Procedure and system for the assessment of a technical installation

Abstract

A method for evaluating a technical installation is proposed. During an operating phase, an operating data set comprising a plurality of measured variables is recorded. Each measured variable comprises a plurality of measured values, and the measured values ​​of each measured variable are recorded sequentially. At least one spectrum of each measured variable in the operating data set is calculated. The calculated spectra of the operating data set are examined using a previously trained anomaly model, and a warning message is issued if an anomaly is detected in the spectra of the measured variables during the examination of the spectra of the operating data set. A system according to the invention for evaluating a technical installation comprises a data memory and an anomaly model. The system is configured to carry out the method according to the invention.

Description

[0001] The invention relates to a method for evaluating a technical installation, wherein an operating data set comprising a plurality of measured variables is recorded during an operating phase, each measured variable comprising a plurality of measured values, and the measured values ​​of each measured variable are recorded sequentially, and at least one spectrum of each measured variable of the operating data set is calculated. The invention also relates to a system for evaluating a technical installation, which is configured to carry out the method according to one of the preceding claims.

[0002] Technical systems, such as drive systems, comprise a multitude of technical components, such as an electric motor, a converter for generating a three-phase alternating voltage for the electric motor, and a gearbox for converting the electric motor's speed. Examples of such technical systems include turntables, conveyor belts, or storage and retrieval machines. The components of such technical systems can malfunction after extended periods of operation due to wear and tear.

[0003] It is known to monitor such technical systems by recording and evaluating specific measured values ​​at specific times. If the recorded measured values ​​deviate too significantly from previously defined target values, a fault in the system is assumed, and a corresponding message is issued to the operator. For example, a current generated by the converter that exceeds a defined limit can indicate that the gearbox is becoming stiff due to wear.

[0004] EP 3 100 064 B1 discloses a method and a device for fault detection in a machine comprising a drive unit. The frequency spectrum of a measured variable characterizing the electrical power consumption of the drive unit is determined, and in an analysis step, the frequency spectrum of the measured variable is evaluated for abnormalities that indicate faults.

[0005] DE 10 2006 058 689 A1 discloses a device and method for diagnosing the condition of a machine component. A vibration of a machine component is detected and converted into an electrical signal, and a frequency spectrum is generated from the electrical signal.

[0006] From DE 10 2019 108 226 A1, a motor shaft arrangement for an internal combustion engine with a motor shaft is known, which has a detection system for detecting a rotational position of the motor shaft and an angle calculation unit for determining a position signal of the motor shaft.

[0007] EP 3 100 064 B1 discloses a device and method for detecting faults in a machine comprising an electric drive unit. The frequency spectrum of a measured variable characterizing the electrical power consumption of the drive unit is determined, and the frequency spectrum of the measured variable is evaluated in an analysis step.

[0008] From DE 11 2019 007 464 T5 a drive noise diagnostic system is known which includes a drive noise detection unit which detects a drive noise of an actuator or a machine to be driven, an operating state detection unit which records a drive position, a drive speed or a force generated by the drive of the actuator in a time series, a sound vibration time series spectrum recording unit, a feature point extraction unit and a cause determination unit.

[0009] The invention is based on the object of developing a method and a system for evaluating a technical system.

[0010] The problem is solved by a method for evaluating a technical system having the features specified in claim 1. Advantageous embodiments and further developments are the subject of the subclaims. The problem is also solved by a system for evaluating a technical system having the features specified in claim 13.

[0011] A method for evaluating a technical system is proposed. During an operating phase, an operating data set comprising a plurality of measured variables is recorded. Each measured variable comprises a plurality of measured values, and the measured values ​​of each measured variable are recorded sequentially. At least one spectrum is calculated for each measured variable in the operating data set.

[0012] The technical system includes a converter for generating alternating voltage for an electric motor. The recorded operating data set includes a speed and a torque as measured variables. The measured values ​​of the speed and torque are recorded synchronously.

[0013] This involves an order reduction of the recorded operating data set before calculating the spectra of the operating data set. A traveled angle is calculated from the measured speed, and the torque is calculated using angular equidistant interpolation.

[0014] The calculated spectra of the operational dataset are examined using a previously trained anomaly model, and a warning message is issued if an anomaly in the spectra of the measured quantities is detected when examining the spectra of the operational dataset.

[0015] An analysis of spectra of measured variables based on fixed, predefined criteria, such as frequency ranges or amplitude, is limited to the predefined criteria. Further criteria that indicate wear or a defect in the technical system and that are not easily defined mathematically cannot be detected using known methods. Examining the calculated spectra using a previously trained anomaly model also allows the detection of anomalies that are not easily defined mathematically.

[0016] The technical system is, for example, a drive system comprising, for example, an electric motor, a converter, and a gear for converting the speed of the electric motor, for example, a turntable, a conveyor belt, or a storage and retrieval machine. The aforementioned measured variables, speed and torque, are measured indirectly via the converter by measuring the frequency and current of the output current to the electric motor. The components of such technical systems can malfunction after extended periods of operation due to wear and tear that occurs during operation. Such wear and tear can be detected early using the method according to the invention.

[0017] The order reduction of the recorded operating data set is carried out in particular when no operating points are defined in which the speed and torque lie in a defined range, i.e. when no ranges with approximately constant speed and constant torque are known.

[0018] According to an advantageous embodiment of the invention, the examination of the spectra of the operational data set is carried out by a neural network configured for anomaly detection, for example an autencoder or a variational autoencoder, a principal component analysis, a k-nearest neighbors, a one-class support vector machine, an isolation forest or ensemble methods of the preceding algorithms.

[0019] According to an advantageous embodiment of the invention, a spectrum is calculated for each measured variable of the operating data set as a frequency spectrum from the measured values ​​of the measured variable.

[0020] The frequency spectrum is preferably calculated from the measured values ​​of the measured quantity using a fast Fourier transformation.

[0021] According to an advantageous embodiment of the invention, envelope curves are determined around the measured values ​​of each measured variable of the operating data set, and for each measured variable of the operating data set, a spectrum is calculated as an envelope spectrum from the determined envelope curves.

[0022] The envelope spectrum is preferably calculated from the determined envelopes using a fast Fourier transformation.

[0023] According to an advantageous embodiment of the invention, the operating data set is assigned by a classifier to a previously defined operating point before the spectra of the operating data set are calculated, wherein in each defined operating point the speed and the torque lie in a defined range.

[0024] At such an operating point, the speed, and thus the frequency of the output current, and the torque, and thus also the current intensity of the output current, are approximately constant. Operating points describe application cycles of the technical system that are performed regularly, such as forward travel, reverse travel, generator operation, motor operation, transport of a light load, and transport of a heavy load.

[0025] According to an advantageous embodiment of the invention, when examining the calculated spectra of the operational data set, an anomaly score is determined based on the previously trained anomaly model, which is then compared with previously determined reference anomaly scores. An anomaly in the spectra of the measured variables is detected if the determined anomaly score deviates from the previously determined reference anomaly scores.

[0026] When the anomaly model is applied to training data during a reference phase, reference anomaly scores are determined when the technical system is in good condition. By evaluating the reference anomaly scores thus determined, a model is advantageously created that classifies anomaly scores determined during the operational phase as normal or anomalous. It is advantageous to determine the anomaly scores during the operational phase using cross-validation, which allows for even more specific anomaly score determination.

[0027] According to an advantageous embodiment of the invention, several reference data sets are recorded in preparation during a reference phase, each of which comprises a plurality of measured variables. Each measured variable comprises a plurality of measured values, and the measured values ​​of each measured variable are recorded sequentially. At least one spectrum of each measured variable of each reference data set is calculated, and the anomaly model is trained using the calculated spectra of the reference data sets in order to examine future operational data sets.

[0028] For example, a spectrum is calculated as a frequency spectrum for each measured variable in each reference data set from the measured values ​​of the measured variable. For example, envelopes are determined around the measured values ​​of each measured variable in each reference data set, and a spectrum is calculated as an envelope spectrum for each measured variable in each reference data set from the determined envelopes. The calculation of the frequency spectrum from the measured values ​​of the measured variable and the calculation of the envelope spectrum from the determined envelopes are preferably carried out using a fast Fourier transform.

[0029] According to an advantageous embodiment of the invention, the technical system comprises a converter for generating an alternating voltage for an electric motor, and the recorded reference data sets each comprise a speed and a torque as measured variables. The measured values ​​of the speed and torque are recorded synchronously.

[0030] According to an advantageous embodiment of the invention, a preprocessing algorithm is used to check whether the speed and torque are constant in the recorded reference data sets during at least one defined time window.

[0031] The aforementioned test using the preprocessing algorithm is performed especially when no order reduction is performed on the recorded reference data sets before the spectra of the reference data sets are calculated. If an order reduction of the recorded reference data sets is performed, a traveled angle is calculated from the measured speed, and the torque is calculated using an interpolation method based on angular equidistance.

[0032] According to an advantageous embodiment of the invention, if the rotational speed and the torque are constant during at least one defined time window, a spectrum of the measured variable of the respective reference data set is calculated only from the measured values ​​of each measured variable of each reference data set recorded during said time window.

[0033] According to an advantageous embodiment of the invention, reference data sets that exhibit a relatively high degree of similarity to one another are identified as belonging to a common cluster. The reference data sets belonging to a common cluster are assigned to a common operating point, and a classifier is trained to assign future operating data sets to the previously defined operating points.

[0034] At such an operating point, the speed and torque are thus approximately constant. Operating points describe application cycles of the technical system that are performed repeatedly on a regular basis, such as forward travel, reverse travel, generator operation, motor operation, transport of a light load, and transport of a heavy load.

[0035] According to an advantageous embodiment of the invention, the anomaly model determines reference anomaly scores from the calculated spectra of the reference data sets, which are comparable with anomaly scores determined in the future.

[0036] The anomaly model is applied to training data during the reference phase, determining reference anomaly scores when the technical system is in good condition. By evaluating the resulting reference anomaly scores, a model is advantageously created that classifies anomaly scores determined during the operational phase as normal or anomalous. The determination of the reference anomaly scores during the reference phase is advantageously carried out using cross-validation, which allows for even more specific determination of the reference anomaly scores.

[0037] A system according to the invention for evaluating a technical installation comprises a data storage device and an anomaly model. The system is configured to execute the method according to the invention.

[0038] The system according to the invention allows an evaluation of the technical system based on a previously trained anomaly model, and in particular the detection of anomalies that cannot be easily defined mathematically.

[0039] The invention is not limited to the combination of features in the claims. Further possible combinations of claims and / or individual claim features and / or features of the description and / or the figures will become apparent to those skilled in the art, particularly from the problem and / or the problem posed by comparison with the prior art.

[0040] The invention will now be explained in more detail with reference to the accompanying drawings. The invention is not limited to the exemplary embodiments shown in the drawings. The drawings only represent the subject matter of the invention schematically. They show: Fig. 1: a flowchart of a recording of a reference data set during a reference phase, Fig. 2: a flowchart of processing recorded reference data sets during the reference phase and Fig. 3: a flowchart of processing recorded operational data records during an operational phase.

[0041] Fig. Figure 1 shows a flowchart of the acquisition of a reference data set 10 in a technical system during a reference phase. The reference phase is carried out preparatory to generate spectra of measured variables and train an anomaly model, which will later be required to evaluate the technical system.

[0042] The technical system in question is a drive system. The drive system comprises an electric motor, a converter for generating a three-phase alternating voltage for the electric motor, and a gearbox for converting the electric motor's speed.

[0043] During the reference phase, a reference data set 10 is recorded. The recorded reference data set 10 comprises a plurality of measured variables. These measured variables include, in particular, the speed and torque of the electric motor. These measured variables, speed and torque, are measured indirectly via the converter, for example, by measuring the frequency and current of the output current to the electric motor.

[0044] Each measured variable comprises a plurality of measured values, for example, 2048 measured values. The measured values ​​of each measured variable are recorded sequentially and synchronously. For example, the measured values ​​of each measured variable are recorded in equidistant time intervals of 5 ms each. The recorded reference data set 10 thus comprises, for example, two different measured variables, each with 2048 measured values.

[0045] In an optional first step 101, a preprocessing algorithm is used to check whether the speed and torque are constant in the recorded reference data set 10 during a defined time window. Such a time window with constant speed and constant torque corresponds to a constant travel of the technical system, i.e., a travel with a constant load and constant speed.

[0046] If such time windows with constant speed and constant torque are found, only the measured values ​​of the measured variables of reference data set 10 recorded during said time windows are retained. The remaining measured values ​​that lie outside said time windows are deleted and therefore not retained.

[0047] The recorded reference data set 10 is stored in a data storage 20 in a second step 102. If measured values ​​were deleted in the optional step 101, the deleted measured values ​​are not stored in the data storage 20.

[0048] During the reference phase, several more reference data sets 10 are recorded. The second step 102 is repeated, and the recorded reference data sets 10 are stored in the data memory 20. Optionally, the first step 101, in which measured values ​​are deleted, is also repeated. Another reference data set 10 is recorded when a defined trigger condition is met. Such a trigger condition is, for example, the start of an application cycle of the technical system, for example when a speed of an electric motor of the technical system exceeds a predetermined minimum speed. Such a trigger condition is also, for example, when the recorded measured values ​​of the measured variables of the reference data set 10 are at least approximately constant over a period of, for example, 10 seconds.

[0049] Fig. Figure 2 shows a flowchart of processing recorded reference data sets 10 during the reference phase. The processing of the recorded reference data sets 10 according to Fig. 2 occurs only once. In a step 103, the previously recorded and stored reference data sets 10 are loaded from the data storage 20.

[0050] Reference data sets 10 that exhibit a relatively high degree of similarity to one another belong to a common cluster and are identified as belonging to a common cluster. In a step 104, reference data sets 10 that belong to a common cluster and have been identified as belonging to a common cluster are assigned to a common operating point.

[0051] The clusters are detected automatically using a clustering method, such as DBSCAN. Alternatively, the clusters are detected by manually assigning a working point. Reference data sets 10 that do not exhibit a sufficiently high similarity to any of the other reference data sets 10 are discarded.

[0052] In a step 105, a classifier is trained using the reference data sets 10 and the operating points assigned to the reference data sets 10 in step 104. The classifier is trained using a machine learning method to assign future recorded operating data sets 12 to the previously defined operating points. The goal is an error-free assignment of recorded operating data sets 12 to previously defined operating points.

[0053] In a step 106, the classifier and the operating points assigned to the reference data sets 10 are stored in the data memory 20.

[0054] Optionally, an order reduction of the recorded reference data sets 10 is performed in a step 107. In this case, a traveled angle of rotation is calculated from the measured rotational speed, and the torque is calculated angularly equidistantly by means of interpolation.

[0055] In a step 108, spectra of each measured variable of each reference data set 10 are calculated. For example, a spectrum of each measured variable of each reference data set 10 is calculated as a frequency spectrum from the measured values ​​of the measured variable. Alternatively or additionally, envelopes are determined around the measured values ​​of each measured variable of each reference data set 10, and a spectrum of each measured variable of each reference data set 10 is calculated as an envelope spectrum from the determined envelopes.

[0056] In a step 109, an anomaly model is trained using the calculated spectra of the reference data sets 10. The anomaly model is trained to examine future operational data sets 12. The anomaly model is also stored in the data storage 20.

[0057] Fig. Figure 3 shows a flowchart of the processing of recorded operational data sets 12 during an operational phase. During the operational phase, the evaluation of the technical system is carried out.

[0058] During the operating phase, an operating data record 12 is recorded. The recorded operating data record 12 is stored in the data memory 20 in a first step 121.

[0059] The recorded operating data set 12 comprises a plurality of measured variables. These measured variables include, in particular, the speed and torque of the electric motor. These measured variables, speed and torque, are measured indirectly via the converter, for example, by measuring the frequency and current of the output current to the electric motor.

[0060] Each measured variable comprises a plurality of measured values, for example, 2048 measured values. The measured values ​​of each measured variable are recorded sequentially and synchronously. For example, the measured values ​​of each measured variable are recorded in equidistant time intervals of 5 ms each. The recorded operating data set 12 thus comprises, for example, two different measured variables, each with 2048 measured values.

[0061] In a step 122, the recorded operating data set 12 is fed to the classifier. The classifier assigns the operating data set 12 to one of the previously defined operating points. The operating data set 12 is assigned to the operating point with the greatest similarity.

[0062] In a step 123, an order reduction of the recorded operating data set 12 is performed. A traveled angle of rotation is calculated from the measured speed, and the torque is calculated with angular equidistance by means of interpolation.

[0063] In a step 124, spectra of each measured variable of each operating data set 12 are calculated. For example, a spectrum of each measured variable of each operating data set 12 is calculated as a frequency spectrum from the measured values ​​of the measured variable. Alternatively or additionally, envelope curves are determined around the measured values ​​of each measured variable of each operating data set 12, and a spectrum of each measured variable of each operating data set 12 is calculated as an envelope spectrum from the determined envelope curves.

[0064] In a step 125, the anomaly model is loaded from the data storage 20. In particular, the anomaly model that corresponds to the operating point to which the operating data set 12 was assigned by the classifier in step 122 is loaded.

[0065] In a step 126, the spectra of the operational data set 12 calculated in step 124 are examined using the previously trained anomaly model.

[0066] In a step 127, a warning message is issued if an anomaly in the spectra of the measured variables is detected during the examination of the spectra of the operating data set 12.

[0067] The aforementioned steps 121 to 127 are executed repeatedly. In particular, the aforementioned steps are repeated when a defined trigger condition is met and when a defined period of time has elapsed since the previous operating data record was recorded.

[0068] Such a trigger condition is, for example, the start of an application cycle of the technical system, if, for example, a speed of an electric motor of the technical system exceeds a specified minimum speed.

[0069] For example, the time span is defined as six hours. Thus, operational data records 12 are recorded and processed only four times per day. List of reference symbols 10 Reference data set 12 Operating data record 20 data storage 101..109 steps during the reference phase 121..127 steps during the operating phase

Claims

[1] Method for evaluating a technical installation, wherein During an operational phase, an operational data record (12) is recorded, which includes a plurality of measured quantities, and wherein Each measurement quantity comprises a plurality of measured values, and wherein the measured values ​​of each measured quantity are recorded sequentially over time, and at least one spectrum of each measured quantity of the operational data set (12) is calculated, characterized by , that the technical system includes a converter for generating an alternating voltage for an electric motor, and that the recorded operating data set includes rotational speed and torque as measured variables, and that The measured values ​​of the rotational speed and torque are recorded synchronously, and that an order reduction of the recorded operating data set (12) is performed before the spectra of the operating data set (12) are calculated, whereby a rotation angle traveled is calculated from the measured rotational speed, and The torque is calculated angularly equidistant by means of interpolation, and that the calculated spectra of the operational data set (12) are examined using a previously trained anomaly model, and that A warning message is issued if an anomaly in the spectra of the measured quantities is detected during the examination of the spectra of the operational data set (12). [2] Method according to any of the preceding claims, characterized by , that the examination of the spectra of the operational data set (12) is performed by a neural network configured for anomaly detection, for example an autoencoder or a variational autoencoder, a principal component analysis, a k-nearest neighbors, a one-class support vector machine, an isolation forest, or ensemble methods of the preceding algorithms. [3] Method according to any of the preceding claims, characterized by , that for each measured quantity of the operational data set (12) a spectrum is calculated as a frequency spectrum from the measured values ​​of the measured quantity. [4] Method according to any of the preceding claims, characterized by , that envelopes around the measured values ​​of each measured quantity of the operational data set (12) are determined, and that For each measured variable of the operating data set (12), a spectrum is calculated as an envelope spectrum from the determined envelopes. [5] Method according to any of the preceding claims, characterized by, that the operating data set (12) is assigned to a previously defined operating point by a classifier before the spectra of the operating data set (12) are calculated, wherein in each defined operating point the rotational speed and torque are within a defined range. [6] Method according to any of the preceding claims, characterized by , that when examining the calculated spectra of the operational data set (12) using the previously trained anomaly model, an anomaly score is determined which is compared with previously determined reference anomaly scores, and that an anomaly in the spectra of the measured quantities is detected if the determined anomaly score differs from the previously determined reference anomaly scores. [7] Method according to any of the preceding claims, characterized by that in preparation during a reference phase several reference data sets (10) are included, which each comprise a plurality of measured quantities, and that Each measurement quantity comprises a plurality of measured values, and that the measured values ​​of each measured quantity are recorded sequentially over time, and that at least one spectrum of each measured quantity of each reference data set (10) is calculated, and that the anomaly model is trained with the calculated spectra of the reference data sets (12) in order to examine future recorded operational data sets (12). [8] Method according to claim 7, characterized by , that the technical system includes a converter for generating an alternating voltage for an electric motor, and that the recorded reference data sets (10) each include a rotational speed and a torque as measured quantities, and that The measured values ​​of the rotational speed and torque are recorded synchronously. [9] Method according to claim 8, characterized by, that a preprocessing algorithm is used to check whether the rotational speed and torque are constant in the recorded reference data sets (10) during at least one defined time window. [10] Method according to claim 9, characterized by , that if the rotational speed and torque are constant for at least a defined time window, only from the measured values ​​of each measurement of each reference data set (10) recorded during the said time window, a spectrum of the measurement of each reference data set (10) is calculated. [11] Method according to any one of claims 8 to 10, characterized by , that reference data sets (10) which exhibit a relatively high similarity to each other are recognized as belonging to a common cluster, and that the reference datasets (10) belonging to a common cluster are assigned to a common operating point, and that a classifier is trained to assign future recorded operational data sets (12) to the previously defined operating points. [12] Method according to any one of claims 8 to 11, characterized by , that reference anomaly scores are determined from the anomaly model from the calculated spectra of the reference data sets (12) which are comparable with anomaly scores determined in the future. [13] System for the evaluation of a technical installation, comprising a data storage device (20) and an anomaly model, whereby the system is set up to carry out the method according to one of the preceding claims.

Images