Method and device for detecting and displaying accidental fire from engine

Abstract

To compare and evaluate an accidental fire detection result through mechanical frequency analysis not considering human hearing sense characteristics, and an accidental fire detection result when the human hearing sense characteristics are taken into consideration.SOLUTION: A method for detecting and displaying an accidental fire from an engine includes: acquiring a sound emitted from an engine by using a microphone to generate sound data, and performing frequency analysis to generate a spectrogram with a sound pressure level as a parameter of a Z-axis; detecting, as first accidental fire detection, an accidental fire from a temporal reduction of the sound pressure level or the frequency in time-series data of suitable harmonics; generating, as second accidental fire detection, a spectrogram with loudness as a parameter of the Z-axis and detecting an accidental fire from a temporal reduction of the loudness or the frequency; and displaying a result of the first accidental fire detection on the left of a screen, displaying a result of the second accidental fire detection on the right of the screen, and comparing the results.SELECTED DRAWING: Figure 3

Description

[Technical field]

[0001] The present invention relates to a method and apparatus for detecting engine misfires based on engine sounds and displaying the detection results, for example in the finished vehicle inspection process on an automobile production line, and in particular to a method and apparatus for performing detection and display in parallel, taking into account the hearing characteristics of inspectors. [Background technology]

[0002] For example, there have been attempts to detect engine misfires from changes in sound or vibration. For example, Patent Document 1 discloses a technology in which the vibration sound of an engine is collected by a microphone, the sound within a certain time range after a predetermined time has elapsed since the generation of an ignition pulse is sampled, frequency analysis is performed, and the presence or absence of an abnormality is determined by comparing the normal vibration waveform with the detected vibration waveform for a specific frequency band. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Publication No. 1-217218 Summary of the Invention [Problem to be solved by the invention]

[0004] For example, engine misfire inspections carried out during the finished vehicle inspection process, which is the final stage of an automobile production line, are generally sensory inspections using the inspector's hearing. For example, an inspector test drives the finished vehicle to be inspected on a free roller, and accelerates the engine so that the engine speed gradually changes from a low speed range (e.g., 600 rpm) to a high speed range (e.g., 6,000 rpm) over an appropriate inspection time (e.g., about 10 to 20 seconds), and the inspector listens to the sounds made by the engine during that time to check for misfires.

[0005] In misfire judgment based on mechanical frequency analysis as in Patent Document 1, the hearing characteristics of humans (that is, inspectors) are not taken into consideration, and therefore it is not possible to compare and evaluate the same with a sensory inspection by a general inspector. [Means for solving the problem]

[0006] The engine misfire detection and indication method according to the present invention comprises: Acquire the sound produced by the engine and generate sound data. This sound data is converted according to frequency to generate frequency data including sound pressure level as a parameter; Based on the frequency data, time series data of harmonics of the fundamental frequency associated with the engine rotation speed is extracted; As a first misfire detection method, a misfire of an engine is detected based on a change in frequency or sound pressure level over time in the extracted harmonic time series data, generating sensory value time series data including sensory values ​​corresponding to human hearing characteristics as parameters based on the sound data; As the second misfire detection, the misfire of the engine is detected from the time change of the sensory value time series data, The detection results of the first misfire detection and the second misfire detection are displayed together.

[0007] That is, in this invention, the acquired sound data is processed in two ways, and misfire detection is performed in parallel based on the time-series data of the harmonic overtones of the fundamental frequency, that is, the frequency or sound pressure level change over time, and based on the time-series data of the sensory value corresponding to the human hearing characteristic change over time. Then, both detection results are displayed together. Effect of the Invention

[0008] According to the present invention, it becomes easy to compare and evaluate the misfire detection result based on mechanical frequency analysis that does not take into account the human hearing characteristics with the misfire detection result that takes into account the human hearing characteristics. [Brief description of the drawings]

[0009] [Figure 1]FIG. 1 is a functional block diagram of a first embodiment of the present invention. [Diagram 2] 4 is a flowchart showing the flow of processing in the first embodiment. [Diagram 3] FIG. 4 is an explanatory diagram showing a display example of the first embodiment. [Figure 4] FIG. 11 is a functional block diagram of a second embodiment. [Diagram 5] 10 is a flowchart showing the flow of processing according to a second embodiment. [Figure 6] FIG. 11 is an explanatory diagram showing a display example of the second embodiment. [Figure 7] FIG. 13 is an explanatory diagram showing a display example of the third embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0010] An embodiment of the present invention will be described below. This embodiment is implemented as a misfire inspection to check that no engine misfire occurs during the finished vehicle inspection process, which is the final stage of an automobile production line. In general, during the finished vehicle inspection process, an inspector test-drives the finished vehicle to be inspected on a free roller, and inspects many items such as the engine, meters, and brakes. During the test drive on the free roller, the engine is accelerated so that the engine speed gradually changes from a low speed range (e.g., 600 rpm) to a high speed range (e.g., 6000 rpm) over an appropriate inspection time (e.g., about 10 to 20 seconds), and a misfire inspection is performed based on the sound generated from the engine during that time. Hereinafter, the device of this embodiment will be referred to as a "misfire inspection device," and this misfire inspection device is configured as a part of the inspection device in the finished vehicle inspection process.

[0011] The engine to be inspected may be either a gasoline engine (i.e., a spark ignition internal combustion engine) or a diesel engine (i.e., a compression ignition internal combustion engine), may be a four-stroke or two-stroke engine, may be an in-line multi-cylinder engine, or may be a V-type multi-cylinder engine.

[0012] 1 shows a functional block diagram of the misfire detection device of the first embodiment. The misfire detection device of the first embodiment includes a sound acquisition unit 1, a vehicle signal acquisition unit 10, a sound signal processing unit 20, a feature extraction unit 30, an auditory characteristic conversion unit 80, a misfire determination unit 40, an input unit 50, a storage unit 60, and an information display unit 70.

[0013] The sound acquisition unit 1 includes a microphone that acquires the sound generated by the engine and converts it into an electric signal, i.e., sound data, and a recording unit that temporarily stores the sound data. The microphone is disposed outside the vehicle so as to collect the engine sound of the vehicle running on the free rollers. The directivity and frequency characteristics of the microphone are selected according to the position of the vehicle to be measured and the frequency band of the engine sound. For example, an omnidirectional microphone or a shotgun-type directional condenser microphone may be used. Sound data from which noise sounds have been removed by localizing the sound source using a microphone array or the like may be obtained. In addition, the engine sound may be acquired using existing microphones of the vehicle, such as a voice recognition microphone or a mobile phone microphone provided in the vehicle's car navigation system or the like.

[0014] As described above, the misfire check is performed while increasing the engine speed for, for example, 10 to 20 seconds, and therefore sound data having a time length of, for example, 10 to 20 seconds is acquired.

[0015] The vehicle signal acquisition unit 10 acquires various vehicle information such as the vehicle speed and engine speed of the vehicle to be inspected from the vehicle side via a communication standard such as CAN communication. When determining a misfiring cylinder, which will be described later, a signal indicating a specific reference crank angle position corresponding to the top dead center position of cylinder #1, for example, is acquired via the vehicle signal acquisition unit 10.

[0016] The sound signal processing unit 20 generates frequency data by converting the sound data acquired by the sound acquisition unit 1 according to frequency using a frequency analysis method such as STFT (Short-Time Fourier Transform), FFT (Fast Fourier Transform), or wavelet analysis. In particular, since it is necessary to obtain the time change of the frequency or sound pressure of a specific sound ultimately, frequency analysis is performed for each relatively short time domain to generate frequency data for each time. When the frequency data is expressed in two dimensions, for example, it is expressed as a frequency spectrum with the horizontal axis representing frequency and the vertical axis representing sound pressure level. In addition, the frequency data may be handled as a so-called spectrogram, which is three-dimensional data including time, by superimposing the conversion results for each time in a time series.

[0017] For example, processing with STFT generates a spectrogram, which is three-dimensional data with the X-axis representing time (seconds), the Y-axis representing frequency (Hz), and the Z-axis representing sound pressure level (dB), as shown in the left half of the display example in Figure 3. The sound pressure level on the Z-axis is represented, for example, by differences in color and brightness.

[0018] The multiple lines appearing in the spectrogram of Fig. 3 represent frequencies showing peaks of sound pressure levels of several orders including the fundamental frequency. In the illustrated example, the line present in the lowest frequency band is characteristic of the fundamental frequency, and several clearly appearing lines are characteristic of frequencies corresponding to harmonics of the fundamental frequency. As described above, the misfire check of one embodiment is performed while the engine speed is increasing, so the fundamental frequency and the harmonics tend to increase with time (i.e., as the engine speed increases).

[0019] The feature extraction unit 30 extracts features and feature quantities for detecting engine misfire from the time-series frequency data processed by the sound signal processing unit 20. For example, based on the above-mentioned spectrogram, it calculates the change in frequency and sound pressure level associated with engine misfire. When an engine misfires in a certain cylinder, the sound pressure level drops instantaneously because no engine explosion occurs. The engine speed also drops slightly instantaneously, causing the frequency to drop. Although such an instantaneous drop is slight in the characteristics of the fundamental frequency, it appears as an expanded change in the harmonics. Therefore, the feature extraction unit 30 calculates the differentiation of the time-series characteristics of the peak frequency of an appropriate harmonic and extracts the downward convex state. As for the sound pressure level, the time-series data of the appropriate harmonic is differentiated to extract the downward convex state.

[0020] Specifically, the fundamental frequency corresponding to the engine speed is found, and a frequency window is set for each harmonic of that frequency to generate time series data for the harmonics. For example, the frequency characteristics of a sound at a certain moment (t) are calculated to obtain multiple frequency peaks. The lowest peak frequency among those frequency peaks is then set as the fundamental frequency (N=1), and the sound that is an integer multiple of this is set as a harmonic to set the frequency window for the harmonics. The frequency characteristics at each moment are converted into time series data using this frequency window for the harmonics, and the time series data for each harmonic is obtained.

[0021] For example, if there is a peak at 33 (Hz) as the fundamental frequency at a certain engine speed, the second harmonic (N=2) will have a sound pressure level peak at 66 (Hz), the third harmonic (N=3) will have a sound pressure level peak at 99 (Hz), the fifth harmonic (N=5) will have a sound pressure level peak at 165 (Hz), and the tenth harmonic (N=10) will have a sound pressure level peak at 333 (Hz). Therefore, the harmonic window is set as follows: 49.5 to 82.5 (Hz) for the second harmonic (N=2), with a range of 66±16.5 (Hz), and 82.5 to 115.5 (Hz) for the third harmonic (N=3), with a range of 99±16.5 (Hz).

[0022] In order to identify frequency changes and sound pressure level changes associated with misfire, it is preferable to use harmonics of about the fifth to tenth harmonic. For example, in the example shown in FIG. 3, misfire detection is performed based on the fifth harmonic.

[0023] The fundamental frequency is determined based on the number of cylinders in the engine, whether the engine is a four-stroke cycle engine or a two-stroke cycle engine, and the engine rotation speed, so the fundamental frequency may be set using a signal of the engine rotation speed.

[0024] The misfire determination unit 40 determines whether or not a misfire has occurred based on the features and feature amounts extracted by the feature extraction unit 30. For example, the amount of frequency drop compared to the amount before and after a downward convex feature portion in the time series data of the peak frequency of the fifth harmonic is compared with a predetermined threshold, and if the drop is equal to or greater than the threshold, it is determined that a misfire has occurred. The misfire determination may also be performed taking into account the slope of the characteristics when the frequency drops.

[0025] The same applies when focusing on the drop in sound pressure level instead of the drop in frequency associated with a misfire. For example, the amount of drop in sound pressure level before and after a characteristic portion that shows a downward convexity in the time series data of the sound pressure level of the fifth harmonic is compared with a predetermined threshold, and if the drop is equal to or greater than the threshold, it is determined that a misfire has occurred.

[0026] The storage unit 60 is a storage means consisting of a HDD or the like for storing data such as the sound data acquired by the sound acquisition unit 1, the vehicle signal acquired by the vehicle signal acquisition unit 10, the data (frequency data, time series data, etc.) processed and generated by the sound signal processing unit 20, and the misfire determination results by the misfire determination unit 40. This is configured, for example, by a server computer in a factory included in the finished vehicle inspection device, or a cloud computer connected via a communication line.

[0027] The input unit 50 is a means for inputting additional information, such as characters such as text, images, videos, etc., to the data thus stored in the storage unit 60. This is not limited to manual input, but includes automatic input. For example, vehicle information (vehicle number, vehicle model name, vehicle specifications, etc.), environmental information such as weather, temperature, humidity, time, location information, data processing method, inspection equipment information, etc. are automatically input and added to the data set. In addition, the inspection time, inspection status, inspector's name or ID number, etc. may be input automatically or manually.

[0028] As described above, the misfire determination that does not take into account the human hearing characteristics is referred to as the "first misfire detection" in this specification.

[0029] In addition, the basic technology for detecting engine misfires based on a momentary drop in frequency or sound pressure level in the above-mentioned harmonic time series data is described in detail in JP 2023-56081 A, which was previously proposed by the present inventor.

[0030] The auditory characteristic conversion unit 80 generates sensory value data including sensory values ​​corresponding to human auditory characteristics as parameters, using the frequency data processed by the sound signal processing unit 20. As the sensory value parameters, for example, loudness (unit: phon), sharpness (unit: acum), roughness (unit: asper), fluctuation strength (unit: vib), etc. may be used. In addition, sound pressure levels weighted by C-weighted or A-weighted auditory characteristics may be used.

[0031] In one example, by using loudness (phon) as the sensory parameter instead of sound pressure level, a spectrogram is generated, which is three-dimensional data with the X-axis representing time (seconds), the Y-axis representing frequency (Hz), and the Z-axis representing loudness (phon), as shown in the right half of the display example in Fig. 3. The loudness on the Z-axis is represented by differences in color and brightness, as in the case of sound pressure level. When loudness is used as the sensory parameter, the time window for determining the loudness of a sound is set to be relatively long, for example, about 30 ms, taking into account the impulse response of the auditory nerve.

[0032] This kind of sensory value time series data is input to the feature extraction section 30, similar to the case of the data using the sound pressure level described above, and the misfire determination section 40 performs a misfire determination.

[0033] As shown in the right half of FIG. 3, a spectrogram using loudness as the sensory value parameter is similar to the spectrogram in the left half of FIG. 3 using sound pressure level as a parameter.

[0034] That is, the multiple lines appearing in the spectrogram represent frequencies showing loudness peaks of several orders including the fundamental frequency, the line present in the lowest frequency band is characteristic of the fundamental frequency, and several clearly appearing lines are characteristic of frequencies corresponding to harmonics of the fundamental frequency. Since the misfire inspection of one embodiment is performed while the engine speed is increasing, the fundamental frequency and harmonics frequencies at which loudness peaks tend to increase with time (i.e., as the engine speed increases).

[0035] However, the characteristics of the peak frequencies (fundamental frequencies and harmonic frequencies) at which loudness peaks do not completely match the characteristics of the peak frequencies (fundamental frequencies and harmonic frequencies) at which the sound pressure level peaks, as shown in the left half of Figure 3.

[0036] In the feature extraction unit 30, similarly to the case of the spectrogram using the above-mentioned sound pressure level, for example, the frequency time series data of the peak frequency of the fifth harmonic related to loudness is differentiated to extract the downward convex state. Alternatively, the loudness time series data of the fifth harmonic is differentiated to extract the downward convex state.

[0037] As mentioned above, when an engine misfire occurs in a cylinder, the loudness drops instantaneously because no engine explosion occurs. The engine speed also drops instantaneously slightly, causing the frequency to drop. This instantaneous drop is slight in the characteristics of the fundamental frequency, but appears as an expanded change in the harmonics. Therefore, the feature extraction unit 30 calculates the derivative of the time series frequency characteristics at the peak frequency of an appropriate harmonic and extracts a downward convex state. With regard to loudness, the time series data of an appropriate harmonic is differentiated to extract a downward convex state.

[0038] The features and feature quantities extracted in this manner are input to the misfire determination unit 40. The misfire determination unit 40 determines whether or not a misfire has occurred based on the features and feature quantities extracted by the feature extraction unit 30. For example, the amount of frequency drop compared to before and after a downwardly convex feature portion in the time-series data of the peak frequency of the fifth harmonic (peak frequency related to loudness) is compared with a predetermined threshold, and if the drop is equal to or greater than the threshold, it is determined that a misfire has occurred. The misfire determination may also be performed taking into consideration the slope of the characteristics when the frequency drops.

[0039] The same applies when focusing on the drop in loudness instead of the drop in frequency associated with a misfire. For example, the loudness before and after a characteristic portion that is convex downward in the time series data of the loudness of the fifth harmonic is compared with a predetermined threshold, and if the drop is equal to or greater than the threshold, it is determined that a misfire has occurred.

[0040] Such a misfire determination based on a sensory value (for example, loudness) that takes into account the characteristics of human hearing is referred to in this specification as a "second misfire detection."

[0041] The results of such second misfire detection and various data related thereto are stored in the memory unit 60, similar to the data related to the first misfire detection. In addition, additional information is input via the input unit 50. In one preferred example, the data related to the first misfire detection and the data related to the second misfire detection are stored as a data set for each vehicle to be inspected.

[0042] The information display unit 70 is a display means for displaying the misfire determination result to relevant parties such as workers in the vicinity of the inspection device, inspectors driving the vehicle, managers, data scientists who utilize the data, etc. For example, it is configured to include a liquid crystal display, an organic EL display, etc. Furthermore, if some kind of notification sound or voice is involved, it is configured to include a sound source, an amplifier, a speaker, etc. for generating and emitting the sound.

[0043] Here, the information display unit 70 displays both the detection result of the first misfire detection that does not take into account the human hearing characteristics and the detection result of the second misfire detection that takes into account the human hearing characteristics. For example, the two detection results may be displayed side by side, vertically or vertically, or the two detection results may be displayed alternately or sequentially by switching the display of one screen. Furthermore, the information display unit 70 may have two displays, and the detection result of the first misfire detection and the detection result of the second misfire detection may be displayed individually on each display.

[0044] FIG. 3 shows an example in which the detection results of the first misfire detection and the detection results of the second misfire detection are displayed side by side on one display screen. The "(1) Measurement equipment evaluation result" on the left side is the detection result of the first misfire detection, and the "(2) Sensory evaluation result (estimated value)" on the right side is the detection result of the second misfire detection. In the center of each of the left and right sections, the spectrograms for the first misfire detection and the spectrograms for the second misfire detection are displayed in color. As shown in the example of "(1) Measurement equipment evaluation result" on the left side, the timing at which a misfire was determined to have occurred is indicated by a red triangle as "misfire sound determination" at the upper edge of the spectrogram. In addition, the upper part of each of the left and right sections shows the thresholds for misfire determination for the feature that appears as a downward convex (frequency threshold and sound pressure level threshold in the left section, and frequency threshold and loudness threshold in the right section). At the bottom of each of the left and right sections, the judgment result is displayed in text form, such as "Engine misfire" or "No engine misfire."

[0045] In the example of Figure 3, the first misfire detection on the left determines that "engine misfire exists," and the second misfire detection on the right determines that "engine misfire does not exist." By displaying the two results together in this way, it becomes easy to compare and evaluate the misfire detection results based on mechanical frequency analysis that does not take into account human hearing characteristics with the misfire detection results that do take into account human hearing characteristics. For example, it is possible to know that a misfire that cannot be detected based on human hearing characteristics is determined to be a misfire by the first misfire detection.

[0046] Next, Fig. 2 is a flow chart showing the process flow of the misfire detection device of the first embodiment. In this example, the first misfire detection focuses on the drop in sound pressure level accompanying a misfire, and the second misfire detection focuses on the drop in loudness accompanying a misfire.

[0047] First, the sound of the engine of the vehicle test-running on the free rollers is collected by the microphone of the sound acquisition unit 1 and acquired as sound data (step 1). Next, the sound signal processing unit 20 converts the sound data according to frequency using a frequency analysis method such as STFT or FFT to generate frequency data (step 2). In step 3, the vehicle signal acquisition unit 10 acquires information on the engine rotation speed.

[0048] Next, as described above, an appropriate frequency window is set to generate time-series data on the sound pressure levels of appropriate harmonics (step 4). The time-series data on the sound pressure levels is then differentiated to extract characteristic portions that are downwardly convex (step 5), and the amount of reduction in the sound pressure levels is calculated (step 6). In step 7, the amount of reduction is compared with a threshold value to determine the presence or absence of a misfire.

[0049] In step 8, the result of the first misfire detection thus obtained is stored in the storage unit 60 and is also displayed on the information display unit .

[0050] Meanwhile, in steps 9 to 13, a second misfire detection process is executed in parallel with the first misfire detection process in steps 4 to 8. That is, an appropriate frequency window suitable for loudness, which is a sensory value, is set to generate time series data of appropriate harmonic loudness (step 9). Then, the loudness time series data is differentiated to extract a downward convex characteristic portion (step 10), and the amount of loudness reduction is calculated (step 11). In step 12, the amount of reduction is compared with a threshold value to determine the presence or absence of a misfire.

[0051] In step 13, the result of the second misfire detection thus obtained is stored in the storage unit 60 and is also displayed on the information display unit .

[0052] Next, Fig. 4 shows a functional block diagram of a misfire inspection device of the second embodiment. As with the first embodiment described above, the misfire inspection device of the second embodiment includes a sound acquisition unit 1, a vehicle signal acquisition unit 10, a sound signal processing unit 20, a feature extraction unit 30, an auditory characteristic conversion unit 80, a misfire determination unit 40, an input unit 50, a storage unit 60, and an information display unit 70. Furthermore, the misfire inspection device of the second embodiment includes a cylinder number estimation unit 100 that estimates the number of cylinders of the engine to be inspected, and a misfiring cylinder determination unit 90 that determines which cylinder is misfiring.

[0053] In the second embodiment, the vehicle signal acquisition unit 10 acquires from the vehicle side an ignition signal for each cylinder and a reference crank angle signal for cylinder discrimination indicating a specific reference crank angle position corresponding to, for example, the top dead center position of cylinder #1.

[0054] The cylinder number estimation unit 100 estimates the number of cylinders based on the number of relatively low peak frequencies that appear between the peak frequency of a certain harmonic (which may be a reference frequency) at which the sound pressure level peaks and the peak frequency of an adjacent harmonic. For example, in the spectrogram shown in the display example of FIG. 6, two characteristic lines showing relatively low peaks are generated between the characteristic line showing the reference frequency of N=1 and the characteristic line showing the harmonic of N=2. Similarly, two characteristic lines showing relatively low peaks are generated between the characteristic line of the adjacent harmonic of N=2 and the characteristic line of the harmonic of N=3. For example, in a three-cylinder engine of a two-stroke cycle type, three explosions occur per rotation, so that frequency peaks corresponding to the explosions of each cylinder are generated. Therefore, if the frequency characteristic includes three peaks including the harmonic peak as shown in the display example of FIG. 6, it can be determined that the engine is a three-cylinder engine of a two-stroke cycle type (or a six-cylinder engine of a four-stroke cycle type). The accuracy of the cylinder number estimation is improved by counting and comparing the number of peaks between multiple harmonics.

[0055] Such cylinder number estimation is used, for example, to confirm whether the sound acquired by the sound acquisition unit 1 is valid as the engine sound of the vehicle being inspected. Since the engine type and number of cylinders of the vehicle being inspected are generally known, if the number of cylinders estimated from the sound differs from the number of cylinders of the vehicle being inspected, it can be considered that the engine sound acquired is erroneously that of another vehicle, such as a vehicle on an adjacent inspection line.

[0056] The misfiring cylinder determination unit 90 determines which cylinder is misfiring by synchronizing the timing of the misfire with the timing of the ignition signal for each cylinder. The number of the cylinder is determined based on a reference crank angle signal for cylinder determination.

[0057] The display example in Fig. 6 is an example in which the result of the first misfire detection is displayed on one screen of the information display unit 70 having, for example, two screens, and on the left side, a spectrogram is displayed together with a red triangle indicating the timing of misfire detection, and below that, it is displayed that the estimated number of cylinders is "3 cylinders." Note that auxiliary lines indicating the harmonic positions are added to the right side of the spectrogram.

[0058] On the right side, the ignition signal for the entire engine and the ignition signal for each cylinder are shown as square waves to indicate misfiring cylinders. Below the ignition signal, time series data of the peak frequency of appropriate harmonics is displayed with a circle indicating the ignition timing, and in particular, the timing determined to be a misfire due to a downward convex temporary drop is shown as a red circle. Note that a time frame corresponding to the period of the right display is shown on the spectrogram, and a thick arrow is drawn from the bottom end of the time frame to the right to easily understand the relationship between the two.

[0059] In the illustrated example of Fig. 6, it is shown that the #2 cylinder is misfiring. Here, Fig. 6 shows only the detection result of the first misfire detection, but the other screen of the information display unit 70 having two screens displays the detection result of the second misfire detection considering the human hearing characteristics. It is also possible to estimate the number of cylinders and determine the misfiring cylinder based on a spectrogram that takes into account the human hearing characteristics, for example, loudness as a parameter.

[0060] Fig. 5 is a flowchart showing the flow of processing of the second embodiment which is performed in addition to the processing shown in the flowchart of Fig. 2. In step 21, the number of cylinders is estimated from the characteristics of the peak frequency generated by the processing of the first embodiment. In step 22, an ignition signal is acquired from the vehicle side by the vehicle signal acquisition unit 10, in step 23, the timing at which the misfire is determined to have occurred is synchronized with the timing of the ignition signal, and in step 24, the misfiring cylinder is determined. Then, in step 25, information on the misfiring cylinder and the number of cylinders is displayed on the information display unit 70.

[0061] The above describes an example in which loudness (phon) is used as the sensory value parameter, but similar processing is possible when sharpness (acum), roughness (asper), fluctuation strength (vib), etc. are used as the sensory value parameter.

[0062] Fig. 7 shows an example of display on the information display unit 70 when sharpness (acum), which corresponds to the pitch of a sound, is used as the sensory value parameter. As in the first embodiment of Fig. 3, the left side of the screen displays the detection results based on the first misfire detection that does not depend on the human hearing characteristics, and the right side displays the detection results based on the second misfire detection that uses sharpness as a parameter.

[0063] In this example, the first misfire detection detects misfire by detecting a decrease in the half-width of the peak frequency of an appropriate harmonic. That is, as described above, when a misfire occurs, the sound pressure level and the peak frequency temporarily decrease, so the peak value of the frequency indicating the peak decreases and its half-width becomes narrower. For this reason, an appropriate threshold value (10 Hz in the illustrated example) is set for the amount of temporary decrease in the half-width associated with a misfire, and if the half-width decreases more than this threshold value, it is determined that a misfire has occurred. In the example on the left side of FIG. 7, misfire is determined at two points in time, and red triangles indicating the misfire detection timing are displayed in two places.

[0064] On the other hand, in the second misfire detection shown on the right side of Fig. 7, time series data of sharpness (acum) is generated, and the amount of the temporary decrease in sharpness (acum) due to misfire is compared with a predetermined threshold value (2 to 4 acum in the illustrated example), and if the decrease is greater than this threshold value, it is determined that a misfire has occurred. In other words, when a misfire occurs, the sound pressure level temporarily decreases and the peak frequency decreases, so the sharpness (acum), which corresponds to the high pitch, also decreases.

[0065] In the example on the right side of FIG. 7, since there is no large temporary drop in the time series data of sharpness (acum), it is determined that there is no misfire.

[0066] In the example of FIG. 7, the detection results of the first misfire detection and the detection results of the second misfire detection are shown by a single characteristic line that is similar to each other, making it easier to compare the two.

[0067] While one embodiment of the present invention has been described above in which the present invention is applied to engine misfire inspection during the finished vehicle inspection process, the present invention is not limited to this and can be applied in a variety of ways. For example, the present invention can be used to easily check for the presence or absence of misfire during inspection and maintenance at a dealership or repair shop.

[0068] In the above embodiment, misfire detection is performed while the engine speed is being increased, but it is also possible to detect misfire while maintaining a constant engine speed. [Explanation of symbols]

[0069] 1...Sound acquisition section 1 10…Vehicle signal acquisition unit 20...Audio signal processing unit 30…Feature extraction section 40...Misfire determination section 50...Input section 60...Storage section 70…Display section 80…Hearing characteristic conversion unit 90…Misfiring cylinder detection section 100…Cylinder number estimation section

Claims

1. Obtain the sound generated from the engine to generate sound data, generate frequency data including the sound pressure level as a parameter by converting this sound data according to the frequency, extract the time-series data of the harmonics of the fundamental frequency related to the engine speed based on this frequency data, as the first misfire detection, detect the misfire of the engine from the change in frequency or sound pressure level over time in the extracted time-series data of the harmonics, generate sensory value time-series data including a sensory value corresponding to human auditory characteristics as a parameter based on the above sound data, as the second misfire detection, detect the misfire of the engine from the change in time of this sensory value time-series data, combine and display the detection result by the first misfire detection and the detection result by the second misfire detection, A method for detecting and displaying misfires of an engine.

2. The above sensory value time-series data is loudness data including the loudness value for each frequency as a sensory value, as the second misfire detection, detect the misfire of the engine from the change in frequency or loudness over time in the loudness data of the frequency related to the engine speed, The method for detecting and displaying misfires of an engine according to Claim 1.

3. The above sensory value time-series data is fluctuation strength data including the value of the fluctuation strength for each frequency as a sensory value, as the second misfire detection, detect the misfire of the engine from the change in frequency or fluctuation strength over time in the fluctuation strength data of the frequency related to the engine speed, The method for detecting and displaying misfires of an engine according to Claim 1.

4. The above sensory value time-series data is sharpness data including the sharpness value based on the above frequency data, as the second misfire detection, detect the misfire of the engine from the change in sharpness over time, The method for detecting and displaying misfires of an engine according to Claim 1.

5. The above sensory value time-series data is roughness data including the roughness value based on the above frequency data, as the second misfire detection, detect the misfire of the engine from the change in roughness over time, The method for detecting and displaying misfires of an engine according to Claim 1.

6. Generate loudness data by converting the sound pressure level in the above frequency data into loudness along with human auditory characteristics, The method for detecting and displaying misfires of an engine according to Claim 2.

7. Extract the loudness data of the harmonics of the fundamental frequency related to the engine's rotational speed, and perform second misfire detection from the time change of this loudness data of the harmonics. The engine misfire detection and display method according to claim 2.

8. The second misfire detection determines misfire when there is a change in the sensory value above a temporary threshold in the sensory value time series data, and the above threshold is set in consideration of the human auditory characteristics related to frequency. The engine misfire detection and display method according to claim 1.

9. Obtain a reference crank angle signal correlated with the top dead center position of cylinder #1 from the engine side, and perform discrimination and display of the misfiring cylinder. The engine misfire detection and display method according to claim 1.

10. Estimate the number of engine cylinders based on the number of frequency peaks generated between the fundamental frequency and the second harmonic, and display this number of cylinders in accordance with the misfire detection result. The engine misfire detection and display method according to claim 1.

11. A sound acquisition unit that acquires the sound generated from the engine and generates sound data, a sound signal processing unit that generates frequency data including the sound pressure level as a parameter by converting this sound data according to frequency, a first misfire detection unit that extracts the time series data of the harmonics of the fundamental frequency related to the engine's rotational speed based on this frequency data, and detects the engine misfire from the time change of the frequency or sound pressure level in this time series data of the harmonics, a second misfire detection unit that generates sensory value time series data including a sensory value corresponding to human auditory characteristics as a parameter based on the above sound data, and detects the engine misfire from the time change of this sensory value time series data, and an information display unit that combines and displays the detection result by the first misfire detection unit and the detection result by the second misfire detection unit. An engine misfire detection and display device comprising the above.

Images