阅读说明：本技术 不均检测装置、不均检测系统、不均检测方法、数据解析装置、及内燃机的控制装置 (Unevenness detection device, unevenness detection system, unevenness detection method, data analysis device, and control device for internal combustion engine ) 是由 武藤晴文 片山章弘 森厚平 桥本洋介 于 2020-02-12 设计创作，主要内容包括：提供一种不均检测装置、不均检测系统、不均检测方法、数据解析装置、及内燃机的控制装置。在取得处理中,取得基于检测曲轴的旋转行为的传感器的检测值的旋转波形变量、和多个第1间隔各自中的空燃比检测变量。在算出处理中,基于以通过取得处理取得的值为输入的映射的输出算出不均变量。不均变量表示内燃机的空燃比的偏差程度。旋转波形变量表示多个第2间隔各自中的作为与曲轴的转速相应的变量的瞬时速度变量彼此之间的差异。(Provided are an unevenness detecting device, an unevenness detecting system, an unevenness detecting method, a data analyzing device, and a control device for an internal combustion engine. In the acquisition process, a rotation waveform variable based on a detection value of a sensor that detects a rotation behavior of the crankshaft and an air-fuel ratio detection variable at each of the plurality of 1 st intervals are acquired. In the calculation process, the variation is calculated based on the output of the map having the value obtained by the acquisition process as an input. The variation variable indicates the degree of deviation of the air-fuel ratio of the internal combustion engine. The rotational waveform amount indicates a difference between instantaneous speed variables, which are variables corresponding to the rotational speed of the crankshaft, in each of the plurality of 2 nd intervals.)

1. A device for detecting unevenness of a substrate is provided,

the unevenness detecting device is applied to a multi-cylinder internal combustion engine,

the unevenness detecting device includes a storage device and an executing device,

the storage device is configured to store map data that is data of a predetermined map that takes as input a rotation waveform variable and an air-fuel ratio detection variable that is a variable corresponding to an output of an air-fuel ratio sensor in each of a plurality of 1 st intervals, the map output being a variation variable that is a variable indicating a degree of variation in an air-fuel ratio of the internal combustion engine,

the execution device is configured to execute:

an acquisition process of acquiring the rotation waveform variable based on a detection value of a sensor that detects a rotation behavior of a crankshaft and the air-fuel ratio detection variable at each of a plurality of 1 st intervals;

a calculation process of calculating the uneven variable based on an output of the map having the value obtained by the acquisition process as an input; and

a handling process for handling a case where the degree of deviation of the air-fuel ratio is large by operating predetermined hardware based on a calculation result of the calculation process,

the rotation waveform variable is a variable representing a difference between instantaneous speed variables as variables corresponding to the rotational speed of the crankshaft in each of the plurality of 2 nd intervals,

the 1 st interval and the 2 nd interval are both angular intervals of the crankshaft smaller than an occurrence interval of compression top dead center,

the rotation waveform variable and the plurality of air-fuel ratio detection variables, both of which are input to the map, are time-series data in a predetermined angular interval larger than the appearance interval.

2. The unevenness detecting device according to claim 1,

the input of the map includes a 0.5 th order component of a rotational frequency of the crankshaft based on a detection value of a sensor that detects a rotational behavior of the crankshaft,

the acquiring process includes a process of acquiring a 0.5-order component variable that is a variable defining the 0.5-order component,

the calculation process is a process of calculating the uneven variable based on an output of the map in which the 0.5 th-order component variable acquired by the acquisition process is also included in an input to the map.

3. The unevenness detecting device according to claim 1 or 2,

an operation point variable that is a variable that defines an operation point of the internal combustion engine is included in the input of the map,

the acquisition process includes a process of acquiring the action point variable,

the calculation process is a process of calculating the variation variable based on an output of the map, the output of the map further including the operating point variable acquired by the acquisition process in an input to the map.

4. The unevenness detecting device according to any one of claims 1 to 3,

an adjustment variable that is a variable for adjusting a combustion speed of an air-fuel mixture in a combustion chamber of the internal combustion engine by an operation of an operation portion of the internal combustion engine is included in an input of the map,

the acquisition process includes a process of acquiring the adjustment variable,

the calculation process is a process of calculating the uneven variable based on an output of the map in which the adjustment variable acquired by the acquisition process is also included in an input to the map.

5. The unevenness detecting device according to any one of claims 1 to 4,

a drive system state variable that is a variable representing a state of a drive system device coupled to the crankshaft is included in an input of the map,

the acquiring process includes a process of acquiring the drive system state variable,

the calculation process is a process of calculating the variation amount based on an output of the map in which the drive system state variable acquired by the acquisition process is also included in an input to the map.

6. The unevenness detecting device according to any one of claims 1 to 5,

the map includes, as inputs to the map, a road surface state variable that is a variable representing a state of a road surface on which a vehicle mounted with the internal combustion engine is traveling,

the acquisition process includes a process of acquiring the road surface state variable,

the calculation process is a process of calculating the unevenness variable based on an output of the map in which the road surface state variable acquired by the acquisition process is also included in an input to the map.

7. The unevenness detecting device according to any one of claims 1 to 6,

the rotation waveform variable is constituted as follows: the difference between the above instantaneous speed variables from each other is represented by the instantaneous speed variable itself in each of a plurality of the 2 nd intervals,

the acquisition processing includes processing of acquiring the instantaneous speed variable in each of a plurality of the 2 nd intervals as the rotation waveform variable,

the calculation process is a process of calculating the variation by inputting the instantaneous speed variable in each of the plurality of 2 nd intervals as the rotation waveform variable to the map.

8. The unevenness detecting device according to any one of claims 1 to 7,

the storage device is provided with a plurality of the mapping data,

the calculation process includes a selection process of selecting the mapping data for calculating the variation in unevenness from the plurality of types of mapping data.

9. The unevenness detecting device according to any one of claims 1 to 8,

the internal combustion engine is provided with:

a canister that traps fuel vapor in a fuel tank that stores fuel injected from a fuel injection valve;

a purge passage interconnecting the canister and an intake passage of the internal combustion engine; and

an adjustment device that adjusts a flow rate of fuel vapor flowing from the canister into the intake passage via the purge passage,

the calculation process includes the following processes:

calculating the variation amount by using data acquired by the acquisition process when the flow rate of the fuel vapor is zero as an input to the map; and

the uneven distribution variable is calculated by using data obtained by the obtaining process when the flow rate of the fuel vapor is larger than zero as an input to the map,

the coping process is a process of operating the predetermined hardware based on the variation calculated by using, as an input to the map, data acquired by the acquiring process when the flow rate of the fuel vapor is greater than zero, when the flow rate of the fuel vapor is greater than zero.

10. The unevenness detecting device according to any one of claims 1 to 8,

the internal combustion engine is provided with:

a canister that traps fuel vapor in a fuel tank that stores fuel injected from a fuel injection valve;

a purge passage interconnecting the canister and an intake passage of the internal combustion engine; and

an adjustment device that adjusts a flow rate of fuel vapor flowing from the canister into the intake passage via the purge passage,

the calculation process includes the following processes:

calculating the non-uniformity variable by using data acquired by the acquisition process when the flow rate of the fuel vapor is zero as an input to the map, and calculating and storing a 1 st learning value from the non-uniformity variable; and

calculating the variation amount by using data obtained by the obtaining process when the flow rate of the fuel vapor is larger than zero as an input to the map, and calculating and storing a 2 nd learning value from the variation amount,

the coping process is a process of operating the predetermined hardware by selectively using the 1 st learning value of the stored 1 st learning value and the 2 nd learning value in a case where the flow rate of the fuel vapor is zero.

11. The unevenness detecting device according to any one of claims 1 to 8,

the internal combustion engine is provided with:

an EGR passage that connects the exhaust passage and the intake passage to each other; and

an EGR valve that adjusts a flow rate of exhaust gas flowing from the exhaust passage into the intake passage via the EGR passage,

the calculation process includes the following processes:

calculating the variation amount by using data obtained by the obtaining process as an input to the map when a flow rate of the exhaust gas flowing into the intake passage is zero; and

the amount of variation is calculated by using data obtained by the acquisition process as input to the map when the flow rate of the exhaust gas flowing into the intake passage is greater than zero,

the coping process is a process of operating the predetermined hardware based on the imbalance calculated by using, as an input to the map, data obtained by the obtaining process when the flow rate of the exhaust gas flowing into the intake passage is larger than zero, when the flow rate of the exhaust gas flowing into the intake passage is larger than zero.

12. The unevenness detecting device according to any one of claims 1 to 8,

the internal combustion engine is provided with:

an EGR passage that connects the exhaust passage and the intake passage to each other; and

an EGR valve that adjusts a flow rate of exhaust gas flowing from the exhaust passage into the intake passage via the EGR passage,

the calculation process includes the following processes:

calculating the non-uniform variable by using data obtained by the obtaining process as an input to the map when a flow rate of the exhaust gas flowing into the intake passage is zero, and calculating and storing a 1 st learning value from the non-uniform variable; and

calculating an uneven variable by using data obtained by the obtaining process when the flow rate of the exhaust gas flowing into the intake passage is larger than zero as an input to the map, and calculating and storing a 2 nd learned value from the uneven variable,

the coping process is a process of operating the predetermined hardware by selectively using the 1 st learned value of the stored 1 st and 2 nd learned values in a case where a flow rate of exhaust gas flowing into the intake passage is zero.

13. The unevenness detecting device according to any one of claims 1 to 12,

the predetermined hardware includes a combustion operation portion for controlling combustion of an air-fuel mixture in a combustion chamber of the internal combustion engine,

the coping process includes an operation process of operating the combustion operation section in accordance with the variation in the air-fuel ratio when the degree of deviation is large.

14. The unevenness detecting device according to claim 13,

the operation processing includes processing for operating fuel injection valves as the combustion operation portions for supplying fuel into a plurality of cylinders, respectively.

15. A system for detecting unevenness of a substrate is provided,

the execution device and the storage device according to any one of claims 1 to 14,

the executing device comprises a 1 st executing device and a 2 nd executing device,

the 1 st execution device is mounted on a vehicle and configured to execute:

the acquisition process;

a vehicle-side transmission process of transmitting the data acquired by the acquisition process to the outside of the vehicle;

a vehicle-side reception process of receiving a signal based on a calculation result of the calculation process; and

the coping process is carried out by performing a coping process,

the 2 nd execution device is disposed outside the vehicle and configured to execute:

an outside-side reception process of receiving data transmitted by the vehicle-side transmission process;

the calculation processing; and

and an external transmission process of transmitting a signal based on a calculation result of the calculation process to the vehicle.

16. A data analysis device is provided, which is capable of analyzing data,

the 2 nd execution device and the storage device according to claim 15 are provided.

17. A control device for an internal combustion engine,

the apparatus according to claim 15 is provided with the 1 st actuator.

18. A method for detecting unevenness of a substrate is provided,

the unevenness detecting method is applied to a multi-cylinder internal combustion engine,

the unevenness detecting method includes:

storing, by a storage device, map data that is data of a predetermined map that takes as input a rotation waveform variable and an air-fuel ratio detection variable that is a variable corresponding to an output of an air-fuel ratio sensor in each of a plurality of 1 st intervals, the map output being a variation variable that is a variable indicating a degree of variation in an air-fuel ratio of the internal combustion engine;

acquiring, by an actuator, the rotation waveform variable based on a detection value of a sensor that detects a rotation behavior of a crankshaft and the air-fuel ratio detection variable at each of a plurality of 1 st intervals;

calculating the variation variable based on an output of the map having the rotation waveform variable and the air-fuel ratio detection variable as inputs; and

operating predetermined hardware based on the variation amount to cope with a case where the degree of deviation of the air-fuel ratio is large,

the rotation waveform variable is a variable representing a difference between instantaneous speed variables as variables corresponding to the rotational speed of the crankshaft in each of the plurality of 2 nd intervals,

the 1 st interval and the 2 nd interval are both angular intervals of the crankshaft smaller than an occurrence interval of compression top dead center,

the rotation waveform variable and the plurality of air-fuel ratio detection variables, both of which are input to the map, are time-series data in a predetermined angular interval larger than the appearance interval.

Technical Field

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

Background

For example, japanese patent application laid-open No. 2013-194685 describes a device for detecting unevenness that is a deviation between actual air-fuel ratios when fuel injection valves are operated to control the air-fuel ratios of the mixture in each of a plurality of cylinders to be equal to each other. The apparatus determines a cylinder whose air-fuel ratio deviates from an expected value based on an air-fuel ratio estimation model using a detection value of the air-fuel ratio sensor as an input, in addition to a difference value of the detection value.

The unevenness includes a phenomenon of deviation from an intended air-fuel ratio to a rich side and a phenomenon of deviation to a lean side. The inventors have also found that the behavior of the detection value of the air-fuel ratio sensor may be approximated between the variation of the deviation to the rich side and the variation of the deviation to the lean side, and in this case, it may be difficult to determine which cylinder has the air-fuel ratio deviated.

Disclosure of Invention

Examples of the present disclosure are described below.

An unevenness detecting apparatus applied to a multi-cylinder internal combustion engine, the unevenness detecting apparatus including a storage device configured to store map data as data of a predetermined map having a rotation waveform variable and an air-fuel ratio detection variable as inputs, the air-fuel ratio detection variable being a variable corresponding to an output of an air-fuel ratio sensor in each of a plurality of 1 st intervals, and an execution device configured to execute: an acquisition process of acquiring the rotation waveform variable based on a detection value of a sensor that detects a rotation behavior of a crankshaft and the air-fuel ratio detection variable at each of a plurality of 1 st intervals; a calculation process of calculating the uneven variable based on an output of the map having the value obtained by the acquisition process as an input; and a coping process for coping with a case where a degree of deviation of the air-fuel ratio is large by operating predetermined hardware based on a calculation result of the calculation process, wherein the rotation waveform variable is a variable indicating a difference between instantaneous speed variables as variables corresponding to a rotation speed of a crankshaft among a plurality of 2 nd intervals, the 1 st interval and the 2 nd interval are both angular intervals of the crankshaft smaller than an appearance interval of a compression top dead center, and the rotation waveform variable and the plurality of air-fuel ratio detection variables, both of which are input to the map, are 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 in each of the plurality of 1 st intervals. The rotation waveform variable is a variable obtained by quantifying a rotation behavior generated by the torque generated in the combustion chamber of each cylinder. Therefore, the rotation waveform variable has sensitivity to a difference in the air-fuel ratio of the mixture. Therefore, even in a case where it is difficult to determine which cylinder's air-fuel ratio is deviated from the air-fuel ratio detection variable in each of the plurality of 1 st intervals, for example, it is possible to determine which cylinder's air-fuel ratio is deviated by considering the rotation waveform variable.

Example 2 the unevenness detecting apparatus according to example 1, wherein the input of the map includes a 0.5-order (japanese: 0.5-order) component of the rotational frequency of the crankshaft based on a detection value of a sensor that detects the rotational behavior of the crankshaft, the acquiring process includes a process of acquiring a 0.5-order component variable that is a variable defining the 0.5-order component, and the calculating process is a process of calculating the unevenness variable based on an output of the map that further includes the 0.5-order component variable acquired by the acquiring process in the input to the map.

It is considered that a linear relationship is established between the irregularity and the 0.5-order amplitude, which is the magnitude of the 0.5-order component of the rotation frequency. Further, the inventors found that the 0.5 order component becomes particularly large in the amplitude of the rotation frequency in the case where the unevenness exists. This is considered to be because: when the imbalance occurs in any of the plurality of cylinders, the deviation of the generated torque occurs once in 1 combustion cycle. In the above configuration, in view of this point, by including the 0.5-order component variable in the input to the map, the variation variable can be calculated with higher accuracy.

Example 3 the unevenness detecting apparatus according to the above-described example 1 or 2, wherein an input of the map includes an operation point variable that is a variable that defines an operation point of the internal combustion engine, the acquiring process includes a process of acquiring the operation point variable, and the calculating process is a process of calculating the unevenness variable based on an output of the map that further includes the operation point variable acquired by the acquiring process in the input to the map.

The control of the internal combustion engine tends to be performed in accordance with the operating point of the internal combustion engine. Therefore, the rotation 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, it is possible to calculate the variation variable while reflecting the fact that the rotational behavior of the crankshaft differs depending on the operating point.

Example 4 the unevenness detecting apparatus according to any one of examples 1 to 3, wherein the input of the map includes an adjustment variable that is a variable for adjusting a combustion speed of an air-fuel mixture in a combustion chamber of the internal combustion engine by an operation of an operating portion of the internal combustion engine, and the acquiring process includes a process of acquiring the adjustment variable, and the calculating process is a process of calculating the unevenness variable based on an output of the map that further includes the adjustment variable acquired by the acquiring process in the input to the map.

The rotational behavior of the crankshaft varies according to the combustion speed of the air-fuel mixture. Therefore, in the above configuration, the adjustment variable for adjusting the combustion speed is included in the input of the map. This makes it possible to calculate the variation while reflecting the fact that the rotation behavior of the crankshaft changes according to the combustion speed.

Example 5 the unevenness detecting apparatus according to any one of examples 1 to 4, wherein the input of the map includes a drive system state variable that is a variable indicating a state of a drive system apparatus connected to the crankshaft, and the acquiring process includes a process of acquiring the drive system state variable, and the calculating process is a process of calculating the unevenness variable based on an output of the map that further includes the drive system state variable acquired by the acquiring process in the input to the map.

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 input of the map, it is possible to calculate the variation variable while reflecting the fact that the rotation behavior of the crankshaft differs depending on the state of the drive system device.

Example 6 the unevenness detecting apparatus according to any one of examples 1 to 5, wherein the map input includes a road surface state variable that is a variable indicating a state of a road surface on which a vehicle on which the internal combustion engine is mounted is traveling, and the acquiring process includes a process of acquiring the road surface state variable, and the calculating process is a process of calculating the unevenness variable based on an output of the map that further includes the road surface state variable acquired by the acquiring process in the map input.

For example, when there is a rough road surface, vibration is generated in the vehicle, and the vibration is transmitted to the crankshaft. In this way, the state of the road surface affects the rotation behavior of the crankshaft. Therefore, in the above configuration, by including the road surface state variable in the input of the map, it is possible to calculate the unevenness variable while reflecting a case where the rotational behavior of the crankshaft changes depending on the state of the road surface.

Example 7. the unevenness detecting apparatus according to any of examples 1 to 6, wherein the rotational waveform variable is configured as follows: the difference between the instantaneous speed variables is expressed by the instantaneous speed variable itself in each of the plurality of 2 nd intervals, the acquiring process includes a process of acquiring the instantaneous speed variable in each of the plurality of 2 nd intervals as the rotation waveform variable, and the calculating process is a process of calculating the variation variable by inputting the instantaneous speed variable in each of the plurality of 2 nd intervals as the rotation waveform variable to the map.

In the case of calculating the difference or ratio between the instantaneous speed variables, it is necessary to separately apply which difference or ratio between the instantaneous speed variables in the 2 nd interval should be calculated. In this regard, in the above configuration, a plurality of instantaneous speed variables are used as the rotation waveform variables. Thus, for example, the amount of work to be applied to the rotational waveform variable can be reduced as compared with a case where the difference or ratio between the instantaneous speed variables is calculated in advance and input to the map.

Example 8 the unevenness detecting apparatus according to any of examples 1 to 7, wherein the storage device includes a plurality of types of the mapping data, and the calculating process includes a selecting process of selecting the mapping data for calculating the unevenness amount from the plurality of types of the mapping data.

For example, if a map is configured that can output an uneven variable with high accuracy in any situation, the structure of the map is likely to be complicated. Therefore, a plurality of kinds of mapping data are set in the above configuration. This makes it possible to select an appropriate map according to the situation. In this case, the respective configurations of the plurality of kinds of maps are easily simplified, as compared with a case where all situations are dealt with by a single map, for example.

Example 9 the unevenness detecting apparatus according to any one of examples 1 to 8, the internal combustion engine including: a canister that traps fuel vapor in a fuel tank that stores fuel injected from a fuel injection valve; a purge passage interconnecting the canister and an intake passage of the internal combustion engine; and an adjustment device that adjusts a flow rate of the fuel vapor flowing from the canister into the intake passage via the purge passage, wherein the calculation process includes both of: calculating the variation amount by using data acquired by the acquisition process when the flow rate of the fuel vapor is zero as an input to the map; and a process of calculating the variation by using, as an input to the map, data acquired by the acquisition process when the flow rate of the fuel vapor is greater than zero, wherein the coping process is a process of operating the predetermined hardware based on the variation calculated by using, as an input to the map, data acquired by the acquisition process when the flow rate of the fuel vapor is greater than zero.

Example 10. the unevenness detecting apparatus according to any of the above examples 1 to 8,

the internal combustion engine is provided with:

a canister that traps fuel vapor in a fuel tank that stores fuel injected from a fuel injection valve;

a purge passage interconnecting the canister and an intake passage of the internal combustion engine; and

an adjustment device that adjusts a flow rate of fuel vapor flowing from the canister into the intake passage via the purge passage,

the calculation process includes the following processes:

calculating the non-uniformity variable by using data acquired by the acquisition process when the flow rate of the fuel vapor is zero as an input to the map, and calculating and storing a 1 st learning value from the non-uniformity variable; and

calculating the variation amount by using data obtained by the obtaining process when the flow rate of the fuel vapor is larger than zero as an input to the map, and calculating and storing a 2 nd learning value from the variation amount,

the coping process is a process of operating the predetermined hardware by selectively using the 1 st learning value of the stored 1 st learning value and the 2 nd learning value in a case where the flow rate of the fuel vapor is zero.

When the flow rate of the fuel vapor is larger than zero, the fuel vapor flows from the canister into the intake passage and flows into each cylinder. However, the inflow amount of fuel vapor to each cylinder varies. The variation cannot be expressed by a variation obtained by quantifying the variation due to individual differences of the fuel injection valves, aging, and the like. Therefore, in the above configuration, the variation according to the presence or absence of the influence of the fuel vapor is individually calculated. This makes it possible to grasp the variation due to individual differences of the fuel injection valves, the variation due to aging, and the variation due to the fuel vapor, and to cope with the situation.

Example 11 the unevenness detecting apparatus according to any one of the above examples 1 to 8, the internal combustion engine including: an EGR passage connecting the exhaust passage and the intake passage; and an EGR valve that adjusts a flow rate of the exhaust gas flowing from the exhaust passage into the intake passage via the EGR passage, wherein the calculation process includes both of: calculating the variation amount by using data obtained by the obtaining process as an input to the map when a flow rate of the exhaust gas flowing into the intake passage is zero; and a process of calculating an uneven variable by using, as an input to the map, data acquired by the acquisition process when a flow rate of the exhaust gas flowing into the intake passage is greater than zero, wherein the coping process is a process of operating the predetermined hardware based on the uneven variable calculated by using, as an input to the map, data acquired by the acquisition process when the flow rate of the exhaust gas flowing into the intake passage is greater than zero.

Example 12. the unevenness detecting apparatus according to any of the above examples 1 to 8,

the internal combustion engine is provided with:

an EGR passage that connects the exhaust passage and the intake passage to each other; and

an EGR valve that adjusts a flow rate of exhaust gas flowing from the exhaust passage into the intake passage via the EGR passage,

the calculation process includes the following processes:

calculating the non-uniform variable by using data obtained by the obtaining process as an input to the map when a flow rate of the exhaust gas flowing into the intake passage is zero, and calculating and storing a 1 st learning value from the non-uniform variable; and

calculating an uneven variable by using data obtained by the obtaining process when the flow rate of the exhaust gas flowing into the intake passage is larger than zero as an input to the map, and calculating and storing a 2 nd learned value from the uneven variable,

the coping process is a process of operating the predetermined hardware by selectively using the 1 st learned value of the stored 1 st and 2 nd learned values in a case where a flow rate of exhaust gas flowing into the intake passage is zero.

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 into the intake passage via the EGR passage and flows into each cylinder. However, the inflow amount of exhaust gas into each cylinder varies. The variation in the combustion state of the air-fuel mixture in each cylinder due to the variation cannot be expressed by an uneven variable obtained by quantifying the variation due to individual differences of the fuel injection valves, aging, and the like. Therefore, in the above configuration, the variation amount corresponding to the presence or absence of the influence of the exhaust gas flowing into the intake passage is calculated. This makes it possible to grasp variations due to individual differences and aging of the fuel injection valves and variations due to exhaust gas flowing into the intake passage, and to cope with the situation.

Example 13 the unevenness detecting apparatus according to any of the above examples 1 to 12, wherein the predetermined hardware includes a combustion operation unit for controlling combustion of an air-fuel mixture in a combustion chamber of the internal combustion engine, and the coping process includes an operation process of operating the combustion operation unit based on the variation when a degree of variation in the air-fuel ratio is large.

In the above configuration, the operation portion for controlling combustion of the air-fuel mixture is operated in accordance with the variation amount. This can improve deterioration of the combustion state caused by a large variation in the air-fuel ratio.

Example 14 the unevenness detecting apparatus according to the above example 13, wherein the operation processing includes processing for operating fuel injection valves as the operation portions for supplying fuel to the plurality of cylinders, respectively.

In the above configuration, the variation in the air-fuel ratio of the air-fuel mixture in each cylinder can be reduced by correcting the injection amount based on the variation.

An embodiment 15 provides a non-uniformity detection system including the actuator and the storage device of any one of embodiments 1 to 14, wherein the actuator includes a 1 st actuator and a 2 nd actuator, and the 1 st actuator is mounted on a vehicle and configured to execute: the acquisition process; a vehicle-side transmission process of transmitting the data acquired by the acquisition process to the outside of the vehicle; a vehicle-side reception process of receiving a signal based on a calculation result of the calculation process; and the coping process, the 2 nd execution device being disposed outside the vehicle and configured to execute: an outside-side reception process of receiving data transmitted by the vehicle-side transmission process; the calculation processing; and an external transmission process of transmitting a signal based on a calculation result of the calculation process to the vehicle.

In the above configuration, the calculation process is executed outside the vehicle. This reduces the computational load on the in-vehicle device.

Example 16a data analysis device comprising the 2 nd execution device and the storage device according to example 15.

Example 17 a control device for an internal combustion engine, including the 1 st actuator described in example 15.

Example 18A system for detecting unevenness, comprising a 1 st execution device, a 2 nd execution device, and a storage device,

the storage device is configured to store map data that is data of a predetermined map that takes as input a rotation waveform variable and an air-fuel ratio detection variable that is a variable corresponding to an output of the air-fuel ratio sensor in each of the plurality of 1 st intervals, the map output being a variation variable that is a variable indicating a degree of variation in air-fuel ratio of the multi-cylinder internal combustion engine,

the 1 st execution device is mounted on a vehicle and configured to execute:

an acquisition process of acquiring the rotation waveform variable based on a detection value of a sensor that detects a rotation behavior of a crankshaft and the air-fuel ratio detection variable at each of a plurality of 1 st intervals;

a vehicle-side transmission process of transmitting the data acquired by the acquisition process to the outside of the vehicle;

a vehicle-side reception process of receiving a signal based on a calculation result of the calculation process; and

a handling process for handling a case where the degree of deviation of the air-fuel ratio is large by operating predetermined hardware based on a calculation result of the calculation process,

the 2 nd execution device is disposed outside the vehicle and configured to execute:

an outside-side reception process of receiving data transmitted by the vehicle-side transmission process;

a calculation process of calculating the uneven variable based on an output of the map having the value obtained by the acquisition process as an input; and

an external transmission process of transmitting a signal based on a calculation result of the calculation process to the vehicle,

the rotation waveform variable is a variable representing a difference between instantaneous speed variables as variables corresponding to the rotational speed of the crankshaft in each of the plurality of 2 nd intervals,

the 1 st interval and the 2 nd interval are both angular intervals of the crankshaft smaller than an occurrence interval of compression top dead center,

the rotation waveform variable and the plurality of air-fuel ratio detection variables, both of which are input to the map, are time-series data in a predetermined angular interval larger than the appearance interval.

Example 19. a method of detecting unevenness applied to a multi-cylinder internal combustion engine, the method including the process of any one of examples 1 to 18.

Example 20 a non-transitory computer-readable storage medium storing a program for causing an execution apparatus to execute the process described in any one of examples 1 to 18.

Example 21 in any one of examples 1 to 20, the variation variable is a variable indicating a degree of deviation between actual air-fuel ratios when a fuel injection valve of the internal combustion engine is operated to control air-fuel ratios of air-fuel mixtures in respective cylinders to be equal to each other.

Example 22 in the above example 5, the drive system state variable is a variable indicating a state of a transmission coupled to the crankshaft or a variable indicating a state of a lock-up clutch.

Example 23 in any of the above examples 1-22, the mapping data includes data learned by machine learning.

Drawings

Fig. 1 is a diagram showing the configuration of a control device and a vehicle drive system according to embodiment 1.

Fig. 2 is a block diagram showing a part of the processing executed by the control device according to the embodiment.

Fig. 3 is a flowchart showing the procedure of the unevenness detecting process according to this embodiment.

Fig. 4 is a flowchart showing the procedure of the processing for coping with unevenness according to the embodiment.

Fig. 5 is a diagram showing a system for generating map data according to this embodiment.

Fig. 6 is a flowchart showing the steps of the learning process of the map data according to the embodiment.

Fig. 7 is a time chart showing the influence of the unevenness on the instantaneous speed variable and the air-fuel ratio.

Fig. 8A is a graph showing the relationship between unevenness and amplitude of order 0.5.

Fig. 8B is a graph showing the relationship of the rotation order number and the amplitude.

Fig. 9 is a flowchart showing the procedure of the unevenness learning value calculation process according to embodiment 2.

Fig. 10 is a flowchart showing the procedure of the processing for coping with unevenness according to the embodiment.

Fig. 11 is a flowchart showing the procedure of the processing for coping with unevenness according to embodiment 3.

Fig. 12 is a flowchart showing the procedure of the map data selection processing according to embodiment 4.

Fig. 13 is a flowchart showing steps of the map data selection process according to embodiment 5.

Fig. 14 is a flowchart showing the procedure of the map data selection processing according to embodiment 6.

Fig. 15 is a flowchart showing the procedure of the map data selection processing according to embodiment 7.

Fig. 16 is a flowchart showing the procedure of the unevenness detecting process according to embodiment 8.

Fig. 17 is a diagram showing a configuration of the unevenness detecting system according to embodiment 9.

Parts (a) and (b) of fig. 18 are flowcharts showing steps of processing executed by the unevenness detecting system of fig. 17.

Detailed Description

< embodiment 1 >

Hereinafter, embodiment 1 relating to the unevenness detecting device will be described with reference to fig. 1 to 8B.

In an internal combustion engine 10 mounted on a vehicle VC shown in fig. 1, a throttle valve 14 is provided in an intake passage 12. The air taken in from the intake passage 12 flows into the combustion chamber 18 of each 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 and an ignition device 22 for generating spark discharge so as to be exposed to the combustion chamber 18. In the combustion chamber 18, a mixture of air and fuel is used for combustion, and energy generated by the combustion is extracted as rotational energy of the crankshaft 24. The air-fuel mixture used for combustion is discharged as exhaust gas to the exhaust passage 28 as the exhaust valve 26 is opened. A catalyst 30 having an oxygen storage capacity is provided in the exhaust passage 28. The exhaust passage 28 communicates with the intake passage 12 via an EGR passage 32. The EGR passage 32 is provided with an EGR valve 34 for adjusting the cross-sectional area of the passage.

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

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

An input shaft 66 of a transmission 64 can be coupled to the crankshaft 24 of the internal combustion engine 10 via a torque converter 60. The torque converter 60 includes a lock-up clutch 62, and the crankshaft 24 and the input shaft 66 are coupled by engaging the lock-up clutch 62. A drive wheel 69 is mechanically coupled to an output shaft 68 of the transmission 64. Further, in the present embodiment, the transmission 64 is a stepped transmission capable of changing the gear ratio from 1 st to 5 th.

A crankshaft rotor 50 is coupled to the crankshaft 24, and teeth 52 indicating each of a plurality of rotation angles of the crankshaft 24 are provided on the crankshaft rotor 50. In the present embodiment, 34 teeth 52 are illustrated. The crank rotor 50 is provided with teeth 52 at intervals of substantially 10 ° CA, but is provided with 1-point tooth-missing portion 54, which is a portion where the interval between adjacent teeth 52 is 30 ° CA. This is used to indicate a rotation angle that becomes a reference of the crankshaft 24.

The control device 70 controls the internal combustion engine 10, and operates operation portions of the internal combustion engine 10 such as 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 variable valve timing device 46 in order to control the torque, the exhaust gas component ratio, and the like, which are control amounts thereof. 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 variable intake-side valve timing device 46, respectively.

When controlling the control amount, the control device 70 refers to the intake air amount Ga detected by the air flow meter 80, the upstream side detection value Afu detected by the air-fuel ratio sensor 82 provided upstream of the catalyst 30, and the output signal Scr of the crank angle sensor 86, and the crank angle sensor 86 outputs pulses at angular intervals between the teeth 52 provided at intervals of 10 ° CA other than the teeth 54. The controller 70 refers to a water temperature THW, which is the temperature of the cooling water of the internal combustion engine 10 detected by the water temperature sensor 88, a shift position Sft of the transmission 64 detected by the shift position sensor 90, and an acceleration Dacc in the vertical direction of the vehicle VC detected by the acceleration sensor 92.

The control device 70 includes a CPU72, a ROM74, a storage device 76 as an electrically rewritable nonvolatile memory, and a peripheral circuit 77, and can communicate with each other via a local area network 78. The peripheral circuit 77 includes a circuit that generates a clock signal that defines an internal operation, a power supply circuit, a reset circuit, and the like.

The control device 70 executes the control of the above-described control amount by the CPU72 executing a program stored in the ROM 74.

A part of the processing realized by the CPU72 executing the program stored in the ROM74 is shown in fig. 2.

The ignition timing operation process M10 is a process of setting a basic value of the ignition timing based on the rotation speed NE and the charging efficiency η that define the operating point of the internal combustion engine 10, and outputting the operation signal MS3 to the ignition device 22 so as to obtain the corresponding ignition timing aig, thereby operating the ignition device 22. Here, the charging efficiency η is a parameter indicating the amount of air charged into the combustion chamber 18, and is calculated by the CPU72 based on the intake air amount Ga and the rotation speed NE. The rotation speed NE is calculated by the CPU72 based on the output signal Scr of the crank angle sensor 86. The rotation speed NE is an average value of rotation speeds when the crankshaft 24 rotates at an angular interval larger than the appearance interval of the compression top dead center (180 ° CA in the present embodiment). Preferably, the rotation speed NE is an average value of rotation speeds when the crankshaft 24 rotates by a rotation angle of 1 rotation or more of the crankshaft 24. The average value here is not limited to a simple average, and may be, for example, an exponential moving average process, and may be calculated using a plurality of sample values of the rotation speed at a rotation angle of 1 rotation or more, for example, at a minute rotation angle interval. However, the present invention is not limited to this, and may be calculated based on a single measured value of the time required to rotate the rotation angle by 1 or more turns.

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

The target clearance calculation process M14 is a process of calculating the target clearance 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 tank 40 into the intake passage 12 by the intake air amount Ga, and the target purge rate Rp is a target value of the purge rate in the control.

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

The base injection amount calculation process M18 is a process of calculating the base injection amount Qb based on the charging efficiency η. The basic injection amount Qb is a basic value of the amount of fuel for making the air-fuel ratio of the air-fuel mixture in the combustion chamber 18 the target air-fuel ratio. More specifically, when the filling efficiency η is expressed by a percentage, for example, the basic injection amount calculation process M18 may be a process of calculating the basic injection amount Qb by multiplying the filling efficiency η by the fuel amount QTH per 1% of the filling efficiency η for setting the air-fuel ratio to the target air-fuel ratio. The base injection amount Qb is an amount of fuel calculated based on the amount of air filled in the combustion chamber 18 in order to control the air-fuel ratio to the target air-fuel ratio. The target air-fuel ratio may be the stoichiometric air-fuel ratio.

The feedback process M20 is a process of calculating a feedback correction coefficient KAF obtained by adding "1" to the correction ratio. The correction ratio is an operation amount for feedback-controlling the upstream side detection value Afu as a feedback control amount to a target value Af. The feedback correction coefficient KAF is a correction coefficient of the base injection amount Qb. Here, when the correction ratio is "0", the correction ratio of the base injection amount Qb is zero. The feedback process M20 performs the increment correction of the base injection amount Qb when the correction ratio is larger than "0", and performs the decrement correction of the base injection amount Qb when the correction ratio is smaller than "0". In the present embodiment, the correction ratio is defined as the sum of the respective output values of the proportional element and the differential element, which have the difference between the target value Af and the upstream detection value Afu as input, and the sum of the output value of the integral element, which outputs the integrated value of the value corresponding to the difference.

The air-fuel ratio learning process M22 is a process of sequentially updating the air-fuel ratio learning value LAF so that the deviation of the correction ratio from "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 converges when the deviation amount of the correction ratio from "0" is equal to or less than a predetermined value.

In the coefficient addition process M24, the air-fuel ratio learning value LAF is added to the feedback correction coefficient KAF.

The clear density learning process M26 is a process of calculating the clear density learning value Lp based on the above correction ratio. The scavenging concentration learning value Lp is a value obtained by converting a correction rate for correcting the deviation of the base injection amount Qb to a scavenging rate of 1%. The correction ratio is used to correct a deviation of the base injection amount Qb from the injection amount required for controlling the fuel-air ratio to the target air-fuel ratio due to the inflow of the fuel vapor from the canister 40 into the combustion chamber 18. In the present embodiment, the factors that cause the feedback correction coefficient KAF to deviate from "1" when the target purge rate Rp is controlled to a value greater than "0" are all considered to be caused by the fuel vapor flowing from the canister 40 into the combustion chamber 18. That is, the correction ratio is regarded as a correction ratio for correcting a deviation of the base injection amount Qb from the injection amount necessary for controlling the fuel vapor to the target air-fuel ratio, which is caused by the inflow of the fuel vapor from the canister 40 into the intake passage 12. However, since the correction ratio depends on the clearance, in the present embodiment, the clearance concentration learning value Lp is set to an amount corresponding to the value "/Rp" of clearance per 1%. Specifically, the removal-concentration learning value Lp is set to an exponential moving average processing value of the removal rate "/Rp" per 1%. It is preferable to execute the purge concentration learning process M26 with the target purge rate Rp set to a value greater than zero on condition that it is determined that the air-fuel ratio learning value LAF converges.

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 density learning value Lp. The clear correction ratio Dp is a value equal to or smaller than zero.

The correction coefficient calculation process M30 is a process of adding the clear correction ratio Dp to the output value of the coefficient addition process M24.

The required injection amount calculation process M32 is a process of calculating the required injection amount Qd by multiplying the base injection amount Qb by the output value of the correction coefficient calculation process M30 to correct the base injection amount Qb.

The injection valve operation process M34 is a process of outputting an operation signal MS2 to the fuel injection valve 20 to operate the fuel injection valve 20 based on the required injection quantity Qd.

Next, a process of detecting variation in the actual air-fuel ratio among the cylinders when the fuel injection valves 20 in the respective cylinders #1 to #4 are operated based on the process of fig. 2 will be described.

The steps of the process related to the detection of the unevenness are shown in fig. 3. The processing shown in fig. 3 is realized by the CPU72 repeatedly executing the unevenness detection program 74a stored in the ROM74 shown in fig. 1, for example, at predetermined cycles. In the following, the step number of each process is represented by a numeral labeled with "S" in the top.

In the series of processing shown in fig. 3, the CPU72 first determines whether or not the execution condition of the unevenness detection processing 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 CPU72 acquires the minute rotation time T30(1), T30(2), …, T30(24), the upstream-side average value Afuave (1), Afuave (2), …, Afuave (24), the rotation speed NE, the charging efficiency η, and the amplitude Ampf/2 of 0.5 step (S12). Further, Ampf/2 denotes a reference numeral of 0.5 order amplitude. The minute rotation time T30 is calculated by the CPU72 counting the time required for the crankshaft 24 to rotate by 30 ° CA based on the output signal Scr of the crank angle sensor 86. Here, when the numbers in parentheses such as the minute rotation time T30(1) and T30(2) are different, the different rotation angle intervals in 720 ° CA, which is 1 combustion cycle, are indicated. That is, the minute rotation times T30(1) to T30(24) represent rotation times in each angular interval obtained by equally dividing the rotation angle region of 720 ° CA (3 rd interval) by 30 ° CA (4 th interval). That is, the minute rotation time T30 is an instantaneous speed parameter that is a parameter corresponding to the number of rotations of the crankshaft 24 at each of a plurality of angular intervals (30 ° CA, i.e., 4 th interval). Here, the 3 rd interval is a rotational angle interval of the crankshaft 24 and is an interval including compression top dead center, and the 4 th interval is an interval smaller than an occurrence interval of compression top dead center. The minute rotation time T30 constitutes time series data as an instantaneous speed parameter in each of a plurality of 4 th intervals that are consecutive and included in the 3 rd interval.

Specifically, the CPU72 counts the time during which the crankshaft 24 has rotated 30 ° CA based on the output signal Scr, and sets the time as the pre-filter processing time NF 30. Next, the CPU72 calculates the post-filter time AF30 by performing digital filter processing with the pre-filter time NF30 as an input. Then, the CPU72 normalizes (japanese: plus size) the post-filter time AF30 so that the difference between the maximum value (maximum value) and the minimum value (minimum value) of the post-filter time AF30 in a predetermined period (for example, 720 ° CA) becomes "1", thereby calculating the minute rotation time T30.

When m is 1 to 24, the upstream average value afuave (m) is an average value of the upstream detection values Afu at the same angular intervals of 30 ° CA as the above-described minute rotation times T30 (m).

The amplitude Ampf/2 of 0.5 th order is the intensity of the 0.5 th order component of the rotational frequency of the crankshaft 24, and is calculated by the CPU72 through the fourier transform of the time series data of the minute rotation time T30.

Next, the CPU72 substitutes the values obtained by the processing of S12 into the input variables x (1) to x (51) of the map of the output unevenness rate Riv (S14). More specifically, "m" is 1 to 24 ", the CPU72 substitutes the minute rotation time T30(m) for the input variable x (m), substitutes the upstream-side average value afuave (m) for the input variable (24+ m), substitutes the rotation speed NE for the input variable x (49), substitutes the filling efficiency η for the input variable x (50), and substitutes the amplitude Ampf/2 of 0.5 step for the input variable x (51).

In the present embodiment, the unevenness rate Riv is "0" in the cylinder in which the fuel of the target injection amount is injected, becomes a positive value when the actual injection amount is larger than the target injection amount, and becomes a negative value when the actual injection amount is smaller than the target injection amount. That is, the variation rate Riv is a variation variable that is a variable indicating the degree of variation between the actual air-fuel ratios when the fuel injection valves 20 are operated to control the air-fuel ratios of the air-fuel mixtures in the plurality of cylinders #1 to #4 to be equal to each other.

Next, the CPU72 inputs the input variables x (1) to x (51) into a map defined by the map data 76a stored in the storage device 76 shown in fig. 1, thereby calculating the respective rates of nonuniformity Riv (1) to Riv (4) of the cylinders # i (i: 1 to 4) (S16). The above-described unevenness ratios Riv (1) to Riv (4) can be grasped as variables indicating the degree of deviation of the air-fuel ratio of the internal combustion engine 10.

In the present embodiment, the map is composed of a neural network having 1 layer as an intermediate layer. The neural network includes an input-side coefficient wFjk (j is 0 to n, k is 0 to 51) and an activation function h (x). The input-side linear map is a linear map defined by input-side coefficients wFjk, and the activation function h (x) is an input-side nonlinear map in which outputs of the input-side linear map are each subjected to nonlinear transformation. In the present embodiment, a hyperbolic tangent function "tanh (x)" is exemplified as the activation function h (x). The neural network includes an output-side coefficient wSij (i is 1 to 4, j is 0 to n) and an activation function f (x). The output-side linear mapping is a linear mapping defined by an output-side coefficient wSij, and the activation function f (x) is an output-side nonlinear mapping in which outputs of the output-side linear mapping are each subjected to nonlinear conversion. In the present embodiment, a hyperbolic tangent function "tanh (x)" is exemplified as the activation function f (x). Further, the value n represents the dimension of the intermediate layer.

When the process at S16 is completed and the process at S10 is determined as no, the CPU72 once ends the series of processes shown in fig. 3.

Fig. 4 shows a procedure of processing using the above-described unevenness ratio riv (i). The processing shown in fig. 4 is realized by the CPU72 repeatedly executing the coping program 74b stored in the ROM74 shown in fig. 1, for example, every time the unevenness rate riv (i) is calculated.

In the series of processes shown in fig. 4, the CPU72 first updates the unevenness learning value liv (i) by an exponential moving average process that takes as input the unevenness rate riv (i) newly calculated by the process of fig. 3 (S20). That is, the CPU72 updates the unevenness learning value Liv by the sum of the value obtained by multiplying the unevenness learning value Liv (i) stored in the storage device 76 by the coefficient α and the value obtained by multiplying the unevenness ratio riv (i) by "1- α" (S20), for example. Further, "0 < α < 1".

Next, the CPU72 determines whether or not the unevenness learning value liv (i) is equal to or higher than the lean limit value LL and equal to or lower than the rich limit value LH (S22). When the CPU72 determines that the unevenness learning value liv (i) is smaller than the lean-side allowable limit value LL or that the unevenness learning value liv (i) is larger than the rich-side allowable limit value (S22: NO), the CPU72 operates the warning lamp 98 to execute a notification process to prompt the user to repair the internal combustion engine 10 or the like (S24).

On the other hand, when the CPU72 determines that the unevenness learning value liv (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 (YES in S22), and the processing of S24 is completed, the CPU72 corrects the required injection amount Qd (# i) for each cylinder (S26). That is, the CPU72 corrects the required injection amount Qd (# i) by adding a correction amount Δ Qd (liv (i)) corresponding to the learning irregularity value liv (i) to the required injection amount Qd (# i) for each cylinder. Here, the correction amount Δ Qd (liv (i)) becomes a negative value when the unevenness learning value liv (i) is larger than zero, and becomes a positive value when the unevenness learning value liv (i) is smaller than zero. When the unevenness learning value liv (i) is zero, the correction amount Δ Qd (liv (i)) is also zero.

Further, the CPU72, upon completion of the process of S26, temporarily ends the series of processes shown in fig. 4. In the present embodiment, when it is determined that the process at S10 is yes and the process at S12 is executed, the process at S26 is temporarily stopped.

Next, a method of generating the map data 76a will be described.

A system for generating mapping data 76a is shown in fig. 5.

As shown in fig. 5, in the present embodiment, a dynamometer (dynamometer)100 is mechanically coupled to the crankshaft 24 of the internal combustion engine 10 via a torque converter 60 and a transmission 64. Then, various state variables at the time of operating the internal combustion engine 10 are detected by the sensor group 102, and the detection results are input to the application device 104 as a computer that generates the map data 76 a. The sensor group 102 includes an air flow meter 80, an air-fuel ratio sensor 82, and a crank angle sensor 86 as sensors for detecting values for generating inputs to the map.

Fig. 6 shows the steps of the map data generation process. The processing shown in fig. 6 is performed by the applicable device 104. The processing shown in fig. 6 may be realized, for example, by providing the application device 104 with a CPU and a ROM, and executing a program stored in the ROM by the CPU.

In the series of processing shown in fig. 6, the application device 104 first acquires the same data as the data acquired in the processing of S12 as training data based on the detection result of the sensor group 102 (S30). In this process, a plurality of fuel injection valves 20 having an unevenness ratio Riv of various values different from zero and 3 fuel injection valves having an unevenness ratio of zero are prepared in advance based on the measurement of the fuel injection valves 20 alone. Then, in a state where 3 fuel injection valves 20 having an imbalance ratio of zero and 1 fuel injection valve 20 having an imbalance ratio different from zero are mounted on the internal combustion engine 10, the process of S30 is performed. The variation ratios Rivt of the mounted fuel injection valves are teacher data. Here, the index of the unevenness as the teacher data is Rivt in which t is added to Riv.

Next, the application device 104 substitutes training data other than the teacher data into the input variables x (1) to x (51) in the approach of the processing of S14 (S32). Then, the application device 104 calculates the unevenness ratios 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 CPU72 determines whether or not the number of samples of the unevenness rate riv (i) calculated by the processing of S34 is equal to or greater than a predetermined number (S36). Here, in order to be equal to or more than a predetermined value, the operating state of the internal combustion engine 10 is changed in a state where a plurality of fuel injection valves having different unevenness ratios Rivt from zero are mounted in each of the cylinders #1 to # 4. Thus, it is required to calculate the unevenness ratio Riv at various operating points defined by the rotation speed NE and the charging efficiency η.

If the application device 104 determines that the number of samples of the unevenness rate riv (i) is not equal to or greater than the predetermined value (no in S36), the process returns to S30. On the other hand, when the CPU72 determines that the number of samples of the unevenness rate riv (i) is equal to or greater than the predetermined number (yes in S36), the input-side coefficient wFjk and the output-side coefficient wSij are updated so that the sum of squares of the differences between the unevenness rate Rivt as the teacher data and each of the unevenness rates riv (i) calculated in S34 is minimized (S38). Then, the application device 104 stores the updated input-side coefficient wFjk and output-side coefficient wSij as the learned map data (S40).

Here, the operation and effect of the present embodiment will be described.

Fig. 7 shows the transition of the crank counter having a cycle of 360 ° CA, the upstream-side average value Afuave, and the minute rotation time T30, when the variation rate Riv (1) of the cylinder #1 is positive and the variation rate Riv (4) of the cylinder #4 is negative, respectively. As shown in fig. 7, there is no large difference in the phase of the upstream-side average value Afuave 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, and it is therefore difficult to determine what kind of abnormality has occurred in which cylinder. However, as is clear from the time series data of the minute rotation time T30, the portion enclosed by the chain line in fig. 7 shows that there is a clear difference between the two abnormalities.

Therefore, in the present embodiment, the CPU72 calculates the unevenness rate Riv using the time series data of the minute rotation time T30 and the time series data of the upstream-side average value Afuave. Thus, for example, even when it is difficult to distinguish which cylinder has such an abnormality based on the upstream side detection value Afu alone, the present embodiment can calculate the unevenness rates Riv (1) to Riv (4) of the respective cylinders.

In the present embodiment, instead of learning a map in which the variation rate Riv is calculated by randomly inputting a large amount of various variables of the internal combustion engine 10 by machine learning, for example, the variables to be input to the map are strictly selected based on the knowledge of the inventors who are skilled in the control of the internal combustion engine 10. Therefore, as compared with the case where the inventors' knowledge is not used, the number of layers in the intermediate layer of the neural network and the dimension of the input variable can be reduced, and the structure of the map for calculating the unevenness ratio riv (i) can be easily simplified.

According to the present embodiment described above, the following operational effects can be obtained.

(1) The rotational speed NE and the charging efficiency η, which are operating point variables that define the operating point of the internal combustion engine 10, are used as map inputs. The operation amounts of the operation portions of the internal combustion engine 10, such as the ignition device 22, the EGR valve 34, and the intake-side variable valve timing device 46, tend to be determined based on the operating point of the internal combustion engine 10. Therefore, the operating point variable is a variable including information on the operation amount of each operation unit. Therefore, by inputting the operating point variable as a map, the unevenness rate riv (i) can be calculated based on information on the operation amount of each operation unit. The unevenness ratio riv (i) can be calculated with higher accuracy.

(2) The upstream-side average value Afuave is included in the input of the map. Thus, more accurate information on the oxygen and unburned fuel flowing into the catalyst 30 can be obtained without increasing the number of data of the time series data, as compared with, for example, the case where the upstream side detection value Afu for each time interval of the time series data is used. The unevenness ratio riv (i) can be calculated with higher accuracy.

(3) By including the amplitude Ampf/2 of 0.5 order in the input to the map, the unevenness rate Riv can be calculated with higher accuracy. That is, as shown in fig. 8A, a linear relationship is established between the unevenness rate Riv and the amplitude Ampf/2 of 0.5 order. As illustrated in fig. 8B, when the amplitude of the rotational frequency of the crankshaft 24 is a cylinder in which the unevenness ratio Riv (1) is not zero, the 0.5-order component is particularly large, as in the case where the unevenness ratio Riv (1) is "1.15". This is considered to be because: when the unevenness rate riv (i) is different from zero in any of the cylinders #1 to #4, a deviation of the generated torque occurs once in 1 combustion cycle. In the present embodiment, the torque variation of the 720 ° CA cycle is taken as the amplitude Ampf/2 of 0.5 step, and the unevenness rate Riv can be calculated with higher accuracy.

< embodiment 2 >

Hereinafter, embodiment 2 will be described mainly focusing on differences from embodiment 1 with reference to fig. 9 and 10.

In the present embodiment, after the unevenness rate Riv is calculated once, the target clearance rate Rp is switched from zero to a value larger than zero, and at this time, the unevenness rate Riv is calculated again. Then, the EGR rate Regr is switched from zero to a value larger than zero, and the unevenness rate Riv is calculated again. Therefore, in the present embodiment, the execution conditions in the process of S10 in fig. 3 are set to three as follows: target clearance Rp is zero and EGR rate Regr is zero; when the target clearance Rp switches from zero to a value greater than zero; when the EGR rate Regr is switched from zero to a value larger than zero.

Fig. 9 shows a procedure of the learning value calculation process based on the unevenness Riv calculated in this manner. The processing shown in fig. 9 is realized by the CPU72 repeatedly executing the coping program 74b stored in the ROM74, for example, at predetermined cycles.

In the series of processing shown in fig. 9, the CPU72 first determines whether or not the unevenness rate Riv is newly calculated (S50). When determining that the unevenness rate Riv is newly calculated (yes in S50), the CPU72 determines whether or not the input variable for calculating the unevenness rate Riv is a variable sampled when the target purge rate Rp is zero and the EGR rate Regr is zero (S52). If the CPU72 determines yes in the process of S52, the same process as S20 is executed (S54).

On the other hand, if the CPU72 determines no in the process of S52, it determines whether or not the input variable for calculating the unevenness rate Riv is the variable sampled when the EGR rate Regr is zero (S56). When the CPU72 determines that the EGR rate Regr is zero (yes in S56), the CPU updates the learned value livp at purge (i) by exponential moving average processing with the currently calculated unevenness rate riv (i) as an input (S58).

On the other hand, when the CPU72 determines that the input variable for calculating the unevenness rate Riv is the variable sampled when the EGR rate Regr is greater than zero (S56: no), the CPU updates the EGR time learning value live (i) by exponential moving average processing with the unevenness rate Riv (i) calculated this time as an input (S60).

When the processing in S54, S58, and S60 is completed and the CPU72 determines no in S50, it temporarily ends the series of processing shown in fig. 9.

Fig. 10 shows the procedure of the process using the above-described learning value. The processing shown in fig. 10 is realized by the CPU72 repeatedly executing the coping program 74b stored in the ROM74 shown in fig. 1, for example, at predetermined cycles. In fig. 10, for convenience, the same step numbers are assigned to the processes corresponding to the process shown in fig. 4.

In the series of processing shown in fig. 10, when the CPU72 determines yes in S22 or the processing in S24 is completed, it determines whether the target purge rate Rp is zero and the EGR rate Regr is zero (S62). When the CPU72 determines that the target purge rate Rp is zero and the EGR rate Regr is zero (yes in S62), it executes the same processing as that of S26 (S64).

On the other hand, if the CPU72 determines no in the process of S62, it determines whether the EGR rate Regr is zero (S66). When the CPU72 determines that the EGR rate Regr is zero (yes in S66), it corrects the required injection amount Qd (# i) for each cylinder based on not only the unevenness learning value liv (i) but also the learning value livp (i) at the time of purge (S68). That is, the CPU72 corrects the required injection amount Qd (# i) by adding a correction amount Δ Qd (liv (i)) corresponding to the imbalance learning value liv (i) and a correction amount Δ Qdp (livp (i)) corresponding to the purge time learning value livp (i) to the required injection amount Qd (# i) for each cylinder.

On the other hand, when determining that the EGR rate Regr is greater than zero (no in S66), the CPU72 corrects the required injection amount Qd (# i) for each cylinder based on the unevenness learning value liv (i), the learning value livp (i) at the time of purge, and the learning value live (i) at the time of EGR (S70). That is, the CPU72 corrects the required injection amount Qd (# i) for each cylinder by the sum of the correction amount Δ Qd (liv (i)) corresponding to the unevenness learning value liv (i), the correction amount Δ Qd (livp (i)) corresponding to the learning value livp (i) at the time of purge, and the correction amount Δ Qde (live (i)) corresponding to the learning value live (i) at the time of EGR.

Further, the CPU72 once completes the processing of S64, S68, S70, and then temporarily ends the series of processing shown in fig. 10. In the present embodiment, when the process of S12 for calculating the unevenness rate Riv used in the process of S58 is executed, the correction amount Δ Qd (livp (i)) corresponding to the learning value livp (i) at the time of clearing is not used in the process of S68. When the process of S12 for calculating the unevenness Riv used in the process of S60 is executed, the correction amount Δ Qd (live (i)) corresponding to the EGR learning value live (i)) is not used in the process of S70.

In the present embodiment, it is preferable that the mapping data 76a is data that is learned as training data in each case of the case where the processing at S62 is determined to be yes, the case where the processing at S66 is determined to be yes, and the case where the processing at S66 is determined to be no. However, instead of the mapping data 76a being single data, the mapping data 76a may be different data between the case where the determination is yes in the process of S62, the case where the determination is yes in the process of S66, and the case where the determination is no in the process of S66. In this case, for each data, learning is performed using only the corresponding training data.

According to the present embodiment described above, in addition to the operational effects described in embodiment 1, the operational effects described below can be obtained.

(4) The unevenness rate Riv when the target purge rate Rp is larger than zero is calculated, and the required injection amount Qd is corrected based on the corresponding purge learning value livp (i). Thus, even when the scavenging ratio is larger than zero, variation in the air-fuel ratio between the cylinders #1 to #4 can be suppressed. That is, when the purge rate is larger than zero, the fuel vapor flows from the canister 40 into the intake passage 12 and flows into the cylinders #1 to #4, but the flow rate varies among the cylinders #1 to # 4. The variation cannot be expressed by the variation learning value liv (i) obtained by quantifying the variation due to the individual difference or the secular change of the fuel injection valve 20. In this regard, in the present embodiment, the deviation of the air-fuel ratio between the cylinders #1 to #4 can be suppressed by correcting the required injection amount Qd based on the purge time learning value livp (i).

(5) A variation rate Riv when the EGR rate Regr is larger than zero is calculated, and the required injection amount Qd is corrected based on an EGR time learning value live (i) corresponding to the variation rate Riv. Thus, even when the EGR rate Regr is larger than zero, variation in the air-fuel ratio among the cylinders #1 to #4 can be suppressed. That is, when the EGR rate Regr is larger than zero, the exhaust gas from the exhaust passage 28 flows into the intake passage 12 via the EGR passage 32 and flows into the cylinders #1 to #4, but the flow rate varies among the cylinders #1 to # 4. The variation cannot be expressed by the variation learning value liv (i) obtained by quantifying the variation due to the individual difference or the secular change of the fuel injection valve 20. In this regard, in the present embodiment, the deviation of the air-fuel ratio between the cylinders #1 to #4 can be suppressed by correcting the required injection amount Qd based on the EGR time learning value live (i).

< embodiment 3 >

Hereinafter, embodiment 3 will be described mainly focusing on differences from embodiment 2 with reference to fig. 11.

In the present embodiment, the ignition timing is operated when the unevenness is not sufficiently eliminated by the injection amount correction.

Fig. 11 shows a procedure of processing regarding the operation of the ignition timing according to the unevenness rate Riv. The processing shown in fig. 11 is realized by the CPU72 repeatedly executing the coping program 74b stored in the ROM74 shown in fig. 1 at predetermined cycles on the condition that, for example, the number of updates of the unevenness learning value liv (i) or the like becomes equal to or more than a predetermined number.

In the series of processing shown in fig. 11, the CPU72 first determines whether or not the absolute value of the unevenness rate riv (i) newly calculated after the processing of S64 in fig. 10 is performed is equal to or greater than the predetermined value Δ th when the target purge rate Rp is zero and the EGR rate Regr is zero (S72). This process is a process of determining whether or not the unevenness is not sufficiently eliminated by the correction amount Δ Qd (liv (i)) based on the unevenness learning value liv (i). When the CPU72 determines that the absolute value of the unevenness rate riv (i) is equal to or greater than the predetermined value Δ th (yes in S72), the ignition timing aig (# i) of each cylinder # i is corrected based on the newly calculated unevenness rate riv (i) (S74). Here, the CPU72 performs processing for advancing the ignition timing aig for a cylinder in which the absolute value of the unevenness rate riv (i) is larger than zero, for example. This is because: when the unevenness rate Riv deviates from zero due to the fact that the air-fuel ratio becomes excessively rich and there is a sign of misfire (misfiring), it is possible to improve combustion by advancing the ignition timing aig.

When the process at S74 is completed and it is determined as no in the process at S72, the CPU72 determines whether or not the absolute value of the unevenness rate riv (i) newly calculated after the process at S68 of fig. 10 is executed is equal to or greater than the predetermined value Δ thp when the EGR rate Regr is zero (S76). This processing is processing for determining whether or not the unevenness is not sufficiently eliminated by the correction amount Δ Qd (livp (i)) based on the learning value livp (i) at the time of elimination. When the CPU72 determines that the absolute value of the unevenness rate riv (i) is equal to or greater than the predetermined value Δ thp (yes in S76), the ignition timing aig (# i) of each cylinder # i is corrected based on the newly calculated unevenness rate riv (i) and the target clearance Rp (S78).

When the process at S78 is completed and it is determined as no in the process at S76, the CPU72 determines whether or not the absolute value of the unevenness rate riv (i) newly calculated after the process at S70 of fig. 10 is performed is equal to or greater than the predetermined value Δ the when the EGR rate Regr is greater than zero (S80). This process is a process of determining whether or not the unevenness is not sufficiently eliminated by the correction amount Δ Qd (live (i)) based on the EGR time learning value live (i). When the CPU72 determines that the absolute value of the unevenness rate riv (i) is equal to or greater than the predetermined value Δ the (yes in S80), the ignition timing aig (# i) of each cylinder # i is corrected based on the newly calculated unevenness rate riv (i), the target purge rate Rp, and the EGR rate Regr (S82).

When the process at S82 is completed and the process at S80 is determined as no, the CPU72 once ends the series of processes shown in fig. 11.

< embodiment 4 >

Hereinafter, referring to fig. 12, embodiment 4 will be described mainly focusing on differences from embodiment 1.

In the present embodiment, a plurality of types of map data are stored in the storage device 76 as map data 76 a. The above-mentioned plural kinds of mapping data are mapping data for each control mode.

The steps of the processing performed prior to the processing of fig. 3 are shown in fig. 12. The process shown in fig. 12 is realized by the CPU72 repeatedly executing the unevenness detection program 74a stored in the ROM74, for example, at predetermined cycles.

In the series of processes shown in fig. 12, the CPU72 determines whether or not it is during the warm-up process of the catalyst 30 (S84). The warm-up process includes a process of retarding the ignition timing aig by a predetermined amount. The warm-up processing is executed by the CPU72 when the logical product of the integrated value of the intake air amount Ga from the time of starting the internal combustion engine 10 is equal to or less than a predetermined value and the water temperature THW is equal to or less than a predetermined temperature is true, for example.

When the CPU72 determines that the period is not the catalyst warm-up processing period (S84: no), it selects map data for warm-up from the map data 76a stored in the storage device 76 (S86). Thus, the CPU72 calculates the unevenness ratio riv (i) using the map data for after-warm-up in the processing of S16. Here, the post-warm-up map data is data generated by processing shown in fig. 6 using training data generated based on detection values obtained by the sensor group 102 shown in fig. 5 after warm-up of the internal combustion engine 10.

On the other hand, if the CPU72 determines that the catalyst warm-up processing period is present (yes in S84), the CPU72 selects map data for warm-up from the map data 76a stored in the storage device 76 (S88). The map data for warm-up is data generated based on training data corresponding to the detection values obtained by the sensor group 102 while the warm-up process of the internal combustion engine 10 is being executed.

Further, the CPU72 once ends the series of processing shown in fig. 12 when the processing of S86 and S88 is completed.

Here, the operation and effect of the present embodiment will be described.

The CPU72 calculates the unevenness rate Riv using the map data for warm-up when the warm-up process of the catalyst 30 is executed, and calculates the unevenness rate Riv using the map data for warm-up when the warm-up process is not executed. During the warm-up process of the catalyst 30, since the ignition timing aig is on the retard side, the rotational behavior of the crankshaft 24 is different from that when the warm-up process is not being performed. However, in the case where a single map is to be generated in which the unevenness rate Riv is calculated with high accuracy, for example, it is preferable to add data that can distinguish between the warm-up period and the non-warm-up period of the catalyst 30 to the input variable x, and in this case, the number of input-side coefficients wFjk and the like increases, so a large amount of training data is required. In particular, whether or not the execution condition of the preheating process is satisfied is not determined only by the two sampled values of the intake air amount Ga and the water temperature THW, and therefore, the dimension of the input variable tends to be excessively large in order to be able to distinguish between the preheating process period and the non-preheating process period of the catalyst 30. Further, when the dimension of the input variable is increased, for example, it is easy to increase the intermediate layer of the neural network, which makes the structure of the map complicated. However, if the number of intermediate layers increases and the structure of the map becomes complicated, the CPU72 has a large computational load. In this regard, in the present embodiment, by dividing the map data between the time of catalyst warm-up and the time after warm-up, it is possible to suppress the number of dimensions of the input variable x from becoming excessively large while ensuring the calculation accuracy of the unevenness ratio Riv, and to easily simplify the structure of the map.

< embodiment 5 >

Hereinafter, referring to fig. 13, the description will be given of embodiment 5 focusing on differences from embodiment 1.

In the present embodiment, a plurality of types of map data are stored in the storage device 76 as map data 76 a. The above-described map data is map data for each rotation speed NE.

The steps of the processing performed before the processing shown in fig. 3 are shown in fig. 13. The process shown in fig. 13 is realized by the CPU72 repeatedly executing the unevenness detection program 74a stored in the ROM74, for example, at predetermined cycles.

In a series of processing shown in fig. 13, the CPU72 first acquires the rotation speed NE (S90). Next, the CPU72 determines whether the rotation speed NE is equal to or less than the low rotation speed threshold NE1 (S92). When the CPU72 determines that the low rotation threshold NE1 is equal to or less than the low rotation threshold NE1 (S92: yes), it selects the low rotation map data from the map data 76a stored in the storage device 76 (S94). Thus, the CPU72 calculates the unevenness rate Riv using the low rotation map data in the processing of S16. Here, the low rotation map data is data generated by processing shown in fig. 6, in which training data is generated from detection values obtained by the sensor group 102, when the rotation speed NE is equal to or less than the low rotation threshold NE1, based on the system illustrated in fig. 5, in which the dynamometer 100 is controlled, for example.

On the other hand, if the CPU72 determines that the rotation speed NE is greater than the low rotation speed threshold NE1 (S92: no), it determines whether the rotation speed NE is equal to or less than the intermediate rotation speed threshold NE2 (S96). The intermediate rotation speed threshold NE2 is set to a value greater than the low rotation speed threshold NE 1. When the CPU72 determines that the rotation speed NE is equal to or less than the intermediate rotation speed threshold NE2 (S96: yes), it selects intermediate rotation speed map data from the map data 76a stored in the storage device 76 (S98). The map data for the intermediate rotation speed is data generated based on training data corresponding to the detection values obtained by the sensor group 102 when the rotation speed NE of the internal combustion engine 10 is greater than the threshold value NE1 for the low rotation speed and equal to or less than the threshold value NE2 for the intermediate rotation speed.

On the other hand, when the CPU72 determines that the rotation speed NE is greater than the intermediate rotation speed threshold NE2 (S96: no), it selects the high rotation speed map data from the map data 76a stored in the storage device 76 (S100). The high-rotation-speed map data is data generated based on training data corresponding to the detection values obtained by the sensor group 102 when the rotation speed NE of the internal combustion engine 10 is greater than the medium-rotation-speed threshold value NE 2.

Further, when the processing of S94, S98, S100 is completed, the CPU72 temporarily ends the series of processing shown in fig. 13.

As described above, in the present embodiment, by using the map data separately according to the rotation speed NE, it is possible to use appropriate map data at each rotation speed NE. Therefore, for example, compared to the case where common map data is used regardless of the magnitude of the rotation speed NE, the present embodiment can easily simplify the structure of the map while ensuring the calculation accuracy of the unevenness ratio Riv.

< embodiment 6 >

Hereinafter, embodiment 6 will be described mainly focusing on differences from embodiment 1 with reference to fig. 14.

In the present embodiment, a plurality of types of map data are stored in the storage device 76 as map data 76 a. The above-described plural kinds of map data are map data for each shift position Sft.

Fig. 14 shows a procedure of the map data selection processing according to the present embodiment. The process shown in fig. 14 is realized by the CPU72 repeatedly executing the unevenness detection program 74a stored in the ROM74, for example, at predetermined cycles.

In a series of processing shown in fig. 14, the CPU72 first acquires the shift position Sft (S102). Next, the CPU72 determines the shift position Sft based on the determination process of whether the shift position Sft is 1 st (S104), 2 nd (S106), 3 rd (S108), and 4 th (S110). When the CPU72 determines that the data is 1 st gear (yes in S104), it selects map data for 1 st gear from the map data 76a stored in the storage device 76 (S112). Thus, the CPU72 calculates the unevenness rate Riv using the mapping data for 1 st stage in the processing of S16. Similarly, the CPU72 selects the mapping data for the 2 nd gear (S114) when it is determined that the gear is the 2 nd gear (YES in S106), selects the mapping data for the 3 rd gear (S116) when it is determined that the gear is the 3 rd gear (S108: YES), and selects the mapping data for the 4 th gear (S118) when it is determined that the gear is the 4 th gear (S110: YES). When the CPU72 determines that the position is the 5 th position (no in S110), it selects the 5 th position mapping data (S120).

Here, when the variable m is set to "1" to "5", the m-range map data is data generated by the processing shown in fig. 6 using training data generated based on the detection values obtained by the sensor group 102 shown in fig. 5 when the transmission 64 is set to the m range.

When the processing in S112 to S120 is completed, the CPU72 once ends the series of processing shown in fig. 14.

According to the present embodiment described above, by using the map data corresponding to the shift position Sft, the unevenness rate Riv can be calculated in a state where the inertia constant from the crankshaft 24 to the drive wheels 69 changes according to the shift position Sft and the rotational behavior of the crankshaft 24 changes according to the shift position Sft is appropriately reflected. Therefore, compared to the case where all the shift positions Sft are dealt with by a single map, the structure of each map can be easily simplified while ensuring the calculation accuracy of the unevenness rate Riv.

< 7 th embodiment >

Hereinafter, embodiment 7 will be described mainly focusing on differences from embodiment 1 with reference to fig. 15.

In the present embodiment, a plurality of types of map data are stored in the storage device 76 as map data 76 a. The plurality of types of map data are map data of the state of each road surface.

The steps of the processing performed prior to the processing of fig. 3 are shown in fig. 15. The processing shown in fig. 15 is realized by the CPU72 repeatedly executing the unevenness detection program 74a stored in the ROM74, for example, at predetermined cycles.

In the series of processing shown in fig. 15, the CPU72 first acquires the acceleration Dacc (S122). Next, the CPU72 determines whether or not the poor road determination flag Fb is "1" (S124). The poor road determination flag Fb is "1" when the undulation of the road surface on which the vehicle VC is traveling is large, and is "0" otherwise. When determining that the poor road determination flag Fb is "0" (yes in S124), the CPU72 determines whether or not a state in which the absolute value of the change amount of the acceleration Dacc is equal to or greater than a predetermined value Δ DH continues for a predetermined time (S126). Here, the amount of change in the acceleration Dacc may be the difference between the previous value and the present value of the acceleration Dacc. This process is a process of determining whether or not there are many undulations on the road surface. When the CPU72 determines that a state in which the absolute value of the amount of change in the acceleration Dacc is equal to or greater than the predetermined value Δ DH has continued for a predetermined time (yes in S126), it substitutes "1" for the bad road determination flag Fb (S128), and selects bad road map data from the map data 76a stored in the storage device 76 (S130). Thus, the CPU72 calculates the unevenness rate Riv using the defective road map data in the processing of S16. For example, the dynamometer 100 shown in fig. 5 may be used to simulate the load torque applied to the crankshaft 24 when vibration is applied to the vehicle, and generate training data for generating poor road map data based on the detection values obtained by the sensor group 102 shown in fig. 5. Further, the internal combustion engine 10, the dynamometer 100, and the like shown in fig. 5 are mounted on a device capable of generating vibration, for example. Further, it is also possible to simulate the load torque applied to crankshaft 24 when vibration is applied to the vehicle by generating vibration, and generate training data based on the detection values obtained by sensor group 102 shown in fig. 5.

On the other hand, if the poor road determination flag Fb is "1" (yes in S124), the CPU72 determines whether or not a state in which the absolute value of the change amount of the acceleration Dacc is equal to or less than a predetermined value Δ DL (< Δ DH) continues for a predetermined time (S132). If the CPU72 determines no in the process at S132, the process proceeds to S130.

On the other hand, when the CPU72 determines that the state in which the absolute value of the amount of change in the acceleration Dacc is equal to or less than the predetermined value Δ DL (< Δ DH) continues for the predetermined time (yes in S132), it substitutes "0" for the bad road determination flag Fb (S134). When the process at S134 is completed and the process at S126 is determined as no, the CPU72 selects normal map data from the map data 76a (S136). The normal map data is, for example, data generated by processing shown in fig. 6 using training data generated based on detection values obtained by the sensor group 102 shown in fig. 5 in a state where intentional vibration is not applied to the internal combustion engine 10 and the dynamometer 100 shown in fig. 5.

When the processing in S130 and S136 is completed, the CPU72 once ends the series of processing shown in fig. 15.

According to the present embodiment described above, by using the map data corresponding to the state of the road surface, the unevenness rate Riv can be calculated in a state in which the rotation behavior of the crankshaft 24 changes depending on the state of the road surface is appropriately reflected. Further, for example, compared to the case where the states of all the road surfaces are dealt with by a single map, the structure of the map can be easily simplified while ensuring the calculation accuracy of the unevenness ratio Riv.

< embodiment 8 >

Hereinafter, referring to fig. 16, embodiment 8 will be described mainly focusing on differences from embodiment 1.

In the present embodiment, the dimension of the input variable x is enlarged.

Fig. 16 shows a procedure of the unevenness detecting process according to the present embodiment. The processing shown in fig. 16 is realized by the CPU72 repeatedly executing the unevenness detection program 74a stored in the ROM74, for example, at predetermined cycles. In fig. 16, for convenience, the same step numbers are assigned to the processes corresponding to the processes shown in fig. 2, and the descriptions thereof are omitted.

In the series of processing shown in fig. 16, if the CPU72 determines yes in the processing of S10, the CPU acquires the ignition timing average value aigave, the EGR rate average value Regrave, the water temperature THW, the shift position Sft, the road surface state variable SR, and the alcohol concentration Dal, in addition to the variables acquired in the processing of S12 (S12 a). Here, the ignition timing average value aigave and the EGR rate average value Regr are average values of the ignition timing aig and the EGR rate Regr in the cycle of the processing in fig. 16. The road surface state variable SR is a parameter calculated by the CPU72 based on the acceleration Dacc, and may be a variable that has a binary value that differs between a case where the road surface undulation is large and a case where the road surface undulation is small, for example. The process of generating the road surface state variable SR may be the same as the process of setting the value of the poor road determination flag Fb in the process of fig. 15, for example. The alcohol concentration Dal may be estimated, for example, from the correction ratio of the feedback process M20 described above.

Next, the CPU72 inputs the data acquired in S12a into the input variables x (1) to x (57) of the map (S14 a). Here, the ignition timing average value aigave and the EGR rate average value Regrave are parameters (adjustment variables) for adjusting the combustion speed of the air-fuel mixture in the combustion chamber 18 by the operation of the operation unit of the internal combustion engine 10. That is, the ignition timing aig is the operation amount of the ignition device 22 as the operation unit, and is an adjustment variable for adjusting the combustion speed by the operation of the ignition device 22. The EGR rate Regr is an adjustment variable for adjusting the combustion speed by operating the EGR valve 34 as an operation unit. The adjustment variable is added to the input variable x in consideration of the fact 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 friction of a sliding portion such as friction between a piston and a cylinder. In view of the fact that the rotational behavior of the crankshaft 24 also changes in accordance with the water temperature THW, in the present embodiment, the water temperature THW is added to the input variable x.

Further, the stoichiometric air-fuel ratio differs depending on the alcohol concentration Dal, and therefore may cause a variation in the degree of variation in the fuel injection amount. In view of this, in the present embodiment, the alcohol concentration Dal is added to the input variable.

When the process of S14a is completed, the CPU72 inputs the input variables x (1) to x (57) to the map defined by the map data 76a, thereby calculating the unevenness rate riv (i) of the cylinders #1 to #4 (S16 a). The map is composed of a neural network having "α" intermediate layers, and the activation functions h1 to h α of the intermediate layers and the activation function f of the output layer are hyperbolic tangent functions. For example, the value of each node in the 1 st intermediate layer is generated by inputting, to the activation function h1, an output obtained when the input variables x (1) to x (57) are input to a linear map defined by coefficients w (1) ji (j is 0 to n1, i is 0 to 57). That is, when m is 1, 2, …, α, the value of each node in the mth intermediate level is generated by inputting the output of the linear map defined by the coefficient w (m) to the activation function hm. Here, n1, n2, …, and n α are the number of nodes in the 1 st, 2 nd, … th, and α -th intermediate layers, respectively. w (1) j0, etc. are bias parameters (bias parameters), and the input variable x (0) is defined as "1".

When the process of S16a is completed, the CPU72 once ends the process of fig. 16.

In the present embodiment, the training data acquired in the process of S30 in fig. 6 includes all the parameters acquired in S12a, and the process corresponding to S14a is executed as the process of S32, and the process of S34 uses a map in which the input variable x is 57-dimensional. However, here, the input variable x (0) is not included in the dimension number. The training data corresponding to the case where the road surface state variable SR indicates that the road surface has a large amount of undulations is the same as the data used to generate the defective road map data shown in fig. 15.

As described above, in the present embodiment, the input variable x is added with 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. Although these parameters are parameters that affect the rotational behavior of crankshaft 24, it is difficult to take into account the conventional method of detecting unevenness based on the difference between the times required for rotations at adjacent angular intervals. That is, when the above parameters are taken into consideration, the amount of work for application is enormous, and the implementation is poor. In this regard, in the present embodiment, by using machine learning, it is possible to construct a logic for calculating the unevenness rate riv (i) by considering them within a range of practical applicable workload.

< embodiment 9 >

Hereinafter, the 9 th embodiment will be described mainly focusing on differences from the 1 st embodiment with reference to fig. 17 and 18.

In the present embodiment, the process of calculating the unevenness ratio riv (i) is performed outside the vehicle.

Fig. 17 shows a unevenness detecting system according to the present embodiment. In fig. 17, for convenience, the same reference numerals are given to components corresponding to those shown in fig. 1.

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

The center 120 parses data sent from a plurality of vehicles VC. The center 120 includes a CPU122, a ROM124, a storage device 126, a peripheral circuit 127, and a communication device 129, and can communicate with each other via a local area network 128. The ROM124 stores a main unevenness detection program 124a, and the storage device 126 stores mapping data 126 a.

The steps of the process performed by the system shown in fig. 17 are shown in fig. 18. The processing shown in part (a) of fig. 18 is realized by the CPU72 executing the unevenness detection subroutine 74c stored in the ROM74 shown in fig. 17. The processing shown in part (b) of fig. 18 is realized by CPU122 executing unevenness detection main program 124a stored in ROM 124. In fig. 18, for convenience, the same step numbers are assigned to the processes corresponding to the process shown in fig. 16. The processing shown in fig. 18 will be described below in terms of the sequence of the unevenness detecting processing.

As shown in part (a) of fig. 18, in the vehicle VC, the CPU72, after completing the process of S12a, transmits the data acquired in the process of S12a to the center 120 by operating the communicator 79 together with the vehicle ID as the identification information of the vehicle VC (S132).

In contrast, as shown in part (b) of fig. 18, the CPU122 of the center 120 receives the transmitted data (S140), and executes the processing of S14a, S16 a. Then, CPU122 operates communicator 129 to transmit a signal relating to unevenness rate riv (i) to vehicle VC that has transmitted the data received in S140 (S142), and temporarily ends the series of processing shown in fig. 18 (b). On the other hand, as shown in fig. 18 a, the CPU72 receives a signal relating to the unevenness rate riv (i) (S134), and once ends the series of processing shown in fig. 18 a.

In this way, in the present embodiment, the center 120 calculates the unevenness rate Riv, and therefore, the calculation load of the CPU72 can be reduced.

< correspondence >)

The correspondence between the matters in the above embodiment and the matters described in the above section of "summary of the invention" is as follows. In the following, the correspondence relationship is shown for each number of the example described in the column of "contents of invention".

[1] And 7 unevenness detecting means corresponds to the control means 70. The execution devices correspond to the CPU72 and the ROM 74. The storage device corresponds to the storage device 76.

The predetermined rotation angle interval corresponds to 720 ° CA, and the 1 st and 2 nd intervals correspond to 30 ° CA. The air-fuel ratio detection variable corresponds to the upstream-side average value Afuave, and the instantaneous speed variable corresponds to the minute rotation time T30. The plurality of rotation waveform variables correspond to minute rotation times T30(1) to T30 (24).

The acquisition processing corresponds to the processing of S12 and S12 a. The calculation processing corresponds to the processing of S14 and S16 and the processing of S14a and S16 a. The coping process corresponds to the processes of S24, S26, S64, S68, S70, S74, S78, S82.

[2] The 0.5 order component variable corresponds to an amplitude Ampf/2 of order 0.5.

[3] The operating point variables correspond to the rotational speed NE and the filling efficiency η.

[4] The adjustment variables correspond to the ignition timing average value aigave, the EGR rate average value Regrave.

[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 processing corresponds to the processing of fig. 12 to 15.

[9] The adjusting device corresponds to the purge valve 44. The variation in accordance with the presence or absence of the influence of the fuel vapor corresponds to the variation rate Riv when the determination in S52 is yes and the variation rate Riv when the determination is no.

[10] The 1 st learning value corresponds to the unevenness learning value Liv shown in fig. 9 and 10, and the 2 nd learning value corresponds to the clear time learning value Livp shown in fig. 9 and 10.

[11] The variation variable corresponding to the presence or absence of the influence of the exhaust gas flowing in through the EGR passage corresponds to the variation rate Riv when it is determined as yes in the process of S56 and the variation rate Riv when it is determined as no.

[12] The 1 st learning value corresponds to the unevenness learning value Liv and the learning value Livp at the time of purge shown in fig. 9 and 10, and the 2 nd learning value corresponds to the EGR learning value Live shown in fig. 9 and 10.

[13] The operation portion for controlling combustion (combustion operation portion) corresponds to the fuel injection valve 20 and the ignition device 22.

[14] The correction amounts different from each other correspond to Δ Qd.

[15] The 1 st execution device corresponds to the CPU72 and the ROM 74. The 2 nd execution means corresponds to the CPU122 and the ROM 124. The calculation processing corresponds to the processing of S14a and S16a in fig. 18. The vehicle-side transmission processing corresponds to the processing of S132, and the vehicle-side reception processing corresponds to the processing of S134. The external-side reception process corresponds to the process of S140, and the external-side transmission process corresponds to the process of S142.

[16] The data parsing means corresponds to the center 120.

[17] The control device of the internal combustion engine corresponds to the control device 70 shown in fig. 16.

[23] The machine learning corresponds to the processing of fig. 5 and 6.

< other embodiments >

The present embodiment can be modified and implemented as follows. The present embodiment and the following modifications can be implemented in combination with each other within a range not technically contradictory.

"about 1 st Interval, 2 nd Interval"

The 1 st interval, which is the sampling interval of the upstream-side average value Afuave as an input to the map, is not limited to 30 ° CA. For example, the angular interval may be smaller than 30 ° CA, such as 10 ° CA. The angular interval is not limited to 30 ° CA or less, and may be, for example, 45 ° CA.

The 2 nd interval, which is the sampling interval of the minute rotation time T30 as an input to the map, is not limited to 30 ° CA. For example, the angular interval may be smaller than 30 ° CA, such as 10 ° CA. The angular interval is not limited to 30 ° CA or less, and may be, for example, 45 ° CA.

In addition, it is not necessary that the 1 st interval and the 2 nd interval are the same size interval.

"about a predetermined rotation angle interval (predetermined angle interval)"

The time series data of the upstream-side average value Afuave and the time series data of the minute rotation time T30, which are input as the map, are not limited to the time series data at the rotation angle interval of 720 °. For example, the time series data may be in a rotation angle interval wider than 720 ° CA. As described in the column of "various types of map data" described below, for example, when a map is used that outputs only the unevenness rate riv (i) of a specific cylinder # i, the map may be time series data having an angular interval narrower than 720 ° CA.

For example, when the time series data of the minute rotation time T30 in the 1 st predetermined period and the time series data of the upstream-side average value Afuave in the 2 nd predetermined period are input to the map, one of the two periods 1 and 2 nd predetermined period may be delayed from the other. Further, it is not essential that the 1 st scheduled period and the 2 nd scheduled period are periods having an equal angular interval to each other.

"about a rotating waveform variable as input to a map"

In the above embodiment, the minute rotation time T30 in each of the plurality of intervals obtained by dividing the rotation angle interval of 720 ° CA, which is 1 combustion cycle, is input to the map, but the present invention is not limited to this. For example, 0 to 20, 40 to 60, 80 to 100, 120 to 140, 160 to 180, …, 700 to 720 ° CA of 0 to 720 ° CA may be set as the 2 nd interval, and the time required for their rotation may be set as the input to the map.

The instantaneous speed variable is not limited to a minute rotation time which is a time required for the 2 nd interval of rotation. For example, the value may be obtained by dividing the 2 nd interval by the minute rotation time. It is not essential that the instantaneous speed variable is a variable subjected to normalization processing in which the difference between the maximum value (maximum value) and the minimum value (minimum value) is a fixed value. The filtering process as the preprocessing performed on the instantaneous speed variable to input the instantaneous speed variable as the map is not limited to the above, and for example, the instantaneous speed variable may be subjected to a process of removing an influence portion of the rotation of the crankshaft 24 through the input shaft 66 based on the minute rotation time of the input shaft 66 of the transmission 64. Also, it is not necessary to perform filtering processing on the instantaneous speed variable as an input of the map.

The rotation waveform variable as an input to the map is not limited to the time series data of the instantaneous speed variable. For example, the time series data may be the difference between a pair of instantaneous speed variables separated by the occurrence interval of compression top dead center. In this case, the instantaneous speed variable is an angular interval equal to or smaller than the interval of occurrence of the compression top dead center as the 2 nd interval, and the minute rotation time, which is the time required for the rotation, is set as the instantaneous speed variable, or a value obtained by dividing the 2 nd interval by the minute rotation time is set as the instantaneous speed variable.

"air-fuel ratio detection variable as input to map"

In the above embodiment, the upstream-side average value Afuave of each of the plurality of intervals obtained by dividing the rotation angle interval of 720 ° CA of 1 combustion cycle is input to the map, but the present invention is not limited thereto. For example, 0 to 20, 40 to 60, 80 to 100, 120 to 140, 160 to 180, …, 700 to 720 of CA degrees may be set as the 1 st interval, and the upstream average Afuave of them may be set as the input to the map.

In the above embodiment, the upstream-side average value Afuave is set to the average value of the plurality of sample values in the 1 st interval of the primary combustion cycle, but the present invention is not limited thereto. For example, the upstream-side average value Afuave may be calculated by sampling the upstream-side detection value Afu at each 1 st interval over a plurality of combustion cycles and averaging the values at each 1 st interval in each 1 st interval.

The air-fuel ratio detection variable is not limited to the upstream-side average value Afuave, and may be an upstream-side detection value Afu.

"operating point variables for internal combustion engines"

In the above embodiment, the operating point variable is defined by the rotation speed NE and the charging efficiency η, but the present invention is not limited thereto. For example, the rotation speed NE and the intake air amount Ga may be used. For example, the filling efficiency η may be replaced with a required torque for the internal combustion engine or an injection amount as the load. The use of the injection amount and the required torque as the load is particularly effective in the compression ignition type internal combustion engine described in the following section "related to internal combustion engines".

"adjustment variables for adjusting the Combustion speed of the mixture" as input to the map "

In the processing of S16a, the ignition timing average value aigave and the EGR rate average value Regrave are used as the adjustment variables that are input to the map and adjust the combustion speed, but only one of the two parameters may be used.

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

For example, when the present invention is applied to a compression ignition type internal combustion engine as described in the column of "internal combustion engine" described below, the injection timing and/or the average value thereof may be used instead of the ignition variable.

The adjustment variable for adjusting the combustion speed as an input to the map is not limited to the above-described variable, and may be, for example, the opening timing and lift amount of the intake valve 16, the opening timing and lift amount of the exhaust valve 26, or the like. For example, the information may be information of a cylinder in which combustion control is stopped during a cylinder deactivation control period in which combustion control of the air-fuel mixture is stopped in a plurality of cylinders. For example, when the internal combustion engine 10 includes a supercharger and a wastegate valve, the opening degree of the wastegate valve may be set.

"with respect to drive System State variables as inputs to the map"

In the process of S16a, the shift position Sft is used as the state variable of the drive system device connected to the crankshaft 24, but the present invention is not limited thereto. For example, the state of the lockup clutch 62 may be set. This state may be, for example, a binary state of an engaged state or a released state, or may be a configuration in which each of the engaged state, the released state, and the slip state is recognized.

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

"with respect to road surface state variables as inputs to the map"

The road surface state variable SR is not limited to a binary variable corresponding to the magnitude of the undulation of the road surface, and may be, for example, a three-value or more variable corresponding to the magnitude of the undulation. 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, an average value of the absolute value of the acceleration Dacc per a predetermined time may be used. Thereby, the road surface state variable SR becomes a continuous quantity.

The road surface state variable SR is not limited to a configuration that is quantified based on a detected value of the acceleration of the vehicle VC in the vertical direction. For example, the vehicle may be quantified based on the lateral acceleration and the front-rear acceleration of the vehicle. Further, as the input of the map, the vehicle may be quantified using any two or all of three accelerations, i.e., the vertical acceleration, the lateral acceleration, and the longitudinal acceleration of the vehicle.

The road surface state variable SR used in the process of S16a is quantified from the acceleration Dacc detected by the acceleration sensor 92, but is not limited thereto. For example, data obtained by imaging the road surface with a camera or data obtained by analyzing data relating to reflected waves of radar may be used. Further, for example, the state of the road surface at the current position may be acquired by using map data having the state of the road surface and a GPS, and the state may be used. In addition, the air pressure of each wheel may be used. That is, the difference in the air pressure between the wheels includes information about the inclination and unevenness of the road surface. Further, for example, a detection value of a sensor that detects a position of a liquid level of a fluid in the vehicle, such as a fuel gauge or an oil gauge, may be used. That is, the fluctuation of the liquid level includes information such as unevenness of the road surface. Further, a detection value of a sensor that detects a hydraulic pressure of the suspension may be used. That is, the hydraulic pressure includes information such as irregularities on the road surface.

The road surface state variable is not limited to a variable for quantifying irregularities on the road surface. For example, the variable may be a variable obtained by quantifying the magnitude of friction on the road surface, such as the presence or absence of icing, or may be a variable for identifying both undulation of the road surface and the magnitude of friction.

"input on mapping"

The input of the map is not limited to the configurations exemplified in the above embodiment and the above modification. For example, parameters relating to the environment in which the internal combustion engine 10 is located may also be used. Specifically, for example, atmospheric pressure may be used. Further, for example, an 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. For example, humidity may be used as a parameter that affects the combustion speed of the air-fuel mixture in the combustion chamber 18. For the determination of the humidity, a humidity sensor may be used, or the state of the wiper or the detection value of a sensor for detecting raindrops may be used. Further, the data may be data relating to the state of an auxiliary machine mechanically coupled to crankshaft 24.

It is not essential that the map input includes the operating point of the internal combustion engine 10. For example, when the internal combustion engine is mounted on a series hybrid vehicle as described in the following section "about the internal combustion engine" and control is assumed that the operating point is limited to a narrow range, the unevenness ratio riv (i) can be calculated with high accuracy without including the operating point.

"mapping data on each control schema"

The map data for each control mode is not limited to map data that is separately set for each of the warm-up control mode and the mode after completion of warm-up of the catalyst 30 as illustrated in fig. 12. For example, in the case of an internal combustion engine provided with both a port injection valve and an in-cylinder injection valve as described in the column of "regarding the internal combustion engine (10)" below, it is also possible to provide a mode in which injection control of fuel by only the port injection valve is performed and a mode in which injection control of fuel using the in-cylinder injection valve is performed separately.

"map data on each state variable of the internal combustion engine as a result of control"

In fig. 13, the number of the map data for each rotation speed NE is 3, but the map data is not limited to this, and may be two or 4 or more.

The state variable of the internal combustion engine 10 is grasped as a result of the control, and the map data of each state variable of the internal combustion engine 10 as a result of the control is not limited to the map data of each rotation speed NE illustrated in fig. 13. For example, it may be set for each filling efficiency η. For example, the engine speed NE and the charging efficiency η may be set for each operating point of the internal combustion engine 10.

"mapping data on State variables of Each drive System"

For example, when a continuously variable transmission is used as the transmission as described in the "other" column below, the available range of the transmission ratio may be divided into a plurality of ranges, and the map data may be provided for each of the ranges.

The map data as the state variable of each drive system coupled to the crankshaft 24 is not limited to the map data for each shift position Sft. For example, instead of using the drive system state variables described in the column "drive system state variables as inputs to the map" as inputs to the map, map data may be provided for each of the state variables. In the case where the mapping data is set for each of the above state variables, it is not necessary that the state variable is not included in the input of the mapping.

"mapping data on each road surface State variable"

The map data as each road surface state variable is not limited to the configuration illustrated in fig. 15. For example, instead of using the road surface state variable SR described in the column "road surface state variable as input of map" as input of map, a map may be provided for each of the road surface state variables SR. Further, in the case where a map is provided for each road surface state variable SR, it is not necessary that the road surface state variable SR is not included in the input of the map.

"about various mapping data"

The setting as the various kinds of map data is not limited to the setting for each state variable of the internal combustion engine 10 as a result of the control, the setting for each control mode, the setting for each drive system state variable, and the setting for each road surface state variable. For example, instead of using the adjustment variables described in the column "adjustment variables for adjusting the combustion speed of the air-fuel mixture, which are inputs to the map", as inputs to the map, map data may be provided for each of the adjustment variables. Further, for example, when map data is provided for each adjustment variable such as an EGR variable, it is not necessary that the adjustment variable is not included in the input of the map. For example, instead of using the parameter related to the environment in which the internal combustion engine is located described in the above-described "input to map" column, map data may be provided for each parameter related to the environment. Also, in the case where the mapping data is set for each parameter relating to the environment, it is not necessary that the parameter is not included in the input of the mapping.

For example, different pieces of map data may be provided when the target clearance rate Rp is zero and when the target clearance rate Rp is greater than zero.

Further, the plurality of types of mapping data may be a combination of the plurality of types of mapping data. That is, for example, the map data for each shift position Sft may be further subdivided into map data for each rotation speed NE.

In addition, the number of pieces of map data defining a map for outputting the unevenness rate Riv may be set as the number of cylinders for the unevenness rate Riv of a specific cylinder.

"about mapping data"

For example, in the processing of S16a in fig. 16, the number of intermediate layers in the neural network may be 2 or more.

In the processing of S16a in each of fig. 16 and 18, the number of intermediate layers of the neural network is shown to be greater than 2 layers in the mathematical expression (α > 2), but the present invention is not limited thereto. In particular, from the viewpoint of reducing the computational load of the control device 70, it is preferable to reduce the number of intermediate layers of the neural network to 1 layer or 2 layers.

In the above embodiment, the activation functions h, f, h1, h2, and … h α are hyperbolic tangent functions, but the present invention is not limited thereto. For example, some or all of the activation functions h, h1, h2, and … h α may be ReLU, or may be logical sigmoid functions (logical signature function), for example.

In fig. 3 and the like, the map of the output unevenness rate Riv is not limited to a configuration in which learning is performed for a phenomenon in which the fuel injection amount deviates from the assumed amount for any cylinder. For example, learning may be performed for a phenomenon in which the fuel injection amount deviates from the assumed amount in two cylinders. In this case, the intensity of the 1 st order component of the rotational frequency of crankshaft 24 may be included as an input to the map.

Further, as an input to the neural network, each dimension is not limited to being constituted by a single physical quantity. For example, as for a part of a plurality of types of physical quantities using the physical quantities and the like exemplified above, instead of using them as direct inputs to the neural network, a plurality of principal components based on principal component analysis thereof may be used as direct inputs to the neural network. In the case where the principal component is input to the neural network, it is not necessary that only a part of the input to the neural network is the principal component, and all of the inputs may be the principal component. In addition, in the case where a principal component is included in the input to the neural network, data that specifies a map that specifies the principal component is included in the map data 76a, 126 a.

"Algorithm for machine learning"

The algorithm for machine learning is not limited to the use of a neural network. Regression equations may also be used, for example. This corresponds to a configuration in which the neural network described above does not have an intermediate layer.

"Generation of mapping data"

In the above embodiment, the data when the internal combustion engine 10 is operated with the crankshaft 24 connected to the dynamometer 100 via the torque converter 60 and the transmission 64 is used as the training data, but the present invention is not limited thereto. For example, data obtained when the internal combustion engine 10 is driven with the internal combustion engine 10 mounted on the vehicle VC may be used as the training data.

"about data parsing device"

The center 120 may execute the process of S22 and a process of notifying the user' S mobile terminal of the presence of an abnormality in place of the process of S24.

For example, the process of part (b) of fig. 18 may be executed by a mobile terminal held by the user.

"about the actuator"

The execution device is not limited to a configuration including the CPU72(122) and the ROM74(124) and executing software processing. For example, a dedicated hardware circuit (e.g., ASIC) may be provided for performing hardware processing on at least a part of the processing performed by software processing in the above-described embodiment. That is, the execution device may be configured as any one of the following (a) to (c). (a) The processing device includes a processing device that executes all of the above-described processing in accordance with a program, and a program storage device (including a non-transitory computer-readable storage medium) such as a ROM that stores the program. (b) The apparatus includes a processing device and a program storage device for executing a part of the above processes in accordance with a program, and a dedicated hardware circuit for executing the remaining processes. (c) The apparatus includes a dedicated hardware circuit for executing all of the above processes. Here, a plurality of software execution devices and dedicated hardware circuits may be provided, each of which includes a processing device and a program storage device.

"about storage device"

In the above embodiment, the storage device storing the map data 76a, 126a is a storage device separate from the storage devices (ROMs 74, 124) storing the unevenness detecting program 74a and the unevenness detecting main program 124a, but is not limited thereto.

"relating to internal combustion engines"

In the above-described embodiment, the in-cylinder injection valve that injects fuel into the combustion chamber 18 is exemplified as the fuel injection valve, but the present invention is not limited thereto. For example, a port injection valve may be used to inject fuel into the intake passage 12. For example, both the port injection valve and the in-cylinder injection valve may be provided.

The internal combustion engine is not limited to a spark ignition type internal combustion engine, and may be, for example, a compression ignition type internal combustion engine using light oil or the like as fuel.

The fact that the combustion engine constitutes the drive system is not necessary per se. For example, the internal combustion engine may be mounted on a so-called series hybrid vehicle in which a crankshaft is mechanically coupled to an in-vehicle generator to cut off power transmission from the internal combustion engine to the drive wheels 69.

"about vehicle"

The vehicle is not limited to a vehicle in which the device generating the propulsive force of the vehicle is only an internal combustion engine, and may be a parallel hybrid vehicle or a series/parallel hybrid vehicle, for example, in addition to the series hybrid vehicle described in the column "related to the internal combustion engine" described above.

"other"

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