JP2010261334A - Fuel injection control device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a fuel injection control device reducing the number of learning injections to the minimum level. <P>SOLUTION: When learning conditions of the injection amount are established, the fuel injection control device commands a fuel injection valve to perform a single injection for learning the injection amount (S400), and senses the change amount of the engine speed as the change amount of the engine operation state by the single injection (S402). The fuel injection control device calculates an actual injection amount Qn, based on the engine speed. It also calculates the average value Qaven of the actual injection amount, and furthermore, it calculates the standard deviation σn as the variance of the actual injection amount so far (S404). The fuel injection control device counts up the number of learning injections (S412). If the standard deviation σn of the actual injection amount is in a range of a target accuracy (S414: Yes) and the number of learning injections is the same as or more than the minimum times of injections (S416: Yes), and a correction amount of a command injection amount of a single injection is calculated based on the deviation of the average value Qaven of the actual injection amount and the target injection amount (S418). <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a fuel injection control device that learns an injection amount of a fuel injection valve.

  Conventionally, the deviation of the actual injection amount from the target injection amount of the fuel injection valve caused by machine differences or changes over time is learned, and the injection command value commanded to the fuel injection valve is corrected so that the actual injection amount becomes the target injection amount (For example, refer to Patent Document 1). In particular, in the case of a diesel engine that performs a small amount of pilot injection before main injection in order to reduce NOx and combustion noise, injection amount learning for correcting the minute injection amount with high accuracy is required. The actual injection amount is estimated based on a change in the operating state of the internal combustion engine when the learning injection is executed, for example, a change in the engine speed or the air-fuel ratio.

  Incidentally, in addition to the injection amount variation between the fuel injection valves due to machine differences and changes over time, as shown in FIG. 7, each fuel injection valve has injection amount variation for each injection and combustion variation for each injection. . In FIG. 7, reference numeral 300 indicates an upper limit value of the injection amount variation for each injection, and reference numeral 310 indicates an upper limit value of the estimated injection amount variation caused by the combustion variation. Reference numeral 302 represents a lower limit value of the injection amount variation for each injection, and reference numeral 312 represents a lower limit value of the estimated injection amount variation caused by the combustion variation.

The injection amount variation for each injection is caused by, for example, variation in fuel pressure during injection. When the injection amount varies, the actual injection amount estimated based on the change in the operating state of the internal combustion engine varies.
Further, the combustion variation for each injection occurs due to variations in fuel injection timing, spray shape, supercharging pressure, etc. in each injection. When combustion variation occurs, the operating state of the internal combustion engine varies even when the same fuel amount is injected, and therefore the actual injection amount estimated based on the change in the operating state of the internal combustion engine varies.

  As described above, since the estimated value of the actual injection amount varies for each injection in each fuel injection valve, one learning injection is executed, and the actual injection amount estimated at that time is compared with the target injection amount. For example, the representative value may deviate greatly from the average value of the actual injection amounts when the learning injection is executed a plurality of times. Based on such representative values, highly accurate injection amount learning cannot be executed.

  Therefore, as disclosed in Patent Document 1, a plurality of learning injections are executed for each injection amount learning, and, for example, an average value is calculated from the actual injection amounts estimated in the plurality of learning injections. Therefore, it is desirable to use a representative value of the actual injection amount.

  As the number of injections increases, the variation of the actual injection amount in each fuel injection valve becomes within the target accuracy range, and the representative value obtained from the average value converges to a constant value. Therefore, the representative of the actual injection amount in each fuel injection valve The value can be obtained with high accuracy.

  Thereby, based on the representative value of the actual injection amount and the target injection amount, the injection command value commanded to the fuel injection valve can be corrected with high accuracy so that the actual injection amount becomes the target injection amount. As a result, it is possible to eliminate the deviation between the target injection amount and the actual injection amount caused by machine differences and changes with time.

JP 2005-146947 A

  By the way, in each fuel injection valve, as described above, the estimated value of the actual injection amount varies due to the injection amount variation for each injection and the combustion variation for each injection. The range of variation in the estimated value of the actual injection amount is different for each fuel injection valve. Therefore, the number of injections when the variation in the actual injection amount is within the target accuracy range is also different for each fuel injection valve.

  Therefore, it is conceivable to set the number of learning injections in accordance with the maximum variation of the actual injection amount so that the variation of the actual injection amount falls within the target accuracy range for all the fuel injection valves.

However, when the number of learning injections is set in accordance with the maximum variation, there is a problem that the time required for injection amount learning becomes long.
The present invention has been made to solve the above problem, and an object of the present invention is to provide a fuel injection control device that reduces the number of injections of learning injection as much as possible.

  According to the first to fourth aspects of the present invention, the injection amount estimating means is based on a change in the operating state of the internal combustion engine when the single injection for learning the injection amount is performed from the fuel injection valve by the command of the injection command means. The actual injection amount of the single injection is estimated, and the variation calculating means sequentially calculates the variation of the actual injection amount of the single injection based on the actual injection amount estimated by the injection amount estimating means in the single injection executed a plurality of times. The learning end determination unit determines the number of injections when the learning injection is ended based on the variation in the actual injection amount calculated by the variation calculation unit and the target accuracy of the variation in the actual injection amount.

  Thereby, the number of injections when the learning injection is terminated is determined according to the range of variation in the actual injection amount in each fuel injection valve. That is, if the variation in the actual injection amount is small, the number of injections for learning the injection amount is reduced. As a result, since the number of injections for learning the injection amount can be reduced as much as possible, the period required for learning the injection amount can be shortened. Therefore, for example, in a vehicle, the distance traveled from the start to the end of injection amount learning can be reduced as much as possible.

  Thereby, for example, when the next injection amount learning is executed at a predetermined travel distance after the previous injection amount learning is completed in the vehicle, the next injection amount learning is started after the previous injection amount learning is started. Since the distance until this is shortened, the injection amount learning is performed more frequently during a certain travel distance. As a result, the period during which the injection amount of the fuel injection valve is controlled with the actual injection amount deviating from the target injection amount can be shortened as much as possible.

  Since the time required to complete the injection amount learning for all the cylinders becomes longer as the number of cylinders increases, an internal combustion engine having a larger number of cylinders is required to shorten the execution period of the injection amount learning.

  According to the second aspect of the present invention, the learning end determination means has the minimum injection required for calculating the variation in the variation in the actual injection amount within the target accuracy range and the number of injections performed for the injection amount learning. If the number is greater than or equal to the number of times, it is determined that the learning injection is terminated.

  In this way, the minimum number of injections necessary to calculate the variation in the actual injection amount is executed, and when the learning injection is terminated when the variation in the actual injection amount is within the target accuracy range, the learning end determination is made. Since the means determines, the representative value of the actual injection amount for comparison with the target injection amount can be calculated with high accuracy while reducing the number of injections for injection amount learning as much as possible.

  By the way, when the actual injection amount is estimated from the change in the operating state of the internal combustion engine, the variation in the actual injection amount is within the target accuracy range in the second and third learning injections, for example, even though the variation in the actual injection amount is large. It may fit in by chance. If the representative value of the actual injection amount for comparison with the target injection amount is calculated based on such an actual injection amount, the representative value may not be calculated with high accuracy.

  Therefore, according to the third aspect of the present invention, the learning end determination unit determines that the injection command is used when the number of injections executed for learning the injection amount is less than the minimum number of injections necessary for calculating the variation in the actual injection amount. Let the means continue the learning injection.

  As described above, by executing the injection more than the minimum number of times necessary for calculating the variation in the actual injection amount, it is possible to prevent the variation in the actual injection amount from accidentally falling within the target accuracy range.

  According to the invention of claim 4, the representative value calculating means calculates the representative value of the actual injection quantity based on the actual injection quantity of the single injection executed a plurality of times, and the correction amount calculating means is the variation of the actual injection quantity. When the learning end determination unit determines that the learning injection is finished within the target accuracy range, the fuel injection valve is based on the deviation between the representative value calculated by the representative value calculation unit and the target injection amount of the actual injection amount. The correction amount of the injection command value for commanding the fuel injection of the target injection amount is calculated.

Thereby, the correction amount of the injection command value can be calculated with high accuracy while reducing the number of injections of the learning injection as much as possible.
The functions of the plurality of means provided in the present invention are realized by hardware resources whose functions are specified by the configuration itself, hardware resources whose functions are specified by a program, or a combination thereof. The functions of the plurality of means are not limited to those realized by hardware resources that are physically independent of each other.

The block diagram which shows the fuel-injection system by this embodiment. The characteristic view which shows an injection amount learning process. The characteristic view which shows the relationship between the frequency | count of injection and the estimated value, dispersion | variation, and average value of actual injection amount. The characteristic view which shows the relationship between the frequency | count of injection by the comparative example, the estimated value of an actual injection quantity, dispersion | variation, and an average value. The characteristic view which shows the relationship between the frequency | count of injection, and the dispersion | variation in an actual injection amount. The flowchart which shows the injection quantity learning routine. FIG. 5 is a distribution diagram for explaining the difference in actual injection amount due to machine differences and changes with time of fuel injection valves, variation in injection amount, and variation in combustion.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.
A fuel injection system according to an embodiment of the present invention is shown in FIG.
(Fuel injection system 10)
The accumulator fuel injection system 10 of this embodiment includes a fuel supply pump 14, a common rail 20, a pressure sensor 22, a pressure limiter 24, a fuel injection valve 30, an electronic control unit (ECU) 40, and the like. Fuel is injected into each cylinder of a four-cylinder diesel engine (hereinafter also simply referred to as “engine”) 2.

  The fuel supply pump 14 includes: a feed pump that sucks fuel from the fuel tank 12; and a high-pressure pump that pressurizes fuel sucked from the feed pump into the pressurizing chamber as the plunger reciprocates as the cam of the camshaft rotates. This is a known pump. By controlling the current value supplied to the metering valve 16 of the fuel supply pump 14 by the ECU 40, the fuel suction amount that the fuel supply pump 14 sucks in the suction stroke is metered. Then, by adjusting the fuel intake amount, the fuel discharge amount of the fuel supply pump 14 is adjusted.

  The common rail 20 accumulates the fuel pumped by the fuel supply pump 14 and maintains the fuel pressure at a predetermined high pressure according to the engine operating state. The pressure of the common rail 20 (hereinafter also referred to as “common rail pressure”) is controlled by the discharge amount of the fuel supply pump 14. The pressure sensor 22 as pressure detecting means detects the pressure of the common rail 20 and outputs it to the ECU 40.

The pressure limiter 24 discharges the fuel in the common rail 20 to the fuel tank 12 side when the common rail pressure exceeds a predetermined pressure, and prevents the common rail pressure from exceeding the predetermined pressure.
The fuel injection valve 30 is installed in each cylinder of the four-cylinder diesel engine 50, and injects the fuel accumulated in the common rail 20 into the cylinder. The fuel injection valve 30 performs multistage injection including pilot injection, main injection, and post injection in one combustion stroke of the diesel engine. The fuel injection valve 30 is a known electromagnetically driven valve that controls the fuel injection amount by controlling the pressure in a control chamber that applies fuel pressure to the nozzle needle in the valve closing direction.

  The ECU 40 is composed of a microcomputer (microcomputer) centered on a rewritable nonvolatile memory such as a CPU, ROM, RAM, and flash memory. The ECU 40 includes an accelerator sensor (not shown) that detects the opening (ACC) of the accelerator pedal, an intake air temperature sensor (not shown), a pressure sensor 22, a NE sensor 32 that detects the engine speed (NE), and the like. The operation state of the diesel engine 50 is acquired from the detection signal of the sensor. The ECU 40 controls energization to the metering valve 16, the fuel injection valve 30, and the like based on the acquired engine operating state in order to control the diesel engine 50 to an optimal operating state.

  The ECU 40 stores a discharge amount characteristic of the discharge amount of the fuel supply pump 14 with respect to the energization amount to the metering valve 16 as a map in a storage device such as a ROM or a flash memory according to the engine operating state. The ECU 40 feedback-controls energization to the metering valve 16 so that the common rail pressure acquired from the pressure sensor 22 becomes the target common rail pressure based on the discharge amount characteristic of the fuel supply pump 14 stored in the storage device. .

  In addition, the ECU 40 controls the injection timing and the injection amount of the fuel injection valve 30 according to the engine operating state obtained from various sensors including the accelerator sensor and the NE sensor 32. The ECU 40 outputs an injection pulse signal as an injection command signal for controlling the injection timing and the injection amount of the fuel injection valve 30. The ECU 40 stores the injection amount characteristic of the injection amount with respect to the pulse width of the injection pulse signal in a storage device such as the above-described flash memory as a map for each common rail pressure that is an injection pressure.

The ECU 40 functions as the following units according to a control program stored in a storage device such as a ROM or a flash memory.
(Learning condition judging means)
As shown in FIG. 2A, the ECU 40 determines that the injection amount learning condition is satisfied if it is a non-injection deceleration state when the accelerator is off. Further, in order to prevent disturbances in the engine operating state, the clutch is disengaged for manual transmission (MT), and the engagement state of the friction engagement element is neutral for automatic transmission (AT). May be added to the learning conditions.

(Injection command means)
The ECU 40 outputs an injection pulse signal that instructs the fuel injection valve 30 on the injection timing and the injection amount based on the engine operating state. As the pulse width of the injection pulse signal becomes longer, the injection amount increases. In addition to the normal fuel injection control for causing the engine 2 to generate torque, the ECU 40 controls the fuel injection valve 30 when the learning condition described above is established for each predetermined travel distance, and the injection amount that does not generate torque. A small amount of single injection for learning is executed several times. The minute amount injected for the injection amount learning corresponds to the pilot injection amount.

  At this time, the ECU 40 acquires an injection pulse signal having a pulse width corresponding to the target injection amount from an injection amount characteristic map or the like representing the relationship between the pulse width and the injection amount. The injection amount characteristic map is set for each pressure region obtained by dividing the use range of the common rail pressure into a plurality of regions. The ECU 40 executes a small amount of learning injection for each region of the common rail pressure when the learning condition is satisfied for all the cylinders of the engine 2. If there is a pressure region of the common rail pressure for which the injection amount learning has not ended, the ECU 40 executes the learning injection 210 by controlling the common rail pressure in the pressure region as indicated by the dotted line 200 in FIG. To do.

(Injection amount estimation means)
The ECU 40 executes a plurality of single-shot learning injections 210 for all the cylinders in a non-injection deceleration state when the accelerator is off and in a clutch disengaged state at MT, for example. The ECU 40 detects a change in the engine speed indicated by a dotted line 220 in FIG. 2C from the output signal of the NE sensor 32 as a change in the engine operating state when the learning injection 210 is executed.

  The ECU 40 converts the change in the engine speed into engine torque, and further converts the engine torque into an injection amount. Thereby, the actual actual injection amount with respect to the injection pulse signal having the pulse width τ1 (see FIG. 2D) obtained from the injection amount characteristic map in order to command the target injection amount can be estimated.

(Variation calculation means)
The ECU 40 executes learning injection a plurality of times for each injection amount learning. Then, as described above, the actual injection amount Qn (see FIG. 3A) by single injection is estimated each time, and in each single injection after the first, actual injection from the first to each time is performed. The amount variation (see FIG. 3B) is calculated sequentially. In the present embodiment, the ECU 40 calculates the standard deviation σn as the variation in the actual injection amount.

(Learning end determination means)
As shown in FIG. 3B, when the standard deviation representing the variation in the actual injection amount falls within a range that satisfies a predetermined target accuracy, ECU 40 performs learning for the number of injections executed at that time. It is determined that the injection is finished. The target accuracy is corrected by correcting the injection pulse signal as the injection command value based on the actual injection amount and the target injection amount of the plurality of learning injections executed when the target accuracy is reached, for example. When the fuel injection is controlled by this, the engine 2 is set to a value that does not misfire.

  Note that the standard deviation of the actual injection amount may accidentally fall within a predetermined target accuracy range, for example, in the second learning injection, although the variation in the actual injection amount is large. If the representative value of the actual injection amount for comparison with the target injection amount is calculated based on such an actual injection amount, the representative value may not be calculated with high accuracy.

  Therefore, in the present embodiment, the ECU 40 determines the actual injection amount when the number of injections executed for learning the injection amount is smaller than the minimum number of injections necessary for calculating the standard deviation of the actual injection amount with a predetermined accuracy or more. Even if the standard deviation falls within the target accuracy range, the fuel injection valve 30 is caused to execute the learning injection until the minimum number of injections. If the standard deviation of the actual injection amount is within the target accuracy range and the number of injections at that time is equal to or greater than the minimum number of injections, the ECU 40 determines to end the learning injection.

(Representative value calculation means)
If the standard deviation of the actual injection amount is within the target accuracy range and the number of injections at that time is equal to or greater than the minimum number of injections, the average value Qaven of the actual injection amount is a certain value as shown in FIG. Almost converges. When the learning injection is completed, the ECU 40 sets the average value Qaven of the actual injection amount executed a plurality of times as a representative value of the actual injection amount.

(Correction amount calculation means)
The ECU 40 calculates the correction amount of the injection amount based on the deviation between the average value of the actual injection amount and the target injection amount when the learning injection is ended, as shown in FIG. The injection amount characteristic map at the target common rail pressure is corrected by the amount. Specifically, the pulse width of the injection pulse signal corresponding to the target injection amount is changed from τ1 to τ2.

(Comparative example)
In contrast to the present embodiment, in the comparative example of FIG. 4B, in the fuel injection valve 30, the injection amount variation for each injection and the maximum variation of the actual injection amount when the combustion variation for each injection is maximized. Assuming that the standard deviation Σn of the maximum variation (indicated by reference numeral 240) falls within the target accuracy range, n2 learning injections are executed uniformly in accordance with the standard deviation Σn of the maximum variation. In the case of the comparative example, the learning injection is performed n2 times even when the variation in the actual injection amount is within the target accuracy range in the n1 (n1 <n2) times shown in FIG. 3 as in the present embodiment. The As a result, in the comparative example, the number of learning injections increases with respect to the present embodiment.

  On the other hand, in the present embodiment, as shown in the upper part of FIG. 5, when the variation in the actual injection amount is maximum, the learning injection is executed up to n2 times as in the comparative example. Since the variation in the actual injection amount is smaller than the maximum value, as shown in the lower part of FIG. 5, the number of injections (n1) in which the variation in the actual injection amount falls within the target accuracy range is smaller than n2.

  Note that, as described above, in the present embodiment, the number of injections performed for injection amount learning calculates the standard deviation of the actual injection amount even if the variation in the actual injection amount is within the target accuracy range. If it is less than the required minimum number of injections, the fuel injection valve 30 is caused to perform learning injection until the minimum number of injections.

(Injection amount learning routine)
Next, injection amount learning in the fuel injection system 10 will be described with reference to FIG. In FIG. 6, “S” represents a step. The routine of FIG. 6 is executed when the injection amount learning condition is satisfied.

  In S400, the ECU 40 controls the discharge amount of the fuel supply pump 14, sets the common rail pressure in a pressure region where the injection amount learning has not ended, and executes single injection for injection amount learning.

In S402, the ECU 40 detects the change amount of the engine speed as the change amount of the engine operation state due to the single injection executed this time.
In S404, the ECU 40 converts the change amount of the engine speed into the engine torque, converts the engine torque into the actual injection amount Qn of the fuel injection valve 30, and calculates the average value Qaven of the actual injection amount so far. Further, in S404, the ECU 40 calculates the standard deviation σn as the variation in the actual injection amount from the first time to the current time in the learning injection executed after the first time.

  In S406, the ECU 40 determines whether or not the actual injection amount calculated in S404 has a deviation of a predetermined value or more with respect to the target injection amount. When the deviation of the predetermined value or more continues for a predetermined number of times (S408: Yes), It is determined that the injection amount learning should be resumed after eliminating the deviation of the actual injection amount from the target injection amount by a predetermined value or more. In S410, based on the deviation between the actual injection amount and the target injection amount, the correction amount of the injection command value for single injection is calculated and reflected in the injection command value. Then, the ECU 40 returns to S400, clears the number of injections executed so far, and re-executes the learning injection.

  When the deviation amount between the actual injection amount and the target injection amount is less than the predetermined value (S408: No), the ECU 40 counts up the number of learning injections executed under the same learning condition in S412.

  Then, the standard deviation σn, which is the variation in the actual injection amount calculated in S404, is within the target accuracy range (S414: Yes), and the number of executed learning injections sets the standard deviation of the actual injection amount to a predetermined value. If the number of injections is equal to or greater than the minimum number of injections required for calculation with the accuracy of (S416: Yes), the ECU 40 determines in S418 that the complementary standard deviation is calculated with a predetermined accuracy or more, and the actual injection amount calculated in S404. Based on the deviation between the average value Qaven and the target injection amount, the correction amount of the injection command value for single injection is calculated as the current learning amount, and the correction amount is determined.

  Whether the standard deviation σn of the actual injection amount calculated in S404 is not within the target accuracy range (S414: No), or the number of executed learning injections calculates the standard deviation of the actual injection amount with a predetermined accuracy or more. If it is less than the minimum number of injections required to perform (S416: No), the ECU 40 determines that it is necessary to continue the injection for learning, and returns the process to S400.

  If the standard deviation σn of the actual injection amount does not fall within the target accuracy range even when the number of injections considering the maximum variation is reached due to an abnormality of the fuel injection valve 30, an abnormality determination routine different from this routine. The abnormality is determined.

  In the present embodiment, the ECU 40 corresponds to a fuel injection control device, an injection command unit, an injection amount estimation unit, a variation calculation unit, a learning end determination unit, a representative value calculation unit, and a correction amount calculation unit of the present invention. In the present embodiment, S400 in FIG. 7 corresponds to the function executed by the injection command means of the present invention, and S402 and S404 in FIG. 7 indicate the injection amount estimation means, variation calculation means, and representative value calculation means of the present invention. 7 corresponds to the function executed by the learning end determination means of the present invention, and S418 of FIG. 7 corresponds to the function executed by the correction amount calculation means of the present invention.

  In the present embodiment described above, the actual injection amount is estimated based on the amount of change in the engine speed when the learning injection is executed, and the standard deviation is within the target accuracy range as the variation in the estimated actual injection amount. When it is settled, the learning injection is terminated. Accordingly, the learning injection can be completed with a smaller number of times than when the number of times of the learning injection is uniformly set in consideration of the maximum variation in the actual injection amount when the learning injection is executed.

  As a result, when the next injection amount learning is executed at a predetermined travel distance interval after the previous injection amount learning is completed in the vehicle, the previous injection amount learning is started and the next injection amount learning is started. Therefore, the injection amount learning is frequently performed during a certain travel distance. As a result, the period during which the injection amount of the fuel injection valve 30 is controlled with the actual injection amount deviating from the target injection amount can be shortened as much as possible.

  In the present embodiment, the four-cylinder engine 2 has been described. However, an engine having a larger number of cylinders than the four cylinders takes a longer time to complete the injection amount learning for all the cylinders. The effect of shortening the execution period of the injection amount learning is increased.

[Other Embodiments]
In the above embodiment, the engine speed is adopted as the amount of change in the engine operating state when the learning injection is executed. On the other hand, an air-fuel ratio or an in-cylinder pressure of the engine may be employed as the amount of change in the engine operating state.

  In the above embodiment, in addition to the standard deviation of the actual injection amount falling within the target accuracy range, the number of learning injections is the minimum required for calculating the standard deviation of the actual injection amount with a predetermined accuracy or more. The end condition of the learning injection was set to be equal to or greater than the number of injections. On the other hand, the fact that the number of injections is equal to or greater than the minimum number of injections may be excluded from the learning injection end condition.

  In the above embodiment, the functions of the injection command means, the injection amount estimation means, the variation calculation means, the learning end determination means, the representative value calculation means, and the correction amount calculation means are realized by the ECU 40 whose functions are specified by the control program. . On the other hand, at least some of the functions of the plurality of means may be realized by hardware whose functions are specified by the circuit configuration itself.

  As described above, the present invention is not limited to the above-described embodiment, and can be applied to various embodiments without departing from the gist thereof.

10: fuel injection system, 14: fuel supply pump, 20: common rail, 30: fuel injection valve, 40: ECU (fuel injection control device, injection command means, injection amount estimation means, variation calculation means, learning end determination means, representative Value calculation means, correction amount calculation means)

Claims (4)

An injection command means for instructing the fuel injection valve to perform learning injection when a learning condition for the injection amount by single injection of the fuel injection valve is satisfied;
Injection amount estimation means for estimating the actual injection amount of the single injection based on a change in the operating state of the internal combustion engine when the single injection for learning the injection amount is performed from the fuel injection valve according to the command of the injection command means When,
In the single injection executed a plurality of times, variation calculating means for sequentially calculating variations in the actual injection amount based on the actual injection amount estimated by the injection amount estimating means;
A learning end determination unit that determines the number of injections when the learning injection is ended based on the variation calculated by the variation calculation unit and the target accuracy of the variation;
A fuel injection control device comprising:   The learning end determination means ends the learning injection when the variation is within the target accuracy range and the number of injections executed for injection amount learning is equal to or greater than the minimum number of injections necessary for calculating the variation. The fuel injection control device according to claim 1, wherein the determination is made.   The learning end determination means causes the injection command means to continue the learning injection when the number of injections executed for injection amount learning is less than the minimum number of injections necessary for calculating the variation. The fuel injection control device according to claim 1 or 2. Representative value calculating means for calculating a representative value of the actual injection amount based on the actual injection amount of the single injection executed a plurality of times;
A deviation between the representative value calculated by the representative value calculating means and the target injection amount of the actual injection amount when the learning end determination means determines that the variation is within the target accuracy range and the learning injection is finished. Correction amount calculating means for calculating a correction amount of an injection command value for commanding fuel injection of the target injection amount to the fuel injection valve, based on
The fuel injection control device according to any one of claims 1 to 3, further comprising:

Images