Control device for internal combustion engine

Abstract

The purpose of the present invention is to provide a control device which is for an internal combustion engine and which can reduce the difference between a corrected air flowrate and an actual air flowrate by appropriately correcting an output value of a flowrate sensor provided in an intake air passage of the internal combustion engine. A control device 100 for an internal combustion engine 10 comprises: a correction amount derivation unit 107 that derives a correction amount for an output value of a flowrate sensor 17; and a flowrate correction unit 113 that calculates a flowrate of air flowing through an intake air passage 20 by correcting the output value of the flowrate sensor 17 on the basis of the correction amount. The correction amount derivation unit 107 includes: a normalization unit 108 that normalizes the output value of the flowrate sensor 17 for each wave form of one wavelength acquired from an output value, of the flowrate sensor 17, acquired during a prescribed period; and an NN calculation unit 109 that calculates a correction amount for the output value of the flowrate sensor 17 during the prescribed period, from the normalized output value of the flowrate sensor 17.

Description

[Technical Field]

[0001] The present invention relates to a control device for an internal combustion engine. [Background technology]

[0002] BACKGROUND ART There is known an invention relating to a control device for an internal combustion engine that calculates the flow rate of air flowing through an intake passage of the internal combustion engine based on the output value of a flow rate sensor (air flow sensor) provided in the intake passage of the internal combustion engine (for example, Patent Document 1).

[0003] Patent Document 1 discloses an internal combustion engine control device having a fundamental frequency derivation unit that derives a fundamental frequency, which is the frequency of a flow rate pulsation waveform that corresponds to the rotation speed of the internal combustion engine; a flow rate amplitude calculation unit that extracts multiple high frequencies of frequencies equal to or higher than the fundamental frequency from a pulsation waveform based on the output value of a flow rate sensor as flow rate high frequencies and calculates the amplitude of the flow rate high frequencies for each frequency; a correction amount derivation unit that derives a correction amount based on the amplitude of the flow rate high frequencies for each frequency; and a flow rate calculation unit that calculates the air flow rate using the output value of the flow rate sensor and the correction amount.

[0004] The correction amount derivation unit disclosed in Patent Document 1 includes a neural network model, in which the amplitude of high-frequency flow for each frequency is set for each unit in the input layer, a weight, a bias, and an activation function are set for each unit in the intermediate layer, and a correction amount is set for each unit in the output layer. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] International Publication No. 2020 / 066548 Summary of the Invention [Problem to be solved by the invention]

[0006] The correction amount derivation unit disclosed in Patent Document 1 may derive the same correction amount if the amplitude of the fundamental wave and the amplitude of the high frequency of one pulsation waveform are the same as the amplitude of the fundamental wave and the amplitude of the high frequency of another pulsation waveform, because the pulsation rates of both are the same, even if the phases of the fundamental wave or the high frequency of both pulsation waveforms are different. In other words, the correction amount derivation unit disclosed in Patent Document 1 may derive the correction amount without distinguishing between pulsation waveforms having different characteristics due to differences in the phase of the fundamental wave or the high frequency. Therefore, the internal combustion engine control device disclosed in Patent Document 1 may produce an error between the corrected air flow rate and the actual air flow rate, leaving room for improvement.

[0007] The present invention has been made in consideration of the above, and aims to provide a control device for an internal combustion engine that can appropriately correct the output value of a flow sensor provided in the intake flow path of an internal combustion engine, thereby reducing the error between the corrected air flow rate and the actual air flow rate. [Means for solving the problem]

[0008] In order to solve the above problem, the control device for an internal combustion engine of the present invention is a control device for an internal combustion engine that calculates the flow rate of air flowing through an intake flow path of the internal combustion engine based on the output value of a flow sensor provided in the intake flow path, and is characterized in that it includes a correction amount derivation unit that derives a correction amount for the output value of the flow sensor, and a flow rate correction unit that calculates the flow rate of the air by correcting the output value of the flow sensor based on the correction amount, and the correction amount derivation unit includes a normalization unit that normalizes the output value of the flow sensor for each waveform of one wavelength obtained from the output value of the flow sensor acquired over a predetermined period, and a calculation unit that calculates the correction amount for the output value of the flow sensor over the predetermined period from the normalized output value of the flow sensor. [Effects of the Invention]

[0009] According to the present invention, the output value of a flow rate sensor provided in an intake passage of an internal combustion engine can be appropriately corrected, thereby reducing the error between the corrected air flow rate and the actual air flow rate. Problems, configurations, and effects other than those described above will become apparent from the following description of the embodiments. [Brief explanation of the drawings]

[0010] [Figure 1] 1 is a diagram showing a schematic configuration of an internal combustion engine system equipped with a control device according to an embodiment of the present invention; [Figure 2] FIG. 2 is a block diagram showing the functional configuration of the control device shown in FIG. [Figure 3] FIG. 4 is a diagram showing an output value of a flow rate sensor. [Figure 4] FIG. 4 is a diagram illustrating a correction amount derivation unit. [Figure 5] FIG. 4 is a diagram illustrating a pulsation determination unit. [Figure 6] 4 is a flowchart of a process performed by a control device. [Figure 7] 10A and 10B are diagrams showing the results of verifying the error between the correction amount derived by the correction amount derivation unit and the correction amount actually required. [Figure 8] 4 is a diagram illustrating an output value of a flow sensor acquired by a method different from that shown in FIG. 3. DETAILED DESCRIPTION OF THE INVENTION

[0011] Hereinafter, embodiments of the present invention will be described with reference to the drawings. Note that components with the same reference numerals in each embodiment have the same functions in each embodiment unless otherwise specified, and description thereof will be omitted.

[0012] FIG. 1 is a diagram showing a schematic configuration of an internal combustion engine system 1 equipped with a control device 100 according to this embodiment.

[0013] The internal combustion engine system 1 includes, for example, an internal combustion engine 10, an intake passage 20, an exhaust passage 30, a turbocharger 40, an exhaust gas recirculation (EGR) passage 50, and a control device 100. The control device 100 is configured, for example, by an electronic control unit (ECU) including a processor such as a CPU and storage devices such as a ROM and a RAM. The functions of the control device 100 are realized by the CPU executing a program stored in the ROM.

[0014] In the following, the configuration of each part of the internal combustion engine system 1 will first be described, and then the configuration of the control device 100 for the internal combustion engine 10 will be described.

[0015] The internal combustion engine 10 includes, for example, an intake valve 11, an exhaust valve 12, a fuel injection valve 13, a spark plug 14, a knock sensor 15, a crank angle sensor 16, and a flow rate sensor 17. The internal combustion engine 10 is connected to an intake passage 20 and an exhaust passage 30. The intake valve 11 and the exhaust valve 12 each have a variable valve mechanism. The variable valve mechanism of the intake valve 11 includes a sensor 11s that detects the opening and closing phase of the intake valve 11, and is configured to continuously vary the phase of the intake valve 11. The variable valve mechanism of the exhaust valve 12 includes a sensor 12s that detects the opening and closing phase of the exhaust valve 12, and is configured to continuously vary the phase of the exhaust valve 12.

[0016] The fuel injection valve 13 is, for example, a direct injection valve that injects fuel directly into the cylinder of the internal combustion engine 10. However, the fuel injection valve 13 may also be a port injection valve that injects fuel into an intake port. The spark plug 14 has an electrode exposed inside the cylinder of the internal combustion engine 10, and ignites a combustible mixture by sparking. The knock sensor 15 is provided in the cylinder block of the internal combustion engine 10 and detects the presence or absence of knock occurring in the combustion chamber. The crank angle sensor 16 is provided on the crankshaft of the internal combustion engine 10 and outputs a signal corresponding to the rotation angle of the crankshaft to the ECU 60 for each combustion cycle as a signal indicating the rotation speed of the crankshaft.

[0017] The intake flow path 20 has, for example, an upstream section 21, a midstream section 22, a downstream section 23, and a bypass section 24. The upstream section 21 is a flow path that connects an air cleaner (not shown) and the turbocharger 40. The midstream section 22 is a flow path that connects the turbocharger 40 and the downstream section 23 of the intake flow path 20. The downstream section 23 is an intake manifold connected to the internal combustion engine 10. The bypass section 24 is a flow path that connects the upstream section 21 and the midstream section 22.

[0018] An upstream portion 21 of the intake air flow path 20 is provided with, for example, a flow rate sensor 17 that measures the flow rate of air flowing through the intake air flow path 20. The flow rate sensor 17 is, for example, an air flow sensor that includes an intake air temperature sensor.

[0019] The flow sensor 17 is, for example, a hot wire airflow sensor. The flow sensor 17 is provided with, for example, a bypass passage, which is a secondary passage that bypasses the air flowing through the intake passage 20, which is the main passage. The flow sensor 17 has a sensor element installed in this bypass passage for detecting the flow velocity of the air. The flow sensor 17 can prevent dust and water from adhering to the sensor element by devising the shape of the bypass passage.

[0020] The flow field shapes are different between the main passage through which the mainstream flows and the bypass passage through which the bypass flow flows. The shape loss coefficient and friction loss coefficient of the mainstream flow field are different from those of the bypass flow field. Therefore, the flow fields of the mainstream and the bypass flow are based on different momentum equations.

[0021] A hot-wire airflow sensor's main component is a heating resistor placed in the airflow to be measured, and a bridge circuit is configured so that the current flowing through the heating resistor increases when the intake airflow rate is high and decreases when the intake airflow rate is low. In other words, the flow sensor 17 is configured to, for example, extract the airflow rate as a voltage signal from the current flowing through the heating resistor. The flow sensor 17 detects the amount of heat dissipated due to the local flow of air near the sensor element in the bypass passage, and outputs a voltage signal corresponding to the flow rate of air through the bypass passage.

[0022] The midstream section 22 of the intake air flow path 20 is provided with, for example, an intercooler 22a, a supercharger temperature sensor 22b, and a throttle valve 22c. The intercooler 22a cools and lowers the temperature of air that has been adiabatically compressed by the compressor 41 of the turbocharger 40. The supercharger temperature sensor 22b is disposed downstream of the intercooler 22a and measures the temperature of the air cooled by the intercooler 22a. The temperature of the air measured by the supercharger temperature sensor 22b is referred to as the supercharger temperature. The throttle valve 22c is disposed downstream of the supercharger temperature sensor 22b and throttles the intake air flow path 20 to control the amount of air flowing into the cylinders of the internal combustion engine 10. The throttle valve 22c is, for example, an electronically controlled butterfly valve whose valve opening can be controlled independently of the accelerator pedal depression amount by the driver.

[0023] The downstream portion 23 of the intake passage 20 is provided with, for example, a boost pressure sensor 23a and a flow enhancement valve 23b. The boost pressure sensor 23a is located downstream of the throttle valve 22c provided in the midstream portion 22. The downstream portion 23 of the intake passage 20, which is an intake manifold connected to the internal combustion engine 10, may be integrated with the intercooler 22a. In this case, the volume from downstream of the compressor 41 of the turbocharger 40 to the cylinder of the internal combustion engine 10 can be reduced, thereby improving acceleration / deceleration responsiveness. The flow enhancement valve 23b increases turbulence in the flow inside the cylinder of the internal combustion engine 10 by causing a biased flow in the intake air.

[0024] The exhaust flow path 30 has, for example, an upstream portion 31, a downstream portion 32, and a bypass portion 33. The upstream portion 31 is an exhaust manifold that connects the internal combustion engine 10 and the turbocharger 40. The downstream portion 32 is a flow path that connects the turbocharger 40 and a muffler (not shown). The bypass portion 33 is a flow path that connects the upstream portion 31 and the downstream portion 32 of the exhaust flow path 30. The downstream portion 32 of the exhaust flow path 30 is provided with, for example, an air-fuel ratio sensor 32a and an exhaust purification catalyst 32b.

[0025] The air-fuel ratio sensor 32a is provided downstream of the turbine 42 of the turbocharger 40, and outputs a signal indicating the detected oxygen concentration, i.e., the air-fuel ratio, to the control device 100 of the internal combustion engine 10. The exhaust purification catalyst 32b is provided downstream of the air-fuel ratio sensor 32a, and purifies harmful exhaust gas components such as carbon monoxide, nitrogen compounds, and unburned hydrocarbons in the exhaust gas through catalytic reaction.

[0026] The turbocharger 40 is composed of a compressor 41 and a turbine 42, and is equipped with, for example, an air bypass valve 43 provided in the bypass section 24 of the intake passage 20 and a wastegate valve 44 provided in the bypass section 33 of the exhaust passage 30. The compressor 41 has compressor vanes, and the upstream section 21 of the intake passage 20 is connected to the upstream side of the compressor vanes, and the midstream section 22 of the intake passage 20 is connected to the downstream side of the compressor vanes.

[0027] The turbine 42 has turbine blades connected to compressor blades, with the upstream portion 31 of the exhaust passage 30 connected to the upstream side of the turbine blades and the downstream portion 32 of the exhaust passage 30 connected to the downstream side of the turbine blades. The turbine 42 converts the energy of the exhaust gas flowing through the exhaust passage 30 into rotational energy using the turbine blades. The compressor 41 compresses the air flowing through the intake passage 20 by rotating the compressor blades.

[0028] The air bypass valve 43 prevents an excessive increase in pressure from downstream of the compressor 41 to the upstream of the throttle valve 22c under the control of the control device 100 of the internal combustion engine 10. When the throttle valve 22c is suddenly closed in a supercharging state, the air bypass valve 43 is opened under the control of the control device 100 of the internal combustion engine 10, causing the compressed intake air downstream of the compressor 41 to flow back to the upstream of the compressor 41 through the bypass portion 24 of the intake passage 20. As a result, it becomes possible to reduce the supercharging pressure.

[0029] The wastegate valve 44 is an electrically operated valve whose opening degree can be freely controlled in response to the boost pressure under the control of the control device 100 of the internal combustion engine 10. The control device 100 of the internal combustion engine 10 adjusts the opening degree of the wastegate valve 44 based on the boost pressure detected by a boost pressure sensor 23a provided in the downstream portion 23 of the intake passage 20. By allowing a portion of the exhaust gas to pass through the bypass portion 33 of the exhaust passage 30, the work that the exhaust gas imparts to the turbine 42 can be reduced, and as a result, the boost pressure can be maintained at a target pressure.

[0030] The EGR flow path 50 has one end connected to the downstream portion 32 of the exhaust flow path 30 and the other end connected to the upstream portion 21 of the intake flow path 20, and diverts exhaust gas from downstream of the exhaust purification catalyst 32b and recirculates it upstream of the compressor 41. The EGR flow path 50 is provided with, for example, an EGR cooler 51, an EGR valve 52, a temperature sensor 53, and a differential pressure sensor 54. The EGR cooler 51 cools the exhaust gas. The EGR valve 52 is provided downstream of the EGR cooler 51 and controls the flow rate of the exhaust gas. The temperature sensor 53 detects the temperature of the exhaust gas upstream of the EGR valve 52. The differential pressure sensor 54 detects the differential pressure between the upstream and downstream sides of the EGR valve 52.

[0031] The control device 100 of the internal combustion engine 10 controls each part of the internal combustion engine system 1 and executes various data processing. The control device 100 is connected to the various sensors and actuators described above. The various actuators drive, for example, the throttle valve 22c, the fuel injection valve 13, the intake valve 11 and the exhaust valve 12 with a variable valve mechanism, and the EGR valve 52. The control device 100 controls the operation of these various actuators. Furthermore, the control device 100 detects the operating state of the internal combustion engine 10 based on signals input from the various sensors, and ignites the spark plug 14 at a timing determined according to the operating state.

[0032] Fig. 2 is a block diagram showing the functional configuration of the control device 100 shown in Fig. 1. Fig. 3 is a diagram showing the output value of the flow rate sensor 17. Fig. 4 is a diagram explaining the correction amount derivation unit 107. Fig. 5 is a diagram explaining the pulsation determination unit 111.

[0033] The control device 100 of the internal combustion engine 10 calculates the flow rate of air flowing through the intake passage 20 based on the output value of a flow rate sensor 17 provided in the intake passage 20 of the internal combustion engine 10 .

[0034] The control device 100 includes an A / D conversion unit 101, a voltage / flow rate conversion unit 102, a data array creation unit 103, a rotational speed calculation unit 104, an angle determination unit 105, an average flow rate calculation unit 106, a correction amount derivation unit 107, and a flow rate correction unit 113.

[0035] The A / D conversion unit 101 performs A / D conversion to sample an analog voltage signal (hereinafter also referred to as "flow sensor signal") output from the flow sensor 17, thereby converting the signal into a digital signal. Specifically, the A / D conversion unit 101 samples the flow sensor signal at regular crank angles. This regular crank angle may be, for example, 6 degrees as shown in FIG. 3. The timing at which this regular crank angle arrives may be notified to the A / D conversion unit 101 by an angle determination unit 105 that determines the crank angle based on a signal output from the crank angle sensor 16 (hereinafter also referred to as "crank angle sensor signal").

[0036] The voltage / flow rate converter 102 converts the voltage value indicated by the digital signal converted by the A / D converter 101 into a flow rate value of air flowing near a sensor element arranged in the bypass flow path of the flow sensor 17. The voltage / flow rate converter 102 may perform voltage / flow rate conversion using a preset voltage / flow rate conversion table. In this embodiment, the output value of the flow rate sensor 17 refers to the flow rate value acquired through each conversion by the A / D converter 101 and the voltage / flow rate converter 102. Each circle in FIG. 3 indicates the output value of the flow rate sensor 17.

[0037] The data array creation unit 103 creates a data array whose elements are the output values ​​of the flow sensor 17 sampled at regular crank angles. The data array creation unit 103 may create a single data array of a plurality of output values ​​of the flow sensor 17 acquired over a predetermined period. This predetermined period is not particularly limited, and may be an interval defined by the crank angle. The predetermined period may be the interval between intake strokes of the internal combustion engine 10. For example, if the internal combustion engine 10 has three cylinders, the predetermined period may be 240 degrees as shown in FIG. 3. The timing at which the predetermined period has elapsed may be notified to the data array creation unit 103 by the angle determination unit 105, for example.

[0038] The rotation speed calculation unit 104 calculates the rotation speed of the internal combustion engine 10 based on the crank angle sensor signal.

[0039] The angle determination unit 105 determines, based on the crank angle sensor signal, that the certain crank angle has arrived, which defines the timing for sampling the output value of the flow rate sensor 17. Furthermore, the angle determination unit 105 determines, based on the crank angle sensor signal, that the predetermined period defining one data array has elapsed.

[0040] The average flow rate calculation unit 106 calculates the average flow rate of air flowing through the intake flow path 20 over a predetermined period of time based on the output value of the flow rate sensor 17. The average flow rate calculation unit 106 may calculate the average value of the output values ​​of the flow rate sensor 17 acquired over the predetermined period of time.

[0041] The correction amount derivation unit 107 derives a correction amount for the output value of the flow rate sensor 17. The correction amount derivation unit 107 includes a normalization unit 108, an NN calculation unit 109, an NN selection unit 110, a pulsation determination unit 111, and a switching unit 112.

[0042] The normalization unit 108 normalizes the output values ​​of the flow sensor 17 for each waveform of one wavelength obtained from the output values ​​of the flow sensor 17 acquired during a predetermined period. For example, as shown in FIG. 4, the normalization unit 108 sets the maximum and minimum values ​​of the waveform of one wavelength obtained from the output values ​​of the flow sensor 17 to 1.0 and 0.0, respectively. The normalization unit 108 then normalizes the waveform of one wavelength so that it falls within the range from 0.0 to 1.0. Here, if the predetermined period is the interval between intake strokes of the internal combustion engine 10, the waveform obtained from the output values ​​of the flow sensor 17 acquired during the predetermined period may represent a waveform of one wavelength. The normalization unit 108 extracts and normalizes the output values ​​of the flow sensor 17 for the predetermined period from the data array created by the data array creation unit 103.

[0043] The NN calculation unit 109 calculates a correction amount for the output value of the flow sensor 17 for a predetermined period from the output value of the flow sensor 17 normalized by the normalization unit 108. The NN calculation unit 109 has a neural network model. The normalized output value of the flow sensor 17 is set in the input layer of the neural network model. A correction amount for the output value of the flow sensor 17 for a predetermined period is set in the output layer of the neural network model. The weight, bias, and activation function of each neuron are set in the intermediate layer of the neural network model. The neural network model is constructed by machine learning the weight and bias of each neuron using an error backpropagation method or the like, and is pre-installed in the NN calculation unit 109. The machine learning may be supervised learning.

[0044] The NN calculation unit 109 may have a machine learning model other than the neural network model. Also, instead of the neural network model, the NN calculation unit 109 may have a correction amount map configured as a table showing the correspondence between the output value of the flow sensor 17 and the correction amount.

[0045] The NN calculation unit 109 of this embodiment has a plurality of neural network models that differ depending on the rotation speed of the internal combustion engine 10. The NN calculation unit 109 calculates a correction amount for the output value of the flow sensor 17 for a predetermined period using a neural network model selected by an NN selection unit 110 from the plurality of neural network models.

[0046] The NN selection unit 110 selects a neural network model to be used for calculating the correction amount from among the multiple neural network models possessed by the NN calculation unit 109, depending on the rotation speed of the internal combustion engine 10. Since the waveform of the air flow rate in the intake passage 20 varies greatly depending on the rotation speed of the internal combustion engine 10, it is considered that the correction amount to be applied to the output value of the flow sensor 17 also varies greatly depending on the rotation speed of the internal combustion engine 10. Therefore, the NN selection unit 110 selects a neural network model depending on the rotation speed of the internal combustion engine 10 so that the NN calculation unit 109 can calculate the correction amount using a neural network model suited to the rotation speed of the internal combustion engine 10. The rotation speed of the internal combustion engine 10 is notified to the NN selection unit 110 by the rotation speed calculation unit 104.

[0047] The pulsation determination unit 111 determines whether or not pulsation is occurring in the air flowing through the intake air flow path 20 during a predetermined period. Specifically, the pulsation determination unit 111 determines that pulsation is not occurring if the difference between the maximum and minimum values ​​(hereinafter also referred to as "variation width") of the output values ​​of the flow rate sensor 17 acquired during the predetermined period is smaller than a threshold value. The pulsation determination unit 111 determines that pulsation is occurring if this variation width is equal to or greater than the threshold value.

[0048] If it is determined that pulsation is occurring, the pulsation determination unit 111 sets the switching unit 112 so that the output value of the flow sensor 17 is corrected using the correction amount calculated by the NN calculation unit 109. Specifically, the pulsation determination unit 111 sets the switching unit 112 so that the correction amount calculated by the NN calculation unit 109 is input to the flow correction unit 113.

[0049] If it is determined that no pulsation is occurring, the pulsation determination unit 111 sets the switching unit 112 so that the output value of the flow sensor 17 is not corrected using the correction amount calculated by the NN calculation unit 109. Specifically, the pulsation determination unit 111 sets the switching unit 112 so that a predetermined default value (1.0) is input to the flow correction unit 113.

[0050] The switching unit 112 switches the correction amount input to the flow rate correcting unit 113 based on the determination result of the pulsation determining unit 111 .

[0051] FIG. 5 shows a case where the variation in the output value of the flow sensor 17 acquired over a predetermined period is smaller than a threshold value. The multiple waveforms shown on the left side of FIG. 5 are all in the same operating range of the internal combustion engine 10 and have the same rotation speed. The variation in the multiple waveforms shown on the left side of FIG. 5 is at most 2.7 g / s. Converting this variation into a pulsation rate (also referred to as a "pulsation amplitude ratio") obtained by dividing the average flow rate is at most about 15%. Even if the output value of the flow sensor 17 having such a pulsation rate is normalized, a waveform with significant characteristics cannot be obtained from the output value of the flow sensor 17, as shown on the right side of FIG. 5. In other words, the inconsistent waveforms shown on the right side of FIG. 5 are not due to changes in the output value of the flow sensor 17 caused by pulsation occurring in the air flowing through the intake passage 20, but are simply due to changes caused by amplification of noise contained in the output value of the flow sensor 17. If correction is performed using a correction amount calculated from the output value of flow sensor 17 that is determined to be free of pulsation, the error between the corrected air flow rate and the actual air flow rate may actually increase.

[0052] Therefore, when pulsation occurs in the air flowing through the intake flow path 20 for a predetermined period, the pulsation determination unit 111 sets the switching unit 112 so that the output value of the flow rate sensor 17 is corrected using the correction amount calculated by the NN calculation unit 109. On the other hand, when pulsation does not occur, the pulsation determination unit 111 sets the switching unit 112 so that the output value of the flow rate sensor 17 is not corrected using the correction amount calculated by the NN calculation unit 109. Note that the threshold value that serves as the criterion for determining whether pulsation is occurring is determined appropriately in advance, and may be, for example, a change width corresponding to a pulsation rate of 50%.

[0053] The flow rate corrector 113 corrects the output value of the flow rate sensor 17 based on the average flow rate calculated by the average flow rate calculator 106 and the correction amount derived by the correction amount deriving unit 107. For example, the flow rate corrector 113 corrects the output value of the flow rate sensor 17 for a predetermined period by multiplying the average flow rate calculated by the average flow rate calculator 106 by the correction amount derived by the correction amount deriving unit 107.

[0054] Specifically, when it is determined that pulsation is occurring, the flow rate correction unit 113 corrects the output value of the flow rate sensor 17 for a predetermined period by multiplying the average flow rate by the correction amount calculated by the NN calculation unit 109. On the other hand, when it is determined that pulsation is not occurring, the flow rate correction unit 113 multiplies the average flow rate by 1.0, which is a default value, without using the correction amount calculated by the NN calculation unit 109. In this case, the flow rate correction unit 113 does not actually correct the output value of the flow rate sensor 17. In this way, the flow rate correction unit 113 can calculate the flow rate of air flowing through the intake flow path 20.

[0055] FIG. 6 is a flowchart of the process performed by the control device 100.

[0056] 6, the NN calculation unit 109 will be described as having three neural network models: a first neural network model, a second neural network model, and a third neural network model. The number of neural network models that the NN calculation unit 109 has is not limited to three. The control device 100 executes a series of processes shown in steps S1 to S13 in FIG. 6 for each constant crank angle (6 degrees).

[0057] In step S1, the control device 100 performs A / D conversion and voltage / flow rate conversion to acquire the output value of the flow rate sensor 17.

[0058] In step S2, the control device 100 adds the acquired output value of the flow rate sensor 17 to the data array.

[0059] In step S3, the control device 100 determines whether a predetermined period () has elapsed. If the internal combustion engine 10 has three cylinders, the predetermined period is, for example, 240 degrees. If the predetermined period has elapsed, the control device 100 proceeds to step S4. If the predetermined period has not elapsed, the control device 100 ends the processing shown in FIG. 6.

[0060] In step S4, the control device 100 determines whether the variation (difference between the maximum and minimum values) in the output value of the flow sensor 17 acquired over a predetermined period is smaller than a threshold value. If the variation is smaller than the threshold value, the control device 100 proceeds to step S5. If the variation is equal to or greater than the threshold value, the control device 100 proceeds to step S5.

[0061] In step S5, the control device 100 sets the correction amount for the output value of the flow rate sensor 17 to a default value (1.0). Thereafter, the control device 100 proceeds to step S12.

[0062] In step S6, the control device 100 normalizes the output value of the flow rate sensor 17 acquired during a predetermined period.

[0063] In step S7, the control device 100 determines whether the rotation speed of the internal combustion engine 10 is greater than a first reference value. If the rotation speed of the internal combustion engine 10 is greater than the first reference value, the control device 100 proceeds to step S8. If the rotation speed of the internal combustion engine 10 is equal to or less than the first reference value, the control device 100 proceeds to step S9.

[0064] In step S8, the control device 100 uses a first neural network model (also referred to as "first NNM") to calculate a correction amount for the output value of the flow rate sensor 17. Thereafter, the control device 100 proceeds to step S12.

[0065] In step S9, the control device 100 determines whether the rotation speed of the internal combustion engine 10 is greater than a second reference value (< the first reference value). If the rotation speed of the internal combustion engine 10 is greater than the second reference value, the control device 100 proceeds to step S10. If the rotation speed of the internal combustion engine 10 is equal to or less than the second reference value, the control device 100 proceeds to step S11.

[0066] In step S10, the control device 100 uses a second neural network model (also referred to as "second NNM") to calculate a correction amount for the output value of the flow rate sensor 17. Thereafter, the control device 100 proceeds to step S12.

[0067] In step S11, the control device 100 calculates a correction amount for the output value of the flow rate sensor 17 using a third neural network model (also referred to as "third NNM").

[0068] In step S12, the control device 100 calculates the average flow rate of air flowing through the intake air flow path 20 during a predetermined period.

[0069] In step S13, the control device 100 corrects the output value of the flow rate sensor 17 for a predetermined period based on the calculated average flow rate and the derived correction amount.

[0070] FIG. 7 is a diagram showing the results of verifying the error between the correction amount derived by the correction amount derivation unit 107 and the correction amount actually required.

[0071] The vertical axis of FIG. 7 represents the correction amount derived by the correction amount derivation unit 107. The horizontal axis of FIG. 7 represents the correction amount actually required (hereinafter also referred to as "required correction amount"). The straight line in FIG. 7 indicates a case where the derived correction amount is equal to the required correction amount and the error is zero. In FIG. 7, the correction amount was derived under the following learning and inference conditions and compared with the required correction amount. That is, the range of the rotation speed of the internal combustion engine 10 from 800 rpm to 6400 rpm was divided into 20 regions, and a neural network model was prepared for each region. Then, using each neural network model, the correction amount was calculated by setting the VTC (valve timing control) to 20 degrees, 40 degrees, and 55 degrees and changing the TVO (throttle valve opening) from 10 degrees to WOF (fully open). Then, the correction amount calculated by each neural network model, i.e., the correction amount derived by the correction amount derivation unit 107, was compared with the required correction amount, and the error between them was calculated.

[0072] 7, the error between the correction amount derived by the correction amount derivation unit 107 and the required correction amount is generally within a range of ±5%, which is at a level that does not pose a problem in practical use. Therefore, the control device 100 can appropriately correct the output value of the flow sensor 17 by correcting the output value of the flow sensor 17 based on the correction amount derived by the correction amount derivation unit 107.

[0073] FIG. 8 is a diagram illustrating the output value of the flow rate sensor 17 acquired by a method different from that of FIG.

[0074] In the above description, the signal output from the flow rate sensor 17 is a voltage signal, and the control device 100 samples the output value of the flow rate sensor 17 at every fixed crank angle. However, the signal output from the flow rate sensor 17 may be a frequency signal.

[0075] Furthermore, the signal output from flow sensor 17 may be transmitted to control device 100 via an external device other than flow sensor 17. In this case, the signal output from flow sensor 17 may be transmitted to control device 100 at regular time intervals according to the communication rate. That is, as indicated by the * marks in FIG. 8, control device 100 essentially samples the output value of flow sensor 17 at regular time intervals (for example, 1 ms). In this case, control device 100 may interpolate between the output values ​​of flow sensor 17 sampled at regular time intervals, as indicated by the circles in FIG. 8. This interpolated value may be the output value of flow sensor 17 sampled at regular crank angles. In this way, control device 100 can flexibly adapt to various types of flow sensor 17 and can appropriately correct the output value of flow sensor 17.

[0076] As described above, the control device 100 of this embodiment is a control device for the internal combustion engine 10 that calculates the flow rate of air flowing through the intake passage 20 based on the output value of the flow sensor 17 provided in the intake passage 20 of the internal combustion engine 10. The control device 100 includes a correction amount derivation unit 107 that derives a correction amount for the output value of the flow sensor 17, and a flow rate correction unit 113 that calculates the flow rate of air flowing through the intake passage 20 by correcting the output value of the flow sensor 17 based on the correction amount. The correction amount derivation unit 107 includes a normalization unit 108 that normalizes the output value of the flow sensor 17 for each waveform of one wavelength obtained from the output values ​​of the flow sensor 17 acquired over a predetermined period, and an NN calculation unit 109 that calculates a correction amount for the output values ​​of the flow sensor 17 over the predetermined period from the normalized output values ​​of the flow sensor 17.

[0077] As a result, even if the waveform obtained from the output value of the flow sensor 17 includes fluctuations related to offset or gain, the control device 100 of this embodiment can eliminate these fluctuations and clearly identify waveforms with different characteristics to derive a correction amount. Furthermore, the control device 100 of this embodiment can clearly identify waveforms with different characteristics even if the pulsation rate is the same due to a phase difference of the fundamental wave or high frequency, and derive a correction amount. Therefore, even if pulsation whose waveform changes in a complex manner occurs due to various factors such as VTC, TVO, EGR rate, or intake temperature, the control device 100 of this embodiment can clearly identify the characteristics of the pulsation waveform from the output value of the flow sensor 17 and derive an appropriate correction amount each time. Therefore, the control device 100 of this embodiment can appropriately correct the output value of the flow sensor 17 provided in the intake passage 20 of the internal combustion engine 10 and reduce the error between the corrected air flow rate and the actual air flow rate.

[0078] In the control device 100 of this embodiment, the predetermined period is the interval between intake strokes of the internal combustion engine 10.

[0079] As a result, the control device 100 of this embodiment can reliably obtain one wavelength of the pulsation waveform from the output value of the flow sensor 17. Therefore, the control device 100 of this embodiment can clearly and reliably identify the characteristics of the pulsation waveform from the output value of the flow sensor 17 and derive an appropriate correction amount. Therefore, the control device 100 of this embodiment can reliably reduce the error between the corrected air flow rate and the actual air flow rate.

[0080] In the control device 100 of this embodiment, the output value of the flow rate sensor 17 is sampled at regular crank angle intervals.

[0081] When the output value of the flow sensor 17 is sampled at regular time intervals, the number of pieces of data input to the NN calculation unit 109 may vary for each calculation depending on the rotation speed of the internal combustion engine 10, making it difficult to stably perform the calculation by the NN calculation unit 109. In contrast, the control device 100 of this embodiment samples the output value of the flow sensor 17 at regular crank angles, so the number of pieces of data input to the NN calculation unit 109 can be kept constant for each calculation. Therefore, the control device 100 of this embodiment can stably perform the calculation by the NN calculation unit 109, and can stably derive an appropriate correction amount for the output value of the flow sensor 17. Therefore, the control device 100 of this embodiment can further reduce the error between the corrected air flow rate and the actual air flow rate.

[0082] Furthermore, in the control device 100 of this embodiment, the NN calculation unit 109 has a neural network model, and the normalized output value of the flow sensor 17 is set in the input layer of the neural network model, and a correction amount for the output value of the flow sensor 17 for a predetermined period is set in the output layer of the neural network model.

[0083] This allows the control device 100 of this embodiment to easily derive the amount of correction for the output value of the flow rate sensor 17. Therefore, the control device 100 of this embodiment can easily reduce the error between the corrected air flow rate and the actual air flow rate.

[0084] Moreover, in the control device 100 of this embodiment, the NN calculation unit 109 has a plurality of neural network models that differ depending on the rotation speed of the internal combustion engine 10. The correction amount derivation unit 107 further includes an NN selection unit 110 that selects a neural network model to be used for calculating the correction amount from the plurality of neural network models that the NN calculation unit 109 has, depending on the rotation speed of the internal combustion engine 10. The NN calculation unit 109 uses the selected neural network model to calculate a correction amount for the output value of the flow sensor 17 for a predetermined period.

[0085] As a result, the control device 100 of this embodiment can accurately identify the characteristics of the pulsation waveform that changes depending on the rotation speed of the internal combustion engine 10 from the output value of the flow sensor 17, and can derive a correction amount using a neural network model that is suited to the pulsation waveform. Therefore, the control device 100 of this embodiment can more appropriately derive a correction amount for the output value of the flow sensor 17. Therefore, the control device 100 of this embodiment can further reduce the error between the corrected air flow rate and the actual air flow rate.

[0086] Furthermore, in the control device 100 of this embodiment, the correction amount derivation unit 107 further includes a pulsation determination unit 111 that determines that no pulsation is occurring in the air flowing through the intake flow path 20 during a predetermined period when the difference between the maximum and minimum values ​​in the output value of the flow rate sensor 17 during the predetermined period is smaller than a threshold value. When it is determined that no pulsation is occurring, the flow rate correction unit 113 calculates the air flow rate without correcting the output value of the flow rate sensor 17 using the correction amount calculated by the NN calculation unit 109.

[0087] As a result, the control device 100 of this embodiment can derive the correction amount while eliminating cases where the output value of the flow rate sensor 17 is changed not by pulsation but by noise, and therefore the control device 100 of this embodiment can further reduce the error between the corrected air flow rate and the actual air flow rate.

[0088] The present invention is not limited to the above-described embodiments and includes various modifications. For example, the above-described embodiments have been described in detail to clearly explain the present invention, and the present invention is not necessarily limited to those including all of the described configurations. Furthermore, it is possible to replace part of the configuration of one embodiment with the configuration of another embodiment, or to add the configuration of another embodiment to the configuration of one embodiment. Furthermore, it is possible to add, delete, or replace part of the configuration of each embodiment with other configurations.

[0089] Furthermore, the above-described configurations, functions, processing units, processing means, etc. may be partially or entirely realized by hardware, for example, by designing them as integrated circuits. The above-described configurations, functions, etc. may also be realized by software, in which a processor interprets and executes a program that realizes each function. Information such as the program, tape, and file that realizes each function can be stored in a memory, a recording device such as a hard disk or solid state drive (SSD), or a recording medium such as an IC card, SD card, or DVD.

[0090] In addition, the control lines and information lines shown are those that are considered necessary for the explanation, and do not necessarily show all the control lines and information lines in the product. In reality, it can be assumed that almost all components are interconnected. [Explanation of symbols]

[0091] 10... internal combustion engine, 17... flow rate sensor, 20... intake passage, 100... control device, 107... correction amount derivation unit, 108... normalization unit, 109... NN calculation unit (calculation unit), 110... NN selection unit (selection unit), 111... pulsation determination unit, 113... flow rate correction unit

Claims

1. A control device for an internal combustion engine that calculates a flow rate of air flowing through an intake passage of the internal combustion engine based on an output value of a flow rate sensor provided in the intake passage, a correction amount deriving unit that derives a correction amount for the output value of the flow rate sensor; a flow rate correction unit that calculates the flow rate of the air by correcting the output value of the flow rate sensor based on the correction amount, The correction amount derivation unit a normalization unit that normalizes the output value of the flow sensor for each waveform of one wavelength obtained from the output value of the flow sensor acquired over a predetermined period; a calculation unit that calculates the correction amount for the output value of the flow sensor for the predetermined period from the normalized output value of the flow sensor. A control device for an internal combustion engine.

2. The predetermined period is the interval between intake strokes of the internal combustion engine.

2. The control device for an internal combustion engine according to claim 1.

3. The output value of the flow rate sensor is sampled at regular crank angle intervals.

3. The control device for an internal combustion engine according to claim 2.

4. the calculation unit has a neural network model, a normalized output value of the flow rate sensor is set in an input layer of the neural network model; The output layer of the neural network model is set with the correction amount for the output value of the flow sensor for the predetermined period.

2. The control device for an internal combustion engine according to claim 1.

5. the calculation unit has a plurality of the neural network models that differ depending on the rotational speed of the internal combustion engine, the correction amount derivation unit further includes a selection unit that selects the neural network model used for calculating the correction amount from among the plurality of neural network models included in the calculation unit in accordance with a rotation speed of the internal combustion engine, The calculation unit calculates the correction amount for the output value of the flow sensor for the predetermined period using the selected neural network model.

5. The control device for an internal combustion engine according to claim 4.

6. the correction amount derivation unit further includes a pulsation determination unit that determines that no pulsation is occurring in the air during the predetermined period when a difference between a maximum value and a minimum value of the output value of the flow rate sensor during the predetermined period is smaller than a threshold value, When it is determined that the pulsation is not occurring, the flow rate correction unit calculates the flow rate of the air without correcting the output value of the flow rate sensor using the correction amount calculated by the calculation unit.

2. The control device for an internal combustion engine according to claim 1.

Images