JP2020133622A - Imbalance detection device, imbalance detection system, data analysis device, and controller for internal combustion engine - Google Patents

Abstract

To provide an imbalance detection device which, even if it is difficult to identify from only a detection value of an air-fuel ratio sensor how an air-fuel ratio of any cylinder is shifted, can identify it.SOLUTION: A CPU 72 calculates an imbalance ratio, which is the deviation of an actual injection amount from an intended injection amount in each of cylinders #1 to #4, using, as inputs, minute rotation duration T30(1) to T30(24), which is time required for each 30° CA rotation in one combustion cycle, and time series data of an average value of detection values of an air-fuel ratio sensor 82 at an interval of 30° CA.SELECTED DRAWING: Figure 1

Description

The present invention relates to an imbalance detection device, an imbalance detection system, a data analysis device, and a control device for an internal combustion engine.

For example, in Patent Document 1 below, imbalance, which is a variation between actual air-fuel ratios when a fuel injection valve is operated in order to control the air-fuel ratios of air-fuel ratios in each of a plurality of cylinders to equal air-fuel ratios, is detected. The device is described. This device identifies a cylinder whose air-fuel ratio deviates from the intended one based on an air-fuel ratio estimation model in which the detected value is input in addition to the difference value of the detected value of the air-fuel ratio sensor.

Japanese Unexamined Patent Publication No. 2013-194685

By the way, as imbalance, there are a phenomenon of shifting to the rich side from the intended air-fuel ratio and a phenomenon of shifting to the lean side. Then, the inventor may approximate the behavior of the detected value of the air-fuel ratio sensor between the imbalance shifted to the rich side and the imbalance shifted to the lean side, and in that case, how the air-fuel ratio of any cylinder shifts. It was found that it may be difficult to identify whether or not it is.

Hereinafter, means for solving the above problems and their actions and effects will be described.
1. 1. An imbalance detection device applied to a multi-cylinder internal combustion engine, comprising a storage device and an execution device, wherein the storage device is a rotation waveform variable and an air-fuel ratio sensor at each of a plurality of first intervals. The air-fuel ratio detection variable, which is a variable according to the output, is input, and the mapping data, which is the data that defines the mapping to output the imbalance variable, which is a variable indicating the degree of variation in the air-fuel ratio of the internal combustion engine, is stored. The execution device is acquired by the acquisition process of acquiring the rotation waveform variable based on the detection value of the sensor that detects the rotation behavior of the crank shaft and the air-fuel ratio detection variable at each of the plurality of first intervals, and the acquisition process. A calculation process for calculating the imbalance variable based on the output of the mapping with a value as an input, and a large variation in the air-fuel ratio by operating a predetermined hardware based on the calculation result of the calculation process. The coping process for coping is executed, and the rotation waveform variable is a variable indicating the difference between the instantaneous speed variables, which are variables corresponding to the rotation speed of the crank shaft at each of the plurality of second intervals, and is the first. The interval and the second interval are both angular intervals of the crank shaft that are smaller than the appearance interval of the compression top dead point, and the rotation waveform variable and the plurality of air fuel ratio detection variables used as input of the mapping are each. , An imbalance detection device that is time-series data within a predetermined angular interval larger than the appearance interval.

In the above configuration, the rotation waveform variable is used in addition to the air-fuel ratio detection variable at each of the plurality of first intervals. Since the rotational waveform variable quantifies the rotational behavior that occurs in response to the torque generated in the combustion chamber of each cylinder, it has sensitivity to differences in the air-fuel ratio of the air-fuel mixture. Therefore, it may be difficult to identify how the air-fuel ratio of which cylinder is deviated depending on the air-fuel ratio detection variable at each of the plurality of first intervals by adding the rotation waveform variable. However, this can be specified.

2. 2. The input of the map includes a 0.5th-order component of the rotation frequency of the crankshaft based on the detection value of the sensor that detects the rotation behavior of the crankshaft, and the acquisition process uses the 0.5th-order component. The calculation process includes the process of acquiring the 0.5th-order component variable, which is a specified variable, and the calculation process further includes the 0.5th-order component variable acquired by the acquisition process in the input to the map. The imbalance detection device according to 1 above, which is a process of calculating the imbalance variable based on the output.

It can be considered that a linear relationship is established between the imbalance rate and the 0.5th-order amplitude, which is the magnitude of the 0.5th-order component of the rotation frequency, and the amplitude of the rotation frequency in the presence of imbalance is It has been found by the inventor that the 0.5th order component is particularly large. It is considered that this is because when imbalance occurs in any one of the plurality of cylinders, the generated torque deviates once in one combustion cycle. In the above configuration, paying attention to this point, the imbalance variable can be calculated with higher accuracy by including the 0.5th-order component variable in the input to the map.

3. 3. The input of the map includes an operating point variable which is a variable defining the operating point of the internal combustion engine, the acquisition process includes a process of acquiring the operating point variable, and the calculation process is the acquisition process. The imbalance detection device according to 1 or 2 above, which is a process of calculating the imbalance variable based on the output of the map, which further includes the operation point variable acquired by the above in the input to the map.

Since the control of the internal combustion engine tends to be performed according to the operating point of the internal combustion engine, the rotational behavior of the crankshaft may differ depending on the operating point of the internal combustion engine. Therefore, in the above configuration, by including the operating point variable of the internal combustion engine in the input of the map, the imbalance variable can be calculated while reflecting that the rotational behavior of the crankshaft differs depending on the operating point.

4. The input of the map includes an adjustment variable which is a variable for adjusting the combustion speed of the air-fuel mixture in the combustion chamber of the internal combustion engine by operating the operation unit of the internal combustion engine, and the acquisition process is the adjustment variable. 1. The calculation process is a process of calculating the imbalance variable based on the output of the map including the adjustment variable acquired by the acquisition process in the input to the map. The imbalance detection device according to any one of 3 to 3.

The rotational behavior of the crankshaft changes depending on the combustion speed of the air-fuel mixture. Therefore, in the above configuration, by including the adjustment variable for adjusting the combustion speed in the input of the map, the imbalance variable can be calculated while reflecting that the rotational behavior of the crankshaft changes according to the combustion speed.

5. The input of the map includes a drive system state variable which is a variable indicating the state of the drive system device connected to the crankshaft, and the acquisition process includes a process of acquiring the drive system state variable. The calculation process is any one of 1 to 4 above, which is a process of calculating the imbalance variable based on the output of the map including the drive system state variable acquired by the acquisition process in the input to the map. The imbalance detection device according to the above.

When the state of the drive system device connected to the crankshaft is different, the rotational behavior of the crankshaft tends to be different. Therefore, in the above configuration, by including the drive system state variable in the mapping input, the imbalance variable can be calculated while reflecting that the rotational behavior of the crankshaft differs depending on the state of the drive system device.

6. The input of the map includes a road surface state variable which is a variable indicating the state of the road surface on which the vehicle on which the internal combustion engine is mounted is traveling, and the acquisition process includes a process of acquiring the road surface state variable. The calculation process is any one of 1 to 5 above, which is a process of calculating the imbalance variable based on the output of the map in which the road surface state variable acquired by the acquisition process is further included in the input to the map. The imbalance detection device according to one.

For example, if the road surface is uneven, vibration will occur in the vehicle, which will be transmitted to the crankshaft. In this way, the condition of the road surface affects the rotational behavior of the crankshaft. Therefore, in the above configuration, by including the road surface state variable in the mapping input, the imbalance variable can be calculated while reflecting the change in the rotational behavior of the crankshaft according to the road surface state.

7. The rotation waveform variable is configured as a variable indicating a difference between the instantaneous velocity variables by the instantaneous velocity variable itself at each of the plurality of second intervals, and the acquisition process is performed as the plurality of rotation waveform variables. The calculation process includes the process of acquiring the instantaneous velocity variable in each of the second intervals, and the calculation process uses the instantaneous velocity variable in each of the plurality of second intervals as the rotation waveform variable as an input to the mapping. The imbalance detection device according to any one of 1 to 6 above, which is a process of calculating a variable.

When calculating the difference or ratio between the instantaneous velocity variables, it is necessary to separately match which two second intervals the difference or ratio between the instantaneous velocity variables should be calculated. On the other hand, in the above configuration, by using a plurality of instantaneous velocity variables as the rotational waveform variables, the rotational waveform variables are more suitable than the case where the difference or ratio between the instantaneous velocity variables is calculated in advance and input to the map. Man-hours can be reduced.

8. The storage device includes a plurality of types of the mapping data, and the calculation processing includes selection processing for selecting the mapping data to be used for calculating the imbalance variable from the plurality of types of mapping data. The imbalance detection device according to any one of the above.

In all situations, when constructing a map capable of outputting imbalance variables with high accuracy, the structure of the map tends to be complicated. Therefore, in the above configuration, a plurality of types of mapping data are provided. This makes it possible to select an appropriate map according to the situation, and in that case, it is easier to simplify the structure of each of the multiple types of maps as compared with the case where a single map deals with all situations. ..

9. The internal combustion engine has a canister that collects fuel vapor in a fuel tank that stores fuel injected from a fuel injection valve, a purge passage that connects the canister and an intake passage of the internal combustion engine, and the purge passage. The calculation process includes the adjusting device for adjusting the flow rate of the fuel vapor flowing from the canister into the intake passage via the canister, and the calculation process obtains the data acquired by the acquisition process when the flow rate of the fuel vapor is zero. A process of calculating the imbalance variable as an input to the mapping, and a process of calculating the imbalance variable by using the data acquired by the acquisition process as an input to the mapping when the flow rate of the fuel steam is larger than zero. In the coping process, when the flow rate of the fuel steam is larger than zero, the data acquired by the acquisition process is transferred to the mapping when the flow rate of the fuel steam is larger than zero. The imbalance detection device according to any one of 1 to 8 above, which is a process of operating the predetermined hardware based on the imbalance variable calculated as an input.

When the flow rate of the fuel vapor is larger than zero, the fuel vapor flows from the canister into the intake passage, and this flows into each cylinder, but the amount of the fuel vapor flowing into each cylinder varies. This variation cannot be expressed by the imbalance variable that quantifies the variation caused by individual differences in fuel injection valves, aging, and the like. Therefore, in the above configuration, the imbalance caused by individual differences and aging of the fuel injection valve and the imbalance caused by the fuel vapor are calculated separately by calculating the imbalance variable according to the presence or absence of the influence of the fuel vapor. It can be grasped, and by extension, it can be dealt with according to the situation.

10. The internal combustion engine is applied to an internal combustion engine including an EGR passage connecting an exhaust passage and an intake passage, and an EGR valve for adjusting the flow rate of exhaust gas flowing from the exhaust passage to the intake passage through the EGR passage. The calculation process includes a process of calculating the imbalance variable by using the data acquired by the acquisition process as an input to the mapping when the flow rate of the exhaust gas flowing into the intake passage is zero, and the intake passage. The coping process includes both a process of calculating an imbalance variable by using the data acquired by the acquisition process as an input to the mapping when the flow rate of the exhaust gas flowing into the exhaust gas is larger than zero. When the flow rate of the exhaust gas flowing into the intake passage is larger than zero, the data acquired by the acquisition process when the flow rate of the exhaust gas flowing into the intake passage is larger than zero is calculated as an input to the mapping. The imbalance detection device according to any one of 1 to 8 above, which is a process of operating the predetermined hardware based on the imbalance variable.

When the flow rate of the exhaust gas flowing into the intake passage is larger than zero, the exhaust gas flows from the exhaust passage to the intake passage through the EGR passage, and although this flows into each of the cylinders, the inflow of the exhaust gas into each cylinder. There are variations in the amount. The variation in the combustion state of the air-fuel mixture in each cylinder due to this variation cannot be expressed by the imbalance variable that quantifies the variation due to individual differences in the fuel injection valve, aging, and the like. Therefore, in the above configuration, by calculating the imbalance variable according to the presence or absence of the influence of the exhaust gas flowing into the intake passage, the imbalance caused by the individual difference of the fuel injection valve, the secular change, etc., and the inflow into the intake passage It is possible to grasp the imbalance caused by the exhaust gas, and to take measures according to the situation.

11. In the coping process, when the degree of variation in the air-fuel ratio is large, the operation unit for controlling the combustion of the air-fuel mixture in the combustion chamber of the internal combustion engine as the predetermined hardware is operated according to the imbalance variable. The imbalance detection device according to any one of 1 to 10 above, which includes an operation process to be performed.

In the above configuration, by operating the operation unit for controlling the combustion of the air-fuel mixture according to the imbalance variable, it is possible to improve the deterioration of the combustion state due to the large variation in the air-fuel ratio.

12. The operation process is the imbalance detection device according to 11 above, which includes a process of operating a fuel injection valve as the operation unit for supplying fuel to each of a plurality of cylinders.
In the above configuration, the variation in the air-fuel ratio of the air-fuel ratio of each cylinder can be reduced by correcting the injection amount according to the imbalance variable.

13. The execution device and the storage device according to any one of the above 1 to 12 are provided, the execution device includes a first execution device and a second execution device, and the first execution device is mounted on a vehicle. The acquisition process, the vehicle-side transmission process for transmitting the data acquired by the acquisition process to the outside of the vehicle, the vehicle-side reception process for receiving a signal based on the calculation result of the calculation process, and the countermeasure process. The second execution device is arranged outside the vehicle and receives the data transmitted by the vehicle-side transmission process, the external reception process, the calculation process, and the calculation process. This is an imbalance detection system that executes an external transmission process for transmitting a signal based on the calculation result of the above to the vehicle.

In the above configuration, the calculation load of the in-vehicle device can be reduced by executing the calculation process outside the vehicle.
14. A data analysis device including the second execution device and the storage device according to the above 13.

15. It is a control device of an internal combustion engine including the first execution device according to 13.

The figure which shows the structure of the control device and the drive system of a vehicle which concerns on 1st Embodiment. The block diagram which shows a part of the processing executed by the control device which concerns on this embodiment. The flow chart which shows the procedure of the imbalance detection processing concerning the same embodiment. The flow chart which shows the procedure of the coping process for imbalance concerning the same embodiment. The figure which shows the system which generates the mapping data concerning this embodiment. The flow chart which shows the procedure of the learning process of the mapping data which concerns on the same embodiment. A time chart showing the effect of imbalance on instantaneous velocity variables and air-fuel ratio. (A) and (b) are diagrams showing the relationship between imbalance and 0.5th-order amplitude. The flow chart which shows the procedure of the calculation process of the imbalance learning value which concerns on 2nd Embodiment. The flow chart which shows the procedure of the coping process for imbalance concerning the same embodiment. The flow chart which shows the procedure of the coping process for imbalance concerning the 3rd Embodiment. The flow chart which shows the procedure of the selection process of the mapping data which concerns on 4th Embodiment. The flow chart which shows the procedure of the selection process of the mapping data which concerns on 5th Embodiment. The flow chart which shows the procedure of the selection process of the mapping data which concerns on 6th Embodiment. The flow chart which shows the procedure of the selection process of the mapping data which concerns on 7th Embodiment. The flow chart which shows the procedure of the imbalance detection processing which concerns on 8th Embodiment. The figure which shows the structure of the imbalance detection system which concerns on 9th Embodiment. (A) and (b) are flow charts showing the procedure of processing executed by the imbalance detection system.

<First Embodiment>
Hereinafter, the first embodiment of the imbalance detection device will be described with reference to the drawings.

In the internal combustion engine 10 mounted on the vehicle VC shown in FIG. 1, a throttle valve 14 is provided in the intake passage 12. The air sucked from the intake passage 12 flows into the combustion chambers 18 of the cylinders # 1 to # 4 by opening the intake valve 16. The internal combustion engine 10 is provided with a fuel injection valve 20 for injecting fuel so as to be exposed to the combustion chamber 18 and an ignition device 22 for generating spark discharge. In the combustion chamber 18, the air-fuel mixture is subjected to combustion, and the energy generated by the combustion is taken out as the rotational energy of the crankshaft 24. The air-fuel mixture used for combustion is discharged to the exhaust passage 28 as exhaust gas when the exhaust valve 26 is opened. The exhaust passage 28 is provided with a catalyst 30 having an oxygen storage capacity. The exhaust passage 28 is communicated with the intake passage 12 via the EGR passage 32. The EGR passage 32 is provided with an EGR valve 34 for adjusting the cross-sectional area of the flow path.

The fuel stored in the fuel tank 38 is supplied to the fuel injection valve 20 via the pump 36. The fuel vapor generated in the fuel tank 38 is collected in the canister 40. The canister 40 is connected to the intake passage 12 via the purge passage 42, and the flow path cross section of the purge passage 42 is adjusted by the purge valve 44.

The rotational power of the crankshaft 24 is transmitted to the intake side camshaft 48 via the intake side valve timing variable device 46. The intake-side valve timing variable device 46 changes the relative rotational phase difference between the intake-side camshaft 48 and the crankshaft 24.

The input shaft 66 of the transmission 64 can be connected to the crankshaft 24 of the internal combustion engine 10 via a torque converter 60. The torque converter 60 includes a lockup clutch 62, and when the lockup clutch 62 is engaged, the crankshaft 24 and the input shaft 66 are connected to each other. Drive wheels 69 are mechanically connected to the output shaft 68 of the transmission 64. In the present embodiment, the transmission 64 is a stepped transmission that can change the gear ratio from the 1st speed to the 5th speed.

A crank rotor 50 provided with tooth portions 52 indicating each of a plurality of rotation angles of the crank shaft 24 is connected to the crank shaft 24. In this embodiment, 34 tooth portions 52 are illustrated. The crank rotor 50 is basically provided with tooth portions 52 at intervals of 10 ° CA, but there is one missing tooth portion 54 at a location where the interval between adjacent tooth portions 52 is 30 ° CA. It is provided. This is for indicating the reference rotation angle of the crankshaft 24.

The control device 70 targets the internal combustion engine 10 and controls the throttle valve 14, the fuel injection valve 20, the ignition device 22, the EGR valve 34, and the purge valve 44 in order to control the torque and the exhaust component ratio, which are the controlled amounts thereof. , Operates the operation unit of the internal combustion engine 10 such as the intake side valve timing variable device 46. Note that FIG. 1 shows operation signals MS1 to MS6 of the throttle valve 14, the fuel injection valve 20, the ignition device 22, the EGR valve 34, the purge valve 44, and the intake side valve timing variable device 46, respectively.

When controlling the control amount, the control device 70 has an intake air amount Ga detected by the air flow meter 80, an upstream side detection value Afu detected by an air-fuel ratio sensor 82 provided on the upstream side of the catalyst 30, and a missing tooth portion. Refer to the output signal Scr of the crank angle sensor 86 that outputs a pulse for each angle interval between the tooth portions 52 provided at every 10 ° CA except 54. The control device 70 is detected by the water temperature THW, which is the temperature of the cooling water of the internal combustion engine 10 detected by the water temperature sensor 88, the shift position Sft of the transmission 64 detected by the shift position sensor 90, and the acceleration sensor 92. Refer to the vertical acceleration Dacc of the vehicle VC.

The control device 70 includes a CPU 72, a ROM 74, a storage device 76 which is an electrically rewritable non-volatile memory, and a peripheral circuit 77, which are made communicable by the local network 78. The peripheral circuit 77 includes a circuit that generates a clock signal that defines the internal operation, a power supply circuit, a reset circuit, and the like.

The control device 70 executes the control of the control amount by the CPU 72 executing the program stored in the ROM 74.
FIG. 2 shows a part of the processing realized by the CPU 72 executing the program stored in the ROM 74.

The ignition timing operation process M10 sets the base value of the ignition timing based on the rotation speed NE and the filling efficiency η that define the operating point of the internal combustion engine 10, and the ignition timing aig is set according to the base value of the ignition timing. This is a process of outputting the operation signal MS3 to operate the ignition device 22. Here, the filling efficiency η is a parameter indicating the amount of air filled in the combustion chamber 18, and is calculated by the CPU 72 based on the intake air amount Ga and the rotation speed NE. Further, the rotation speed NE is calculated by the CPU 72 based on the output signal Scr of the crank angle sensor 86. The rotation speed NE is an average value of the rotation speeds when the crankshaft 24 rotates at an angle interval larger than the appearance interval of the compression top dead center (180 ° CA in this embodiment). The rotation speed NE is preferably an average value of the rotation speeds when the crankshaft 24 rotates by a rotation angle of one rotation or more of the crankshaft 24. The average value here is not limited to the simple average, and may be, for example, exponential movement averaging processing. When the crankshaft 24 rotates by a rotation angle of one rotation or more, for example, a plurality of rotation speeds at minute rotation angle intervals. It may be calculated by the sampling value of. However, the present invention is not limited to this, and it may be calculated based on a single measurement value of the time required to rotate by a rotation angle of one rotation or more.

The EGR control process M12 is based on the operating point of the internal combustion engine 10, and is based on the sum of the flow rate of the air sucked into the intake passage 12 and the flow rate of the exhaust gas flowing into the intake passage 12 through the EGR passage 32. This is a process of outputting an operation signal MS4 to the EGR valve 34 to operate the opening degree of the EGR valve 34 in order to control the EGR rate Regr rate, which is a ratio of the flow rate.

The target purge rate calculation process M14 is a process of calculating the target purge rate Rp * based on the filling efficiency η. Here, the purge rate is a value obtained by dividing the flow rate of the fluid flowing from the canister 40 into the intake passage 12 by the intake air amount Ga, and the target purge rate Rp * is a control target value of the purge rate.

The purge valve operation process M16 is a process of outputting an operation signal MS5 to the purge valve 44 in order to operate the purge valve 44 so that the purge rate becomes the target purge rate Rp * based on the intake air amount Ga. Here, in the purge valve operation process M16, when the target purge rate Rp * is the same, the smaller the intake air amount Ga, the smaller the opening degree of the purge valve 44. This is because even if the pressure in the canister 40 is the same, the pressure in the intake passage 12 is lower as the intake air amount Ga is smaller, so that the pressure in the canister 40 is higher than the pressure in the intake passage 12. This is because the fluid easily flows from the canister 40 to the intake passage 12.

The base injection amount calculation process M18 is a process of calculating the base injection amount Qb, which is the base value of the fuel amount for setting the air-fuel ratio of the air-fuel mixture in the combustion chamber 18 as the target air-fuel ratio, based on the filling efficiency η. Specifically, in the base injection amount calculation process M18, for example, when the filling efficiency η is expressed as a percentage, the filling efficiency η is set to the fuel amount QTH per 1% of the filling efficiency η for setting the air-fuel ratio as the target air-fuel ratio. The process may be such that the base injection amount Qb is calculated by multiplying. The base injection amount Qb is a fuel amount calculated to control the air-fuel ratio to the target air-fuel ratio based on the amount of air filled in the combustion chamber 18. By the way, the target air-fuel ratio may be, for example, the theoretical air-fuel ratio.

The feedback processing M20 is a process of calculating the feedback correction coefficient KAF by adding "1" to the correction ratio δ, which is the operation amount for feedback-controlling the upstream detection value Afu, which is the feedback control amount, to the target value Af *. .. The feedback correction coefficient KAF is a correction coefficient for the base injection amount Qb. Here, when the correction ratio δ is “0”, the correction ratio of the base injection amount Qb is zero. When the correction ratio δ is larger than “0”, the base injection amount Qb is increased and corrected, and when the correction ratio δ is smaller than “0”, the base injection amount Qb is reduced. In the present embodiment, the output of the integrating element that outputs the sum of the output values of the proportional element and the differential element that input the difference between the target value Af * and the upstream detection value Afu and the integrated value of the values corresponding to the difference. Let the sum of the values be the correction ratio δ.

The air-fuel ratio learning process M22 is a process of sequentially updating the air-fuel ratio learning value LAF so that the deviation between the correction ratio δ and “0” becomes small during the air-fuel ratio learning period. The air-fuel ratio learning process M22 includes a process of determining that the air-fuel ratio learning value LAF has converged when the deviation amount of the correction ratio δ from “0” is equal to or less than a predetermined value.

The coefficient addition process M24 adds the air-fuel ratio learning value LAF to the feedback correction coefficient KAF.
The purge concentration learning process M26 is a process for calculating the purge concentration learning value Lp based on the correction ratio δ. The purge concentration learning value Lp is a correction ratio that corrects the deviation of the base injection amount Qb with respect to the injection amount required to control the target air-fuel ratio due to the inflow of fuel vapor from the canister 40 to the combustion chamber 18. It is a value converted per 1% of the rate. Here, in the present embodiment, all the factors that cause the feedback correction coefficient KAF to deviate from "1" when the target purge rate Rp * is controlled to a value larger than "0" flow into the combustion chamber 18 from the canister 40. It is considered to be due to the fuel vapor. That is, the correction ratio δ is regarded as a correction ratio for correcting the deviation of the base injection amount Qb with respect to the injection amount required for controlling the target air-fuel ratio due to the inflow of fuel vapor from the canister 40 to the intake passage 12. However, since the correction ratio δ depends on the purge rate, in the present embodiment, the purge concentration learning value Lp is set to an amount corresponding to the value “δ / Rp” per 1% of the purge rate. Specifically, the purge concentration learning value Lp is used as the exponential moving average processing value of the value “δ / Rp” per 1% of the purge rate. It is desirable that the target purge rate Rp * is set to a value larger than zero and the purge concentration learning process M26 is executed on condition that the air-fuel ratio learning value LAF is determined to have converged.

The purge correction ratio calculation process M28 is a process of calculating the purge correction ratio Dp by multiplying the target purge rate Rp * by the purge concentration learning value Lp. The purge correction ratio Dp is a value of zero or less.

The correction coefficient calculation process M30 is a process of adding the purge correction ratio Dp to the output value of the coefficient addition process M24.
The required injection amount calculation process M32 is a process of correcting the base injection amount Qb by multiplying the base injection amount Qb by the output value of the correction coefficient calculation process M30, and calculating the required injection amount Qd.

The injection valve operation process M34 is a process of outputting an operation signal MS2 to the fuel injection valve 20 in order to operate the fuel injection valve 20 based on the required injection amount Qd.
Next, when the fuel injection valves 20 in each of the cylinders # 1 to # 4 are operated based on the process of FIG. 2, a process of detecting an imbalance in which the actual air-fuel ratio shifts between the cylinders will be described.

FIG. 3 shows a procedure for processing related to imbalance detection. The process shown in FIG. 3 is realized by the CPU 72 repeatedly executing the imbalance detection program 74a stored in the ROM 74 shown in FIG. 1, for example, at a predetermined cycle. In the following, the step number of each process is represented by a number prefixed with "S".

In the series of processes shown in FIG. 3, the CPU 72 first determines whether or not the execution condition of the imbalance detection process is satisfied (S10). In the present embodiment, the execution condition includes a condition that the target purge rate Rp * is zero and the EGR rate Regr is zero.

Next, the CPU 72 has a minute rotation time T30 (1), T30 (2), ..., T30 (24), upstream average value Amplitude (1), Afave (2), ..., Afave (24), rotation speed NE. , Filling efficiency η, and 0.5th order amplitude Ampf / 2 are obtained (S12). The minute rotation time T30 is calculated by the CPU 72 measuring the time required for the crankshaft 24 to rotate 30 ° CA based on the output signal Scr of the crank angle sensor 86. Here, when the numbers in parentheses are different, such as the minute rotation times T30 (1) and T30 (2), it indicates that they have different rotation angle intervals within 720 ° CA, which is one combustion cycle. That is, the minute rotation times T30 (1) to T30 (24) indicate the rotation times at each angle interval obtained by equally dividing the rotation angle region of 720 ° CA by 30 ° CA.

Specifically, the CPU 72 clocks the time when the crankshaft 24 rotates by 30 ° CA based on the output signal Scr, and sets this as the pre-filtering time NF30. Next, the CPU 72 calculates the post-filter processing time AF30 by performing digital filter processing with the pre-filter processing time NF30 as an input. Then, the CPU 72 normalizes the post-filter processing time AF30 so that the difference between the maximum value and the minimum value of the post-filter processing time AF30 in a predetermined period (for example, 720 ° CA) is “1”, thereby causing the minute rotation time T30. Is calculated.

Further, when m = 1 to 24, the upstream side average value Afuave (m) is the average value of the upstream side detection value Afu at the same angle interval of 30 ° CA as each of the minute rotation times T30 (m).

The 0.5th-order amplitude Ampf / 2 is the intensity of the 0.5th-order component of the rotation frequency of the crankshaft 24, and is calculated by the CPU 72 by the Fourier transform of the time-series data of the minute rotation time T30.

Next, the CPU 72 substitutes the value acquired by the process of S12 into the input variables x (1) to x (51) of the map that outputs the imbalance rate Riv (S14). Specifically, the CPU 72 substitutes the minute rotation time T30 (m) for the input variable x (m) and substitutes the upstream average value Afave (m) for the input variable (24 + m) with "m = 1 to 24". , The rotation speed NE is substituted into the input variable x (49), the filling efficiency η is substituted into the input variable x (50), and the 0.5th order amplitude Ampf / 2 is substituted into the input variable x (51).

In the present embodiment, the imbalance rate Riv is set to "0" in the cylinder in which the fuel of the target injection amount is injected, and becomes a positive value when the actual injection amount is larger than the target injection amount. If it is small, it will be a negative value.

Next, the CPU 72 inputs the input variables x (1) to x (51) to the mapping defined by the mapping data 76a stored in the storage device 76 shown in FIG. 1, thereby causing the cylinder #i (i = 1 to 1). The imbalance rates Riv (1) to Riv (4) of 4) are calculated (S16).

In the present embodiment, this map is composed of a neural network having a single mesosphere. The above neural network is an input side that non-linearly transforms each of the input side coefficient wFjk (j = 0 to n, k = 0 to 51) and the output of the input side linear map which is a linear map defined by the input side coefficient wFjk. It includes an activation function h (x) as a non-linear map. In this embodiment, the hyperbolic tangent "tanh (x)" is exemplified as the activation function h (x). Further, the neural network non-linearly transforms each of the output side coefficient wSij (i = 1 to 4, j = 0 to n) and the output of the output side linear map which is a linear map defined by the output side coefficient wSij. It includes an activation function f (x) as an output-side non-linear map. In this embodiment, the hyperbolic tangent "tanh (x)" is exemplified as the activation function f (x). The value n indicates the dimension of the intermediate layer.

When the process of S16 is completed or when a negative determination is made in the process of S10, the CPU 72 temporarily ends the series of processes shown in FIG.
FIG. 4 shows a procedure of processing using the imbalance rate Rivi (i). The process shown in FIG. 4 is realized by repeatedly executing the coping program 74b stored in the ROM 74 shown in FIG. 1 each time the CPU 72 repeatedly executes, for example, the imbalance rate Riv (i).

In the series of processes shown in FIG. 4, the CPU 72 first obtains the imbalance learning value Live (i) by the exponential moving average process in which the imbalance rate Riv (i) newly calculated by the process of FIG. 3 is input. Update (S20). That is, the CPU 72 has, for example, a value obtained by multiplying the imbalance learning value Live (i) stored in the storage device 76 by a coefficient α and a value obtained by multiplying the imbalance rate Riv (i) by “1-α”. The imbalance learning value Live is updated by the sum (S20). In addition, "0 <α <1".

Next, the CPU 72 determines whether or not the imbalance learning value Live (i) is equal to or greater than the lean side allowable limit value LL and equal to or less than the rich side allowable limit value LH (S22). When the CPU 72 determines that the imbalance learning value Live (i) is less than the lean side allowable limit value LL or is larger than the rich side allowable limit value (S22: NO), the CPU 72 repairs the user. The warning light 98 is operated to execute the notification process (S24).

On the other hand, when the CPU 72 determines that the lean side allowable limit value LL or more and the rich side allowable limit value LH or less (S22: YES), or when the processing of S24 is completed, the request of each cylinder is required. The injection amount Qd (#i) is corrected (S26). That is, the CPU 72 adds the required injection amount Qd (# i) of each cylinder to the correction amount ΔQd (Live (i)) corresponding to the imbalance learning value Live (i), thereby causing the required injection amount Qd (# i). ) Is corrected. Here, the correction amount ΔQd (Liv (i)) becomes a negative value when the imbalance learning value Liv (i) is larger than zero, and becomes a positive value when it is smaller than zero. When the imbalance learning value Liv (i) is zero, the correction amount ΔQd (Liv (i)) is also zero.

When the CPU 72 completes the process of S26, the CPU 72 temporarily ends the series of processes shown in FIG. By the way, in the present embodiment, when the affirmative determination is made in the process of S10 and the process of S12 is executed, the process of S26 is temporarily stopped.

Next, a method for generating mapping data 76a will be described.
FIG. 5 shows a system for generating mapping data 76a.
As shown in FIG. 5, in the present embodiment, the dynamometer 100 is mechanically connected to the crankshaft 24 of the internal combustion engine 10 via the torque converter 60 and the transmission 64. Then, various state variables when the internal combustion engine 10 is operated are detected by the sensor group 102, and the detection results are input to the matching device 104, which is a computer that generates mapping data 76a. The sensor group 102 includes an air flow meter 80, which is a sensor for detecting a value for generating an input to a map, an air-fuel ratio sensor 82, and a crank angle sensor 86.

FIG. 6 shows a procedure for generating mapping data. The process shown in FIG. 6 is performed by the matching device 104. The process shown in FIG. 6 may be realized, for example, by equipping the conforming device 104 with a CPU and a ROM and executing the program stored in the ROM by the CPU.

In the series of processes shown in FIG. 6, the matching device 104 first acquires the same data as the training data acquired in the process of S12 based on the detection result of the sensor group 102 (S30). In this process, a plurality of fuel injection valves 20 having various values different from zero in the imbalance rate Riv and three fuel injection valves having an imbalance rate of zero are measured in advance by a single unit. This is performed with three fuel injection valves 20 having a zero imbalance rate and one fuel injection valve 20 having a different imbalance rate from zero mounted on the internal combustion engine 10. The imbalance rate Rivt of each of the mounted fuel injection valves is the teacher data.

Next, the matching device 104 substitutes training data other than the teacher data into the input variables x (1) to x (51) in the same manner as in S14 (S32). Then, the matching device 104 calculates the imbalance rates Riv (1) to Riv (4) using the input variables x (1) to x (51) obtained by the processing of S32 in the same manner as the processing of S16 (S34). ). Then, the CPU 72 determines whether or not the number of samples of the imbalance rate Rivi (i) calculated by the process of S34 is equal to or greater than a predetermined number (S36). Here, in order to be equal to or higher than a predetermined value, the operating state of the internal combustion engine 10 is changed with each of a plurality of fuel injection valves having an imbalance rate Rivt different from zero mounted on each of the cylinders # 1 to # 4. It is required that the imbalance rate Riv is calculated at various operating points defined by the rotation speed NE and the filling efficiency η.

When the conforming device 104 determines that the value is not equal to or higher than the predetermined value (S36: NO), the process returns to the process of S30. On the other hand, when the CPU 72 determines that the value is equal to or higher than a predetermined value (S36: YES), the difference between the imbalance coefficient Riv as teacher data and the imbalance coefficient Riv (i) calculated by the processing of S34 is different. The input-side coefficient wFjk and the output-side coefficient wSij are updated so as to minimize the sum of squares (S38). Then, the matching device 104 stores the updated input-side coefficient wFjk, output-side coefficient wSij, and the like as learned mapping data (S40).

Here, the action and effect of this embodiment will be described.
FIG. 7 shows a period of 360 ° CA for each of the case where the imbalance rate Riv (1) of the cylinder # 1 is positive and the case where the imbalance rate Riv (4) of the cylinder # 4 is negative. The transition of the crank counter, the upstream average value Afave, and the minute rotation time T30 is shown. As shown in FIG. 7, there is no large difference in the phase of the upstream average value Afave between the case where the fuel injection amount is excessive in the cylinder # 1 and the case where the fuel injection amount is insufficient in the cylinder # 4. It is difficult to identify in which cylinder the anomaly is occurring. However, according to the time series data of the minute rotation time T30, as can be seen from the place surrounded by the alternate long and short dash line in FIG. 7, there is a clear difference between these two types of abnormalities.

Therefore, in the present embodiment, the CPU 72 calculates the imbalance rate Riv by using the time series data of the minute rotation time T30 and the time series data of the upstream average value Afave. As a result, even if it is difficult to distinguish which cylinder has such an abnormality based only on the upstream detection value Afu, the imbalance rate of each cylinder Riv (1) to RIv ( 4) can be calculated.

Moreover, in the present embodiment, the invention is familiar with the control of the internal combustion engine 10 instead of learning the mapping for calculating the imbalance rate Riv by randomly and a large number of various variables of the internal combustion engine 10 by machine learning. Based on the knowledge of the person, the variables to be input to the map were carefully selected. Therefore, compared with the case where the inventor's knowledge is not used, it is possible to reduce the number of layers of the intermediate layer of the neural network and the dimension of the input variable, and the structure of the map for calculating the imbalance rate Riv (i) can be reduced. Easy to simplify.

According to the present embodiment described above, the effects described below can be further obtained.
(1) Rotation speed NE and filling efficiency η as operating point variables that define the operating point of the internal combustion engine 10 were used as mapping inputs. The amount of operation of the operation unit of the internal combustion engine 10 such as the ignition device 22, the EGR valve 34, and the intake side valve timing variable device 46 tends to be determined based on the operating point of the internal combustion engine 10. Therefore, the operating point variable is a variable that includes information on the amount of operation of each operation unit. Therefore, by using the operating point variable as the input of the map, the imbalance rate Riv (i) can be calculated based on the information on the operation amount of each operation unit, and the imbalance rate Riv (i) can be calculated with higher accuracy. Can be calculated.

(2) The upstream average value Afave was included in the input of the map. As a result, compared to the case where the upstream detection value Afu for each time interval of the time series data is used, the oxygen and unburned fuel flowing into the catalyst 30 are more accurate without increasing the number of data of the time series data. Information can be obtained, and the imbalance rate Riv (i) can be calculated with higher accuracy.

(3) By including the 0.5th-order amplitude Ampf / 2 in the input to the map, the imbalance rate Riv can be calculated with higher accuracy. That is, as shown in FIG. 8A, a linear relationship is established between the imbalance rate Riv and the 0.5th order amplitude Ampf / 2. Further, as shown in FIG. 8 (b) as an example of the case where the imbalance rate Riv (1) is "1.15", the amplitude of the rotation frequency of the crankshaft 24 is that of a cylinder in which the imbalance rate Riv is not zero. If present, the 0.5th order component is particularly large. It is considered that this is because when the imbalance rate Rivi (i) is different from zero in any one of the cylinders # 1 to # 4, the generated torque deviates once in one combustion cycle. In the present embodiment, the imbalance rate Riv can be calculated with higher accuracy by incorporating the torque fluctuation of the 720 ° CA cycle as the 0.5th order amplitude Ampf / 2.

<Second embodiment>
Hereinafter, the second embodiment will be described with reference to the drawings, focusing on the differences from the first embodiment.

In the present embodiment, after the imbalance rate Riv is calculated once, the target purge rate Rp * is switched from zero to a value larger than zero, and at that time, the imbalance rate Riv is calculated again. Then, after that, the EGR rate Regr is switched from zero to a value larger than zero, and the imbalance rate Riv is calculated again. Therefore, in the present embodiment, the execution conditions in the process of S10 in FIG. 3 are when the target purge rate Rp * is set to zero and the EGR rate Regr is set to zero, and when the target purge rate Rp * is zero. There are three cases, one is when the value is switched to a value larger than zero and the other is when the EGR rate Regr is switched from zero to a value larger than zero.

FIG. 9 shows a procedure for calculating the learning value based on the imbalance rate Riv calculated in this way. The process shown in FIG. 9 is realized by the CPU 72 repeatedly executing the coping program 74b stored in the ROM 74, for example, at a predetermined cycle.

In the series of processes shown in FIG. 9, the CPU 72 first determines whether or not the imbalance rate Riv is newly calculated (S50). When the CPU 72 determines that the calculation has been performed (S50: YES), the input variable used for calculating the imbalance rate Riv is sampled when the target purge rate Rp * is zero and the EGR rate Regr is zero. It is determined whether or not it has been done (S52). When the CPU 72 determines affirmatively in the process of S52, the CPU 72 executes the same process as S20 (S54).

On the other hand, when making a negative determination in the processing of S52, the CPU 72 determines whether or not the input variable used for calculating the imbalance rate Riv is sampled when the EGR rate Regr is zero. (S56). Then, when the CPU 72 determines that the sample was sampled when the EGR rate Regr is zero (S56: YES), the CPU 72 is purged by an exponential moving average process using the imbalance rate Riv (i) calculated this time as an input. The hourly learning value Livep (i) is updated (S58).

On the other hand, when the CPU 72 determines that the input variable used for calculating the imbalance rate Riv is sampled when the EGR rate Regr is larger than zero (S56: NO), the imbalance calculated this time. The learning value Live (i) at the time of EGR is updated by the exponential moving average processing in which the rate Riv (i) is input (S60).

When the processing of S54, S58, and S60 is completed, or when a negative determination is made in the processing of S50, the CPU 72 temporarily ends the series of processing shown in FIG.
FIG. 10 shows a procedure of processing using the learning value. The process shown in FIG. 10 is realized by the CPU 72 repeatedly executing the coping program 74b stored in the ROM 74 shown in FIG. 1, for example, at a predetermined cycle. In FIG. 10, the same step numbers are assigned to the processes corresponding to the processes shown in FIG. 4 for convenience.

In the series of processes shown in FIG. 10, when the CPU 72 makes a negative determination in S22 or completes the process in S24, whether or not the target purge rate Rp * is set to zero and the EGR rate Regr is set to zero. Is determined (S62). Then, when the CPU 72 determines that the target purge rate Rp * is zero and the EGR rate Regr is zero (S62: YES), the CPU 72 executes the same process as the process of S26 (S64).

On the other hand, when a negative determination is made in the process of S62, the CPU 72 determines whether or not the EGR rate Regr is zero (S66). Then, when the CPU 72 determines that the EGR rate Regr is zero (S66: YES), the CPU 72 sets the required injection amount Qd (#i) of each cylinder not only as the imbalance learning value Live (i) but also as a learning value during purging. Correction is made based on Livep (i) (S68). That is, the CPU 72 responds to the required injection amount Qd (#i) of each cylinder, the correction amount ΔQd (Live (i)) according to the imbalance learning value Live (i), and the purging learning value Livep (i). The required injection amount Qd (#i) is corrected by adding the correction amount ΔQdp (Livep (i)).

On the other hand, when the CPU 72 determines that the EGR rate Regr is larger than zero (S66: NO), the CPU 72 sets the required injection amount Qd (#i) of each cylinder to the imbalance learning value Live (i) and the purging learning value Livep. Correction is made based on (i) and the EGR learning value Live (i) (S70). That is, the CPU 72 sets the required injection amount Qd (#i) of each cylinder according to the correction amount ΔQd (Live (i)) according to the imbalance learning value Live (i) and the purging learning value Livep (i). The correction amount is corrected by the sum of the correction amount ΔQdp (Live (i)) and the correction amount ΔQde (Live (i)) corresponding to the EGR learning value Live (i).

When the processes of S64, S68, and S70 are completed, the CPU 72 temporarily ends the series of processes shown in FIG. By the way, in the present embodiment, when the process of S12 for calculating the imbalance rate Riv used in the process of S58 is executed, the correction amount ΔQd according to the learning value Livep (i) at the time of purging in the process of S68. (Live (i)) is not used. Further, when the processing of S12 for calculating the imbalance rate Riv used in the processing of S60 is executed, in the processing of S70, the correction amount ΔQd (Live (i)) corresponding to the learning value Live (i) at the time of EGR is executed. ) Is not used.

Further, in the present embodiment, the mapping data 76a is training data for data at the time when the affirmative judgment is made in the processing of S62, when the positive judgment is made in the processing of S66, and when the negative judgment is made in the processing of S66. It is desirable that it was learned as. However, instead of using the mapping data 76a as a single data, affirmative judgment is made in the processing of S62, affirmative judgment is made in the processing of S66, and negative judgment is made in the processing of S66. It may be data. In that case, each data shall be learned using only the corresponding training data.

According to the present embodiment described above, in addition to those described in the first embodiment, the effects described below can be obtained.
(4) Purging by calculating the imbalance rate Riv when the target purge rate Rp * is made larger than zero and correcting the required injection amount Qd based on the learning value Livep (i) at the time of purging corresponding to this. Even when the rate is larger than zero, the variation in the air-fuel ratio of cylinders # 1 to # 4 can be suppressed. That is, when the purge rate is larger than zero, fuel vapor flows from the canister 40 into the intake passage 12, and although this flows into each of cylinders # 1 to # 4, the amount of inflow is in each of cylinders # 1 to # 4. There are variations in. This variation cannot be expressed by the imbalance learning value Live (i) that quantifies the variation caused by individual differences in the fuel injection valve 20, secular change, and the like. On the other hand, in the present embodiment, the variation in the air-fuel ratio of the cylinders # 1 to # 4 can be suppressed by correcting the required injection amount Qd based on the learning value at the time of purging Livep (i).

(5) The imbalance rate Riv when the EGR rate Regr is made larger than zero is calculated, and the required injection amount Qd is corrected based on the EGR learning value Live (i) corresponding to the calculation, so that the EGR rate Regr Even when is larger than zero, the variation in the air-fuel ratio of cylinders # 1 to # 4 can be suppressed. That is, when the EGR ratio Regr is larger than zero, the exhaust gas from the exhaust passage 28 flows into the intake passage 12 through the EGR passage 32, and although this flows into each of the cylinders # 1 to # 4, the cylinder # The amount of inflow varies from 1 to # 4. This variation cannot be expressed by the imbalance learning value Live (i) that quantifies the variation caused by individual differences in the fuel injection valve 20, secular change, and the like. On the other hand, in the present embodiment, the variation in the air-fuel ratio of the cylinders # 1 to # 4 can be suppressed by correcting the required injection amount Qd based on the EGR learning value Live (i).

<Third embodiment>
Hereinafter, the third embodiment will be described with reference to the drawings, focusing on the differences from the second embodiment.

In the present embodiment, when the imbalance is not sufficiently eliminated by the injection amount correction, the ignition timing is manipulated.
FIG. 11 shows a procedure of processing related to the operation of the ignition timing according to the imbalance rate Riv. The process shown in FIG. 11 is performed at a predetermined cycle on the condition that the CPU 72 updates the coping program 74b stored in the ROM 74 shown in FIG. 1, for example, the imbalance learning value Live (i) or the like at a predetermined number or more. It is realized by repeating the execution.

In the series of processes shown in FIG. 11, the CPU 72 was newly calculated when the process of S64 was executed when the target purge rate Rp * was set to zero and the EGR rate Regr was set to zero. It is determined whether or not the absolute value of the imbalance rate Rivi (i) is equal to or greater than the predetermined value Δth (S72). This process is a process of determining whether or not the imbalance is not sufficiently eliminated depending on the correction amount ΔQd (Liv (i)) based on the imbalance learning value Live (i). When the CPU 72 determines that the value is Δth or more (S72: YES), the CPU 72 corrects the ignition timing aid (#i) of each cylinder #i based on the newly calculated imbalance rate Rivi (i) ( S74). Here, for example, the CPU 72 executes a process of advancing the ignition timing aig for a cylinder in which the absolute value of the imbalance rate Rivi (i) is larger than zero. This is because if the imbalance rate Riv deviates from zero due to the air-fuel ratio becoming excessively rich and misfire, there is a possibility that combustion will be improved by advancing the ignition timing aig. Because there is.

When the processing of S74 is completed, or when a negative determination is made in the processing of S72, the CPU 72 has an imbalance rate Rivi newly calculated when the processing of S68 is executed when the EGR rate Regr is zero. It is determined whether or not the absolute value of (i) is equal to or greater than the predetermined value Δthp (S76). This process is a process of determining whether or not the imbalance is not sufficiently eliminated depending on the correction amount ΔQd (Livep (i)) based on the learning value at the time of purging Livep (i). When the CPU 72 determines that the value is Δthp or more (S76: YES), the CPU 72 sets the ignition timing aid (#i) of each cylinder #i to the newly calculated imbalance rate Rivi (i) and the target purge rate Rp. Correct based on * (S78).

When the processing of S78 is completed, or when a negative determination is made in the processing of S76, the CPU 72 newly calculates the imbalance rate when the processing of S70 is executed when the EGR rate Regr is larger than zero. It is determined whether or not the absolute value of Rivi (i) is equal to or greater than the predetermined value Δthe (S80). This process is a process of determining whether or not the imbalance is not sufficiently eliminated depending on the correction amount ΔQd (Live (i)) based on the EGR learning value Live (i). When the CPU 72 determines that the value is equal to or greater than the predetermined value Δthe (S80: YES), the CPU 72 sets the ignition timing aid (#i) of each cylinder #i to the newly calculated imbalance rate Rivi (i) and the target purge rate Rp. Correction is made based on * and the EGR rate Regr (S82).

When the processing of S82 is completed or when a negative determination is made in the processing of S80, the CPU 72 temporarily ends the series of processing shown in FIG.
<Fourth Embodiment>
Hereinafter, the fourth embodiment will be described with reference to the drawings, focusing on the differences from the first embodiment.

In the present embodiment, as the mapping data 76a, a plurality of types of mapping data are stored in the storage device 76. The plurality of types of mapping data are mapping data for each control mode.
FIG. 12 shows a procedure of processing performed prior to the processing of FIG. The process shown in FIG. 12 is realized by the CPU 72 repeatedly executing the imbalance detection program 74a stored in the ROM 74, for example, at a predetermined cycle.

In the series of processes shown in FIG. 12, the CPU 72 determines whether or not the catalyst 30 is warming up (S84). The warm-up process includes a process of retarding the ignition timing aim by a predetermined amount. The warm-up process is performed, for example, when the logical product of the integrated value of the intake air amount Ga from the start of the internal combustion engine 10 is not more than a predetermined value and the water temperature THW is not more than a predetermined temperature is true. It is executed by the CPU 72.

When the CPU 72 determines that it is not during the catalyst warm-up process (S84: NO), the CPU 72 selects post-warm-up mapping data from the mapping data 76a stored in the storage device 76 (S86). As a result, the CPU 72 calculates the imbalance rate Rivi (i) using the post-warm-up mapping data in the processing of S16. Here, as the post-warm-up mapping data, training data is generated based on the detection values obtained by the sensor group 102 shown in FIG. 5 after the internal combustion engine 10 is warmed up, and the training data is used in FIG. It is the data generated by the processing shown.

On the other hand, when the CPU 72 determines that the catalyst warm-up process is in progress (S84: YES), the CPU 72 selects the warm-up mapping data from the mapping data 76a stored in the storage device 76 (S88). .. The warm-up mapping data is data generated based on training data corresponding to the detected values obtained by the sensor group 102 while the warm-up process of the internal combustion engine 10 is being executed.

When the processes of S86 and S88 are completed, the CPU 72 temporarily ends the series of processes shown in FIG.
Here, the action and effect of this embodiment will be described.

When the warm-up process of the catalyst 30 is executed, the CPU 72 calculates the imbalance rate Riv using the map data for warm-up, and when the warm-up process is not executed, the CPU 72 is used after warm-up. The imbalance rate Riv is calculated using the mapping data. Since the ignition timing aid is set to the retard side during the warm-up treatment of the catalyst 30, the rotational behavior of the crankshaft 24 is different from that when the warm-up treatment is not performed. Nevertheless, when trying to generate a single map that calculates the imbalance rate Riv with high accuracy, the input variable x is the data that can distinguish between when the catalyst 30 is warmed up and when it is not. It is desirable to add it, and in that case, since the number of input-side coefficients wFjk and the like will be increased, a large amount of training data is required. In particular, whether or not the execution condition of the warm-up treatment is satisfied is not determined only by the two sampling values of the intake air amount Ga and the water temperature THW, so that the catalyst 30 can be distinguished between the warm-up treatment and the non-warm-up treatment. In order to make this possible, the dimension of the input variable tends to be excessive. Then, when expanding the dimension of the input variable, there is a tendency to request to complicate the structure of the map, for example, to increase the intermediate layer of the neural network. However, when the structure of the mapping is complicated, such as by increasing the number of layers of the intermediate layer, the calculation load of the CPU 72 becomes large. On the other hand, in the present embodiment, by separating the mapping data during the catalyst warm-up and after the catalyst warm-up, the number of dimensions of the input variable x becomes excessive while ensuring the calculation accuracy of the imbalance rate Riv. Can be suppressed, and the structure of the map can be easily simplified.

<Fifth Embodiment>
Hereinafter, the fifth embodiment will be described with reference to the drawings, focusing on the differences from the first embodiment.

In the present embodiment, as the mapping data 76a, a plurality of types of mapping data are stored in the storage device 76. These mapping data are mapping data for each rotation speed NE.
FIG. 13 shows a procedure of the process performed prior to the process shown in FIG. The process shown in FIG. 13 is realized by the CPU 72 repeatedly executing the imbalance detection program 74a stored in the ROM 74, for example, at a predetermined cycle.

In the series of processes shown in FIG. 13, the CPU 72 first acquires the rotation speed NE (S90). Next, the CPU 72 determines whether or not the rotation speed NE is equal to or lower than the low rotation threshold value NE1 (S92). Then, when the CPU 72 determines that the low rotation threshold is NE1 or less (S92: YES), the CPU 72 selects the low rotation mapping data from the mapping data 76a stored in the storage device 76 (S94). As a result, the CPU 72 calculates the imbalance rate Riv using the low rotation mapping data in the processing of S16. Here, the low rotation mapping data was obtained by the sensor group 102 when the dynamometer 100 or the like was controlled in the system illustrated in FIG. 5 to realize a state in which the rotation speed NE was equal to or less than the low rotation threshold NE1. This is the data generated by the process shown in FIG. 6 by generating training data based on the detected values.

On the other hand, when the CPU 72 determines that the rotation speed NE is larger than the low rotation threshold value NE1 (S92: NO), the CPU 72 determines whether or not the rotation speed NE is equal to or less than the medium rotation threshold value NE2 (S96). The medium rotation threshold value NE2 is set to a value larger than the low rotation threshold value NE1. Then, when the CPU 72 determines that the threshold value NE2 or less for medium rotation is equal to or lower (S96: YES), the CPU 72 selects the mapping data for medium rotation from the mapping data 76a stored in the storage device 76 (S98). The medium rotation mapping data is the training data according to the detection value obtained by the sensor group 102 when the rotation speed NE of the internal combustion engine 10 is larger than the low rotation threshold NE1 and equal to or less than the medium rotation threshold NE2. It is the data generated based on.

On the other hand, when the CPU 72 determines that it is larger than the medium rotation threshold value NE2 (S96: NO), the CPU 72 selects the high rotation mapping data from the mapping data 76a stored in the storage device 76 (S100). The high rotation mapping data is data generated based on training data according to the detection value obtained by the sensor group 102 when the rotation speed NE of the internal combustion engine 10 is larger than the medium rotation threshold NE2.

When the processes of S94, S98, and S100 are completed, the CPU 72 temporarily ends the series of processes shown in FIG.
As described above, in the present embodiment, by properly using the mapping data according to the rotation speed NE, it is possible to use the appropriate mapping data at each rotation speed NE. Therefore, regardless of the magnitude of the rotation speed NE, it is easy to simplify the structure of the mapping while ensuring the calculation accuracy of the imbalance rate Riv as compared with the case of using the common mapping data.

<Sixth Embodiment>
Hereinafter, the sixth embodiment will be described with reference to the drawings, focusing on the differences from the first embodiment.

In the present embodiment, as the mapping data 76a, a plurality of types of mapping data are stored in the storage device 76. The plurality of types of mapping data are mapping data for each shift position Sft.
FIG. 14 shows a procedure for selecting mapping data according to the present embodiment. The process shown in FIG. 14 is realized by the CPU 72 repeatedly executing the imbalance detection program 74a stored in the ROM 74, for example, at a predetermined cycle.

In the series of processes shown in FIG. 14, the CPU 72 first acquires the shift position Sft (S102). Next, the CPU 72 determines whether the shift position Sft is the 1st speed (S104), the 2nd speed (S106), the 3rd speed (S108), or the 4th speed (S108). The shift position Sft is determined by the determination process of S110). Then, when the CPU 72 determines that the speed is 1st speed (S104: YES), the CPU 72 selects the 1st speed mapping data from the mapping data 76a stored in the storage device 76 (S112). As a result, the CPU 72 calculates the imbalance rate Riv using the 1st-speed mapping data in the processing of S16. Similarly, when the CPU 72 determines that the speed is 2nd (S106: YES), selects the 2nd speed mapping data (S114), and determines that the 3rd speed is 3rd speed (S108: YES), the 3rd speed mapping When data is selected (S116) and it is determined that the speed is 4th speed (S110: YES), mapping data for 4th speed is selected (S118). Further, when the CPU 72 determines that the speed is 5th (S110: NO), the CPU 72 selects the mapping data for the 5th speed (S120).

Here, assuming that the variables m are "1" to "5", the map data for m speed is the detected value obtained by the sensor group 102 shown in FIG. 5 when the transmission 64 is set to m speed. Training data is generated based on the above, and the data is generated by the process shown in FIG. 6 using the training data.

When the processes of S112 to S120 are completed, the CPU 72 temporarily ends the series of processes shown in FIG.
According to the present embodiment described above, by using the mapping data corresponding to the shift position Sft, the inertial constant from the crankshaft 24 to the drive wheel 69 changes according to the shift position Sft, and eventually the rotation of the crankshaft 24. The imbalance rate Riv can be calculated by appropriately reflecting the change in behavior. Therefore, as compared with the case where all shift positions Sft are dealt with by a single map, it is easy to simplify the structure of each map while ensuring the calculation accuracy of the imbalance rate Riv.

<7th Embodiment>
Hereinafter, the seventh embodiment will be described with reference to the drawings, focusing on the differences from the first embodiment.

In the present embodiment, as the mapping data 76a, a plurality of types of mapping data are stored in the storage device 76. The plurality of types of mapping data are mapping data for each road surface condition.
FIG. 15 shows a procedure of processing performed prior to the processing of FIG. The process shown in FIG. 15 is realized by the CPU 72 repeatedly executing the imbalance detection program 74a stored in the ROM 74, for example, at a predetermined cycle.

In the series of processes shown in FIG. 15, the CPU 72 first acquires the acceleration Dacc (S122). Next, the CPU 72 determines whether or not the bad road determination flag Fb is "1" (S124). The rough road determination flag Fb is set to "1" when there are many undulations on the road surface on which the vehicle VC is traveling, and is set to "0" when the road surface is not. When the CPU 72 determines that the rough road determination flag Fb is "0" (S124: YES), the CPU 72 determines whether or not the state in which the absolute value of the change amount of the acceleration Dacc is equal to or greater than the predetermined value ΔDH has continued for a predetermined time. (S126). Here, the amount of change may be the difference between the previous value and the current value of the acceleration Dacc. This process is a process of determining whether or not the road surface has many undulations. When the CPU 72 determines that the continuation has continued for a predetermined time (S126: YES), the CPU 72 substitutes "1" for the rough road determination flag Fb (S128), and the map data 76a stored in the storage device 76 is used for the rough road. The mapping data is selected (S130). As a result, the CPU 72 calculates the imbalance rate Riv using the rough road mapping data in the processing of S16. The training data for generating the mapping data for rough roads is shown in FIG. 5 while simulating the load torque applied to the crankshaft 24 when vibration is applied to the vehicle by the dynamometer 100 shown in FIG. 5, for example. It may be generated based on the detected value obtained by the sensor group 102. Further, for example, the internal combustion engine 10 and the dynamometer 100 shown in FIG. 5 are placed on a device capable of generating vibration, and the vibration is generated to simulate the load torque applied to the crankshaft 24 when the vehicle is vibrated. While doing so, training data may be generated based on the detection values obtained by the sensor group 102 shown in FIG.

On the other hand, when the CPU 72 determines that the rough road determination flag Fb is "1" (S124: YES), the state in which the absolute value of the change amount of the acceleration Dacc is equal to or less than the specified value ΔDL (<ΔDH) is a predetermined time. It is determined whether or not it has continued (S132). Then, when the CPU 72 determines negative in the process of S132, the CPU 72 shifts to the process of S130.

On the other hand, when it is determined that the CPU 72 has continued for a predetermined time (S132: YES), the CPU 72 substitutes "0" for the bad road determination flag Fb (S134). Then, when the processing of S134 is completed or when a negative determination is made in the processing of S126, the CPU 72 selects normal mapping data from the mapping data 76a (S136). For the normal mapping data, for example, training data is generated based on the detection values obtained by the sensor group 102 shown in FIG. 5 without intentionally applying vibration to the internal combustion engine 10 and the dynamometer 100 shown in FIG. It is the data generated by the process shown in FIG. 6 using the training data.

When the processes of S130 and S136 are completed, the CPU 72 temporarily ends the series of processes shown in FIG.
According to the present embodiment described above, by using the mapping data according to the road surface condition, the imbalance rate Riv is appropriately reflected that the rotational behavior of the crankshaft 24 changes according to the road surface condition. Can be calculated. Moreover, as compared with the case where a single mapping is used to deal with all road surface conditions, it is easy to simplify the structure of the mapping while ensuring the calculation accuracy of the imbalance rate Riv.

<8th Embodiment>
Hereinafter, the eighth embodiment will be described with reference to the drawings, focusing on the differences from the first embodiment.

In this embodiment, the dimension of the input variable x is expanded.
FIG. 16 shows the procedure of the imbalance detection process according to the present embodiment. The process shown in FIG. 16 is realized by the CPU 72 repeatedly executing the imbalance detection program 74a stored in the ROM 74, for example, at a predetermined cycle. Note that, in FIG. 16, the processes corresponding to the processes shown in FIG. 2 are given the same step numbers for convenience, and the description thereof will be omitted.

In the series of processes shown in FIG. 16, when the CPU 72 makes an affirmative decision in the process of S10, in addition to the variables acquired in the process of S12, the ignition timing average value agiave, the EGR rate average value Regave, the water temperature THW, and the shift position The Sft, the road surface state variable SR, and the alcohol concentration Dal are acquired (S12a). Here, the ignition timing average value agiave and the EGR rate average value Regave are average values of the ignition timing ai g and the EGR rate Regr in the processing cycle of FIG. Further, the road surface state variable SR is a parameter calculated by the CPU 72 based on the acceleration Dacc, and may be, for example, a variable that can take different binary values depending on whether the road surface has many undulations or few undulations. Incidentally, the process of generating the road surface state variable SR may be the same process as the process of setting the value of the bad road determination flag Fb in the process of FIG. 15, for example. Further, the alcohol concentration Da may be estimated from, for example, the correction ratio δ of the feedback process M20 described above.

Next, the CPU 72 inputs the data acquired in S12a into the mapping input variables x (1) to x (57) (S14a). Here, the average ignition timing value aigave and the average EGR rate retrieve are parameters (adjustment variables) for adjusting the combustion speed of the air-fuel mixture in the combustion chamber 18 by operating the operation unit of the internal combustion engine 10. That is, the ignition timing aig is an operation amount of the ignition device 22 as an operation unit, and is an adjustment variable for adjusting the combustion speed by operating the ignition device 22. The EGR rate Regr is an adjustment variable that adjusts the combustion speed by operating the EGR valve 34 as an operating unit. The reason why these adjustment variables are added to the input variable x is that the behavior of the crankshaft 24 changes depending on the combustion speed of the air-fuel mixture.

On the other hand, the water temperature THW is a state variable of the internal combustion engine 10, and is a parameter that changes the friction of the sliding portion, such as the friction between the piston and the cylinder. In view of the fact that the rotational behavior of the crankshaft 24 also changes depending on the water temperature THW, in the present embodiment, the water temperature THW is added to the input variable x.

Further, since the stoichiometric air-fuel ratio differs depending on the alcohol concentration Dal, the alcohol concentration Dal is added to the input variable in the present embodiment in view of the fact that the degree of variation in the fuel injection amount may change.

When the CPU 72 completes the processing of S14a, the CPU 72 inputs the input variables x (1) to x (57) into the mapping defined by the mapping data 76a, so that the imbalance rate Riv (i) of the cylinders # 1 to # 4 ) Is calculated (S16a). This map is composed of a neural network in which the number of intermediate layers is "α" and the activation functions h1 to hα of each intermediate layer and the activation function f of the output layer are hyperbolic tangents. For example, the value of each node in the first intermediate layer is a linear map defined by the coefficient w (1) ji (j = 0 to n1, i = 0 to 57), and the input variables x (1) to x ( It is generated by inputting the output when 57) is input to the activation function h1. That is, assuming that m = 1, 2, ..., α, the value of each node in the mth mesosphere is generated by inputting the output of the linear map defined by the coefficient w (m) into the activation function hm. Will be done. Here, n1, n2, ..., Nα are the number of nodes in the first, second, ..., And α intermediate layers, respectively. Incidentally, w (1) j0 and the like are bias parameters, and the input variable x (0) is defined as “1”.

When the CPU 72 completes the process of S16a, the CPU 72 temporarily ends the process of FIG.
In the present embodiment, the training data acquired in the process of S30 in FIG. 6 includes all the parameters acquired in S12a, the process corresponding to S14a is executed as the process of S32, and the input variable x is set in the process of S34. A 57-dimensional map is used. However, here, the input variable x (0) is not included in the number of dimensions. Further, the training data corresponding to the case where the road surface state variable SR indicates that the road surface has many undulations is the same as that used for generating the rough road mapping data shown in FIG.

As described above, in the present embodiment, the adjustment variable for adjusting the combustion speed, the state variable of the internal combustion engine 10, the drive system state variable, the road surface state variable SR, and the alcohol concentration Dal are added to the input variable x. Although these are parameters that affect the rotational behavior of the crankshaft 24, it has been difficult to add them by the conventional method of detecting imbalance based on the difference in time required for rotation of adjacent angular intervals. It is a thing. That is, when these parameters are taken into consideration, the matching man-hours become enormous and the feasibility is poor. On the other hand, in the present embodiment, by using machine learning, it is possible to construct a logic for calculating the imbalance rate Riv (i) in the range of practical matching man-hours in consideration of these.

<9th embodiment>
Hereinafter, the ninth embodiment will be described with reference to the drawings, focusing on the differences from the first embodiment.

In the present embodiment, the imbalance rate Rivi (i) is calculated outside the vehicle.
FIG. 17 shows an imbalance detection system according to this embodiment. In FIG. 17, the members corresponding to the members shown in FIG. 1 are designated by the same reference numerals for convenience.

The control device 70 in the vehicle VC shown in FIG. 17 includes a communication device 79. The communication device 79 is a device for communicating with the center 120 via the network 110 outside the vehicle VC.

The center 120 analyzes the data transmitted from the plurality of vehicle VCs. The center 120 includes a CPU 122, a ROM 124, a storage device 126, a peripheral circuit 127, and a communication device 129, which can be communicated by the local network 128. The imbalance detection main program 124a is stored in the ROM 124, and the mapping data 126a is stored in the storage device 126.

FIG. 18 shows a procedure of processing executed by the system shown in FIG. The process shown in FIG. 18A is realized by the CPU 72 executing the imbalance detection subprogram 74c stored in the ROM 74 shown in FIG. Further, the process shown in FIG. 18B is realized by the CPU 122 executing the imbalance detection main program 124a stored in the ROM 124. Note that the processes corresponding to the processes shown in FIG. 16 in FIG. 18 are assigned the same step numbers for convenience. In the following, the process shown in FIG. 18 will be described along with the time series of the imbalance detection process.

As shown in FIG. 18A, when the CPU 72 completes the processing of S12a in the vehicle VC, the CPU 72 operates the communication device 79 to obtain the data acquired in the processing of S12a as the vehicle ID which is the identification information of the vehicle VC. It is transmitted to the center 120 together with (S132).

On the other hand, as shown in FIG. 18B, the CPU 122 of the center 120 receives the transmitted data (S140) and executes the processes of S14a and S16a. Then, by operating the communication device 129, the CPU 122 transmits a signal regarding the imbalance rate Rivi (i) to the vehicle VC to which the data received by the processing of S140 is transmitted (S142), and is shown in FIG. 18 (b). The series of processing is temporarily terminated. On the other hand, as shown in FIG. 18A, the CPU 72 receives the signal relating to the imbalance rate Rivi (i) (S134), and temporarily ends the series of processes shown in FIG. 18A.

As described above, in the present embodiment, since the imbalance rate Riv is calculated at the center 120, the calculation load of the CPU 72 can be reduced.
<Correspondence>
The correspondence between the matters in the above embodiment and the matters described in the column of "Means for Solving the Problem" is as follows. In the following, the correspondence is shown for each number of the solution means described in the column of "Means for solving the problem". [1,7] The imbalance detection device corresponds to the control device 70. The execution device corresponds to the CPU 72 and the ROM 74. The storage device corresponds to the storage device 76. The predetermined rotation angle interval corresponds to 720 ° CA, the first interval and the second interval correspond to 30 ° CA, the air-fuel ratio detection variable corresponds to the upstream average value Wave, and the instantaneous velocity variable corresponds to. It corresponds to the minute rotation time T30, and the plurality of rotation waveform variables correspond to the minute rotation times T30 (1) to T30 (24). The acquisition process corresponds to the processes of S12 and S12a. The calculation process corresponds to the processing of S14 and S16 and the processing of S14a and S16a. The coping process corresponds to the process of S24, S26, S64, S68, S70, S74, S78, S82. [2] The 0.5th-order component variable corresponds to the 0.5th-order amplitude Ampf / 2. [3] The operating point variable corresponds to the rotation speed NE and the filling efficiency η. [4] The adjustment variables correspond to the average ignition timing value agiave and the average EGR rate Regave. [5] The drive system state variable corresponds to the shift position Sft. [6] The road surface state variable corresponds to the road surface state variable SR. [8] The selection process corresponds to the process of FIGS. 12 to 15. [9] The adjusting device corresponds to the purge valve 44. The imbalance variable depending on the presence or absence of the influence of the fuel vapor corresponds to the imbalance rate Riv when the affirmative judgment is made and the imbalance rate Riv when the negative judgment is made in the processing of S52. [10] The imbalance variables according to the presence or absence of the influence of the exhaust gas flowing in through the EGR passage are the imbalance rate Riv when affirmatively determined and the imbalance rate Riv when negatively determined in the processing of S56. Correspond. [11] The operation unit for controlling combustion corresponds to the fuel injection valve 20 and the ignition device 22. [12] The correction amounts different from each other correspond to ΔQd. [13] The first execution device corresponds to the CPU 72 and the ROM 74. The second execution device corresponds to the CPU 122 and the ROM 124. The calculation process corresponds to the processes of S14a and S16a of FIG. The vehicle-side transmission process corresponds to the process of S132, and the vehicle-side reception process corresponds to the process of S134. The external reception process corresponds to the process of S140, and the external transmission process corresponds to the process of S142. [14] The data analysis device corresponds to the center 120. [15] The control device for the internal combustion engine corresponds to the control device 70 shown in FIG.

<Other Embodiments>
In addition, this embodiment can be implemented by changing as follows. The present embodiment and the following modified examples can be implemented in combination with each other within a technically consistent range.

・ "About the 1st interval and the 2nd interval"
The first interval, which is the sampling interval of the upstream average value Afave that is the input to the map, is not limited to 30 ° CA. The angle interval may be smaller than 30 ° CA, for example, 10 ° CA. However, the interval is not limited to 30 ° CA or less, and may be, for example, 45 ° CA.

The second interval, which is the sampling interval of the minute rotation time T30 that is the input to the map, is not limited to 30 ° CA. The angle interval may be smaller than 30 ° CA, for example, 10 ° CA. However, the interval is not limited to 30 ° CA or less, and may be, for example, 45 ° CA.

It is not essential that the first interval and the second interval have the same size.
・ "Regarding the predetermined rotation angle interval"
The time-series data of the upstream average value Afave and the time-series data of the minute rotation time T30, which are the inputs of the mapping, are not limited to the time-series data at the rotation angle interval of 720 °. For example, it may be time series data at a rotation angle interval wider than 720 ° CA. Further, for example, as described in the column of "Multiple types of mapping data" below, when a mapping that outputs only the imbalance rate Riv (i) of a specific cylinder #i is used, the angle is narrower than 720 ° CA. It may be time series data of intervals.

Further, for example, when the time series data of the minute rotation time T30 in the first predetermined period and the time series data of the upstream average value Afave in the second predetermined period are input to the mapping, the first predetermined period and The other may be delayed relative to one of the two periods of the second predetermined period. It is not essential that the first predetermined period and the second predetermined period have an angular interval equal to each other.

・ "About the rotation waveform variable as an input to the map"
In the above embodiment, the minute rotation time T30 at each of the plurality of intervals in which the rotation angle interval of 720 ° CA, which is one combustion cycle, is divided is used as an input to the map, but the present invention is not limited to this. For example, of 0 to 720 ° CA, 0 to 20, 40 to 60, 80 to 100, 120 to 140, 160 to 180, ..., 700 to 720 are set as the second intervals, and the time required for their rotation is set. May be used as an input to the map.

The instantaneous velocity variable is not limited to the minute rotation time, which is the time required for the rotation of the second interval. For example, it may be a value obtained by dividing the second interval by a minute rotation time. It is not essential that the instantaneous velocity variable has been normalized so that the difference between the maximum value and the minimum value is a fixed value. Further, the filter processing as the preprocessing for inputting the map is not limited to the above-mentioned one, and the crankshaft 24 is rotated by the input shaft 66 based on, for example, the minute rotation time of the input shaft 66 of the transmission 64. It may be considered that the processing for removing the affected part has been performed. However, it is not essential that the instantaneous velocity variable as the input of the map is filtered.

The rotation waveform variable as an input to the map is not limited to the time series data of the instantaneous velocity variable. For example, it may be time series data of the difference between a pair of instantaneous velocity variables separated by the appearance interval of the compression top dead center. At this time, the instantaneous velocity variable is a value obtained by dividing the minute rotation time, which is the time required for the rotation, or the second interval by the minute rotation time, with the angle interval equal to or less than the appearance interval of the compression top dead center as the second interval. ..

・ "About the air-fuel ratio detection variable as an input to the map"
In the above embodiment, the upstream average value Afave at each of the plurality of intervals in which the rotation angle interval of 720 ° CA, which is one combustion cycle, is divided is used as an input to the map, but the present invention is not limited to this. For example, out of 0 to 720 ° CA, 0 to 20, 40 to 60, 80 to 100, 120 to 140, 160 to 180, ..., 700 to 720 are set as the first intervals, and the upstream average value in them is set. Average may be used as an input to the map.

In the above embodiment, the upstream average value Afave is set to the average value of a plurality of sampling values in one first interval, but the present invention is not limited to this. For example, the upstream average value Afuave may be calculated by sampling the upstream detection value Afu in each of the first intervals over a plurality of combustion cycles and averaging them for each first interval.

The air-fuel ratio detection variable is not limited to the upstream average value Afuave, but may be the upstream side detection value Afu.
・ "About operating point variables of internal combustion engines"
In the above embodiment, the operating point variable is defined by the rotation speed NE and the filling efficiency η, but the present invention is not limited to this. For example, the rotation speed NE and the intake air amount Ga may be used. Further, for example, as the load, the injection amount or the required torque for the internal combustion engine may be used instead of the filling efficiency η. Using the injection amount and the required torque as the load is particularly effective in the compression ignition type internal combustion engine described in the column of "About the internal combustion engine" below.

・ "Regarding the adjustment variable that adjusts the combustion rate of the air-fuel mixture as the input of the map"
In the processing of S16a, the ignition timing average value aidave and the EGR rate average value Regave were used as the adjustment variables for inputting the map and adjusting the combustion rate, but for these two parameters, for example, only one was used. You may use it.

The ignition variable as a variable indicating the ignition timing is not limited to the average ignition timing value aigave, and may be the ignition timing aig. The variable indicating the EGR rate is not limited to the EGR rate average value Regave, and may be the EGR rate Regr.

Further, for example, when applied to a compression ignition type internal combustion engine as described in the column of "About the internal combustion engine" below, the injection timing or its average value may be used instead of the ignition variable.
The adjustment variables that input the map and adjust the combustion speed are not limited to the above variables, but may be, for example, the valve opening timing and lift amount of the intake valve 16, the valve opening timing and lift amount of the exhaust valve 26, and the like. Good. Further, for example, during cylinder deactivation control in which the combustion control of the air-fuel mixture is stopped in some cylinders, the information may be the information of the cylinder in which the combustion control is stopped. Further, for example, in the case of an internal combustion engine 10 including a supercharger and a wastegate valve, the opening degree of the wastegate valve may be used.

・ "About drive system state variables as input of mapping"
In the process of S16a, the shift position Sft was used as a state variable of the drive system device connected to the crankshaft 24, but the present invention is not limited to this. For example, the lockup clutch 62 may be in the state. This state may be, for example, a binary state of a fastened state or a released state, but may identify each of the fastened state, the released state, and the slip state.

Further, for example, the oil temperature of the transmission 64 may be used. Further, for example, the rotation speed of the drive wheels 69 may be used. Further, as described in the column of "About the vehicle" below, when the internal combustion engine 10 and the motor generator are provided as the devices for generating the thrust, the torque and the rotation speed of the motor generator may be included. The state variable of the drive system device that is the input of the map is not limited to one.

・ "About road surface state variables as input of mapping"
The road surface state variable SR is not limited to a binary variable according to the magnitude of undulations, and may be, for example, a variable having three or more values according to the magnitude of undulations. Further, the method of generating the road surface state variable SR based on the acceleration Dacc is not limited to the method illustrated in FIG. 15, and for example, the average value of the absolute values of the acceleration Dacc per predetermined time may be used. According to this, the road surface state variable SR is a continuous quantity.

The road surface state variable SR is not limited to the one quantified based on the detected value of the vertical acceleration of the vehicle VC. For example, it may be quantified by the lateral acceleration or the front-rear acceleration of the vehicle. Further, as the input of the map, it is quantified using any two or all of the three accelerations of the vertical acceleration of the vehicle, the lateral acceleration and the front-rear acceleration. May be good.

The road surface state variable SR used in the processing of S16a is quantified by the acceleration Dacc detected by the acceleration sensor 92, but is not limited to this. For example, analysis result data of road surface imaging data by a camera or data related to radar reflected waves may be used. Further, for example, the map data having the road surface condition and GPS may be used to acquire the road surface condition at the current location and use this. Further, the air pressure of each wheel may be used. That is, the difference in air pressure between the wheels includes information on the inclination and unevenness of the road surface. Further, for example, a sensor such as a fuel gauge or an oil gauge that detects the position of the liquid level of the fluid in the vehicle may be used. That is, the fluctuation of the liquid level includes information such as unevenness of the road surface. Further, the detection value of the sensor that detects the hydraulic pressure of the suspension may be used. That is, this hydraulic pressure includes information such as unevenness of the road surface.

The road surface state variable is not limited to a quantified road surface unevenness or the like. For example, it may be a quantification of the magnitude of friction such as the presence or absence of freezing, or it may be a variable that distinguishes both the magnitude of undulation and the magnitude of friction.

・ "About map input"
The input of the map is not limited to the one illustrated in the above-described embodiment and the above-mentioned modified example. For example, parameters related to the environment in which the internal combustion engine 10 is placed may be used. Specifically, for example, atmospheric pressure may be used. Further, for example, the intake air temperature, which is a parameter that affects the combustion speed of the air-fuel mixture in the combustion chamber 18, may be used. Further, for example, humidity, which is a parameter that affects the combustion speed of the air-fuel mixture in the combustion chamber 18, may be used. A humidity sensor may be used to grasp the humidity, but the state of the wiper and the detection value of the sensor that detects raindrops may be used. Further, it may be data on the state of the auxiliary machine mechanically connected to the crankshaft 24.

It is not essential to include the operating point of the internal combustion engine 10 in the mapping input. For example, if the internal combustion engine is installed in a series hybrid vehicle and the operating point is limited to a narrow range as described in the "About internal combustion engine" column below, the operating point is set. Even if it is not included, the imbalance rate Riv (i) can be calculated with high accuracy.

・ "About mapping data for each control mode"
As illustrated in FIG. 12, the mapping data for each control mode is not limited to the data provided for each of the warm-up control mode of the catalyst 30 and the mode after the warm-up is completed. For example, in the case of an internal combustion engine having both a port injection valve and an in-cylinder injection valve as described in the column of "About the internal combustion engine 10" below, a mode in which fuel injection control is performed only by the port injection valve. , It may be provided separately for each mode in which fuel injection control using the in-cylinder injection valve is performed.

・ "About the mapping data for each state variable of the internal combustion engine as a result of control"
In FIG. 13, the mapping data for each rotation speed NE is set to 3 types, but the present invention is not limited to this, and may be 2 types or 4 or more types.

The mapping data for each state variable of the internal combustion engine 10 as a result of the control is not limited to the mapping data for each rotation speed NE illustrated in FIG. For example, it may be provided for each filling efficiency η. Further, for example, it may be provided for each operating point of the internal combustion engine 10, such as for each rotation speed NE and filling efficiency η.

・ "About mapping data for each drive system state variable"
For example, as described in the "Other" column below, when a continuously variable transmission is used as the transmission, it is sufficient to provide mapping data for each of a plurality of divided regions where the gear ratio can be obtained.

The mapping data for each drive system state variable connected to the crankshaft 24 is not limited to the mapping data for each shift position Sft. For example, instead of using the drive system state variable described in the column of "driving system state variable as a mapping input" as the mapping input, mapping data may be provided for each of the same state variables. However, when mapping data is provided for each of these state variables, it is not essential that the state variables are not included in the mapping input.

・ "About mapping data for each road surface state variable"
The mapping data for each road surface state variable is not limited to that illustrated in FIG. For example, instead of not using the road surface state variable SR described in the column of "road surface state variable as input of mapping" as the input of the mapping, a mapping may be provided for each road surface state variable SR. However, when a mapping is provided for each road surface state variable SR, it is not essential that the road surface state variable SR is not included in the input of the mapping.

・ "About multiple types of mapping data"
As for providing a plurality of types of mapping data, those provided for each state variable of the internal combustion engine 10 as a result of control, those provided for each control mode, those provided for each drive system state variable, and those provided for each road surface state variable. Not limited to things. For example, instead of not using the adjustment variable described in the column of "adjustment variable for adjusting the combustion rate of the air-fuel mixture as the input of the mapping" as the input of the mapping, mapping data may be provided for each adjustment variable. However, when mapping data is provided for each adjustment variable such as an EGR variable, it is not essential not to include the adjustment variable in the mapping input. Further, for example, instead of using the parameters related to the environment in which the internal combustion engine is placed described in the column of "input of map" as the input of the map, the map data may be provided for each parameter related to the environment. However, when mapping data is provided for each environment-related parameter, it is not essential that the parameter is not included in the mapping input.

Further, for example, different mapping data may be provided depending on whether the target purge rate Rp * is zero or larger than zero.
Further, the plurality of types of mapping data may be a combination of some of the above mapping data. That is, for example, the mapping data for each shift position Sft may be further subdivided into mapping data for each rotation speed NE.

In addition, mapping data that defines a mapping that outputs the imbalance rate Riv of a specific cylinder may be provided for the number of cylinders.
・ "About mapping data"
For example, in the process of FIG. 16, the number of layers of the intermediate layer of the neural network may be two or more.

In the processing of S16a, the number of layers of the intermediate layer of the neural network is expressed as more than two layers, but the present invention is not limited to this. In particular, from the viewpoint of reducing the calculation load of the control device 70, it is desirable to limit the number of layers of the intermediate layer of the neural network to one or two layers.

In the above embodiment, the activation functions h, f, h1, h2, ... Hα are set as hyperbolic tangents, but the present invention is not limited to this. For example, a part or all of the activation functions h, h1, h2, ... Hα may be ReLU, or may be, for example, a logistic jigmoid function.

In FIG. 3 and the like, the map that outputs the imbalance rate Riv is not limited to the one learned for the phenomenon that the fuel injection amount deviates from the assumed one only for any one cylinder. For example, it may be learned for a phenomenon in which the fuel injection amount deviates from the assumed one in two cylinders. In this case, the intensity of the primary component of the rotation frequency of the crankshaft 24 may be included in the input to the map.

The input to the neural network is not limited to one in which each dimension consists of a single physical quantity. For example, for some of the multiple types of physical quantities using the physical quantities illustrated above, instead of using them as direct inputs to the neural network, some principal components by their principal component analysis are directly input to the neural network. It may be input. However, when the principal component is the input of the neural network, it is not essential that only a part of the input to the neural network is the principal component, and the entire input may be the principal component. When the input to the neural network includes a principal component, the mapping data 76a and 126a include data that defines the mapping that defines the principal component.

・ "About machine learning algorithms"
The machine learning algorithm is not limited to the one using a neural network. For example, a regression equation may be used. This corresponds to the neural network having no intermediate layer.

・ "Generation of mapping data"
In the above embodiment, the data when the internal combustion engine 10 is operated with the dynamometer 100 connected to the crankshaft 24 via the torque converter 60 and the transmission 64 is used as training data, but the training data is not limited to this. For example, data when the internal combustion engine 10 is driven while mounted on the vehicle VC may be used as training data.

・ "About data analysis equipment"
At the center 120, the process of S22 and the process of notifying that there is an abnormality in the user's mobile terminal instead of the process of S24 may be executed.

The process of FIG. 18B may be executed by, for example, a mobile terminal owned by the user.
・ "About the execution device"
The execution device is not limited to the one provided with the CPU 72 (122) and the ROM 74 (124) to execute software processing. For example, a dedicated hardware circuit (for example, ASIC or the like) for hardware processing at least a part of the software processed in the above embodiment may be provided. That is, the executing device may have any of the following configurations (a) to (c). (A) A processing device that executes all of the above processing according to a program and a program storage device such as a ROM that stores the program are provided. (B) A processing device and a program storage device that execute a part of the above processing according to a program, and a dedicated hardware circuit that executes the remaining processing are provided. (C) A dedicated hardware circuit for executing all of the above processes is provided. Here, there may be a plurality of software execution devices including a processing device and a program storage device, and a plurality of dedicated hardware circuits.

・ "About storage device"
In the above embodiment, the storage device that stores the mapping data 76a and 126a is different from the storage device (ROM74, 124) that stores the imbalance detection program 74a and the imbalance detection main program 124a. , Not limited to this.

・ "About internal combustion engine"
In the above embodiment, as the fuel injection valve, an in-cylinder injection valve that injects fuel into the combustion chamber 18 is exemplified, but the present invention is not limited to this. For example, it may be a port injection valve that injects fuel into the intake passage 12. Further, for example, both a port injection valve and an in-cylinder injection valve may be provided.

The internal combustion engine is not limited to the spark ignition type internal combustion engine, and may be, for example, a compression ignition type internal combustion engine that uses light oil or the like as fuel.
It is not essential that the internal combustion engine constitutes the drive system. For example, it may be mounted on a so-called series hybrid vehicle in which a crankshaft is mechanically connected to an in-vehicle generator and power transmission is cut off from the drive wheels 69.

・ "About the vehicle"
The vehicle is not limited to a vehicle in which the device that generates the propulsive force of the vehicle is only an internal combustion engine. For example, in addition to the series hybrid vehicle described in the "About internal combustion engine" column, a parallel hybrid vehicle or a series parallel hybrid It may be a car.

·"others"
The drive system device interposed between the crankshaft and the drive wheels is not limited to a stepped transmission, and may be, for example, a continuously variable transmission.

10 ... internal combustion engine, 12 ... intake passage, 14 ... throttle valve, 16 ... intake valve, 18 ... combustion chamber, 20 ... fuel injection valve, 22 ... ignition device, 24 ... crankshaft, 26 ... exhaust valve, 28 ... exhaust passage , 30 ... catalyst, 32 ... EGR passage, 34 ... EGR valve, 36 ... pump, 38 ... fuel tank, 40 ... canister, 42 ... purge passage, 44 ... purge valve, 46 ... intake side valve timing variable device, 48 ... intake side Cam shaft, 50 ... crank rotor, 52 ... tooth, 54 ... missing tooth, 60 ... torque converter, 62 ... lockup clutch, 64 ... transmission, 66 ... input shaft, 68 ... output shaft, 69 ... drive wheel, 70 ... control device, 72 ... CPU, 74 ... ROM, 74a ... imbalance detection program, 74b ... coping program, 74c ... imbalance detection subprogram, 76 ... storage device, 76a ... mapping data, 77 ... peripheral circuit, 78 ... Local network, 79 ... communication equipment, 80 ... air flow meter, 82 ... air fuel ratio sensor, 86 ... crank angle sensor, 88 ... water temperature sensor, 90 ... shift position sensor, 92 ... acceleration sensor, 98 ... warning light, 100 ... dynamometer , 102 ... sensor group, 104 ... conforming device, 110 ... network, 120 ... center, 122 ... CPU, 124 ... ROM, 124a ... imbalance detection main program, 126 ... storage device, 126a ... mapping data, 127 ... peripheral circuit, 128 ... local network, 129 ... communicator.

Claims (15)

An imbalance detector applied to multi-cylinder internal combustion engines.
Equipped with a storage device and an execution device,
The storage device receives a rotation waveform variable and an air fuel ratio detection variable which is a variable corresponding to the output of the air fuel ratio sensor at each of the plurality of first intervals, and is a variable indicating the degree of variation in the air fuel ratio of the internal combustion engine. It stores mapping data, which is the data that defines the mapping that outputs a certain imbalance variable.
The execution device is acquired by the acquisition process of acquiring the rotation waveform variable based on the detection value of the sensor that detects the rotational behavior of the crankshaft and the air fuel ratio detection variable at each of the plurality of first intervals, and the acquisition process. A calculation process for calculating the imbalance variable based on the output of the map with a value as an input, and a large variation in the air-fuel ratio by operating a predetermined hardware based on the calculation result of the calculation process. Execute the corrective action to deal with it,
The rotation waveform variable is a variable indicating the difference between the instantaneous speed variables, which are variables corresponding to the rotation speed of the crankshaft at each of the plurality of second intervals.
The first interval and the second interval are both angular intervals of the crankshaft smaller than the appearance interval of the compression top dead center.
An imbalance detection device in which the rotation waveform variable and the plurality of air-fuel ratio detection variables used as input of the map are time-series data within a predetermined angular interval larger than the appearance interval, respectively. The mapping input includes a 0.5th order component of the rotation frequency of the crankshaft based on the detection value of the sensor that detects the rotation behavior of the crankshaft.
The acquisition process includes a process of acquiring a 0.5th-order component variable, which is a variable that defines the 0.5th-order component.
The calculation process is the process of calculating the imbalance variable based on the output of the map, which further includes the 0.5th-order component variable acquired by the acquisition process in the input to the map. Imbalance detector. The input of the map includes an operating point variable, which is a variable that defines the operating point of the internal combustion engine.
The acquisition process includes a process of acquiring the operating point variable.
The input according to claim 1 or 2, wherein the calculation process is a process of calculating the imbalance variable based on the output of the map, which further includes the operating point variable acquired by the acquisition process in the input to the map. Balance detector. The input of the mapping includes an adjustment variable which is a variable for adjusting the combustion speed of the air-fuel mixture in the combustion chamber of the internal combustion engine by operating the operation unit of the internal combustion engine.
The acquisition process includes a process of acquiring the adjustment variable.
The calculation process is any one of claims 1 to 3, which is a process of calculating the imbalance variable based on the output of the mapping, which further includes the adjustment variable acquired by the acquisition process in the input to the mapping. The imbalance detector according to the section. The input of the map includes a drive system state variable which is a variable indicating the state of the drive system device connected to the crankshaft.
The acquisition process includes a process of acquiring the drive system state variable.
The calculation process is any of claims 1 to 4, which is a process of calculating the imbalance variable based on the output of the map in which the drive system state variable acquired by the acquisition process is further included in the input to the map. The imbalance detection device according to item 1. The input of the map includes a road surface state variable which is a variable indicating the state of the road surface on which the vehicle on which the internal combustion engine is mounted is traveling.
The acquisition process includes a process of acquiring the road surface state variable.
The calculation process is any one of claims 1 to 5, which is a process of calculating the imbalance variable based on the output of the map in which the road surface state variable acquired by the acquisition process is further included in the input to the map. The imbalance detection device according to item 1. The rotation waveform variable is configured as a variable indicating the difference between the instantaneous velocity variables themselves by the instantaneous velocity variables themselves at each of the plurality of second intervals.
The acquisition process includes a process of acquiring the instantaneous velocity variable at each of the plurality of second intervals as the rotation waveform variable.
The calculation process is any one of claims 1 to 6, which is a process of calculating the imbalance variable by inputting the instantaneous velocity variable at each of the plurality of second intervals as the rotation waveform variable to the mapping. The imbalance detector according to. The storage device includes a plurality of types of mapping data.
The imbalance detection device according to any one of claims 1 to 7, wherein the calculation process includes a selection process for selecting the map data to be used for calculating the imbalance variable from the plurality of types of map data. .. The internal combustion engine has a canister that collects fuel vapor in a fuel tank that stores fuel injected from a fuel injection valve, a purge passage that connects the canister and an intake passage of the internal combustion engine, and the purge passage. An adjusting device for adjusting the flow rate of fuel steam flowing from the canister to the intake passage through the canister is provided.
The calculation process includes a process of calculating the imbalance variable by using the data acquired by the acquisition process as an input to the map when the flow rate of the fuel steam is zero, and a process in which the flow rate of the fuel steam is more than zero. It includes both a process of calculating an imbalance variable by using the data acquired by the acquisition process when it is large as an input to the map.
In the coping process, when the flow rate of the fuel steam is larger than zero, the imbalance calculated by inputting the data acquired by the acquisition process to the map when the flow rate of the fuel steam is larger than zero. The imbalance detection device according to any one of claims 1 to 8, which is a process of operating the predetermined hardware based on a variable. The internal combustion engine is applied to an internal combustion engine including an EGR passage connecting an exhaust passage and an intake passage, and an EGR valve for adjusting the flow rate of exhaust gas flowing from the exhaust passage to the intake passage through the EGR passage. ,
The calculation process includes a process of calculating the imbalance variable by using the data acquired by the acquisition process as an input to the map when the flow rate of the exhaust flowing into the intake passage is zero, and a process of calculating the imbalance variable to the intake passage. And the process of calculating the imbalance variable by using the data acquired by the acquisition process as the input to the map when the flow rate of the inflowing exhaust is larger than zero.
In the coping process, when the flow rate of the exhaust gas flowing into the intake passage is larger than zero, and when the flow rate of the exhaust gas flowing into the intake passage is larger than zero, the data acquired by the acquisition process is mapped. The imbalance detection device according to any one of claims 1 to 8, which is a process of operating the predetermined hardware based on the imbalance variable calculated as an input to. In the coping process, when the degree of variation in the air-fuel ratio is large, the operation unit for controlling the combustion of the air-fuel mixture in the combustion chamber of the internal combustion engine as the predetermined hardware is operated according to the imbalance variable. The imbalance detection device according to any one of claims 1 to 10, which includes an operation process to be performed. The imbalance detection device according to claim 11, wherein the operation process includes a process of operating a fuel injection valve as the operation unit for supplying fuel to each of a plurality of cylinders. The executing device and the storage device according to any one of claims 1 to 12 are provided.
The execution device includes a first execution device and a second execution device.
The first execution device is mounted on the vehicle and transmits the acquisition process, the vehicle-side transmission process for transmitting the data acquired by the acquisition process to the outside of the vehicle, and the signal based on the calculation result of the calculation process. Execute the receiving process on the vehicle side to receive and the above-mentioned coping process,
The second execution device is arranged outside the vehicle and receives the data transmitted by the vehicle-side transmission process, the external reception process, the calculation process, and a signal based on the calculation result of the calculation process. An imbalance detection system that executes an external transmission process that transmits data to the vehicle. A data analysis device including the second execution device and the storage device according to claim 13. A control device for an internal combustion engine including the first execution device according to claim 13.