DE112019007464T5 - ENGINE NOISE DIAGNOSTIC SYSTEM, ENGINE NOISE DIAGNOSTIC METHOD, MACHINE LEARNING DEVICE FOR ENGINE NOISE DIAGNOSTIC SYSTEM, PROGRAM, RECORDING MEDIA AND MEMORY - Google Patents

Abstract

A drive noise diagnostic system (10) includes a drive noise detection unit (11) that detects a drive noise of an actuator or a machine to be driven, an operating state detection unit (12) that detects a drive position, a drive speed or a force generated by driving the actuator detected in a time series, a sound vibration time-series spectrum detecting unit (15), a feature point extracting unit (16), and a cause determining unit (18). The sound vibration time-series spectrum acquisition unit outputs a time-series spectrum in which the power of a frequency spectrum calculated from the time-series data of driving noise is linked with frequency and time. The feature point extraction unit outputs feature point data in which the frequency of a feature point in the time-series spectrum, the time, and the driving position, the driving speed, or the force generated by driving the actuator at the time are combined as a set. The cause determining unit determines a cause of the driving noise by comparing a numeric value of the feature point data with a numeric range when combining the frequency or time including the feature point and the operation data at the time for each phenomenon as multidimensional data.

Description

area

The present invention relates to a powertrain noise diagnosis system that diagnoses a powertrain noise of an actuator or a machine to be driven by the actuator, a powertrain noise diagnosis method, and a machine learning apparatus for the powertrain noise diagnosis system.

background

It is well known that in a device having an energy source such as e.g. B. contains a motor, a driving noise of the energy source or a machine as a driving object of the energy source contains numerous information about a state of the energy source and the driving object. For example, when an abnormality occurs in the power source or the driving object, noise or vibration different from that in a normal state is generated. For this reason, a diagnostic device that diagnoses whether the noise or vibration is abnormal or not is known. However, when the device is put into operation for the first time or when it is operated immediately after changing the machine configuration and the driving pattern of the device, it is difficult to immediately determine whether the driving noise is normal. Therefore, there is a need for a technique that reduces engine noise, i. H. Noise or vibration generated by the device is detected using a sensor and easily identifies the cause of the drive noise. In particular, there is a need for a technique that performs the diagnosis using the engine noise in a simpler and more versatile manner without restricting the machine configuration and the setting of the device.

Patent Literature 1 discloses a technique that measures noise or vibration generated by a device including a rotating device and identifies the presence or absence of an abnormality or a cause of an abnormality in the device. According to this technique, first, before actually operating the device, a set of characteristics related to frequency and time is previously acquired for each cause of an anomaly regarding noise generated by a power source and a machine to be driven by the power source at the time , at which an abnormal noise is generated. The feature related to the frequency is a frequency at which a peak appears from a temporal change of a spectrum obtained by performing short-time Fourier transform or the like on time-series data of the measured driving noise. The feature with respect to time is a time interval in which a peak occurs for each frequency feature value. Subsequently, at the time of actual operation of the device, the characteristics in terms of frequency and time acquired during actual operation by a similar method are compared with the characteristics in terms of frequency and time acquired in advance, so that the presence or to identify the absence and cause of an anomaly.

List of documents cited

patent literature

Patent Literature 1: Japanese Patent Application Laid-Open No. H10-274558

abstract

Technical problem

However, the technique described in Patent Literature 1 identifies the presence or absence and the cause of an abnormality in the device only based on data obtained by measuring the driving noise generated by the actuator or the machine to be driven. Therefore, there is a problem that it is difficult to identify the cause of an abnormality depending on the position or the rotation speed of the actuator or the machine to be driven.

The present disclosure was made in view of the above, and an object of the present invention is to provide a driving noise diagnostic system capable of identifying a cause of an abnormality depending on a position or rotation speed of an actuator or a machine to be driven .

the solution of the problem

To solve the above problem and achieve the object, a powertrain noise diagnosis system according to the present invention includes a powertrain noise detection unit, an operating state detection unit, a sound vibration time-series spectrum acquisition unit, a feature point extraction unit, and a cause determination unit. The driving noise detection unit detects driving noise, ie noise or mechanical noise Vibrations generated in an actuator or a machine to be driven by the actuator. The operating state detection unit detects a drive position, a drive speed, or a force generated by driving the actuator in a time series. The sound vibration time-series spectrum acquisition unit calculates a frequency spectrum corresponding to each time point of sound vibration data, that is, time-series data of detected driving noise, and outputs a time-series spectrum in which the power of the calculated frequency spectrum is related to frequency and time as a set. The feature point extraction unit extracts, as a feature point, a point where a waveform of the power of the time-series spectrum satisfies a predetermined condition in terms of frequency and time, and outputs feature point data in which the frequency and time of the feature point, the waveform of the feature point and operational data, ie, the driving position, the driving speed, or the force generated by driving the actuator at the time corresponding to the feature point are combined as a set. The cause determining unit determines a cause of the detected driving noise by comparing a cause determining condition with a numerical value of the feature point data, the cause determining condition for each phenomenon occurring in the actuator or the machine to be driven as a cause of the driving noise defining a first numerical range when at least one of the frequency and time including the feature point occurring with the phenomenon and the operation data, ie the driving position, the driving speed or the force generated by driving the actuator at the time, are combined as multi-dimensional data.

Advantageous Effects of the Invention

The driving noise diagnosis system according to the present invention has an effect of being able to identify the cause of the abnormality depending on the position or the rotational speed of the actuator or the machine to be driven.

character list

• 1 14 is a block diagram showing an example of a functional configuration of a powertrain noise diagnostic system according to a first embodiment.
• 2 14 is a diagram showing an example of a hardware configuration when the drive noise diagnosis system according to the first embodiment is applied to a hoist.
• 3 12 is a block diagram showing an example of a hardware configuration of an arithmetic processing unit according to the first embodiment.
• 4 FIG. 12 is a block diagram schematically showing an example of a functional configuration of the arithmetic processing unit of FIG 2 according to the first embodiment.
• 5 14 is a flowchart showing an example of a processing operation of a driving noise diagnosing method according to the first embodiment.
• 6 14 is a graph showing an example of sound vibration data according to the first embodiment.
• 7 14 is a diagram showing an example of operational data according to the first embodiment.
• 8th 14 is a flowchart showing an example of a procedure of synchronization processing for the acoustic vibration data and the operational data according to the first embodiment.
• 9 12 is a diagram for explaining a process of synchronizing the sound vibration data and the operational data according to the first embodiment.
• 10 Fig. 12 is a diagram showing an example in which the operation data of 7 divided into operating modes.
• 11 Fig. 13 is a three-dimensional graph with three axes of time, frequency and power showing an example of time-series spectral data obtained by frequency transforming the sound vibration data of Figs 6 were received.
• 12 Fig. 13 is a three-dimensional plot with three axes for time, frequency and power, showing an example of one from the time-series spectral data of 11 extracted peak of the noise.
• 13 14 is a diagram showing an example of a hardware configuration when a drive noise diagnosis system according to a second embodiment is applied to an extraction unit.
• 14 14 is a block diagram schematically showing an example of a functional configuration of a server device according to the second embodiment.
• 15 14 is a table showing an example of recording device data according to the second embodiment.
• 16 14 is a block diagram schematically showing an example of a functional configuration of a user terminal according to the second embodiment.
• 17 Fig. 12 is a block diagram showing an example of a hardware configuration of the server device and the user terminal.
• 18 14 is a block diagram showing an example of a functional configuration of a cause determination unit in a powertrain noise diagnostic system according to a third embodiment.
• 19 14 is a block diagram showing an example of a functional configuration of a machine learning device for a powertrain noise diagnosis system according to a fourth embodiment.
• 20 14 is a block diagram schematically showing an example of a functional configuration of a server device according to the fourth embodiment.
• 21 14 is a block diagram schematically showing an example of a functional configuration of a user terminal according to the fourth embodiment.

Description of Embodiments

Hereinafter, a powertrain noise diagnosis system, a powertrain noise diagnosis method, and a machine learning apparatus for the powertrain noise diagnosis system according to embodiments of the present invention will be described in detail with reference to the drawings. It should be noted that the present invention is not limited to these embodiments.

First embodiment.

1 14 is a block diagram showing an example of a functional configuration of a powertrain noise diagnostic system according to a first embodiment. A driving noise diagnosis system 10 includes a driving noise detection unit 11, an operating state detection unit 12, a time synchronization unit 13, an operating mode extraction unit 14, an acoustic vibration time-series spectrum acquisition unit 15, a feature point extraction unit 16, an operating vibration extraction unit 17 and a cause determination unit 18

The driving noise detection unit 11 detects noise or vibration generated by an actuator and a machine to be driven, i. H. a machine driven by the actuator. In the following, the noise or vibrations generated by the actuator and the machine to be driven are referred to as drive noise. The driving noise detection unit 11 is a sensor that detects noise or vibration. A microphone that detects a sound and a vibration sensor such as An acceleration sensor, for example, are examples of the driving sound detection unit 11. The driving sound detection unit 11 can detect a sound generated by a driving device that drives the actuator.

The operating state detection unit 12 detects an operating state of the actuator connected to the drive device in a time series. The operational state of the actuator includes a driving position, a driving speed, or a force generated by driving the actuator. The actuator is, for example, a motor. Examples of the operating condition detection unit 12 include an encoder that detects the rotation angle of the motor, a linear scale that detects the position of a linear motor, a position sensor, an odometer, a distance sensor, a speed detector, a current detector, an acceleration sensor, a gyro sensor, and a force sensor.

The time synchronization unit 13 synchronizes the time of sound vibration data, i. H. Time-series data of the driving noise detected by the driving noise detection unit 11, and the time of operation data, i. H. Time-series data of the operating state detected by the operating state detection unit 12 . In one example, the time synchronization unit 13 obtains a difference between reference times of the sonic vibration data and the operational data, corrects the time of the sonic vibration data or the operational data using the difference, and extracts data for a time interval in which the sonic vibration data and the operational data are collected together, thereby obtaining the sonic vibration data and collecting operational data that is synchronized.

An example of the operational data is time-series data for current, generated torque, position, or speed. The operational data may be information about the propulsion device or an estimated value based on certain detectable data. A motor drive device may store a position command, a speed command, a torque command, a current command, and a voltage command of the motor, and the drive noise diagnostic system 10 may acquire them from the motor drive device and use them as operation data as they are, or it may convert the operation data by a comparison calculation with detectable data or an estimation calculation including filter processing.

In addition, the operational data is not limited to an analog signal. An example of run data is binary data that turns on when the speed reaches the speed command. In addition, the operational data need not be a set of time and value. An example of operational data is a function that takes as an argument a time representing a speed pattern at each point in time.

The operation mode extraction unit 14 divides the operation state into two or more periods based on the operation data detected by the operation state detection unit 12 . Specifically, the operation mode extraction unit 14 identifies the type of operation state of the actuator through analysis of the operation data, and extracts a period in which the operation state of the actuator is of a specific type as the operation mode. As for the operation mode, the operation state can be identified from the operation data after being synchronized with the sound vibration data by the time synchronization unit 13, or can be identified from the operation data before being synchronized with the sound vibration data by the time synchronization unit 13.

The acoustic vibration time-series spectrum acquiring unit 15 performs frequency transformation on the acoustic vibration data time-synchronized by the time synchronization unit 13 and acquires a time-series spectrum in which a spectrum of the acoustic vibration data corresponding to each time point is obtained. The time series spectrum is a set in which the power of the calculated frequency spectrum is related to frequency and time. When performing the frequency transformation, the acoustic vibration time-series spectrum acquisition unit 15 can select the timing at which the frequency transformation is performed by performing the frequency transformation only on a portion of a specific operation mode among the results extracted by the operation mode extraction unit 14. For example, the operating mode in which the actuator is stopped and drive noise is less likely to be measured may be excluded from the frequency transformation target. In this case, the amount of data to be subjected to the frequency transformation processing is reduced, so a reduction in working memory used at the time of calculation and a reduction in processing time can be expected.

When a waveform related to the frequency and time of power of the time-series spectrum obtained by the acoustic vibration time-series spectrum acquisition unit 15 satisfies a predetermined condition, the feature point extraction unit 16 extracts that point as a feature point. The feature point extraction unit 16 combines the frequency, time, and power of the feature point, the waveform of the feature point, and the operation data at the time corresponding to the feature point as feature point data.

The operational vibration extraction unit 17 extracts a vibration component from the operational data subjected to time synchronization, and acquires vibration data including the frequency or the amplitude of the vibration component.

The cause determining unit 18 determines a cause of engine noise by comparing a cause determination condition, which is a numerical range defined for each phenomenon occurring as a cause of engine noise and including a feature point occurring with the phenomenon, with the compares feature point data extracted by the feature point extraction unit 16 . Here, the cause determination condition may include a numerical range predetermined for each phenomenon occurring as a cause of a driving noise, including vibration data occurring with the phenomenon. In this case, the cause determination unit 18 estimates a cause of a driving noise by comparing the cause determination condition with the feature point data extracted by the feature point extraction unit 16 and the vibration data extracted by the operational vibration extraction unit 17 .

It should be noted that when the numerical range containing the feature point is given for each device abnormality, the numerical range containing the feature point for each cause of device abnormality must be obtained for all types of machines to be driven. Similarly, when the configuration of the machine to be driven is changed, the numerical range containing the feature point for each cause of an abnormality of the device must be found. Since there are numerous types of machines to be driven and There are numerous kinds of changes in the configuration of the machines to be driven, it is not realistic to obtain the numerical range containing the feature point for each cause of an abnormality of the device for all of them.

However, in the first embodiment, the numerical range including the feature point is defined not for each anomaly of the device but for each phenomenon appearing as a cause of driving noise. i.e. the feature value is not determined by a specific drive pattern such as e.g. B. defines the value of the speed of the motor, but it is expressed by a conditional expression independent of the drive pattern, such as. B. by a conditional expression that uses the speed of the motor as a variable. Therefore, even if another driven machine is used or the configuration of the driven machine is changed, the occurring phenomenon is the same if the cause of the driving noise is the same, so that regardless of the type or configuration of the driven machine, the same cause determination condition is used can be. As a result, compared to the case where the numerical range including the feature point is defined for each abnormality of the device, the time required for making the diagnosis can be shortened, and the diagnosis of the driving noise can be performed in a versatile manner.

The feature point data is the set of frequency, time and power of the feature point, the waveform of the feature point and the operation data at the time corresponding to the feature point. Therefore, the feature point data contains not only information about the noise or vibration but also information about the position or speed of the actuator or machine to be driven. i.e. in determining a cause of the driving noise, the cause determination unit 18 makes the determination by considering the position or the rotation speed of the actuator or the machine to be driven, so that the cause of an abnormality depending on the position or the rotation speed of the actuator or the machine to be driven , can be easily identified.

In addition, according to 1 the time synchronization unit 13, the operation mode extraction unit 14 and the operation vibration extraction unit 17 can be incorporated into the engine noise diagnosis system 10 or removed from the engine noise diagnosis system 10 as needed, depending on the performance or device configuration required for the diagnosis. In one example, when the acoustic vibration data and the operational data are acquired by the same measurement instrument or device, the acquired data can be synchronized using analog-to-digital (AD) converters that have the same acquisition clock. i.e. the sound vibration data and the operational data can be synchronized without the time synchronization unit 13 . In this case, by removing the time synchronization unit 13, a reduction in the working memory used by the engine noise diagnosis system 10 and a reduction in the processing time can be expected.

2 14 is a diagram showing an example of a hardware configuration when the drive noise diagnosis system according to the first embodiment is applied to a hoist. As in 2 shown, a diagnosis object 100 includes a motor 110, ie an actuator, a machine 120 to be driven and a drive device 130 which drives the motor 110. The diagnosis object 100 is further provided with a drive noise diagnosis system 10</b>A including the drive noise detection unit 11 , the operating state detection unit 12 , and an arithmetic processing unit 140 .

The motor 110 is a servo motor that controls the current of an armature winding based on a difference between a drive command and rotation information. Motor 110 need only be an actuator that generates power by receiving energy or an electrical signal from a device that drives motor 110 . In this example, the motor 110 is controlled by the driving device 130 . However, the state quantity of the motor 110 controlled by the driving device 130 is not limited to the current. Hydraulic pressure, pneumatic pressure, heat, and ultrasonic waves are examples of the state variables of the motor 110 that are controlled by the driving device 130 . In addition, the motor 110 is not limited to a motor that generates a rotational force, but may be a linear motor that causes a drive in a translational direction.

The machine to be driven 120 is a lifting device that conveys a driving object in a vertical direction according to the rotation of the motor 110 . The machine to be driven 120 includes a bracket 121 that secures the motor 110 to a frame, a gearbox 122 that amplifies the rotational force generated by the motor 110 and transmits the amplified rotational force to a ball screw 123, the ball screw 123 that rotates the motor 110 converts to a movement in the vertical direction, and a clutch 124 connecting the gear 122 and the ball screw 123.

The lifting device further includes a carriage 125 which is driven in the vertical direction by the rotation of the ball screw 123, a platform 126 which is fixed to the carriage 125 and on which a workpiece can be placed, a linear guide 127 which controls the movement of the carriage 125 slidably in the vertical direction, and a bracket 128 rotatably fixing the ball screw 123 to the linear guide 127 via a bearing.

Although the machine 120 to be driven is the lifting device including the ball screw 123 in this case, the machine 120 to be driven need only be a machine that generates noise or vibration by the rotation of the motor 110 . A screw, belt, gear, cam, linkage, bearing or seal, or a machine combining these elements are examples of the machine 120 to be driven.

The driving device 130 is connected to the motor 110 via a cable 151 . The drive device 130 includes a motor driver 131, a motor controller 132 and a display 133. The motor driver 131 provides power to the motor 110 to drive the motor 110. FIG. Motor driver 131 also drives motor 110 in accordance with a drive command transmitted from motor controller 132 .

The motor controller 132 controls the amount and timing of the current provided by the motor driver 131 to the motor 110 by an electrical signal such as. B. a commanded position or a commanded speed is sent to the motor drive 131. The display 133 displays various states of the entire system including the powertrain noise diagnosis system 10A and the diagnosis item 100, and notifies a user of the states.

In the first embodiment, the motor driver 131 is connected to the driving sound detection unit 11 and the operating state detection unit 12 and has a function of relaying communication between the motor controller 132 and each of the driving sound detection unit 11 and the operating state detection unit 12 .

In addition, the motor controller 132 is a controller that gives the motor driver 131 a drive command that includes a position or speed pattern of the motor 110 . Motor controller 132 is a controller that includes a programmable logic controller (PLC), a motor drive central processing unit (CPU), a digital signal processor (DSP), a pulse generator, or the like.

Further, the display 133 detects a state of the entire system including the engine noise diagnosis system 10A of at least one of the engine driver 131, the engine controller 132 and the arithmetic processing unit 140 through communication and displays the state in a form easy for a user to understand see is. Display 133 may include a liquid crystal display. By including the display 133 in the system, a result of diagnosing an engine noise can be displayed on the display 133 through communication after the diagnosis of an engine noise is made by the arithmetic processing unit 140, and a user can be informed of the result in an easy-to-understand manner .

The driving device 130 need only be a device that drives at least one of the motor 110, and can be formed by combining a part or a plurality of parts of the present embodiment.

The arithmetic processing unit 140 is a processing device capable of performing diagnostic processing of the powertrain noise diagnostic system 10A, which will be described later as software. In the first embodiment, a CPU of a microcomputer incorporated in the engine controller 132 implements the function of the arithmetic processing unit 140. 3 12 is a block diagram showing an example of a hardware configuration of the arithmetic processing unit according to the first embodiment. The arithmetic processing unit 140 includes a processor 141 and a main memory 142. The processor 141 and the main memory 142 are connected via a bus line 143. FIG. A CPU and graphics processing units (GPU) are examples of the processor 141.

4 FIG. 12 is a block diagram schematically showing an example of a functional configuration of the arithmetic processing unit of FIG 2 according to the first embodiment. The arithmetic processing unit 140 includes the time synchronization unit 13, the operation mode extraction unit 14, the acoustic vibration time-series spectrum acquisition unit 15, the feature point extraction unit 16, the operation vibration extraction unit 17 and the cause determination unit 18 as software executed by the CPU of the microcomputer. It should be noted that the same components as those of 1 and 2 are denoted by the same reference numerals as those assigned to these components ten in 1 and 2 were assigned, and therefore their description is omitted.

Even though 2 shows the configuration in which the arithmetic processing unit 140 is integrated with the motor controller 132, the embodiment is not limited to this configuration. The arithmetic processing unit 140 may be integrated with the motor driver 131 or the display 133 which is another device in the driving device 130 . Alternatively, another microcomputer may be attached to engine controller 132 to implement arithmetic processing unit 140 . As a further alternative, the arithmetic processing unit 140 may be connected as another device independent of the driving device 130 .

The driving noise detection unit 11 is a microphone that detects driving noise generated from the diagnosis object 100 . The detected driving noise is transmitted to the arithmetic processing unit 140 via the cable 152 and the driving device 130 . In the example of 2 the driving sound detection unit 11 is fixed to the hoist. However, the driving noise detection unit 11 only needs to be able to detect the driving noise of the diagnosis object 100 , and a method for installing the driving noise detection unit 11 is not limited to fixing it to the machine 120 to be driven. The driving noise detection unit 11 may be fixed to the engine 110 or installed at a distance from the machine 120 to be driven where the driving noise can be detected.

When the space where the noise of a phenomenon to be diagnosed is generated is limited, the area subjected to noise detection or the area where noise can be detected can be limited by e.g. B. placing the microphone adjacent to the place where the noise is generated or by using a directional microphone. Limiting noise detection can reduce ambient noise affecting diagnosis and reduce misdiagnosis.

Furthermore, a variety of microphones can be used as the driving sound detection unit 11 . Capturing the same sounds or vibrations with multiple microphones can reduce misdiagnosis due to noise.

In addition, a recording device can be used to record the driving noise generated by the diagnosis object 100 in a time series as sound vibration data. A smartphone or a voice recorder is an example of the recording device. In addition, a moving image can be picked up by a camera or the like, and it is possible to take out only the noise part thereof as sound vibration data.

In addition, the driving noise detection unit 11 may be provided with a storage unit that stores the driving noise generated in a time series as sound vibration data and transmits the stored sound vibration data to an upper computing device as needed. In this case, the driving noise detection unit 11 and the arithmetic processing unit 140 do not communicate with each other when the driving noise diagnosis is not performed, and the time-series acoustic vibration data stored by the driving noise detection unit 11 is transmitted collectively when the driving noise diagnosis is performed. Such a configuration can reduce processing related to communication between the driving noise detection unit 11 and the arithmetic processing unit 140 .

The operation state detection unit 12 is an encoder that is attached to the motor 110 and detects the rotation angle of the motor 110 . The data of the rotation angle detected by the encoder is transmitted to the arithmetic processing unit 140 via the cable 153 and the driving device 130 as an operating state.

The installation location of the operation state detection unit 12 is not limited to the case where it is fixed to the engine 110 . The operating state detection unit 12 can be connected to a drive unit of the machine 120 to be driven, i. H. of the machine driven by engine 110. Like the driving noise detection unit 11, the operating state detection unit 12 may be provided with a storage unit that stores a detected time-series operating state as operating data and transmits the stored operating data to a higher-level computing device as needed.

The arithmetic processing unit 140 receives the driving noise detected by the driving noise detection unit 11 and the operating state detected by the operating state detecting unit 12 via the cables 152 and 153 and diagnoses a cause of the driving noise.

Since the arithmetic processing unit 140 a corresponding data exchange with the drive noise detection unit 11 and the operating state detection unit 12 performs, it is desirable to have an environment in which the arithmetic processing unit 140 is connected to the driving sound detection unit 11 and the operating state detection unit 12 in a high-speed network without data delay.

Next, the operation of the diagnosis object 100 will be described. A user of the hoisting machine 120 to be driven inputs a command into the motor controller 132 to drive the motor 110 to convey a workpiece (not shown) through the hoisting machine. The motor controller 132 transmits a drive command to the motor driver 131 based on information including a drive pattern and an operation schedule inputted by the user. The motor driver 131 controls a motor driving current according to the received driving command and drives the motor 110 . When the motor 110 rotates as a power source in the elevator, the ball screw 123 rotates, and the carriage 125 connected to the ball screw 123 and the platform 126 connected to the carriage 125 move up and down.

In one example, when the position of the platform 126 is moved to a vertically lower part of the linear guide 127 by the operation described above, the operator of the elevator places the workpiece on the platform 126 . The platform 126 moves to a vertically higher part of the linear guide 127 as the motor 110 rotates. With this movement, the elevator conveys the workpiece placed on the platform 126 .

The diagnosis object 100 generates driving sound when driven by the motor 110 . In particular, the diagnostic object 100 generates driving noise due to the torque ripple of the motor 110, the translational and torsional rigidity of the ball screw 123, the connection rigidity of the coupling 124, the meshing rigidity, machine movement, deformation or collision, the sliding friction between the linear guide 127 and the slider 125 and the same. The driving noise generated by the diagnosis object 100 is detected by the driving noise detection unit 11, and the rotation angle of the motor 110 is detected by the operating state detection unit 12, so that the detected driving noise and the detected rotation angle are transmitted via the cables 152 and 153 to the motor driver 131, of the engine controller 132 and the arithmetic processing unit 140 are transmitted. However, the transmission method is not necessarily wired, but may be wireless or via a recording medium.

The display 133 communicates with the engine controller 132 accordingly, and displays on the display 133 various kinds of information that the user needs, including the rotation angle of the engine 110 detected by the operating state detection unit 12 and the presence or absence of an abnormality in the entire system , including diagnostic object 100.

In the first embodiment, the time synchronization unit 13, the operation mode extraction unit 14, the acoustic vibration time-series spectrum acquisition unit 15, the feature point extraction unit 16, the operation vibration extraction unit 17, and the cause determination unit 18 are configured in the arithmetic processing unit 140, which is a single piece of hardware, but they can also be divided into separate pieces of hardware. For example, the time synchronization unit 13 may be separate from the arithmetic processing unit 140 and configured as software in a microcomputer of the motor driver 131 . In this case, the motor driver 131, which executes the control processing at a higher speed than the motor controller 132, performs time synchronization processing with strict timing, and the arithmetic processing unit 140 in the motor controller 132 performs the processing of the operation mode extraction unit 14, the sound vibration time-series spectrum -acquisition unit 15, the feature point extraction unit 16, the operational vibration extraction unit 17 and the cause determination unit 18 with relatively less time constraint. This can reduce the power required of the arithmetic processing unit 140, and even a device with less arithmetic processing capacity can perform the diagnosis of a cause of the driving noise.

In addition, the implementation of each function is not limited to implementation in the form of software in the CPU of the microcomputer, but an electronic circuit such as an electronic circuit may be used. B. an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or a complex programmable logic device (CPLD) can be used.

Next, a driving noise diagnosing method in the driving noise diagnosing system 10A for noise or vibration generated by the diagnosing object 100 will be described. 5 14 is a flowchart showing an example of a processing operation of a driving noise diagnosing method according to the first embodiment. As described above, in the first embodiment, when the machine 120 to be driven is driven by the motor 110 is driven, driving noise is generated due to movement, deformation, collision, rigidity, friction or the like of the machine. In addition, depending on the method of driving the motor 110, the driving noise has a different sound pressure and frequency. Therefore, in the first embodiment, a method for diagnosing the driving noise generated by the machine 120 to be driven will be described.

The driving noise detection unit 11 of the driving noise diagnostic system 10A detects the driving noise (step S11), and the operating state detecting unit 12 detects an operating state of the engine 110 (step S12).

The time synchronization unit 13 imports the driving noise from the driving noise detection unit 11 as time-series acoustic vibration data (step S13). 6 14 is a graph showing an example of the sound vibration data according to the first embodiment. In this graph, the horizontal axis represents time and the vertical axis represents amplitude.

The time synchronization unit 13 also imports the operation state from the operation state detection unit 12 as time-series operation data (step S14). 7 FIG. 14 is a diagram showing an example of operational data according to the first embodiment. In this graph, the horizontal axis represents time, and the vertical axis represents rotation angle, speed, and current. 7 12 shows, as operating data, a rotation angle "Po" of the motor 110, a rotational speed "w" of the motor 110, and a motor current "i" from a reference position of the ball screw 123.

In this case, data for the same time interval must be imported to collect the sound vibration data and to collect the operational data. However, as long as the data for the same time interval is imported, the time of collection of the data to be imported may be different, or the time of collection of a part of the data may be different from this time interval.

In addition, when the frequency band of the noise or vibration generated by the diagnosis object 100 is predicted by some knowledge, it is desirable to reduce the number of samples of the sound vibration data and the operational data up to the target frequency band. Reduction processing can reduce the amount of memory used and the processing time. When the reduction processing is performed, the frequency transformation of the sound vibration data is performed in the processing after the processing of the driving noise detection unit 11 and the operating state detection unit 12, so it is desirable to perform filtering using a noise removal filter. A low-pass filter, a high-pass filter, a band-pass filter, a notch filter, and a band elimination filter are examples of the noise removal filter. In addition, one or a plurality of the filters shown can be used as noise removal filters. Thereby, an effect of reducing aliasing noise due to the reduction can be expected.

Next, the time synchronization unit 13 performs synchronization processing on the sound vibration data and the operational data (step S15). 8th 14 is a flowchart showing an example of a procedure of synchronization processing for the acoustic vibration data and the operational data according to the first embodiment.

First, a difference between reference times of the sound vibration data and the operational data is determined (step S31). Here, the reference time is a time set to time zero in all data. When the sound vibration data and the operational data are not synchronized with each other, the timing at which the time is reset to zero in all the data is adjusted individually, so even if the data collection times are the same, the actual collection times may be different. Therefore, the time of all data is corrected by finding the difference between the actual times of the schedule used to zero the time in all data.

As a method for finding the difference between the reference times, there is a method in which the data collection schedules are set to a constant time difference in advance and the constant time difference is used as the difference between the reference times. There is also a method in which a common master clock is accessed in advance to set the same acquisition timing for all data and zero the difference between the reference times. Alternatively, there is a method that stores the time of a common master clock at the time of starting data collection as a time stamp, and uses a difference between the time stamps of the acoustic vibration data and the operational data as the difference between the reference times.

9 13 is a diagram for explaining a method of synchronizing the sound vibration data and the operational data according to the first embodiment. In the example of 9 it is assumed that the zero time of the sound vibration data “Da” is the data collection time and the zero time of the operational data “Do” is the data collection time. It is also assumed that the zero time point of the sound vibration data “Da” is in the reference time “t1” and the zero time point of the operational data “Do” is in the reference time “t2”. In this case, a difference Δt between the reference times is “t2-t1”.

Next, the acquisition time of each data is corrected based on the detected difference between the reference times (step S32). Specifically, the detected difference between the reference times is added as a correction to the collection time of all data, namely those among the sound vibration data and the operational data, which collection was first started.

In the example of 9 the data that was first started to be collected is the sound vibration data “Da”. Therefore, the reference time "t2" of the operation data is obtained by adding the difference Δt between the reference times to the detection time "t1" of the sound vibration data "Da".

up again 8th Referring to this, a time interval in which the sound vibration data and the operational data are collected together is calculated (step S33). An example of a specific calculation method is a method in which the time from the acquisition start time of the data whose acquisition is started later in the processing described above to the acquisition time which is an earlier point in time between the latest acquisition times of the sound vibration data and of the operational data is defined as the time interval in which the sound vibration data and the operational data are collected together. In the example of 9 the period from the collection start time "t2" of the operational data "Do" to the last collection time "t3" of the sound vibration data "Da" is a time interval Δtc in which the sound vibration data and the operational data are collected together.

Subsequently, the data of a point in time that is not included in the calculated time interval in which the acoustic vibration data and the operational data are collected together and which is not common to the acoustic vibration data and the operational data is discarded (step S34). In the example of 9 the sound vibration data "Da" before time "t2" and the operating data "Do" after time "t3" are discarded. In this way, the sound vibration data and the operational data can be synchronized with each other, and the synchronization processing with the sound vibration data and the operational data ends.

up again 5 Referring to this, the operation mode extraction unit 14 estimates an operation state of the machine 120 to be driven based on the operation data subjected to the time synchronization and extracts an operation mode (step S16). In the first embodiment, the operation mode of the machine to be driven 120 is divided into the three modes stopped, constant speed operation, and acceleration/deceleration.

<Stopped>

A portion where the rotation of the motor 110 is stopped is defined as stopped. Here, the rotation of the motor 110 is stopped means that a sum of the changes in the data of the rotation angle of the motor 110 detected by the operating state detection unit 12 is within a specific threshold value during a certain period of time. At this time, a portion where the sum of changes in the data of the rotation angle is within the specific threshold is defined as a portion where the position is stopped. The specific period of time included in the condition may be defined according to an average value of detection cycles of the sound vibration data and the operational data.

<Constant speed operation>

A portion in which the motor 110 rotates at a constant speed and is not stopped is defined as being in constant speed operation. Here, the speed being constant means that a sum of the changes in the speed data over a certain period of time is within a specific threshold value, as is the case with the stopped state. At this time, a period in which the sum of changes in the rotational speed data is within the specific threshold value is defined as a period longer than a predetermined period in which the rotational speed is constant.

<Acceleration/Deceleration>

A section that is neither the stopped state nor the constant speed operation is defined as acceleration/deceleration. Acceleration/deceleration can be distinguished such that acceleration that is positive with respect to the direction of travel is defined as acceleration, and acceleration that is negative with respect to the direction of travel is defined as defined as deceleration. 10 is a diagram showing an example in which the operation data of 7 divided into operating modes. In this example, the operation modes include acceleration/deceleration from time zero to time t11, constant speed operation from time t11 to time t12, acceleration/deceleration from time t12 to time t13, and the stopped state after time t13.

up again 5 Referring to this, the acoustic vibration time-series spectrum acquiring unit 15 performs frequency transformation on the acoustic vibration data subjected to the time synchronization by the time synchronization unit 13 and calculates a spectrum of the acoustic vibration data for each time point (step S17). The spectrum of the sound vibration data is time-series spectral data in which the power of the calculated frequency spectrum is related to frequency and time as a set.

The frequency transform is desirably done using Short Time Fourier Transform (STFT) or Wavelet Transform. The time-series spectrum is obtained by these methods, so labor and processing for filter design or the like can be reduced.

Here, when the frequency band of the driving noise generated by the diagnosis object 100 is predicted by some knowledge, a filter that extracts a frequency component in the predicted band can be applied to the sound vibration data before frequency transformation. The application of the filter enables a more precise diagnosis of the drive noise.

In the first embodiment, the acoustic vibration time-series spectrum acquisition unit 15 performs the frequency transformation only on the acoustic vibration data corresponding to the period of constant speed operation extracted by the operation mode extraction unit 14 from the acoustic vibration data subjected to the time synchronization. and discards data from the other periods.

11 Fig. 13 is a three-dimensional graph with three axes of time, frequency and power showing an example of the time-series spectral data obtained by frequency transforming the acoustic vibration data of Figs 6 were received. In this graph, the vertical axis represents power, and the two orthogonal axes in a plane perpendicular to the vertical axis represent time and frequency.

Now will up again 5 Referring to where, for the spectrum of the acoustic vibration data represented as time-series data acquired by the acoustic vibration time-series spectrum acquisition unit 15, the feature point extraction unit 16 extracts a feature point at which the frequency and time of the power of the spectrum have a predetermined Meet the condition (step S18). The feature point extraction unit 16 further generates feature point data for the extracted feature point (step S19). The feature point data is a set of the frequency, time, power and waveform of the feature point and the operation data at the time corresponding to the feature point.

Examples of the condition used when the feature point extraction unit 16 extracts the feature point include a peak, a top, a ridge, and a saddle point. A case where a peak is detected as a feature point will be described.

Since the spectrum of the time-series acoustic vibration data includes noise, a low-pass filter is first applied to the time axis and the frequency axis. A time constant of the low-pass filter is determined by the sampling period of the acoustic vibration data and the resolution of the frequency transform. Instead of applying the low pass filter, a Hilbert transform can be applied to the time series spectrum.

Next, a power threshold is determined for the time-series spectrum after the low-pass filter is applied, and the time-series spectrum is divided into an area with higher power than the threshold and an area with lower power than the threshold. The median of the time series spectrum is an example of the performance threshold.

1. (1) Punkt x1 (t_x+1, f_x, p₁) mit der Frequenz „f_x“ und der Erfassungszeit „t_x+1“ nach „tx“
2. (2) Punkt x2 (t_x-1, f_x, p₂) mit der Frequenz „f_x“ und der Erfassungszeit „t_x-1“ unmittelbar vor „tx“
3. (3) Punkt x3 (tx, f_x+1, p₃) mit der Erfassungszeit „tx“ und der Frequenz „f_x+1“ nach „f_x“
4. (4) Punkt x4 (t_x, f_x-1, p₄) mit der Erfassungszeit „tx“ und der Frequenz „f_x-1“ unmittelbar vor „f_x“

After that, the following four points are found from the points x=(t _x , f _x , p _x ) where the power is higher than the threshold. Here “t _x ”, “f _x ” and “p _x ” represent the time, frequency and power at point x, respectively.

1. (1) Point x1 (t _x+1 , f _x , p ₁ ) with frequency "f _x " and acquisition time "t _x+1 " after "tx"
2. (2) Point x2 (t _x-1 , f _x , p ₂ ) with frequency "f _x " and acquisition time "t _x-1 " immediately before "tx"
3. (3) Point x3 (tx, f _x+1 , p ₃ ) with acquisition time "tx" and frequency "f _x+1 " after "f _x "
4. (4) Point x4 (t _x , f _x-1 , p ₄ ) with acquisition time “tx” and frequency “f _x-1 ” just before “f _x ”

Next, the performance of a total of five points including the four points x1, x2, x3 and x4 and the origin point x is compared. If the performance of point x is higher than that of the other four points among the five points, i.e. H. if the point x surrounded by the other four points x1, x2, x3, and x4 among the five points has the highest power, the point x is determined as a candidate for the peak. Note that this example illustrates the case where the point with the highest power is extracted, taking the point x and the two points adjacent thereto in the time axis direction and the two points adjacent thereto in the frequency axis direction with the point x as center can be used, but the embodiment is not limited thereto. i.e. any method can be used as long as the method uses the point x and a plurality of points surrounding the point x in a plane formed by the time axis and the frequency axis and extracts the point x as a maximum power point, if the point x surrounded by the plurality of points has the highest power.

Finally, the peak candidates are ranked in descending order of power, and a certain number of points from the highest are extracted as maxima.

Here, the candidate points are arranged in descending order of power, and a certain number of points detected from the highest one can be set as peaks, or all candidate points whose power exceeds a predetermined threshold can be set as peaks. These methods can also be combined.

By detecting the peak of the waveform with respect to the frequency and time of the power of the time-series spectrum as a feature point, the amount of calculation in the cause diagnosis described later can be reduced in a low-processing method.

If the operational data is not collected at the time corresponding to the feature point, the operational data at the time corresponding to the feature point can be determined by using the operational data at the time closest to the time point of the feature point, or by interpolating the operational data at the time corresponding to the feature point based on the operational data at a plurality of times before and after the time point of the feature point.

12 Fig. 13 is a three-dimensional plot with three axes for time, frequency and power, showing an example of one from the time-series spectral data of 11 extracted peak of the noise. This graph shows peaks 1, 2 and 3 extracted by the method described above.

up again 5 Referring to this, the operational vibration extracting unit 17 extracts a vibration component from the operational data subjected to the time synchronization and acquires vibration data (step S20). The frequency, amplitude or phase of the vibration component is an example of the vibration data. This processing can be performed independently of the processing for acquiring the acoustic vibration time-series spectrum in step S17.

An example of a method for extracting the vibration component is a method that obtains a peak-to-peak time of the waveform of the operating data during the constant speed operation.

Next, the cause determination unit 18 compares the feature point data generated in step S19 and the vibration data acquired in step S20 with the cause determination condition registered for each cause of the driving noise, thereby determining a cause of the driving noise (step S21). Here, as for the feature point data, each cause determination condition registered for each cause of the driving noise defines a numerical range for each phenomenon occurring in the motor 110 or the machine 120 to be driven as a cause of the driving noise when at least one of the frequency and time, which includes the feature point occurring with this phenomenon, and the operational data, i. H. the drive position, the drive speed or the speed caused by the drive of the motor 110, i. H. of the actuator, generated force at that time, can be combined into multidimensional data. As for the vibration data, each cause determination condition for each phenomenon occurring in the engine 110 or the machine 120 to be driven as a cause of the driving noise defines a numerical range including the vibration component generated with that phenomenon.

As for the feature point data, the cause determination unit 18 determines the cause of the sound vibration data by compares the collected feature point data to the numeric range for each cause. As for the vibration data, the cause determination unit 18 determines the cause of the vibration data by comparing the detected vibration data with the numerical range for each cause, that is, by comparing the vibration component of the operation data with the peak of the noise.

For example, when dirt adheres to the linear guide 127 and causes friction between the carriage 125 and the linear guide 127 to increase and friction noise to be generated, the friction noise is generated only when the carriage 125 is within a specific positional range. When the position of the motor 110 is concentrated at a specific position or within a section P of a certain area at the time the feature point occurs, the cause determination unit 18 determines that the diagnosis object 100 generates noise in the P section. Here, the area of the section P can be determined by a ratio to a difference between the maximum value and the minimum value of position data in the time-series operation data. In addition, the cause determination condition can be any condition that makes it possible to check whether the distribution of the position of the motor 110 at the time corresponding to the feature point is concentrated on a specific position. For example, the number or the ratio of the number of feature points outside the section P having a certain range, the average and the variance of the position at the time corresponding to the feature point, or the correlation coefficient between the frequency and the position of the feature point can be calculated to check whether the calculated value is within a predetermined range.

As another example, it is known that when the centers of two axles connected by the clutch 124 are misaligned, a characteristic noise is generated at a frequency that is twice the frequency of the number of revolutions of the clutch 124 . Thus, the speed of the clutch 124 is calculated by multiplying a gear ratio of the transmission 122 by the speed of the engine 110 in the operation data at the time the feature point occurs, and the cause determining unit 18 determines that the midpoints of the two are divided by the Axes connected by coupling 124 are misaligned when twice the frequency of the calculated speed and the frequency of the feature point are directly proportional to each other. At this time, instead of the rotation speed of the motor 110, the current of the motor 110 may be detected as operation data, and the value may be accumulated to replace the rotation speed.

As still another example of the condition, when the machine has a mechanical resonance by driving the motor 110, a resonance noise having a specific resonance frequency “f” is generated at a specific speed “v” at which the mechanical resonance is excited. If the speed of the engine 110 at the time the feature point occurs is concentrated at a specific speed or section "V" with a certain range and the frequency of the feature point is concentrated at a specific frequency or section "F" with a certain Therefore, the cause determination unit 18 determines that the noise of the frequency “f” is generated in the diagnosis object 100 due to the mechanical resonance at the rotation speed “v”.

Further, when the frequency of vibrations detected by the operational vibration extraction unit 17 and the feature point corresponding to the frequency “f” due to the mechanical resonance have the same time cycle, the cause determination unit 18 determines that the vibration component of the engine 110 is the mechanical Resonance at regular intervals excited by the rotation of the motor 110. i.e. by comparing the vibration component of the operating data with the peak of the noise, the cause determining unit 18 determines a phenomenon in which the mechanical resonance is periodically excited by the vibration component of the engine and the resonance is excited.

As described above, the cause determination unit 18 uses the registered cause identification condition to check the feature point data extracted by the feature point extraction unit 16, or the feature point data and the vibration data extracted by the operational vibration extraction unit 17, thereby determining the cause of the actuator position or rotation speed or the drive noise dependent on the machine 120 to be driven is identified. As for the registered cause identification condition, similar conditions can be organized as a binary search tree in advance to perform the check. This can reduce the number of conditions to be checked in root cause identification and shorten the time for identification. This completes the processing of the engine noise diagnostic process.

The drive noise diagnosis systems 10 and 10A according to the first embodiment syn synchronize the time-series acoustic vibration data for the driving noise generated by the actuator or the machine to be driven 120 with time-series operational data for the operational state of the actuator and extract the operational mode. In addition, the feature point is extracted from the spectrum of the time-series sound vibration data obtained by time-frequency analysis of the sound vibration data, and the feature point data is generated in which the frequency, time, power and waveform of the feature point and the Operational data at the time corresponding to the feature point are combined as a set. Then, the cause of the engine noise is identified by comparing the feature point data with the cause identification condition prepared in advance. Since the cause of the driving noise is identified based on the sound vibration data and the operation data as described above, it is possible to identify the place where the driving noise is generated due to the attachment of foreign matter or the like to the machine 120 to be driven, and it is possible to easily determine the cause of an anomaly that depends on the position or the speed of the actuator or the machine 120 to be driven. It is also possible to identify the cause of a change in the frequency of the noise generated in accordance with the rotational speed of the actuator or the machine 120 to be driven. Furthermore, even in a case where the entire device vibrates, it can be determined whether the noise or vibration is due to the operation.

In addition, the drive noise diagnosis systems 10 and 10A diagnose the cause of the noise or vibration based on the sound vibration data of the feature point, the operation data and the time of their occurrence. Therefore, even if there is a device that generates a loud operation noise near the diagnosis object, it is possible to prevent or reduce misdiagnosis due to the causal relationship between the occurrence time and the operation data.

Furthermore, since the drive noise diagnosis systems 10 and 10A diagnose the cause of the noise by combining the frequency spectrum of the sound vibration data and the operation data, it is possible to eliminate the influence of a change in the spectrum of the sound vibration data due to a change in the operation pattern when the change of the spectrum of the sound vibration data is determined. In particular, even when the operation pattern changes due to the state of the workpiece, the influence on the spectrum of the acoustic vibration data can be eliminated by the operation data, so that the driving noise diagnosis systems 10 and 10A can perform appropriate diagnosis.

In addition, the drive noise diagnosis systems 10 and 10A diagnose the cause of the drive noise by substituting the feature value obtained through the time-frequency analysis of the sound vibration data into the conditional expression independent of the drive pattern. Thus, the operational data in a normal or abnormal state need not be prepared in advance. Therefore, even if the configuration of the driving device or the engine is changed, the diagnosis of the driving noise can be performed in a versatile and straightforward manner.

$$\begin{array}{l}\text{Spitzenfrequenz nach der Ver}\ddot{\text{a}}\text{nderung}=20\times \\ \ddot{\text{U}}\text{bersetzungsverh}\ddot{\text{a}}\text{ltnis des Getriebs 122}\times \text{2}\pm \text{Fehler}\\ \left[\text{Hz}\right]\end{array}$$

A case where the centers of two axes connected by the coupling 124 are misaligned will be described as an example. It is known that a characteristic noise is generated at the frequency twice the frequency of the clutch 124 rotational speed. A change in the mode of operation of the machine is believed to have caused a change in drive pattern from a motor 110 speed of 10 rpm to a motor 110 speed of 20 rpm. In the technique described in Patent Literature 1, a threshold of the peak frequency of the noise in the case of detecting an abnormality needs to be set for each drive pattern. An example of the peak frequency threshold value is expressed by the following expression (1) before the drive pattern is changed, and is expressed by the following expression (2) after the drive pattern is changed. $$\begin{array}{l}\text{peak frequency before the ver}\ddot{\text{a}}\text{change}=10\times \\ \ddot{\text{u}}\text{translation ratio}\ddot{\text{a}}\text{122 gear ratio}\times \text{2}\pm \text{error}\\ \left[\text{Hz}\right]\end{array}$$

$$\begin{array}{l}\text{Peak Frequency after Ver}\ddot{\text{a}}\text{change}=20\times \\ \ddot{\text{u}}\text{translation ratio}\ddot{\text{a}}\text{122 gear ratio}\times \text{2}\pm \text{error}\\ \left[\text{Hz}\right]\end{array}$$

On the other hand, in the present embodiment, the peak frequency, ie, the characteristic value of the noise generated by the misalignment of the coupling 124 is expressed by the following expression (3). $$\begin{array}{l}\text{peak frequency}=\text{frequency of rotation}\times \\ \ddot{\text{u}}\text{translation ratio}\ddot{\text{a}}\text{122 gear ratio}\times \text{2}\pm \text{error}\\ \left[\text{Hz}\right]\end{array}$$

i.e. the peak frequency of the noise generated by the misalignment of the coupling 124, including the expressions (1) and (2), is expressed by a conditional expression using the speed of the engine 110 as a variable, and it is not necessary to use the conditional expression for each drive pattern, in this case for each motor 110 speed. Since the frequency of the rotation speed can be determined from the detected operation data, misalignment of the clutch 124 in any operation pattern can be determined by using expression (3). i.e. the frequency, which is the feature value of the noise, is substituted into the drive pattern independent conditional expression for identification. It should be noted that although the noise produced when the centers of the two axles connected by the coupling 124 are misaligned has been described here, the cause of the drive noise can also be diagnosed for other causes by examining the time Frequency analysis of the vibration data is substituted into the conditional expression independent of the drive pattern.

In addition, the cause is diagnosed based on the driving noise generated by the machine 120 to be driven. By using the engine noise, the diagnosis can be performed on any type of engine even if the engine rigidity is low or the like.

In addition, the drive noise diagnosis systems 10 and 10A diagnose the cause of the noise or vibration using a result of time-frequency analysis of the sound vibration data. Performing time-frequency analysis can reduce false detection by the sensor or misdiagnosis of the cause due to noise.

Further, the engine noise diagnosis systems 10 and 10A diagnose the cause of the engine noise based on the feature value obtained through the time-frequency analysis of the sound vibration data. This can reduce the data to be diagnosed and the processing time. The cause of the engine noise can also be estimated using the frequency.

In addition, in the drive noise diagnosis systems 10 and 10A, the operation mode extraction unit 14 estimates the operation state of the device based on the operation data and extracts the operation mode. Then, the frequency transformation is performed only on the sonic vibration data corresponding to the period during the constant speed operation, so the sonic vibration data can be discarded during acceleration/deceleration where driving of the actuator is difficult to stabilize. Therefore, compared to a case where the acoustic vibration data is used during acceleration/deceleration, the accuracy of the diagnosis is improved, and the amount of work memory used in the arithmetic processing and the processing time can be reduced by using the acoustic vibration data, unsuitable for diagnosis are discarded.

Further, the driving noise diagnosis systems 10 and 10A determine, based on the operation data, the period when the machine to be driven 120 is not operated or the like, and omits the sound vibration data in the period when the machine to be driven is not operated, whereby a false detection due to non-drive noise can be prevented. In addition, the processing cost can be reduced by omitting the frequency transformation in the state where the machine 120 to be driven is not operated.

In addition, when examining the cause of the engine noise, the operation mode including the engine noise to be examined is specified so that the influence of another phenomenon on the engine noise can be eliminated. In particular, in a case where a plurality of actuators are provided, the actuator that causes the driving noise can be estimated.

In addition, in the drive noise diagnosis systems 10 and 10A, the operation mode extraction unit 14 determines that the constant speed operation is performed when the sum of changes in the speed data as the operation data during a certain period of time is within a specific threshold value. Thereby, in a state where the rotational speed of the actuator is within a certain range, the driving noise generated by the machine to be driven 120 is stable and tends to be in a homogeneous state, so that the cause of the noise can be determined more accurately. Since the noise can be extracted in the steady state without specifying the rotation speed moreover, the same root cause diagnosis can be performed as long as the constant speed operation is in progress, even if the driving patterns are different. Further, the noise can be extracted when it is stable at multiple rotation speeds of the actuator or the machine 120 to be driven, which is desirable for cause estimation.

Furthermore, since the drive noise diagnosis systems 10 and 10A include the operational vibration extraction unit 17 that extracts the vibration component from the operational data, the cause determination unit 18 can compare the vibration component with the occurrence timing of the feature point. Thereby, it is possible to detect the phenomenon in which the mechanical resonance is periodically excited by the vibration component of the engine and the resonance is excited.

In addition, the drive noise diagnosis systems 10 and 10A detect that the distribution of the feature point and the position of the engine 110 at the time corresponding to the feature point is concentrated at a specific position. This makes it possible to determine whether the detected feature point is a feature point that was generated by a cause dependent on the operating data or not.

In addition, the powertrain noise diagnosis systems 10 and 10A extract, as a feature point, a point where the power of the time-series spectrum corresponds to the peak of the waveform in terms of frequency and time. The extraction reduces the number of items to be diagnosed, so the processing performed by the cause determination unit 18 can be reduced. By calculating the correlation coefficient between the peak of the spectrum and the operational data, it can be determined whether or not the detected peak is due to a cause depending on the operational data.

In addition, the driving noise diagnosis systems 10 and 10</b>A determine the cause of the driving noise generated by the machine 120 to be driven, and provide notification to a user of the hoist via the display 133 . Thereby, the hoist operator can determine that there is a problem with the driving noise of the driven machine 120 and take an appropriate measure based on the diagnosis result.

Second embodiment.

13 14 is a diagram showing an example of a hardware configuration when a drive noise diagnosis system according to a second embodiment is applied to an extraction unit. In this example, a diagnosis object 200 is an extraction unit that conveys a workpiece 291 transported on a belt conveyor 225 to another belt conveyor 226 .

As in 13 1, the diagnosis object 200 includes a plurality of motors 211, 212, 213, and 214 and actuators 215, a machine to be driven 220, and a driving device 230. A driving noise diagnosis system 10B is provided in the diagnosis object 200. FIG. The drive noise diagnosis system 10B includes the drive noise detection unit 11, the operating state detection unit 12, a wireless network device 240, a server device 250, a user terminal 260 and a network device 270. The server device 250 and the user terminal 260 are connected via a communication line 280. Note that in the machine 220 to be driven, a vertical direction is defined as a Z direction, an extending direction of the belt conveyors 225 and 226 in a plane perpendicular to the Z direction is defined as a Y direction, and a direction perpendicular to the Y direction and to the Z-direction is defined as the X-direction.

The machine 220 to be driven is the extraction unit. The machine to be driven 220 includes a linear rail 221, a ball screw 222, a linear guide 223, a head 224 and the two belt conveyors 225 and 226.

The linear rail 221 is a rail that fixes a driving direction of the motor 212, which is a linear motor. In this example, the linear rail 221 is arranged so as to extend in the X-direction over the two belt conveyors 225 and 226 arranged in parallel. The head 224 is connected to the linear rail 221 through the motor 212 , the ball screw 222 and the linear guide 223 . A direction in which the motor 212 can be moved when the motor 212 is driven is limited to the extending direction of the linear rail 221 . In the example of 13 the direction in which the motor 212 can be moved is limited to the X direction.

The ball screw 222 is connected to the motor 211 and drives the head 224 in the Z direction by the rotation of the motor 211 . The linear guide 223 is a guide that guides a driving direction of the ball screw 222 restricted to the Z-direction. The linear guide 223 is fixed to the motor 211 and the ball screw 222 while being fixed to a movable portion of the motor 212 . Thereby, the ball screw 222 is driven along the linear rail 221 by driving the motor 212 .

The head 224 is a platform driven by driving the ball screw 222 in the Z direction. The head 224 has a shape extending in the Y-direction and includes the actuators 215 including a mechanism for holding the workpiece 291 in a lower portion. The actuators 215 are vacuum pads that hold the workpiece 291 by a vacuum suction mechanism. The workpiece 291 held by the actuators 215 can be moved up and down by the ball screw 222 . The workpiece 291 is an unloading target of the unloading unit.

The belt conveyor 225 is a feeder that transports the workpiece 291 in the Y direction from a positive side to a negative side. The motor 213 is connected to the belt conveyor 225 . When a built-in belt is driven by the rotation of the motor 213, the workpiece 291 is conveyed on the belt.

The belt conveyor 226 is an unloader that transports the workpiece 291 in the Y direction from the positive side to the negative side. Motor 214 is connected to belt conveyor 226 . When a built-in belt is driven by the rotation of the motor 214, the workpiece 291 is conveyed on the belt.

The motor 211 is connected to the ball screw 222 . The motor 211 is a servomotor that receives a current controlled by the driving device 230 and rotates a shaft.

The motor 212 is connected to the linear rail 221 . The motor 212 is a linear servomotor that receives a current controlled by the driving device 230 and is driven in the X-direction.

Motor 213 is connected to belt conveyor 225 and motor 214 is connected to belt conveyor 226 . The motors 213 and 214 are each a stepper motor which receives a pulsed electrical signal from the drive device 230 and rotates a shaft.

The actuators 215 are a plurality of vacuum pads that receive an electrical signal from the drive device 230 and suck or release the workpiece 291 . The workpiece 291 to be taken out is held or released by the suction or release operation.

The drive device 230 includes a plurality of motor drives 231, 232 and 233 and a motor controller 234.

The motor driver 231 is connected to the motor 211 via a cable and supplies the motor 211 with driving energy depending on a rotation angle of the motor 211. The motor driver 232 is connected via a cable to the motor 212 and supplies the motor 212 with driving energy depending on a position of the motor 212. The motor driver 233 is connected to the motors 213 and 214 by cables and transmits a pulse command to the motors 213 and 214.

The motor controller 234 performs integral control of the motor drivers 231, 232 and 233 and controls driving of the motors 211, 212, 213 and 214. The motor drivers 231, 232 and 233 and the motor controller 234 are connected by a cable and can receive information through communication exchange with each other. The motor controller 234 is also connected to the actuators 215 by a cable (not shown) and controls the suction and release of the workpiece 291 by the actuators 215 using an electrical signal. A vacuum pump is an example of the actuator 215 since the workpiece 291 is sucked or released. The workpiece 291 is sucked when the vacuum pump is operated, and the workpiece 291 is released when the vacuum pump is not operated.

The driving noise detection unit 11 is a microphone disposed adjacent to the machine 220 to be driven, and detects noise generated by the diagnosis object 200 . The driving sound detection unit 11 is provided with a communication unit 241A. An example of the communication unit 241A is a wireless communication device. The driving noise detection unit 11 combines the detected driving noise with a time stamp, i. H. with the time when the noise is detected as a sentence, and transmits the sentence to the server device 250 via the communication unit 241A.

The operating state detection unit 12 is a logger that is connected to the drive device 230 via a cable and records information about the drive device 230 as the operating state. The operating state detection unit 12 communicates with the driving device 230 as necessary, and acquires and stores various data including the operating state. The operating add Status detection unit 12 detects a speed command for the motors 211, 212, 213 and 214 and an intake state of the actuators 215 as the operating state of the diagnosis object 200 and outputs them to the time synchronization unit 13. Binary data for aspiration or release is an example of the aspiration state.

The operating state detection unit 12 is provided with a communication unit 241B. An example of the communication unit 241B is a wireless communication device. The operating state detection unit 12 combines the detected operating state with a time stamp, i. H. with the time when the operation state is detected as a sentence, and transmits the sentence to the server device 250 via the communication unit 241B.

Even though 13 shows a case in which a single driving noise detection unit 11 and a single operating state detection unit 12 are provided for the motors 211, 212, 213 and 214 and the actuators 215 of the plurality of axles, a plurality of driving noise detection units 11 or one A plurality of operating state detection units 12 may be provided. In particular, in a case where the machine 220 to be driven has a configuration that eliminates interference due to the driving noise of another motor through a shield or the like, more accurate diagnosis can be expected by using the driving noise detection unit 11 for each of the motors 211, 212, 213 and 214 and the actuators 215 is provided.

The wireless network device 240 is a device that performs wireless communication with the communication units 241A and 241B. The wireless network device 240 includes a communication unit 241C. A wireless local area network (LAN) is formed by the wireless network device 240 and the communication units 241A, 241B and 241C, with the wireless network device 240 serving as an access point of the communication units 241A, 241B and 241C. The wireless network device 240 also serves as a router and relays communication between each terminal and a network via the communication line 280 . Note that the communication units 241A, 241B, and 241C and the wireless network device 240 perform time synchronization between the wireless devices including the communication units 241A, 241B, and 241C and the wireless network device 240 when necessary. This is to set an accurate time for the time to be used by the driving noise detection unit 11 and the operating state detection unit 12 .

In the second embodiment, a function processing unit that performs cause identification using the sound vibration data and the operation data is distributed between the server device 250 and the user terminal 260 . The server device 250 and the user terminal 260 are connected via the communication line 280 .

When instructed by the user terminal 260 to execute the cause identification processing for the driving noise of the machine to be driven 220, the server device 250 performs processing that generates device data including feature point data, using the sound vibration data, i. H. time-series data of the driving noise detected from the machine 220 to be driven, and the operation data, i. H. Time series data of operating status used.

14 12 is a block diagram schematically showing an example of a functional configuration of the server device according to the second embodiment. The server device 250 is an information processing device installed on the network using the communication line 280 . The server device 250 includes functions of the time synchronization unit 13, the operation mode extraction unit 14, the acoustic vibration time-series spectrum acquisition unit 15, and the feature point extraction unit 16. That is, the server device 250 includes, as applications in the server device 250, the time synchronization unit 13, the operation mode extraction unit 14, the acoustic vibration time-series spectrum acquisition unit 15, and the feature point extraction unit 16. Note that the second embodiment shows a configuration in which the Operating vibration extraction unit 17 is omitted.

The server device 250 further includes a communication unit 251 that performs communication with the driving sound detection unit 11 and the operation state detection unit 12 and with the user terminal 260 .

The server device 250 further includes a device data storage unit 252. The device data storage unit 252 is a database such as e.g. B. a Relational Database Management System (RDBMS) or a Not only SQL. The server device 250 collects device data including the driving sound and the operating state, which the driving sound detection unit 11 and the operating ass tands detection unit 12 are transmitted to the network, in a database, processes the device data as required and then stores the device data. For example, as described in the first embodiment, the acoustic vibration data and the operation data are synchronized with each other, the time-series spectrum data is acquired for the acoustic vibration data, the feature point is extracted from the spectrum data, and the device data including the feature point data in which the frequency, the time, the power and the waveform of the feature point and the operation data at the time corresponding to the feature point are combined as a set are stored in the device data storage unit 252 .

15 14 is a table showing an example of recording device data according to the second embodiment. A record of device data includes registration time, operation mode, feature point data, and operation data. It should be noted that by analyzing the data collected in the server device 250 through deep learning or the like, the cause identification can be performed with higher accuracy.

The user terminal 260 instructs the server device 250 to perform the cause identification on the driving noise of the machine to be driven 220, performs the cause identification processing using the device data obtained from the server device 250, and displays the result. The user terminal 260 is an information processing device owned by a user of the machine to be driven 220. An example of the user terminal 260 is a laptop or a desktop personal computer.

16 14 is a block diagram schematically showing an example of a functional configuration of the user terminal according to the second embodiment. The user terminal 260 has the function of the cause determination unit 18 described in the first embodiment. i.e. the user terminal 260 contains the cause determination unit 18 as an application in the user terminal 260.

The user terminal 260 further includes a communication unit 261, an input unit 262, and a display unit 263. The communication unit 261 communicates with the server device 250 via the communication line 280. For example, a command for performing cause identification processing inputted by the user into the input unit 262 is transmitted to the server device 250, and various kinds of information including the device data are received from the server device 250. The communication unit 261 may also be connected to the driving noise detection unit 11 and the operating condition detection unit 12 to transmit the sound vibration data, i. H. the time-series data of the engine noise, and the operational data, d. H. the time series data of the operating status.

The input unit 262 is an input interface with a user. A keyboard or a mouse is an example of the input unit 262. A command to perform the cause identification processing on the machine to be driven 220 is input through the input unit 262, for example.

The display unit 263 displays information necessary for executing the cause identification processing on the machine 220 to be driven. A liquid crystal display is an example of the display unit 263. The display unit 263 displays a result of the cause identification, the device data acquired via the network, or the like.

The network device 270 is a communication device that relays the communication between the driving sound detection unit 11 and the operating state detection unit 12 provided in the machine 220 to be driven, the server device 250 and the user terminal 260 . A router is an example of network device 270.

By configuring the engine noise diagnosis system 10B via the network as in 13 1, the driving noise can also be diagnosed at a remote location other than the factory where the machine 220 to be driven is installed.

The server device 250 and the user terminal 260 that perform the diagnosis of the engine noise are each implemented by the information processing device as described above. 17 Fig. 12 is a block diagram showing an example of a hardware configuration of the server device and the user terminal. The server device 250 and the user terminal 260 each include a computing device 401, a working memory 402, a memory 403, a communication device 404, an input device 405 and a display device 406. The computing device 401, the working memory 402, the memory 403, the communication device 404, the Input device 405 and display device 406 are connected via a bus line 407 .

The arithmetic device 401 is a processor including a CPU that performs arithmetic processing. The working memory 402 functions as a work area for storing data used by the computing device 401 during computation processing. In the memory 403, a computer program, information and the like are stored. The communication device 404 includes a function of communicating with another device connected to a network. The input device 405 receives input from an operator. The input device 405 is a keyboard, a mouse, or the like. The display device 406 outputs a display screen. The display device 406 is a monitor, display, or the like. Note that a touch panel in which the input device 405 and the display device 406 are integrated can be used.

The functions of the time synchronization unit 13, the operation mode extraction unit 14, the acoustic vibration time-series spectrum acquisition unit 15 and the feature point extraction unit 16 shown in 14 are implemented by the computing device 401 reading and executing the computer program stored in the memory 403.

Also the function of the in 16 Cause determining unit 18 shown is implemented by computing device 401 reading and executing the computer program stored in memory 403 .

Next, the operation of the diagnosis object 200 will be described. A user of the pick-up unit inputs drive commands for the motors 211, 212, 213 and 214 and the actuators 215 into the motor controller 234 to feed the workpiece 291 into the pick-up unit. For example, the user inputs a command to motor controller 234 through user terminal 260 . The command is then transmitted to the motor controller 234 via the network device 270, the communication line 280, the server device 250, the wireless network device 240, the communication unit 241C, and the communication unit 241B. The motor controller 234 generates a drive command according to the inputted command and transmits it to each of the motor drivers 231, 232 and 233.

The motor drivers 231, 232, and 233 control the motor drive current according to the received drive command, and drive the motors 211, 212, 213, and 214 and actuators 215, respectively.

When the motor 211 is driven, the ball screw 222 connected to the motor 211 rotates, and the head 224 and actuators 215 connected to the ball screw 222 move in the Z direction.

When the motor 212 is driven, the linear guide 223 connected to the motor 212, the motor 211 and the device connected to the motor 211 move in the X direction.

When the motor 213 is driven, the belt conveyor 225 rotates and the workpiece 291 on the belt conveyor 225 moves in the Y direction from the positive side to the negative side. Similarly, when the motor 214 is driven, the belt conveyor 226 rotates and the workpiece 291 on the belt conveyor 226 moves in the Y direction from the positive side to the negative side.

The actuators 215 hold the workpiece 291 by sucking the workpiece 291 moving on the belt conveyor 225 directly under the actuators 215 at the time of contact. This allows the workpiece 291 to remain in contact with the actuators 215 even if the actuators 215 move up after being sucked by the motor 211 being driven.

Thereafter, motor 211 is rotated to cause actuators 215 to move upward. At this time, since the workpiece 291 is sucked by the actuators 215 , the workpiece 291 rises up with the rise of the actuators 215 and is separated from the belt conveyor 225 .

Next, by driving the motor 212 to the positive side in the Y direction, the workpiece 291 moves just above the belt conveyor 226. Thereafter, the motor 211 is rotated again to cause the workpiece 291 to be placed on the belt conveyor 226. The actuators 215 place the workpiece 291 on the belt conveyor 226 by releasing the workpiece 291 at a timing when the held workpiece 291 comes into contact with the belt conveyor 226 . Upon release by actuators 215, motor 211 and motor 212 are driven to cause actuators 215 to return to the start position. Finally, the motor 214 is rotated to cause the workpiece 291 to be conveyed out toward the negative side in the Y direction of the belt conveyor 226 .

In the course of the operation described above, the diagnosis object 200 generates a driving sound when driven. The drive noise generated by the diagnosis object 200 and the rotation angles of the motors 211, 212, 213, and 214 are recorded and detected by the driving sound detection unit 11 and the operating state detection unit 12, respectively, and transmitted to the server device 250 by wireless communication.

The server device 250 communicates with the driving sound detection unit 11 and the operating condition detection unit 12 via the wireless network device 240 as needed, acquires various types of information related to the diagnosis object 200 including the driving sound and the operating condition, and stores the information in the device data storage unit 252 As pre-processing for storage in the device data storage unit 252, the time synchronization unit 13, operation mode extraction unit 14, acoustic vibration time-series spectrum acquisition unit 15, and feature point extraction unit 16 perform processing on the acoustic vibration data and operation data as acquired time-series data, and calculate the mode of operation and the feature point data. The device data storage unit 252 stores the feature point data calculated by the time synchronization unit 13, the operation mode extraction unit 14, the acoustic vibration time-series spectrum acquisition unit 15 and the feature point extraction unit 16, the operation data in which the time of the feature point has been synchronized, and other necessary information regarding the diagnosis object 200.

In response to a request from the user, the user terminal 260 collects various types of information about the system that the user needs, such as: B. the presence or absence of an abnormality of the system, through the server device 250 and displays the information on the display unit 263 of the user terminal 260. The user can also perform analysis necessary for the maintenance and operation of the diagnosis object 200 and the powertrain noise diagnosis system 10B using a database stored in the device data storage unit 252 of the server device 250 .

Next, a driving noise diagnosing method in the driving noise diagnosing system 10B according to the second embodiment will be described. It should be noted that the basic driving noise diagnosing method is basically similar to that of the first embodiment, so the differences from the first embodiment will be mainly described below.

In the second embodiment, the time synchronization unit 13 of the server device 250 synchronizes the sound vibration data and the operational data using the time stamp in the set combined by the driving noise detection unit 11 and the operational state detection unit 12 . Thereby, the sound vibration data and the operation data can be synchronized even when real-time communication cannot be secured due to poor communication quality or the like.

In addition, when the operation mode extraction unit 14 or the like of the server device 250 requires data of a specific time interval, data is extracted with a time stamp of the time interval specified by the time stamp. The extraction by this method enables the analysis of the acoustic vibration data and the operational data as synchronized data even if the acquisition periods of the acoustic vibration data and the operational data are different.

As another method, in the case where the detection periods of the sound vibration data and the operational data are different, a method is conceivable that reduces the data after performing the filter processing to have the same period, or on the contrary, a method is conceivable the data in between by a technique such. B. linear interpolation extracted after filter processing.

In addition, the operation mode extraction unit 14 sets the operation mode by dividing the steps according to which motor of the plurality of motors 211, 212, 213, and 214 and the actuators 215 is mainly operated. Specifically, the operation mode extraction unit 14 determines the operation mode by identifying from the speed commands of the motors 211, 212, 213 and 214 and the data of the pressure contact state of the actuators 215 based on the operation data which of the motors 211, 212, 213 and 214 or of the actuators 215 is operated. In one example, during loading of the workpiece, motor 213 is mainly operated, motors 211 and 212 are mainly operated during lifting of the workpiece, motor 214 is mainly operated during ejection of the workpiece, and actuators 215 are mainly operated during pumping operated. Accordingly, in one example, the operation mode extraction unit 14 divides the operation mode into work-in, work-up, work-out, and pumping.

The cause determination unit 18 of the user terminal 260 compares the feature point data with the numerical range generated by the server device 250 as the cause determination condition, and identifies the cause of the driving noise. The server device 250 may generate the numerical range based on a parameter estimated as a cause based on knowledge, or may generate the numerical range through machine learning or the like based on the operation data and the sound vibration data. Alternatively, the server device 250 may dynamically generate the cause determination condition immediately before the cause determination condition is identified by the cause determination unit 18 . In this case, the cause diagnosis can be performed based on a more urgent case, so that a more accurate diagnosis can be performed.

It is to be noted that in the second embodiment, unlike the configuration of FIG 1 , the operating vibration extracting unit 17 is not provided. The cause determination unit 18 thus identifies the cause using the feature point data without using the vibration data.

The engine noise diagnosis system 10B according to the second embodiment diagnoses the cause of the engine noise by comparing the feature point data obtained through a time-frequency analysis of the sound vibration data and the operation data with the numerical range provided by the server device 250 for each operation mode was generated. Thereby, it is not necessary to prepare the operational data in a normal state or in an abnormal state in advance, so that even if the configuration of the driving device or the engine is changed, the diagnosis of the driving noise can be performed in a versatile and immediate manner.

Further, the driving noise diagnosis system 10B according to the second embodiment can immediately determine the cause of the driving noise generated by the machine 220 to be driven, which is driven by the plurality of motors 211, 212, 213 and 214 and actuators 215. In particular, it is easier to determine which part is the cause and which motor is driving it. Thereby, when the driving noise of the machine 220 to be driven is problematic, the user can take appropriate action depending on the cause, and the time required for the action can be reduced.

Third embodiment.

18 14 is a block diagram showing an example of a functional configuration of a cause determination unit in a powertrain noise diagnostic system according to a third embodiment. The powertrain noise diagnostic system according to the third embodiment is obtained by replacing the cause determination unit 18 of the arithmetic processing unit 140 in the first embodiment with a cause determination unit 18A that identifies a cause of the powertrain noise using a learning device that has learned the cause determination of the powertrain noise. It should be noted that in the following description, the same components as in the first embodiment are denoted by the same reference numerals as assigned to those components in the first embodiment, so that the description of those components and parts different from those will be omitted differ from the first embodiment will be described.

The cause determination unit 18A includes a learning result storage unit 31 and a cause inference unit 32. The learning result storage unit 31 stores a learning result obtained by performing machine learning in advance to identify a cause of the driving noise for the feature point data extracted by the feature point extraction unit 16 appreciate. The cause inference unit 32 performs arithmetic processing based on the learning result in the learning result storage unit 31 .

The cause determination unit 18A uses the feature point data extracted by the feature point extraction unit 16 as input and performs the cause identification on the driving noise by performing the arithmetic processing based on the learning result in the learning result storage unit 31 by the cause inference unit 32 .

Here, the learning result stored in the learning result storage unit 31 may be a machine learning result using all the feature point data or a machine learning result using part of the feature point data. In addition, the cause inference unit 32 extracts input data and performs arithmetic processing according to the feature point data used at the time of learning from the learning result storage unit 31 .

In addition, the learning models of machine learning belonging to the one in the Learning result storage unit 31 leads stored learning result, the K-Nearest Neighbor Algorithm, the decision tree, the support vector machine and the kernel approximation. Deep learning can also be used as a learning model.

The cause determination unit 18A of the third embodiment performs the cause identification on the driving noise using an informed learning device. This makes it possible to provide a more accurate determination of the cause of the engine noise. In addition, the cause determination unit 18A uses a learning device that has performed machine learning using the same data as the cause determination unit 18, thereby being able to perform the identification regardless of the driving pattern. Thereby, the learning result stored in the learning result storage unit 31 can be applied to various drive devices.

Fourth embodiment.

19 14 is a block diagram showing an example of a functional configuration of a machine learning device for a powertrain noise diagnosis system according to a fourth embodiment. A machine learning device 50 is obtained by allowing the server device 250 and the user terminal 260 of the second embodiment to have a machine learning function, wherein the server device 250 and the user terminal 260 include the functional processing unit that performs cause identification using the Performs sound vibration data and operational data. Note that in the following description, the same components as in the first and second embodiments are denoted by the same reference numerals as assigned to those components in the first and second embodiments, so the description of those components is omitted and Parts different from those of the first and second embodiments will be described.

The machine learning device 50 includes the driving noise detection unit 11, the operating state detection unit 12, the sound vibration time-series spectrum detection unit 15, the feature point extraction unit 16, a cause detection unit 51, a cause learning unit 52 and a learning result storage unit 53.

The cause acquisition unit 51 acquires data related to a cause of a driving noise. For example, the cause acquisition unit 51 may acquire data related to a cause of a measured engine noise inputted by an action of a designer or a user. Alternatively, the cause detection unit 51 may detect a diagnosis result of a cause of engine noise according to the first or second embodiment, or may acquire a diagnosis result of a cause of engine noise in another engine noise diagnosis system.

The cause learning unit 52 learns the cause of the engine noise from a training data set based on a combination of the feature point data extracted by the feature point extracting unit 16 and the data related to the cause of the engine noise acquired by the cause acquiring unit 51 . A learning model used in the cause learning unit 52 includes, for example, the learning model described in the third embodiment. Diagnostic data from a large number of machines 220 to be driven can be used as diagnostic data for the training data set. In one example, the diagnosis data is collected by connecting to the plurality of driven machines 220 via the communication line 280 or the like, so that more accurate diagnosis can be performed. The learning result storage unit 53 stores a result of learning by the cause learning unit 52.

20 14 is a block diagram schematically showing an example of a functional configuration of a server device according to the fourth embodiment. A server device 250</b>A is an information processing device installed on a network using the communication line 280 . The server device 250A includes the time synchronization unit 13, the operation mode extraction unit 14, the acoustic vibration time-series spectrum acquisition unit 15, the feature point extraction unit 16, the cause learning unit 52, the learning result storage unit 53, and the communication unit 251.

The server device 250A processes device data, including the driving sound and the operating state, transmitted from the driving sound detection unit 11 and the operating state detection unit 12 to the communication line 280 as necessary, and then learns and stores the device data. For example, first, as in the second embodiment, all or part of the feature point data in which the frequency, time, power and waveform of the feature s point and the operation data at the time corresponding to the feature point combined as a set are input to the cause learning unit 52 . Next, the cause of the driving noise is acquired from the user terminal 260 through the communication line 280 and the communication unit 251 and combined to obtain the training data set. Finally, the cause learning unit 52 learns the cause of the engine noise using the training data set and stores the result in the learning result storage unit 53.

21 14 is a block diagram schematically showing an example of a functional configuration of a user terminal according to the fourth embodiment. A user terminal 260A includes the communication unit 261, an input unit 262A, the display unit 263, the cause determination unit 18, and the cause detection unit 51.

The cause acquisition unit 51 acquires data related to the cause of the driving noise from a user, for example. The acquired data related to the cause of the engine noise is transmitted to the server device 250A via the communication unit 261 .

The input unit 262A is an input interface with the user. The input unit 262A serves as an interface when the user inputs an instruction for executing the cause identification processing for the machine to be driven 220 and when the user inputs the data acquired by the cause acquisition unit 51 related to the cause of the driving noise.

The engine noise diagnosis system according to the fourth embodiment has both a function of machine learning the cause of the engine noise and a function of identifying the cause using a result of the machine learning function. Thereby, the accuracy of the cause diagnosis can be increased by continuing learning while using the cause diagnosis of the engine noise.

In addition, by connecting with the plurality of machines 220 to be driven or the user terminal 260A via the communication line 280, more accurate cause diagnosis can be used in a plurality of places regardless of the installation site of the machine 220 to be driven. This makes it possible to learn a very versatile diagnosis.

The powertrain noise diagnosis system according to the fourth embodiment has both the function of machine learning the cause of the powertrain noise and the function of identifying the cause using the result of the machine learning function, however, the machine learning device 50 of the powertrain noise diagnosis system may can also be configured alone, and the identification of the cause using the learning result can be a function of another device. An example of a device for identifying the causes using the learning result is a device with the in 18 illustrated configuration of the third embodiment.

The configuration shown in the above embodiment is merely an example of the content of the present invention, and therefore may be combined with another known technique, or partially omitted and/or modified without departing from the scope of the present invention.

Reference List

10, 10A, 10B

drive noise diagnostic system;

11

drive noise detection unit;

12

operating state detection unit;

13

time synchronization unit;

14

operating mode extraction unit;

15

sound vibration time-series spectrum acquisition unit;

16

feature point extraction unit;

17

operational vibration extraction unit;

18, 18A

cause determination unit;

31, 53

learning result storage unit;

32

cause inference unit;

50

machine learning device;

51

cause detection unit;

52

cause lesson;

100, 200

diagnosis object;

110, 211, 212, 213, 214

Engine;

120, 220

machine to be driven;

130, 230

drive device;

131, 231, 232, 233

motor drive;

132, 234

engine control;

133

Advertisement;

140

arithmetic processing unit;

215

actuator;

250, 250A

server device;

251, 261

communication unit;

252

device data storage unit;

260, 260A

user terminal;

270

network device;

280

communication line.

QUOTES INCLUDED IN DESCRIPTION

This list of the documents cited by the applicant was generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Patent Literature Cited

• JP H10274558 [0004]

Claims (8)

Drive noise diagnostic system, comprising: a driving noise detection unit for detecting a driving noise, d. H. noise or mechanical vibrations generated in an actuator or a machine to be driven by the actuator; an operating state detection unit for detecting a driving position, a driving speed or a force generated by driving the actuator in a time series; an acoustic vibration time-series spectrum acquisition unit for calculating a frequency spectrum corresponding to each time point of acoustic vibration data, d. H. of time-series data of the detected engine noise, and for outputting a time-series spectrum in which the power of the calculated frequency spectrum is associated with a frequency and time as a set; a feature point extracting unit for extracting, as a feature point, a point where a waveform of power of the time-series spectrum satisfies a predetermined condition in terms of frequency and time, and for outputting feature point data in which the frequency of the feature point, time, the waveform of the feature point and operational data, d. H. the driving position, the driving speed, or the force generated by driving the actuator at the time corresponding to the feature point are combined as a set; and a cause determining unit for determining a cause of the detected driving noise by comparing a cause determining condition with a numerical value of the feature point data, wherein the cause determining condition for each phenomenon occurring in the actuator or the machine to be driven as a cause of the driving noise defines a first numerical range when at least one of the frequency and the time including the feature point occurring with the phenomenon and the operational data, d. H. the driving position, the driving speed, or the force generated by driving the actuator at the time when multidimensional data is combined. drive noise diagnostic system claim 1 , further comprising an operation mode extraction unit for dividing an operation state of the machine to be driven into two or more periods based on the operation data, wherein the acoustic vibration time-series spectrum acquisition unit calculates the frequency spectrum using the acoustic vibration data of at least one point falling between a start time and an end time of the section is detected. drive noise diagnostic system claim 2 , wherein the operation mode extracting unit extracts a portion where the rotational speed of the actuator falls within a specified range during a period longer than a predetermined period. Drive noise diagnostic system according to one of the Claims 1 until 3 , further comprising an operational vibration extracting unit for extracting a vibration component from the operational data, wherein the cause determination unit determines the cause of the detected driving noise by, for each phenomenon, in addition to the comparison between the numerical value of the feature point data and the first numerical range, the vibration component with a second numerical area including the generated vibration component accompanying the phenomenon is compared. Drive noise diagnostic system according to one of the Claims 1 until 4 , wherein the feature point to be examined by the feature point extraction unit is a point corresponding to a peak in the waveform of the power of the time-series spectrum with respect to frequency and time. Drive noise diagnostic system according to one of the Claims 1 until 5 wherein the cause determination unit includes: a learning result storage unit for storing a learning result obtained by performing machine learning to estimate the feature point data and a cause of the noise for the operation data; and a cause inference unit for receiving the multi-dimensional data as an input, performing arithmetic processing based on the learning result stored in the learning result storage unit, and determining the cause of the driving noise. A driving noise diagnosing method comprising: a step of detecting driving noise, ie noise or mechanical vibration generated in an actuator or a machine to be driven by the actuator; a step of detecting a driving position, a driving speed, or a force generated by driving the actuator in a time series; a step of calculating a frequency spectrum corresponding to each time point of sound vibration data, ie, time-series data of detected driving noise, and outputting a time-series spectrum in which the power of the calculated frequency spectrum is related to a frequency and time as a set; a step of extracting, as a feature point, a point where a waveform of power of the time-series spectrum satisfies a predetermined condition in terms of frequency and time, and outputting feature point data in which the frequency of the feature point, time, waveform of the feature point and operation data, ie, the driving position, the driving speed, or the force generated by driving the actuator at the time corresponding to the feature point, are combined as a set; and a step of determining a cause of the detected drive noise by comparing a cause determination condition with a numerical value of the feature point data, the cause determination condition for each phenomenon occurring in the actuator or the machine to be driven as a cause of the drive noise defining a first numerical range when at least one of the frequency and time including the feature point occurring with the phenomenon and the operation data, ie the driving position, the driving speed or the force generated by driving the actuator at the time are combined as multidimensional data. Machine learning apparatus for a powertrain noise diagnostic system, comprising: a driving noise detection unit for detecting a driving noise, d. H. noise or mechanical vibrations generated in an actuator or a machine to be driven by the actuator; an operating state detection unit for detecting a driving position, a driving speed or a force generated by driving the actuator in a time series; an acoustic vibration time-series spectrum acquisition unit for calculating a frequency spectrum corresponding to each time point of acoustic vibration data, d. H. of time-series data of the detected engine noise, and for outputting a time-series spectrum in which the power of the calculated frequency spectrum is associated with a frequency and time as a set; a feature point extracting unit for extracting, as a feature point, a point where a waveform of power of the time-series spectrum satisfies a predetermined condition in terms of frequency and time, and for outputting feature point data in which the frequency of the feature point, time, the waveform of the feature point and operational data, d. H. the driving position, the driving speed, or the force generated by driving the actuator at the time corresponding to the feature point are combined as a set; a cause learning unit for learning a cause of the driving noise based on multidimensional data obtained by using at least one of the frequency and time including the feature point for each phenomenon occurring in the actuator or the machine to be driven as a cause of the driving noise , which occurs with the phenomenon, and the operational data, i. H. the driving position, the driving speed or the force generated by driving the actuator are combined at the time; a learning result storage unit for storing a learning result learned by the cause learning unit; and a cause acquiring unit for acquiring the cause of the driving noise to be transmitted to the cause learning unit.

Images