Method and apparatus for determining a damage state of a component

Abstract

The invention relates to a method for determining a damage state of a component, wherein, for the component, the damage state is determined from operating states of a total system comprising the component in such a way that, for each operating state, a time-normalized load set and / or load set (ZBB) is determined, by means of which, for the component, the damage state is determined from the operating states occurring for the total system, and / or the future damage state is determined from the predicted future occurring operating states of the total system comprising the component, and / or, for another component which is identically constructed to the component, the damage state is determined from the occurring operating states of another total system comprising the other component and identically constructed to the total system, and / or the future damage state is determined from the predicted future occurring operating states of the other total system comprising the other component.

Description

Technical Field

[0001] The present invention relates to a method and apparatus for determining the damage state of a component. Background Technology

[0002] As described in DE 100 60 694 A1, a method for detecting the operating condition of an internal combustion engine in a motor vehicle is known from the prior art. In this method, measured variables and events of the motor vehicle's operation are detected, thereby forming basic variables. Furthermore, graded variables are formed in the form of composite grades based on the residence times of each pair of measured variables. The basic variables and the composite graded variables are stored in a memory and provided as corresponding signals for retrieval. Based on the information thus obtained, an appropriate evaluation is performed, with the aim of using a computer to determine the expected assessment quantities for, for example, the consumption of operating materials such as engine oil and fuel, and comparing them accordingly with the consumption measurements, a comparison that allows conclusions to be drawn regarding the wear condition of the internal combustion engine or its components. Summary of the Invention

[0003] The purpose of this invention is to provide an improved method and apparatus compared to the prior art for determining the damage state of a component.

[0004] In the method for determining the damage state of a component according to the invention, for a component, the damage state of the component is determined based on the operating state of the overall system including the component. This is achieved by detecting the time progression of the load and / or load on the component and the time progression of signals related to the determination of the operating state of the overall system during the operation of the overall system; determining the operating state of the overall system that has occurred from the detected signals, particularly the time progression of the occurring operating state; determining damaging events from the detected load and / or load and correlating the events with the operating states; determining the load set and / or load set for each operating state; and determining the time-normalized load set and / or load set for each operating state based on the duration of the corresponding operating state. This detection, particularly the detection of the time progression of the load and / or load on the component and the time progression of signals related to the determination of the operating state of the overall system, is performed, for example, during one or more tests, particularly during one or more test runs and / or one or more test bench tests and / or one or more simulations. The detection of the load and / or the time progression of the load on the component, particularly the detection of the load and / or the load on the component, is performed here, especially by a corresponding unit for detecting the load and / or the load on the component, particularly by a sensor system on the component, which, for example, includes at least one or more sensors, that is, the sensor system advantageously forms the unit for detecting the load and / or the load on the component. The sensor system is constructed and configured for detecting the load and / or the load on the component and is suitably mounted on the component. The test run or each test run does not necessarily require specific pre-setting and corresponding execution of a predetermined operating state to achieve a particular pre-defined operating state; rather, the operating state, particularly the time progression of the operating state, is determined by the time progression of the detected signals related to determining the operating state of the overall system. Thus, such testing, particularly test runs, can also be performed, for example, during normal operation of the overall system, but in this case, a sensor system for detecting the load and / or the load on the component, particularly its time progression, is required.

[0005] The correspondence between damaging events and operating states is achieved in particular by proportionally mapping damage events caused by the loads and / or loads, such as mechanical, particularly material mechanical, damage events, such as load changes, in rainflow classification to all operating states involved in their formation. This allows for the automated mapping of damage characteristic values ​​to operating states. Load changes are one possible form of damaging event.

[0006] Now, the damage state is determined by time-normalized load sets and / or load sets for each operating state, for example, for the component, based on the operating state of the overall system and / or for example, based on the predicted future operating state of the overall system including the component, and / or for example, for another component constructed identically to the component, based on the operating state of another overall system including the other component and constructed identically to the overall system, and / or for example, based on the predicted future operating state of the other overall system including the other component. This determination, particularly based on the actual operating state, advantageously can be performed during normal operation of the overall system; that is, test runs and simulations are no longer required, and in particular, sensors for determining the load and / or load of the component are no longer needed. Other variations of the determination based on the predicted future operating state are not associated with the actual operation of the overall system, but rather with the predicted, corresponding, future operation of the overall system as the starting point. In the variant scheme where the operating state is determined based on the actual operating state, the operating state is also detected based on the time process of the signals of the overall system related to the determined operating state. That is, the time process of the signals of the overall system related to the determined operating state is detected during the operation of the overall system, especially during normal operation, and the operating state of the overall system is determined from the detected signals, especially the time process of the operating state.

[0007] As clearly demonstrated by the description of the signals related to the determination of the operating state of the overall system and the time process of determining the operating state of the overall system, especially the time process of the operating state, it is advantageous to continuously determine the operating state over the entire detected time process; that is, to assign a corresponding operating state to each moment of the operation of the overall system that determines the signals related to the determination of the operating state of the overall system. The operation of the overall system is therefore advantageously a continuous process of operating states. At each moment, exactly one operating state is valid, that is, the state derived from the overall system in response to the current signal related to the determination of the operating state is valid.

[0008] In summary, thus, in the first method component, a time-normalized set of loads and / or loads for determining the operating state is determined based on the time progression of the load and / or loads and the time progression of signals related to determining the operating state; and in the second method component, the damage state of the component is determined based on the time-normalized set of loads and / or loads for each operating state determined in the first method component and the currently occurring operating state. The second method component can be performed during normal operation of the overall system, and in particular, it does not require equipping the overall system with an additional sensor system for detecting the time progression of the load and / or loads.

[0009] The overall system is, for example, a vehicle, a machine, or an overall system in other forms.

[0010] Determining the load and / or load set is mandatory only during the determination of the time-normalized load set and / or load set, and thus only during the first method component described above; that is, it is not necessary to detect the load and / or load during normal operation. For this reason, the determination of the load and / or load set, and in particular the measurement, during the determination of the time-normalized load set and / or load set is preferably performed by the additional sensor system already mentioned above, which is not included in the overall system according to the standard, that is, by a sensor system that is not necessary to provide during normal operation of the overall system. This additional sensor system, in particular, includes, for example, one or more strain gauges, force sensors, and / or temperature sensors. During the determination of the time-normalized load set and / or load set, that is, during the first method component mentioned above, and advantageously also during normal operation of the overall system, i.e., in the second method component mentioned above, signals of the overall system related to determining the operating state are detected, and in the second method component, these signals are particularly used to determine the operating state of the overall system so that, at least when the operating state is not predicted to be a future operating state, a damage state can be determined based on the operating state, as described above as an alternative. Especially in the case of a vehicle as the main system, the signals related to determining the operating status of the main system can be, for example, driving speed or the position of a switch.

[0011] Alternatively or additionally, the determination of the time-normalized load set and / or load set can be based entirely or partially on a simulated operation of the overall system, in which the load and / or load, and in particular the signals of the overall system related to the determination of the operating state, are also simulated, and these signals can also be obtained during the normal operation of the overall system.

[0012] According to the method of the invention, damaging events are determined based on the operating states of the overall system. This allows the damage state of a component or other identical components to be determined solely based on the operating state of the corresponding component and the corresponding duration of that operating state, or solely based on the predicted future operating state of the corresponding component and the corresponding duration of that operating state. This is because the solution according to the invention allows the load set and / or load set to be reconstructed for a specific component solely by detecting the operating state of the overall system including that 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 having to determine the load progression and / or load integration on the component itself. As explained above, the prerequisite for this reconstruction is that a time-normalized load set and / or load set is predetermined for each operating state occurring on another component of the same and / or similarly constructed overall system. According to the processing method of the invention, particularly according to the first method component described above, the load set and / or load set has been determined and used for reconstruction.

[0013] In particular, the present invention provides a state-dependent load analysis and / or load analysis for reconstructing a future and / or historical load set and / or load set for a component. By determining the state-dependent load set and / or load set, and thereby determining the future and / or historical load set and / or load set, a description of the current damage state of a component can be made, without necessarily performing load measurements and / or load measurements on said component.

[0014] The method can be used for components to which load sets and / or load groups can be assigned based on load-time processes and / or load-time processes, that is, based on the time process of detected loads and / or loads. These load sets and / or load groups must be determined by identifying damage-related events from the load-time process or by determining damage-related events from the load-time process. The definition of damage-related events is determined based on the fundamental damage mechanism of the component under consideration and the corresponding method used to describe the current damage state due to the occurrence of damage-related events. For example, for the damage mechanism “material mechanical fatigue due to cyclic loads and / or loads,” damage-related events are defined as so-called load changes when described according to nominal stress design for operational stability. Other methods, each with its own definition of damage-related events, are also known to those skilled in the art for describing such and / or other damage mechanisms.

[0015] For example, loads and / or loads can be mechanical stress, a quantity proportional to mechanical stress, temperature, pressure, voltage, current, or other physical quantities. Loads and / or loads specifically refer to physical measurement parameters that directly constitute damage-related measurement parameters based on the damage mechanism associated with the component and / or represent one or more such measurement parameters, which can be converted into directly damage-related measurement parameters through calculation and / or simulation models.

[0016] The principle of the method according to the present invention is based on the correspondence between the operating state and the relevant load measurement values ​​and / or load measurement values, and the generation of separate damage feature groups, that is, groups of damage events.

[0017] The corresponding damaging events are advantageously distributed proportionally to the operational states involved in the events.

[0018] In one possible implementation of the method, damaging events, in particular individual damaging events, are allocated in equal proportions to the operational states involved in the events.

[0019] In another, particularly alternative, possible embodiment of the method, the corresponding damaging events are weighted and assigned to the operating states involved in their formation. Here, in one possible embodiment of the method, the weight of the corresponding operating state is determined by determining the time period in which the damaging event occurs and the total duration of the corresponding operating state within that time period, and the weight is derived as the ratio of the total operating state duration to the total duration of the damaging events. For example, the time period of the total duration of the damaging events, according to the nominal stress design for operational stability, can be considered as the time period between the first and last reversal points of the load transformation, or alternatively as the time period between the first and last reversal points of the load transformation minus the time period attributable to the included load transformations. The sum of the weights of all operating states must also correspond to the original frequency of the damaging events.

[0020] The allocation of corresponding damaging events according to the above-described implementation of the method can be explained by the following formula:

[0021] For weighted allocation, the following applies:

[0022] H i =G i ·H0 (1)

[0023] Here H i G is the frequency of damaging events obtained in operating state i. iH is the weight of running state i, and H0 is the original frequency of damaging events before assignment.

[0024] The condition is that the following formula is satisfied:

[0025]

[0026] Where, n B It represents the number of running states.

[0027] For weighted allocation based on characteristic points (in the case of load transformation, the reversal point of the load transformation), the following applies:

[0028]

[0029] Here, N i N is the number of characteristic measurement points in running state i, and N0 is the number of characteristic measurement points in the entire damage event.

[0030] For time-weighted allocation, the following applies:

[0031]

[0032] Here, t i t0 is the dwell time of operating state i during the damaging event, and t0 is the total duration of the damaging event.

[0033] For weighted allocation based on incremental changes in load and / or load:

[0034]

[0035] Here, ΔBi is the increment of load and / or load accumulation in operating state i, and ΔB0 is the total increment of load during the damaging event.

[0036] To obtain the time progression of the emerging operating states, it is advantageous to classify the signals of the overall system and thereby define the possible operating states. That is, the operating state is determined by the signals detected by the overall system, particularly by signals also obtained during normal operation of the overall system. In classifying, for example, the numerical range of all signals to be examined in the overall system is divided into multiple levels, i.e., intervals, for each signal individually and completely. Then, each possible combination of the levels of all examined signals in the overall system is defined as an operating state; that is, the possible operating states are defined. This can thus be done, for example, during normal operation and / or after normal operation of the overall system. Therefore, it is not necessary to pre-define the operating states that the overall system must experience, i.e., the operating states experienced in the form of standardized test runs. The method described here is thus significantly more flexible and can be used in the normal customer operation of the overall system, such as a vehicle.

[0037] Advantageously, the duration of a corresponding operating state is determined by forming a multi-dimensional histogram for the signals of the overall system using the levels into which the signals are input, and determining the duration of the corresponding operating state based on the histogram. Advantageously, the histogram is determined by counting the relevant instantaneous values ​​of multiple parameters. That is, at each measurement point where a load measurement and / or load measurement has been detected, the value of the detected signal (i.e., signals detected other than load and / or load) related to the determination of the operating state of the overall system is determined, and the level to which the signal belongs is determined. For each combination of the levels, there exists exactly one element in the histogram, the value of which is relative to the previous measurement point, increasing by the time interval. Each element of the histogram here formally constitutes an operating state. The result is the duration of the operating state of the overall system.

[0038] The alternatives to the processing methods described in the preceding two paragraphs are, for example, the hierarchical classification of the time process for defining possible operating states in order to obtain the operating states that occur, and the determination of the duration of the corresponding operating states. Alternatively, the operating states can be determined by signals related to the determination of the operating states using artificial intelligence methods.

[0039] In order to determine the duration of operating states during normal operation of the overall system, that is, especially for operation where it is not necessary to determine the load and / or load, it is preferable to use a single-parameter dwell time count, in which the dwell time of the overall system in each operating state is determined by the operating state-time process determined according to one of the methods described above, thereby obtaining the duration of each operating state of the overall system during normal operation.

[0040] To this end, it is advantageous to determine the damaging events for each operating state by using a counting method appropriate for converting the parameters from a time range to a frequency range for the existing load and / or load amounts, in which damaging events are determined by the load and / or load detected in the time range within the frequency range, and assigned to the operating state. For this purpose, load analysis and / or load analysis are advantageously performed, in which damaging events are determined by an appropriate counting method and a quantitative characteristic of the damage contribution of the determined damaging events is assigned. Here, all damaging events can be determined first, and in a next step, a damage contribution can be assigned to all events, or a damage contribution can be assigned directly after determining one damaging event, and then the other events can be determined.

[0041] In the load analysis and / or load assessment, particularly according to the nominal stress design or notched stress design for operational stability, it is advantageous to correlate each load and / or load with its associated operational state based on the signals of the overall system. To determine the damaging events that result in load transformations, rainflow classification is performed on the loads and / or loads, generating a separate rainflow matrix for each operational state. Load transformations counted in the rainflow matrix are proportionally allocated to the load states that participated in the formation of the load transformations. This is done, for example, by dividing the corresponding load transformation into two half-load transformations and storing them separately in the rainflow matrix of the operational state corresponding to their reversal points. For the purpose of considering residuals, depending on the type of consideration—specifically, whether there is no load transformation, a half-load transformation, or a complete load transformation—the corresponding load transformation is allocated to the rainflow matrix of the participating operational states, particularly proportionally, in one possible embodiment of the method, at 50% each. The relevant reversal points are particularly the first two reversal points of the load transformation, which are removed from the load-time process and / or load-time process after the load transformation count in the classic rainflow classification.

[0042] In another possible embodiment of the method, the allocation of corresponding load transformations and / or residuals to the participating operating states can be implemented according to one of the methods described above, in order to weight the damaging events. Here, in the load analysis and / or load analysis, it is advantageous to generate a separate rainflow matrix for each operating state, in which load transformations counted in the rainflow matrix are proportionally allocated to the operating states that participated in the formation of the load transformations, in such a way that the corresponding load transformations are weighted according to the operating states in which they were formed and are stored accordingly in the rainflow matrix of that operating state. To take residuals into account, depending on the type of consideration, in particular whether there is no load transformation, a half load transformation, or a complete load transformation, the corresponding load transformations are weighted according to the operating states in which they were formed and allocated to the rainflow matrices of the participating operating states.

[0043] For example, the weight of a corresponding operating state is determined by determining the time period during which the corresponding load change occurs and the total duration of the corresponding operating state within that time period, and then using the ratio of the total operating state duration to the total duration of the corresponding load change as the weight. For example, the total duration of the corresponding load change is considered, according to the nominal stress design, as the time period between the first and last reversal points of the load change, or alternatively, as the time period between the first and last reversal points of the load change minus the time period specified for the occurrence of the included load change. The sum of the weights of all operating states must also correspond to the original frequency of the load change.

[0044] In another possible implementation, the weight of a corresponding operating state is determined as the ratio of the cumulative value of the load increments and / or load increments of the relevant load changes in the corresponding operating state to the total load increments and / or load increments of the corresponding load changes. The total load increments and / or load increments of the relevant load changes are derived from twice the difference between the measured load or load at the reversal point of the relevant load change. In this implementation of the method, the corresponding load changes and / or residuals can be allocated to the participating operating states based on the cumulative value of the load increments and / or load increments, where the total load and / or load increments of the load changes correspond to twice the difference between the measured load or load at the reversal point, and the weight of an operating state corresponds to the ratio of the cumulative value of the load increments and / or load increments occurring in the corresponding operating state to the total load increments and / or load increments of the load changes. This processing corresponds to a weighting of the energy motive of an operating state based on the mechanical work done proportionally within that operating state.

[0045] Thus, in this embodiment of the method, in the load analysis and / or load analysis, a separate rainflow matrix is ​​generated for each operating state, in which counted load changes are assigned to the operating states that participated in the load changes according to the accumulated value of the load increments and / or load increments, where the total load increments and / or load increments of the corresponding load changes correspond to twice the difference between the upper load or load measurement value and the lower load or load measurement value at the reversal point of the load change, and the weight of the corresponding operating state corresponds to the ratio of the accumulated value of the load increments and / or load increments occurring in the corresponding operating state to the total load increments and / or load increments of the load changes. To account for residual values, depending on the type of load change considered, particularly whether there is no load change, a half load change, or a full load change, the corresponding load change is assigned to the rainflow matrix of the participating operating states based on the accumulated value of the load increments and / or load increments. The total load increment and / or load increment of the corresponding load change corresponds to twice the difference between the upper and lower load or load measurement values ​​at the reversal point of the load change, and the weight of the operating state corresponds to the ratio of the accumulated value of the load increments and / or load increments occurring in the corresponding operating state to the total load increment and / or load increment of the load change. Thus, the sum of all weights is always equal to 1, and these weights are multiplied by the original frequency as described above. That is, in the case of residual values, multiplied by 0, 1 / 2, or 1 depending on the specific form of the residual value considered, particularly whether there is no load change, a half load change, or a full load change, and always multiplied by 1 for normal load changes. This applies particularly regardless of what kind of event the damaging event is.

[0046] This can also be expressed by formula (5), where the following applies to the total load increment and / or load increment ΔB0:

[0047] ΔB0=2(B Ob -B Ub (6)

[0048] The weights of running state i thus conform to formulas (5) and (6):

[0049]

[0050] Here G i It is the weight of running state i, ΔB i It is the cumulative value of load increment and / or load change, B Ob It is the measured value of the load and / or the measured value of the load, and B Ub These are the measured values ​​of the load and / or the load measurement.

[0051] In this variant, for normal load transformation H0 = 1, and for residuals H0 ∈ {0, 0.5, 1} based on no / half / full load transformation.

[0052] The load set and / or load set for each operating state is determined by evaluating the rainflow matrix separately for each operating state, in particular by extracting the detected load transformations.

[0053] In an alternative approach, advantageously, each load and / or load of a mechanical stress or strain form is assigned a corresponding operating state based on the signal from the overall system, and rainflow classification is performed on the load and / or load to determine the occurrence of damaging events in the load transformation form. Here, the determined load transformation is directly converted into damage parameters after its determination and / or considering the order of its occurrence, particularly according to notched strain design for operational stability, and the operating states involved in the relevant load transformation (here, particularly the operating states corresponding to the reversal points of the load transformation) are respectively added as damage parameters, particularly proportionally, in one possible embodiment of the method, at half the original frequency, i.e., 1 / 2, to the set of damage parameters for the relevant operating states. The set of damage parameters for each operating state here constitutes the set of loads and / or loads for each operating state.

[0054] In one possible implementation of the method, the allocation of damage parameters to the operating state is weighted, preferably in the presence of one of the methods described above for determining the weights of damage events or load changes, particularly as described in formulas (1) to (7).

[0055] For example, in load analysis and / or load analysis, the determined load transformation is directly converted into damage parameters after the load transformation is determined and / or when the order of occurrence of the load transformation is considered. These damage parameters are then weighted according to the operating states in which they are formed and added to the damage parameter set of the operating states in which the relevant load transformation is involved. The damage parameter set of each operating state constitutes the load set and / or load set of each operating state.

[0056] The weight of each corresponding operating state is determined by identifying the time period in which the corresponding load change occurs and the total duration of the corresponding operating state within that time period. The weight is then calculated as the ratio of the total operating state duration to the total duration of the corresponding load change. In another possible embodiment of the method, the weight of the corresponding operating state is determined as the ratio of the cumulative value of the load increment and / or load increment of the relevant load change in the corresponding operating state to the total load increment and / or load increment of the relevant load change. The total load increment and / or load increment is calculated as twice the difference between the measured load or load value at the reversal point of the relevant load change. Therefore, in this implementation, the allocation of damage parameters to the participating operating states is based on the load increments and / or the accumulated values ​​of load changes as a basis. Here, the total load increment and / or load increment of a load change corresponds to twice the difference between the upper or lower load measurement value at the reversal point, and the weight of an operating state is determined as the ratio of the accumulated value of the load increments and / or load increments of the load changes occurring in the corresponding operating state to the total load increment and / or load increment of that load change. This processing method corresponds to the weighting of the energy motive of the operating state based on the mechanical work done proportionally within an operating state.

[0057] This has been illustrated for damage events in formulas (1) to (7) above, where load transformation is a specific form of damage event.

[0058] Therefore, the method according to the invention can be used to perform operating state-related load analysis and / or operating state-related load analysis starting from load analysis and / or load analysis suitable for describing damage mechanisms according to the corresponding prior art, but particularly without limitation, is load analysis and / or load analysis according to nominal stress design, notch stress design and / or notch strain design for operating stability. Those skilled in the art also know methods for performing load analysis and / or load analysis based on important damage mechanisms, in which damage events are determined based on load-time processes and / or load-time processes in order to describe the damage state, and the method according to the invention can also be used in these methods.

[0059] Advantageously, the time-normalized load set and / or load set for each operating state can be determined by normalizing the load set and / or load set for each operating state according to the frequency of its respective set level over the duration of the corresponding operating state, and / or by normalizing the load set and / or load set for each operating state based on the duration of the relevant operating state. Here, the set level is defined as a damage characteristic having a damage feature value that quantitatively describes the damage content and the frequency of occurrence of damaging events having said damage feature value. For example, the set level according to the nominal stress design or notch stress design for operating stability is described by load changes having a quantitative damage feature value “amplitude” and the frequency of load changes having this damage feature. Alternatively, according to the nominal strain design, the set level can also be given as an item in the rainflow matrix, in which case the quantitative damage feature value exists by the row and column indices and the frequency of the corresponding values ​​as matrix elements. In the latter case, the normalization of the load set and / or load set can be performed based on extrapolation of the rainflow matrix known to those skilled in the art.

[0060] In an alternative implementation, the frequency normalization of each set level can also be performed for other parameters, such as the distance traveled for a moving total system and / or the area processed for a total system used for land processing in the field of agricultural machinery. Thus, the load set and / or load group can also be referred to as distance normalization and / or area normalization, respectively. The duration of determining the operating state is correspondingly replaced and / or supplemented by determining the distance traveled and / or the area processed for each operating state.

[0061] The method can be implemented, for example, by a device that is directly mounted on the main system, i.e., airborne, or partially mounted on and partially external to the main system, i.e., located outside the main system. The device outside the main system particularly includes a processing unit, such as a computer-like processing unit. For a device fully mounted on the main system, the processing unit is, for example, a controller or other computer unit. The method, outside the main system, can be implemented, for example, based on externally recorded data, particularly the load and / or the time progression of the load on the component under investigation, and signals from the main system related to determining the operating state. For this purpose, the device includes, for example, a corresponding detection unit on the main system that transmits data to external units, particularly to the processing unit for recording and processing. Alternatively, the data can be recorded first at the main system and then transmitted to the processing unit later for evaluation. This data transmission can be, for example, wireless, e.g., via radio, or it can be wired.

[0062] The device according to the invention is thus constructed and configured for performing the method.

[0063] The apparatus includes, for example: at least one unit for detecting the load and / or load of the component, particularly the sensor system mentioned above; at least one unit for detecting signals of the overall system related to determining the operating state; and at least one processing unit. The processing unit is specifically configured and arranged to perform the steps of the method described above, particularly for processing the detected load and / or load of the component and the signals of the overall system related to determining the operating state, especially in the manner described above.

[0064] As already mentioned, the device may, for example, be entirely located on or can be located on the overall system, or at least one component of the device, particularly the at least one unit for detecting the load and / or load of the component and / or the at least one unit for detecting signals of the overall system related to determining the operating state, may be located on or can be located on the overall system. The processing unit is, for example, a component outside the overall system of the device, such as a computer. Attached Figure Description

[0065] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0066] To avoid misunderstanding, it should be noted that the terms “load set, load analysis, load-time progression, load curve, load measurement, and load measurement value” applicable above refer to the use of loads to perform the method. Similarly, the method may alternatively or additionally be performed using loads, in which case the alternatively or additionally applicable terms above: load set, load analysis, load-time progression, load curve, load measurement, and load measurement value shall apply accordingly.

[0067] in:

[0068] Figure 1 The schematic diagram illustrates a method flow for determining the damage state of at least one component according to the present invention. Detailed Implementation

[0069] Figure 1This diagram schematically illustrates a method flow for determining the damage state of components, particularly mechanical components. This method enables load analysis in relation to operational conditions to reconstruct a set of future and / or historical loads for the component. In particular, this invention enables the determination of a set of operational conditions-related loads to reconstruct future and / or historical load sets to describe the current damage state of the component, without having to perform load measurements on the component. This is contingent upon the prior determination of a time-normalized, operational condition-related set of loads for identical or similarly constructed components by corresponding load measurements on the component.

[0070] Alternatively or additionally, the method, as described below, determines the damage state of a component, particularly a mechanical component, not through load, but through load analysis. This method then enables load analysis related to the operating state, for reconstructing a set of future and / or historical loads on the component. Specifically, this method can determine a set of operating state-related loads to reconstruct future and / or historical load sets to describe the current damage state of the component, without having to perform load measurements on the component for this purpose. Therefore, in the following description involving loads, the corresponding term or part of the term "load" can be replaced by the term or part of the term "load" to describe methods similar to those for load analysis.

[0071] The method can be applied to components to which a load set can be assigned based on a load-time process (BSV), i.e., based on the time progression of the loads occurring on the component. This load set must be obtained from the load-time process (BSV) by determining the damage-related events. Loads and / or loads can be, for example, mechanical stress, a quantity proportional to mechanical stress, temperature, pressure, voltage, current, or any other physical quantity.

[0072] As is known from existing technology, load analysis of machinery is performed in the sense of operational stability. To this end, events related to material mechanical damage are detected starting from the load-time function and processed into a set of loads, such as a set of load amplitudes (nominal stress design, notched stress design) or a set of damage parameters (notched strain design).

[0073] First, the mechanical load is classified by rainflow. The resulting rainflow matrix determines the detected load transformations based on amplitude, average value, and frequency. Residual values ​​remain in the rainflow classification (points where no load transformation reversal occurs). These residual values ​​are either ignored, considered as partial load transformations, or considered as complete load transformations, depending on the application. The resulting amplitude-average value set can be transformed into a substitute amplitude set using a Hertzsprung's diagram, thereby obtaining the set of loads to be discovered.

[0074] However, since the measurement period is usually significantly shorter than the service life of the component, the obtained load set must be extrapolated to obtain a design set (Bemessungskollektiv) for the entire service life of the component. According to current technology, this extrapolation method is primarily based on the ratio of the measurement period to the planned service life. It is assumed that the load will behave as it does during the service life, just as it did during the measurement. This requires accurate knowledge of future usage patterns in the test plan to correctly reflect the composition of different usage scenarios and obtain a design set that is as representative as possible.

[0075] The resulting design set can be used to perform theoretical and experimental lifespan analyses.

[0076] Determining the current damage state of a component requires a current load set that reflects the component's load history. This necessitates either measuring the load over the entire service life period or reconstructing it based on a previously determined design set.

[0077] This reconstruction of the current load set has so far been primarily based on the ratio of current to planned service life. Consequently, the reconstructed current load set has no difference in shape from the design set, and when considering relative service life using the damage accumulation hypothesis, only the relevant service life is effectively compared. Based on this fact, using this reconstructed load set for damage assessment is problematic in the sense of preventative maintenance, that is, predictive maintenance.

[0078] Furthermore, existing technologies also employ methods that obtain separate load sets from fixed test runs for selected, predefined operating states, and then weight these load sets to form a design set based on usage profiles. Reconstruction of the current load set based on usage profiles is thus possible. However, this approach leads to high overhead in test planning and execution. The number of operating states that can be handled is also significantly limited.

[0079] Therefore, the method described below extends the load analysis of components, enabling the determination of damage-related events based on the operating states of the overall system. If the time-normalized load set for the operating states is known, the method thus allows the load set for a specific component to be reconstructed solely from the detection of the operating states of the overall system. This load set has advantageously been predetermined according to the method described herein. This can, for example, be used for online monitoring of the current damage state of a component without having to determine the load progression on the component itself.

[0080] In this method, for a component, the damage state is advantageously determined based on the operating state of the overall system containing the component. This is done by detecting the time progression of the component's load and the time progression of a signal S related to determining the operating state of the overall system during the operation of the overall system. The detected signal S determines the operating state of the overall system, particularly the time progression of the operating state. The detected load determines a damaging event and associates the damaging event with the operating state. A load set BB for each operating state is determined, and a time-normalized load set ZBB for each operating state is determined based on the duration of the corresponding operating state. Using the time-normalized load set ZBB for each operating state, the damage state can now, for example, be determined based on the operating state of the overall system for that component. Alternatively or additionally, for example, a future damage state can be determined based on a predicted future operating state of the overall system containing the component. Alternatively or additionally, for example, for another component constructed identically to the component, the damage state can be determined based on the operating state of another overall system containing the other component and constructed identically to the overall system. Alternatively or additionally, for example, the future damage state can be determined based on the predicted future operating state of another overall system containing the other component.

[0081] The following explanation uses the mechanical stress of a component as an example to illustrate this processing method, where the mechanical stress exists as a discrete time process. Therefore, the load-time process (BSV) of the component is detected as data; that is, the load is detected specifically as a discrete time process, i.e., the time process of the measured load value of the component. The load-time process (BSV), constructed as a load curve in this example, is also used as additional data to detect the signal S of the overall system above the component. Specifically, the time process of this signal S is detected. This is particularly relevant to determining the operating state of the overall system. The signal S is particularly useful during the normal operation of the overall system.

[0082] In the first method step VS1, the signal S is first divided into multiple ranges, particularly levels, that characterize the operating states, in order to define the possible operating states of the overall system. The result is the operating state-time process BZV, which is the time process of the operating states that have occurred.

[0083] For the signal S with such defined levels, a multidimensional histogram is then formed in the second method step VS2, and the duration of the corresponding operating state is determined based on the histogram, wherein the histogram has one dimension for each of the signals S.

[0084] For example, to analyze components of a vehicle constructed as a machine, this signal could be, for instance, the speed of travel. A range is given by a defined operating point. Thus, for example, four levels can be defined, within which the speed of travel typically falls:

[0085] stationary (speed = 0 km / h),

[0086] City driving (speed between 0km / h and 50km / h),

[0087] Off-road driving (speeds between 50km / h and 100km / h),

[0088] Driving on highways (speed greater than 100km / h).

[0089] The histogram is determined in the second method step VS2 by counting the relevant instantaneous values ​​of multiple parameters. That is, at each measurement point where a load measurement is detected, the values ​​of all detected signals S (i.e., signals detected other than mechanical stress) related to the determination of the overall system's operating state are determined, and the level to which said signal S belongs is determined. For each combination of these levels, there exists exactly one element in the histogram whose value increases by a time interval compared to the previous measurement point. Each element of the histogram here formally represents an operating state. Therefore, the result of the second method step VS2 is the duration DB of the overall system's operating state.

[0090] Therefore, the actual progress of each operational state and the actual duration of the corresponding operational state are determined, without relying on a pre-defined sequence and duration of operational states from the overall system. This method can thus be executed, for example, during the actual operation of the overall system, thereby achieving more realistic results compared to pre-defined operational states executed during test runs. In particular, its advantage over current methods is achieved by automatically identifying transitions in operational states without considering or intentionally inducing such transitions.

[0091] As the next and therefore third method step, VS3, a load analysis is now performed, which is modified from the load analysis described above that has been used so far. That is, the determination of damage-related events, i.e., events that damage the components, is modified compared to the processing method known in the prior art described above. To this end, each load measurement value is first mapped to its associated operating state based on the signal S of the overall system. The input data are thus the determined operating state-event curve BZV and load-time process BSV. Thereafter, a modified form of rainflow classification is performed. Here, the determination of damage events, in this example, load changes, is performed according to the classic rainflow classification. The actual modification is made when storing these events that damage the components.

[0092] According to the classical method, load transformations are stored in a rainflow matrix. In contrast, in the method described here, a separate rainflow matrix is ​​generated for each operating state, in which counted load transformations are proportionally allocated to the operating states that respectively participated in the formation of the load transformations.

[0093] The load transformation, according to the classic rainflow classification, includes two reversal points removed from the load curve after counting the load transformations. These two reversal points are characterized by load measurements and / or load measurements, respectively. However, in the method described here, in addition to the load measurements at the reversal points, the relevant operating conditions are also known. Therefore, the corresponding load transformation is divided into two half-load transformations and stored in the rainflow matrix corresponding to the operating conditions at the reversal points.

[0094] The consideration of residuals is similar. Here, the load transformation is also distributed 50% to the rainflow matrix of the participating operating states, depending on the form of consideration: no load transformation, half load transformation, or full load transformation. Only for the special case where only one operating state occurs, the modified rainflow classification result corresponds to the classical rainflow classification result described above.

[0095] In an alternative implementation, the allocation is based on a weighted average of the energy motives described above. For this purpose, unlike traditional rainflow classification, not only the reversal point is considered, but all load measurements and their corresponding operating states are examined over the duration of the load change. For each time interval, the value of the load increment, i.e., the load difference between two moments, is assigned to the operating state at the start of the time interval. The weight of the energy motive for an operating state over the entire load change is then calculated as the sum of the load increments assigned to that operating state divided by twice the difference between the highest and lowest load values ​​of that load change. The load change is then added as the product of its frequency and the weight of the energy motive for that operating state to the corresponding entry in the rainflow matrix. The position of the load change in the rainflow matrix of the operating states, i.e., the row and column numbers, remains unchanged relative to the prior art. Except for the last measurement point, all measurement points of the load change are then removed from the load-time progression and the operating state-time progression, thereby removing the included load changes from further counting, as also done in classical rainflow classification. Based on this processing method, for the special case where only one operating state occurs, the result also corresponds to the result of the classic rainflow classification.

[0096] The result of the third method step VS3 is the event EB that damages the component in each operating state, that is, the allocation of the event that damages the component to each operating state.

[0097] In this way, the rainflow matrix is ​​advantageously evaluated individually for each operating state in at least one other, here fourth method step VS4, to obtain a load set BB for each operating state as a result. Here, the detected load transformation, in the illustrated embodiment, is mechanical stress, extracted according to load amplitude, average load, and frequency. The consideration of average load is preferably performed on the amplitude of the damage equivalent, without average load, according to the Hertzsprung's diagram. In this way, the load set BB for each operating state is obtained as an equivalent amplitude set. The load set is normalized to the duration of the corresponding operating state at its respective set level in another, here fifth method step VS5. That is, the range of the load set is normalized using the duration of each relevant operating state. Thus, in this example, the load amplitude as an average load transformation gives a quantitative description of the damage content of the damage characteristics, the load amplitude also known by the name "equivalent load amplitude" as is known to those skilled in the art.

[0098] In one possible implementation of the method, this can be added to affect the load magnitude of each set level during the normalization process. This is done while considering the static characteristics of the distribution of the highest load values ​​and the ratio of the operating state to the predicted total lifetime of the overall system.

[0099] The result is the time-normalized load set ZBB for each operating state. In other words, there is a time-normalized load set for each operating state. This represents the relevant, time-dependent damage characteristics for the corresponding operating state.

[0100] The method flow for determining the time-normalized load set ZBB for each operating state by the load-time process BSV for the component and the time process of the signal S of the total system including the component is a coupled load analysis GB.

[0101] The described processing method is exemplarily used for nominal stress design or notched stress design based on operational stability to evaluate mechanical stress-time progression, but is not limited to this application. The principle of the method described herein is based on the correspondence between the overall system's operating state and the relevant load measurements of the components and the generation of corresponding individual damage characteristic groups for the corresponding operating state, with damage events being equally distributed across the operating states involved in the time.

[0102] By determining the time-dependent, i.e. time-normalized load set ZBB for each operating state as described above, the load set can now be reconstructed for each arbitrary time period for a component and / or for the same component of the same overall system, using the load set as described above and in... Figure 1 The method shown in the diagram determines the time-normalized load set ZBB for each operating state. Therefore, the occurrence of the operating state of the components whose load sets should be reconstructed, specifically the components of the overall system having said components, is only necessary within the time interval under consideration, that is, the time interval of interest.

[0103] For example, the occurrence of the operating state can be determined by detecting the actual operating state of the total system according to the time process of the detection of the signal S of the total system according to method steps VS1 and VS2 as described above, or by pre-defining an operating state process, for example, derived from usage predictions for the total system. The time-dependent, i.e., time-normalized load set ZBB for each operating state is now extrapolated to the duration of the corresponding operating state, for example, by multiplying its frequency by the duration of the corresponding operating state, and then superimposed over all operating states to form the total load set.

[0104] In one possible implementation of the method, this can also affect the magnitude of the load and / or load at each set level during the extrapolation process. This is done taking into account the normalized static characteristic values ​​of the distribution of the highest load values ​​and, for example, the ratio of an operating state to the predicted total lifetime of the overall system.

[0105] In another possible implementation of the method, normalization and extrapolation are performed based on the extrapolation of the rainflow matrix, where method step VS4 is not performed initially. Instead, normalization of the rainflow matrix is ​​performed in method step VS5 as an extrapolation to the duration of a lower operating state. Before determining the total load set, the rainflow matrix of the operating state is first extrapolated based on the duration of the corresponding operating state. Next, method step VS4 is executed to determine the current load set of the operating state based on the current rainflow matrix, and finally, all current load sets of the resulting operating states are superimposed.

[0106] This superimposed total load set is the reconstructed load set for the component, which is, for example, the load set described above and in... Figure 1 The method shown in the figure determines the components of the time-normalized load set ZBB for each operating state, and / or the same component for the same other total system.

[0107] A particular advantage of this approach is that, in order to reconstruct the load set, it is not necessary to re-implement load measurements within the corresponding time frame. Based on this, the current damage state of a component can be derived by appropriately selecting the reconstructed load set. For this purpose, a load set is needed that corresponds, in the sense of the failure criteria, to the loads experienced over the entire service life of the component. This can be determined, for example, by constructing a histogram for the occurrence of operating states through extrapolation of measurements for the expected service life, or by taking measurements over the expected service life across a set of identical overall systems, such as an entire fleet, and superimposing the individual histograms of all systems, such as all vehicles. The design set for the component is determined from the histogram by reconstructing the time-dependent, i.e., time-normalized load set for each operating state.

[0108] Furthermore, a histogram of operating states is determined over the total service life of the overall system, such as a vehicle. Based on this histogram, the load set of the overall system, such as the vehicle, can be reconstructed up to the current moment. By comparing the currently reconstructed load set with the design set, the current damage to components can be determined comparatively with the design set. A suitable method is constructed based on Palmgern and Miner's linear damage accumulation hypothesis, for example, according to the "Miner basic" improvement scheme.

[0109] The applicability of this method for determining damage presupposes that knowledge is available. The curve. This curve can be determined experimentally or estimated using appropriate methods (e.g., according to FKM guidelines). The total damage is calculated by linear damage accumulation for both the rebuilt load set and the design set; the ratio of the two damage sums represents the current damage and current damage state of the component. Therefore, the damage state of the component can be monitored online by continuously updating the currently rebuilt load set via a histogram showing the updated operating status.

[0110] The processing method is used, for example, to determine the damage of the equivalent amplitude set as a load set based on nominal stress design or notched stress design based on operational stability, but is not limited to this application. The principle is based on individually determining the current reconstructed load set and design set according to appropriate methods for determining and comparing the damage thus obtained.

[0111] As another embodiment, the following describes a processing method for improved load analysis based on notch stress design. In this case, only the specific implementation of steps VS3 and VS4 differs from the processing method described above. In step VS3, damaging events in the form of load transformations are determined by rainflow classification. However, these events are not stored in the rainflow matrix as described above, but are directly converted into load characteristic values ​​(here: damage parameters) in step VS4 according to the notch stress design for operational stability, taking into account the order of occurrence of the events. The damage parameters thus determined are added proportionally, for example, as damage parameters at 50% of the original frequency (i.e., 0.5) to the set of damage parameters for the relevant operational states, based on the operational states involved in the relevant load transformations (e.g., operational states associated with the reversal point of the load transformation). In this way, the load set BB for each operational state is obtained as the set of damage parameters for each operational state.

[0112] In an alternative implementation, the allocation of damage parameters to operating conditions is preferably performed according to the weighting described above, particularly the weighting of energy motives.

[0113] Another approach is at least substantially similar to the example above. The determination of damage is partly the only difference here. Based on notch stress design, instead of the conventional... The curve requires the damage parameters corresponding to the damage parameters used. Curve. The determination of the total damage is similar to that of the Miner rule.

[0114] The method can be implemented, for example, by a device that is directly mounted on the main system, i.e., airborne, or partially mounted on the main system and partially designed to be external, i.e., located outside the main system. Outside the main system, the device particularly includes a processing unit, such as a computer-like processing unit. For a device fully mounted on the main system, the processing unit is, for example, a controller or other computer unit. Outside the main system, the method can be implemented, for example, based on externally recorded data, particularly the time progression of the load on the examined component, and signals from the main system related to determining its operating state. For example, the device includes a corresponding detection unit on the main system that transmits the data to an external unit, particularly to the processing unit, for recording and processing. Alternatively, the data can be recorded on the main system first, and then transmitted to the processing unit later for evaluation. This data transmission can be wireless, e.g., via radio, or wired.

[0115] This invention is thus constructed and configured for performing the method.

[0116] The apparatus, for example, includes at least one unit for detecting the load on a component, at least one unit for detecting a signal S related to determining the operating state of the overall system, and at least one processing unit. The processing unit is specifically configured and arranged to perform the steps of the method described above, particularly for processing the detected load on the component and processing the signal S related to determining the operating state of the overall system, especially in the manner described above.

[0117] As already mentioned, the device may be entirely or can be installed on the overall system, or at least components of the device, particularly at least one unit for detecting the load on the component, and / or at least one unit for detecting signals S related to determining the operating state of the overall system, may be installed on the overall system. In this case, the processing unit is, for example, a component outside the overall system of the device, such as a computer.

[0118] List of reference numerals

[0119] BB Load set for each operating state

[0120] BSV Load-Time Process

[0121] BZV running status - time process

[0122] DB running time

[0123] Damaging events in each operating state of EB

[0124] GB Coupled Load Analysis

[0125] S signal

[0126] ZBB's time-normalized load set for each running state

[0127] VS1 First Method Steps

[0128] VS2 Second Method Steps

[0129] VS3 Third Method Steps

[0130] VS1 Fourth Method Steps

[0131] VS5 Fifth Method Steps

Claims

1. A method for determining a damage state of a component, wherein, For a component, the damage state of the component is determined from the operating states of the overall system comprising the component in such a way that, during operation of the overall system, the time course of the load and / or the loading of the component is detected and the time course of signals (S) of the overall system relevant for determining the operating states is detected, from the detected signals (S) the occurring operating states of the overall system are determined, from the detected load and / or loading damage events are determined and are correlated to the operating states, a load set and / or a loading set (BB) is determined for each operating state and from the duration of the respective operating state a time-normalized load set and / or loading set (ZBB) is determined for each operating state, by means of the time-normalized load set and / or loading set (ZBB) of each operating state: for the component, the damage state is determined from the occurring operating states of the overall system and / or the future damage state is determined from the predicted future occurring operating states of the overall system comprising the component, and / or for another component which is identically constituted as the component, the damage state is determined from the occurring operating states of another overall system comprising the other component and identically constituted as the overall system, and / or the future damage state is determined from the predicted future occurring operating states of the other overall system comprising the other component, wherein from the detected load and / or loading damage events are determined and are correlated to the operating states, and from this damage events (EB) are determined for each operating state in such a way that a load analysis and / or a loading analysis is carried out, wherein in the load analysis and / or the loading analysis: each load and / or loading is correlated to the operating state associated therewith on the basis of the signals (S) of the overall system, a rainflow classification is carried out for the loads and / or the loadings for determining the occurrence of damage events in the form of load transitions, and wherein in the load analysis and / or the loading analysis: for each operating state a separate rainflow matrix is generated in which the counted load transitions are proportionally distributed to the operating states which respectively participated in forming the load transitions, for taking account of residual values, the respective load transitions are distributed to the rainflow matrix of the participating operating states depending on the type of consideration, or, in the load analysis and / or the loading analysis, directly after determining the load transitions and / or in consideration of the sequence in which the load transitions occurred, the determined load transitions are converted into damage parameters and are added to the damage parameter set of the operating states participating in the relevant load transitions, the damage parameter set of each operating state forming the load set and / or the loading set (BB) of each operating state.

2. The method of claim 1, wherein, a classification of the signals (S) of the overall system is carried out and from this the definition of the possible operating states is carried out in order to obtain the time course of the occurring operating states.

3. The method of claim 2, wherein, The duration of the respective operating state is determined in that, for the signal (S) of the overall system, a multi-dimensional histogram is formed from the signal fractions, and the duration of the respective operating state is determined from the histogram.

4. The method of claim 3, wherein, The determination of the histogram takes place by counting the relevant instantaneous values of the multi-parameter.

5. The method of claim 1, wherein, In the load analysis and / or the load analysis, For each operating state, a separate rainflow matrix is generated in which the counted load changes are proportionally distributed to the operating states which respectively participated in the formation of the load changes, in that the respective load change is divided into two half load changes and is respectively stored in the rainflow matrix of the operating state which belongs to the reversal point, In order to take account of the residual value, the respective load change is distributed to the rainflow matrix of the participating operating state in accordance with the type of consideration, 50% respectively to the rainflow matrix of the participating operating state.

6. The method of claim 1, wherein, In the load analysis and / or the load analysis, For each operating state, a separate rainflow matrix is generated in which the counted load changes are proportionally distributed to the operating states which respectively participated in the formation of the load changes, in that the respective load change is weightedly distributed to the operating state which participated in its formation and is respectively stored in the rainflow matrix of this operating state, In order to take account of the residual value, the respective load change is weightedly distributed to the rainflow matrix of the participating operating state in accordance with the type of consideration.

7. The method of claim 6, wherein, The weight of the respective operating state is determined in that the time period in which the respective load change occurred and the total duration of the respective operating state within the time period are determined and the weight is derived as the ratio of the total operating state duration of the respective operating state to the total load increment and / or load increment of the respective load change.

8. The method of claim 6, wherein, The weight of the respective operating state is determined as the ratio of the accumulated load increment and / or load increment of the relevant load change in the respective operating state to the total load increment and / or load increment of the relevant load change, the total load increment and / or load increment being derived from the double difference of the upper load or load measurement value and the lower load or load measurement value of the reversal point of the relevant load change.

9. The method of claim 1, wherein, The load set and / or the load set (BB) of each operating state is determined in that the rainflow matrix is evaluated separately for each operating state.

10. The method of claim 1, wherein, In the load analysis and / or the load analysis, the determined load changes are converted into damage parameters directly after the determination of the load changes and / or taking account of the sequence in which the load changes occurred, and are respectively added to the damage parameter set of the operating states which participated in the relevant load change as damage parameters at half the original frequency, the damage parameter set of each operating state forming the load set and / or the load set (BB) of each operating state. In the load analysis and / or the load analysis, the determined load changes are converted into damage parameters directly after the determination of the load changes and / or taking account of the sequence in which the load changes occurred, and are respectively added to the damage parameter set of the operating states which participated in the relevant load change as damage parameters at half the original frequency, the damage parameter set of each operating state forming the load set and / or the load set (BB) of each operating state.

11. The method of claim 1, wherein, In the load analysis and / or the load analysis, the determined load changes are converted into damage parameters directly after the determination of the load changes and / or in consideration of the sequence in which the load changes occur, and are added to the damage parameter set of the operating state participating in the relevant load change as damage parameters weighted by the operating state, respectively, and are distributed in each operating state's load set and / or load set (BB) consisting of the operating states participating in the relevant load change.

12. The method of claim 11, wherein, The weight of the respective operating state is determined in such a way that the time period in which the respective load change occurs and the total duration of the respective operating state within said time period are determined, and said weight is derived as the ratio of the total operating state duration of the respective operating state to the total duration of the respective load change.

13. The method of claim 11, wherein, The weight of the respective operating state is determined as the ratio of the cumulative value of the load increments and / or load increments of the relevant load change in the respective operating state to the total load increment and / or load increment of the relevant load change, said total load increment and / or load increment being derived from the double difference of the upper load or load measurement value and the lower load or load measurement value of the reversal point of said relevant load change.

14. The method of any one of claims 1 to 4, wherein, A time-normalized load set and / or load set (ZBB) is determined for each operating state in such a way that the load set and / or load set (BB) of each operating state is normalized to the duration of the relevant operating state in terms of the frequency of its individual set levels.

15. Apparatus, which is constructed and arranged to carry out the method according to any one of claims 1 to 4.

16. Apparatus according to claim 15, comprising: at least one unit for detecting the load and / or load of the component, at least one unit for detecting the signal (S) of the total system relevant to the determination of the operating state, and at least one processing unit.

17. The apparatus of claim 15, wherein, The apparatus is completely arranged or can be arranged on the total system, or at least one component of the apparatus is arranged or can be arranged on the total system.

Images