properties. The test specimen can also be a component or other mechanical structure. A test setup can, for example, comprise the test rig according to one of the previously described embodiments and the test specimen.

[0035] To determine the damage energy, the load and deformation must be selected such that they correspond to a stress direction of the test specimen. For axial tensile-compressive loading, the load and stress directions are identical. For loads that cause bending, the load and deflection directions do not correspond to the stress direction and are therefore not suitable for determining the damage energy.

[0036] The test rig can therefore be set up in such a way that a change in length of the test specimen occurring under load and / or a stress generated by the load within the test specimen occurs along a direction that corresponds to the direction in which the test specimen is subjected to the load.

[0037] Finally, regarding the first aspect, a procedure for operating a control unit for a test rig to perform a fatigue test on a test specimen will be described below. This procedure comprises the following steps:

[0038] Capturing a load signal indicating a load acting on the test specimen,

[0039] Capturing a deformation signal that indicates a deformation of the test specimen,

[0040] Determining damage energy using the load signal and the deformation signal,

[0041] Determining an operating parameter of an actuator for applying a load to the test specimen as a function of the damage energy, and

[0042] Providing an actuator signal depending on the determined operating parameter.

[0043] The following section will describe in more detail the procedure for performing a fatigue test on a test specimen in accordance with the second aspect of this disclosure.

[0044] The method comprises the following steps: cyclic loading and unloading of the test specimen by means of at least one actuator to generate a large number of vibration cycles, and repeated adjustment of an operating parameter of the at least one actuator, preferably by means of a control loop, so that a damage energy experienced by the test specimen through the cyclic loading and unloading corresponds to a predetermined damage energy reference value.

[0045] The method shares the advantages with the control and / or regulating unit according to the first aspect of this disclosure.

[0046] The following describes embodiments of this method with additional optional features.

[0047] In one embodiment of the method, the repeated adjustment of an operating parameter of the at least one actuator can be carried out by the following method steps:

[0048] Capturing a load signal indicating a load acting on the test specimen,

[0049] Detecting a deformation signal that indicates a deformation of the test specimen,

[0050] Determining damage energy using the load signal and the deformation signal,

[0051] Determining an operating parameter of at least one actuator as a function of the damage energy and the specified damage energy, and

[0052] Providing an actuator signal depending on the determined operating parameter at the at least one actuator.

[0053] Furthermore, the procedure can be combined with the approach already described in relation to the control and / or regulation unit, for example to determine the damage energy.

[0054] Furthermore, in one embodiment of the method, the specified damage energy reference value can be a constant damage energy reference value. In this embodiment, the method can also include the additional step of determining a tolerable number of load cycles for the test specimen at the constant, specified damage energy reference value. By determining the tolerable number of load cycles at constant damage energy per load cycle, a better assessment can be made of how many load cycles a test specimen can withstand before material fatigue than is the case with tests using, for example, a constant load amplitude.

[0055] In addition, in variants of this embodiment, the method can be repeated for a multitude of different predetermined damage energy reference values ​​in order to determine a respective tolerable number of load cycles for each of the multitude of predetermined damage energy reference values. From this multitude of tests, a service life curve can then be determined from the relationship between tolerable number of load cycles and predetermined damage energy per load cycle. Within the scope of this disclosure, the relationship between tolerable number of load cycles and damage energy is also referred to as a modified Wöhler curve, based on the Wöhler curve determined at constant load amplitude.

[0056] In a further embodiment of the method, deformation of the test specimen and the load acting on it can be detected during cyclic loading. Furthermore, in this embodiment, using the detected deformation, the detected load, and the predetermined damage energy, a relationship between the load acting on the test specimen and its deformation can be determined as a function of the accumulated damage energy of the test specimen. The accumulated damage energy can be defined as the damage energy that the test specimen has accumulated up to a given time over all load cycles, i.e., over all loading and unloading cycles. This method allows for more precise information on the change in the load-deformation relationship as a function of the accumulated damage energy.Here too, the specified damage energy reference value can be a constant damage energy reference value.

[0057] As an alternative to a constant, predetermined damage energy reference value, in one embodiment of the method the predetermined damage energy reference value can be a time-varying damage energy reference value corresponding to a stress-time function. This is advantageous when a test specimen is to be subjected to a vibration cycle on the test rig, which, with regard to damage, is similar to a vibration cycle that would be expected when using the test specimen under operating conditions. This can be particularly advantageous if the test specimen is a mechanical component.

[0058] In the following, a computer-implemented method for performing a numerical fatigue test on a test specimen according to the third aspect of this disclosure will be described.

[0059] The process includes the following steps:

[0060] Providing a test specimen geometry that specifies a spatial extent of the test specimen;

[0061] Providing a load-deformation mapping for a material of the test specimen, which specifies a mapping between a load on the material and a deformation of the material;

[0062] Providing a damage energy-cycle number assignment for the material, which specifies an assignment between a tolerable number of cycles at a constant damage energy per cycle;

[0063] Providing a load-time function that specifies a time-based load distribution to be applied to the test specimen; for one or more test specimen sections of the test specimen, determining a time-based profile of a local damage energy that indicates damage to the respective test specimen section caused by the loading, whereby a time-based profile of a local load and a time-based profile of a local deformation of the respective test specimen section under loading is determined according to the load-time function and using the test specimen geometry and the load-deformation mapping.

[0064] Determining a time course of a local damage energy using the time course of the local load and the time course of the local deformation; and

[0065] Determining a lifetime for the one or more test specimen sections using the time course of the damage energy and the damage energy-cycle count assignment.

[0066] The local load and the local deformation indicate how large the load is that a respective section of the test specimen is subjected to when the test specimen is subjected to a load according to the load-time function, or how large the respective deformation of the respective section of the test specimen is due to the load.

[0067] The specified method has the advantage that by taking into account the local damage energy of a test specimen section, a more accurate prediction of the service life of the test specimen sections can be made than is the case if only the local load and the local deformation are taken into account.

[0068] The following describes embodiments of the computer-implemented method with additional, optional features.

[0069] In one embodiment of the method, the load-deformation mapping can be given as a stress-strain mapping, and the local load and local deformation can be calculated as local stress and local strain, respectively.

[0070] Furthermore, in one embodiment of the method, the load-deformation relationship can be dependent on accumulated damage energy. In this embodiment, a locally accumulated damage energy for the respective test specimen section can also be determined based on the temporal profile of the local damage energy. Moreover, when determining the temporal profile of the local load and / or the local deformation, the dependence of the load-deformation relationship on the accumulated damage energy can be taken into account. By considering the dependence of the load-deformation relationship on the damage to the material in the respective test specimen section, an even more realistic modulation of the test specimen behavior can be achieved.The accumulated damage energy describes how much total damage energy has already been introduced into the respective test specimen section at a given time through loading and unloading. In one embodiment of the method, the damage energy can also be determined via the area of ​​a hysteresis curve and / or via the area between the hysteresis east and the load-deformation coordinate system, as already described above with regard to the control unit.

[0071] Furthermore, in embodiments, the service life can be determined by linear damage accumulation according to Palmgren and Miner (Miner rule), whereby for a multitude of values ​​for the local damage energy, the number of cycles with this value of local damage energy is determined, and the number of cycles is compared to the tolerable number of cycles from the damage energy-cycle assignment for the respective value of local damage energy.

[0072] Furthermore, the described computer-implemented method can be given in the form of a computer program, wherein the computer program includes instructions which, when the program is executed by a computer, cause it to execute the above-described method in one of its embodiments.

[0073] Furthermore, the computer-implemented method can also be in the form of a computer-readable storage medium, wherein the computer-readable storage medium includes instructions which, when executed by a computer, cause it to execute the method described above in one of its embodiments.

[0074] Furthermore, the provided damage energy-revolution number assignment may have been determined according to one of the preceding methods.

[0075] Furthermore, the damage energy-cycle number assignment may have been determined in a physically carried out procedure in which a test specimen was subjected to a load in such a way that a change in length of the test specimen occurring under load and / or a stress generated by the load within the test specimen occurred along a direction corresponding to the direction in which the test specimen was subjected to the load.

[0076] Furthermore, the damage energy can include the energy absorbed by the material through elastic and / or plastic deformation.

[0077] The following examples will be explained with reference to the accompanying figures. First, a brief overview of the figures will be given:

[0078] Fig. 1 shows a control and / or regulation unit for performing a fatigue test on a test specimen,

[0079] Fig. 2a shows a load-deformation graph to illustrate a first method for calculating damage energy,

[0080] Fig. 2b shows a load-deformation graph to illustrate a second method for calculating damage energy,

[0081] Fig. 3 shows the calculation of the damage energy for a hysteresis leak according to the second calculation method.

[0082] Fig. 4 shows a test rig for performing a fatigue test on a test specimen, which includes the control and / or regulation unit shown in Fig. 1,

[0083] Fig. 5 shows a schematic drawing of a method for performing a fatigue test on a test specimen,

[0084] Fig. 6 shows an illustration of a control loop as it can be used within the framework of the method according to Fig. 5,

[0085] Fig. 7 shows a measurement diagram 600 for a lifetime curve, which was determined using the method from Fig. 5, and

[0086] Fig. 8 shows a procedure flow for a computer-implemented method for performing a numerical fatigue test on a component.

[0087] First, Figures 1 to 4 will be used to explain embodiments of the control and / or regulation unit and the corresponding test rig. Figures 5 to 8 will then be used to illustrate examples of the method for performing a fatigue test on a test specimen and the computer-implemented method for performing a numerical fatigue test on a test specimen, specifically a component.

[0088] Here, as in the rest of the description of the figures, the same reference symbols are used wherever possible to represent an object in several figures.

[0089] Fig. 1 shows a control and / or regulation unit 100 for performing a fatigue test on a test specimen.

[0090] The control and / or regulation unit comprises an input interface 110, a computing unit 120, a parameter determination unit 130 and an output interface 140.

[0091] The input interface 110 is designed to detect a load signal 102, which indicates a load acting on the test specimen, and a deformation signal 104, which indicates a deformation of the test specimen.

[0092] The computing unit 120 receives the load signal 102 and the deformation signal 104 and is designed to determine a damage energy 122 using the load signal 102 and the deformation signal 104.

[0093] The parameter determination unit 130 is designed to receive the determined damage energy 122 and to determine an operating parameter 132 of an actuator for applying a load to the test body as a function of the damage energy 122.

[0094] The output interface 140 receives the operating parameter 132 and is configured to provide an actuator signal 142 depending on the determined operating parameter 132.

[0095] In the illustrated embodiment, the load signal 102 indicates the load acting on the test specimen in the form of a force acting on the test specimen. In other embodiments, however, the load can additionally or alternatively be specified as a moment acting on the test specimen, such as a bending moment or a torsional moment. It is also possible for the load to be specified in the form of a stress, such as a nominal stress, a technical stress, and / or a true stress.

[0096] The deformation signal can also be presented in various forms. In the embodiment shown in Fig. 1, the deformation is in the form of strain. This strain includes both positive strain (extension) and negative strain (contraction). In other embodiments of the control unit, however, the deformation can also be presented, for example, as shear, a twist angle, and / or a vibration displacement of the test specimen. The way the load and deformation are presented can thus be adapted to the specific parameters of a given fatigue test.

[0097] Optionally, the control and / or regulation unit 100 can also include a storage unit 150, as is the case for the control and / or regulation unit 100 according to Fig. 1.

[0098] The storage unit 150 is configured to provide a damage energy reference value 152. In the example of the damage and / or control unit 100, this is received by the parameter determination unit 130. Furthermore, the parameter determination unit 130 is configured to determine a reference value deviation between the damage energy reference value 150 and the determined damage energy 122 and to determine the operating parameter as a function of the reference value deviation.

[0099] The damage energy reference value 152 can be stored as a constant damage energy reference value in the storage unit 150. Alternatively, a time-varying damage energy reference value can also be stored. As an alternative to storing the damage energy reference value 152 in the storage unit 150, the damage energy reference value 152 can also be received via an interface.

[0100] In the control and / or regulation unit 100, the damage energy 122 is used as an indicator of the damage to the test specimen. The damage energy can be determined in two different ways. This will be explained below with reference to Figures 2a and 2b.

[0101] In the first case, the damage energy is determined over an area of ​​a closed load-deformation hysteresis curve, as is first explained with reference to Fig. 2a.

[0102] Fig. 2a shows a load-deformation graph 200A to illustrate a first calculation method for damage energy.

[0103] A deformation is plotted on the abscissa 202 of graph 200A, and the load is plotted on the ordinate 204 of graph 200A. The time course 210 of load and deformation corresponds to a hysteresis curve, as is typically observed in elastic-plastic material behavior. The hysteresis curve 210 extends between load reversal point 214 and load reversal point 216. A segment of the load-deformation curve 210 that connects two temporally adjacent load reversal points is often referred to as a hysteresis branch. For Fig. 2a, this results in two hysteresis branches 218 and 220. The area of ​​the hysteresis curve 210 is indicated as a hatched region 212 in the load-deformation graph 200. By integrating the load-deformation curve 210, an area 212 of the hysteresis curve 210 can be determined, which corresponds to the damage energy according to the first calculation method.The damage energy determined in this way corresponds to damage suffered by the test body during an oscillation cycle.

[0104] Alternatively, it is also possible to calculate the damage energy for half cycles. This is defined as the integral over a hysteresis angle between two successive load reversal points, as illustrated in Fig. 2b.

[0105] Fig. 2b shows a load-deformation graph 200B to illustrate a second method for calculating damage energy.

[0106] The load-deformation hysteresis curve 210, already shown in Fig. 2a, is also shown in Fig. 2b. Fig. 2b also shows the hysteresis branches 218 and 220. Furthermore, a coordinate system 240, whose origin lies at the load reversal point 214, is indicated for integration over hysteresis branch 218. Integration over hysteresis branch 220 is performed in the coordinate system 250, whose origin lies at the load reversal point 216.

[0107] The coordinate system 240 is simply shifted relative to the original coordinate system formed by abscissa 202 and ordinate 204. Abscissa 242 of coordinate system 240, over which the deformation is plotted, and ordinate 244 of coordinate system 240, over which the load is plotted, are thus parallel to abscissa 202 and ordinate 204, respectively. Coordinate system 250 is shifted relative to the coordinate system formed by abscissa 202 and ordinate 204 and is additionally rotated by 180°. The abscissa 252 of coordinate system 250, over which the deformation is plotted, and the ordinate 254 of coordinate system 250, over which the load is plotted, are thus antiparallel to the abscissa 202 and the ordinate 204, respectively. The temporal evolution of the respective hysteresis branch therefore points in the same direction as the abscissas 242 and 252 in both coordinate system 240 and coordinate system 250.

[0108] By rotating the coordinate system 250, it is possible to transform the damage energies for two hysteresis branches of a hysteresis curve, determined using the second calculation method, into the damage energy of a closed hysteresis curve, as determined using the first calculation method. The relationship is as follows:

[0109] W N = WHI + W H2 - (Ae AL).

[0110] Here, WN is the damage energy of the closed hysteresis of a vibration cycle, WHI and WH2 are the damage energies of one of the hysteresis branches, Ae is a deformation amplitude, and AL is a load amplitude of the hysteresis curve.

[0111] The advantage of the second calculation method lies in the fact that a service life curve can be expressed as a correlation between half-cycles and damage energies, which can be advantageous in service life assessment. Firstly, the second calculation method allows for the determination of damage energies even for small cycles, whose hysteresis loops cannot be measured at the macro level in cyclic testing due to measurement limitations, but are present at the micro level. Secondly, within the framework of a numerical stress simulation, the load paths from load reversal point to load reversal point, and thus half-cycles, must be simulated, since hysteresis loops only rarely lead directly to closed hysteresis loops under operating loads.

[0112] Regardless of the advantages and disadvantages of both approaches, the calculation of the area proceeds in two steps using both methods. In the first step, pairs of values ​​are created, consisting of a value for the load and a value for the deformation.

[0113] The load acting on the test specimen and its deformation are typically not continuously measured by sensors, but rather provided as discrete values ​​over time. Therefore, in this embodiment, the processing unit 120 of the control unit 100 is additionally configured to acquire the load acting on the test specimen and its deformation as a time-correlated pair of values. In other words, this means that the processing unit 120 is configured, based on given discrete values ​​for load and deformation, to generate pairs of values ​​consisting of a load value and a deformation value that occurred simultaneously on the test specimen.

[0114] In the second step, the area of ​​the corresponding surface is then determined. In the embodiment shown in Fig. 1, the computing unit 120 therefore has a first calculation mode and is configured in this calculation mode to calculate the area of ​​the surface bounded by a closed hysteresis curve in order to determine the damage energy. This corresponds to the first calculation method for calculating the damage energy. In addition, in the embodiment shown here, the computing unit 120 has a second calculation mode and is configured in this calculation mode to calculate the area of ​​the surface bounded by a segment of the load-deformation curve between two temporally successive load reversal points and coordinate axes of a load-deformation coordinate system in order to determine the damage energy.This corresponds to the second method of calculation for calculating the damage energy.

[0115] Details of this area calculation will be described below using Fig. 3 as an example for the second calculation mode.

[0116] Fig. 3 shows the calculation of the damage energy for a hysteresis leak according to the second calculation method.

[0117] Figure 3 shows the coordinate system 240 and, for simplification, only the hysteresis east 218 of the hysteresis curve 210. The curve representing the hysteresis east 218 in Figure 3 is shown with a dashed line to indicate that it is only for illustrative purposes. The data actually used by the processing unit 120 for the area calculation are the pairs of values, each represented as a circle in Figure 3 and labeled with the reference symbols 360A-L, where the value pair 360A coincides with the load reversal point 214 and the value pair 360L coincides with the load reversal point 216.

[0118] According to the second calculation method, in the second calculation mode of the computing unit 120, the area of ​​the surface bounded by the abscissa 242 of the coordinate system 240 and the hysteresis east 218 is calculated to determine the damage energy. Since the course of the hysteresis east 218 is given by the discrete pairs of values ​​360A-L, the Riemann sum is suitable for calculating the area, whereby different shapes, e.g., rectangles or trapezoids, can be used as basis functions for the approximate determination of the damage energy.

[0119] The use of rectangles is suitable for the approximate calculation of the damage energy of a hysteresis branch 218. When using rectangles, a distinction can be made between upper, middle, and lower sums.

[0120] However, due to the expected shape of the hysteresis branch, the trapezoidal approach can also be advantageous, as it often approximates the area better than the rectangular approach. Alternatively, Simpson's rule can also be used to calculate the area.

[0121] The accuracy of the approximated damage energy depends on the measurement rate of load and deformation. Accordingly, the limiting case analysis when transitioning from the Riemann sum to the integral can be performed by increasing the measurement rate.

[0122] The described procedure for determining the damage energy results in a quasi-continuous signal that indicates the damage energy and reflects changes in the input variables load and deformation. This signal therefore fulfills the requirement to be used as a control variable in cyclic tests.

[0123] The control and / or regulation unit 100 can be used in a test rig to perform a fatigue test on a test specimen. An embodiment of such a test rig will be described below with reference to Fig. 4.

[0124] Fig. 4 shows a test stand 400 for carrying out a fatigue test on a test specimen 410, which includes the control and / or regulation unit 100 shown in Fig. 1.

[0125] The test stand 400 comprises a frame 402, which includes an upper cross member 402A and a lower cross member 402B.

[0126] Furthermore, the test stand 400 includes a holding device 420, which is configured to receive the test specimen 410 and comprises an upper holding end 410A and a lower holding end 410B. A load sensor 430 is mechanically connected in the force flow from the frame 402, actuator 440, and holding device 420 in such a way that it indicates a load applied to the test specimen 410. It is advantageous, but not essential, for the load sensor to be arranged on a fixed side of the test stand. This is the case in the illustrated embodiment. Specifically, the upper holding end 420A of the holding device 420 is mechanically connected to the crossbeam 402A via the load sensor 430. Furthermore, the load sensor 430 is configured to provide a load signal 102 indicating a load applied to the test specimen 410.

[0127] Furthermore, the test stand 400 has an actuator 440, which in this example is a hydraulic cylinder. The hydraulic cylinder 440 is connected via its piston 440A to the lower holding end 420B of the holding device 420 and is designed to apply a load, in this example optionally a tensile or a compressive force, to the test specimen via the lower holding end 420B.

[0128] The actuator 440 is arranged in the test rig 400 such that the tensile or compressive force generated by the actuator acts in a perpendicular direction. Furthermore, the holding device 420 of the test rig 400 is designed such that any change in length of the test specimen 410 occurring under load and any stress generated within the test specimen 410 by the load also occur perpendicularly, i.e., along the direction of the acting force.

[0129] Furthermore, the test rig 400 includes a deformation sensor 450, which in the example of the test rig 400 shown in Fig. 4 is a strain gauge and is designed to provide a deformation signal 104 that indicates a deformation of the test specimen 410 in the form of a positive or negative strain.

[0130] Finally, the test rig 400 also includes the control and / or regulation unit 100, whose input interface is connected to the load sensor 430 and the deformation sensor 450 to receive the load signal 102 and the deformation signal 104, and whose output interface is connected to the hydraulic cylinder 440 to provide the actuator signal 142 to the hydraulic cylinder 440. The test specimen 410 shown in Fig. 4 is a material specimen for investigating the mechanical properties of a material. However, the test specimen 410 can also be a component or similar item to be tested.

[0131] Using the test rig shown in Fig. 4, a method for carrying out a fatigue test on a test specimen can be performed, which will be described in more detail below with reference to Figs. 5 to 7.

[0132] Fig. 5 shows a schematic drawing of method 570 for performing a fatigue test on a test specimen.

[0133] Method 570 comprises two steps that are performed alternately or simultaneously. In a first step 580, a test specimen is cyclically loaded and unloaded by means of at least one actuator to generate a large number of vibration cycles. Parallel to the first step, in a second step 590, an operating parameter of the at least one actuator is repeatedly adjusted so that the damage energy experienced by the test specimen due to the cyclic loading and unloading corresponds to a predetermined damage energy reference value. The adjustment of the operating parameter can be carried out by means of a control loop. An example of such a control loop is described in more detail with reference to Fig. 6.

[0134] Fig. 6 shows an illustration of a control loop 500, as it can be used in the procedure according to Fig. 5.

[0135] The control loop 500 shown in Fig. 6 begins with the specification of a predetermined damage energy reference value 502. Depending on this damage energy reference value 502, an operating parameter of the at least one actuator is determined by means of a parameter determination unit. This parameter is transmitted as an actuator signal 506 to another actuator, which then applies stress to the test specimen according to the actuator signal 506. The test specimen and actuator are combined here as a test rig 530. The actual damage energy 512 introduced into the test specimen by the actuator can then be determined by means of sensors.By comparing the specified damage energy 502 with the actually measured damage energy 512, a damage energy deviation 504 can then be determined using a comparison unit 510 and, based on this damage energy deviation 504, generally a difference value between damage energy and damage energy reference value, the operating parameters can be redefined.

[0136] The actual damage energy introduced by the actuator is often not directly measurable. However, it is possible, for example as with the control and / or regulating unit 100, to determine a load acting on the test specimen and the resulting deformation, and to provide this as a measurement signal 508 to a conversion unit 540. This conversion unit 540 then calculates the actual damage energy 512 from the measurement signal. The conversion unit 540 and the comparator unit 510 of Fig. 5 are implemented in the control and / or regulating unit 100 by the processing unit.

[0137] This procedure allows for the determination of various properties of the test specimen. One example is the determination of the test specimen's lifetime curve. This is illustrated in Fig. 7.

[0138] Fig. 7 shows a measurement diagram 600 for a lifetime curve, which was determined using the method from Fig. 5.

[0139] The procedure illustrated in Fig. 5 was carried out several times for different predetermined constant damage energy reference values. Furthermore, the tolerable number of load cycles was determined for each run, i.e., the number of load cycles at a constant predetermined damage energy that led to failure of the test specimen.

[0140] Measurement diagram 600 shows an evaluation of these fatigue tests.

[0141] The abscissa 602 of measurement diagram 600 indicates the tolerable number of stress cycles. The ordinate 604 indicates the specified damage energy. Measurements were carried out for a total of four different values ​​of the specified damage energy, the results of which are shown as measurement point sets 610A-D in measurement diagram 600. TI

[0142] In a final step, it is possible to determine a service life curve 620 by fitting the measuring point sets 610A-D, which reflects a functional relationship between damage energy and tolerable number of load cycles. In the literature, the relationship between constant load amplitude and tolerable service life is referred to as the S-N curve. Therefore, the relationship between damage energy and tolerable number of load cycles is also referred to as a modified S-N curve within the scope of this disclosure.

[0143] Alternatively or in addition to determining the modified Wöhler curve, it is also possible to determine a dependence of a load-deformation relationship of the test specimen on the total damage energy introduced into the test specimen, i.e. the accumulated damage energy, using the method illustrated in Fig. 5.

[0144] Furthermore, it is also possible to specify the damage energy as a function of time. This allows fatigue tests to be carried out on a test specimen under conditions that correspond to the expected damage energies.

[0145] By determining material parameters as a function of the damage energy, numerical simulations of test specimens, e.g., components, can be made more realistic. A method based on this idea is described below with reference to Fig. 8.

[0146] Fig. 8 shows a process flow for a computer-implemented method 700 for performing a numerical fatigue test on a component. The method comprises six steps.

[0147] In step 702, a component geometry is provided that specifies the spatial extent of the component.

[0148] In step 704, a load-deformation mapping is provided for a material of the component, specifying a relationship between a load on the material and a corresponding deformation of the material. In step 706, a damage energy-cycle number mapping is provided for the material, specifying a relationship between a tolerable number of cycles at constant damage energy.

[0149] In step 708, a load time function is provided, which specifies a time-based load distribution with which the component is to be subjected.

[0150] Subsequently, in a two-stage approach, a time course of a local damage energy, indicating damage to the respective component section caused by loading, is determined for one or more component sections of the component. The two-stage approach comprises a step 710 and a step 712, as illustrated in Fig. 8 by the dashed outlines of steps 710 and 712.

[0151] In a first stage, in step 710, a time course of a local load and a time course of a local deformation of the respective component section under load of the component are determined according to the load-time function and using the component geometry and the load-deformation assignment.

[0152] In a second stage, in step 712, a time course of a local damage energy is determined using the time course of the local load and the time course of the local deformation.

[0153] Finally, in step 714, a lifetime of one or more component sections is determined using the respective time course of the damage energy and the damage energy-cycle number assignment.

[0154] The computer-implemented method described above can, for example, be part of a finite element approach, in which a component is subdivided by meshing, i.e., by dividing it into individual small component sections, and relevant local quantities such as load, deformation, and damage energy are calculated for each of these sections. In a simulation, all calculable properties of the component or its sections are accessible. In particular, stress and strain can be easily determined through simulation. Therefore, the load-deformation relationship is often given as a stress-strain relationship, and the local load and local deformation are calculated as local stress and local strain, respectively.

[0155] The service life can be determined using the linear damage accumulation method according to Palmgren and Miner (Miner's rule), which is already known from the literature. However, unlike the prior art method, Method 700 does not use the S-N curve, but rather the modified S-N curve, i.e., the damage energy-cycle number of loads assignment. For this purpose, the local damage energy per load cycle is first calculated for each component section according to the load-time function. Each of the values ​​for the local damage energy per load cycle is then compared to the tolerable number of load cycles from the damage energy-cycle number of loads assignment. The total damage for the respective component section can be calculated by summing these ratios. If the sum of the ratios reaches a value greater than 1, local material failure is to be expected.

[0156] To make the simulation of a component even more realistic, the load-deformation relationship can be dependent on accumulated damage energy. In the case of elastic-plastic material behavior, the relationship between load and deformation is not constant, but changes depending on the extent of damage to the material. This dependency can be taken into account by making the load-deformation relationship dependent on accumulated damage, thus transforming it into a relationship based on load, deformation, and accumulated damage energy. The accumulated damage energy is the sum of the damage energies introduced into the component or component section over all load cycles. The accumulated damage can also be calculated for each component section using Method 700.This allows for recourse at any time to a load-deformation relationship corresponding to the accumulated damage energy.

[0157] In summary, this disclosure describes a control and / or regulation unit (100) for a test rig for performing a fatigue test on a test specimen. The control and / or regulation unit (100) comprises the following components: an input interface (110) configured to acquire a load signal (102) indicating a load acting on the test specimen and a deformation signal (104) indicating a deformation of the test specimen; a processing unit (120) configured to determine a damage energy (122) using the load signal (102) and the deformation signal (104); a parameter determination unit (130) configured to determine an operating parameter (132) of an actuator for applying a load to the test specimen as a function of the damage energy (122); and an output interface.

[0158] (140), which is designed to provide an actuator signal (142) depending on the determined operating parameter (132).

Claims

Claims 1. Control and / or regulation unit (100) for a test rig for performing a fatigue test on a test specimen, comprising: an input interface (110) configured to acquire a load signal (102) indicating a load acting on the test specimen and a deformation signal (104) indicating a deformation of the test specimen; a computing unit (120) configured to determine a damage energy (122) using the load signal (102) and the deformation signal (104); a parameter determination unit (130) configured to determine an operating parameter (132) of an actuator for applying a load to the test specimen as a function of the damage energy (122); and an output interface (140) configured to provide an actuator signal (142) as a function of the determined operating parameter (132).

2. Control and / or regulation unit (100) according to claim 1, wherein the computing unit (120) is configured to detect the load acting on the test body and the deformation of the test body as time-correlated pairs of values ​​(360A-L), wherein the pairs of values ​​form a load-deformation curve (210), and to calculate an area of ​​a surface (212) bounded by the load-deformation curve (210) in order to determine the damage energy (122).

3. Control and / or regulation unit (100) according to claim 2, wherein the computing unit (120) is configured in a first calculation mode to calculate the area of ​​the surface bounded by a closed hysteresis curve (210) in order to determine the damage energy (122).

4. Control and / or regulation unit (100) according to claim 2 or 3, wherein the computing unit (120) is configured in a second calculation mode to determine the damage energy (122) by calculating the area of ​​the surface bounded by a section (218, 220) of the load-deformation curve (210) between two successive load reversal points (214, 216) and coordinate axes (242, 244, 252, 254) of a load-deformation coordinate system (240, 250).

5. Control and / or regulation unit (100) according to one of claims 2 to 4, wherein the computing unit (120) is configured to calculate the area using the upper, middle, lower sum approach, trapezoidal approach and / or Simpson's rule.

6. Control and / or regulation unit (100) according to one of the preceding claims, wherein the parameter determination unit (130) is configured to determine a reference value deviation as a function of the damage energy (122) and a predetermined damage energy reference value (152), and to determine the operating parameter as a function of the reference value deviation.

7. Control and / or regulation unit (100) according to one of the preceding claims, wherein the load indicated by the load signal (102) is in the form of a force acting on the test specimen and / or a torque acting on the test specimen and / or a nominal voltage and / or technical voltage and / or true voltage of the test specimen.

8. Control and / or regulation unit (100) according to one of the preceding claims, wherein the deformation of the test body indicated by the deformation signal (104) is in the form of an elongation, a shear, a twist angle and / or a vibration displacement of the test body.

9. Test rig (400) for performing a fatigue test on a test specimen (410), comprising: a holding device (420) configured to receive the test specimen (410), at least one actuator (440) mechanically connectable to the test specimen (410) via the holding device (420) and configured to apply a load to the test specimen (410), a load sensor (430) configured to provide a load signal (102) indicating a load applied to the test specimen (410), a deformation sensor (450) configured to provide a deformation signal (104) indicating a deformation of the test specimen (410), and a control and / or regulation unit (100) according to one of the preceding claims, the input interface (110) of which is connected to the load sensor (430) and the deformation sensor (450). is to receive the load signal (102) and the deformation signal (104),and whose output interface (140) is connected to the at least one actuator (440) in order to provide the actuator signal (142) to the at least one actuator (440).

10. Method (570) for performing a fatigue test on a test specimen, comprising: cyclic loading and unloading of the test specimen by means of an actuator to generate a large number of vibration cycles (580), and repeated adjustment of the at least one actuator by means of a control loop, such that a damage energy experienced by the test specimen through the cyclic loading and unloading corresponds to a predetermined damage energy reference value (590).

11. Method (570) according to claim 10, wherein the predetermined damage energy reference value is a constant damage energy reference value and an acceptable number of load cycles for the test specimen is determined for the constant damage energy reference value.

12. Method (570) according to claim 10 or 11, wherein during cyclic loading of the test body a deformation of the test body and a load acting on the test body is detected, and using the detected deformation, the detected Using the load and the specified damage energy reference value, a relationship is determined between the load acting on the test specimen and the deformation of the test specimen as a function of an accumulated damage energy of the test specimen.

13. Method (570) according to claim 10 or 12, wherein the predetermined damage energy reference value is a predetermined damage energy reference value that varies over time according to a stress-time function.

14. Computer-implemented method (700) for performing a numerical fatigue test on a test specimen, comprising the following steps: Providing a test body geometry that specifies a spatial extent of the test body (702); Providing a load-deformation mapping for a material of the test specimen, which specifies a mapping between a load of the material and a deformation of the material (704); Providing a damage energy-cycle number assignment for the material, which specifies an assignment between a tolerable number of cycles at a constant damage energy per cycle (706); Providing a load-time function that specifies a time-distribution load to which the test specimen is to be subjected (708); for one or more test specimen sections of the test specimen, determining a time course of a local damage energy that specifies damage to the respective test specimen section caused by loading the test specimen, wherein a time course of a local load and a time course of a local deformation of the respective test specimen section at The load on the test specimen is determined according to the load-time function and using the test specimen geometry and the load-deformation mapping (710), and the time course of the local damage energy is determined based on the time course of the local load and the local deformation (712); and Determining a lifetime for the one or more test body sections using the respective time course of the damage energy and the damage energy-cycle number assignment (714).

15. Computer-implemented method (700) according to claim 14, wherein the load-deformation mapping has a dependence on an accumulated damage energy, a local accumulated damage energy for the respective test specimen section is determined on the basis of the temporal course of the local damage energy, and when determining the temporal course of the local load and / or the local deformation, the dependence of the load-deformation mapping on the accumulated damage energy is taken into account.

Images