JP2006177711A - Method and apparatus of inspecting pivot allophone of engine valve gear structure - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a pivot allophone inspection technique, with which the presence of a pivot allophone in an engine valve gear structure, having a pivot-type rocker arm, can be evaluated from an engine sound. <P>SOLUTION: The pivot allophone inspection device of the engine valve gear structure having the pivot pe rocker arm. This device includes an FFT unit 42, which takes out the attention signal of a pivot sound frequency band, including a pivot sound from an engine sound signal acquired through a microphone 3; an operation part 43 which sequentially measures multiple times instrumentation one by one, by making occurrences in predetermined time spacing of an amplitude value more than a threshold or higher in the attention signal into an allophone parameter and calculates the amount of statistical descriptions of a plurality of measured allophone parameters; and an evaluating section 45 which performs pivot priority determining, based on the comparison of the amount of the statistical description with a criterion for evaluation. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a method and apparatus for inspecting a pivot noise of an engine valve structure having a pivot type rocker arm.

  In the automotive engine test process, the signal based on the operating sound emitted from the engine during operation is separated into signals of a predetermined frequency band, and the presence or absence of abnormal noise is determined based on the separated signals, and the final engine A technique for performing pass / fail evaluation is known, and one of them is an average value detector that averages signals in a predetermined frequency band every predetermined time to obtain an instantaneous signal voltage, and is averaged every predetermined time. The peak value detector that detects a high voltage value protruding from the signal waveform formed by the average value of the signal voltage, and the average of the signal waveform formed by the signal voltage averaged every predetermined time to average the entire waveform An averaging circuit that obtains a voltage value, a peak frequency detector that compares the output signal from the average value detector and the output signal from the averaging circuit, and obtains the appearance frequency of the protruding signal, and an operation discriminator Provided, Is calculated as an abnormal noise factor from the peak value and the average value based on the output signals from the average value detector and the peak value detection unit, and abnormal when the abnormal noise factor exceeds a set threshold There is an abnormal sound discriminating device that discriminates that there is an operating noise and calculates the peak frequency per rotation as the peak occurrence frequency based on the output signals from the tachometer and the peak frequency detector (for example, Patent Documents) 1).

  Furthermore, the engine mounted on the test bench is started, and the engine sound from this engine and the engine speed are used to determine the quality of the engine sound based on the determination criteria. And input the rotation speed detection signal of the tachometer, analyze the frequency under the condition of constant rotation speed, detect the engine sound exceeding the threshold, calculate by weighting by frequency, and determine whether the engine sound is good or bad based on the judgment criteria There is also known an engine sound quality determination device that determines whether or not (see, for example, Patent Document 2).

Sho 61-234332 (page 3-5, Fig. 1) No. 04-242138 (page 2-3, FIG. 2)

The conventional engine pass / fail judgment device evaluates the entire engine instead of evaluating a specific part of the engine from the engine sound, converts the engine sound as a temporal waveform signal into the frequency domain, It seems that the objective was achieved only by evaluating the calculated values obtained by performing arithmetic operations such as averaging and integration on each signal value. If the valve structure is being examined, such a method will not give good results. In other words, if the engine valve structure with a pivot-type rocker arm does not use a lash adjuster that automatically adjusts the valve clearance, a pivot sound is generated from the pivot, but this sound is irregularly generated in a specific frequency range. Because it is a discontinuous sound that occurs, it is not possible to discriminate the pivot noise among the pivot sounds by simply calculating the engine sound signal expressed in the frequency domain directly and comparing it with the threshold value. It is.
In view of the above situation, an object of the present invention is to provide a pivot noise inspection technique capable of evaluating the presence of a pivot noise in an engine valve structure having a pivot type rocker arm from the engine sound.

  In order to solve the above-mentioned problem, in the pivot abnormal sound inspection device of the engine valve structure having a pivot type rocker arm according to the present invention, the attention signal in the pivot sound frequency band including the pivot sound from the engine sound signal acquired through the microphone. A spectrum decomposing unit for taking out noise, and measuring the number of occurrences of an amplitude value equal to or greater than a threshold value in the signal of interest at a predetermined time interval sequentially as an abnormal noise parameter and a statistical feature amount of the measured abnormal noise parameter And an evaluation unit for determining whether the pivot is good or not based on a comparison between the statistical feature quantity and the evaluation criterion.

  According to this configuration, a signal (herein called a signal of interest) of a pivot sound frequency band, which is derived from an experiment in advance and output from a spectrum decomposition unit (generally using an FFT device), is directly evaluated. Rather, the signal of interest is sampled sequentially at predetermined time intervals, and the number of occurrences (appearances) of amplitude values above the threshold (preset) on the sampling time axis is measured as an abnormal noise parameter. Pivot quality is evaluated by comparing a statistical feature obtained by performing statistical computation on the parameter group with an evaluation standard determined by experimental and empirical consideration. Rather than immediately evaluating a signal in a specific frequency band as in the prior art, the attention signal in the pivot sound frequency band is further expanded on the time axis to measure an abnormal sound parameter group, and this abnormal sound parameter group is characterized. By performing final evaluation based on this, it has become possible to perform pivot abnormal sound inspection, which is very difficult to judge whether it is good or bad, with high reliability.

  According to an experimental study by the inventor of the present application, the average value of a plurality of measured abnormal noise parameters is found to be effective as a statistical feature quantity of the abnormal noise parameters that determine pass / fail judgment in pivot abnormal noise inspection. Yes. However, depending on subtle differences in valve structure and engine test conditions, the sum of multiple abnormal parameters may be used as the statistical feature amount, or the number of occurrences of abnormal noise parameters that are greater than or equal to a predetermined value. This is also possible with advantages. In addition, when there is a margin in the arithmetic processing, it is also possible to obtain a statistical feature amount from a variance analysis or a histogram analysis of the obtained abnormal sound parameter group.

  In the pivot abnormal sound inspection apparatus according to the present invention, the selection of the frequency band of the pivot sound is an important element for performing the pass / fail judgment with high reliability. However, according to the examination based on the experiment by the present inventor, the signal in the 7500 Hz band is used. Is suitable, and the allowable allowable band is preferably about 7500 Hz plus or minus 500 Hz.

  When collecting sound from the engine during the driving test, sounds unrelated to the pivot abnormal sound inspection, for example, disturbance noise that is ambient noise, are also included. For this reason, there is a possibility that the reliability of the above-described pivot noise inspection is lowered by such disturbance noise. In order to suppress the influence of such disturbance noise, in one preferred embodiment of the present invention, a peak value other than the pivot sound frequency band in the converted signal converted into the frequency domain by the spectrum decomposition unit is set. An auxiliary calculation unit to be obtained as an auxiliary parameter is further provided, and the evaluation unit performs a pivot quality determination based on a comparison between the auxiliary parameter and the statistical feature value with each evaluation criterion. That is, instead of obtaining a final pass / fail result with the statistical feature amount of the above-described abnormal sound parameter group, a peak value in a frequency band including ambient noise other than the pivot sound frequency band is also obtained, By adding the peak value to the pass / fail judgment condition, the adverse effect of disturbance noise, which is ambient noise, on the pivot pass / fail judgment is suppressed.

Furthermore, in order to solve the above-mentioned problem, in the pivot abnormal sound inspection method for an engine valve structure having a pivot type rocker arm according to the present invention, the attention signal in the pivot sound frequency band including the pivot sound from the acquired engine sound signal. Taking out, measuring the number of occurrences at a predetermined time interval of the amplitude value greater than or equal to the threshold in the signal of interest as a noise parameter and calculating a statistical feature quantity of the measured noise parameters, Pivot quality determination is performed based on the comparison between the statistical feature quantity and the evaluation criteria. Of course, such a method for inspecting a pivot noise can also be provided with the effects and all the additional features and effects described for the pivot noise inspection apparatus described above.
Other features and advantages of the present invention will become apparent from the following description of embodiments using the drawings.

Embodiments of the present invention will be described with reference to the drawings.
FIG. 1 shows an engine inspection station that starts an engine 2 mounted on a test bench 1 and performs various dynamic performance tests of the engine 2. The engine 2 has an engine valve structure including a pivot type rocker arm 21 schematically shown in FIG. This engine valve operating structure opens and closes the valve by transmitting to the valve 24 the movement of the pivot-type rocker arm 21 that swings around the pivot portion 23 by the cam 22. A valve clearance is provided in the middle of the mechanism in which the cam 22 pushes the valve 24 through the pivot type rocker arm 21 and is returned by the valve spring 25. This valve clearance can be absorbed by the lash adjuster. However, the introduction of the lash adjuster is abandoned in order to avoid the complexity of the engine valve structure and the resulting increase in cost. A pivot type rocker arm in an engine valve structure that does not use such a lash adjuster may emit an abnormal pivot sound due to an assembly error or the like. In order to detect such an abnormal pivot sound, that is, an abnormal pivot sound and determine whether the engine is good or bad, the engine inspection station receives the microphone 3 for collecting the engine sound and the engine sound captured by the microphone 3. A pivot abnormal sound inspection device 4 is introduced which performs a pass / fail determination by processing.

  As shown in FIG. 3, the pivot noise detector 4 receives an amplifier 41 that appropriately amplifies the engine sound signal acquired through the microphone 3, and inputs the amplified engine sound signal. An FFT (Fast Fourier Transform) unit 42 serving as a spectrum decomposing unit that converts the signal into a region and outputs the signal of interest that is a pivot sound frequency band signal (7500 Hz plus or minus 500 Hz in this embodiment) output from the FFT unit 42. A calculation unit 43 for generating an abnormal sound parameter group to be described in detail later and calculating a statistical feature amount of the abnormal sound parameter group, and a band other than the pivot sound frequency band output from the FFT unit 42 Auxiliary calculation unit 44 for obtaining a peak value of the spectrum by inputting a signal, and an abnormal sound output from calculation unit 43 The evaluation unit 45 that determines the quality by comparing the statistical feature quantity of the parameter group and the peak value output from the auxiliary calculation unit 44 with the evaluation criteria, and the above-described control that constitutes the core of the pivot abnormal sound inspection device 4 A control unit 46 that controls the elements and an input / output device 48 that exchanges various data with the evaluation unit 45 and the control unit 46 via the I / O interface 47 are included.

  The calculation unit 43 includes an abnormal sound parameter measurement unit 43a that generates an abnormal sound parameter group and a feature amount calculation unit 43b that calculates a statistical feature amount from the abnormal sound parameter group. The abnormal sound parameter measurement unit 43a monitors the amplitude value of the attention signal input using the threshold value set by the threshold value setting unit 46a provided in the control unit 46, and counts the amplitude value exceeding the threshold value. It has a function of using the count number within a predetermined time as an abnormal sound parameter. For example, when an attention signal having a waveform as shown in FIG. 4 is input to the abnormal sound parameter measurement unit 43a from the engine 2 operating at 750 RPM as the idling speed, each amplitude is compared with the threshold value SH. Threshold value: The number of occurrences of amplitude exceeding SH (indicated by D1, D2, D3... In the figure) is counted. FIG. 4 is a waveform signal showing the temporal variation of the sound pressure in the pivot sound frequency band, and the amplitude here means the length between successive peaks and valleys.

  The feature amount calculation unit 43b develops the abnormal sound parameters obtained at predetermined time intervals by the abnormal sound parameter measurement unit 43a on the time axis (see FIG. 5), and calculates the statistical feature amount of the abnormal sound parameter group for a fixed time. However, in this embodiment, an average value that is easy to calculate is used as the statistical feature amount. In FIG. 5, the time axis development diagram of the abnormal noise parameter indicated by the engine valve structure finally evaluated as defective is a solid line, and the abnormal parameter time axis indicated by the engine valve structure evaluated as good is shown in FIG. The development diagram is shown as a dotted line. It can be understood that the average value of the abnormal noise parameter indicated by the solid line is clearly higher than the average value of the abnormal noise parameter indicated by the dotted line. For this reason, the evaluation unit 45 uses the first evaluation threshold set by the threshold setting unit 46a as the pass / fail judgment criterion for the statistical feature quantity, and shows a higher average value than the first evaluation threshold. Is given a bad evaluation regarding statistical features. In addition to the average value, the processing algorithm of the feature amount calculation unit 43b includes the sum of all abnormal parameters, the number of occurrences of all abnormal parameters that are greater than or equal to a predetermined value, and the variance analysis and histogram analysis of the abnormal parameter group. These can be used and any of them may be combined. Further, the feature amount calculation unit 43b may be provided with a plurality of types of processing algorithms so that the feature amount calculation unit 43b can be selected.

  The auxiliary calculation unit 44 has a function of obtaining, as an auxiliary parameter, a peak value in a band other than the pivot sound frequency band in the converted signal converted into the frequency domain by the spectrum decomposition unit by the FFT unit 42. As a result, the auxiliary parameter represents the characteristic of the disturbance sound that is ambient noise. Therefore, in the evaluation unit 45, when the value of the auxiliary parameter (peak value) is larger than the second evaluation threshold set by the threshold setting unit 46a, it is regarded as a region having a large disturbance noise influence. It is considered that the area is small, and this determination criterion (area) is also used to make a final pass / fail determination in consideration of the influence of disturbance noise in the pass / fail determination regarding the statistical feature amount. FIG. 6 schematically shows such a determination criterion. As is clear from this, in order to be finally determined to be a non-defective product, the influence of disturbance noise is small and the statistical feature value is related. It is necessary to determine that the quality is good. The case where the failure is reliably determined is a case where it is determined that the failure is good in the quality determination regarding the statistical feature quantity even though the influence of the disturbance noise is large. Others are inappropriate for judgment, and re-examination is performed as necessary.

A flow of a series of processes of the pivot noise inspection performed using the above-described pivot noise inspection apparatus 4 will be described below with reference to the flowcharts shown in FIGS.
First, the engine 2 is mounted on the test bench 1, and the engine is started upon receiving fuel supply or power supply (# 01). Various dynamic performance tests including a pivot sound test are performed on the engine 2 driven on the test bench 1. However, since the present invention is a pivot sound test, description other than the pivot sound test is omitted. .

  The engine sound is collected by the microphone 3 arranged on the test bench 1, and the engine sound signal is sent to the pivot abnormal sound inspection device 4 (# 02). The engine sound signal is subjected to FFT to generate a signal (spectrum signal) obtained by converting the time axis in the FFT sampling time into a frequency (# 03). The spectrum signal includes a broadband spectrum, and a signal in the 7500 Hz band (a signal of interest) is sent to the abnormal sound parameter measurement unit 43a, where an abnormal sound parameter group is measured (# 04). The obtained abnormal sound parameter group is sent to the feature quantity calculation unit 43b, where the statistical feature quantity, here, the average value is calculated (# 05). A signal other than the 7500 Hz band is sent to the auxiliary calculation unit 44, and the spectrum peak value is calculated as an auxiliary parameter (# 06). The statistical feature amount (average value) of the abnormal sound parameter group and the peak value as the auxiliary parameter are sent to the evaluation unit 45 and compared with each evaluation criterion (threshold value), and the determination method as shown in FIG. On the basis of the above, the pass / fail determination regarding the pivot noise is made (# 7). The determination result output from the evaluation unit 45 is sent to the input / output device 48 via the I / O interface 47 and used for notification of the determination result by the monitor 48a and the buzzer 48b (# 08).

  An abnormal sound parameter group can be measured by various methods, and an example thereof is shown in the flowchart of FIG. In this method, N noise parameters are measured. For this reason, 1 is set as an initial setting for a variable: n that restricts measurement to N (# 40). The signal of interest in the 7500 Hz band sent from the FFT unit 42 is sampled at a predetermined time, the waveform signal as shown in FIG. 4 is captured (# 41), and the amplitude value of the wave is calculated (# 42). . The calculated amplitude value is compared with a preset threshold value, and if the amplitude value exceeds the threshold value (# 43 Yes branch), the count value is incremented by one (# 44). If the amplitude value is equal to or less than the threshold value (# 43 No branch), the count value is not incremented, the next wave amplitude value check is performed, and this wave amplitude value check is repeated until all wave amplitude value checks are completed. (# 45). When the amplitude value check of all the waves is completed (# 45 Yes branch), the count value is substituted into the variable: abnormal sound parameter (n) as the abnormal sound parameter value (# 46). Further, the variable: n is incremented by 1 (# 47), and the measurement of the noise parameter from step # 41 to step # 47 is repeated until the value of n exceeds N (# 48). When the measurement of N noise parameters is completed (# 48 Yes branch), the feature value is calculated using the obtained N variable values, the noise parameters (1)... The data is output to the unit 43b (# 49).

  In the description of the embodiment described above, the statistical parameters of the abnormal sound parameters obtained by the computing unit 43 are added to the auxiliary parameters (spectrum peak values excluding the 7500 Hz band) obtained by the auxiliary computing unit 44 to obtain a final result. The pivot abnormal sound quality determination was performed, but the auxiliary calculation unit 44 is omitted, and the final pivot abnormal sound quality determination is adopted only by the statistical feature amount of the abnormal sound parameter obtained by the calculation unit 43. May be.

1 is an external view showing an engine inspection station incorporating a pivot abnormal noise inspection apparatus according to the present invention. The figure which shows typically the engine valve operating structure which has the pivot type rocker arm Functional block diagram of a pivot abnormal sound inspection apparatus according to the present invention The figure which shows typically the attention signal of 7500Hz band output from a FFT unit Diagram showing the distribution of abnormal noise parameters on the time axis The figure which shows the standard of the pivot abnormal sound quality judgment based on the statistical feature quantity and auxiliary parameter of the abnormal sound parameter Flowchart showing the flow of a series of processing for pivot abnormal sound inspection Flowchart showing an abnormal sound parameter group measurement processing subroutine

Explanation of symbols

3: Microphone 4: Pivot noise detector 42: FFT unit (spectral decomposition unit)
43: calculation unit 43a: abnormal sound parameter measurement unit 43b: feature amount calculation unit 44: auxiliary calculation unit 45: evaluation unit

Claims (7)

In the pivot noise inspection device of the engine valve structure with the pivot type rocker arm,
A spectral decomposition unit that extracts a target signal in a pivot sound frequency band including a pivot sound from an engine sound signal acquired through a microphone, and the number of occurrences of an amplitude value equal to or greater than a threshold value in the target signal at a predetermined time interval as sequential noise parameters An arithmetic unit that measures a plurality of times and calculates a statistical feature amount of the measured abnormal noise parameters, and an evaluation unit that performs a pivot quality determination based on a comparison between the statistical feature amount and an evaluation criterion A pivot abnormal sound inspection apparatus characterized by being provided. The pivot noise inspection apparatus according to claim 1, wherein the statistical feature amount is an average value of a plurality of noise parameters. The pivot noise inspection apparatus according to claim 1, wherein the statistical feature amount is a sum of a plurality of noise parameters. 2. The pivot abnormal sound inspection apparatus according to claim 1, wherein the statistical characteristic amount is the number of appearances of an abnormal sound parameter having a predetermined value or more. The pivot noise inspection apparatus according to claim 1, wherein the pivot sound frequency band is 7500 Hz plus or minus 500 Hz. An auxiliary calculation unit that obtains, as an auxiliary parameter, a peak value other than the pivot sound frequency band in the converted signal converted into the frequency domain by the spectrum decomposition unit, and the evaluation unit includes the auxiliary parameter and the statistical feature. The pivot abnormal noise inspection apparatus according to claim 1, wherein the pivot quality determination is performed based on a comparison of each quantity with each evaluation criterion. In the method of inspecting the pivot noise of an engine valve structure having a pivot type rocker arm,
Taking out the attention signal of the pivot sound frequency band including the pivot sound from the acquired engine sound signal, measuring the number of occurrences of the amplitude value of the threshold value or more in the attention signal at a predetermined time interval sequentially as an abnormal sound parameter, and A pivot noise inspection method, comprising: calculating a statistical feature amount of a plurality of measured abnormal noise parameters; and determining whether the pivot is good or bad based on a comparison between the statistical feature amount and an evaluation criterion.

Images