Method for setting an alarm level for a machine

Abstract

A method for setting an alarm level of a machine, comprising: defining at least one state indicator reflecting a state of the machine with respect to a defect of the machine to be monitored, the at least one state indicator being defined from kinematic data of the machine, recording measurements of process-related parameters during a predetermined period of normal operation of the machine, calculating at least one state indicator value for the at least one state indicator using machine state data, determining a graph of the at least one state indicator value as a function of a first process-related parameter selected from the measured process-related parameters, dividing the graph into operating levels, each operating level representing a different operating state of the machine, calculating an alarm level value for each operating level, setting the determined alarm level value at a midpoint of the operating level for each operating level.

Description

Technical Field

[0001] This invention relates to machine diagnostics, and more particularly, to setting alarm levels for different operating conditions to detect machine defects. Background Technology

[0002] Machine diagnostics is used to monitor the health of machines. The main purpose of machine diagnostics is to detect defects in a mechanism as early as possible.

[0003] A common problem is that the machine's operating status is constantly changing, and therefore the measurement results of the status indicators are also changing, making it difficult to set appropriate alarm levels and increasing the risk of false alarms.

[0004] An inappropriate definition of alarm levels may result in less reliable alarms that are at risk of missing, defective, or false alarms being triggered.

[0005] One known method is to manually set different alarm levels for multiple operational states performed by machine condition monitoring experts.

[0006] However, this method requires highly skilled technicians and a large amount of manual work.

[0007] Document US2018 / 158314 discloses a method for trend analysis and adjustment of alarm parameters for machines, and document US2019 / 332102 discloses an automatic diagnostic method for performing extensive machine health monitoring on machine parts by analyzing the measurement results of machine parts to detect defects in the machine parts.

[0008] However, the determination of alert levels is not accurate enough.

[0009] At least some of the previously mentioned drawbacks need to be avoided, especially by enhancing the determination of alarm levels. Summary of the Invention

[0010] Based on one aspect, a method for setting alarm levels for machines is proposed.

[0011] The method includes:

[0012] – Define at least one state index that reflects the state of the machine with respect to the defects to be monitored in the machine, the at least one state index being defined by machine kinematic data.

[0013] - Record machine status data and process measurement results of relevant parameters during the predetermined period of normal machine operation.

[0014] - Calculate a status index value for the at least one status index based on the measurement results of each recorded machine status data.

[0015] - Determine a graph representing the value of at least one state index as a function of a first processing-related parameter selected from the measured processing-related parameters.

[0016] - Divide the chart into operation levels, with each operation level representing a different operating state of the machine.

[0017] - Calculate the alarm level value for each operational level.

[0018] - For each operation level, a defined alarm level value is set at the midpoint of the operation level.

[0019] Alarm levels are determined for each operational level, thereby enhancing the accuracy of alarm levels to detect machine defects at an early stage and avoid false alarms.

[0020] Kinematic data is used to calculate frequencies generated by defective machines or machine components.

[0021] Kinematic data include, for example, the machine's shaft speed, the number of teeth on the machine's gears, and the machine's bearing dimensions, which include the number of rolling elements or the number of blades on the machine's impeller.

[0022] Advantageously, the method also includes connecting the alarm level values ​​via linear interpolation.

[0023] Preferably, the alarm level value for each operation level is calculated from the average and standard deviation of at least one status indicator value among the considered operation levels, as well as a detection factor.

[0024] Advantageously, dividing the chart into the operation levels includes determining a lower bound and an upper bound for each operation level such that the change in at least one state indicator value in the operation level under consideration is less than a threshold defined for the operation level.

[0025] Preferably, when the learning period ends, if M out of the N values ​​of the at least one status indicator value exceed the alarm level value, an alarm is triggered.

[0026] Advantageously, the processing-related parameters include the speed of the machine and / or the load applied to the machine and / or the vibration of the machine.

[0027] On the other hand, a system for setting alarm levels for machines is proposed.

[0028] The system includes:

[0029] - Define a component for defining at least one state index, said at least one state index reflecting the state of the machine with respect to a monitored defect of the machine, said at least one state index being defined by the measurement results of machine kinematic data.

[0030] - A recording component for recording machine status data and processing measurement results of relevant parameters during a predetermined period of normal operation of the machine.

[0031] - A first computing unit is used to calculate a status index value for the at least one status index based on the measurement results of each recorded machine status data.

[0032] - A determining component for determining a graph of at least one state index value, represented as a function of a first processing-related parameter selected from the measured processing-related parameters.

[0033] - A division component is used to divide the chart into operation levels, each operation level representing a different operating state of the machine.

[0034] - The second calculation unit is used to calculate the alarm level value for each operational level.

[0035] - A setting component for setting a defined alarm level value at the midpoint of each operation level.

[0036] Preferably, the system further includes an interpolation component for connecting the alarm level values ​​via linear interpolation. Attached Figure Description

[0037] Other advantages and features of the invention will become apparent upon review of the detailed description of the embodiments (in no way limiting) and the accompanying drawings, in which:

[0038] [ Figure 1 An example of an embodiment of the machine according to the present invention is shown schematically;

[0039] [ Figure 2 An embodiment of the method for setting alarm levels for a machine according to the present invention is shown; and

[0040] [ Figure 3 The image shows an example of a chart showing the status indicators. Detailed Implementation

[0041] Reference Figure 1 This represents an example of an implementation of a machine 1 including sensor 2 and a condition monitoring system 3 connected to sensor 2.

[0042] Sensor 2 includes at least one machine status sensor 4 and at least one process related parameters sensor 5.

[0043] Machine status sensor 4 generates machine status data for machine 1, such as the power output of machine 1.

[0044] Machine status sensor 4 includes, for example, a power sensor.

[0045] Sensor 5 processes relevant parameters and measures the relevant parameters of machine 1.

[0046] The relevant parameters include, for example, operating speed, load, or vibration.

[0047] Sensor 5 includes, for example, a speed sensor, a load sensor, and / or a sensor configured to measure vibrations applied to machine 1.

[0048] The status monitoring system 3 includes a defining means component, a recording component, a first calculation component, a determining component, a dividing component, a second calculation component, a setting component, and an interpolation component.

[0049] Figure 2 This describes an implementation of a method for setting the alarm level for machine 1.

[0050] In step 10, the component defines at least one status indicator (CI), which reflects the status of the machine with respect to the defects to be monitored in machine 1.

[0051] The state index is defined by machine kinematic data.

[0052] The condition index CI includes, for example, the gear meshing frequency of the gears of machine 1, which is equal to the gear spindle speed multiplied by the number of teeth of the gear bearing, which is a kinematic data of machine 1.

[0053] In step 11, during the learning period (time period / cycle) of normal operation of machine 1, the recording unit records the measurement results of the processing-related parameters for the predetermined time period.

[0054] During the learning period, machine 1 operates under all operating conditions of normal machine use.

[0055] In step 12, after the learning period, the first computing unit calculates the state index value for the state index CI for each recorded machine state data measurement result.

[0056] In step 13, the determining component determines a graph GR representing the state index value as a function of a first processing-related parameter selected from the measured processing-related parameters. Assume, hereinafter, that the first processing-related parameter is the operating speed of machine 1.

[0057] Then, in step 14, the partitioning component divides the chart GR into operating classes OC, each operating class representing a different operating state of the machine.

[0058] Determine the lower bound and upper bound for each operation level such that the change in the state indicator value in the considered operation level is less than the threshold defined for that operation level.

[0059] For example, the operational level definition threshold is equal to 10% of the average of all CI values ​​obtained during the learning period.

[0060] The chart GR is divided into 5 to 10 operational levels, each of which includes, for example, 10 to 20 measurement results.

[0061] In step 15, the second calculation unit calculates the alarm level value AL for each operation level, and the alarm level value AL is equal to:

[0062] AL=μ+X.σ (1)

[0063] Where μ is the average value, σ is the standard deviation of the status index value in the considered operating level, and X is the detection factor.

[0064] X is, for example, included between 1 and 10.

[0065] In step 16, a defined alarm level value is set at the midpoint of each operating level by setting the component.

[0066] In step 17, the interpolation unit connects the alarm level values ​​together through linear interpolation, thereby further improving the accuracy (or precision) of alarm levels between the midpoints of the operation levels.

[0067] The status monitoring system 3 classifies the operation status into multiple operation levels based on the monitored processing parameters.

[0068] Figure 3 An example of a graph GR showing the relationship between the state index CI and the first processing-related parameter PP1.

[0069] The measurement results used to define the state index CI are represented by points.

[0070] The GR icon is divided into 7 operational levels, OC1 to OC7.

[0071] Each operational level OC1 to OC7 comprises calculated alarm level values ​​AL1 to AL7 represented by a cross at their midpoint, and alarm level values ​​AL1 to AL7 are connected together by linear interpolation.

[0072] When machine 1 is operating and the learning period ends, if M out of N status indicator values ​​exceed the alarm level values ​​AL1 to AL7, an alarm is triggered, where M and N are integers.

[0073] For example, M equals 4 and N equals 7.

[0074] Without any manual input, alarm levels are automatically set for different operation states, and each operation state is represented by the operation level on the chart GR.

[0075] Alarm levels are determined for each operational level to enhance the accuracy of alarm levels, thereby detecting defects in machine 1 at an early stage and avoiding false alarms ( / warnings).

[0076] In the example shown, a status index reflecting the state of the machine regarding the defects to be monitored was analyzed.

[0077] To detect more defects, as described above, more charts were created that were divided into operational levels and included alarm level values. Each chart included different status indicators relative to the first processing-related parameter PP1. The status indicators reflected the machine's status with respect to the different defects that would be monitored.

Claims

1. A method for setting an alarm level for a machine (1), the method comprising: - Define at least one status index (CI), which reflects the state of the machine with respect to the monitored defect of the machine (1), and the at least one status index is defined by machine kinematic data. - Record machine status data and process measurement results of relevant parameters during the predetermined period of normal machine operation. - Calculate a status index value for the at least one status index (CI) based on the measurement results of each recorded machine status data. - Determine a graph (GR) of the at least one state index value, representing a function of a first treatment-related parameter (PP1) selected from the measured treatment-related parameters. - The chart is divided into operation levels (OC1, OC2, OC3, OC4, OC5, OC6, OC7), each operation level representing a different operation state of the machine (1). - Calculate alarm level values ​​(AL1, AL2, AL3, AL4, AL5, AL6, AL7) for each operation level (OC1, OC2, OC3, OC4, OC5, OC6, OC7). - For each operation level (OC1, OC2, OC3, OC4, OC5, OC6, OC7), set a defined alarm level value (AL1, AL2, AL3, AL4, AL5, AL6, AL7) at the midpoint of the operation level.

2. The method according to claim 1, characterized in that, It also includes connecting the alarm level values ​​(AL1, AL2, AL3, AL4, AL5, AL6, AL7) via linear interpolation.

3. The method according to any one of claims 1 and 2, characterized in that, The alarm level value (AL1, AL2, AL3, AL4, AL5, AL6, AL7) for each operating level (OC1, OC2, OC3, OC4, OC5, OC6, OC7) is calculated from the average and standard deviation of at least one status indicator value among the considered operating levels and a detection factor, wherein the detection factor is included between 1 and 10.

4. The method according to any one of claims 1 and 2, characterized in that, Dividing the chart into operational levels (OC1, OC2, OC3, OC4, OC5, OC6, OC7) includes: determining the lower and upper bounds of each operational level (OC1, OC2, OC3, OC4, OC5, OC6, OC7) such that the change in at least one status index (CI) value in the considered operational level is less than the operational level defined threshold.

5. The method according to any one of claims 1 and 2, characterized in that, When the learning period ends, if M out of the N values ​​of the at least one status indicator exceed the alarm level value (AL1, AL2, AL3, AL4, AL5, AL6, AL7), an alarm is triggered.

6. The method according to any one of claims 1 and 2, wherein, The processing-related parameters include the speed of the machine and / or the load applied to the machine and / or the vibration of the machine.

7. A system (3) for setting alarm levels for a machine (1), the system comprising: - Define a component for defining at least one condition index (CI), said at least one condition index (CI) reflecting the state of the machine with respect to a monitored defect of the machine (1), said at least one condition index being defined by the measurement results of machine kinematic data. - A recording component for recording machine status data and processing measurement results of relevant parameters during a predetermined period of normal operation of the machine. - A first computing unit is used to calculate a status index value for the at least one status index (CI) based on the measurement results of each recorded machine status data. - A determining component for determining a graph (GR) of the at least one state index value, represented as a function of a first processing-related parameter (PP1) selected from the measured processing-related parameters. - A division component is used to divide the chart into operation levels (OC1, OC2, OC3, OC4, OC5, OC6, OC7), each operation level representing a different operation state of the machine (1). - The second calculation unit is used to calculate the alarm level value (AL1, AL2, AL3, AL4, AL5, AL6, AL7) for each operation level (OC1, OC2, OC3, OC4, OC5, OC6, OC7). - Setting component for setting the determined alarm level value (AL1, AL2, AL3, AL4, AL5, AL6, AL7) at the midpoint of each operation level (OC1, OC2, OC3, OC4, OC5, OC6, OC7).

8. The system according to claim 7, characterized in that, It also includes an interpolation component for connecting the alarm level values ​​(AL1, AL2, AL3, AL4, AL5, AL6, AL7) via linear interpolation.

Images