JP5557286B2 - Knocking determination method and apparatus - Google Patents

Description

  The present invention relates to a knocking determination method and apparatus, and more particularly, to a knocking determination method and apparatus for determining knocking of an engine in an engine test apparatus.

  Knocking (or knocking) is a phenomenon in which unburned gas is compressed by the combustion gas in the engine cylinder and self-ignited and rapidly burns and resonates. When knocking occurs, the combustion gas propagates heat by vibration. May damage the engine. Although it is necessary to delay the ignition timing in order to suppress knocking, there is a problem that if the ignition timing is delayed, there is a problem in that the fuel consumption is lowered. Therefore, it is necessary to accurately measure and determine the occurrence of knocking.

  As a knock determination method, auditory evaluation is widely used (see, for example, Patent Document 1). This is a method in which the tester actually listens to the engine operating sound, listens to the magnitude and frequency of abnormal noise in the engine, and evaluates the knock state. However, there is a problem that the auditory evaluation has a large variation due to the examiner and the test environment and the reproducibility is low. In particular, whether or not weak sounds, high-frequency sounds, and instantaneous sounds can be heard depends largely on the individual's hearing ability and experience, and it is difficult to accurately identify them unless they are skilled. For this reason, the number of people who can perform auditory evaluation is limited. Moreover, since auditory evaluation is performed by a tester, there is a problem that it cannot be applied to an automatic knock control system or an automatic adaptive measurement system.

  Therefore, in recent years, a method based on in-cylinder pressure measurement has been widely used in order to quantitatively perform knock determination (see, for example, Patent Document 2). This method is a method for determining knock by measuring the pressure in the cylinder and performing various signal processing on the in-cylinder pressure signal. For example, the acquired in-cylinder pressure signal is filtered to acquire the natural vibration component of the knock, and the waveform of the crank angle range where the knock occurs is cut out, and the maximum amplitude, the root mean square of the amplitude, the square integration, etc. A representative value is obtained and used as a knock index. Thereby, quantitative knock determination is performed.

  As another quantitative knock determination method, there is a method using a vibration sensor (see, for example, Patent Document 3). This is a system mounted on an actual vehicle, which is a method of performing knock determination by attaching a vibration sensor to a crankcase to measure vibration and performing various signal processing on the vibration measurement signal. For example, the vibration measurement signal is filtered to obtain the natural vibration component of knock, and the waveform of the crank angle range where the knock occurs is cut out to obtain a representative value such as the average of the absolute amplitude, which is used as the knock index. Thereby, quantitative knock determination is performed.

JP2003-232675A JP2000-110652A Japanese Patent No. 3054919

  However, the above-described two methods have a problem that calibration by auditory evaluation is necessary because the determination result does not match the result of auditory evaluation. For example, the in-cylinder pressure measurement method excludes mechanical noise from the measurement, so the calculation result does not match the result of the auditory evaluation. It becomes. On the other hand, the vibration sensor method calibrates the threshold value of the amplitude value corresponding to the occurrence of a knocking sound by auditory evaluation for each condition because the component of mechanical vibration included in the vibration measurement value varies depending on the driving conditions and individual engines. There is a need.

  As another disadvantage, the in-cylinder pressure measurement method requires an expensive in-cylinder pressure sensor as many as the number of cylinders, which is expensive, and requires a prior investigation on the knock frequency and the filter used. There is a problem. On the other hand, the method using the vibration sensor has a problem in that it is necessary to investigate the natural frequency and the filter.

  The present invention has been made in view of such circumstances, and it is an object of the present invention to provide a knock determination method and apparatus that can quantitatively determine knock and can obtain results equivalent to evaluation by audibility. To do.

  In order to achieve the above object, the invention according to claim 1 is a sensor for detecting engine sound, and a time mask process based on a signal immediately before and / or immediately after a signal obtained by the sensor, Calculating means for calculating the power ratio of the background sound to the power of the background sound by performing frequency mask processing based on a signal in the adjacent frequency band, and calculating the power ratio to the power of the background sound as the knock intensity; knocking based on the calculated knock intensity There is provided a knocking judging device characterized by comprising: a judging means for judging the presence or absence.

  According to the present invention, the time mask process and the frequency mask process are performed in the same manner as the process performed by the human auditory sense, so that it is possible to obtain a result equivalent to the auditory evaluation and to obtain a quantitative result by signal processing. Knock determination can be performed.

  According to a second aspect of the present invention, in the first aspect, the calculation unit obtains the power of the background sound subjected to the time mask processing and the power of the background sound subjected to the frequency mask processing, and The power ratio is calculated by selecting the larger of the powers and dividing the power of the entire sound. According to the present invention, the power ratio is obtained using the larger one of the power of the background sound subjected to the time mask process and the power of the background sound subjected to the frequency mask process, or the frequency at which the power ratio exceeds the threshold is obtained. Since the presence or absence of knocking is determined, a result very close to human hearing can be obtained.

  According to a third aspect of the present invention, in the second aspect, the calculation means extracts the signal obtained by the sensor by dividing it into a plurality of frequency bands, calculates the power ratio for each frequency band, and calculates the calculated power. The largest value among the ratios is selected as the knock strength. According to the present invention, it is possible to obtain a result very close to human hearing by selecting the one with the largest power ratio.

  According to a fourth aspect of the present invention, in any one of the first to third aspects, a sound signal of a pseudo knock sound is created by adding a reverberant sound intensity to the knock intensity and applying a vibration of a predetermined frequency. Pseudo knock sound generation means is provided. According to the present invention, since reverberation is added to the knock intensity and the frequency vibration is applied, a pseudo knock sound that is easy to perform auditory evaluation is generated. Therefore, it is possible to reliably perform auditory evaluation without causing individual differences. Moreover, since quantitative evaluation and auditory evaluation can be performed simultaneously, the reliability of evaluation can be improved.

  In order to achieve the above object, the invention according to claim 5 detects engine sound, extracts the detected signal by dividing it into a plurality of frequency bands, and immediately and / or immediately after each of the extracted frequency bands. Performs a time mask process based on the above to determine the power of the background sound, and performs a frequency mask process based on a signal in the adjacent frequency band to determine the power of the background sound, and selects the larger of the two background sound powers The power ratio is obtained by dividing the power of the entire sound, and the largest value among the power ratios obtained for each frequency band is set as the knock strength, and the presence or absence of knock is determined based on the set knock strength. A knocking determination method characterized by determining the above is provided.

  According to the present invention, since the time mask process and the frequency mask process are performed in the same way as the process performed by human auditory sense, it is possible to obtain a result equivalent to the auditory evaluation and to obtain a quantitative result by signal processing. Can make a knock determination.

The figure which shows typically the structure of the knock determination apparatus of this Embodiment. Schematic of the control flow of the knock determination device of the present embodiment The figure which shows the detail of the knock intensity | strength calculation flow of FIG. 2, and a knock determination flow Illustration about frequency extraction Diagram showing an example of signal waveform and power Explanation of delay compensation Illustration of auditory sensation A characteristics Illustration of time mask effect Explanatory drawing of knock determination The figure which shows the detail of the output flow of the pseudo knock sound of FIG. Illustration of pseudo knock sound

  A preferred embodiment of a knocking determination method and apparatus according to the present invention will be described with reference to the accompanying drawings. The knocking determination method and apparatus according to the present invention is an apparatus for obtaining a knock determination result equivalent to human auditory sensation evaluation. In order to effectively explain the operation, first, the function of human hearing will be described.

  When the engine sound is sensed by hearing, the engine sound passes through the ear canal, vibrates the eardrum, and the ear ossicles, and reaches the cochlea. The cochlea is formed in a spiral shape with a non-constant diameter, and reacts at different portions for each frequency, so the engine sound is divided and transmitted by the frequency. At that time, the present sound is partially erased by the reverberation of the past sound (time mask effect), and the effect is partially erased by the sound of the adjacent sound range (frequency mask effect). The time mask effect is an effect that disturbs the hearing during a certain period of time because it takes a certain amount of time to calm down from the state where the human auditory nerve is excited. The mask effect also occurs just before the sound begins. On the other hand, the frequency mask effect is an effect in which hearing of neighboring frequencies is hindered by the presence of nerves that hear sounds in the neighboring sound range in close proximity.

  The sound that remains without being drowned out by these two mask effects excites the auditory nerve. Here, the sound corresponding to the excited state by the mask effect is defined as the background sound, and the remaining sound is defined as the singular sound. The knocking sound is recognized when the singular sound is larger than the background sound. In addition, the sound with the highest specificity among the sounds heard for each sound range is typically recognized. Here, the frequency at which the sound with the highest specificity is observed does not necessarily match the natural frequency of knocking, and the actual knocking sound is generally more than the sound of the natural frequency called canning that is expressed as knocking. The sound repellent sound with a wide frequency range is dominant. In particular, in the trace knock state, the former sound is hardly observed. It is considered that the natural frequency sound is generated from a standing wave in the cylinder, and the sound repellent sound is generated from a sudden stepwise pressure increase in the cylinder. When listening to such a knocking sound, a highly specific sound may be observed in a wider frequency range.

  When the observer hears a particularly loud noise, it determines that it is a knock and counts its occurrence frequency. Then, subjectively judging from the magnitude of the knocking sound and the occurrence frequency, whether or not the knocking is constantly generated or whether the level of the knocking state is evaluated.

  Next, the configuration of knocking determination apparatus 10 of the present embodiment will be described based on FIG. A knocking determination device 10 shown in FIG. 1 is a device that determines knocking of the engine 12, and mainly includes a sensor 14, a processor 16, a speaker 18, and a display 20.

  The sensor 14 is a microphone for detecting engine sound, and is installed in the vicinity of the engine 12. The sensor 14 is not limited to a microphone, and any sensor that can determine the sound pressure may be used. For example, you may use a vibration sensor by combining with the processing apparatus which performs unit conversion. In this case, the vibration sensor may be directly attached to the engine 12.

  The sensor 14 is connected to the processor 16 via an amplifier 22 and an AD converter 24. The detection signal of the sensor 14 is amplified by the amplifier 22, converted into a digital signal by the AD conversion 24, and input to the processor 16.

  The processor 16 is a device that performs various signal processing described later, and is connected to the display device 20. The display device 20 is a device that displays the knock intensity, the knock determination result, and the like obtained by the processor 16 in real time, and displays measurement conditions and the like as necessary. The knock strength and the knock determination result may be output to the recording medium 28 and recorded on the recording medium 28. Further, it may be output to a printer (not shown) and printed.

  The processor 16 is connected to the speaker 18 via the DA converter 26. The speaker 18 is a device that outputs the pseudo knock sound signal generated by the processor 16 as sound, and the pseudo knock sound is reproduced from the speaker 18 toward the observer in real time.

  Next, various signal processing performed by the processor 16 will be described with reference to FIG. FIG. 2 shows a rough control flow of the processor 16. As shown in the figure, the processor 16 first calculates the knock intensity (step S1). Here, the knock intensity is a value indicating the loudness of the engine sound that is recognized as a knock, and, as will be described later, is specific to the power of the sound existing as the background (hereinafter referred to as background sound). It is given as a ratio to the power of the changing sound (hereinafter referred to as singular sound).

  Next, the presence or absence of knocking is determined based on the calculated knocking intensity (step S2). Further, a pseudo knock sound signal is generated based on the calculated knock intensity, and the pseudo knock sound is output from the speaker 18 (step S3). Each flow will be specifically described below.

  Details of the above-described knock intensity calculation flow (step S1) and knock determination flow (step S2) are shown in FIG. As shown in the figure, the processor 16 first separates the input signal into a plurality of frequency bands (step S11). At that time, separation may be performed using a plurality of bandpass filters. Further, the bandwidth and the number of bands of the frequency band to be separated are not particularly limited, but may be defined as shown in FIG. 4, for example. That is, the bandwidth is defined as 1/6 Octave based on the balance between the detection accuracy of the singular sound and the calculation load of the processor 16, and the 4 kHz to 20 Hz range is divided into 15 by the bandwidth and further overlapped by being located in the middle. Add 14 bands and define a total of 29 bands. Note that high-frequency sounds of about 10 kHz or more are difficult to hear for humans, and may be ignored in some cases.

  The following processing from step S12 to step S18 is performed for each separated frequency band and is simultaneously calculated in real time. First, the separated signals are squared and the sound power is calculated (step S12). An example of the calculated sound power is shown in FIG. The upper part of FIG. 5 is a waveform of a certain sound signal, and the lower part of FIG. 5 is a power waveform obtained by squaring it. In obtaining the mean square, the average width is preferably set as small as possible so that the peak is not erased by the averaging operation, and is set to, for example, five bands of the band center frequency.

  Next, in order to establish synchronization between the frequency bands separated in step S11, a delay correction process is performed (step S13). As shown in FIG. 6, the bandpass filters described above have their own delays, and the power peaks of the bands corresponding to the fluctuations in the input signal appear at different times. Therefore, the peak generation time is synchronized by applying a correction delay for each band signal.

  Next, the audibility A characteristic is applied (step S14). Specifically, the power corresponding to the audibility is obtained by multiplying the attenuation rate of the characteristic as shown in FIG.

Next, a time mask effect is calculated (step S15). The time mask effect can be obtained from the time change of the sound. That is, when the sound increases, it can be approximated that the power of the background sound follows with a constant first-order lag, and when the sound decreases, it can be approximated as the power of the background sound attenuates with a constant attenuation factor. Therefore, the power M _t [Pa ² ] of the background sound is calculated based on the following formula 1.

上記の式１において、Ｍ_ｔ：時間マスクのパワー［Ｐａ^２］、Ｍ：背景音のパワー［Ｐａ^２］、Ｓ：音のパワー［Ｐａ^２］、ｎ：サンプル番号、Δｔ：サンプリング時間［sec］、ａ：一次遅れ時定数［sec］、ｂ：減衰率［sec^−１］とする。ただし、Ｍ_ｔ（ｎ−１）＞Ｓ（ｎ）、且つ、Ｍ_ｔ（ｎ）＜Ｓ（ｎ）となった場合は、Ｍ_ｔ（ｎ）＝Ｓ（ｎ）とする。

In the above equation 1, M _t : time mask power [Pa ² ], M: background sound power [Pa ² ], S: sound power [Pa ² ], n: sample number, Δt: sampling time [sec. ], A: primary delay time constant [sec], b: attenuation rate [sec ⁻¹ ]. However, when M _t (n−1)> S (n) and M _t (n) <S (n), M _t (n) = S (n).

  In the above formula 1, the indices a and b may be obtained by experiments or the like. As an experiment, for example, when a sound of 60 dB is generated for T seconds and then a pulse sound of 80 dB is generated for 0.01 second, a single pulse of what dB is compared with a case where a pulse sound is generated alone. Ask if it sounds the same as the sound. Further, after stopping the generation of the sound of 60 dB, the case where the pulse sound of 60 dB is generated for 0.01 seconds after T seconds, compared with the case where the pulse sound is generated for 0.01 seconds alone, how many dB It asks if it sounds the same as a single pulse sound. The above experiment is repeated while changing T, and the change in the mask effect is examined. In the test by the present inventor, as a result of conducting an experiment on the case where the volume and frequency are different, a is about 0.005 and b is about 23.04. However, this value is an example and is not particularly limited.

  Note that this embodiment is an example in which only the mask effect (Post-Masking) by the immediately preceding sound is obtained, and the mask effect (Pre-Masking) by the immediately following sound is not performed. However, the present invention is not limited to this, and the mask effect by the sound immediately after may be handled.

Next, the frequency mask effect is calculated (step S16). As a method for calculating the frequency mask effect, for example, a loudness calculation chart is used. Regarding the use of this chart, the target conditions may be limited. For example, the frequency range to be analyzed is limited to 4 kHz to 20 kHz, the volume range is limited to 40 to 100 dB, and the measurement room is reverberated. Under such a premise, the power of the background sound is approximated as being attenuated at a constant attenuation rate as the frequency difference (logarithmic scale) is opened, and the lower frequency is treated as having no mask effect. Therefore, the power M _f [Pa ² ] of the background sound is calculated based on the following formula 2.

上記の式２において、Ｍ_ｆ：周波数マスクのパワー［Ｐａ^２］、Ｍ：背景音のパワー［Ｐａ^２］、ｆ：周波数［Ｈｚ］、ｆ₀：マスク効果の原因となる周波数［Ｈｚ］、ｃ：オクターブ差に対する減衰率とする。ただし、最低周波数のバンドにおいては、Ｍ_ｆ＝０とする。また、指数のｃは、ｄＢスケールにおける減衰量を２４ｄＢ／ Octaveと近似し、ｃ =５．５２１５とする。

In the above equation 2, M _f : power of frequency mask [Pa ² ], M: power of background sound [Pa ² ], f: frequency [Hz], f ₀ : frequency [Hz] causing mask effect, c: Decay rate relative to octave difference. However, M _f = 0 in the lowest frequency band. The exponent c approximates the attenuation on the dB scale to 24 dB / Octave, and is c = 5.5215.

  Next, the power of the background sound is calculated (step S17). Specifically, the larger one of the background sound power obtained in step S15 and the background sound power obtained in step S16 and the frequency mask obtained in step S16 is set as the background sound power. Note that the obtained background sound power may be used to calculate the frequency mask effect in the adjacent frequency band.

  Next, the power of the entire sound is divided by using the power of the obtained background sound to calculate a power ratio (step S18). Specifically, the power of the entire sound before calculating the mask effect is divided by the power of the background sound obtained in step S17. If the result is smaller than 1, the power is corrected to 1. An example of the power ratio thus obtained is shown in FIG. The upper part of FIG. 8 shows the time waveform of the power of the entire sound and the time waveform of the power of the background sound, and the lower part of FIG. 8 shows the time waveform of the power ratio. In addition, the process from the above step S12 to step S18 calculates simultaneously in each frequency band. Thereby, the power ratio is calculated in all frequency bands.

  Next, the maximum value is acquired from the power ratios obtained in the respective frequency bands (step S19). This is the knock strength. Knock strength means the sound power waveform when the background sound is normalized to 1.

  Next, the obtained knock intensity is converted into a dB scale (step S20). The dB scale is calculated based on 1. This corresponds to the dB volume of the singular sound itself. Hereinafter, the equivalent value of knock strength in dB is referred to as knock dB strength. Although omitted in the control flow of FIG. 3, the knock intensity or the knock dB intensity may be output to the display 20 and displayed as necessary.

  Next, the knock dB intensity is compared with a threshold value (step S21). Here, the threshold corresponds to a tester's auditory judgment standard and is typically about 10 dB, although there are individual differences. However, as an exception, if some kind of sound that is inaudibly indistinguishable from the knocking sound is generated from the mechanical system, it is impossible for humans to distinguish unless a knocking sound larger than that sound is generated. It is necessary to set a larger threshold. Further, when resetting the threshold value based on audibility, a pseudo knock sound described later may be referred to. The installation position and type of the sensor 14 (for example, a directional microphone) may be changed according to the threshold value.

  When the knock dB intensity exceeds the threshold value set in this way, it is determined that the momentary knock state. FIG. 9 shows a time waveform of knock dB intensity in the trace knock state and a determination result of the occurrence of knock. The portion indicated by a circle in the figure shows that the knock dB intensity exceeds the threshold value, and can be determined as an instantaneous knock state.

  Next, the time that is in an instantaneous knock state within a certain time is counted, a knock frequency is obtained (step S22), and the knock frequency is compared with a threshold value (step S23). This threshold is arbitrarily determined based on audibility. When the knock frequency exceeds the threshold value, it is determined that a steady knock state exists. This result corresponds to the result of human auditory evaluation, and the knock determination is completed as described above. The obtained knock determination result may be output to the display 16 as necessary. In the present embodiment, the presence / absence of knock is determined based on the frequency at which the power ratio exceeds the threshold. However, the knock determination method is not limited to this, and various determination methods using the knock intensity are possible. . For example, the knock determination may be performed by comparing a statistic such as the fourth power average value of the knock intensity with a threshold value.

  Next, the pseudo knock sound output flow shown in FIG. 10 is performed. In the output flow of the pseudo knock sound, first, the knock intensity is amplified (step S31). For example, assuming that the amplification factor is d, the knock intensity is raised to the d power. This is equivalent to multiplying the knock dB intensity by d. The amplification factor d may be arbitrarily set to a value that makes it easy to hear the pseudo knock sound to be output. For example, d = 4 is set. As a result, a power waveform obtained by amplifying knock intensity can be obtained.

  Next, reverberation that attenuates at a constant attenuation rate is added to the amplified knock intensity (step S32). Specifically, the arithmetic processing shown by the following formula 3 is performed.

上記の式３において、Ｒ：残響を付加した信号、Ｋ：増幅されたノック強度、ｎ：サンプル信号、Δｔ：サンプリング時間［sec］、ｒ：残響の減衰率［sec^−１］とする。ただし、Ｒ(n-1)＞Ｋ(n)且つＲ(n)＜Ｋ(n)となった場合は、Ｒ(n)＝Ｋ(n)とする。

In the above Equation 3, R: signal with reverberation added, K: amplified knock intensity, n: sample signal, Δt: sampling time [sec], r: reverberation decay rate [sec ⁻¹ ]. However, when R (n-1)> K (n) and R (n) <K (n), R (n) = K (n).

  By applying reverberation, a pseudo knock power signal can be obtained. For example, the signal shown in the upper part of FIG. 11 can be obtained by amplifying the knock intensity in the middle part of FIG. As a result, since the power waveform is such that the instantaneous sound can be heard more continuously, the output sound can be easily heard. If the reverberation is too small, the effect cannot be obtained. If the reverberation is too large, hearing of a knocking sound at another time is disturbed. Therefore, the reverberation rate r is set to an appropriate value of about several hundreds, for example, r = 500.

  Next, after taking the square root of the power waveform and obtaining the effective value (step S33), a process of multiplying the sine wave is performed (step S34). At that time, as the sine wave, a sine wave having an effective value of 1 and a typical knock frequency (for example, 5 kHz) may be applied. For example, the arithmetic processing shown by the following formula 4 is performed.

上記の式４において、Ｔ：擬似ノック音、Ｅ：擬似ノック音の実効値、ｆ：擬似ノック音の周波数［Ｈｚ］、ｎ：サンプル番号、Δｔ：サンプリング時間［sec］とする。なお、本実施の形態では、ｆ＝５０００とする。

In the above Equation 4, T: pseudo knock sound, E: effective value of pseudo knock sound, f: frequency of pseudo knock sound [Hz], n: sample number, Δt: sampling time [sec]. In this embodiment, f = 5000.

  Thereby, a sound signal as shown in the lower part of FIG. 11 is obtained. As a result, even a singular sound generated in a high frequency band near the audible limit is converted to a standard frequency. Therefore, it becomes easier to hear when output as sound.

  Next, a gain corresponding to the output volume is applied (step S35). The pseudo knock sound signal generated in this way is output as a pseudo knock sound from the speaker 18 of FIG. This pseudo knock sound is a signal obtained by amplifying knock intensity, adding reverberation, and converting the frequency. Therefore, even if the original knocking sound is a weak sound, a high-frequency sound, or an instantaneous sound, it is converted into a sound that is easy to hear as a pseudo knocking sound and output.

  Next, the operation of the knock determination device 10 configured as described above will be described.

  As described above, in the present embodiment, the signal detected by the sensor 14 is separated into a plurality of frequency bands, and the power of the background sound is calculated by performing time mask processing based on the immediately preceding signal for each separated frequency band. At the same time, the power of the background sound is calculated by performing frequency mask processing based on the adjacent frequency band, and the larger one of the powers of the two background sounds is selected, and the power of the entire sound is set to the power of the selected background sound. The power ratio is calculated by dividing. Then, the maximum power ratio obtained in a plurality of frequency bands is set as the knock intensity, and the presence or absence of knock is determined based on whether the magnitude and frequency exceed the threshold. Thus, by performing the time mask process and the frequency mask process to determine the presence or absence of knocking, a result equivalent to the auditory evaluation can be obtained. Moreover, since the presence or absence of knock is determined by converting into a numerical value called knock intensity, knock determination can be performed quantitatively.

  Further, in the present embodiment, after knock intensity is amplified, reverberation is added, and an effective value is obtained, a pseudo knock sound signal is generated by multiplying by a sine wave, and this is output from the speaker 18 as sound. Output. Since the pseudo knock sound output in this way is output based on the knock intensity, which is the power ratio with respect to the background sound, the knock sound is emphasized. In addition, since reverberation is added to the knock intensity, the instantaneous knock sound is converted into a sound that is easy to hear. Furthermore, since a sine wave of a predetermined frequency is multiplied, it is converted to a frequency that is easy to hear. Therefore, according to the present embodiment, since the pseudo knock sound that is converted so as to be very easy to hear is output, it is possible to easily and accurately evaluate the audibility based on the audibility.

  In addition, according to the present embodiment, since the quantitative evaluation by the knock intensity and the audibility evaluation by the pseudo knock sound are simultaneously performed, the knock determination can be performed while confirming the reliability of the quantitative knock evaluation. it can.

  In the above-described embodiment, the pseudo knock sound is generated. However, the present invention is not limited to this, and only knock strength calculation and knock determination based on the knock strength may be performed.

  DESCRIPTION OF SYMBOLS 10 ... Knock determination apparatus, 12 ... Engine, 14 ... Sensor, 16 ... Processor, 18 ... Speaker, 20 ... Display, 22 ... Amplifier, 24 ... AD conversion, 26 ... DA conversion, 28 ... Recording medium

Claims (5)

A sensor for detecting engine sound;
The signal obtained by the sensor is subjected to time mask processing based on the signal immediately before and / or immediately after the signal and frequency mask processing based on a signal in an adjacent frequency band to determine the power of the background sound, Calculating means for calculating a power ratio of the background sound to the power as a knock intensity;
Determining means for determining the presence or absence of knock based on the calculated knock intensity;
A knocking determination device characterized by comprising:   The calculation means obtains the power of the background sound subjected to the time mask processing and the power of the background sound subjected to the frequency mask processing, selects the larger one of the powers of the two background sounds, and The knocking determination device according to claim 1, wherein the power ratio is calculated by dividing the power of the knocking.   The calculation means extracts the signal obtained by the sensor by dividing it into a plurality of frequency bands, calculates the power ratio for each frequency band, selects the largest value among the calculated power ratios, and The knock determination device according to claim 2, wherein the knock strength is a knock strength.   4. A pseudo knock sound generating means for creating a sound signal of a pseudo knock sound by adding a reverberant sound intensity to the knock intensity and applying a vibration of a predetermined frequency. The knock determination device according to any one of the above. Detect engine sound,
The detected signal is extracted by dividing it into a plurality of frequency bands,
For each of the extracted frequency bands, time mask processing based on the immediately preceding and / or immediately following signals is performed to obtain the power of the background sound, and frequency mask processing based on the signals in the adjacent frequency bands is performed to reduce the power of the background sound. Seeking
Choose the larger of the powers of the two background sounds and determine the power ratio by dividing the power of the whole sound,
Of the power ratio obtained for each frequency band, the largest value is set as the knock strength,
A knocking determination method characterized by determining the presence or absence of a knock based on the set knock strength.

Images