MONITORING DEVICE AND MONITORING METHOD FOR DETECTING A DIAGNOSTIC STATE

Abstract

A monitoring device for detecting a diagnostic state of an industrial machine on the basis of vibrations associated with the industrial machine, includes at least one measurement transducer for detecting at least one first signal characterizing the vibration of the industrial machine in a time period and for detecting at least one second signal characterizing the vibration of the industrial machine in the same time period, wherein the at least one measurement transducer has at least one signal processor for evaluating the at least first signal as a first characteristic value and for evaluating the at least second signal as a second characteristic value and wherein the at least one first characteristic value and the at least one second characteristic value form a characteristic value tuple, further having a storage unit, in which different diagnostic states describing the state of the industrial machine are saved, wherein the diagnostic states are determined as a function of the characteristic value tuple, and a predefined scalar range is saved with predetermined different fixed scalar values, which define respectively an existing limit potential, wherein a respective limit potential characterizes a change in the diagnostic state, and wherein between adjacent limit potentials a diagnostic partition is formed and wherein different scalar values from the scalar range are assigned to the different diagnostic partitions and wherein different respective diagnostic partitions represent different diagnostic states of the industrial machine, and wherein the monitoring device has at least one interpolation function for mapping the characteristic value tuple on a common, unique, time-dependent scalar value in the predefined scalar range, so that by way of the time-dependent scalar value a time monitoring of the diagnostic state of the industrial machine is achieved.

Claims

1. A monitoring device for detecting a diagnostic state of an industrial machine on the basis of vibrations associated with the industrial machine, comprising at least one measurement transducer for detecting at least one first signal characterizing the vibration of the industrial machine in a time period and for detecting at least one second signal characterizing the vibration of the industrial machine in the same time period, and wherein the at least one measurement transducer has at least one signal processor for evaluating the at least first signal as a first characteristic value and for evaluating the at least second signal as a second characteristic value and wherein the at least one first characteristic value and the at least one second characteristic value form a characteristic value tuple, further having a storage unit, in which different diagnostic states describing the state of the industrial machine are saved, wherein the diagnostic states are determined as a function of the characteristic value tuple, and a predefined scalar range is saved with predetermined different fixed scalar values, which define respectively an existing limit potential, wherein a respective limit potential characterizes a change in the diagnostic state, and wherein between adjacent limit potentials a diagnostic partition is formed and wherein different scalar values from the scalar range are assigned to the different diagnostic partitions and wherein different respective diagnostic partitions represent different diagnostic states of the industrial machine, and wherein the monitoring device has at least one interpolation function for mapping the characteristic value tuple onto a common unique, time-dependent scalar value in the predetermined scalar range, so that by way of the time-dependent scalar value a time monitoring of the diagnostic state of the industrial machine is achieved.

2. The monitoring device according to claim 1, wherein the respectively adjacent diagnostic partitions with respect to the criticality consecutive diagnostic states of the industrial machine and wherein the consecutive diagnostic states are formed at least in ascending order as a false measurement, a good state, a warning state and an alarm state.

3. The monitoring device according to claim 2, wherein the values of the time-dependent scalar values increase with respect to the criticality of the sequence of states.

4. The monitoring device according to claim 1, wherein the monitoring device is further configured, in the presence of at least one characteristic value tuple, which corresponds to a change between the diagnostic states, to map the characteristic value tuple to the corresponding limit potential unique to the change of the diagnostic state.

5. The monitoring device according to claim 4, wherein a plurality of potentials with intermediate sub-partitions are assigned to a respective diagnostic partition, wherein a potential is uniquely assigned to a predetermined scalar value and wherein the monitoring device is configured to map, in the presence of at least one characteristic value tuple, which indicates a change between the sub-partitions, the characteristic value tuple to the corresponding scalar value uniquely associated with the change.

6. The monitoring device according to claim 1, wherein the monitoring device has a plurality of interpolation functions and is further configured to select the interpolation function as a function of the characteristic value tuple.

7. The monitoring device according to claim 6, wherein the interpolation functions are assigned to the different diagnostic partitions, wherein the monitoring device is further configured to map, in the presence of at least one characteristic value tuple, which lies within one of the diagnostic partitions, the characteristic value tuple to the unique scalar value by the interpolation function assigned to the diagnostic partition.

8. The monitoring device according to claim 7, wherein a respective diagnostic partition is assigned a plurality of potentials with sub-partitions in between, wherein the interpolation functions are assigned to the different sub-partitions, wherein the monitoring device is further configured, in the presence of at least one characteristic value tuple, which lies within one of the sub-partitions, to map the characteristic value tuple by the interpolation function assigned to the sub-partition to the unique scalar value.

9. The monitoring device according to claim 1, wherein an aggregating unit is provided for aggregating the first characteristic values and second characteristic values determined during the predefined time period as characteristic value tuple and coupling the characteristic value tuple to the time period.

10. The monitoring device according to claim 1, wherein the at least one measurement transducer is configured to generate periodically at least one first signal and a second signal to consecutive time periods.

11. The monitoring device according to claim 1, wherein a trigger unit is provided, which triggers the at least one measurement transducer, in order to generate at least the first signal and at least the second signal in consecutive time periods.

12. The monitoring device according to claim 1, wherein the at least one measurement transducer is a piezoelectric vibration transducer.

13. The monitoring device according to claim 1, wherein a plurality of measurement transducers are provided respectively with a signal processor for detecting at least a first signal and second signal characterizing the vibrations of the industrial machine.

14. The monitoring device according to claim 1, wherein the at least one measurement transducer is configured for detecting a first signal and second signal characterizing the vibrations of the industrial machine and for detecting further signals characterizing the vibrations of the industrial machine and for evaluating in a plurality of characteristic values.

15. A monitoring method for detecting a diagnostic state of an industrial machine on the basis of vibrations associated with the industrial machine, wherein at least one measurement transducer is provided for detecting at least one first signal characterizing the vibration of the industrial machine in a time period and for detecting at least one second signal characterizing the vibration of the industrial machine in the same time period, and wherein the at least one measurement transducer has at least one signal processor for evaluating the at least first signal as a first characteristic value and for evaluating the at least second signal as a second characteristic value, and wherein the at least one first characteristic value and the at least one second characteristic value form a characteristic value tuple, wherein furthermore a storage unit is provided, in which different diagnostic states describing the state of the industrial machine are saved, wherein the diagnostic states are determined as a function of the characteristic value tuple, and a predefined scalar range is saved with predefined different fixed scalar values, which define respectively an existing limit potential, wherein a respective limit potential characterizes a change in the diagnostic state, and wherein between adjacent limit potentials a diagnostic partition is formed, and wherein different scalar values from the scalar range are assigned to the different diagnostic partitions, and wherein different respective diagnostic partitions represent different diagnostic states of the industrial machine, and wherein at least one interpolation function is provided for mapping the characteristic value tuple on a common, unique, time-dependent scalar value in the predefined scalar range, so that by way of the time-dependent scalar value a time monitoring of the diagnostic state of the industrial machine is achieved.

16. The monitoring method according to claim 15, wherein the respective adjacent diagnostic partitions represent, with respect to the criticality, consecutive diagnostic states of the industrial machine and wherein the consecutive diagnostic states are formed at least in ascending order as a false measurement, a good state, a warning state and an alarm state.

17. The monitoring method according to claim 15, wherein in the presence of at least one characteristic value tuple, which corresponds to a change between the diagnostic states, the characteristic value tuple is mapped onto the corresponding scalar value unique to the limit potential associated with the change of diagnostic state.

18. The monitoring method according to claim 17, wherein a diagnostic partition is assigned a plurality of potentials with sub-partitions lying in between, wherein a potential is assigned uniquely to a predefined scalar value and in the presence of at least one characteristic value tuple, which specifies a change between the sub-partitions, the characteristic value tuple is mapped to the corresponding scalar value unique to the change of potential.

19. The monitoring method according to claim 15, wherein a plurality of interpolation functions are provided and the interpolation functions are selected as a function of the characteristic value tuple, and wherein the interpolation functions are assigned to the different diagnostic partitions, and wherein in the presence of at least one characteristic value tuple, which lies within one of the diagnostic partitions, the characteristic value tuple is mapped to the unique scalar value by the interpolation function assigned to the diagnostic partition.

20. The monitoring method according to claim 19, wherein a plurality of potentials are assigned to a respective diagnostic partition with sub-partitions in between, wherein the interpolation functions are assigned to the different sub-partitions, wherein in the presence of at least one characteristic value tuple, which lies within one of the sub-partitions, the characteristic value tuple is mapped to the unique scalar value by the interpolation function assigned to the sub-partition.

Description

BRIEF DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWINGS

[0039] Further features, properties and advantages of the present disclosure are given in the following description with reference to the accompanying Figures. In schematic form:

[0040] FIG. 1 shows a monitoring device in a first embodiment,

[0041] FIG. 2 shows an example of a measurement transducer in a first embodiment,

[0042] FIG. 3 shows schematically the diagnostic partition as a function of a characteristic value tuple,

[0043] FIG. 4 shows an aggregating unit in a first embodiment,

[0044] FIG. 5 shows schematically a mapping of two characteristic values,

[0045] FIG. 6 shows schematically an interpolation function,

[0046] FIG. 7 shows a further configuration of a monitoring device,

[0047] FIG. 8 shows two characteristic values as time-dependent scalar values in a diagram,

[0048] FIG. 9 shows possible applications of a determined scale, and

[0049] FIG. 10 shows schematically the monitoring method.

DETAILED DESCRIPTION

[0050] FIG. 1 shows a monitoring device 1 for detecting a diagnostic state of an industrial machine 2 on the basis of vibrations associated with the industrial machine 2 by at least one measurement transducer 3 in a first embodiment.

[0051] The monitoring device 1 can be arranged in one embodiment as a fixed stationary device on the machine 2. In another embodiment the monitoring device 1 can be installed semi-permanently. In a further embodiment the monitoring device 1 is configured as a portable device for setting up and detecting vibrations.

[0052] In one embodiment, industrial machines 2 can be understood as any vibration generating industrial machines, for example machines with rotating elements. Examples of such machines include pumps, generators or wind turbines, industrial plants or roller bearings.

[0053] In general, the monitoring device 1 can be applied to all types of industrial machines 2, which allow the detection of a diagnostic state based on the vibrations generated by the industrial machine 2.

[0054] Without being limited to this, measurement transducers 3 can be contactless measurement transducers or seismic measurement transducers with suitable signal processing functions or piezoelectric vibration transducers, which are capable for example of measuring accelerations typically in the order of 0.1 m/s.sup.2 to 1000 m/s.sup.2 in a frequency range of 0.1 Hz to 25 kHz.

[0055] The measurement transducers 3 can be accelerometers, speed transducers, contactless probes or a combination thereof.

[0056] Other configurations are also possible.

[0057] FIG. 2 shows a measurement transducer 3 for detecting at least a first signal S1 and second signal S2 characterizing the vibration in a predetermined time period, by which a time stamp is formed, of the industrial machine 2. In a further embodiment further signals can also be detected by the measurement transducer 3. The measurement transducer 3 has a signal processor 4 for evaluating the at least first signal S1 as a first characteristic value K1 and the at least second signal S2 as a second characteristic value K2. Typical characteristic values can be mean values and the effective value of the acceleration as a root mean square, maximum, zero-peak-characteristic values (0-to-peak, amplitudes) for the given time period. The effective value of acceleration indicates which power has a vibration signal and is obtained by bandpass filtering (e.g. in the range of 10 Hz-1 kHz).

[0058] A zero-peak-characteristic value can be understood as the maximum of the magnitude of a detected signal over a predefined measurement period (time period). This characteristic value can be used as a measure for the peak content of a signal.

[0059] Furthermore, also more than two characteristic values can be provided. The combination of two or more characteristic values allows a more reliable assessment compared to the use of a single characteristic value, in particular for machine states which can only be monitored by looking at two or more characteristic values at the same time.

[0060] The detected characteristic values K1, K2 can be represented as time-dependent characteristic value tuples.

[0061] The monitoring device 1 further comprises a storage unit 11.

[0062] In one embodiment, the storage unit 11 can also be integrated into the measurement transducer 3.

[0063] In the storage unit 11 different, consecutive diagnostic states describing the state of the industrial machine 2 are saved.

[0064] Diagnostic states can be, in one embodiment for example, a false measurement, a good state, a warning state, and an alarm state.

[0065] Furthermore, the storage unit 11 has predefined different scalar values, which define respectively at least available limit potentials, in this embodiment G1, G2, G3 (FIG. 3), wherein a limit potential G1, G2, G3 characterizes a change of the diagnostic state, wherein between adjacent limit potentials G1, . . . , G3 a diagnostic partition is formed, in this embodiment as diagnostic partitions D1, . . . , D4 (FIG. 3).

[0066] It should be noted that more than three limit potentials (n-limit potentials, with n natural number) and more than four diagnostic partitions (m-diagnostic partitions with m natural number) can be provided.

[0067] In this embodiment the diagnostic partitions D1,D2,D3,D4 (FIG. 7) are assigned to the diagnostic states: false measurement as D1, good state as D2, warning state as D3 and alarm state as D4. Furthermore, the limit potentials, here G1, G2,G3, are each assigned a scalar value, here for example the scalar values 1, 5 and 8. It should be noted here that these are simplified examples of the present disclosure and its embodiments.

[0068] FIG. 3 shows schematically the diagnostic partitions D1,D2,D3,D4 as a function of the first characteristic value K1 and a second characteristic value K2 and the limit potentials G1,G2,G3, which can also be considered as contour lines for example.

[0069] Every combination of characteristic values K1,K2, i.e. to each characteristic value tuple, is thus assigned uniquely to one of the four diagnostic partitions D1,D2,D3,D4 and thus four discrete diagnostic states or one of the limit potentials G1,G2,G3.

[0070] Here a characteristic value tuple can comprise at least two characteristic values K1,K2 or also a plurality of characteristic value tuples, which can be formed by one or more measurement transducers 3a, . . . , 3n.

[0071] FIG. 4 shows schematically a plurality of measurement transducers 3a, . . . , 3n. The latter can be in the form of vibration transducers or other sensors. Each of these measurement transducers 3a, . . . , 3n detects one of the signals S1, . . . , Sn characterizing the vibration of the industrial machine 2, which are converted into characteristic values K1, . . . , Kn by way of a corresponding signal processor 4a, . . . 4n. In another embodiment, the individual measurement transducers can also detect a plurality of signals.

[0072] In this embodiment, a trigger unit 5 can be provided, which triggers the measurement transducers 3a, . . . 3n to detect the signals S1, . . . , Sn in consecutive time periods.

[0073] The trigger unit 5 thus triggers for example the generation of signals S1, . . . , Sn in the existing measurement transducer 3a, . . . , 3n. The trigger signal thus activates the existing measurement transducers 3a, . . . , 3n.

[0074] In other embodiments, the measurement transducer(s) 3a, . . . 3n can be designed to detect signals periodically or continuously.

[0075] Furthermore, the embodiment in FIG. 4 has an aggregating unit 6, for aggregating the characteristic values K1, . . . , Kn determined during the predefined time periods as a characteristic value tuple. The aggregating unit 6 waits until each measurement transducer 3a, . . . 3n or each signal processor 4a, . . . , 4n has delivered a characteristic value K1, . . . , Kn, and then passes these on aggregated as a characteristic value tuple. Here, depending on the characteristic value K1, . . . , Kn the time period between triggering and the finished characteristic value K1, . . . , Kn can be very different, for example due to different settling periods and the different lengths of a required recording interval.

[0076] In a further embodiment, the characteristic value tuple is uniquely mapped to a scalar value by one or more interpolation functions.

[0077] For mapping the characteristic value tuple to a common unique, time-dependent scalar value in the predefined scalar range by at least one interpolation function, the monitoring device 1 has, as usual, a computing module/processor or similar hardware.

[0078] FIG. 5 shows schematically a mapping of the at least two characteristic values K1 and K2 to a unique, time-dependent scalar value by interpolation functions.

[0079] In one embodiment, different interpolation functions F1,F2,F3,F4 are provided for the different diagnostic partitions D1,D2,D3,D4, as shown for example in FIG. 3.

[0080] Firstly, it is checked in which diagnostic partition D1,D2,D3,D4 the characteristic value tuple K1,K2 is located or whether it is located on one of the limit potentials G1,G2,G3. For each diagnostic partition D1,D2,D3, D4 an interpolation function F1,F2,F3,F4, which maps the characteristic values K1,K2 to a scalar value is used. In one embodiment, these interpolation functions F1,F2,F3,F4 are continuous and smooth. Furthermore, in one embodiment, within the diagnostic partition D1,D2,D3,D4 the interpolation functions F1,F2,F3,F4 either have no extremes or map all characteristic values K1,K2 on the same scalar value. In other embodiments, for example for a plurality of diagnostic partitions D1,D2,D3,D4, the same interpolation function can be used.

[0081] In this example, the characteristic value tuple K1,K2 if present in the diagnostic partition D1 is mapped to a scalar value between 0 and 1 by way of the interpolation function F1. Here the range 0 and 1 is selected at random and is used only for illustration.

[0082] If the characteristic value tuple lies on the limit potential G1, the characteristic value tuple K1,K2 is mapped directly to the scalar 1. Here the scalar value 1 is chosen at random.

[0083] If the characteristic value tuple K1,K2 lies in the diagnostic partition D2, the characteristic value tuple K1,K2 is mapped to a scalar value between 1 and 5 by way of the interpolation function F2. Here the range 1 and 5 is selected at random and is used only for illustration.

[0084] If the characteristic value tuple K1,K2 lies on the limit potential G2, the characteristic value tuple K1,K2 is mapped directly to the scalar value 5. The value 5 is chosen here at random.

[0085] If the characteristic value tuple K1,K2 lies in the diagnostic partition D3, the characteristic value tuple K1,K2 is mapped to a scalar value between 5 and 8 by way of the interpolation function F3. Here the range 5 and 8 is selected at random and is used only for illustration.

[0086] If the characteristic value tuple K1,K2 lies on the limit potential G3, the characteristic value tuple K1,K2 is mapped directly to the scalar value 8. The value 8 is chosen here at random.

[0087] If the characteristic value tuple K1,K2 lies in the diagnostic partition D4, the characteristic value tuple K1,K2 is mapped to a scalar value between 8 and 15 by way of the interpolation function F4. Here the range 8 and 15 is selected at random and is used only for illustration.

[0088] Thus the individual scalar values can give an approximate measure of the distance to the corresponding limit potentials G1,G2,G3.

[0089] Furthermore, also more than two characteristic values can be included.

[0090] The respective scalar values are linked to the time period as time stamps.

[0091] FIG. 6 shows for the characteristic values K1=effective value of acceleration as a root mean square, and K2=zero-peak-characteristic value (0-to-peak) such an interpolation function F. Here the effective value of the acceleration indicates which power has a vibration signal and is obtained by bandpass-filtering (e.g. to 10 Hz-1 kHz), wherein a detailed explanation is not given. The two characteristic values effective value of acceleration, zero-peak-characteristic value can thus be represented simply as a time series and shown as a function of the limit potentials G1,G2,G3 for example.

[0092] There are only scalar values, which are assigned to one of the diagnostic partitions D1, . . . , D4 or a limit potential G1,G2,G3.

[0093] FIG. 7 shows a further configuration. Here a plurality of potentials Pot_a, Pot_b . . . , Pot_n with sub-partitions Part_a, . . . , Part_n in between are assigned to one of the diagnostic partitions D1,D2,D3,D4, wherein a predefined scalar value is uniquely assigned to each potential Pot_a, Pot_b . . . , Pot_n.

[0094] Here the sub-partitions Part_a, . . . , Part_n as well as the diagnostic partitions D1,D2,D3,D4 can be arranged so that only two sub-partitions Pot_a, Pot_b . . . , Pot_n or two diagnostic partitions D1,D2,D3,D4 are adjacent to one another. Furthermore, advantageously only diagnostic partitions D1,D2,D3,D4 of consecutive states can be adjacent to one another. This can also be transferred to the sub-partitions Part_a, . . . , Part_n.

[0095] In addition, each of the sub-partitions Part_a, . . . , Part_n can be assigned an interpolation function, or a plurality of sub-partitions Part_a, . . . , Part_n have a common interpolation function. The characteristic values K1,K2, which lie on a potential Pot_a, Pot_b . . . , Pot_n, can be mapped to the scalar value defined for these potentials Pot_a, Pot_b . . . , Pot_n, just like the characteristic values K1,K2, which lie on a limit potential G1, G2,G3, can be mapped to the scalar value defined therefor.

[0096] FIG. 8 shows as a diagram 7 the characteristic values K1,K2 as time-dependent scalar values as a function of the limit potentials G1,G2,G3 with the diagnostic partitions D1,D2,D3,D4, plotted over time t, wherein the diagnostic partitions D1,D2,D3,D4 in this example represent consecutive states, i.e. false measurement, good state, warning state and alarm state.

[0097] This enables a time-dependent monitoring of the at least two characteristic values K1,K2. The monitoring device 1 can also be used for a plurality of characteristic values K1, . . . , Kn.

[0098] The combination of a plurality of characteristic values enables a more reliable assessment compared to the use of a single characteristic value; in particular, vibrations of components can also be monitored which can only be described by several characteristic values.

[0099] By evaluating the time paths of the scalar values for example an estimate can be made as to when a bearing may fail.

[0100] Diagram 7 can be seen as a continuous mapping of the characteristic values to a scalar value which represents the time development of the diagnostic state of the industrial machine 2. Thus a scalar value is assigned to each location in the multi-dimensional diagram 7. This value can then be used for trending. Here when trending the rate at which the vibration values change over the operating time is examined.

[0101] FIG. 9 shows further possible applications of the scalar value determined in this way, which is formed by two or more characteristic values. Thus this can be entered into a signal device 8, which emits a warning in time and/or is stored in a database 9 or logging device 10 to enable long-term monitoring.

[0102] FIG. 10 shows schematically the monitoring method for detecting a diagnostic state of an industrial machine 2 on the basis of vibrations which are associated with the industrial machine 2.

[0103] In step S1 two or more signals S1, . . . , Sn are detected by one or more measurement transducers 3a, . . . 3n in a time period and in a second step S2 converted by a signal processor 4a, . . . 4n into two or more characteristic values K1, . . . , Kn.

[0104] Afterwards, these characteristic values K1, . . . Kn are mapped as characteristic value tuple in a step S3 by one or more interpolation functions to scalar values, so that a time-dependent monitoring of the diagnostic state of the industrial machine 2 is achieved by the time-dependent scalar value.

TABLE-US-00001 REFERENCE SIGNS 1 monitoring device 2 industrial machine 3, 3a, . . . 3n    measurement transducer 4, 4a, . . . , 4n  signal processor 5 trigger unit 6 aggregating unit 7 diagram 8 signal device 9 database 10  logging device 11  storage unit S1, . . . , Sn signals K1, . . . , Kn characteristic values D1, . . . , D4 diagnosis partition F1, . . . , F4 interpolation functions G1, G2, G3 limit potentials Part_a,, Part_n sub-partitions Pot_a, . . . , Pot_n potentials

[0105] The various embodiments described above can be combined to provide further embodiments. These and other changes can be made to the embodiments in light of the above-detailed description. In general, in the following claims, the terms used should not be construed to limit the claims to the specific embodiments disclosed in the specification and the claims, but should be construed to include all possible embodiments along with the full scope of equivalents to which such claims are entitled.