MACHINE CONDITION MONITORING

Description

[0001] In the operation of complex machinery, there is an increasing trend to replace the previously common reactive or preventive maintenance with dynamic maintenance oriented towards the machine's actual state of health. Preventive maintenance involves shutting down, inspecting, and servicing the machine at regular intervals. However, this often results in the replacement of components that are still intact and whose remaining service life has not yet been exhausted. Reactive maintenance involves shutting down the machine when a malfunction occurs, repairing it, and servicing it if necessary. However, unexpected malfunctions and the resulting downtime can lead to problems, especially if the malfunction occurs during a critical operating phase.

[0002] Modern monitoring systems, known as "condition monitoring" systems, enable condition-based maintenance strategies that allow for better utilization of the service life of critical machine components and more flexible alignment of maintenance measures with the machine's operational requirements. In particular, such monitoring systems can measure the stresses on machine components or parts and / or count load cycles. Based on established damage accumulation models, the remaining service life of machine components or parts can then be statistically estimated, from which optimized maintenance cycles can be derived.

[0003] A fundamental problem lies in measuring or determining indicators that are meaningful regarding the stress or health of the machine. However, especially with complex machines, critical machine components are often difficult to access for measurement.

[0004] It is usually technically very complex to install sensors on the often highly stressed rotor of a motor and to transport sensor signals from there to a monitoring device.

[0005] It is common practice to evaluate sensor readings from easily accessible locations on a machine for condition monitoring. These readings can then be used, in many cases with the aid of physical simulation models, to infer the stress on less accessible machine components. However, a further problem is that measured stresses often have very different effects on the machine's health. In particular, these effects can vary considerably depending on the machine's location, configuration, modifications, maintenance performed, and / or environmental conditions.

[0006] D1 discloses a method for managing the health status of at least one rotating system. The method includes receiving operational data associated with the rotating system in real time from one or more sensor units via a processing unit. The operational data includes parameter values ​​corresponding to the operation of the rotating system. Furthermore, a virtual copy of the rotating system is configured using the operational data. The behavior of the rotating system is simulated on a simulation instance of the rotating system based on the configured virtual copy. The simulation results are analyzed to determine an abnormality in the health status of the rotating system. The abnormality corresponds to the health status of an internal component of the rotating system.

[0007] D2 discloses a system analysis device comprising a correlation model storage unit and a display control unit. The correlation model storage unit stores the correlation model, which expresses correlations between metrics in the system. The display control unit subdivides a display area into n subdivided areas such that an area of ​​one subdivided area is equal to or greater than the area of ​​a subdivided area. The display control unit assigns each of the plurality of clusters obtained by tracing correlations contained in the correlation model to the subdivided area.

[0008] D3 discloses a controller that is operatively connected to an electric actuator via a network. The controller is configured to receive torque curve data from the electric actuator via the network, wherein the torque curve data is associated with an actuation of an electric motor of the electric actuator, the actuation causing a flow control element of a motor-driven valve to move, the flow control element being mechanically coupled to the electric motor and movable between an open position and a closed position.The controller is configured to determine an area under a torque curve based on the torque curve data, determine a variance between the area under the torque curve and an area under a reference curve, and generate an initial control signal indicating that the electric actuator is healthy if it is determined that the variance does not exceed the variance threshold.

[0009] The object of the present invention is to provide a method and a monitoring device for monitoring the condition of a machine, by which the health status of the machine can be better determined.

[0010] This problem is solved by a method with the features of claim 1, by a monitoring device with the features of claim 13, and by a computer program product with the features of claim 14. For condition monitoring of a machine, in particular a motor, a robot, a machine tool, a production plant, a turbine, a 3D printer, an internal combustion engine, and / or a motor vehicle, current operating data of the machine are continuously recorded. Based on the operating data, the current operating behavior of a machine component is continuously simulated by a concurrent simulation module. Furthermore, performance values ​​quantifying the current performance of the machine component are continuously derived from the simulated operating behavior and stored over time.Furthermore, a performance norm is regularly determined based on a large number of previously derived performance values. The operating data and / or performance values ​​are also monitored to detect whether a predefined initial change pattern occurs. Upon detection of this initial change pattern, a performance reference value is updated to reflect the current performance norm. Additionally, the current performance values ​​are continuously compared to the current performance reference value. Depending on the comparison result, the current health status of the machine component is then displayed.

[0011] To carry out the method according to the invention, a monitoring device, a computer program product and a computer-readable, preferably non-volatile storage medium are provided.

[0012] The method and monitoring device according to the invention can be carried out or implemented, for example, by means of one or more computers, processors, application-specific integrated circuits (ASIC), digital signal processors (DSP) and / or so-called "Field Programmable Gate Arrays" (FPGA).

[0013] A particular advantage of the invention is that performance reference values ​​can be dynamically and automatically adjusted to changes in the machine's operating conditions. In this way, operational performance changes, for example, due to maintenance of individual components, a modified configuration, or altered environmental conditions, which often do not significantly affect the machine's health, can be taken into account when assessing its health status.

[0014] Advantageous embodiments and further developments of the invention are specified in the dependent claims.

[0015] According to an advantageous embodiment of the invention, the first change pattern can specify a threshold for a fluctuation amplitude, a temporal gradient, and / or a temporal variance of the performance values ​​and / or the operating data. Alternatively or additionally, the first change pattern can specify a distance of a performance value from the current performance reference value and / or a duration during which a fluctuation amplitude of the performance values ​​and / or the operating data lies below a threshold. Using such a first change pattern, operating phases in which performance changes only slightly and which are therefore suitable as reference operating phases can be specifically detected. Consequently, an update of the performance reference value can be initiated when such a reference operating phase is detected.

[0016] According to a further advantageous embodiment of the invention, current environmental data of the machine can be continuously recorded. The comparison between the current performance values ​​and the current performance reference value can then be made based on the current environmental data. Environmental data can include, in particular, ambient temperature, humidity, installation location, time of day, day of the week, calendar date, and / or season. In this way, environmental influences on performance can be taken into account when assessing the machine's health.

[0017] Furthermore, the environmental data can be monitored to detect whether a predefined second change pattern occurs. Upon detection of this second change pattern, the performance reference value can be updated with the current performance norm. Such second change patterns can be used to specify thresholds, fluctuation amplitudes, gradients, variances, and / or durations within the time course of the environmental data. Using such a second change pattern, operating phases can be specifically detected in which one or more environmental conditions of the machine change only slightly and which are therefore suitable as reference operating phases. Consequently, an update of the performance reference value can be initiated when such a reference operating phase is detected.

[0018] Similarly, the machine can be monitored to detect whether a machine component and / or part is replaced or modified. Upon detection of such a replacement or modification, the performance reference value can be updated with the current performance norm, optionally with a predefined or calculated time delay. In this way, changes to the machine can be taken into account when assessing its health status.

[0019] According to a further advantageous embodiment of the invention, an operating phase can be detected to determine the performance normal value. This phase is characterized by a below-average fluctuation amplitude in the performance values ​​and / or a value below a predetermined threshold. Additionally or alternatively, a moving average of the performance values ​​can be calculated to determine the performance normal value. In particular, the performance normal value can be set to a current moving average of the performance values ​​if an operating phase with low fluctuations in the performance values ​​is detected. As mentioned above, such an operating phase is often suitable as a reference operating phase.

[0020] According to an advantageous further development of the invention, a parallel predictive module can continuously predict the future health status of the machine component based on the simulated operating behavior. In particular, a time of probable damage occurrence, a probable remaining service life of the machine component, and / or a temporal development of fatigue phenomena can be predicted.

[0021] The predictive module continuously determines the load on the machine component based on simulated operating behavior. From this load, the future health status can then be predicted using a dynamic load model, wear model, degradation model, or lifetime model. A variety of well-known dynamic models and efficient numerical methods for their evaluation are available for such predictions. In particular, a combination of extrapolating historical health states and one or more physical degradation models can be used. The latter dynamically transfer loads such as stresses, vibrations, or high temperatures to structural mechanics and / or their material properties. Furthermore, empirical remaining service life curves can also be considered in the prediction.

[0022] Advantageously, the dynamic load model, wear model, or lifetime model can be used to determine forecast uncertainty, confidence interval, and / or probability of damage. Forecast uncertainty can arise from inaccuracies in the forecast model(s) used, inaccuracies in the acquisition of operational data, and uncertainties in boundary conditions. In this way, the reliability of a predicted health status can generally be assessed much more accurately.

[0023] According to a further advantageous embodiment of the invention, when comparing a performance value with the performance reference value, a distance between the performance value and the performance reference value, a temporal gradient of the distance, and / or a temporal variance of the distance can be determined. The current health status can then be displayed depending on the distance, the gradient, and / or the variance. In particular, during phases with a comparatively high variance of the distance, the current distance and / or the current gradient of the distance can be weighted less heavily when assessing the health status. In this way, it can often be avoided that short-term fluctuations in the distance impair the display of the health status through artifacts.

[0024] According to a further advantageous embodiment of the invention, several machine components can be monitored, wherein for each of the machine components The simulation module can have a component-specific simulation model for simulating the respective machine component, the prediction module can have a component-specific prediction model for predicting a future health state of the respective machine component, component-specific performance values ​​can be derived, a component-specific performance normal value can be determined by a component-specific calculation method, a component-specific performance reference value can be used, component-specific change patterns can be specified, a component-specific comparison can be carried out, and / or a component-specific health state can be displayed.

[0025] In particular, component-specific processing pipelines can be implemented, each with a component-specific simulation model, a component-specific performance value determination, a component-specific reference value determination, a component-specific performance value comparison, and / or a component-specific forecasting model. Such processing pipelines can be executed in parallel and can be easily extended to include new processing pipelines for additional machine components.

[0026] According to a further advantageous embodiment of the invention, an overall health status of the machine can be derived and displayed by linking the component-specific health states. This linking can preferably be performed using logical operators.

[0027] An embodiment of the invention is explained in more detail below with reference to the drawing. The drawings illustrate each embodiment schematically. Figure 1 shows a monitoring of a machine by a monitoring device according to the invention, Figure 2 shows a predicted course of a performance value and Figure 3 shows several component-specific processing pipelines of a monitoring device according to the invention.

[0028] Insofar as the figures use the same or corresponding reference symbols, these reference symbols denote the same or corresponding entities, which may be described, implemented or designed in particular as in connection with the figure in question.

[0029] Figure 1Figure 1 illustrates the monitoring of a machine M by a monitoring device MON according to the invention coupled to the machine M. The machine M can be, in particular, a robot, a machine tool, a production plant, a turbine, a 3D printer, an internal combustion engine, and / or a motor vehicle, or comprise such a machine. The machine M has several machine components C1, C2, ... which are to be monitored component-specifically. For the sake of clarity, only two machine components, C1 and C2, are explicitly shown in the drawing. If the machine M is an electric motor, a respective machine component C1 or C2 can, for example, be a rotor, an axle bearing, a cooling system, a stator, a winding, or insulation.

[0030] The MON monitoring facility is in Figure 1The monitoring device MON is shown externally to machine M. Alternatively, it can also be fully or partially integrated into machine M.

[0031] The monitoring device MON comprises one or more processors PROC for executing process steps of the invention and one or more memory MEMs for storing data to be processed. Such monitoring devices MON are also frequently referred to as "condition monitoring" systems.

[0032] Machine M is equipped with sensors S for continuously measuring and / or recording operating data BD of machine M and environmental data UD from the machine M's environment. Operating data BD and / or environmental data UD can also be recorded by other means besides the sensors S.

[0033] Operating data (BD) can include, in particular, current operating signals, sensor data, and / or measured values ​​that quantify the power, rotational speed, torque, speed of movement, applied force, emissions, and / or temperature of one or more machine components over time. Similarly, environmental data (UD) can specify, for example, the ambient temperature or humidity in the machine's environment over time. Furthermore, environmental data (UD) can specify the machine's location and / or installation type, time of day, calendar date, and / or season. Both the operating data (BD) and the environmental data (UD) are continuously recorded and transmitted to or received by the monitoring device (MON).

[0034] According to the invention, the monitoring device MON has a parallel simulation module SIM for real-time simulation of the current operating behavior of the machine M and / or its components, here C1 and C2, in parallel with the ongoing operation of the machine M. In particular, the simulation module SIM continuously simulates loads on the machine components C1 and C2. For the purpose of the simulation, the operating data BD and the environmental data UD are continuously fed into the simulation module SIM.

[0035] In the present embodiment, the simulation module SIM has several component-specific simulation models (not shown) for the component-specific simulation of machine component C1 or C2. Each component-specific simulation model can preferably comprise several domain-specific sub-models, e.g., a mechanical sub-model, an electrical sub-model, and / or a thermal sub-model. A large number of efficient domain-specific simulation models are available for such mechanical, electrical, or thermal simulations. Typical implementations of such simulation modules can easily include 30 or 40 sub-models.

[0036] The simulation module SIM and its simulation models are initialized by structural data SD of machine M from a database DB linked to the monitoring device MON. This structural data SD specifies, in particular, the geometry, physical properties, and other operating parameters of machine elements of machine M, and especially of machine components C1 and C2. Based on the structural data SD, simulations can be performed using established simulation methods with the operating data BD and environmental data UD. For this purpose, both detailed physical simulation models, such as finite element models, and efficient data-driven surrogate models are available.A neural network or other machine learning model can be used as a surrogate model, provided it has been previously trained using a precise physical simulation model to reproduce its simulation results as accurately as possible. Evaluating such surrogate models typically requires significantly fewer computing resources than a detailed physical simulation model.

[0037] From the simulated current operating behavior of machine components C1 and C2, a performance value PV is continuously derived, quantifying the current performance of each component. The performance of a given machine component C1 or C2 can relate to, in particular, rotational speed, power, resource consumption, yield, efficiency, precision, emissions, stability, vibrations, wear, load, and / or other target parameters of the machine M and / or its components, here C1 and C2. Specifically, performance values ​​PV are used to determine indicators of the load on machine components C1 and C2, such as stresses, vibrations, high forces, high temperatures, or high pressures. The simulation thus effectively creates virtual sensors for the performance of machine components C1 and C2.

[0038] The component-specific performance values ​​PV are stored as time series.

[0039] The performance values ​​PV are transmitted from the simulation module SIM to a reference value module REF of the monitoring device MON. The reference value module REF serves to dynamically determine and adjust component-specific reference values ​​PR for the performance values ​​PV. To determine a respective component-specific reference value PR, a component-specific performance normal value PN is regularly derived from the transmitted performance values ​​PV. The performance normal value PN is intended to serve as a benchmark for the respective performance in a normal or target state of the machine M. For this purpose, a moving average, e.g., over several hours, days, or weeks, or another average value of the performance values ​​PV is preferably calculated during operating phases with low temporal fluctuations and stored as the performance normal value PN.

[0040] Based on the regularly derived performance normal values ​​PN, the performance reference value PR for a respective machine component C1 or C2 is dynamically updated.

[0041] At the start of operation of machine M or at other initialization times, the performance reference value PR is initialized by an initial component-specific performance value PRI, retrieved, for example, from the database DB. Since operating conditions can change depending on the installation location, installation type, environmental influences, maintenance work, or reconfigurations, the performance reference value is dynamically updated according to the invention when predefined change patterns occur and / or when a replacement or modification of machine elements is detected.

[0042] The REF reference module has a detector DT for detecting change patterns and for detecting the replacement or modification of machine elements. The detector DT specifically monitors the performance values ​​PV and, if applicable, the operating data BD to determine whether a predefined first change pattern occurs. Furthermore, the detector DT monitors the environmental data UD to determine whether a predefined second change pattern occurs. For this purpose, the environmental data UD and, if applicable, the operating data BD are transmitted to the REF reference module and fed into the detector DT. The first and second change patterns are in Figure 1 The reference symbol CP represents both change patterns. Each CP change pattern is component-specific, meaning it is specifically defined for each of the components C1 and C2.

[0043] The first change pattern specifies, in particular, those changes to the operational data BD and / or performance values ​​PV that should trigger an update of the performance reference value PR. For this purpose, the first change pattern can, for example, specify a threshold for a fluctuation amplitude, a temporal gradient, and / or a temporal variance of the operational data BD and / or the performance values ​​PV. In this way, an update of the performance reference value PR can be triggered when a fluctuation amplitude, gradient, and / or variance of the performance values ​​PV and / or the operational data BD remains relatively low over a predefined period, thus indicating a normal operating phase suitable as a reference. Accordingly, the first change pattern can also specify a period of time with a low fluctuation amplitude.Alternatively or additionally, the first change pattern can specify a distance between a performance value PV and the performance reference value PR, which triggers an update if exceeded or fallen below.

[0044] The second change pattern preferably defines, on a component-specific basis, which changes to the environmental data (UD) should trigger an update of the performance reference value (PR). For this purpose, the second change pattern, analogous to the first change pattern, can specify a threshold for a fluctuation amplitude, a temporal gradient, and / or a temporal variance of the environmental data (UD) in order to define a normal operating phase suitable for reference.

[0045] As soon as the detector DT detects a change pattern CP or a change or replacement of a machine element based on the operating data BD, the environmental data UD, and the performance values ​​PV, the detector DT generates a component-specific trigger signal TS. This component-specific trigger signal TS triggers an update of the performance reference value PR for the respective machine component C1 or C2 by the performance normal value PN of the respective machine component C1 or C2. The trigger signal TS can be generated, in particular, when all normal state indicators coincide. Furthermore, an update can also be initiated at predefined times, e.g., at night or on weekends.

[0046] The component-specific performance reference value PR serves as a current reference for assessing the current health status of a given machine component C1 or C2. For this purpose, the current performance values ​​PV are continuously compared with the current performance reference value PR for each component. This results in a current component-specific difference D between a current performance value PV and the performance reference value PR. This difference D can be, for example, the absolute value or the square of the difference PV - PR, or a relative difference D - |PV - PR| / PR.

[0047] The determined distance D is fed into an evaluation module EV of the monitoring device MON as a comparison result. Based on the component-specific distances D, and preferably also based on their temporal profiles, the evaluation module EV assesses the current health status of a respective machine component C1 or C2. For this purpose, distances D can preferably be compared at different, especially successive, points in time. This allows a temporal gradient of successive distances to be generated, for example, to detect increases in the distances D and assess them as a deterioration of the health status. Furthermore, a temporal variance of the distances D can also be determined to distinguish normal fluctuations in the distance D from exceptional fluctuations indicating a deterioration of the health status.Furthermore, the distances D can be compared by the assessment unit EV with predefined or calculated thresholds that define critical health conditions. In addition, the assessment module EV can also directly compare the performance values ​​PV with a predefined performance threshold, the exceeding or falling below of which is considered damage; for example, if the performance of a machine component falls below a mandatory minimum performance level.

[0048] Depending on the distances D, the evaluation module EV determines a current component-specific health value HS for each component C1 and C2, specifying the health status of the respective component C1 or C2. In the present embodiment, the evaluation unit EV checks, on a component-specific basis, whether a given distance D and / or its temporal gradient, possibly taking into account a temporal variance, exceeds a first component-specific threshold and a second component-specific threshold. The first threshold defines, in a component-specific manner, a maximum deviation from a target behavior that is considered non-critical. In contrast, the second threshold defines, on a component-specific basis, a deviation from the target behavior, the exceeding of which requires immediate or timely user action.The current health value HS indicates, in particular, whether the first threshold and whether the second threshold is exceeded. Furthermore, the current health value HS can also specify the current deviation D from the target behavior, a specific impairment, a failure, and / or a probability of the occurrence of impairment, a malfunction, or a failure.

[0049] From the current health data HS for the individual components C1 and C2, the evaluation module EV preferably derives a current overall health data HSM, which specifies the current overall health status of the machine M. For this purpose, individual current health data HS can be combined, for example, using logical operators to form the current overall health data HSM. Thus, a critical overall status of the machine M can be indicated as soon as a critical health status is determined for one of the machine components C1 or C2.

[0050] The current health data HS for components C1 and C2, as well as the current overall health data HSM, are transmitted by the evaluation module EV to an output unit OUT of the monitoring device MON. The output unit OUT can include a display, an alarm device, an online notification system, or another form of signaling or message output. The output unit OUT can output a monitoring signal, an alarm signal, an operating recommendation, an error signal, a diagnostic signal, and / or a maintenance signal.

[0051] In the present embodiment, the output unit OUT includes a traffic light indicator for displaying the component-specific health information HS and the overall health information HSM. The traffic light indicators each comprise a green indicator G, a yellow indicator Y, and a red indicator R. The green indicator G is activated when the relevant distance D is below the first threshold. In this case, the health status of the component in question is not critical. Similarly, the red indicator R is activated when the relevant distance D is above the second threshold, indicating a critical health status of the component and requiring immediate or timely user action. The yellow indicator Y is activated when the relevant distance D is between the first and second thresholds.In this way, monitoring of the relevant machine component can be recommended.

[0052] The current overall health information (HSM) can be displayed analogously using a traffic light system.

[0053] The monitoring device MON also features a concurrent predictive module PM with several component-specific predictive models. Performance values ​​PV and the performance reference value PR are fed into the PM predictive module. Preferably, the current health data HS and, if applicable, the current overall health data HSM are also transmitted from the evaluation module EV to the PM predictive module. The PM predictive module serves the purpose of continuously predicting the future health status of a respective machine component C1 or C2 based on the operating behavior simulated by the simulation module SIM during machine operation. The predictive models can include, in particular, a dynamic load model, a dynamic wear model, a dynamic degradation model, or a dynamic lifetime model.A variety of methods are available for the application and numerical evaluation of such forecasting models.

[0054] The PM prediction module continuously determines the load on machine components C1 and C2 based on the performance values ​​PV. Using its predictive models, the PM module then forecasts the future performance development of machine components C1 and C2 or the development of fatigue phenomena over time. This allows, in particular, the prediction of the remaining service life of machine components C1 and C2 and / or the time until damage occurs. If necessary, no development of the performance reference value PR can be predicted over time.

[0055] To determine a future health status, predicted performance values ​​(PV) can be compared with a current or predicted performance reference value (PR), and a difference between the predicted performance values ​​(PV) and the current or predicted performance reference value (PR) can be determined. From this difference, a future health status for a respective machine component (C1) and (C2) can then be derived as described above.

[0056] The PM forecasting module generates a value HP for each machine component C1 or C2 regarding its future health status. Furthermore, the PM forecasting module generates a value HPM regarding the future overall health status of the machine M and transmits this, along with the component-specific HP values, to the output unit OUT. There, the forecasted values ​​HP and HPM can be displayed using traffic light indicators, similar to the current values ​​HS and HSM. The HPM value can be determined by linking the component-specific HP values, as described above in connection with the current health values ​​HS and HSM.

[0057] The HP and HPM values ​​can preferably be determined for one or more future time intervals.

[0058] In addition to forecasts of future health conditions, the PM forecasting module preferably also determines a forecast uncertainty, a confidence interval, and / or a probability of damage. In determining these quantities, uncertainties in physical boundary conditions or measurement inaccuracies, as well as intrinsic inaccuracies of the simulation models, can be taken into account.

[0059] Figure 2 This illustrates a predicted performance value PV as generated by the PM forecasting module. The predicted trend is plotted against time T. Time T=0 represents the current time, while future times are plotted to the right of it.

[0060] While the predicted course of the performance value PV is illustrated by a solid line, uncertainty in the time-based forecast is indicated by dashed lines. Furthermore, a performance threshold TH is shown, the undercutting of which by the performance value PV is considered a loss. The predicted course of the performance value PV falls below the performance threshold TH at time TS, which can thus be determined as the probable time of loss occurrence. Based on the course of the forecast uncertainty, a confidence interval CI can be determined, within which the loss is most likely or with a given probability to occur. The confidence interval CI can be calculated from the points where the dashed lines intersect the performance threshold TH.

[0061] A component-specific confidence interval (CI) can preferably be displayed together with the HP information about a future health status of a respective component.

[0062] Figure 3 Figure 1 shows several component-specific processing pipelines P1 and P2 of a monitoring device according to the invention. Processing pipeline P1 serves to monitor the health status of machine component C1, while processing pipeline P2 serves to monitor the health status of machine component C2. Component-specific operating data BD and environmental data UD are optionally fed into both processing pipelines P1 and P2.

[0063] The processing pipeline P1 includes a simulation module SIM1 for the concurrent simulation of the current operating behavior of machine component C1. Furthermore, a reference value determination REF1 specific to machine component C1 is provided. In addition, the processing pipeline P1 performs a comparison CMP1 specific to machine component C1 of the performance values ​​specific to this component with the performance reference value specific to this component C1. The reference value determination REF1, and in particular the updating of the relevant performance reference value, is carried out component-specifically as described above. Depending on the component-specific comparison result, a current health state HS1 specific to machine component C1 is determined. Furthermore, a future health state HP1 specific to machine component C1 is predicted by a forecasting model P1 specific to this component.

[0064] The processing pipeline P2, which is specific to the machine component C2, works in a similar way.

[0065] The processing pipelines P1 and P2 are executed in parallel and in real time during the operation of machine M. Due to the separation of the processing pipelines, here P1 and P2, for the individual components, here C1 and C2, the monitoring device MON can be extended to monitor further machine components in a particularly simple manner.

Claims

1. Computer-implemented method for monitoring the state of a machine (M), wherein a) current operating data (BD) of the machine (M) are continuously captured, b) a current operating behaviour of a machine component (C1, C2) is continuously simulated based on the operating data (BD) by a concurrent simulation module (SIM), c) performance values (PV) quantifying a current performance of the machine component (C1, C2) are continuously derived from the simulated operating behaviour and stored over time, characterized in that d) a performance normal value (PN) is determined regularly based on a multiplicity of previously derived performance values, e) the operating data (BD) and / or the performance values (PV) are monitored for whether a predefined first change pattern (CP) occurs, f) an update of a performance reference value (PR) with the current performance normal value (PN) is prompted in succession of the detection of the first change pattern (CP), g) the respective current performance values (PV) are continuously compared with the respective current performance reference value (PR), and h) a current health state (HS) of the machine component (C1, C2) is displayed on the basis of the comparison result.

2. Method according to Claim 1, characterized in that by the first change pattern (CP) - a threshold value for a fluctuation amplitude, for a temporal gradient and / or for a temporal variance of the performance values (PV) and / or the operating data (BD) is specified, - a distance of a performance value (PV) from the current performance reference value (PR) is specified and / or - a period of time is specified, in which a fluctuation amplitude of the performance values (PV) and / or the operating data (BD) is below a threshold value.

3. Method according to either of the preceding claims, characterized in that current environmental data (UD) of the machine (M) are continuously captured, and in that the comparison between the current performance values (PV) and the current performance reference value (PR) is made on the basis of the current environmental data (UD).

4. Method according to either of the preceding claims, characterized in that current environmental data (UD) of the machine (M) are continuously captured, in that the environmental data (UD) are monitored for whether a predefined second change pattern occurs, and in that upon detection of the second change pattern an update of the performance reference value (PR) with the current performance normal value (PN) is prompted.

5. Method according to either of the preceding claims, characterized in that the machine (M) is monitored for whether a machine component (C1, C2) and / or a part of the machine (M) is replaced or changed, and in that upon detection of such replacement and / or such a change an update of the performance reference value (PR) with the current performance normal value (PN) is prompted.

6. Method according to either of the preceding claims, characterized in that, in order to determine the performance normal value (PN), - an operating phase in which a fluctuation amplitude of the performance values (PV) is below average and / or is below a predefined threshold value is detected, and / or - a moving average of the performance values (PV) is determined.

7. Method according to either of the preceding claims, characterized in that by a concurrent prediction module (PM) a future health state (HP) of the machine component (C1, C2) is continuously predicted based on the simulated operating behaviour.

8. Method according to Claim 7, characterized in that by the prediction module - based on the simulated operating behaviour a load on the machine component (C1, C2) is continuously determined, and - the future health state (HS) is predicted from the load by means of a dynamic load model, wear model or service life model.

9. Method according to Claim 8, characterized in that a prediction uncertainty, a confidence interval (CI) and / or a damage probability is determined by means of the dynamic load model, wear model or service life model.

10. Method according to one of the preceding claims, characterized in that within the comparison between a performance value (PV) and the performance reference value (PR) a distance (D) between the performance value (PV) and the performance reference value (PR), a temporal gradient of the distance (D) and / or a temporal variance of the distance (D) is determined, and in that the current health state (HS) is displayed on the basis of the distance (D), the gradient and / or the variance.

11. Method according to one of the preceding claims, characterized in that multiple machine components (C1, C2) are monitored, wherein, for each of the machine components (C1, C2), - the simulation module (SIM) comprises a component-specific simulation model (SIM1, SIM2) for simulating the respective machine component (C1, C2), - component-specific performance values are derived, - a component-specific performance normal value is determined by a component-specific computation method, - a component-specific performance reference value is used, - component-specific change patterns are predefined, - a component-specific comparison is performed, and / or - a component-specific health state (HS1, HS2) is displayed.

12. Method according to Claim 11, characterized in that by linking the component-specific health states (HS1, HS2) an overall health state (HSM) of the machine (M) is derived and displayed.

13. Monitoring device (MON) for monitoring the state of a machine (M), configured to carry out a method according to one of the preceding claims.

14. Computer program product comprising commands that, when the program is executed by a computer, cause said computer to carry out the steps of the method according to one of Claims 1 to 12.

15. Computer-readable storage medium comprising a computer program product according to Claim 14.