JP2019196737A - Knocking detector and knocking detection method - Google Patents

Abstract

To suppress wrong detection for knocking.SOLUTION: A knocking detector comprises an acquisition unit, a calculation unit, and a detection unit. The acquisition unit acquires a knock signal associated with knocking of an internal combustion engine. The calculation unit calculates an integrated value for each predetermined section for the knock signal acquired by the acquisition unit. In a case where the maximum value of the integrated value calculated by the calculating unit exists in a predetermined target section, and the knock intensity that is the sum of the integrated values of consecutive sections including the section corresponding to the maximum value exceeds a threshold value, the detection unit detects the case as knocking.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a knocking detection device and a knocking detection method.

  Conventionally, there is a knocking detection device that detects knocking based on a detection result of a knock sensor provided in an internal combustion engine. Such a knocking detection device performs feedback control on the ignition timing of the spark plug based on the knocking detection result (see, for example, Patent Document 1).

Japanese Patent Laid-Open No. 2006-183662

  However, in the related art, when a noise peak is superimposed on the knock signal, there is a possibility that knocking may be erroneously detected based on the noise peak, and improvement in knocking detection accuracy is desired.

  The present invention has been made in view of the above, and an object of the present invention is to provide a knocking detection device and a knocking detection method that can suppress erroneous detection of knocking.

  In order to solve the above-described problems and achieve the object, the knocking detection device according to the embodiment includes an acquisition unit, a calculation unit, and a detection unit. The acquisition unit acquires a knock signal associated with knocking of the internal combustion engine. The calculation unit calculates an integrated value for each predetermined section for the knock signal acquired by the acquisition unit. The detection unit has a knock intensity, which is a sum of integrated values of consecutive sections including a section corresponding to the maximum value, in which the maximum value of the integrated value calculated by the calculation section exists in a predetermined target section. When the threshold value is exceeded, it is detected as knocking.

  According to the present invention, erroneous detection of knocking can be suppressed.

FIG. 1 is a diagram showing an outline of a knocking detection method. FIG. 2 is a block diagram of the knocking detection device. FIG. 3 is a diagram illustrating a specific example of the baseline. FIG. 4 is a diagram illustrating a specific example of processing by the detection unit. FIG. 5 is a diagram illustrating the relationship between the engine speed and / or load and the number of sections of knock strength. FIG. 6 is a flowchart illustrating a processing procedure executed by the knocking detection device.

  Hereinafter, a knocking detection device and a knocking detection method according to an embodiment will be described in detail with reference to the accompanying drawings. In addition, this invention is not limited by this embodiment.

  First, an outline of the knocking detection method according to the embodiment will be described with reference to FIG. FIG. 1 is a diagram showing an outline of a knocking detection method. This knocking detection method is executed by the knocking detection apparatus 1 shown in FIG.

  The knocking detection method according to the embodiment detects knocking based on a knocking signal accompanying knocking of the engine 100 (an example of an internal combustion engine).

  By the way, a noise component may be superimposed on the knock signal. For example, when the intensity of such a noise peak is large, such a noise component may be erroneously detected as knocking. Therefore, in order to detect knocking with high accuracy, it is necessary to separate vibration components and noise based on knocking from the knock signal.

  For example, as a method of separating vibration components and noise peaks based on knocking from a knock signal, a learning model that learns the waveform shape of the knock signal based on knocking is generated, and the learning model and the waveform shape of the actual knock signal are The method of comparing is mentioned. However, since this method requires complicated control, a high-performance microcomputer is required, which may increase manufacturing costs.

  Therefore, in the knocking detection method according to the embodiment, it is determined whether knocking or noise peak is performed by simpler processing. That is, in the knocking detection method according to the embodiment, when knocking occurs, the knocking signal is detected by paying attention to the feature that the amplitude of the knocking signal is maximized and then the amplitude gradually converges.

  Specifically, in the knocking detection method according to the embodiment, as shown in FIG. 1, first, a knock signal is acquired (step S1). The knock signal is a signal accompanying the knocking of engine 100. For example, as a sensor for detecting a knock signal, there are a knock sensor provided for each engine 100 or for each bank of the engine 100, and an in-cylinder pressure sensor for detecting a pressure in each cylinder.

  Subsequently, in the knocking detection method according to the embodiment, an integrated value is calculated for each predetermined section of the knock signal (step S2). For example, in the knocking detection method according to the embodiment, the integrated value is calculated for each predetermined crank angle (for example, 5 CA). The crank angle indicates the rotation angle of the crankshaft.

  Subsequently, in the knocking detection method according to the embodiment, knocking is detected based on the integrated value (step S3). Specifically, in the knocking detection method according to the embodiment, the maximum integrated value P1, which is the maximum integrated value, exists in the target section, and the sum of the integrated values of consecutive sections including the section corresponding to the maximum value. Knock is detected when the knock intensity exceeds the threshold.

  Here, the target section is, for example, a section where the crank angle is from a top dead center (TDC) to a predetermined crank angle after the top dead center. That is, the target section is a section corresponding to a part of the combustion stroke in which the gas expands in the cylinder after the ignition plug is ignited in the cylinder of engine 100.

  Since knocking occurs in the first half of the combustion stroke, the target section is preferably in the range from, for example, top dead center to a predetermined crank angle after top dead center (for example, 45 CA). Alternatively, the target section may be, for example, a range from a first predetermined crank angle after top dead center (including 0CA) to a second crank angle larger than the first predetermined crank angle. In this way, by appropriately setting the target section, it is possible to accurately determine the vibration component and noise peak associated with knocking.

  In the example shown in FIG. 1, the maximum integrated value P1 exists in the target section. For this reason, the maximum integrated value P1 is provisionally determined as a vibration component accompanying knocking. In other words, when the maximum integrated value P1 does not exist in the target section, the maximum integrated value P1 can be determined as a noise peak.

  In the knocking detection method according to the embodiment, when the knock intensity, which is the sum of the integrated values of consecutive sections (sections shown by hatching) including the section corresponding to the maximum integrated value P1, exceeds the threshold, judge.

  That is, in the case of knocking, the amplitude of the knock signal gradually converges, while the noise peak occurs once. Therefore, the knock intensity based on knocking is larger than the knock intensity based on noise peaks.

  That is, in the knocking detection method according to the embodiment, even if a noise peak occurs in the target section, it is possible to determine whether or not the noise peak is based on the knock intensity based on the integrated value.

  As described above, in the knocking detection method according to the embodiment, when the maximum integrated value P1 exists in the target section and the knock intensity by the integrated value exceeds the threshold value, it is detected as knocking.

  In other words, knocking is detected after dividing the peak based on knocking and the noise peak in two stages. Thereby, in the knocking detection method according to the embodiment, erroneous detection of knocking can be suppressed.

  In addition, in the knocking detection method according to the embodiment, as described above, it is possible to suppress erroneous detection of knocking without performing complicated processing using a learning model or the like. As a result, knocking can be accurately detected using an inexpensive microcomputer, and therefore the manufacturing cost of the knocking detection device 1 can be reduced.

  Next, a configuration example of the knocking detection device 1 according to the embodiment will be described with reference to FIG. FIG. 2 is a block diagram of the knocking detection device 1. In the example shown in FIG. 2, a case where knocking detection device 1 is arranged in engine control device 50 in engine ECU (Electronic Control Unit) 200 is shown. However, knocking detection device 1 may be provided outside engine ECU 200, and knocking detection device 1 may be provided outside engine control device 50. FIG. 2 also shows a knock sensor 101 and a crank angle sensor 102 in addition to the engine ECU 200.

  One knock sensor 101 is provided for each engine 100 or for each bank of the engine 100, and is a vibration sensor that converts vibration in the cylinder into an electrical signal, and outputs the electrical signal to the knocking detection device 1. Knock sensor 101 is, for example, a piezoelectric element, and is fixed to a cylinder block of engine 100 or the like.

  The crank angle sensor 102 is a sensor that detects the rotation angle of the crankshaft. The crank angle sensor 102 outputs a crank signal corresponding to the rotation angle of the crankshaft to the knocking detection device 1.

  Engine ECU 200 is an ECU that controls engine 100. The configuration of engine ECU 200 shown in FIG. 2 is an example, and the present invention is not limited to this.

  The knocking detection device 1 includes a control unit 2 and a storage unit 3. The control unit 2 includes a filter unit 21, an acquisition unit 22, an extraction unit 23, a learning unit 24, a calculation unit 25, and a detection unit 26.

  The control unit 2 includes, for example, a computer having a central processing unit (CPU), a read only memory (ROM), a random access memory (RAM), a hard disk drive (HDD), an input / output port, and various circuits.

  The CPU of the computer functions as the filter unit 21, the acquisition unit 22, the extraction unit 23, the learning unit 24, the calculation unit 25, and the detection unit 26 of the control unit 2, for example, by reading and executing a program stored in the ROM. To do.

  In addition, at least some or all of the filter unit 21, the acquisition unit 22, the extraction unit 23, the learning unit 24, the calculation unit 25, and the detection unit 26 of the control unit 2 may be ASIC (Application Specific Integrated Circuit) or FPGA (Field Programmable). Gate Array) or the like can also be configured.

  The storage unit 3 corresponds to, for example, a RAM or a flash ROM. The RAM and flash ROM can store determination condition information 31, baseline information 32, integrated value information 33, and various program information. The knocking detection device 1 may acquire the above-described program and various information via another computer or a portable recording medium connected via a wired or wireless network.

  The filter unit 21 of the control unit 2 is, for example, a band pass filter. The filter unit 21 passes only an electrical signal in a frequency band (for example, 7 to 8 kHz) including a signal component due to knocking to the electrical signal input from the knock sensor 101 and outputs the electrical signal to the acquisition unit 22. Hereinafter, the electrical signal that has passed through the filter unit 21 is referred to as a knock signal.

  Acquisition unit 22 acquires a knock signal associated with knocking of engine 100. The extraction unit 23 determines a gate section that is a section for detecting knocking based on the crank signal, and extracts a knock signal in the gate section. The gate section is a section including the above-described target section, for example, a section from top dead center to 90 CA after top dead center. The extraction unit 23 outputs the extracted knock signal of the gate section to the learning unit 24.

  The learning unit 24 learns a baseline corresponding to the noise intensity for each section for the knock signal. Specifically, the learning unit 24 performs knock signal baseline learning when the engine 100 is operating in an operation region where knocking is unlikely to occur.

  First, the learning unit 24 acquires accelerator opening information related to the current accelerator opening from the engine control device 50. Then, the learning unit 24 refers to the determination condition information 31 stored in the storage unit 3 and determines whether or not knocking is likely to occur at the current accelerator opening.

  Here, the determination condition information 31 is information associated with whether or not knocking is likely to occur in the operating region of the engine 100 based on the accelerator opening. More specifically, the determination condition information 31 takes into account the engine speed and load according to the accelerator opening, the intake air amount in the cylinder, the fuel injection amount, the ignition timing in the cylinder, and the like. Is created in advance.

  Then, when it is difficult for knocking to occur, the learning unit 24 learns the baseline of each section based on the knock signal, and stores the learned baseline as the baseline information 32 in the storage unit 3. Since the knock intensity by the knock signal extracted in a situation where knocking is difficult to occur can be estimated as the noise intensity, the baseline of each section is set based on the noise intensity extracted in each section.

  FIG. 3 is a diagram illustrating a specific example of the baseline. As shown in FIG. 3, as shown in FIG. 3, in the present embodiment, the baseline BL is learned for each section (for example, in increments of 5 CA).

  For example, the value (intensity) of the baseline BL is relatively small in a section where the S / N ratio of the knock signal is good, and is large in a region where the S / N ratio is bad.

  Thus, the learning unit 24 can improve the reliability of the integrated value calculated by the calculation unit 25 described later by learning the baseline BL for each section. As a result, knocking can be detected with high accuracy.

  Here, the case where the baseline BL has a different value for each section is shown, but the baseline BL may have the same value in all sections. In such a case, for example, the learning unit 24 can learn the baseline BL having the same value in all the sections by, for example, leveling the baseline BL.

  Returning to the description of FIG. 2, the calculation unit 25 will be described. The calculation unit 25 calculates an integrated value for each predetermined section (every 5 CA) for the knock signal acquired by the acquisition unit 22.

  Specifically, the calculation unit 25 accumulates the absolute value of the strength of the knock signal with reference to the baseline BL learned by the learning unit 24 for each predetermined section. More specifically, the knock signals equal to or higher than the baseline BL are integrated for each predetermined section as an integration target. Then, the calculation unit 25 stores the calculated integrated value as integrated value information 33 in the storage unit 3.

  Here, as described above, the baseline BL is a value learned for each section. That is, in the present embodiment, the baseline BL is set based on the S / N ratio of the knock signal for each section.

  Therefore, the calculation unit 25 can calculate the integrated value in a state where the influence of the noise component included in the knock signal is reduced by calculating the integrated value based on the baseline BL. Thereby, the calculation unit 25 can calculate the integrated value with high reliability.

  The detecting unit 26 has a maximum integrated value (maximum integrated value P1) calculated by the calculating unit 25 in a predetermined target section, and a sum of integrated values of consecutive sections including a section corresponding to the maximum value. Is detected as knocking when the knock intensity exceeds the threshold.

  First, the detection unit 26 determines whether or not the maximum integrated value P1 exists in the target section. If the maximum integrated value P1 exists in the target section, the process proceeds to a process described later. If the maximum integrated value P1 exists outside the target section, the maximum integrated value P1 is determined as a noise peak.

  Here, for example, as described above, the target section is a section from the top dead center to 45 CA after the top dead center, but the detection unit 26 determines that the target section is an operation state of the actuator provided in the engine 100. It is also possible to change based on this.

  For example, when the case where the actuator is a variable intake / exhaust valve is described, the knock sensor 101 may detect vibration of the variable intake / exhaust valve provided in the surrounding cylinder as a noise peak.

  For example, when the variable intake / exhaust valve is closed, vibration may occur in the cylinder, and a noise peak based on vibration applied to the knock signal may be superimposed. That is, it is possible to predict the timing at which a noise peak occurs based on the operating state of the actuator.

  For this reason, the detection part 26 can also predict the generation timing of a noise peak, and can also set the area where a noise peak is hard to be included as said object area. That is, the detection unit 26 can correct the target section in accordance with the operation state of the actuator.

  Thereby, since the target section from which the noise peak is eliminated can be set, it is possible to suppress erroneous detection of knocking. The actuator is not limited to the variable intake / exhaust valve, and may be another actuator.

  FIG. 4 is a diagram illustrating a specific example of processing by the detection unit 26. Note that A in FIG. 4 indicates a crank signal, and # 1 to # 4 indicate each cylinder. 4B shows the waveform of the knock signal, and C in FIG. 4 shows the integrated value of the knock signal. Further, D in FIG. 4 shows a detection result by the detection unit 26.

  For example, in # 1, # 4, and # 2, the maximum integrated value P1 exists in the target section, and in # 3, the maximum integrated value P1 exists outside the target section. Therefore, the detection unit 26 detects # 3 as a noise peak.

  Subsequently, the detection unit 26 calculates knock magnitudes for # 1, # 4, and # 2. The knock intensity is the sum of the integrated values of consecutive sections including the section corresponding to the maximum integrated value P1, and is indicated by hatching in C of FIG.

  Here, since # 1 and # 2 have large knock strength and exceed the threshold value, detection unit 26 detects knocking for # 1 and # 2. On the other hand, since the knock intensity is small for # 4 and does not exceed the threshold value, the detection unit 26 detects # 4 as a noise peak.

  As described above, the detection unit 26 can accurately determine whether the vibration component is based on knocking or the noise peak by detecting knocking based on the position of the maximum integrated value P1 and the knocking intensity based on the integration.

  By the way, when knocking occurs, the amplitude of the knock signal becomes maximum, and thereafter, the amplitude gradually decreases. That is, when knocking occurs, the integrated value gradually decreases from the maximum integrated value P1.

  For this reason, the detection unit 26 may determine the sum of the integrated values in the section following the section corresponding to the maximum integrated value P1 as the knock intensity. That is, you may decide not to use the integrated value of the area before the maximum integrated value P1 for calculation of knock intensity | strength.

  Thereby, since the noise peak generated immediately before the maximum integrated value P1 can be excluded and knocking can be detected, the detection accuracy of knocking can be improved.

  Further, as described above, the integrated value is calculated at a predetermined crank angle (5 CA increments). For this reason, the higher the engine speed and / or load, the shorter the time interval for calculating the integrated value.

  Therefore, as the rotational speed and / or load of engine 100 is higher, a vibration component based on knocking is included in the integrated value of more sections. In other words, the vibration component based on knocking is included in a smaller section as the rotational speed and / or load of the engine 100 is lower.

  In other words, when the engine speed and / or load is low, if the integrated value of many sections is the knock intensity, the possibility that the knock intensity includes a noise peak increases. In addition, when the engine speed and / or load of engine 100 is high, if the integrated value in a small section is the knock strength, the knock strength may be smaller than the actual knock strength.

  Therefore, detection unit 26 can change the number of sections for calculating the knock magnitude according to the rotation speed of engine 100 and / or the load. FIG. 5 is a diagram showing the relationship between the rotational speed and / or load of engine 100 and the number of sections of knock strength.

  It is noted that detection unit 26 can determine the rotation speed and / or load of engine 100 based on the accelerator opening. That is, when the accelerator opening is large, that is, when the accelerator is stepped on strongly, the intake air amount in the cylinder increases, and the rotation speed and / or load of engine 100 increases.

  As shown in FIG. 5, the detection unit 26 increases the number of knock strength sections as the engine speed and / or load of the engine 100 is higher, and decreases the number of knock strength sections as the load of the engine 100 is lower. .

  By doing so, the detection unit 26 can obtain the knock magnitude based on the integrated value of an appropriate section according to the load of the engine 100. In other words, it is possible to eliminate the noise peak and noise component and calculate the knock intensity using only the vibration component based on knocking.

  Therefore, since the detection accuracy of knocking can be improved, it is possible to suppress erroneous detection of knocking. The number of sections shown in FIG. 5 is an example, and can be arbitrarily changed.

  Returning to the description of FIG. 2, the engine control device 50 will be described. The engine control device 50 controls the ignition timing of the spark plug provided in each cylinder based on the detection result of the knocking detection device 1.

  For example, the engine control device 50 performs advance angle control that advances the ignition timing when knocking is not detected, and performs retard angle control that delays the ignition timing when knocking is detected.

  As described above, the knocking detection device 1 according to the present embodiment can accurately detect knocking. For this reason, the engine control device 50 can appropriately control the ignition timing, so that the fuel consumption of the engine 100 can be improved.

  Next, a processing procedure executed by the knocking detection apparatus 1 according to the embodiment will be described with reference to FIG. FIG. 6 is a flowchart showing a processing procedure executed by the knocking detection apparatus 1.

  As shown in FIG. 6, first, the acquisition unit 22 of the control unit 2 acquires a knock signal (step S101), and the extraction unit 23 extracts a knock signal in the gate section (step S102).

  Subsequently, the calculation unit 25 calculates an integrated value for each predetermined section with respect to the knock signal (step S103). Subsequently, the detection unit 26 determines whether or not the maximum integrated value P1 exists in the target section (step S104).

  When the maximum integrated value P1 is present in the target section (step S104, Yes), the detection unit 26 calculates the knock intensity by the sum of the integrated values of the predetermined number of sections described above (step S105).

  Subsequently, the detection unit 26 determines whether or not the knock magnitude is larger than the threshold (step S106). When the knock magnitude is larger than the threshold (step S106, Yes), the knocking is detected as knocking (step S107). The process ends.

  On the other hand, when the maximum integrated value is outside the target section (step S104, No), the detection unit 26 ends the process as it is, and when the knock intensity is equal to or less than the threshold value (step S106, No), the process ends. . That is, it is determined that there is no knock if the knock intensity is equal to or less than the threshold.

  As described above, the knocking detection device 1 according to the embodiment includes the acquisition unit 22, the calculation unit 25, and the detection unit 26. Acquisition unit 22 acquires a knock signal associated with knocking of engine 100 (an example of an internal combustion engine). The calculation unit 25 calculates an integrated value for each predetermined section for the knock signal acquired by the acquisition unit 22.

  The detecting unit 26 has a maximum value of the integrated value calculated by the calculating unit 25 in a predetermined target section, and a knock intensity that is a sum of integrated values of consecutive sections including a section corresponding to the maximum value has a threshold value. When exceeding, it detects as knocking. Therefore, according to the knocking detection device 1 according to the embodiment, erroneous detection of knocking can be suppressed.

  By the way, in the above-described embodiment, the case where the learning unit 24 learns the baseline BL and calculates the knock intensity based on the baseline BL has been described. For example, the threshold value for the knock magnitude may be changed according to the load of the engine 100.

  For example, when the engine speed and / or load of the engine 100 is large, the threshold value is set to a high value, and when the engine speed and / or load of the engine 100 is small, the threshold value is set to a low value. As a result, an appropriate threshold value can be set according to the rotational speed and / or load of engine 100, so that erroneous detection of knocking can be suppressed.

  In the above-described embodiment, the case where the integrated value is calculated based on the absolute value of the knock signal with reference to the baseline BL has been described. However, based on the baseline obtained by adding an offset amount to the baseline BL. Thus, the integrated value may be calculated. Further, the offset amount can be calculated using machine learning based on the rotation speed and / or load of the engine 100, for example.

  Further effects and modifications can be easily derived by those skilled in the art. Thus, the broader aspects of the present invention are not limited to the specific details and representative embodiments shown and described above. Accordingly, various modifications can be made without departing from the spirit or scope of the general inventive concept as defined by the appended claims and their equivalents.

DESCRIPTION OF SYMBOLS 1 Knock detection apparatus 21 Filter part 22 Acquisition part 23 Extraction part 24 Learning part 25 Calculation part 26 Detection part 50 Engine control apparatus 100 Engine (an example of an internal combustion engine)
101 Knock sensor 102 Crank angle sensor 200 Engine ECU

Claims (6)

An acquisition unit for acquiring a knock signal associated with knocking of the internal combustion engine;
A calculation unit that calculates an integrated value for each predetermined section for the knock signal acquired by the acquisition unit;
When the maximum value of the integrated value calculated by the calculation unit exists in a predetermined target section, and the knock intensity that is the sum of the integrated values of consecutive sections including the section corresponding to the maximum value exceeds a threshold value And a detection unit that detects the knocking as a knocking detection device. The detector is
The knocking detection device according to claim 1, wherein a range from the top dead center of the internal combustion engine to a predetermined crank angle after the top dead center is set as the target section. The detector is
The knock detection device according to claim 1 or 2, wherein a sum of the integrated values in a section subsequent to a section corresponding to the maximum value is used as the knock intensity. The detector is
4. The knocking detection device according to claim 1, wherein the number of consecutive sections is changed in accordance with a rotational speed and / or a load of the internal combustion engine. A learning unit for learning a baseline corresponding to the noise intensity for each section of the knock signal;
The calculation unit includes:
The knocking detection apparatus according to any one of claims 1 to 4, wherein the integrated value is calculated based on the baseline learned by the learning unit. An acquisition step of acquiring a knock signal associated with knocking of the internal combustion engine;
A calculation step of calculating an integrated value for each predetermined section for the knock signal acquired by the acquisition step;
When the maximum value of the integrated value calculated by the calculating step exists within a predetermined target section, and the knock intensity that is the sum of the integrated values of consecutive sections including the section corresponding to the maximum value exceeds a threshold value And a detection step of detecting as knocking of the internal combustion engine.

Images