JP2009025096A - Product inspection device based on operation data by operation of inspection object product, and inspection method of inspection object product - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a product inspection device based on operation data by operation of an inspection object product, and an inspection method of the inspection object product. <P>SOLUTION: Quality determinations are performed from frequency characteristic data of a measured frequency-sound pressure level or corresponding overall trend data or partial overall trend data, by using a plurality of determination methods, respectively, based on operation data generated during an inspection operation time of the inspection object product P, and then the quality determinations are displayed. A determination means 10 preserves the operation data, the frequency characteristic data, the corresponding overall trend data or the partial overall trend data, and the quality determinations, relative to each corresponding inspection object product P. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a product inspection device and an inspection method for a product to be inspected based on operation data from operation of the product to be inspected.

  In general, rotary machines used for air conditioners and the like include compressors, blowers, etc., but these rotary machines may generate abnormal noise during operation due to variations in processing accuracy. Such defective products need to be selected and eliminated during inspection of finished products.

  For this purpose, the factory performs an abnormal noise inspection upon completion. For this abnormal noise inspection, for example, a method has been adopted in which an operator listens to the operating sound generated by operating a compressor and judges the quality. However, since it is difficult to make the above determination unless it is a skilled worker, automation of abnormal noise inspection, that is, in-line and unmanned is being sought.

  As one of the methods for automating abnormal noise inspection, for example, it is conceivable to determine the quality of a product based on a power spectrum (for example, linear display or Log display) obtained by frequency analysis of compressor operation sound data. However, even if it is a non-defective product, it is not easy to simply classify it into a non-defective product and a defective product if the individual difference is large. For example, two non-defective product data expressed by a linear power spectrum have a large feature difference, and classification including these is extremely difficult.

  Also, one of the methods used for abnormal sound inspection is to use an appropriate one based on statistics such as the peak frequency component in the power spectrum obtained by frequency analysis of driving sound data and the average and variance of driving sound data. Some select one and use it as a feature value for discrimination, and perform pass / fail discrimination based on a threshold value of this feature amount.

  However, when the feature quantity is used as a criterion for discrimination, there is a problem that only one aspect of the sound can be expressed, which does not match the human sense of discrimination.

  Therefore, Patent Document 1 proposes an abnormality diagnosis method and apparatus that can determine even the type of abnormality by extracting more information from the waveform data.

  In other words, in this abnormality diagnosis method, waveform data such as vibration or sound generated when the rotating device rotates is detected, the time change of the spectrum of this waveform data is obtained, and the frequency at which the peak is generated from the time change of the spectrum The time interval at which a peak occurs for each frequency feature quantity is obtained as a time feature quantity, and a set of frequency feature quantity-time feature quantity when a rotating device is abnormal is registered in advance as reference data for each cause of abnormality. The presence / absence of abnormality and the type of abnormality are specified from the collation result by collating the frequency feature amount-time feature amount pair obtained from the waveform data during rotation of the rotating device with the reference data.

JP-A-10-274558

However, in order for the vibration or sound generated when the rotating device rotates to be regarded as abnormal noise, it is possible to analyze it if it is a continuous sound generated continuously by driving, that is, a continuous sound. The sudden occurrence of vibration or sound cannot be predicted and the generation time is instantaneous, so that it is difficult to take out instantaneous data.
The present invention has been proposed to overcome the above-described problems, and can analyze even a sudden vibration or sound that occurs suddenly, and automatically performs a wide range of abnormal noise determination. An object of the present invention is to provide a product inspection device and an inspection method for a product to be inspected based on operation data obtained by operating the product to be inspected.

  In order to solve the above-mentioned problem, in the invention according to claim 1, an acquisition means (M) for acquiring operation data generated during the inspection operation of the inspection target product (P), and an inspection based on the acquired operation data Determination means (10) for determining the quality of the target product (P), and the determination means (10) performs waveform analysis from the operation data acquired by the acquisition means (M) and outputs the waveform analysis data And overall trend data, and based on these waveform analysis data or overall trend data, a plurality of different judgment methods are used to make pass / fail judgments and display the pass / fail judgment results respectively. It is characterized by that.

  Thereby, not only continuous operation data but also sudden operation data can be captured and analyzed, and a wide range of abnormal sound determination can be automatically performed.

  Further, in the invention according to claim 2, the determination means (10) converts the operation data acquired by the acquisition means (M) and the waveform analysis data or overall trend data for each corresponding product to be inspected (P). It is characterized in that the result of pass / fail judgment based on is stored.

  Thereby, even after the product is assembled, the inspection target product (P) can be traced and used for later maintenance inspection.

  Moreover, in invention of Claim 3, an acquisition means (M) is arrange | positioned in the indoor space (S) of the product evaluation soundproof room (1) arrange | positioned in the production line (L) of the product (P) to be inspected. Then, in this indoor space (S), the operation data generated during the inspection operation of the inspection target product (P) is acquired, and used for the quality determination of the inspection target product (P) based on the acquired operation data. It is characterized by that.

  As a result, in the indoor space (S) of the product evaluation soundproof room (1) in which sound insulation is ensured, the operation data generated during the inspection operation of the inspection target product (P) is acquired, and the inspection is performed based on the acquired operation data. Since it is used for the quality determination of the target product (P), the quality determination can be performed effectively.

  In the invention according to claim 4, the product to be inspected (P) is operated to acquire operation data based on the operation, and the FFT analysis is continuously performed at a predetermined measurement frequency, sampling number, and time resolution. To obtain frequency characteristic data, obtain the overall value from the frequency characteristic data, display the frequency characteristic data and overall trend data, and make multiple judgments based on these frequency characteristic data or overall trend data The method is characterized in that pass / fail judgment is performed using a technique and the result of pass / fail judgment is displayed.

  Thereby, not only continuous operation data but also sudden operation data can be captured and analyzed, and a wide range of abnormal sound determination can be automatically performed.

  The invention according to claim 5 is characterized in that a determination method for determining pass / fail based on waveform analysis data or overall trend data can be set in accordance with a product to be inspected (P).

  Thereby, it can use for abnormal sound inspection of various inspection object products (P).

  The invention according to claim 6 is characterized in that the result of pass / fail judgment based on operation data and frequency characteristic data or overall trend data is stored for each corresponding product to be inspected (P).

  Thereby, even after the product is assembled, the inspection target product (P) can be traced and used for later maintenance inspection. Further, since the stored data can be reproduced, it is possible to reconstruct the method for the determination together with the feedback of the result of the follow-up survey.

  In the invention according to claim 7, the maximum value in the analysis frequency full range is detected from the frequency characteristic data of the measurement frequency-sound pressure level output by the continuous FFT analysis, and the preset target product is operated. The quality determination is performed by comparing with the standard value of the sound pressure level of the frequency characteristic data based on the operation sound at the time.

  Accordingly, if the peak is large compared to the standard value based on the operation sound and no change is observed in the corresponding overall trend data, it can be determined that the continuous sound is abnormal.

  In the invention according to claim 8, the frequency range is set from the frequency characteristic data of the measurement frequency-sound pressure level output by the continuous FFT analysis, and the maximum value in the section is detected, The quality determination is performed by comparing with the standard value of the sound pressure level of the frequency characteristic data based on the operation sound when the set target product is operated.

  As a result, if the peak is large compared to the standard value based on the operation sound and no change is seen in the corresponding partial overall trend data, it can be determined that the continuous sound is abnormal. .

  In the invention according to claim 9, it is determined whether the average value in the analysis frequency full range is compared with the set predetermined reference value from the frequency characteristic data of the measurement frequency-sound pressure level output by the continuous FFT analysis. Judgment is performed.

  As a result, if the difference between the average value of the sound pressure level in the full analysis frequency range and its maximum value exceeds a predetermined value and no change is found in the corresponding overall trend data, abnormal sounds are generated in the continuous sound. Can be determined as being.

  In the invention according to claim 10, the frequency range is set from the frequency characteristic data of the measurement frequency-sound pressure level output by the continuous FFT analysis, and the average value in the section is detected and set. The quality determination is performed by comparing with a predetermined reference value.

  As a result, if the difference between the average value of the sound pressure level within the frequency range and the maximum value exceeds the specified value and no change is found in the corresponding partial overall trend data, abnormal sounds are generated in the continuous sound. Can be determined as being.

  In the invention described in claim 11, the frequency range is set from the frequency characteristic data of the measurement frequency-sound pressure level output by the continuous FFT analysis, and three lines from the start frequency and three lines before the end frequency are set. A difference between the average value and the maximum value in the section is calculated, and compared with a reference value having a predetermined size, and quality determination is performed.

  Thus, in the specific frequency band, the difference between the average value of the three lines from the start frequency and the three lines before the end frequency and the maximum value in the section is calculated, and the average value in the section is detected, and the frequency is detected. If the difference from the maximum value in the band exceeds a predetermined value and there is no change in the corresponding partial overall trend data, it can be determined that an abnormal sound has occurred in the continuous sound.

  The invention according to claim 12 is characterized in that a pass / fail judgment is made by comparing with an allowable value of an overall value using overall corresponding to frequency characteristic data of a measured frequency-sound pressure level output by continuous FFT analysis. To do.

  Thereby, it can be determined that abnormal noise has occurred by detecting the sudden sound by looking at the presence of the peak value of the overall value.

  According to the invention of claim 13, a frequency range is set from the frequency characteristic data of the measurement frequency-sound pressure level output by the continuous FFT analysis, and the overall value is obtained using the partial overall in the section from the start frequency to the end frequency. It is characterized in that a pass / fail judgment is made by comparing with an allowable value.

  Thereby, it can be determined that abnormal noise has occurred by detecting the sudden sound by looking at the presence of the peak value of the overall value at the specific time.

  In the invention of claim 14, the overall value corresponding to the frequency characteristic data of the measurement frequency-sound pressure level output by the continuous FFT analysis is used, and the overall value at the specific time on the time axis and The difference between the average value of the overall values in the time period and the difference threshold value (allowable difference value) is compared with the allowable value of the overall value at the same time, and quality determination is performed.

  Thereby, sudden sound generation can be detected more reliably, and it can be determined that an abnormal sound has occurred.

  Furthermore, in the invention of claim 15, the frequency range is set from the frequency characteristic data of the measurement frequency-sound pressure level output by the continuous FFT analysis, and the partial overall in the section from the start frequency to the end frequency is used. The difference between the overall value at the specified time on the axis and the average value of the overall value in the time period before the specified time is compared with the difference threshold (allowable difference value), and at the same time, the overall value is allowed. It is characterized in that the quality is determined by comparing with the value.

  As a result, sudden sound at a specific time can be reliably detected, and it can be determined that an abnormal sound has occurred.

  In addition, the code | symbol in the bracket | parenthesis of each said means is an example which shows a corresponding relationship with the specific means as described in embodiment mentioned later.

In FIG. 1, an example of the product evaluation soundproof room 1 installed in the product production line L is shown.
The product evaluation soundproof room 1 is equipped with an inspection device (to be described later) that performs abnormal sound quality determination as an inspection method for the inspection target product based on the operation data of the operation of the inspection target product P according to the present invention.
The inspection target product P here is, for example, an in-vehicle unit of a vehicle air conditioner.
Further, as shown in FIG. 2, the product production line L has an assembly zone Z1 for assembling each part of the in-vehicle unit of the vehicle air conditioner that is the inspection target product P, and an inspection for inspecting the in-vehicle unit that has been assembled. It is constructed by zone Z2. The product evaluation soundproof room 1 is installed in the inspection zone Z2.

The product evaluation soundproof room 1 is constructed of a partition wall 2 made of a known method, a sound absorbing material or a sound insulation material (for example, background noise 30 dB), and has a substantially rectangular room S that allows an inspector to enter the room. Is secured.
The soundproof room 1 is provided with a door for an inspector to enter and exit, although not shown.
Furthermore, the entrance shutter 3 and the exit shutter 4 are respectively provided on the mutually opposing side walls of the soundproof room 1, and a conveyor to be described later passes therethrough so that the inspection target product P can be carried in and out. In this case, the entrance shutter 3 and the exit shutter 4 are doors that automatically open and close to the left and right in conjunction with the movement of the product P to be inspected, and are made of a sound absorbing material or a sound insulating material, like the partition wall 2. Thereby, the sound insulation property which suppresses the penetration | invasion of the sound from the outdoor is ensured.

Since the soundproof room 1 is an example as described above, the size of the soundproof room 1 can be changed according to the product P to be inspected, and the space in the soundproof room 1 is not necessarily accessible to inspectors. It does not have to be a size.
Further, the shape of the soundproof room 1 may not be linear from the entrance shutter 3 to the exit shutter 4. For example, a right-angled shape or a U-turn arrangement is possible. In the case of a U-turn arrangement, it is possible to share the entrance / exit shutter and provide a single location.

In addition, in the product evaluation soundproof room 1 as described above, the inspection target product P is individually placed on the product transport pallet 6 by the conveyor 5 provided along the product production line L, and enters the soundproof room 1. It comes in and out.
Although not shown in the drawings, a product connection connector for connecting to the inspection target product P is provided on the product transport pallet 6, and a fixed-side automatic connection unit 7 connected to the product connection connector is mounted. Yes. On the other hand, a movable automatic connection unit 8 is disposed on the conveyor 5 side. The movable automatic connection unit 8 is equipped with a plurality of drive / control information transmission terminals (not shown). The movable automatic connection unit 8 is connected to an automatic connection installed in the soundproof room 1 and a control command means 9 for sending a drive command and a control command for the inspection target product P.

With the configuration described above, when the inspection target product P is placed on the product transport pallet 6 and transported to the inspection position, the inspection target product P is connected to the inspection target product P by connecting the product connection connector 7 thereto. Is transferred to the inspection position in the soundproof room 1, the drive / control information transmission terminal in the movable automatic connection unit 8 arranged on the side of the conveyor 5 is connected to the fixed automatic connection unit 7 on the product transfer pallet 6. In response to the automatic connection command of the control command means 9, the drive / control information receiving terminal can be automatically connected by means not shown.
In this state, a drive command and a control command are transmitted from the control command means 9 to the inspection target product P to operate the inspection target product P, and an abnormal sound indicating an abnormal operation is applied from the operation sound to the inspection apparatus described later. Data is extracted and used for noise detection.
The control command means 9 can be a known personal computer.

Above the inspection position on the conveyor 5 in the soundproof room 1, microphones M and M for acquiring operation sound, which is operation data, generated when the inspection target product P is operated are suspended. Note that the positions of the movable parts (for example, motors, links, blowers, etc.) mounted on the inspection target product P may differ depending on the type of the final product to be incorporated. It is arranged to correct the operating sound acquisition conditions.
The microphones M and M are connected to a determination unit 10 that constitutes an inspection apparatus that performs data processing, storage, and pass / fail determination based on operation sounds acquired by the microphones M and M.

Note that the operation sound obtained when the inspection target product P is operated, which is the operation data acquired by the microphones M and M, may include the following continuous sounds and sudden sounds as abnormal sounds.
These continuous sounds and sudden sounds are based on operation sounds and state sounds from mounted movable parts (for example, motors, links, blowers, etc.).
Examples of continuous sounds include so-called magnetic noise, fan low / high rotation noise, metal mixing noise, metal noise, door seal wind leakage noise, noise due to component damage, siren noise, fitting wind noise, Examples include rustling sounds and door self-excited vibration noise.
Examples of the sudden sound (instantaneous sound) include a so-called link click sound, an abnormal sound of film door operation, an elastomer door noise, a servo motor noise, and a gear scratch sound.

  As the determination means 10, as a well-known system, abnormal sound quality determination software for acquiring operation sounds as operation data from the microphones M and M as acquisition means and processing the data and executing various determination procedures is provided. You can use an installed PC.

Therefore, the determination unit 10 will be described.
The determination means 10 performs waveform analysis from the operation data acquired by the microphones M and M, displays the operation data and the output waveform analysis data and overall trend data, and also displays the waveform analysis data or overall data. Based on the trend data, a plurality of different determination methods (described later) are used to perform pass / fail judgments and display the pass / fail judgment results.

That is, as a signal processing unit, the determination unit 10 uses an operation sound (waveform data) acquired by the microphones M and M as a continuous FFT (Fast) in a predetermined frequency range (hereinafter, measurement frequency), the number of samplings, and time resolution. An FFT processing unit 11 that performs Fourier Transform analysis and outputs frequency characteristic data (power spectrum) of the measurement frequency-sound pressure level, and obtains an overall value from the frequency characteristic data, thereby obtaining an overall time-sound pressure level (partial overall) OA conversion section 12 that outputs trend data of the above-mentioned) and display means 13 for displaying waveform data acquired by the microphones M and M, frequency characteristic data, and overall trend data (FIG. 3). reference).
As a continuous FFT analysis, as a result of examination of optimum conditions for continuous FFT, the measurement frequency is set to 5 kH, the number of samplings is 1024, the time axis resolution is 80 ms, and the frequency resolution is 12.5 Hz. This is because the frequency and time-axis resolution are inversely related, so the characteristics are rough when priority is given to the time-axis resolution, and the analysis time is delayed if priority is given to the frequency resolution. Because it is necessary to find out.

  The determination unit 10 is equipped with a storage unit 14 that stores a plurality of different determination procedures for determining pass / fail based on the frequency characteristic data or the overall trend data.

Here, when abnormal sound is included in the operation data acquired by the microphones M, M, the frequency characteristic data of the measurement frequency-sound pressure level to be output, as well as the corresponding overtrend data, as follows: It is known that changes are seen in the partial overtrend data.
For example, when the continuous sound includes abnormal sound (for example, magnetic sound, door seal portion wind leakage abnormal sound), the operation data is subjected to continuous FFT analysis, and the frequency characteristic data (FFT waveform) of the output measurement frequency-sound pressure level. 4), a steep waveform appears in the specific frequency band as shown in FIGS. In these cases, no change is observed in the corresponding overtrend data and partial overtrend data (OA and POA waveforms).
Therefore, in such a case, since the peak difference is large in the specific frequency band or the sound pressure level is high in the specific frequency band, it is only necessary to grasp that there is an abnormal sound in the continuous sound.

On the other hand, when an unusual sound includes sudden sound (for example, a link sound, an elastomer sound), the FFT waveform has a certain frequency band and an abnormal sound as shown in FIGS. Appears with a large difference in sound pressure levels when none. In the corresponding OA and POA waveforms, a steep waveform appears in a specific frequency band.
Therefore, in such a case, the sound pressure level in the specific frequency band is high, the peak difference of OA (POA) is large, or the OA (POA) value is large, so that it is grasped that abnormal sound exists in the sudden sound. You can do it.

From the above, the determination means 10 stores the following determination procedure.
That is, using the trend data of the frequency characteristic data (power spectrum) of the measured frequency-sound pressure level or the overall of time-sound pressure level (partial overall) derived from the frequency characteristic data, (1) maximum, that is, the maximum value (2) Average peak, that is, determination procedure using average value, (3) Determination procedure using difference peak (difference between average value and maximum value), (4) Time peak (time axis) (5) Time difference peak (Overall value at specific time on the time axis and average value of overall values in the time-time range before the specific time) A determination procedure using (difference) is set.

(1) In the determination procedure using the maximum, that is, the maximum value, the operation sound (waveform data) acquired by the microphones M and M is subjected to continuous FFT analysis at a predetermined measurement frequency, sampling number, and time resolution, The maximum value in the analysis frequency full range is detected from the frequency characteristic data output here, and the quality is determined by comparing with a preset threshold value. Note that the threshold value refers to a standard value (allowable value) of sound pressure level of frequency characteristic data based on operation sound when the target product is operated, which is known in advance.
Alternatively, a frequency range may be set, a maximum value in the section may be detected, and pass / fail judgment may be performed by comparison with a threshold value (see FIG. 8).

(2) In the determination procedure using the average peak, that is, the average value, the pass / fail determination is made by using the average value in the analysis frequency full range from the frequency characteristic data output by the continuous FFT analysis and comparing it with the set threshold value. I do.
Alternatively, a pass / fail judgment may be made by setting a frequency range, detecting an average value in the section, and comparing with a threshold value. (See FIG. 9). Here, the threshold value is a reference value of a magnitude that an abnormal sound is present if a predetermined value is exceeded compared to an average value within the frequency range.

(3) In the determination procedure using the difference peak, the frequency range of the obtained operation sound (waveform data) is set from the frequency characteristic data of the measurement frequency-sound pressure level output by the continuous FFT analysis, The difference between the average value of the three lines from the start frequency and the three lines before the end frequency and the maximum value in the section is calculated and compared with a threshold value to determine pass / fail (see FIG. 10). Here, the threshold value is a reference value of a magnitude that an abnormal sound is present if a predetermined value is exceeded compared to an average value within the frequency range.
Such a determination procedure is effective when a frequency range in which abnormal noise exists is predicted in advance.

(4) In the determination procedure using the time peak, a frequency range is set, and a threshold value (overall value of the overall value) is set using a partial sum (partial overall, see FIG. 11) or overall in a section from the start frequency to the end frequency. And pass / fail judgment is made. That is, in this determination procedure, the existence of a peak value of the overall value at a specific time is observed.
This determination procedure is effective for detecting sudden sound.

(5) In the case of a determination procedure using a time difference peak, a frequency range is set, and a partial sum (partial overall) or overall in a section from the start frequency to the end frequency is used before the current value (5 to 5). The difference value between the average value of (15 data) and the current value is compared with a difference threshold value (allowable difference value) to determine whether or not it is acceptable.
At the same time, a frequency range is set and compared with a threshold value using a partial sum (partial overall) or overall in a section from the start frequency to the end frequency, and a pass / fail judgment is also performed (see FIG. 12). . In FIG. 12, for convenience, the threshold value (allowable value for the overall value) and the difference threshold value (allowable difference value) are shown at the same level.

Then, as shown in FIG. 13, the display means 13 in the judging means 10 executes the above-described pass / fail judgment procedure, thereby obtaining waveform data, frequency characteristic data, and overall trend data acquired by the microphones M and M. In addition, the date, the product number, the process number, and the serial number are displayed as the product number information of the product P to be inspected, the abnormal sound quality determination result (OK, NG), and the data NO at the time of data reproduction. .
In the product number information, serial number, etc., specification, vehicle type, charge, and inspection date are displayed, and by storing these information together with the measurement data, it can be used to specify the product P to be inspected. Can be used for follow-up surveys.

The abnormal sound quality determination apparatus according to the present invention is configured as described above, and the operation thereof will be described next.
The inspection target product P flows through the assembly zone Z1 in the product production line L and is completed by assembling the component parts, and the inspection target product P is placed on the product transport pallet 6 by the conveyor 5 in the inspection zone Z2. Is sent into the soundproof room 1 (see FIG. 2).
When the inspection target product P is placed on the product transport pallet 6, a product connection connector is connected to the inspection target product P.

  When the product P to be inspected is carried into the indoor space S of the soundproof room 1 and the entrance is closed by the entrance shutter 3, the inside and outside of the soundproof room 1 are shielded by the sound absorbing material, the wall composed of the sound insulation material, and the entrance shutter 3. Intrusion of sound into the soundproof room 1 can be suppressed to ensure quietness, and an abnormal sound detection system is established. In this case, since quietness of about 30 dB can be ensured in the indoor space S with background noise, small abnormal noise can be captured with the second microphones M and M together with the operation sound (see FIG. 1).

When the product transport pallet 6 moves on the conveyor 5 and the inspection target product P is brought to the inspection position in the soundproof room 1, the movable automatic connection unit 8 disposed on the side of the conveyor 5 and the product transport pallet 6 are The fixed-side automatic connection unit 7 can be automatically connected by an automatic connection command of the control command means 9. For this reason, it is not necessary for the inspector to enter the indoor space S of the soundproof room 1 and perform connection work.
The control command means 9 can send drive and control information to the inspection target product P to operate the product P (motor, link, blower, etc.).

When drive and control information is given to the inspection target product P and the inspection target product P is operated, operation noise and vibration are generated by operating the motor, link, and blower which are movable elements. At this time, the operation sound (waveform data) collected by the microphones M and M is a time axis waveform and includes the operation sounds of all the movable elements, and the waveform data is also superimposed.
The operation sound generated when the motor, the link, and the blower, which are movable elements, operate includes a continuous sound and a sudden sound (instantaneous sound).

Here, in inspecting the inspection target product P in the determination means 10, the control command means 9 applies the inspection target product P to the inspection target product P in order to inspect abnormal noise by focusing on specific items as follows. The product P can be operated by transmitting drive and control information (motor, link, blower, etc.).
For example: In order to inspect the servo motor and door operation noise (step 1), the control command means 9 outputs a reciprocating operation command to the servo motor with the blower motor stopped for 60 seconds for the product P to be inspected. The operation data is obtained from the inspection target product P.
In addition, 2. In order to inspect the blowing noise (step 2), the blower motor is operated for 20 seconds from the control command means 9 to the inspection target product P, while the one-way operation command is output to the servo motor, and the inspection target product P Get the operation data from
Furthermore, 3. In order to inspect the magnetic sound (step 3), the blower motor is operated for 10 seconds from the control command means 9 to the inspection target product P, while the stop instruction is output to the servo motor to operate from the inspection target product P. Get data.
Note that these inspection items can be narrowed down by clicking various condition setting display sections on the display screen of the display means 13 (see FIG. 13) with a mouse or the like.

  With the command input of the inspection items as described above, in the determination means 10, the operation sound (waveform data) collected by the microphones M, M is converted into a predetermined measurement frequency (5 kH), sampling number (1024), and time resolution ( 80 ms), the FFT processing unit 11 performs continuous FFT analysis and outputs frequency characteristic data (power spectrum) of the measured frequency-sound pressure level, and the OA converter 12 determines the overall time-sound pressure level from the frequency characteristic data. The trend data of (partial overall) is output, and the frequency characteristic data of the measurement frequency-sound pressure level and the trend data of overall (partial overall) are displayed on the display means 13 together with the waveform data (see FIG. 13), By performing multiple judgment procedures in parallel, various It is possible to perform quality determination by capturing the abnormal sound.

  The determination means 10 uses (1) the maximum, that is, the frequency characteristic data (power spectrum) of the measurement frequency-sound pressure level and the trend data of the overall of the time-sound pressure level (partial overall) from the frequency characteristic data. Determination procedure using the maximum value, (2) Average peak, that is, determination procedure using the average value, (3) Determination procedure using the difference peak, (4) Determination procedure using the time peak, (5) Time difference A determination procedure using peaks can be executed.

(1) In the determination procedure using the maximum, that is, the maximum value, the operation sound (waveform data) acquired by the microphones M, M is converted into a predetermined measurement frequency (5 kH), the number of samplings (1024), and the time resolution (80 ms). ), The FFT processing unit 11 performs continuous FFT analysis, detects the maximum value in the analysis frequency full range from the frequency characteristic data output here, and compares the threshold value with a preset threshold value to determine pass / fail.
In addition, a frequency range is set, the maximum value in the section is detected, and pass / fail judgment is performed by comparison with a threshold value (see FIG. 8).
In the frequency characteristic data (power spectrum) of the measurement frequency-sound pressure level used in this determination procedure, for example, as shown in FIGS. 4 and 5, a peak is seen in the sound pressure level in a specific frequency band, and the threshold is reached. Compared with the value (previously known standard value (allowable value) of the sound pressure level of the frequency characteristic data based on the operation sound when the target product is operated), the peak is large, and the corresponding overall trend data If there is no change in the partial overall trend data, it can be determined that continuous sound (magnetic sound, door seal part wind noise) is generated, and it can be output as NG and displayed on the display means 13.

(2) In the determination procedure using the average peak, that is, the average value, the pass / fail determination is made by using the average value in the analysis frequency full range from the frequency characteristic data output by the continuous FFT analysis and comparing it with the set threshold value. I do.
At that time, a frequency range is set, an average value in the section is detected, and whether the difference from the maximum value in the frequency band exceeds a predetermined value is determined.
In this determination procedure, the difference between the average value of the sound pressure level in a specific frequency band and the maximum value in that frequency band exceeds a predetermined value, and there is no change in the corresponding overall trend data or partial overall trend data. For example, it can be determined that a continuous sound (magnetic noise, door seal portion wind leakage abnormal sound) is generated, output as NG, and displayed on the display means 13.

(3) In the determination procedure using the differential peak, the frequency range is set from the frequency characteristic data of the measurement frequency-sound pressure level, the average value of the three lines from the start frequency and the three lines before the end frequency and the interval The difference between the maximum value and the average value in the section is calculated and the average value in the section is detected, and whether or not the difference with the maximum value in the frequency band exceeds a predetermined value is determined.
According to this determination procedure, in a specific frequency band, the difference between the average value of the three lines from the start frequency and the three lines before the end frequency and the maximum value in the section is calculated, and the average value in the section is detected. If the difference from the maximum value in the frequency band exceeds the specified value, and there is no change in the corresponding partial overall trend data, a continuous sound (magnetic sound, abnormal noise from the door seal part) is generated. It can be determined that the image has been displayed, and it can be output as NG and displayed on the display means 13.

(4) In the determination procedure using the time peak, a frequency range is set, and a pass / fail determination is performed by comparing with a threshold value using a partial sum (partial overall) or overall in a section from the start frequency to the end frequency. .
In this case, the frequency characteristic data of the measured frequency-sound pressure level has a large sound pressure level over the whole, and in the trend data of partial overall or overall, a waveform with a steep shape appears in various places. Compared with the maximum value of this value and a preset threshold value (allowable value for overall value), sudden abnormal noises such as link tune noise, film door operation noise, and elastomer noise are generated. Can be determined and output as NG and displayed on the display means.

further,
(5) In the case of a determination procedure using a time difference peak, a frequency range is set, and a partial sum (partial overall) or overall in a section from the start frequency to the end frequency is used before the current value (5 to 5). The difference value with the average value of (15 data) is compared with the difference threshold value, and the quality is determined.
At the same time, a frequency range is set, and a partial sum (partial overall) or overall in the section from the start frequency to the end frequency is compared with a threshold value (overall value allowable value), and pass / fail judgment is also performed. To do.
Also by such a determination procedure, it can be determined that a sudden abnormal noise such as a link sound is generated.

In addition to displaying the waveform data acquired by the microphones M and M, the frequency characteristic data obtained by the FFT processing unit 11, and the overall trend data on the display unit 13 by executing the above-described determination procedure. The date, product number, process number, and serial number can be displayed as the product number information of the product P to be inspected, the abnormal sound quality determination result (OK, NG), and the data NO at the time of data reproduction.
In the product number information, serial number, etc., specification, vehicle type, charge, and inspection date are displayed, and by storing these information together with the measurement data, it can be used to specify the product P to be inspected. Can be used for follow-up.

  In addition, when storing the measurement data together with information such as serial number, specifications, vehicle type, charge, inspection date, etc. in the product number information, a new problem arises as a result of the follow-up survey of the inspection target product P The stored data can be reproduced, and the judgment conditions (threshold value, etc.) can be reviewed by feedback of the results of the follow-up survey to provide a better judgment procedure.

  As described above, in the present invention, the operation sound (waveform data) collected by the microphones M and M is subjected to the FFT processing unit 11 at a predetermined measurement frequency (5 kH), the number of samplings (1024), and the time resolution (80 ms). A continuous FFT analysis is performed to output frequency characteristic data (power spectrum) of the measured frequency-sound pressure level, and the OA converter 12 obtains trend data of time-sound pressure level overall (partial overall) from the frequency characteristic data. By outputting and executing a plurality of different determination procedures in parallel, it is possible to perform pass / fail determination by capturing various abnormal sounds.

As described above, the example in which the inspection apparatus for performing the abnormal sound quality determination according to the present invention is used in the product evaluation soundproof room 1 has been described.
At that time, as an example of the inspection target product P, an in-vehicle unit of a vehicle air conditioner has been described and described, but other operation units are of course possible. At this time, since the movable element changes and the operation sound associated therewith is different, the determination means 10 is set in advance with various inspection conditions such as a threshold value.

Further, the present inspection apparatus can be used not only for abnormal sound inspection but also for other product evaluation processes such as operation of product P and functional inspection.
In this case, the operation data output from the inspection target product P is not limited to the operation sound, and an acceleration sensor, a vibration meter, or a noise meter can be used as the acquisition unit 13 so that vibration can be captured.

At that time, in addition to the continuous FFT analysis, for example, other wavelet analysis or analysis by Wigner distribution can be considered.
This Wigner distribution is not directly subjected to the complementary constraint of time resolution-frequency resolution, so that a good time-frequency resolution of the power spectrum can be obtained on the frequency-time plane.

It is typical structure explanatory drawing which shows an example of the product evaluation soundproof room which has arrange | positioned the product inspection apparatus concerning this invention. It is the typical top view which showed an example of the production line which has arrange | positioned the product evaluation soundproof room shown in FIG. It is a block diagram of the product inspection apparatus concerning this invention. An example of FFT waveform data and corresponding OA and POA waveform data when abnormal sounds are included in the continuous sound in the operation data of the product to be inspected shown in the display means in the product inspection apparatus according to the present invention is shown. It is a graph. Another example of FFT waveform data and corresponding OA and POA waveform data when abnormal sounds are included in the continuous sound in the operation data of the product to be inspected shown in the display means in the product inspection apparatus according to the present invention. It is a graph to show. It is a graph which shows an example of FFT waveform data and corresponding OA, POA waveform data in case the sudden sound in the operation data of the product to be inspected is shown in the display means in the product inspection apparatus according to the present invention. It is a graph which shows another example of FFT waveform data and corresponding OA, POA waveform data when the sudden sound in the operation data of the product to be inspected is shown in the display means in the product inspection apparatus according to the present invention. . It is a graph with which it uses for description of the determination procedure using the maximum value among the quality determination procedures in the product inspection apparatus concerning this invention. It is a graph with which it uses for description of the determination procedure using the average value among the quality determination procedures in the product inspection apparatus concerning this invention. It is a graph with which it uses for description of the determination procedure by a difference peak among the quality determination procedures in the product inspection apparatus concerning this invention. It is a graph with which it uses for description of the determination procedure by a time peak among the quality determination procedures in the product inspection apparatus concerning this invention. It is a graph with which it uses for description of the determination procedure by a time difference peak among the quality determination procedures in the product inspection apparatus concerning this invention. It is a figure which shows an example of the display screen displayed on the display means in the product inspection apparatus concerning this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Product evaluation soundproof room 2 Partition wall 3 Entrance shutter 4 Exit shutter 5 Conveyor 6 Product conveyance pallet 7 Fixed side automatic connection unit 8 Movable automatic connection unit 9 Control command means 10 Determination means 11 FFT processing part 12 OA conversion part 13 Display means P Product to be inspected S Indoor space R Transport route M Microphone

Claims (15)

Acquisition means (M) for acquiring operation data generated during the inspection operation of the inspection target product (P);
A determination means (10) for determining the quality of the inspection target product (P) based on the acquired operation data;
The determination means (10) performs waveform analysis from the operation data acquired by the acquisition means (M), displays the output waveform analysis data and overall trend data, and also displays the waveform analysis data or overall data. Product inspection based on operation data from the operation of the product to be inspected, characterized in that it uses a plurality of different determination methods based on the trend data and each displays pass / fail results. apparatus.   The determination unit (10) stores the result of pass / fail determination based on the operation data acquired by the acquisition unit (M) and the waveform analysis data or overall trend data for each corresponding inspection target product (P). The product inspection device based on operation data obtained by operating the product to be inspected according to claim 1, wherein   The acquisition means (M) is arranged in the indoor space (S) of the product evaluation soundproof room (1) arranged in the production line (L) of the inspection object product (P), and this indoor space (S) The operation data generated during the inspection operation of the inspection target product (P) is acquired and used for the quality determination of the inspection target product (P) based on the acquired operation data. The product inspection apparatus based on the operation data by the driving | operation of the inspection object product of Claim 1 or 2.   Operate the product to be inspected (P), acquire operation data based on the operation, perform continuous FFT analysis at a predetermined measurement frequency, number of samplings, and time resolution to obtain frequency characteristic data and frequency characteristics The overall value is obtained from the data, the frequency characteristic data and the overall trend data are displayed, and based on the frequency characteristic data or the overall trend data, a plurality of different judgment methods are used to perform pass / fail judgments. A method for inspecting a product to be inspected, wherein the result of pass / fail judgment is displayed.   5. The inspection of the inspection target product according to claim 4, wherein a determination method for determining pass / fail based on waveform analysis data or overall trend data can be set according to the inspection target product (P). Method.   6. The product to be inspected according to claim 4 or 5, wherein for each corresponding product to be inspected (P), the result of pass / fail judgment based on the operation data and the frequency characteristic data or overall trend data is stored. Inspection method.   The maximum value in the analysis frequency full range is detected from the frequency characteristic data of the measurement frequency-sound pressure level output by the continuous FFT analysis, and the frequency characteristic data based on the operating sound when the preset target product is operated is detected. 7. The inspection method for a product to be inspected according to any one of claims 4 to 6, wherein the quality determination is performed in comparison with a standard value of the sound pressure level.   When the frequency range is set from the frequency characteristic data of the measurement frequency-sound pressure level output by the continuous FFT analysis, the maximum value in the section is detected, and the preset target product is operated 7. The inspection method for a product to be inspected according to any one of claims 4 to 6, wherein the quality is determined by comparing with a standard value of the sound pressure level of the frequency characteristic data based on the operation sound.   The pass / fail judgment is performed by comparing an average value in a full analysis frequency range with a set predetermined reference value from frequency characteristic data of a measurement frequency-sound pressure level output by the continuous FFT analysis. The inspection method for a product to be inspected according to any one of 4 to 6.   From the frequency characteristic data of the measured frequency-sound pressure level output by the continuous FFT analysis, a frequency range is set, an average value in the section is detected, and compared with a set predetermined reference value. 7. The inspection method for a product to be inspected according to any one of claims 4 to 6, wherein quality determination is performed.   The frequency range is set from the frequency characteristic data of the measurement frequency-sound pressure level output by the continuous FFT analysis, and the average value of the three lines from the start frequency and the three lines before the end frequency and the maximum value in the section 7. The inspection method for a product to be inspected according to any one of claims 4 to 6, wherein the difference is calculated, compared with a reference value having a predetermined size, and quality is determined.   7. The pass / fail determination is performed by comparing with an allowable value of an overall value using an overall corresponding to frequency characteristic data of a measurement frequency-sound pressure level output by the continuous FFT analysis. Among them, the inspection method for the inspection target product according to any one of the above.   The frequency range is set from the frequency characteristic data of the measurement frequency-sound pressure level output by the continuous FFT analysis, and compared with the allowable value of the overall value using the partial overall in the section from the start frequency to the end frequency. The method for inspecting a product to be inspected according to any one of claims 4 to 6, wherein the determination is performed.   Using the overall corresponding to the frequency characteristic data of the measurement frequency-sound pressure level output by the continuous FFT analysis, the overall value at the specific time on the time axis and the overall value in the time time range before the specific time The difference between the average value and the difference threshold value (allowable difference value) is compared with the allowable value of the overall value at the same time, and the pass / fail judgment is performed. An inspection method for a product to be inspected according to any one of the above.   The frequency range is set from the frequency characteristic data of the measurement frequency-sound pressure level output by the continuous FFT analysis, and the overall at the specific time on the time axis is set using the partial overall in the section from the start frequency to the end frequency. The difference between the value and the average value of the overall value in the time period before the specified time is compared with the difference threshold (allowable difference value), and at the same time, compared with the allowable value of the overall value, the pass / fail judgment The inspection method for a product to be inspected according to any one of claims 4 to 6, wherein the inspection method is performed.

Images