JP2020084851A - Fuel injection control device of internal combustion engine - Google Patents

Abstract

To provide a fuel injection control device of an internal combustion engine which can perform minute injection amount learning while sufficiently securing learning accuracy even if the internal combustion engine is in an operation state.SOLUTION: An ECU 50 comprises: a learning period setting part for setting a learning execution period for learning a fuel injection amount of an injector 33 within an in-cylinder pressure stabilization period in which a pressure variation in a cylinder 32 in one combustion cycle is smaller than a prescribed in-cylinder pressure variation amount; an injection control part for performing fuel injection by the injector 33 different from main injection during the learning execution period which is set by the learning period setting part; an injection amount calculation part for calculating an actual injection amount by the fuel injection which is performed by the injection control part; and an injection amount learning part for learning a fuel injection amount on the basis of the actual injection amount calculated by the injection amount calculation part, and a drive pulse when the injector 33 is valve-open driven by the injection control part.SELECTED DRAWING: Figure 1

Description

The present invention relates to a fuel injection control device for an internal combustion engine.

BACKGROUND ART Conventionally, as a fuel injection control device for an engine, various techniques related to injection amount learning have been proposed in order to correct a deviation in a relationship between a drive pulse for injecting a desired amount of fuel from a fuel injection valve and an actual fuel injection amount. (For example, see Patent Document 1). Patent Document 1 discloses that fuel pressure feeding from a fuel pump is stopped at the time of deceleration of a vehicle, fuel injection is performed, and the injection amount is estimated from the amount of decrease in rail pressure before and after the injection.

Further, for example, in a diesel engine, multi-stage injection is performed in which fuel injection is performed a plurality of times per combustion cycle in order to improve ignitability and exhaust emission.

US Pat. No. 7,788,015

In order to sufficiently obtain the above-mentioned improvement effect by the multi-stage injection, it is important to precisely adjust the fuel injection other than the main injection, that is, the fuel injection amount by the minute injection. Further, in order to perform the fuel injection control with high accuracy, it is desirable to complete the injection amount learning as soon as possible after the ignition is turned on. However, in the technique disclosed in Patent Document 1, fuel pressure feeding from the fuel pump is stopped, fuel injection is performed, and the injection amount is estimated based on the rail pressure change amount before and after the injection. It is conceivable that the injection amount learning can be carried out only during the specified period. In this case, there is concern that sufficient learning opportunities cannot be secured.

The present invention has been made in view of the above problems, and a main object thereof is to carry out the small injection amount learning while sufficiently securing the learning accuracy even when the internal combustion engine is in the operating state. An object of the present invention is to provide a fuel injection control device for an internal combustion engine that can achieve the above.

Hereinafter, the means for solving the above problems, and the operation and effect thereof will be described.

The present invention relates to a fuel pump (20) for pressurizing and discharging fuel, a pressure accumulator (14) for accumulating high pressure fuel discharged from the fuel pump, and an internal combustion engine () for accumulating high pressure fuel stored in the pressure accumulator. The fuel injection system includes a fuel injection valve (33) for directly injecting into the cylinder (32) of (30).

The invention according to claim 1 is a learning execution period for learning the fuel injection amount of the fuel injection valve during an in-cylinder pressure stable period in which the pressure fluctuation in the cylinder is smaller than a predetermined in-cylinder pressure fluctuation amount in one combustion cycle. And a fuel injection by the fuel injection valve, in addition to the main injection that contributes to the generation of the output torque of the internal combustion engine, in the learning period setting unit that sets An injection control unit, an injection amount calculation unit that calculates the actual injection amount actually injected from the fuel injection valve by the fuel injection performed by the injection control unit, and an actual injection amount calculated by the injection amount calculation unit, An injection amount learning unit that learns the fuel injection amount of the fuel injection valve based on a drive pulse when the fuel injection valve is driven to open by the injection control unit.

Paying attention to the fact that the actual injection amount for the same drive pulse fluctuates due to the fluctuation in the cylinder pressure, and that there is a period in which the fluctuation in the cylinder pressure is small within one combustion cycle. The configuration is such that the injection amount learning of the minute injection amount is performed during the in-cylinder pressure stable period in which the in-cylinder pressure fluctuation is smaller than a predetermined in-cylinder pressure fluctuation amount. With such a configuration, even when the internal combustion engine is operating, it is possible to carry out the small injection amount learning while ensuring sufficient learning accuracy.

The schematic block diagram of a fuel injection system. The figure which shows the schematic structure of an injector. The longitudinal cross-sectional view which expanded the front-end|tip part of an injector. The figure which shows the change characteristic of the injection amount with respect to cylinder pressure change. The figure which shows the change of the in-cylinder pressure in 1 combustion cycle. The figure which compares the in-cylinder pressure in a different crank angle position in 1 combustion cycle. The time chart which shows the setting method of the study implementation period. The flowchart which shows the process sequence of a small injection amount learning process. The flowchart which shows the process sequence of learning precondition satisfaction determination processing. The flowchart which shows the process sequence of the data sampling determination process for learning. The flowchart which shows the process sequence of a learning implementation period calculation process. The flowchart which shows the process sequence of an injection implementation point calculation process. The time chart which shows the concrete mode of small amount injection learning. The flowchart which shows the process sequence of a minute injection amount estimation process. FIG. 6 is a diagram illustrating learning of a minute injection amount based on a change amount of rail pressure. The flowchart which shows the process sequence of a shift amount calculation process. The figure which shows the injector injection characteristic in the injection amount range of minute injection. The flowchart which shows the process sequence of a learning value storage process. The time chart which shows the setting method of the study implementation period of other embodiment.

Hereinafter, embodiments will be described with reference to the drawings. This embodiment builds a fuel injection system that supplies fuel from a fuel injection valve to a vehicle-mounted engine as an internal combustion engine. The engine is a multi-cylinder engine, and in this embodiment, a 4-cylinder engine is installed. This system implements various controls with an electronic control unit (hereinafter referred to as an ECU) as the center. FIG. 1 shows an overall schematic configuration diagram of the fuel injection system. In each of the following embodiments, the same or equivalent portions are denoted by the same reference numerals in the drawings, and the description of the portions having the same reference numeral is cited.

In the fuel injection system 60 of FIG. 1, the fuel tank 11 is connected to the fuel pump 20 via the low pressure fuel pipe 12. The fuel pump 20 includes a low-pressure pump 21 that pumps fuel from the fuel tank 11 through the low-pressure fuel pipe 12, and a high-pressure pump 22 that boosts the pressure of the fuel pumped by the low-pressure pump 21. Although the low-pressure pump 21 and the high-pressure pump 22 are integrated in this embodiment, they may be provided separately.

The high-pressure pump 22 is a mechanical pump that sucks and discharges fuel as the engine 30 rotates. Specifically, the high-pressure pump 22 sucks low-pressure fuel to increase its pressure by reciprocating the plunger in synchronization with the rotation of the output shaft 31 (crankshaft) of the engine 30, and the high-pressure fuel (high-pressure fuel ) Is discharged. The fuel discharge portion of the high-pressure pump 22 is provided with a discharge metering valve (PCV) 23 as an electromagnetically driven fuel metering valve, and the discharge start timing is adjusted by controlling the energization of the discharge metering valve 23. Thus, the amount of fuel discharged from the fuel pump 20 (pump discharge amount) is adjusted.

The high-pressure pump 22 includes two metering valves, a first PCV and a second PCV, as the discharge metering valves 23. Of the two discharge adjustment valves 23, the first PCV controls the fuel discharge to the first cylinder #1 and the fourth cylinder #4, and the second PCV controls the second cylinder #2 and the third cylinder #3. Control the discharge of fuel into the. The present system is a one-injection, one-pressure feed fuel supply system in which fuel is pumped by the fuel pump 20 every time fuel injection is performed. As the fuel metering valve, instead of the discharge metering valve 23, a suction metering valve (SCV) arranged in the fuel suction part of the high-pressure pump 22 may be used.

A common rail 14 as a pressure accumulator is connected to the fuel pump 20 via a high-pressure fuel pipe 13. The high pressure fuel discharged from the fuel pump 20 is sequentially supplied to the common rail 14, and the fuel is held in a high pressure state. The common rail 14 is provided with a rail pressure sensor 15 as a pressure detecting means, and the rail pressure sensor 15 detects the fuel pressure (rail pressure) in the common rail 14.

The engine 30 is a multi-cylinder diesel engine. The engine 30 is provided with an injector 33 for each cylinder 32. The high-pressure fuel in the common rail 14 is supplied to each injector 33 through the high-pressure fuel pipe 16. By driving the injector 33, high-pressure fuel is directly injected into each cylinder 32 of the engine 30. A reflux pipe 17 is connected to the fuel pump 20 and the injector 33, and excess fuel in the fuel pump 20 and the injector 33 is returned to the fuel tank 11 via the reflux pipe 17.

As shown in FIG. 2, the injector 33 includes a tubular nozzle body 34 and a nozzle needle 35 that is housed in the internal space of the nozzle body 34 so as to be reciprocally movable in the axial direction. An injection hole 44 that is opened and closed by the nozzle needle 35 is provided at the tip of the nozzle body 34. The injection hole 44 communicates with the high-pressure fuel pipe 16 (see FIG. 1) via the fuel passage, and stores the high-pressure fuel in the middle of the fuel passage that connects the high-pressure fuel pipe 16 and the injection hole 44. A storage chamber 36 is provided. The tip of the nozzle needle 35 is arranged in the storage chamber 36. A back pressure chamber 37 connected to the high-pressure fuel pipe 16 is provided at an end of the nozzle needle 35 opposite to the storage chamber 36. High pressure fuel from the common rail 14 is introduced into the back pressure chamber 37.

The back pressure chamber 37 is connected to the low pressure fuel pipe 24 communicating with the fuel tank 11 via an orifice 39. The low-pressure fuel pipe 24 is provided downstream of the orifice 39 with a normally-closed control valve 42 that opens when the coil 41 is energized. When the coil 41 is not energized, the sum of the force in the valve closing direction due to the fuel pressure in the back pressure chamber 37 and the urging force of the spring 38 becomes larger than the fuel pressure in the storage chamber 36, so that the nozzle needle 35 is displaced in the direction of closing the injection hole 44, and the injector 33 is closed. In the valve closed state, when the coil 41 is energized and the control valve 42 is opened, the fuel in the back pressure chamber 37 flows out to the low pressure fuel pipe 24 side through the orifice 39. As a result, the fuel pressure in the storage chamber 36 becomes larger than the total force of the fuel pressure in the back pressure chamber 37 and the biasing force of the spring 38, so that the nozzle needle 35 opens the injection hole 44. The injector 33 is displaced and the injector 33 is opened. Further, when the injector 33 is opened, the high pressure fuel in the storage chamber 36 is injected from the injection hole 44.

FIG. 3 is an enlarged vertical cross-sectional view of the tip portion of the injector 33. A conical taper portion 48 and a sack portion 46 formed on the tip side of the taper portion 48 are provided at the tip portion inside the nozzle body 34. The taper portion 48 has a valve seat portion 43 on which the tip of the nozzle needle 35 is seated. Further, the nozzle needle 35 includes a conical needle body 35a and a tapered portion 35b provided on the tip side of the needle body 35a. An annular seal portion 45 is provided at an axially intermediate portion of the taper portion 35b, and the seal portion 45 can be seated on the valve seat portion 43.

The sack portion 46 is, for example, a space having a spherical inner wall surface. The sack portion 46 communicates with the storage chamber 36 (see FIG. 2) via an annular fuel passage 47 formed between the inner surface of the nozzle body 34 and the side surface of the nozzle needle 35. As a result, the high-pressure fuel in the storage chamber 36 flows into the sack portion 46 through the fuel passage 47 and is stored therein. A plurality of injection holes 44 are formed at the tip of the nozzle body 34 so as to connect the inner surface of the sack portion 46 and the outer surface of the nozzle body 34. When the nozzle needle 35 moves in the direction toward the injection hole 44 and the seal portion 45 of the nozzle needle 35 is seated on the valve seat portion 43, the inflow of fuel to the sack portion 46 is blocked, and the injector 33 is The valve is closed. In this valve closed state, when the nozzle needle 35 moves in the direction away from the injection hole 44 and the seal portion 45 separates from the valve seat portion 43, the high pressure fuel flows into the suck portion 46 through the fuel passage 47, Fuel is injected from the injection holes 44.

The ECU 50 is an electronic control device including a well-known microcomputer including a CPU, a ROM, a RAM, and the like. In addition to the detection signal of the rail pressure sensor 15 described above, the ECU 50 includes a crank angle sensor 51 for detecting the rotation speed of the engine 30, a cylinder discrimination sensor 52 for discriminating which stroke each cylinder is in, and an accelerator. Detection signals are sequentially input from various sensors such as an accelerator sensor 53 that detects an operation amount, a water temperature sensor 54 that detects an engine cooling water temperature, and a fuel temperature sensor 55 that detects a temperature of fuel in the fuel pump 20.

The ECU 50 determines an optimum fuel injection amount and injection timing based on engine operating information such as engine rotation speed and accelerator operation amount, and outputs an injection control signal according to the determined fuel injection amount and injection magnetism to the injector 33. .. As a result, the fuel injection control by the injector 33 is performed in each cylinder 32. Further, the ECU 50 sets a target rail pressure that is a target value of the rail pressure based on the engine rotation speed and the fuel injection amount at each time, and the actual rail pressure detected by the rail pressure sensor 15 becomes the target rail pressure. Control the pump discharge rate. By controlling the rail pressure, the injection pressure of the fuel injected from the injector 33 is controlled.

In the fuel injection control of this system, multi-stage injection is performed. Specifically, prior to the main injection (main injection) of fuel that contributes to the generation of the output torque of the engine 30, a small amount compared to the main injection is used. (For example, 2 to 3 mm3/st) of fuel is injected. In this embodiment, one pilot injection and one main injection are performed per combustion cycle (intake stroke→compression stroke→expansion stroke→exhaust stroke).

In the pilot injection, it is required to inject a small amount of fuel with high precision in order to sufficiently exert effects such as reduction of combustion noise and suppression of NOx generation. On the other hand, the injector 33 has a command injection amount to the injector 33, that is, an energization pulse period and a fuel injection amount that is actually injected from the injector 33 during the energization pulse period (due to individual differences, deterioration over time, and the like). A deviation appears from the actual injection amount). Therefore, in this system, the deviation amount between the command value (driving pulse) corresponding to the minute injection amount (instruction injection amount Qtarget) to be injected by the pilot injection and the actual injection amount is detected, and the fuel amount injected from the injector 33 is corrected. The small injection amount learning is performed.

Hereinafter, the small injection amount learning performed by this system will be described in detail. In this system, when the cylinder pressure of the engine 30 (hereinafter, also referred to as “cylinder pressure”) fluctuates, the actual injection amount differs even if the same drive pulse command is given to the injector 33, and In the case of minute injection, we paid attention to the fact that the in-cylinder pressure has a large effect and the learning accuracy has a large effect.

FIG. 4 shows the change characteristic of the injection amount with respect to the change in the cylinder pressure. In FIG. 4, (a) shows the in-cylinder pressure [MPa] and the actual injection amount [when the fuel injection is commanded by the drive pulse of the same length so that the fuel of the designated injection amount Qtarget is injected from the injector 33. [mm3/st]. (B) is a diagram showing the relationship between the in-cylinder pressure and the actual injection amount according to time and injection rate. “P1” and “P2” represent in-cylinder pressure values, and P1<P2. The dashed-dotted line in FIG. 4B shows the injection characteristic when the in-cylinder pressure is P1, and the solid line shows the injection characteristic when the in-cylinder pressure is P2. Further, the dotted line indicates the injection characteristic when the in-cylinder pressure is P2 so that the injection start timing t1 of the injection characteristic when the in-cylinder pressure is P1 and the injection start timing t2 of the injection characteristic when the in-cylinder pressure is P2 match. Is a slide in the time axis direction.

As shown in FIGS. 4A and 4B, even when the instructed injection amount is the same value Qtarget, when the in-cylinder pressure is P1, the actual injection amount is Q1, but the in-cylinder pressure is higher. In the case of P2, the actual injection amount becomes Q2, which is larger than Q1. That is, the higher the in-cylinder pressure, the larger the actual injection amount from the injector 33 for the same drive pulse. In such an event, the higher the in-cylinder pressure, the more the nozzle needle 35 is likely to be positioned on the rising side, which causes the nozzle needle 35 to start moving in the valve opening direction (that is, the injection start timing) and the nozzle needle 35 in the valve closing direction. It is considered that this is due to the timing of ending the movement to (i.e., the injection end timing) and the moving speed of the nozzle needle 35 depending on the in-cylinder pressure.

Therefore, in the present embodiment, in the state where the engine 30 is operating, in a period in which the fluctuation of the in-cylinder pressure is smaller than a predetermined in-cylinder pressure fluctuation amount in one combustion cycle, separately from the normal injection (pilot injection and main injection). , Fuel injection for learning (learning injection) is implemented. Then, the actual injection amount by the learning injection is calculated, and the fuel injection amount of the injector 33 is learned using the calculated actual injection amount and the driving pulse for the learning injection. According to such small injection amount learning, it is possible to carry out the small injection amount learning while ensuring sufficient learning accuracy even during normal operation when the engine 30 is burning.

FIG. 5 shows changes in the in-cylinder pressure in one combustion cycle when the normal injection (one pilot injection and main injection) is performed in the first cylinder #1. The horizontal axis represents the crank angle with reference to the compression top dead center. In one combustion cycle, the in-cylinder pressure is stable in the latter half of the expansion stroke and the exhaust stroke, and in period C in the latter half of the expansion stroke and exhaust stroke, period A before compression top dead center and period B after compression top dead center Compared with, the amount of change in the cylinder pressure is small and the cylinder pressure is stable (see FIGS. 5 and 6). Focusing on this point, in the present embodiment, a period TA after the compression top dead center θa (for example, 70 to 80° C. A after the compression top dead center) in one combustion cycle is used, and within the period TA. A learning implementation period for the small injection amount learning (hereinafter, also referred to as “learning implementation period TP”) is set, and the small injection amount learning is implemented in the set learning implementation period TP.

It should be noted that the in-cylinder pressure stable period in which the fluctuation of the in-cylinder pressure is small is also in the period of the intake stroke and the first half of the compression stroke in one combustion cycle, but in order to avoid changing the mode of the normal injection by implementing the learning injection, In the embodiment, the learning execution period TP is set in the latter half of the compression stroke and the exhaust stroke after the main injection is completed.

A specific mode of the setting method of the learning execution period TP of the small injection amount learning of the present embodiment will be described with reference to the time chart of FIG. 7. This small injection amount learning is performed in a state in which the engine 30 is operating, more specifically, in an engine operating state in which fuel discharge from the fuel pump 20 and main injection by the injector 33 are performed in one combustion cycle. To be done.

In the present embodiment, in the period TA, the period in which the fuel discharge of the high-pressure pump 22 and the fuel injection of the injector 33 are not performed (hereinafter, also referred to as “drive stop period”) has a predetermined fuel pressure in which the rail pressure varies. It is assumed that the period is smaller than the variation amount, and the minute injection amount learning is performed during the drive stop period of the period TA. This is to avoid variations in the rail pressure due to variations in the pump pressure feed amount and fuel injection amount, and variations in the injection amount of the learning injection due to variations in the rail pressure. Further, in the present embodiment, since the injection amount learning is performed based on the relationship between the drive pulse of the injector 33 and the change amount of the rail pressure, it is possible to prevent the detection value of the rail pressure from varying and to detect the rail pressure. This is to ensure sufficient accuracy.

In FIG. 7, “TB” indicates a period during which fuel is being discharged from the fuel pump 20 in response to a drive command to the discharge metering valve 23 (first PCV and second PCV), and “TC” is one of four cylinders. The period during which normal injection is performed by the drive command to the injector 33 of the cylinder is shown. “J1” indicates pilot injection, and “J2” indicates main injection.

The learning implementation period TP is set for each cylinder in a period that does not overlap with any of the periods TB and TC of the period TA of each cylinder (hereinafter, also referred to as “learning data sampling possible period TS”). That is, in the period TA, the learning execution period TP is set in a period in which fuel is not being injected from the injector 33 and fuel is not being discharged from the high-pressure pump 22. In addition, when there are a plurality of learning data sampling possible periods TS in each cylinder, the ECU 50 performs learning of one of the plurality of learning data sampling possible periods TS according to the following conditions (a) and (b). Select as the period TP.
(A) The longest period (longest sampling period) is selected as the learning implementation period TP from the plurality of learning data sampleable periods TS.
(B) When there are a plurality of longest sampling periods, the longest sampling period on the most advanced side is selected as the learning implementation period TP.

In FIG. 7, in the first cylinder #1, three learning data sampling possible periods TS (period t(#1-1), period t(#1-2), and period t(#1) are included in one combustion cycle. -3)) is present. The period t(#1-1) and the period t(#1-3) have the same length. The ECU 50 first selects the period t(#1-1) and the period t(#1-3) as the longest sampling period from the three learning data sampling possible periods TS. Subsequently, of the period t(#1-1) and the period t(#1-3) selected as the longest sampling period, the period t(#1-1) on the most advanced side is set as the learning execution period TP. To do.

For the second cylinder #2 to the fourth cylinder #4, a period having the same positional relationship as the first cylinder #1 is set as the learning execution period TP. This prevents the learning execution periods TP of the cylinders from overlapping each other. In FIG. 7, the period closest to the advance angle side is set as the learning implementation period TP.

Specifically, for the second cylinder #2, three learning data sampling available periods TS (period t(#2-1), period t(#2-2) and period t(#2-3)). Of these, the period t (#2-1) is set as the learning implementation period TP. For the third cylinder #3, of the period t (#3-1), the period t (#3-2), and the period t (#3-3), the period t (#3-1) is the learning execution period TP. For the fourth cylinder #4, of the period t(#4-1), the period t(#4-2) and the period t(#4-3), the period t(#4-1) is It is set as a learning implementation period TP. Then, for each cylinder, the learning injection is performed in the learning execution period TP in at least one combustion cycle after the next time, and the learning value is acquired based on the actual injection amount and the drive pulse. The learning execution period TP may be set for the second cylinder #2 to the fourth cylinder #4 by performing the same processing as the first cylinder #1.

The reason for providing the condition (b) is that the present system includes two injector drive circuits, and one injector drive circuit controls the drive of the two injectors 33. Therefore, in each cylinder, the crank angle period in which the drive pulse can be output to the injector 33 is predetermined. Specifically, the crank angle period (for example, a period of 180° C. A) from the crank angle position θp on the advance side of the compression top dead center to the crank angle position θq on the retard side of the compression top dead center is It is set as a period during which a drive pulse can be output to the injector 33 of each cylinder. Therefore, in the present embodiment, the learning injection can be reliably performed by setting the learning data sampling possible period TS on the advance side as much as possible as the learning execution period TP.

Next, the processing procedure of the small injection amount learning of the present embodiment will be described with reference to FIGS. First, the entire flowchart will be described with reference to FIG. This processing is executed by the microcomputer of the ECU 50 in a predetermined cycle.

In FIG. 8, in step S100, a process for determining whether or not a learning precondition, which is a precondition for implementing the small injection amount learning while the engine 30 is operating, is determined (learning precondition satisfaction determination process). carry out. In the following step S200, a process (learning data sampling determination process) of calculating the learning data sampling possible period TS as a period in which the learning data for learning the small injection amount can be acquired is executed. In step S300, a process of calculating the learning execution period TP (learning execution period calculation process) is executed, and in step S400, a process of determining the implementation point of the minute injection amount is executed.

In a succeeding step S500, a process for performing learning injection and estimating a minute injection amount (a minute injection amount estimation process) is performed. In step S600, a process (deviation amount calculation process) of calculating the deviation amount of the drive pulse is performed using the minute injection amount estimated in the process of step S500. In step S700, it is determined whether or not a predetermined number of deviation amounts have been acquired by performing the learning injection a plurality of times. When it is determined that the predetermined number of deviation amounts have not been acquired, the process of step S500 and the process of step S600 are repeatedly performed until an affirmative determination is made in step S700. When an affirmative decision is made in step S700, the operation proceeds to step S800, the learning value of the minute injection amount is calculated, the learning value is stored and stored in the storage section (learning value storage processing), and this routine ends.

Next, each process of the overall flowchart of FIG. 8 will be described. First, the learning precondition satisfaction determination processing in step S100 will be described with reference to FIG. The learning precondition satisfaction determination process is a process for determining whether or not the operating state of the engine 30 is stable.

In FIG. 9, in step S101, the data necessary for determining the success or failure of the learning precondition is read. Specifically, the engine rotation speed Ne detected by the crank angle sensor 51, the actual rail pressure Pc detected by the rail pressure sensor 15, the engine water temperature detected by the water temperature sensor 54, and the fuel detected by the fuel temperature sensor 55. The temperature, the accelerator opening detected by the accelerator sensor 53, the injection amount instruction value QFIN, and the rail pressure instruction value PFIN are respectively acquired.

In the following step S102, the amount of change ΔX of each data acquired in step S101 per unit time is calculated. Here, for each data, the change amount ΔX is calculated from the difference between the previous-previous value X(i-2) and the previous value X(i-1) stored in the storage unit. In a succeeding step S103, it is determined whether or not the change amount ΔX is equal to or less than the threshold value Xth for each data. At this time, if ΔX>Xth for at least one of the data, the process returns to step S102, and the processes of steps S102 and S103 are repeated until a positive determination is made in step S103. Then, when ΔX≦Xth for all the data, an affirmative decision is made in step S103, and this routine is ended.

Next, the learning data sampling determination process of step S200 of FIG. 8 will be described with reference to the flowchart of FIG. In FIG. 10, in step S201, the crank angle detected by the crank angle sensor 51, the fuel pressure feed flag F1 and the normal injection flag F2 are read. Here, the fuel pressure feed flag F1 is a flag that is turned on during the period when fuel is being discharged from the fuel pump 20 by the drive command to the discharge metering valve 23 and is turned off during the period when fuel is not being discharged from the fuel pump 20. Yes (see FIG. 7). The normal injection flag F2 is a flag that is turned on during the period when fuel injection by the pilot injection and the main injection is performed by the drive command to the injector 33, and turned off during the period when fuel is not injected from the injector 33 (FIG. 7).

In the following step S202, for each cylinder, in the period TA after the compression top dead center θa in one combustion cycle of the engine 30, the fuel pressure feeding flag F1 is off and the normal injection flag F2 is off. Or not. When the fuel pressure feed flag F1 is off and the normal injection flag F2 is off during the period TA after θa after the compression top dead center and the normal injection flag F2 is off, the process proceeds to step S203, and the learning data sampling flag F3. Is turned on (see FIG. 7). The learning data sampling flag F3 is a flag indicating whether or not to execute the process of acquiring the learning data for learning the small injection amount, and is a flag set for each cylinder (see FIG. 7). In step S203, it is determined whether the setting process of the learning data sampling flag F3 is completed (that is, whether one combustion cycle has elapsed), and it is determined that the setting process of the learning data sampling flag F3 is completed. When this is done, this routine ends.

Next, the learning execution period calculation process of step S300 of FIG. 8 will be described with reference to the flowchart of FIG. In FIG. 11, in step S301, the length of the learning data sampling available period TS is calculated for each cylinder. Here, the length (° C. A) of the period during which the learning data sampling flag F3 is on is calculated for each cylinder. When there are a plurality of periods in which the learning data sampling flag F3 is on in one combustion cycle, the period length (°C A) is calculated for each of the plurality of periods. For example, in FIG. 7, the period t(#k-1), the period t(#k-2), and the period t(#k-3) (where k is an integer of 1 to 4 representing the cylinder number). ), the period length (°C A) is calculated for each of the three periods.

In the following step S302, the longest sampling period is selected from the learning data sampling available period TS of the specific cylinder for the specific cylinder (first cylinder #1 in the present embodiment) among the plurality of cylinders. Here, when there is one learning data sampleable period TS, that period is selected as the longest sampling period. On the other hand, when there are a plurality of learning data sampling possible periods TS, the period having the longest period length is selected as the longest sampling period from the plurality of learning data sampling possible periods TS. At this time, if there are a plurality of learning data sampleable periods TS having the same length and the period length is the longest in the acquired learning data sampleable periods TS, that is, if there are a plurality of longest sampling periods. , A plurality of learning data sampling available periods TS are selected as the longest sampling period.

In a succeeding step S303, it is determined whether or not there are a plurality of longest sampling periods in the specific cylinder. When there are no plural longest sampling periods, the process proceeds to step S305, and the learning data sampling possible period selected as the longest sampling period is set as the learning implementation period TP. On the other hand, when there are a plurality of longest sampling periods, the process proceeds to step S304, and the period on the most advanced side among the plurality of longest sampling periods is set as the learning implementation period TP.

In the following step S306, the same position as that of the first cylinder #1 in the learning data sampling possible period is set for each of the remaining cylinders (second cylinder #2 to fourth cylinder #4 in this embodiment) other than the specific cylinder. The related period is set as the learning implementation period TP. Then, this process ends.

Next, the injection execution point calculation processing of step S400 of FIG. 8 will be described using the flowchart of FIG. In FIG. 12, in step S401, a period length LA (unit: °C A) that is the length of the learning execution period TP is calculated for the specific cylinder (first cylinder #1 in the present embodiment). In the following step S402, the unit of the period length LA is converted into time, and the period length tA (unit: microsecond) is calculated. In step S403, the number of injections n by which the minute injection can be performed within the learning execution period TP is calculated using the period length tA.

Here, the injection amount of the minute injection is an extremely small amount that does not contribute to the output torque, and if the minute injection amount learning is to be performed by only one injection, the injector 33 will inject the learning fuel. However, there is a concern that the amount of change in the parameter for calculating the learning value (rail pressure in the present embodiment) is small and learning accuracy cannot be sufficiently secured. Therefore, in the present embodiment, by setting the longest sampling period as the learning execution period TP, the length of the learning execution period TP is sufficiently secured, and the minute injection (single shot injection) is executed a plurality of times within the learning execution period TP. As a result, the amount of change in rail pressure due to learning injection is amplified. As a result, the learning accuracy of the small injection amount learning is improved.

In a succeeding step S404, the information necessary for executing the minute injection dedicated to learning (learning injection information) is set in the ECU 50 for each cylinder based on the period length tA and the injection number n. The learned injection information includes the number of injections n calculated in step S403, the injection period TU, the injection interval TI, the instruction injection amount Qtarget in a minute injection per one time, the drive pulse TQ, the injection timing, and the like. The command injection amount Qtarget corresponds to a value within the injection amount range of a minute amount used in pilot injection, and a fixed value of, for example, 1 to 3 mm3/st is set. After setting the learning injection information in the ECU 50, the present processing is ended.

A specific embodiment of the small injection amount learning of this embodiment will be described with reference to the time chart of FIG. FIG. 13 shows one learning execution period TP of the first cylinder #1.

In FIG. 13, in the first cylinder #1, in one combustion cycle, from the first crank angle position θ10 (for example, 75 to 85° C. A after compression top dead center) to the second crank angle position θ11 (for example, compression upper extension). The learning data sampling flag F3 is turned on during the period from 170 to 180° C.). In this case, first, the length of the learning implementation period TP is converted from the crank angle into time (LA [degCA]→tA [μs]). Then, the number of injections n is calculated based on the period length tA, the injection period TU (for example, 400 to 500 microseconds), and the injection interval TI (for example, 1800 to 2300 microseconds). The number of injections n is the maximum number of injections permitted by the conditions of the period length tA of the learning execution period TP, the injection period TU, and the injection interval TI. In this embodiment, the injection period TU and the injection interval TI are predetermined constant values. The injection period TU is a fuel injection time corresponding to the fuel injection amount of the pilot injection, and corresponds to a minute injection amount such that torque is not output from the engine 30. Therefore, the longer the learning implementation period TP, the larger the number of injections of the minute injection is set.

Next, the minute injection amount estimation process of step S500 of FIG. 8 will be described with reference to the flowchart of FIG. In FIG. 14, in step S501, the amount of change in rail pressure for each cylinder in the period length tA in the state where fuel injection is not performed is measured, and the measured value is stored as the reference amount of change ΔP1. Here, the reference change amount ΔP1 of each cylinder is measured according to a predetermined combustion order (for example, #1→#3→#4→#2).

In the following step S502, micro injection is performed n times in each learning execution period TP based on the learning injection information set in the injection execution procedure calculation process of FIG. 12 in each cylinder. In step S503, the amount of change in the rail pressure before and after the micro injection is performed n times is measured, and the measured value is stored as the injection change amount ΔP2.

In a succeeding step S504, the pressure change amount ΔPinj by the nth minute injections is calculated using the reference change amount ΔP1 and the injection time change amount ΔP2. Here, the injection change amount ΔP2 is subtracted from the reference change amount ΔP1, and the value after the subtraction (ΔP1−ΔP2) is set as the pressure change amount ΔPinj. In step S505, the total injection amount Qtotal by n small injections during the learning execution period TP is calculated. The total injection amount Qtotal is estimated using the pressure change amount ΔPinj, the bulk elastic coefficient K, and the high pressure part volume Vsys.

In step S506, the total injection amount Qtotal calculated in step S505 is divided by the number of injections n to obtain the actual injection amount Qinj (=Qtotal/n) as the injection amount per one injection by n minute injections. calculate. The processes of steps S502 to S506 are performed for each cylinder. In step S507, it is determined whether or not the calculation of the actual injection amount Qinj of all the cylinders is completed. If the calculation is not completed, the processes of steps S502 to S506 are executed for the incomplete cylinder. When the calculation of the actual injection amount Qinj of all cylinders is completed, an affirmative decision is made in step S507, and this processing ends.

The small injection amount learning using the change amount of the rail pressure will be described with reference to FIG. In FIG. 15, (a) shows the reference change amount ΔP1, and (b) shows the injection change amount ΔP2. In FIG. 15, it is assumed that a period of length LA1 is set as the learning implementation period TP. The reference change amount ΔP1 is a minute leak amount (static leak amount) in which the fuel flows out to the low-pressure fuel passage side through a gap or the like of the injector 33 when the injector 33 is not driven. In the present embodiment, the learning accuracy of the minute injection amount learning is improved by calculating the actual injection amount Qinj per one minute injection using the value obtained by subtracting the decrease in rail pressure due to the static leak amount. I have to.

Next, the shift amount calculation process of step S600 of FIG. 8 will be described with reference to the flowchart of FIG. In FIG. 16, in step S601, the difference ΔQ(j) (ΔQ(j)=Qtarget−Qinj) between the instruction injection amount Qtarget per n small injections (learning injections) and the actual injection amount Qinj. To calculate. The calculation of ΔQ(j) is performed for each cylinder. j is an integer of 1 to 4 representing the cylinder number.

In a succeeding step S602, the difference ΔQ(j) is used to calculate the deviation amount ΔTQ(j) of the drive pulse in the injection characteristic representing the relationship between the injection amount and the drive pulse for each injector 33 of each cylinder.

FIG. 17 shows the injection characteristic of the injector 33 at a predetermined rail pressure in the injection amount range of the minute injection. In FIG. 17, the horizontal axis represents the injection amount and the vertical axis represents the drive pulse. The solid line shows the master characteristic, and the broken line shows the actual characteristic. In FIG. 17, the injection amount and the drive pulse are in a proportional relationship with the inclination α (α>0), and the drive pulse may be made longer to increase the injection amount. When the actual characteristic deviates from the master characteristic and the ejection amount for the same drive pulse deviates by a difference ΔQ(j), the drive pulse deviates by ΔTQ(j). Therefore, in the deviation amount calculation process, the deviation amount ΔTQ(j) (=ΔQ(j)/α) is calculated from the difference ΔQ(j) using the inclination α of the injection characteristic corresponding to the current rail pressure. After the calculation of the deviation amount ΔTQ(j) of all the cylinders is completed, this processing is ended.

Finally, the learning value storage process of step S800 of FIG. 8 will be described with reference to the flowchart of FIG. 18, in step S801, the small injection amount learning is performed for each cylinder a predetermined number of times (Y times) to obtain the deviation amount ΔTQ(j) Y times, and the average value ΔTQave(j) thereof is acquired for each cylinder. Calculate to. In subsequent step S802, the average value ΔTQave(j) is stored and stored in the storage unit as a learning value for the small injection amount learning, and the present processing is ended.

According to the embodiment described in detail above, the following excellent effects can be obtained.

During a period in which the fluctuation of the in-cylinder pressure is smaller than a predetermined in-cylinder pressure fluctuation amount in one combustion cycle (in-cylinder pressure stable period), more specifically, in the latter half of the expansion stroke of one combustion cycle, injection of a small injection amount The configuration is such that quantity learning is carried out. According to this configuration, even when the engine 30 is in operation, it is possible to carry out the small injection amount learning while ensuring sufficient learning accuracy.

A configuration in which the learning execution period TP is set in a period in which fuel is not being injected from the injector 33 and fuel is not being discharged from the high-pressure pump 22 in the in-cylinder pressure stable period (that is, the learning data sampling possible period TS). And During the period when at least one of the injector 33 and the high-pressure pump 22 is driven, it is conceivable that the rail pressure may fluctuate due to fluctuations in the pump pressure feed amount and fuel injection amount. Therefore, if the fuel injection with the small injection amount learning is performed while at least one of the injector 33 and the high-pressure pump 22 is being driven, the dispersion of the actual injection amount and the dispersion of the detected value of the rail pressure with respect to the same drive pulse are large. Therefore, there is a concern that learning accuracy will decrease. By paying attention to this point and adopting the above-mentioned configuration, even if the normal injection by the injector 33 and the fuel discharge from the fuel pump 20 are not bothered to be stopped in order to implement the small injection amount learning, the learning accuracy is ensured and the small amount is maintained. Injection quantity learning can be implemented.

When there are a plurality of learning data sampleable periods TS in one combustion cycle, the longest period of the plurality of learning data sampleable periods TS is set as the learning execution period TP. According to this configuration, the fuel injection amount for the small injection amount learning can be performed as much as possible within one learning execution period TP to amplify the variation amount of the rail pressure, and the learning accuracy can be further improved. ..

In the learning implementation period TP, the injector 33 is configured to inject a small amount of fuel corresponding to pilot injection a plurality of times. In microinjection such as pilot injection, there is a concern that the fuel injection amount per injection is extremely small and learning accuracy cannot be sufficiently secured. In this respect, according to the above configuration, by performing the fuel injection corresponding to the minute injection a plurality of times, the amount of change in the rail pressure with respect to the learning injection can be amplified, and the learning accuracy can be ensured.

The fuel injection is performed the number of times corresponding to the length of the learning execution period TP. Specifically, the longer the learning implementation period TP, the larger the number of injections of the minute injection. With such a configuration, it is possible to inject as much fuel as possible within the learning execution period TP, and it is possible to amplify changes in the rail pressure as much as possible. Thereby, the learning accuracy can be further improved.

The learning implementation period TP is set to a period after the latter half of the expansion stroke in one combustion cycle. According to this configuration, the minute injection amount learning period is performed after the combustion of the fuel for torque generation is completed and the in-cylinder pressure is stable, so that the engine 30 is operating. Also, the variation in the injection amount is small, and the learning accuracy can be sufficiently secured. Further, since the fuel injection for learning the small injection amount is carried out after the combustion ends, the execution of the normal fuel injection (pilot injection and main injection) is hindered, or the mode of the normal fuel injection (injection timing etc.) It is possible to carry out the small injection amount learning while avoiding changing or the like. At this time, in order not to affect the normal fuel injection in the next combustion cycle, in the period after the latter half of the expansion stroke in one combustion cycle, as much as possible on the advance side, in order to learn the small injection amount. It is desirable to carry out the fuel injection.

In the learning execution period TP, the fuel injection dedicated to learning is performed separately from the normal injection to learn the minute injection amount. According to this configuration, since the learning injection is performed separately from the normal injection, it is possible to perform the small injection amount learning without changing the mode of the normal injection.

(Other embodiments)
The present invention is not limited to the contents of the above embodiment, and may be implemented as follows, for example.

In the above-described embodiment, the case of applying to the fuel supply system of one-injection/one-pressure delivery including the four-cylinder engine has been described as an example. However, the number of cylinders of the engine 30 and the fuel discharge of the high-pressure pump 22 in one combustion cycle The relationship between the number of times and the number of times of the injector 33 is not limited to this. For example, the present invention may be applied to a fuel supply system in which two-injection one-pressure feed or one-injection two-pressure feed in a four-cylinder engine, or three-injection two-pressure feed or three-injection four-pressure feed in a three-cylinder engine. The present invention may be applied to a system. FIG. 19 shows a time chart in the case of being applied to a fuel supply system of a 4-cylinder engine and 2-injection/1-pressure feeding. The description of the symbols in FIG. 19 is the same as in FIG. 7. The injector 33 shown in the above embodiment is an example, and the injector 33 is not limited to the configuration of the above embodiment.

In the above embodiment, the latter half of the expansion stroke and the exhaust stroke of one combustion cycle are learned by regarding the latter half of the expansion stroke and the exhaust stroke as a cylinder pressure stable period in which the fluctuation of the cylinder pressure is smaller than a predetermined cylinder pressure fluctuation amount. Although the implementation period TP is set, an in-cylinder pressure sensor for detecting the in-cylinder pressure is provided in each cylinder, and the period when the detected value of the in-cylinder pressure sensor becomes equal to or less than the threshold value after the end of the main injection is learned as the in-cylinder pressure stable period. The period TP may be set.

In the above embodiment, the learning execution period TP is set for the latter half of the expansion stroke and the exhaust stroke with the latter half of the expansion stroke and the exhaust stroke of one combustion cycle as the in-cylinder pressure stabilization period, but at least one of the first half of the intake stroke and the first half of the compression stroke is set. As the in-cylinder pressure stable period (see FIG. 5), the learning execution period TP may be set in at least one of the intake stroke and the first half of the compression stroke.

The learning execution period TP may be set within the exhaust stroke, and the minute injection amount learning may be performed during the exhaust stroke. Since the fuel injection is performed after the main injection is completed, there is a concern that the injected fuel adheres to the inner wall of the cylinder and the adhered fuel flows down and mixes into the oil pan. Therefore, in order to avoid mixing of fuel with oil, the minute injection amount learning may be performed in the exhaust stroke.

In the above embodiment, the learning implementation period TP is the longest period, but a period at a predetermined position may be set as the learning implementation period TP. For example, when there are a plurality of learning data sampleable periods TS, a period located on the most advanced side may be set as the learning implementation period TP among the plurality of learning data sampleable periods TS.

In the above-described embodiment, the minute injection amount learning is performed by detecting the amount of change in the rail pressure with respect to the learning injection, but the amount of change in the rail pressure is not limited to other values that correlate with the amount of fuel injection from the injector 33. The configuration may be such that the small injection amount learning is performed using parameters. Other parameters include, for example, the air-fuel ratio of exhaust gas (λ value), the exhaust gas temperature, and the like.

In the above embodiment, as a specific example of the multi-stage injection, the case where the injection amount of the pilot injection is learned as the minute injection amount learning by applying to the fuel injection system that performs the pilot injection and the main injection has been described. The mode and the minute injection to be learned are not limited to this. For example, the present invention is applied to a fuel injection system that performs at least one of pre-injection, after-injection, and post-injection with or without performing pilot injection as minute injection, and includes pre-injection, after-injection, and post-injection. The configuration may be such that the minute injection amount learning is performed for at least one of them.

The rail pressure drops when a small amount of fuel is injected, so if the small amount of fuel is injected a plurality of times with the same drive pulse length, the injection amount may vary from time to time. In order to reduce such variations, during the learning injection in the learning execution period TP, the rail pressure immediately before the injection of each injection time is detected, and the length of the drive pulse for the injector 33 is detected based on the detected rail pressure. The configuration may be such that the height is variably set. With this configuration, it is possible to carry out the highly accurate learning of the minute injection amount. At this time, the drive pulse may be lengthened as the detected rail pressure decreases.

An upper limit may be set for the number of minute injections during the learning implementation period TP. With such a configuration, it is possible to reduce the influence on fuel economy. In this case, the upper limit number of times may be set to the number of times that the total injection amount for sufficiently ensuring the learning accuracy can be achieved. Alternatively, a fixed value may be set as the number of injections of minute injections performed in the learning execution period TP.

A configuration in which the injection amount learning value is calculated by subtracting the pressure change amount ΔPdl corresponding to the dynamic leak amount, which is the fuel leak amount due to the valve opening of the injector 33, from the pressure change amount ΔPinj due to the nth minute injection calculated above. May be

-Instead of performing the learning injection, a minute injection amount of the normal injection (for example, post injection) may be used to perform the minute injection amount learning. In this case, the post-injection may be performed by the divided injection to amplify the parameter (for example, rail pressure) for calculating the learning value, which changes with the minute injection.

-Each component described above is a conceptual one, and is not limited to the above embodiment. For example, the function of one constituent element may be distributed to a plurality of constituent elements, or the function of a plurality of constituent elements may be realized by one constituent element.

14... Common rail (accumulation container), 20... Fuel pump, 22... High pressure pump, 30... Engine (internal combustion engine), 33... Injector (fuel injection valve), 50... ECU (learning period setting unit, injection control unit, injection amount) Calculation unit, injection amount learning unit, injection number setting unit), 60... Fuel injection system.

Claims (6)

A fuel pump (20) for pressurizing and discharging fuel, a pressure accumulator (14) for accumulating high-pressure fuel discharged from the fuel pump, and a cylinder of an internal combustion engine (30) for accumulating high-pressure fuel stored in the pressure accumulator. And a fuel injection valve (33) for directly injecting the fuel into the fuel injection system (32).
A learning period setting unit that sets a learning implementation period for learning the fuel injection amount of the fuel injection valve in a cylinder pressure stable period in which the pressure fluctuation in the cylinder is smaller than a predetermined cylinder pressure fluctuation amount in one combustion cycle;
In the learning implementation period set by the learning period setting unit, an injection control unit that performs fuel injection by the fuel injection valve separately from the main injection that contributes to the generation of output torque of the internal combustion engine,
An injection amount calculation unit that calculates an actual injection amount actually injected from the fuel injection valve by the fuel injection performed by the injection control unit,
An injection amount for learning the fuel injection amount of the fuel injection valve based on the actual injection amount calculated by the injection amount calculation unit and a drive pulse when the fuel injection valve is driven to open by the injection control unit. Learning part,
And a fuel injection control device for an internal combustion engine. The learning period setting unit sets the learning execution period in a drive stop period in which no fuel is injected from the fuel injection valve and no fuel is discharged from the fuel pump, among the in-cylinder pressure stable period. A fuel injection control device for an internal combustion engine according to claim 1. The internal combustion engine according to claim 2, wherein the learning period setting unit sets the longest period among the plurality of drive stop periods to the learning execution period when a plurality of drive stop periods exist in one combustion cycle. Engine fuel injection control device. It is applied to a fuel injection system that performs control to inject fuel in multiple stages within one combustion cycle,
The injection control unit injects a minute amount of fuel corresponding to a fuel injection different from the main injection among the multi-stage injections from the fuel injection valve a plurality of times during the learning execution period. 2. A fuel injection control device for an internal combustion engine according to claim 1. According to the length of the learning implementation period set by the learning period setting unit, further comprises an injection number setting unit for setting the number of injections from the fuel injection valve by the injection control unit,
The fuel injection control device for an internal combustion engine according to claim 4, wherein the injection control unit performs the fuel injection of the injection number set by the injection number setting unit during the learning execution period. The fuel for an internal combustion engine according to claim 1, wherein the learning period setting unit sets the learning execution period in a period that is at least one of a latter half of an expansion stroke and an exhaust stroke in one combustion cycle. Injection control device.

Images