JP2017190690A - Learning device - Google Patents

Abstract

A technique for maintaining the accuracy of supercharging pressure in air output from a supercharger is provided. An engine 1 generates a driving force of a vehicle by mixing and burning fuel with supply air. The supercharger 2 is driven by the exhaust from the engine 1 and compresses the supply air to increase the supply air pressure. The boost pressure sensor 14 measures the boost pressure of the supply air compressed by the turbocharger 2. The nozzle control unit 106 controls the throttle amount of the nozzle for adjusting the flow rate of the exhaust gas that drives the supercharger 2. The determination unit 101 performs predetermined learning for the vehicle state to learn about the supercharging pressure. It is determined whether the condition is satisfied. When the determination unit 101 determines that the state of the vehicle satisfies the learning condition, the learning unit 102 detects the theoretical value Pt of the supercharging pressure at the nozzle throttle amount N and the actual excess pressure measured by the boosting pressure sensor 14. A function that determines the relationship with the supply pressure Pr is calculated and stored. [Selection] Figure 1

Description

  The present invention relates to a learning device, and more particularly to a learning device for controlling the supercharging pressure of a supercharger.

  It is common practice to attach a turbocharger to an engine. Generally, a supercharger includes a turbine driven by kinetic energy of exhaust gas discharged from an engine, and a compressor driven by torque of the turbine to compress air. If the supercharger is used, the filling amount of air supplied to the combustion chamber of the engine can be increased, so that the output torque of the engine can be increased. Some of such turbochargers include a learning control unit that learns the relationship between the rate of change in the rotational speed of the turbine and the pressure of the air compressed by the compressor.

JP 2009-150322 A

  The initial output performance of a turbocharger changes due to changes over time. Even if the relationship between the rate of change in turbine speed and the pressure of the air compressed by the compressor can be learned, the original target is the air pressure output by the turbocharger with performance different from the initial output performance. In order to obtain a value, feedback control is required. For this reason, in a turbocharger whose output performance has changed due to aging, etc., it takes time to make the air pressure the original target pressure, and as a result, it takes time to increase the output torque of the engine. Will also increase.

  This invention is made | formed in view of these points, and it aims at providing the technique of maintaining the precision of the supercharging pressure in the air which a supercharger outputs.

  One embodiment of the present invention is a learning device. The apparatus includes an engine that generates a driving force of a vehicle by mixing and burning fuel with supply air, and a supercharger that is driven by exhaust from the engine and compresses the supply air to increase a supply air pressure. A supercharging pressure sensor that measures a supercharging pressure of the air that is compressed by the supercharger, and a nozzle control unit that controls a throttle amount of a nozzle for adjusting a flow rate of exhaust gas that drives the supercharger, A determination unit that determines whether or not the vehicle state satisfies a predetermined learning condition for learning about the supercharging pressure, and the vehicle state satisfies the learning condition by the determination unit If it is determined, a learning unit that calculates and stores a function that determines a relationship between the theoretical value Pt of the supercharging pressure at the nozzle throttle amount N and the actual supercharging pressure Pr measured by the supercharging pressure sensor. And comprising.

  The nozzle control unit may change the throttle amount N of the nozzle based on the function calculated by the learning unit so that the actual supercharging pressure Pr becomes the theoretical value Pt.

  The vehicle may further include a distance measuring unit that measures a travel distance of the vehicle, wherein the learning unit performs the learning condition after the travel distance traveled by the vehicle after executing the previous learning becomes longer than a predetermined distance. If the above condition is satisfied, the function may be recalculated and stored.

  The learning unit plots the theoretical value Pt and the actual supercharging pressure in a rectangular coordinate system having the actual supercharging pressure Pr as a first axis and the theoretical value Pt as a second axis. A coefficient of a Kth order polynomial (K is an integer of 3 or more) that approximates the relationship with Pr may be calculated.

  The learning conditions are: (1) the outside air temperature is within a predetermined temperature range, (2) the exhaust gas temperature is equal to or higher than a predetermined threshold temperature, (3) the recirculation control valve is closed, and (4) fluctuations in the engine speed. Is within a predetermined rotational speed variation range, (5) the variation in the injection amount of fuel injected into the cylinder of the engine is within a predetermined injection amount range, and (6) the vehicle has the exhaust amount and the air supply The amount of the air supply may be adjusted by feedback control based on the amount of air.

  ADVANTAGE OF THE INVENTION According to this invention, the technique which maintains the precision of the supercharging pressure in the air which a supercharger outputs can be provided.

It is a figure which shows typically the principal part of the vehicle to which this invention is applied. It is a flowchart for demonstrating the flow of the learning process which the learning part which concerns on embodiment performs. It is a figure which shows typically the graph which plotted the learning data which the learning part which concerns on embodiment acquired on the two-dimensional orthogonal coordinate system. It is a figure for demonstrating the correction process of the aperture amount of the nozzle which the nozzle control part which concerns on embodiment performs.

An outline of an embodiment of the present invention will be described. Embodiments of the present invention relate to a learning apparatus that aims to maintain the initial performance of a supercharger that is provided mainly to increase the amount of air supplied to a diesel engine mounted on a vehicle. The learning device includes a model for calculating the supercharging pressure in the initial performance of the supercharger, and learns the relationship between the supercharging pressure in the initial performance and the supercharging pressure in the current performance. The learning process in this learning device is realized by an ECU (Engine Control Unit).
Hereinafter, an outline of a main part of a vehicle including a supercharger will be described first as a premise of a learning process performed by the learning device according to the embodiment.

<Overview of main parts of vehicle>
FIG. 1 is a diagram schematically showing main parts of a vehicle to which the present invention is applied.
The engine 1 generates driving force of the vehicle by mixing fuel with supply air and burning it. The injector 12 injects and supplies fuel to the engine 1. The control unit 10 controls the amount of fuel that the injector 12 supplies to the engine 1.

  Hereinafter, in the present specification, it is assumed that the vehicle is a large vehicle such as a bus or a truck, and the engine 1 is a diesel engine. The control unit 10 is an ECU and controls the operation of the entire vehicle including the engine 1. In FIG. 1, an arrow indicated by a broken line indicates a control signal between the control unit 10 and its control target, and a flow of measurement data measured by various sensors for use by the control unit 10 for control.

The supercharger 2 is driven by exhaust from the engine 1 and compresses the supply air to increase its pressure. The boost pressure sensor 14 measures the boost pressure of the supply air compressed by the turbocharger 2. The boost pressure sensor 14 outputs the measured boost pressure to the control unit 10.
The air supply pipe 4 extends from the compressor outlet of the supercharger 2 to the engine 1 and is opened and closed by an air supply throttle 5. The recirculation adjusting valve 3 adjusts the amount of exhaust gas recirculation (EGR) for recirculating a part of the exhaust from the engine 1 to the air supply pipe 4 for supplying air to the engine 1.

  The exhaust pipe 6 extends from the turbine outlet of the supercharger 2 to the outside of the vehicle. The exhaust throttle 7 opens and closes the exhaust pipe 6. The duct 8 is installed at a position from the outside air to the compressor inlet of the supercharger 2. The MAF sensor 9 is installed in the duct 8 and measures a MAF (MAss Flow) flow rate that is a flow rate of gas sucked into the air supply pipe 4 from the outside of the vehicle.

  The supercharger 2 according to the embodiment is also referred to as a VGS (Variable Geometry Turbocharger System) turbo, and is a supercharger that can efficiently supercharge even when the exhaust energy is low, such as when the engine 1 is running at a low speed. For this reason, the turbine which comprises the supercharger 2 is the turbine 15 with a nozzle provided with the nozzle for adjusting the flow velocity of exhaust_gas | exhaustion.

  The rotational speed measurement unit 13 measures the rotational speed of the engine 1. The rotation speed measurement unit 13 outputs the measured rotation speed to the control unit 10. The differential pressure sensor 16 measures the pressure difference between the front and rear gases with the circulation adjustment valve 3 interposed therebetween. The differential pressure sensor 16 outputs the measured pressure difference to the control unit 10.

The nozzle control unit 106 in the control unit 10 controls the throttle amount of the nozzle of the above-described nozzle-equipped turbine 15 in order to adjust the flow rate of the exhaust gas that drives the supercharger 2. Further, the opening degree control unit 107 in the control unit 10 controls the opening degree of the recirculation adjustment valve 3.
In addition to these, the vehicle is provided with known sensors, actuators, controllers, etc., but description thereof is omitted.

  In order to make it possible to set the supercharging pressure of the air compressed by the supercharger 2 to a target value after a vehicle equipped with each part shown in FIG. Calibration of the amount and the supercharging pressure at that time is performed. Thereby, the difference of the external environment which is different depending on the individual difference of the vehicle and the sales place of the vehicle is absorbed. As a result, the nozzle control unit 106 can control the supercharging pressure of the air compressed by the supercharger 2 with high accuracy.

  The supercharger 2 is provided to improve the output of the engine 1. The relationship between the nozzle throttle amount and the supercharging pressure obtained by the above calibration can be said to be a master model for determining a theoretical value for maintaining the initial performance of the vehicle.

  The master model is stored in the storage unit 105 in the control unit 10. Although not shown, in the master model, the relationship between the nozzle throttle amount N in the nozzle-equipped turbine 15 and the supercharging pressure P at that time is stored in the storage unit 105 in a table format. The nozzle control unit 106 refers to the table stored in the storage unit 105 to obtain the nozzle throttle amount N for the compressed air pressure output from the supercharger 2 to be the target supercharging pressure P. can do.

  Hereinafter, in this specification, the relationship between the nozzle throttle amount N and the supercharging pressure P determined by calibration after the vehicle is produced may be referred to as a “master model”. Further, the supercharging pressure calculated according to the master model may be described as “theoretical value Pt” or “supercharging pressure Pt” of the supercharging pressure. Furthermore, the pressure of the air actually output from the supercharger 2 may be described as “supercharging pressure Pr”.

  Due to the calibration described above, all vehicles have similar initial performance. However, generally, the performance of the supercharger 2 gradually changes as the vehicle is used by the user.

  Specifically, the nozzle-equipped turbine 15 of the supercharger 2 is driven by the kinetic energy of the exhaust gas discharged from the engine 1. Since the exhaust gas is a gas generated by combustion of a mixture of fuel and air, soot may be included. When the bag or the like sticks to the nozzle-equipped turbine 15 as the vehicle is used, the smooth movement of the nozzle-equipped turbine 15 may be impaired. As a result, even if the command amount instructed by the nozzle control unit 106 is the same, the nozzle throttle amount N may be larger than expected and the exhaust gas flow rate may be slower than expected. If such a thing happens, the rotation speed of the turbine 15 with a nozzle will become smaller than expected, and as a result, the supercharging pressure of the air which the supercharger 2 outputs will fall rather than assumption. On the other hand, when the nozzle throttle amount N is reduced, the rotational speed of the nozzle-equipped turbine 15 is increased even if the instruction amount indicated by the nozzle control unit 106 is the same, and the supercharging of the air output from the supercharger 2 is performed. The pressure rises.

It is considered that the amount of soot contained in the exhaust gas varies depending on the fuel used for combustion, and also varies depending on how the engine 1 is operated, that is, how the fuel is burned.
As described above, the relationship between the nozzle throttle amount N and the supercharging pressure P gradually deviates from the master model in accordance with the aging of the vehicle, and the degree of the divergence also varies depending on the use situation of each vehicle. As a result, the initial performance of the vehicle may not be maintained.

  In order to cope with such a situation, the initial performance of the vehicle is set sufficiently high with respect to the performance standard to be satisfied by the vehicle. More specifically, it is designed so that the standard can be satisfied even if the initial performance of the vehicle cannot be exhibited. However, such a response can be a factor in increasing the cost of the vehicle.

Therefore, the learning unit 102 according to the embodiment calculates a function that determines the relationship between the supercharging pressure Pt that is the theoretical value in the master model and the supercharging pressure Pr that is the air pressure that is actually output from the supercharger 2. And remember. Based on the function calculated by the learning unit 102, the nozzle control unit 106 changes the nozzle throttle amount N so that the actual supercharging pressure Pr by the supercharger 2 becomes the theoretical value Pt according to the master model. . Thereby, even if the relationship between the nozzle throttle amount N and the supercharging pressure P deviates from the master model due to factors such as aging of the vehicle, the storage unit 105 can obtain the supercharging pressure according to the master model. .
Hereinafter, learning performed by the learning unit 102 according to the embodiment will be described in more detail.

<Learning flow>
FIG. 2 is a flowchart for explaining a flow of learning processing executed by the learning unit 102 according to the embodiment. The processing in this flowchart starts when a vehicle on which the engine 1 is mounted is started, for example.

  The determination unit 101 determines whether or not the state of the vehicle on which the engine 1 is mounted satisfies a predetermined learning execution condition for the learning unit 102 to perform learning related to the supercharging pressure (S2). The “predetermined learning execution condition” is a condition for the learning result by the learning unit 102 to be a reliable result. Specifically, it is a condition for the vehicle to be in a stable state.

The determination unit 101 determines that the learning execution condition is satisfied when at least all of the conditions listed below are satisfied.
(Condition 1) The outside air temperature is within a predetermined temperature range.
If the temperature of the outside air measured by an outside air temperature sensor (not shown) is less than a preset lower limit value or exceeds an upper limit value, the learning unit 102 does not learn. This is because the reading value of the pressurizing pressure sensor 14 is affected when the density of the air greatly changes depending on the temperature.

(Condition 2) The temperature of the exhaust gas is equal to or higher than a predetermined threshold temperature.
If the temperature of the exhaust gas measured by the exhaust gas temperature sensor (not shown) is below a preset threshold temperature, the learning unit 102 does not learn. This is because when the temperature of the exhaust gas is low, the exhaust density is also low, so that the driving force of the nozzle-equipped turbine 15 is affected.

(Condition 3) The recirculation valve 3 is closed.
During the EGR operation, the learning unit 102 does not learn. This is because the nozzle-equipped turbine 15 is driven by exhaust gas, so that the driving force of the nozzle-equipped turbine 15 is affected during the EGR operation.

(Condition 4) The fluctuation of the rotational speed of the engine 1 is within a predetermined rotational speed fluctuation range.
When the engine 1 is hunting and the fluctuation of the engine speed is large, the amount of intake air is not stable and learning is not performed. In addition, since the specific value of the rotation speed fluctuation range varies depending on the engine 1, it may be determined by experiment in consideration of these performances and the like, but is within a range assumed in normal use of the vehicle.

(Condition 5) The variation in the injection amount of the fuel injected into the cylinder of the engine 1 is within a predetermined injection amount range.
When the fuel injection amount changes extremely, it is considered that the fluctuation of the engine speed increases. In this case as well, learning is not performed because the amount of intake air is not stable. In addition, since the specific value of the injection amount fluctuation range varies depending on the engine 1, it may be determined by an experiment in consideration of these performances and the like, but is within a range assumed in normal use of the vehicle.

(Condition 6) The vehicle is adjusting the amount of air supply by feedback control based on the amount of exhaust and the amount of air supply.

  While it is determined that the state of the vehicle does not satisfy the learning execution condition (No in S4), the determination unit 101 returns to step S2 and continues the determination of the learning execution condition. When the determination unit 101 determines that the state of the vehicle satisfies the learning execution condition (Yes in S4), the learning unit 102 determines the nozzles of the nozzle 15 with the nozzle for controlling the rotation speed of the nozzle 15 with the nozzle. An aperture amount N is set (S6), and the nozzle controller 106 controls the aperture amount N of the nozzle of the turbine 15 with nozzle.

  The learning unit 102 acquires the supercharging pressure measured by the boosting pressure sensor 14, that is, the supercharging pressure Pr that is the air pressure actually output from the supercharger 2 (S <b> 8), and stores it in the storage unit 105. The learning unit 102 reads the master model stored in the storage unit 105, acquires the theoretical value Pt of the supercharging pressure at the set nozzle throttle amount N (S10), and stores it in association with the supercharging pressure Pr. Stored in the unit 105. A combination of the theoretical value Pt and the actual supercharging pressure Pr becomes the learning data of the learning unit 102. If the actual supercharging pressure Pr and the theoretical value Pt are regarded as the X coordinate and the Y coordinate in the two-dimensional orthogonal coordinate system, respectively, one combination of the theoretical value Pt and the actual supercharging pressure Pr is two-dimensional orthogonal coordinates. One point in the system.

  Although details will be described later, the learning unit 102 according to the embodiment approximates the relationship between the theoretical value Pt of the supercharging pressure and the actual supercharging pressure Pr using a cubic polynomial. Since the cubic polynomial is uniquely determined by four coefficients, the learning unit 102 can specify the cubic polynomial by changing the set value of the nozzle aperture amount N and acquiring at least four pieces of learning data.

  If the learning data is not ready (No in S12), the learning unit 102 returns to step S6 and continues the learning data acquisition process from step S6 to step S10. When all the learning data are obtained (Yes in S12), the learning unit 102 determines a coefficient of a cubic polynomial representing the relationship between the theoretical value Pt of the supercharging pressure and the actual supercharging pressure Pr (S14), and the storage unit 105. (S16). When the learning unit 102 finishes determining the coefficient of the third-order polynomial, the processing in this flowchart ends.

  As an example, the learning unit 102 is in a state in which the nozzle of the turbine 15 with nozzle is completely closed, that is, a state where the nozzle opening is 0%, and a state where the nozzle is fully open, that is, a state where the nozzle opening is 100%. The nozzle aperture amount N is set so as to increase by 10% until. Further, the learning unit 102 acquires 10 pieces of learning data for the aperture amount N of the same nozzle. In this case, the learning unit 102 acquires 110 points of learning data.

  FIG. 3 is a diagram schematically illustrating a graph in which learning data acquired by the learning unit 102 according to the embodiment is plotted in a two-dimensional orthogonal coordinate system. More specifically, FIG. 3 shows the actual supercharging pressure Pr of the supercharger 2 measured by the boosting pressure sensor 14 as the X axis, and the theoretical value Pt of the supercharging pressure calculated from the master model by the learning unit 102 as Y. A graph plotted as an axis is shown. In FIG. 3, a curve indicated by a symbol F is a cubic function, which is an approximate curve F representing the relationship between the theoretical value Pt of the boost pressure and the actual boost pressure Pr. When the theoretical value Pt of the supercharging pressure coincides with the actual supercharging pressure Pr, the approximate curve F is a straight line having an inclination of 45 degrees.

Assume that the approximate curve F can be expressed by the following equation (1) using four coefficients A, B, C, and D.
Y (X) = AX ³ + BX ² + CX + D (1)
Here, X is the actual supercharging pressure Pr, and Y represents the theoretical value Pt of the supercharging pressure. The learning unit 102 calculates four coefficients that specify the approximate curve F from the acquired learning data. Hereinafter, the coefficient calculation process executed by the learning unit 102 will be described.

<Coefficient calculation processing of approximate curve F>
The learning unit 102 calculates a coefficient of the approximate curve F representing the acquired learning data by using the least square method.
Now, let Xn be the actual value of the supercharging pressure Pr, and Yn be the theoretical value Pt of the supercharging pressure corresponding to Xn. n is a natural number of 1 or more and N or less, and is an index uniquely determined for each of N points of learning data. In the above example, since the learning unit 102 acquires 110 points of learning data, n is any integer value between 1 and 110. Substituting Xn and Yn into equation (1) yields the following equation (2).

When equation (2) is rewritten using a matrix, the following equation (3) is obtained.

  In Expression (3), the left side is a column vector in which the theoretical values Pt of the supercharging pressure are arranged, and is known. The first term on the right side is a matrix that is configured using the actual boost pressure Pr measured by the boost pressure sensor 14 and is known. The second term on the right side is a column vector in which four coefficients specifying the approximate curve F are arranged, and is unknown. The problem of obtaining A, B, C, and D from Equation (3) is a dominant decision problem where the number of unknowns is 4 and the number of equations is 110.

When the column vector in which the theoretical values Pt of the supercharging pressures are arranged is T, the matrix constituted by using the actual supercharging pressure Pr is R, and the column vector in which the four coefficients are arranged is m, the equation (3) is Equation (4) is obtained.
T = Rm (4)
The vector m can also be said to be a model parameter for explaining the theoretical value Pt of the supercharging pressure using the actual supercharging pressure Pr.

If m is set appropriately, the error between T and Rm becomes small. Therefore, a column vector e in which T, Rm, and errors are arranged is defined by the following equation (5).
e = T-Rm (5)
The vector e can be said to be a modeling error of a model for explaining the theoretical value Pt of the supercharging pressure with the actual supercharging pressure Pr.

ここで、「Ｔ」はベクトル又は行列の転置を表し、「−１」は逆行列を表す。ベクトル及び行列の要素を用いて式（６）を書き下すと以下の式（７）を得る。

The optimal solution m _opt in the sense of minimizing the 2-norm square of the error vector e shown in Equation (5), that is, e ^T e is known as a least-square solution, and is obtained by the following Equation (6). It is done.

Here, “T” represents transposition of a vector or matrix, and “−1” represents an inverse matrix. When formula (6) is written down using the elements of the vector and matrix, the following formula (7) is obtained.

  By using Expression (7), the learning unit 102 plots the actual supercharging pressure Pr in a rectangular coordinate system having the first axis (X axis) and the theoretical value Pt as the second axis (Y axis). In this case, a coefficient of a cubic polynomial that approximates the relationship between the theoretical value Pt and the actual supercharging pressure Pr can be calculated. The learning unit 102 stores and stores the four coefficients A, B, C, and D calculated based on Expression (7) in the storage unit 105.

<Correction of boost pressure using approximate curve F>
FIG. 4 is a diagram for explaining the nozzle aperture amount correction processing executed by the nozzle control unit 106 according to the embodiment. 4 is a straight line represented by Y = X, and is a graph when the actual supercharging pressure Pr by the supercharger 2 coincides with the theoretical value Pt according to the master model. Hereinafter, for convenience of explanation, a broken line indicated by a symbol L may be referred to as a “theoretical straight line L”.

  In the example shown in FIG. 4, the approximate curve F is above the theoretical line L while the Y-axis value, that is, the theoretical value Pt according to the master model is from 0 to Pe. This indicates that the theoretical value Pt according to the master model is higher than the actual supercharging pressure Pr by the supercharger 2. Therefore, even if the nozzle control unit 106 sets the nozzle throttle amount N so as to output the air of the theoretical value Pt according to the master model, the actual supercharging pressure Pr by the supercharger 2 is the theoretical value Pt according to the master model. Means lower. Therefore, the region where the value of the supercharging pressure Pt in FIG. 4 is from 0 to Pe can be said to be a “supercharging pressure insufficient region”.

  On the other hand, in the region where the value of the theoretical value Pt is equal to or greater than Pe, the approximate curve F is below the theoretical line L. This indicates that the actual supercharging pressure Pr by the supercharger 2 is larger than the theoretical value Pt according to the master model. Therefore, the region where the supercharging pressure Pt is equal to or higher than Pe can be said to be a “supercharging pressure excessive region”.

  When the target theoretical value Pt is included in the excessive boost pressure region, the nozzle control unit 106 changes the nozzle throttle amount N so that the rotational speed of the nozzle-equipped turbine 15 is reduced. On the other hand, when the target theoretical value Pt is included in the supercharging pressure insufficient region, the nozzle control unit 106 changes the nozzle throttle amount N so that the rotational speed of the nozzle-equipped turbine 15 increases.

  For example, in the example shown in FIG. 4, when the theoretical value Pt that is the target of the boost pressure is P1, P1 is included in the excessive boost pressure region. Therefore, the actual supercharging pressure P2 corresponding to the theoretical value P1 in the approximate curve F shown in FIG. 4 is larger than P1 by ΔP. That is, when the nozzle controller 106 sets the nozzle throttle amount N so that the supercharging pressure becomes P1 according to the master model, the actual supercharging pressure P2 by the supercharger 2 is only ΔP with respect to the theoretical value P2. Become more.

  Therefore, the nozzle control unit 106 adjusts the nozzle amount N so that the rotational speed of the nozzle-equipped turbine 15 is reduced in order to set the actual supercharging pressure Pr to P1. More specifically, as shown in FIG. 4, the theoretical value Pt when the actual supercharging pressure Pr is P1 in the approximate curve F is P3 (<P1). The nozzle throttle amount N is set so that the supercharging pressure according to the model is P3.

  The learning unit 102 updates the coefficient of the approximate curve F by learning each time the determination unit 101 determines that the state of the vehicle satisfies the learning execution condition. Accordingly, even if the relationship between the nozzle throttle amount N and the supercharging pressure P deviates from the master model due to factors such as aging of the vehicle, the nozzle control unit 106 supercharges the supercharging pressure air according to the master model. The machine 2 can output. As a result, the initial performance of the vehicle can be maintained.

<Effect of Embodiment>
As described above, according to the learning device according to the embodiment of the present invention, it is possible to provide a technique for maintaining the accuracy of the supercharging pressure in the air output from the supercharger 2.
In particular, the nozzle control unit 106 obtains the relationship between the result obtained by the learning of the learning unit 102, that is, the theoretical value Pt of the supercharging pressure according to the master model and the supercharging pressure Pr of the air actually output from the supercharger 2. Since the throttle amount of the nozzle of the turbine with nozzle 15 is controlled according to the approximate curve F that is expressed, the initial performance of the supercharger 2 can be maintained.

  In the learning device according to the embodiment of the present invention, a state in which the vehicle is stabilized is set as a learning execution condition in advance, and the learning unit 102 executes learning when the vehicle state satisfies the learning execution condition. For this reason, the learning unit 102 can accurately obtain the relationship between the theoretical value Pt of the supercharging pressure and the actual supercharging pressure Pr, and the correction of the throttle amount of the nozzle 15 of the nozzle-equipped turbine 15 after learning is also accurate. .

  Furthermore, in the learning device according to the embodiment of the present invention, the learning unit 102 updates the coefficient that determines the approximate curve F every time learning is performed. For this reason, the time-dependent change of the supercharger 2, the difference in the fuel which a user uses can be absorbed.

  As mentioned above, although this invention was demonstrated using embodiment, the technical scope of this invention is not limited to the range as described in the said embodiment. It will be apparent to those skilled in the art that various modifications or improvements can be added to the above embodiment. It is apparent from the scope of the claims that the embodiments added with such changes or improvements can be included in the technical scope of the present invention. Such modifications will be described below.

<First Modification>
The case has been described above where the learning unit 102 updates the coefficient of the approximate curve F by learning each time the determination unit 101 determines that the vehicle state satisfies the learning execution condition. However, the learning execution condition by the learning unit 102 is not limited to the above (condition 1) to (condition 6), and further conditions may be added. For example, in addition to (Condition 1) to (Condition 6), a condition related to the travel distance of the vehicle may be added. The determination unit 101 also executes a determination process regarding the added condition.

  In this case, the learning unit 102 learns the state of the vehicle according to the above (Condition 1) to (Condition 6) after the travel distance traveled by the vehicle after the previous learning is performed becomes longer than a predetermined distance. When the execution condition is satisfied, the coefficient of the function that determines the approximate curve F may be recalculated and stored in the storage unit 105. The travel distance of the vehicle may be measured by providing a distance measurement unit (not shown).

  The predetermined distance may be determined in consideration of an assumed usage mode of the vehicle, and is, for example, 500 km. Thereby, it can suppress that the learning part 102 recalculates the coefficient of the approximated curve F many times within a short period.

  Further, the learning unit 102 may add a condition related to the travel time of the vehicle instead of or in addition to the condition related to the travel distance of the vehicle. In this case, the learning unit 102 also sets the time for which the engine 1 of the vehicle has been operating since the previous learning to be longer than a predetermined time (for example, one month) as a condition for executing the learning. Thereby, it can suppress that the learning part 102 performs recalculation of the coefficient of the approximated curve F many times within a short period.

<Second Modification>
The case where the learning unit 102 stores the coefficient of the approximate curve F calculated by learning in the storage unit 105 has been described above. Here, the learning unit 102 may leave the coefficient stored in the storage unit 105 as time-series data in association with the date and time when the calculation is performed without overwriting the coefficient. Since the approximate curve F is information that directly indicates the performance of the vehicle, the time variation of the performance of the vehicle can be monitored by storing the coefficient of the approximate curve F as time series data.

<Third Modification>
The case where the approximate curve F is approximated by a cubic polynomial has been described above. However, the order of the polynomial to be approximated is not limited to the third order, and may be the Kth order (K is an integer of 3 or more). The order of the polynomial to be approximated may be determined by experiment in consideration of the shape of the graph representing the relationship between the actual supercharging pressure Pr and the theoretical value Pt, the processing performance of the learning unit 102, and the like.

DESCRIPTION OF SYMBOLS 1 ... Engine 2 ... Supercharger 3 ... Circulation adjustment valve 4 ... Supply pipe 5 ... Supply throttle 6 ... Exhaust pipe 7 ... Exhaust throttle 8 ... Duct 9 ... MAF sensor 10 ... Control unit 12 ... Injector 13 ... Rotational speed measurement unit 14 ... Pressure supply pressure sensor 15 ... Turb with nozzle 16 ... Differential pressure sensor 101 ... Determination Unit 102 ... learning unit 105 ... storage unit 106 ... nozzle control unit 107 ... opening degree control unit

Claims (5)

An engine that generates driving force for the vehicle by mixing and burning fuel with the supply air; and
A turbocharger that is driven by exhaust from the engine and compresses the supply air to increase the supply air pressure;
A supercharging pressure sensor for measuring the supercharging pressure of the air compressed by the supercharger;
A nozzle control unit for controlling a throttle amount of a nozzle for adjusting a flow rate of exhaust gas for driving the supercharger;
A determination unit that determines whether or not a state of the vehicle satisfies a predetermined learning condition for learning about the supercharging pressure;
When the determination unit determines that the state of the vehicle satisfies the learning condition, the theoretical value Pt of the supercharging pressure at the throttle amount N of the nozzle and the actual overpressure measured by the supercharging pressure sensor. A learning unit that calculates and stores a function that determines a relationship with the supply pressure Pr;
A learning apparatus comprising: The nozzle control unit changes the throttle amount N of the nozzle based on the function calculated by the learning unit so that the actual supercharging pressure Pr becomes the theoretical value Pt.
The learning apparatus according to claim 1. A distance measuring unit for measuring the travel distance of the vehicle;
The learning unit recalculates and stores the function when the learning condition is satisfied after the mileage traveled by the vehicle after the previous learning is performed is longer than a predetermined distance,
The learning apparatus according to claim 1, wherein the learning apparatus is a learning apparatus. The learning unit plots the theoretical value Pt and the actual supercharging pressure in a rectangular coordinate system having the actual supercharging pressure Pr as a first axis and the theoretical value Pt as a second axis. Calculating a coefficient of a K-degree polynomial (K is an integer of 3 or more) approximating the relationship with Pr;
The learning device according to claim 1, wherein the learning device is a learning device. The learning conditions are: (1) the outside air temperature is within a predetermined temperature range, (2) the exhaust gas temperature is equal to or higher than a predetermined threshold temperature, (3) the recirculation control valve is closed, and (4) fluctuations in the engine speed. Is within a predetermined rotational speed variation range, (5) the variation in the injection amount of fuel injected into the cylinder of the engine is within a predetermined injection amount range, and (6) the vehicle has the exhaust amount and the air supply The amount of the air supply is adjusted by feedback control based on the amount of
The learning device according to claim 1, wherein the learning device is a learning device.

Images