Method and device for determining the damage state of components

Abstract

Method for determining the damage state of components, wherein damage states for a component are determined as a function of operating states of a total system comprising the component, by • in a first procedural step - during operation of the overall system, a time history of stresses and / or loads on the component and a time history of signals (S) of the overall system relevant for determining operating states are recorded, - from the recorded signals (S) the operating states of the overall system that occurred and a time history of the operating states that occurred are determined, - from the recorded stresses and / or loads, damaging events in the form of load changes are determined and assigned to the operating states, and thus damaging events are determined for each operating state (EB) by performing a stress analysis and / or load analysis in which each stress and / or load is assigned its corresponding operating state based on the signals (S) of the overall system, and a rainflow classification of the stresses and / or loads is performed to determine the occurrence of damaging events in the form of load changes. - Stress collectives and / or load collectives per operating condition (BB) are determined and - time-normalized stress collectives and / or load collectives per operating state (ZBB) are determined based on the duration of the respective operating state, • wherein in a second process component, the time-normalized stress collectives and / or load collectives per operating state (ZBB) are used - the damage state of the component is determined depending on the operating conditions of the overall system that have occurred and their respective durations, and / or the future damage state is determined depending on the predicted future operating conditions of the overall system encompassing the component and their respective durations, and / or - for a further component designed in the same way as this component, the damage state is determined depending on the operating conditions of a further overall system comprising this further component and designed in the same way as the overall system, and their respective duration, and / or depending on the predicted future operating conditions of the further overall system comprising the further component and their respective duration, the future damage state is determined.

Description

[0001] The invention relates to a method and a device for determining the damage state of components.

[0002] From the prior art, as described in DE 100 60 694 A1, a method for recording the operating status of internal combustion engines in motor vehicles is known. In this method, measured variables and events of the operation of the motor vehicle are recorded and base variables are formed from them. Furthermore, classification variables are formed from the residence times of two measured variables in the form of composite classifications. Both the base variables and the composite classification variables are stored in a memory and made available for retrieval as corresponding signals. On the basis of the information thus obtained, a suitable evaluation is carried out with the aim of mathematically determining a target evaluation variable for, for example, the consumption of an operating fluid such as engine oil and fuel and a corresponding comparison with a consumption measurement value, which allows conclusions to be drawn about the state of wear of the internal combustion engine or components of the internal combustion engine.

[0003] DE 10 2015 120 203 A1 describes a method and a system for determining the stress on a vehicle and / or at least one vehicle component. Driving maneuver detection is performed using sensor data that is measured and available during driving in the vehicle. The driving maneuvers are classified according to their intensity. Stress indicators are assigned to the classified intensities of the driving maneuvers. Based on the determined driving maneuvers and their intensities and stress indicators, the stress on the vehicle and / or at least one vehicle component is determined.

[0004] A model-based lifetime monitor is known from DE 102 57 793 A1. It determines system loads from the existing system sensors, in particular from vehicle sensors, and processes the sensor signals from the system sensors into load-time profiles. Local component stresses are determined from the system model and the load-time profiles. The remaining lifetime of the components included in the system model is calculated from the accumulated component damage using a fatigue strength analysis.

[0005] DE 102 15 865 A1 describes a method and a control unit for determining the failure probability of a motor vehicle component. During operation of the component, data in the form of damage-relevant influencing parameters that influence the failure probability of the component are recorded and stored. To determine the failure probability, a load spectrum is determined as a key influencing factor for verifying the fatigue strength of the component. The influencing parameters are evaluated during operation of the component in a control unit of the motor vehicle to determine the failure probability.

[0006] DE 10 2007 002 801 A1 discloses a method for determining the probability of failure of at least one component of a drive device. The method comprises the steps of determining at least one strength of the component, determining at least one load on the component in multiple operating states of the drive device, determining frequencies of the operating states of a plurality of similar drive machines over a specific period of time, and determining the frequency of damage to the component using a stochastic method based on the strength and load of the component as well as the frequencies of the operating states.

[0007] The invention is based on the object of specifying a method and a device for determining the damage state of components that is improved compared to the prior art.

[0008] The object is achieved according to the invention by a method for determining a damage state of components with the features of claim 1 and a device for determining a damage state of components with the features of claim 9.

[0009] Advantageous embodiments of the invention are the subject of the subclaims.

[0010] In a method according to the invention for determining a damage state of components, damage states are determined for a component as a function of operating states of an overall system comprising the component, in that in a first part of the method, during operation of the overall system, a time profile of stresses and / or strains on the component and a time profile of signals of the overall system relevant for determining operating states are recorded, from the recorded signals, operating states of the overall system that have occurred and a time profile of the operating states that have occurred are determined, damaging events are determined from the recorded stresses and / or strains and are assigned to the operating states, stress collectives and / or load collectives are determined for each operating state and, based on a duration of the respective operating state, time-normalized stress collectives and / or load collectives are determined for eachoperating condition are determined. By means of these time-normalized stress collectives and / or load collectives for each operating condition, the damage state is now determined in a second part of the process, for example for the component depending on the operating conditions that have occurred in the overall system and / or the future damage state is determined, for example, depending on the predicted future operating conditions of the overall system comprising the component and / or the damage state is determined, for example, for another component designed similarly to this component, depending on the operating conditions that have occurred in a further overall system comprising this further component and designed similarly to the overall system and / or the future damage state is determined, for example, depending on the predicted future operating conditions of the further overall system comprising the further componentDamage status determined.

[0011] The overall system is, for example, a vehicle, a machine or another overall system.

[0012] By means of the method according to the invention, the damaging events are determined depending on the operating states of the overall system that have occurred. This makes it possible, in the second part of the method, to determine a damage state of the component or a similar additional component exclusively based on the operating states of the respective overall system that have occurred and their respective duration, or to predict a damage state of the component or a similar additional component exclusively based on predicted future operating states of the respective overall system and their respective duration. This is because the solution according to the invention enables the reconstruction of a stress collective and / or load collective for the respective specific component solely from the recording of the operating states of the overall system that includes this component.The method according to the invention can thus be used, for example, for online monitoring of the current damage state of the component without the stress and / or load profile having to be determined on the component itself in the second part of the method.

[0013] The method according to the invention provides, in particular, an operating-state-dependent stress analysis and / or load analysis for reconstructing future and / or historical stress collectives and / or load collectives of a component. Operating-state-dependent stress collectives and / or load collectives are determined, and from these, future and / or historical stress collectives and / or load collectives can be determined to describe the current damage state of a component, without having to perform a stress measurement and / or load measurement on this component in the second part of the method.

[0014] The method is applicable to components to which a stress collective and / or load collective can be assigned based on a stress-time history and / or load-time history, i.e., based on a recorded time history of stresses and / or loads. The stress collective and / or load collective must be determined by identifying damage-relevant events from the stress-time history or from the load-time history. The stress and / or load can be, for example, a mechanical stress, a quantity proportional to the mechanical stress, a temperature, a pressure, an electrical voltage, an electrical current, or another physical quantity.

[0015] The principle of the method according to the invention is based on the assignment of the operating states to the corresponding stress measurement value and / or load measurement value and the generation of a respective individual set of damage characteristics, whereby a damaging event is divided equally among the operating states involved.

[0016] In order to obtain the temporal progression of the operating states that have occurred, it is advantageous to classify the signals of the overall system and thereby define possible operating states. This means that the operating states are determined from the recorded signals of the overall system. This can therefore also be done, for example, during normal operation and / or after normal operation of the overall system. It is therefore not necessary to specify operating states that the entire system must go through, i.e., in the form of a standardized test drive. The method described here is therefore significantly more flexible and can also be used, for example, in normal customer operation of overall systems, e.g., vehicles.

[0017] The duration of the respective operating state is advantageously determined by creating a multi-dimensional histogram for the signals of the overall system and the classes into which they are divided, based on which the duration of the respective operating state is determined. The histogram is advantageously determined using multi-parameter, related instantaneous value counting. This means that for each measuring point at which a stress measurement and / or load measurement value was recorded, the values ​​of all recorded signals of the overall system relevant for determining the operating states (i.e. the recorded signals with the exception of stress and / or load) are determined, and the associated signal classes are determined. For each combination of classes, there is exactly one element in the histogram, the count value of which is incremented by the time interval from the previous measuring point. Each element of the histogram formally represents an operating state.The result is therefore a duration of the operating states of the entire system.

[0018] As an alternative to the procedure described in the two previous paragraphs, i.e. as an alternative to the described classification for defining possible operating states in order to obtain the time course of the operating states that have occurred and to the described determination of the duration of the respective operating state, the operating states can also be determined, for example, using methods of artificial intelligence.

[0019] According to the invention, the damaging events for each operating condition are determined by determining damaging events in the frequency range from the recorded stresses and / or loads in the time domain, using a counting method suitable for the given stress and / or load quantity to convert the quantity from the time domain to the frequency range, and assigning them to the operating conditions. For this purpose, a stress analysis and / or load analysis is carried out according to the invention, in which the damaging events are determined using a suitable counting method, and a quantitative characteristic is assigned to the determined damaging events regarding their contribution to damage.This can be done either by identifying all damaging events first and then allocating the damage contribution to all events in a subsequent step, or by assigning the damage contribution for a damaging event immediately after identifying a damaging event before identifying further events.

[0020] In the stress analysis and / or load analysis, each stress and / or load is assigned its corresponding operating state based on the signals of the overall system and a rainflow classification of the stresses and / or loads is carried out to determine the occurrence of damaging events in the form of load changes.

[0021] In particular, a separate rainflow matrix is ​​generated for each operating state in which counted load changes are proportionally distributed among the operating states involved in the generation of the load changes by dividing the respective load change into two half load changes and storing them in the rainflow matrices of the operating states belonging to its reversal points, and for taking a residue into account, the respective load change is distributed 50% each among the rainflow matrices of the operating states involved, depending on the type of consideration, in particular whether there is no load change, a half load change or a full load change.

[0022] The stress collectives for each operating condition are advantageously determined by evaluating the rainflow matrices individually for each operating condition, in particular by extracting the recorded load changes.

[0023] In an alternative approach, each stress and / or load in the form of mechanical stress or mechanical strain is assigned its corresponding operating state based on the signals of the overall system, and a rainflow classification of the stresses and / or loads is performed to determine the occurrence of damaging events in the form of load cycles. The determined load cycles are converted into a damage parameter immediately after their determination, in particular according to the notch strain concept of fatigue strength, and added to the operating states involved in the associated load cycle (operating states associated with the reversal points of the load cycle) as damage parameters with half the original frequency – i.e., ½ – to the damage parameter collectives of the associated operating states.The damage parameter collectives for each operating condition represent the stress collectives and / or load collectives for each operating condition.

[0024] The time-normalized stress collectives and / or load collectives for each operating condition are advantageously determined by normalizing the stress collectives and / or load collectives for each operating condition in the frequency of their individual collective stages to the duration of the corresponding operating condition.

[0025] The method can be carried out, for example, by means of a device which is arranged, for example, directly on the overall system, i.e. onboard, or, for example, partly on the overall system and partly offboard, i.e. external to the overall system. External to the overall system, the device in particular comprises a processing unit, for example in the form of a computer. In the case of a device arranged entirely on the overall system, the processing unit is, for example, a control unit or another computer unit. External to the overall system, the method can be carried out, for example, on the basis of externally recorded data, in particular time profiles of the stress and / or load on the component under consideration and the signals of the overall system relevant for determining the operating states.For example, the device comprises corresponding acquisition units in the overall system, which transmit this data to an external unit, in particular to the processing unit, for recording and processing. Alternatively, this data can be initially recorded in the overall system and later transmitted to the processing unit for evaluation. Data transmission can be wireless, for example, via radio, or via a cable.

[0026] Such a device according to the invention is thus designed and configured to carry out the method.

[0027] The device comprises, for example, at least one unit for detecting the stresses and / or loads on the component, at least one unit for detecting the signals of the overall system relevant for determining operating states, and at least one processing unit. The processing unit is particularly designed and configured to carry out the above-mentioned method steps, in particular to process the detected stresses and / or loads on the component and the signals of the overall system relevant for determining operating states, in particular in the manner described above.

[0028] For example, as already mentioned, the device can be arranged or can be arranged entirely within the overall system, or at least one component of the device, in particular the at least one unit for detecting the stresses and / or loads on the component and / or the at least one unit for detecting the signals of the overall system relevant for determining operating states, can be arranged or can be arranged within the overall system. The processing unit is then, for example, a component of the device external to the overall system, such as a computer.

[0029] Embodiments of the invention are explained in more detail below with reference to a drawing.

[0030] To avoid misunderstandings, it should be noted that the terms "stress collective," "stress analysis," "stress-time curve," "stress curve," "stress measurement," and "stress measurement" used above all refer to the implementation of the method using stresses. Analogously, the method can also be implemented using loads, either alternatively or additionally, to which the terms "stress collective," "stress analysis," "stress-time curve," "stress curve," "stress measurement," and "stress measurement" used above refer accordingly.

[0031] It shows: Fig. 1 schematically shows a sequence of a method for determining a damage state of at least one component.

[0032] Fig. Figure 1 schematically shows a process flow for determining the damage state of components, particularly mechanical components. This process enables an operating-condition-dependent stress analysis to reconstruct future and / or historical stress collectives of a component. In particular, the process enables the determination of operating-condition-dependent stress collectives to reconstruct future and / or historical stress collectives to describe the current damage state of a component, without having to perform a stress measurement on the component in a second process component.

[0033] Alternatively or additionally, the procedure described below for determining the damage state of components, particularly mechanical components, is carried out not by means of stresses, but by means of loads. This procedure then enables an operating-condition-dependent load analysis to reconstruct future and / or historical load collectives of a component. In particular, the procedure then enables the determination of operating-condition-dependent load collectives to reconstruct future and / or historical load collectives to describe the current damage state of a component, without having to perform a load measurement on the component in the second part of the procedure.Therefore, in the following description relating to stresses, the respective term or component of the term stress can be replaced by the term or component of the term load in order to describe the analogous procedure for loads and the load analysis.

[0034] The method is applicable to components to which a stress collective can be assigned based on a stress-time history (BSV), i.e., a time history of stresses encountered on the component. The stress collective must be determined by identifying damage-relevant events from the stress-time history (BSV). The stress and / or load can be, for example, a mechanical stress, a quantity proportional to the mechanical stress, a temperature, a pressure, an electrical voltage, an electrical current, or any other physical quantity.

[0035] It is already known from the state of the art to conduct a mechanical stress analysis for the purpose of fatigue strength. For this purpose, material-mechanical damage-relevant events are recorded based on a stress-time function and processed into a stress collective, for example, a stress amplitude collective (nominal stress concept, notch stress concept) or a damage parameter collective (notch strain concept).

[0036] The mechanical stress is first subjected to a rainflow classification. From the resulting rainflow matrix, recorded load cycles are determined according to amplitude, mean, and frequency. Rainflow classification leaves a residual (turning points that do not generate load cycles), which, depending on the application, is either ignored, considered as half a load cycle, or considered as a full load cycle. The resulting amplitude-mean collective can be transformed into a substitute amplitude collective using a Haigh diagram, thus yielding the desired stress collective.

[0037] However, since a measurement period is usually significantly shorter than the component's service life, the resulting stress spectrum must be extrapolated to obtain a design spectrum for the component's entire service life. According to the current state of the art, this extrapolation is primarily based on the relationship between the measurement period and the planned service life. It is assumed that the stress behavior during the service life is the same as it was during the measurement. This requires precise knowledge of a future application profile right from the test planning stage in order to accurately represent the composition of the various application scenarios and obtain the most representative design spectrum possible.

[0038] The design spectrum thus obtained can be used to carry out theoretical and experimental lifetime analyses.

[0039] To determine the current damage state of a component, a current stress spectrum is required that reflects the component's stress history. Currently, this requires either a measurement of the stress over the entire service life or a reconstruction based on the previously determined design spectrum.

[0040] Such a reconstruction of a current load spectrum has so far been based primarily on the relationship between the current and planned service life. This means that the spectra of the reconstructed current load spectrum and the design load spectrum do not differ, and in the case of a relative service life analysis using a damage accumulation hypothesis, only the corresponding service lives are effectively compared. Due to this fact, the use of such reconstructed load spectrums for damage assessment in the sense of predictive maintenance is questionable.

[0041] Furthermore, state-of-the-art approaches exist that determine separate stress collectives from defined test operations for selected, previously defined operating conditions and superimpose them, weighted according to a usage profile, to form a design collective. This also makes it possible to reconstruct a current stress collective based on a usage profile. However, this approach involves considerable effort in test planning and execution. The number of manageable operating conditions is severely limited.

[0042] The method described below therefore extends the stress analysis of a component to determine the damage-relevant events depending on the operating states of an overall system. The method thus enables the reconstruction of a stress collective for a specific component solely from the recording of the operating states of the overall system. This can be used, for example, for online monitoring of the current damage state of the component, without the need to determine the stress profile on the component itself in the second part of the method.

[0043] Advantageously, in this method, damage states are determined for a component as a function of operating states of an overall system comprising the component, in that in a first part of the method, during operation of the overall system, a time profile of stresses on the component and a time profile of signals S of the overall system relevant for determining operating states are recorded, operating states of the overall system that have occurred are determined from the recorded signals S, in particular a time profile of the operating states that have occurred, damaging events are determined from the recorded stresses and assigned to the operating states, stress collectives for each operating state BB are determined and time-normalized stress collectives for each operating state ZBB are determined based on a duration of the respective operating state.Using these time-normalized stress collectives per operating condition ZBB, the damage state can now be determined, for example, for the component depending on the operating conditions that have occurred in the overall system. Alternatively or additionally, the future damage state can be determined, for example, depending on predicted future operating conditions of the overall system comprising the component. Alternatively or additionally, the damage state can be determined, for example, for another component designed similarly to this component depending on the operating conditions that have occurred in another overall system that includes this additional component and is designed similarly to the overall system.Alternatively or additionally, the future damage state can be determined, for example, depending on predicted future operating conditions of the further overall system comprising the further component.

[0044] The procedure is described below using the example of a load on a component in the form of mechanical stress, which is present as a discrete time curve. The data recorded is therefore a stress-time curve BSV of the component, i.e. a, in particular, discrete, time curve of the stresses, i.e. of stress measurements, of the component. In addition to this stress-time curve BSV, which in the example described here is in the form of a stress curve, signals S from the higher-level overall system, which includes the component, are recorded as further data. More precisely, a time curve of these signals S is recorded. These are, in particular, signals S that are relevant for determining the operating states of the overall system.

[0045] In a first process step VS 1, these signals S are initially divided into areas, specifically classes, that are characteristic of the operating states occurring in order to define possible operating states of the overall system. The result is an operating state-time profile (BZV), i.e., a time profile of the operating states that have occurred.

[0046] For the signals S used with the classes defined in this way, a multidimensional histogram is then created in a second process step VS2, on the basis of which the duration of the respective operating state is determined.

[0047] For example, for the analysis of a component designed as a mechanical vehicle component, such a signal S could be the driving speed. The relevant ranges would be defined by designated operating points. For example, four classes can be defined in which the driving speed typically falls: Standstill (speed = 0 km / h), City driving (speed between 0 km / h and 50 km / h), Overland travel (speed from 50 km / h to 100 km / h), Motorway driving (speed greater than 100 km / h).

[0048] The histogram is determined in the second process step VS2 using a multi-parametric, related instantaneous value count. This means that for each measuring point at which a stress measurement was recorded, the values ​​of all recorded signals S of the overall system relevant for determining the operating states (i.e., the recorded signals S with the exception of mechanical stress) are determined, and the associated classes of the signals S are identified. For each combination of classes, there is exactly one element in the histogram, whose count is incremented by the time interval from the previous measuring point. Each element of the histogram formally represents an operating state. The result of this second process step VS2 is thus a duration of the operating states DB of the overall system.

[0049] This determines a real sequence of operating states and the actual duration of each operating state, rather than a predetermined sequence and duration of operating states being processed by the entire system. The method can therefore be implemented, for example, during real operation of the entire system, allowing more realistic results to be achieved during test operation compared to a test system and the predetermined operating states implemented with it.

[0050] The next and third process step VS3 is a stress analysis which is modified from the stress analysis used previously described above. This means that the damage-relevant events, i.e. those which damage the component, are determined in a manner which is modified from the prior art procedure described above. For this purpose, each measured stress value is first assigned its corresponding operating state based on the signals S of the overall system. The input data is therefore the determined operating state-time curve BZV and the stress-time curve BSV. A modified form of rainflow classification is then carried out. The damaging events, in this example the load change, are determined using classic rainflow classification. The actual modification occurs when these determined events which damage the component are saved.

[0051] According to the conventional method, the load changes would be stored in a rainflow matrix. In contrast, the method described here creates a separate rainflow matrix for each operating state, in which the counted load changes are proportionally distributed among the operating states involved in the generation of the load changes.

[0052] A load cycle according to the classic rainflow classification consists of two reversal points, each characterized by a measured stress and / or load value. In the method described here, however, in addition to the measured stress values ​​of the reversal points, the corresponding operating states are also known. Therefore, each load cycle is divided into two half load cycles, each of which is stored in the rainflow matrices of the operating states corresponding to the reversal points.

[0053] The residual is considered in a similar way. Here, too, the load change is divided 50% each among the rainflow matrices of the operating states involved, depending on the type of consideration—that is, whether there is no load change, a half load change, or a full load change. Only in the special case where only one operating state has occurred does the result of the modified rainflow classification correspond to the result of the classic rainflow classification described above.

[0054] The result of this third process step VS3 are the events damaging the component per operating state EB, ie the distribution of the events damaging the component across the operating states.

[0055] In the example shown, the rainflow matrices obtained in this way are advantageously evaluated individually for each operating condition in at least one further, here fourth, method step VS4 in order to obtain stress collectives for each operating condition BB as a result. In this case, the recorded load changes, in the exemplary embodiment described here, the mechanical stress are extracted according to stress amplitude, mean stress and frequency. The mean stress is taken into account according to the Haigh diagram, preferably for a damage-equivalent, mean stress-free amplitude. In this way, the stress collectives for each operating condition BB are obtained as substitute amplitude collectives for each operating condition. In a further, here fifth, method step VS5, these are normalized in terms of the frequency of their individual collective levels to the duration of the associated operating condition. I.e.The extent of the stress collectives is standardized with the duration of the respective operating condition.

[0056] In one possible embodiment of the method, the level of stress at the individual collective stages can also be influenced during this standardization process. This is done by taking into account statistical parameters of the distribution of the maximum stress values ​​and the ratio of the duration of an operating condition to the predicted total service life of the entire system.

[0057] The result is a time-related, ie, time-normalized, stress collective for each operating condition ZBB, ie, a time-normalized stress collective for each operating condition. This is the related, ie, time-related, damage characteristic of the respective operating condition.

[0058] The described procedure for determining the time-normalized stress collective for each operating condition ZBB from the stress-time curve BSV for the component and the time curve of the signals S of the overall system comprising the component is a coupled stress analysis GB.

[0059] The procedure described serves as an example for the evaluation of a mechanical stress measurement based on the nominal stress concept or the notch stress concept of fatigue strength, but is not limited to these applications. The principle of the method described here is based on assigning the operating states of the overall system to the corresponding stress measurement value of the component and generating an individual set of damage characteristics for each operating state, whereby a damaging event is distributed equally among the operating states involved.

[0060] By means of the time-related, ie time-normalized, stress collectives for each operating condition ZBB determined in the manner described, a stress collective for a component can now be reconstructed for any period of time, for example for the component by means of which the time-normalized stress collectives for each operating condition ZBB are calculated in the manner described above and in Fig. 1, and / or similarly for another similar component of another similar overall system. For this purpose, only a histogram for the occurrence of the operating states of this component, for which the stress collective is to be reconstructed, or more precisely, of the entire system including this component, is required in a particular time interval under consideration, i.e., the time interval of interest.

[0061] The occurrence of the operating states can be determined, for example, by recording actually occurring operating states of this overall system, for example analogous to the procedure described above according to process steps VS1 and VS2 from a recorded time history of signals S of this overall system, or an operating state history can be specified, for example resulting from a usage forecast for the overall system. The time-related, i.e. time-normalized, load collectives for each operating state ZBB are then extrapolated to the duration of the respective operating state by multiplying their frequency by the duration of the respective operating state, and then superimposed to form an overall load collective across all operating states.

[0062] In one possible embodiment of the method, the level of stress and / or load of the individual collective stages can also be influenced in this extrapolation process. This is done by taking into account the statistical parameters used for standardization, the distribution of the maximum stress values, and the ratio of the duration of an operating condition to the predicted total service life of the entire system.

[0063] This superimposed total stress collective is the reconstructed stress collective for the component, as mentioned above, for example, for the component, by means of which the time-normalized stress collectives for each operating state ZBB are calculated to the value described above and in Fig. 1, and / or for a similar further component of a similar further overall system.

[0064] The particular advantage of this approach is that reconstructing a stress spectrum eliminates the need to perform repeated stress measurements in the corresponding time range. Based on this, the current damage state of a component can be derived by selecting appropriate reconstructed stress spectrums. This requires a stress spectrum that corresponds to the stress to be endured over the entire service life of the component, in the sense of a failure criterion.This can be determined, for example, by constructing a histogram for the occurrence of operating conditions for the expected service life by extrapolating from a measurement, or by performing a measurement across a group of identical overall systems, for example, across an entire vehicle fleet, over the expected service life and superimposing the individual histograms of all overall systems, for example, all vehicles. From this histogram, a design collective for the component is determined by reconstruction using the time-related, i.e., time-normalized, stress collectives for each operating condition ZBB.

[0065] Furthermore, a histogram of the occurrence of operating conditions over the entire service life of an entire system, for example, a vehicle, can be determined. From the reconstruction of the stress collective based on this histogram, the stress collective of the entire system, for example, the vehicle, can be reconstructed up to the current point in time. By comparing this currently reconstructed stress collective with the design collective, the current damage of the component can be determined in comparison to the design collective. A suitable method for this is the linear damage accumulation hypothesis according to Palmgren and Miner. Due to its simplicity, the "Miner-Elementary" modification is used in this example.

[0066] The prerequisite for the applicability of this method for determining damage is knowledge of the Wöhler curve. This is either determined experimentally or estimated using suitable methods (e.g., according to the FKM guideline).

[0067] The linear damage accumulation calculation results in a damage sum for each of the reconstructed stress collective and the design collective, with the ratio of the two damage sums representing the current damage of the component. By continuously updating the current reconstructed stress collective by updating the histogram of the occurrence of the operating states, online monitoring of the component's damage status is possible.

[0068] The described procedure is an example for determining the damage of an equivalent amplitude collective as a stress collective based on the nominal stress concept or the notch stress concept of fatigue strength, but is not limited to these applications. The principle is based on the separate determination of the damage of the current reconstructed stress collective and the design stress collective using a suitable method for determining the damage and the comparison of the resulting damage values.

[0069] As a further exemplary embodiment, the procedure for the modified stress analysis based on the notch strain concept is described below. In this case, only the concrete implementation of process steps VS3 and VS4 differs from the procedure described above. In step VS3, the damaging events are determined in the form of load cycles using rainflow classification. However, these are not stored in rainflow matrices as described above, but are converted into a stress parameter (here: damage parameter) directly after their recording in step VS4 according to the notch strain concept of fatigue strength. The damage parameter determined in this way is used as a damage parameter with 50% of the original frequency, i.e.0.5 - is added to the damage parameter collectives of the corresponding operating conditions. In this way, the stress collectives for each operating condition BB are obtained as damage parameter collectives for each operating condition.

[0070] The further procedure is, at least essentially, analogous to the above example. The only difference here is partly the determination of the damage. According to the notch strain concept, damage parameter Wöhler curves corresponding to the damage parameter used are required instead of conventional Wöhler curves. The damage sum is determined analogously to the Miner's rule.

[0071] The method can be carried out, for example, by means of a device which is arranged, for example, directly on the overall system, i.e., onboard, or, for example, partially on the overall system and partially offboard, i.e., external to the overall system. External to the overall system, the device comprises, in particular, a processing unit, for example, in the form of a computer. In the case of a device arranged entirely on the overall system, the processing unit is, for example, a control unit or another computer unit. External to the overall system, the method can be carried out, for example, on the basis of externally recorded data, in particular, time profiles of the stress on the component under consideration and the signals of the overall system relevant for determining the operating states.For example, the device comprises corresponding acquisition units in the overall system, which transmit this data to an external unit, in particular to the processing unit, for recording and processing. Alternatively, this data can be initially recorded in the overall system and later transmitted to the processing unit for evaluation. Data transmission can be wireless, for example, via radio, or via a cable.

[0072] Such a device according to the invention is thus designed and configured to carry out the method.

[0073] The device comprises, for example, at least one unit for detecting the stresses on the component, at least one unit for detecting the signals S of the overall system relevant for determining operating states, and at least one processing unit. The processing unit is particularly designed and configured to carry out the above-mentioned method steps, in particular to process the detected stresses on the component and the signals S of the overall system relevant for determining operating states, in particular in the manner described above.

[0074] For example, as already mentioned, the device can be arranged or can be arranged entirely within the overall system, or for example, at least one component of the device, in particular the at least one unit for detecting the stresses on the component and / or the at least one unit for detecting the signals S of the overall system relevant for determining operating states, can be arranged or can be arranged within the overall system. The processing unit is then, for example, a component of the device external to the overall system, such as a computer. LIST OF REFERENCE SYMBOLS BB stress collectives per operating condition BSV stress-time curve BZV operating state time history DB Duration of operating states EB damaging events per operating condition GB coupled stress analysis S signals ZBB time-normalized stress collective per operating condition VS1 first procedural step VS2 second procedural step VS3 third procedural step VS4 fourth procedural step VS5 fifth procedural step

Claims

[1] Method for determining a damage state of components, whereby damage states are determined for a component as a function of operating states of an overall system comprising the component, by • in a first stage of the procedure - during operation of the entire system, a time course of stresses and / or loads on the component and a time course of signals (S) of the entire system relevant for determining operating states are recorded, - the operating states of the entire system that have occurred and a time history of the operating states that have occurred are determined from the recorded signals (S), - damaging events in the form of load changes are determined from the recorded stresses and / or loads and assigned to the operating conditions, thus determining damaging events for each operating condition (EB) by carrying out a stress analysis and / or load analysis in which each stress and / or load is assigned its associated operating condition on the basis of the signals (S) of the overall system and a rainflow classification of the stresses and / or loads is carried out to determine the occurrence of damaging events in the form of load changes, - stress collectives and / or load collectives are determined for each operating condition (BB) and - time-standardised stress collectives and / or load collectives for each operating condition (ZBB) are determined based on the duration of the respective operating condition, • whereby in a second part of the process, the time-standardized stress collectives and / or load collectives for each operating condition (ZBB) - the damage state is determined for the component depending on the operating conditions of the overall system that have occurred and their respective duration, and / or the future damage state is determined depending on the predicted operating conditions of the overall system comprising the component that are to occur in the future and their respective duration, and / or - for a further component of a similar design to this component, the damage state is determined depending on the operating conditions that have occurred in a further overall system comprising this further component and of a similar design to the overall system and their respective duration, and / or the future damage state is determined depending on the operating conditions that are forecast to occur in the future in the further overall system comprising the further component and their respective duration. [2] Method according to claim 1, wherein a classification of the signals (S) of the overall system and thereby a definition of possible operating states is carried out in order to obtain the time course of the operating states that have occurred. [3] Method according to claim 2, wherein the duration of the respective operating state is determined by forming a multidimensional histogram for the signals (S) of the overall system with the classes into which they are divided, on the basis of which the duration of the respective operating state is determined. [4] Method according to claim 3, wherein the histogram is determined via a multi-parametric, related instantaneous value count. [5] Method according to one of the preceding claims, wherein in the stress analysis and / or load analysis - a separate rainflow matrix is ​​generated for each operating state, in which counted load changes are distributed proportionally among the operating states involved in the generation of the load changes, by dividing the respective load change into two half load changes and storing them in the rainflow matrices of the operating states belonging to the reversal points, and - to take a residual into account, the respective load change is divided 50% each between the rainflow matrices of the operating states involved, depending on the type of consideration, in particular whether there is no load change, half a load change or a full load change. [6] Method according to claim 5, wherein the stress collectives and / or load collectives are determined for each operating condition (BB) by individually evaluating the rainflow matrices for each operating condition, in particular by extracting the recorded load changes. [7] Method according to one of claims 1 to 4, wherein in the stress analysis and / or load analysis the determined load changes are converted into a damage parameter directly after their determination and are each added as a damage parameter with half the original frequency to damage parameter collectives of the operating states involved in the associated load change, wherein the damage parameter collectives for each operating state represent the stress collectives and / or load collectives for each operating state (BB). [8] Method according to one of the preceding claims, wherein the time-normalized stress collectives and / or load collectives per operating condition (ZBB) are determined by normalizing the stress collectives and / or load collectives per operating condition (BB) in the frequency of their individual collective stages to the duration of the associated operating condition. [9] Device designed and arranged to carry out a method according to one of the preceding claims. [10] Device according to claim 9, comprising - at least one unit for recording the stresses and / or loads on the component, - at least one unit for recording the signals (S) of the overall system relevant for determining operating states, - at least one processing unit. [11] Device according to claim 9 or 10, wherein the device is arranged or can be arranged entirely on the overall system or wherein at least one component of the device, in particular the at least one unit for detecting the stresses and / or loads of the component and / or the at least one unit for detecting the signals (S) of the overall system relevant for determining operating states, is / are arranged or can be arranged on the overall system.

Images