JP2010117336A - Abnormal noise determination tool - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an abnormal noise determination tool capable of easily and surely knowing and analyzing abnormal noise conditions of a machine to be measured by a simple operation. <P>SOLUTION: A sensor unit 2 comprising a microphone is mounted on a standard machine 1 in which generated abnormal noise satisfies non-defective reference. A picked-up abnormal signal is fast-Fourier-transformed 5. The maximum value of amplitude belonging to a plurality of preset frequency bands is stored 6 and 7 as a determination reference signal of each band together with frequency. In addition, the sensor unit is mounted on a machine to be measured whose abnormal noise level is determined, and a measured abnormal noise signal which is picked up is fast-Fourier-transformed. A converted signal is classified into the frequency bands. The maximum value of amplitude belonging to each band is stored together with the frequency. The stored determination reference signal of each band and the measured abnormal noise signal of the measured machine are subject to comparison operation 8, and the results are displayed. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

Spectrum conversion is performed on a judgment reference signal picked up from a standard device that is judged that the generated abnormal noise satisfies the non-defective standard and a measured abnormal noise signal picked up from the device under test, and the frequency band to be evaluated is divided into a plurality of frequency bands. By subtracting the maximum value of the amplitude belonging to each frequency band, comparing and processing the difference for each frequency band, and displaying the result, the abnormal sound state of the device under test can be displayed with a simple operation. An allophone judgment tool that can easily and reliably judge.

As one of the judgments as to whether the device is operating normally, abnormal noise detection during operation has been put into practical use. In electrical equipment, many abnormal noises such as abnormal noise generated from hard disks, compact disks, fans, solenoids, power supplies, etc., or abnormal noise caused by rubbing caused by magnetic tape, paper running, etc. occur. In addition to being uncomfortable, there are many cases where some kind of malfunction has occurred in the equipment. Whether or not the noise that is generated is in the normal range is often judged by a skilled worker by directly hearing it, but in recent years, sensors such as accelerometers, vibration sensors, microphones, etc. Various methods have been put into practical use, such as fast Fourier transform of the signal picked up by attaching and analyzing the picked up signal, but the abnormal sound picked up varies greatly depending on the measurement location, and the standard setting for judging good or bad Therefore, there is a strong demand for a pass / fail judgment method that can easily and reliably grasp and analyze the abnormal sound state of the device under test by simple operations. Yes.

In a method in which an operator listens to various abnormal sounds generated by the operation of the device with his / her ears, there are individual differences in the determination criteria, and the criteria are ambiguous. On the other hand, the method of collecting sound with an accelerometer, vibration sensor, microphone, etc., converting the spectrum and analyzing it is possible to make a quantitative judgment, but the picked-up noise is the position of various parts on the equipment. The relationship between the devices under test is simple, easy and reliable, and there are various problems such as the determination of pass / fail criteria and difficult pass / fail judgment methods. Provide an allophone judgment tool that can grasp and analyze the sound state.

A standard device that is judged that the generated abnormal noise satisfies the non-defective standard is attached to a sensor unit consisting of a microphone and the abnormal noise signal picked up is subjected to spectrum conversion by fast Fourier transform, and a plurality of preset frequencies. The maximum amplitude value belonging to the band is stored together with the frequency as a judgment reference signal for each band, and the measured abnormal noise signal picked up by attaching the sensor unit to the device under test for determining the abnormal noise level is fast Fourier transformed. The spectrum is converted by, and the converted signal is classified into frequency bands, the maximum amplitude value to which each band belongs is stored together with the frequency, and the stored reference signal for each band is compared with the measured abnormal signal of the device under test. Calculate and display the result.

Because it does not require strict adherence compared to vibration sensors, abnormal noise signals are picked up by a sensor unit consisting of a microphone that can be easily attached to equipment, and abnormal sound quality judgment is not an absolute value evaluation. By comparing and calculating the maximum amplitude value for each frequency band of the standard device and the device under test, at which frequency the abnormal sound signal of the device under test exceeds the abnormal signal of the standard device with a small data storage capacity, By grasping what kind of difference is occurring and evaluating the maximum value of amplitude for each frequency band obtained by continuously picking up multiple parts of the equipment with a sensor unit that is easy to install, It is possible to determine the occurrence of abnormal noise easily and reliably over the entire range of equipment.

FIG. 1 is a conceptual diagram of a configuration of an abnormal sound determination tool according to the present invention. Reference numeral 1 denotes a standard device in which the generated abnormal noise is determined to satisfy a non-defective product standard, or a device to be measured that determines an abnormal noise level, and 2 denotes a sensor unit that picks up the abnormal noise signal with a microphone. The picked-up signal is amplified as required by the amplifier circuit 3, spectrum-converted by the A / D conversion circuit 4 and FFT processing circuit 5, and the judgment reference signal obtained from the standard device is stored in the storage circuit 6, and the device to be measured The measurement abnormal noise signal obtained from the above is stored in the storage circuit 7, the determination reference signal and the measurement abnormal noise signal are processed by the comparison calculation processing circuit 8 as necessary, and the required result is displayed by the display circuit 9. FIG. 2 shows a sensor unit configured by attaching a condenser microphone 10 to a piece 11 of a heavy metal material such as brass that shields ambient noise. To do.

FIG. 3 is a conceptual diagram showing an arrangement example of various parts such as HDD, POW, FDD, FAN and the like mounted on an abnormal sound measurement target 13 such as a personal computer or a car navigation system. An outline of a member that generates an abnormal noise in an abnormal noise measurement target, a mounting position, a basic frequency of the generated abnormal noise, and the like is known in advance for each device. However, the abnormal sound level picked up by the sensor unit is slightly different depending on the mounting position of the sensor unit, and there is also a delicate difference regarding the fundamental frequency of the abnormal sound generated by each part. Therefore, the value obtained in a single measurement at a specific location is picked up by small noise from parts far from the measurement position and has a large measurement error. A pickup method is preferred in which data is obtained by checking multiple times in the vicinity of each part using a simple microphone-type sensor unit.

In the method of checking abnormal noise multiple times, storing all the picked up abnormal noise signals one after another requires enormous storage capacity and complicates the method of analyzing and judging the occurrence of abnormal noise. Therefore, it is not preferable. In general, the noise frequency generated by various parts mounted on equipment indicates the vicinity of a specific frequency specific to each part. Therefore, the frequency band is a frequency band with a predetermined width centered on the specific frequency of each part. And the peak value of the amplitude in each band is stored along with the frequency, and if necessary, the storage data is referred to and the abnormal sound situation is compared and processed. .

FIG. 4 shows the frequency bands of X1, X2, X3, and X4, where F is the frequency band of the abnormal sound to be evaluated, and the basic abnormal sound of the HDD is fa, FDD is fb, FAN is fc, and POW is fd. And L indicates the spectrum amplitude value. Stores the maximum amplitude and frequency in each frequency band to which the spectrum of the abnormal noise picked up belongs, and in multiple pickups, the maximum value and frequency each time a larger maximum amplitude is obtained with the pickup. Are updated and stored, and the difference between the judgment reference signal and the measured abnormal signal of the device under test is compared and processed as necessary. In this way, the FFT output is classified into frequency bands having a predetermined bandwidth, and the maximum value in each band is analyzed. Therefore, even if the frequency of abnormal noise generated by each part is slightly different, The maximum value is evaluated in the band. For this reason, even when multiple locations are picked up multiple times, the resulting abnormal sound signal is processed for each frequency band, so there is no need to memorize which location is picked up and how many times it is picked up. It is only necessary to check what frequency and magnitude the maximum amplitude value can be, and simple and reliable determination of the abnormal sound state of the device under test can be made with a simple operation.

The second aspect of the present invention stores a plurality of amplitude values, together with the maximum amplitude value in each frequency band, as a reference signal for the standard device and a measurement abnormal sound signal for the device under measurement together with the frequency, and measures the device under measurement. Since the comparison calculation process is performed to monitor and analyze how much the abnormal sound signal is different from the determination reference signal, it is possible to make a more detailed situation determination rather than a determination based on the maximum value. In particular, it is effective for the determination when various parts mounted on the device emit a plurality of different sounds having different frequencies within the same frequency band. Further, according to the third aspect of the present invention, in the abnormal sound determination method of the present invention, the measured abnormal sound signal of the device under test is picked up without comparing and processing the measured abnormal sound signal of the device under test and the determination reference signal of the standard device. This is useful when there is a need to evaluate directly.

When using a vibration sensor or the like as a sensor unit, close contact with the device under test is required, and a large error may occur depending on the close contact state. Therefore, a condenser microphone with a simple close contact and excellent operability is used. Is made of high specific gravity brass to efficiently collect noise generated by the device under test and to be collected by the microphone, and has a thickness of 4 mm, a maximum diameter of 40 mm, and a height of 30 mm. A hard disk was used as a storage medium for storing the judgment reference signal and the measured abnormal sound signal. As a method of comparing and calculating the measured abnormal sound signal of the device under test and the judgment reference signal of the standard equipment, a comparison arithmetic processing that evaluates how many times the maximum amplitude of the abnormal sound signal is the maximum amplitude of the judgment reference signal is adopted. However, the absolute value of the difference can be evaluated. As the result display, various methods such as “difference multiple display”, “difference absolute value display”, “abnormal sound level OK / NG display”, and the like can be considered. The frequency band to be set may be a method of monitoring only the frequency band range including the fundamental frequency of the abnormal sound generated by each part mounted on the equipment without covering the entire frequency band of the abnormal sound to be evaluated, Bands may be set intermittently. Furthermore, various modifications can be made without departing from the spirit of the present invention, such as using iron as the material of the piece and semiconductor memory as the storage medium.

Multiple frequency bands in the frequency band to be evaluated by spectrum conversion of the judgment reference signal picked up from a standard device that is judged that the generated abnormal noise satisfies the non-defective product standard and the measured abnormal noise signal picked up from the device under test By displaying the maximum amplitude value in each frequency band by comparing and calculating the difference for each frequency band, the abnormal sound state of the device under test can be easily and reliably displayed with simple operations. An abnormal noise judgment tool that can judge.

  1 is a conceptual diagram of a configuration of an abnormal sound determination tool according to the present invention.   The external appearance perspective view of the pickup sensor by a condenser microphone is shown.   The arrangement | positioning conceptual diagram of the various parts mounted in an apparatus is shown.   The arrangement | positioning conceptual diagram of a frequency band is shown.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Apparatus 10 Microphone 2 Pickup sensor 11 Piece 3 Amplification circuit 12 Cable 4 A / D conversion circuit 13 Abnormal sound measurement object 5 FFT conversion circuit 6 Storage circuit HDD Hard disk 7 Storage circuit FDD Floppy disk 8 Comparison calculation processing circuit FAN Cooling fan 9 Display Circuit POW power supply
L Spectrum amplitude value
F Evaluation frequency band

Claims (3)

Sensor unit composed of microphone, A / D conversion, Fast Fourier Transform (FFT), storage, calculation processing, display, and composed of multiple frequency bands that subdivide and form the frequency band where abnormal noise should be evaluated, When picking up an abnormal sound signal, A / D conversion and high-speed conversion of the abnormal sound signal picked up by mounting a sensor unit on a standard device where the generated abnormal sound is judged to satisfy the non-defective product standard and the device under test The spectrum is transformed by Fourier transform (FFT), and the maximum value of the amplitude belonging to each frequency band is stored together with the frequency as a judgment reference signal by a standard device and a measured abnormal sound signal by a device to be measured, and continuously several times. When picking up, if a maximum value greater than the value stored immediately before is obtained, the memory is updated to the latest value and In the determination of the number, the difference between the memorized reference signal of the standard device and the measured abnormal sound signal of the device under test is compared and processed for each frequency band, and the result is displayed. An abnormal noise judgment tool that can easily and reliably evaluate the abnormal state of the device under test. In the pickup of an abnormal sound signal, the maximum value of the amplitude belonging to each frequency band and a plurality of amplitude values are stored together with the frequency, and in the determination of the abnormal sound signal, the determination reference signal stored in each frequency band and The abnormal noise determination tool according to claim 1, wherein the difference between the measured abnormal noise signal is subjected to comparative calculation processing and the result is displayed. The abnormal noise determination tool according to claim 1 or 2, wherein the difference between the determination reference signal and the measured abnormal noise signal of the device under test stored in each frequency band is displayed without performing a comparison calculation process.

Images