Injection control device

Abstract

To provide an injection control device capable of improving the accuracy of correction in a multi-stage injection.SOLUTION: An injection control device 1 includes an injection controller 20 for controlling a fuel injection valve 3, which is driven by charge voltage obtained by boosting battery voltage, to perform a multi-stage injection in which fuel injection is performed a plurality of times in one cycle for each cylinder. The injection controller 20 learns the second and subsequent minute injections in the multi-stage injection to correct the injection amount.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present disclosure relates to an injection control device for controlling a fuel injection valve. [Background technology]

[0002] Multi-injection, in which fuel is injected multiple times during one cycle for each cylinder, is a known technology for reducing nitrogen oxides and particulate matter emitted from internal combustion engines. Hereinafter, the series of processes of intake, compression, combustion, and exhaust in each cylinder will be referred to as one cycle.

[0003] The total required injection amount during one cycle is determined based on operating conditions, etc. Therefore, in multi-injection, minute injections are performed in which the amount of fuel injected per injection is less than the total required injection amount. It is anticipated that minute injections will result in variations in the injection amount due to the mechanical characteristics of the fuel injection valve. For this reason, Patent Document 1, for example, proposes correcting the injection amount by improving the detection accuracy of the valve closing timing of the fuel injection valve. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Laid-Open No. 2017-201157 Summary of the Invention [Problem to be solved by the invention]

[0005] However, when minute injections are performed at relatively short intervals, as in multiple injections, the behavior of the fuel injector in the second and subsequent injections may change due to factors such as the electromotive force of the previous injection. In this case, even if the learning results of the first injection are applied to correct the second and subsequent injections, the injection accuracy may not improve. The present disclosure has been made in consideration of the above circumstances, and has an object to provide an injection control device that can improve the accuracy of correction in multi-injection. [Means for solving the problem]

[0006] The injection control device (1) according to the present disclosure includes an injection control unit (20) that controls a fuel injection valve (3) driven by a charging voltage obtained by boosting a battery voltage, by multi-injection in which fuel is injected in multiple times during one cycle in each cylinder. Among the multiple micro-injections that take place within It is performed from the second time onwards. Ta Learn about micro-injection , the second and subsequent micro-injections during the next multi-injection This corrects the injection amount. Even if the behavior of the fuel injection valve (3) in the second or subsequent injections differs from the behavior in the first injection due to the influence of the electromotive force of the previous injection, the difference in behavior can be corrected by learning, and the accuracy of correction in multiple injections can be improved. [Brief explanation of the drawings]

[0007] [Figure 1] FIG. 1 is a diagram showing an example of the configuration of an injection control device according to an embodiment; [Figure 2] FIG. 1 shows an example of the configuration of a fuel injection valve. [Figure 3] FIG. 10 is a diagram showing an example of a drive signal given to a fuel injection valve; [Figure 4] Functional block diagram of injection control device [Figure 5] Diagram explaining the calculation flow by the Ti calculation block [Figure 6] Flowchart showing the process of determining whether learning is necessary [Figure 7] An example of a learning Tc map [Figure 8] Flowchart showing the process of learning the later stage injection [Figure 9] A diagram explaining the calculation flow by the subsequent required Q correction unit DETAILED DESCRIPTION OF THE INVENTION

[0008] Hereinafter, an embodiment will be described with reference to the drawings. As shown in Fig. 1, an injection control device 1 of this embodiment controls a fuel injection valve 3 provided in an internal combustion engine 2. If this internal combustion engine 2 is, for example, a four-cylinder internal combustion engine for a vehicle, the fuel injection valve 3 is provided in each of a first cylinder designated as #1, a second cylinder designated as #2, a third cylinder designated as #3, and a fourth cylinder designated as #4. However, the configuration of the internal combustion engine 2 shown in Fig. 1 is merely an example, and the number of cylinders may be different, for example.

[0009] As shown in Fig. 2, the fuel injection valve 3 houses a solenoid coil 302, a valve element 303, a fixed core 304, and a movable core 305 inside a body 301. The valve element 303 is cylindrical overall, with a conical tip on the lower side in the figure, allowing it to move in the axial direction. The body 301 has an injection hole 306 for injecting fuel at the tip on the lower end side in the figure.

[0010] The fixed core 304 is made of a magnetic material and has a cylindrical shape, and its inner periphery forms a space through which fuel flows as the valve element 303 moves. A movable core 305 made of a metallic magnetic material and formed into a disk shape with a through-hole in the center is disposed on the injection hole 306 side of the fixed core 304. The movable core 305 is provided so as to be movable in the axial direction on the inner periphery side of the solenoid coil 302, with the valve element 303 passing through its inner periphery. When the solenoid coil 302 is not energized, the movable core 305 is located in an initial position facing the fixed core 304 with a predetermined gap between them. A locking portion 307 fixed to the valve element 303 is in contact with the upper surface of the movable core 305.

[0011] A first spring 308 that elastically presses the valve element 303 toward the injection hole 306 is provided on the inner peripheral side of the fixed core 304 and is wound around the valve element 303. On the other hand, a second spring 309 that elastically pulls the movable core 305 to its initial position is provided on the injection hole 306 side of the movable core 305 and is fixed to the body 301.

[0012] When the solenoid coil 302 of the fuel injection valve 3 is energized, the movable core 305 moves toward the fixed core 304 against the elastic force of the second spring 309. As the movable core 305 moves, the locking portion 307 is pushed up, causing the valve element 303 to move in the axial direction. Even if the movable core 305 moves and abuts against the fixed core 304, the valve element 303 and the locking portion 307 can still move in the axial direction inside the fixed core 304. In other words, the fuel injection valve 3 is structured such that the mechanical characteristics of the elastic forces of the first spring 308 and the second spring 309 have an effect when injecting fuel.

[0013] Fuel is supplied to each fuel injection valve 3 from a fuel supply system (not shown). This fuel supply system is provided with a pressure sensor 4 that detects the pressure of the fuel and a temperature sensor 5 that detects the temperature of the fuel. The fuel pressure detected by the pressure sensor 4 and the fuel temperature detected by the temperature sensor 5 are input to the injection control device 1 as parameters for calculation.

[0014] 1, the injection control device 1 includes a control circuit 10, a storage unit 11, a drive circuit 12, and a boost circuit 13. The control circuit 10 is configured as a microcomputer, and controls the entire injection control device 1 by executing a program stored in the storage unit 11. The storage unit 11 is configured as a non-volatile storage medium such as a flash memory, and stores various data used by the injection control device 1 and data such as learning results, which will be described later.

[0015] The drive circuit 12 generates and outputs a drive signal for driving the fuel injector 3 based on a command value output from the control circuit 10, and also outputs various parameters acquired from the fuel injector 3 to the control circuit 10. The drive signal is generated in pulse form by a charging voltage obtained by boosting the battery voltage supplied from a battery 14 by a boost circuit 13. For example, as shown in FIG. 3, in the case of a four-cylinder internal combustion engine, the injection order is #1 → #3 → #4 → #2, and energization pulses #1 to #4 of drive signals are output to the fuel injector 3 of each cylinder according to the injection order.

[0016] When the energization pulse is turned on, the solenoid coil 302 of the fuel injector 3 is energized, causing the valve element 303 to move and inject fuel. On the other hand, when the energization pulse is turned off, the solenoid coil 302 of the fuel injector 3 is de-energized, causing the fuel injection to stop. As shown in an enlarged view, each energization pulse switches between on and off states multiple times during one cycle, where one intake stroke, compression stroke, combustion stroke, and exhaust stroke in one cylinder constitutes one cycle. Note that the injection order shown in FIG. 3 is an example.

[0017] In this way, the fuel injection valve 3 is controlled by multi-injection, which divides fuel injection into multiple injections during one cycle for each cylinder. In the following explanation, the first injection during one cycle will be designated as N=1, the second injection as N=2, the third injection as N=3, the fourth injection as N=4, and the fifth injection as N=5. In addition, the first injection during multi-injection, i.e., the injection where N=1, will also be referred to as the first-stage injection, and the second and subsequent injections, i.e., injections where N≧2, will be referred to as subsequent-stage injections.

[0018] The control circuit 10 includes an injection control unit 20 and a platform 21 shown as PF in Fig. 1. The injection control unit 20 and the platform 21 are configured with software by executing a program in the control circuit 10. The platform 21 is a software group made up of driver software and the like that exchanges signals and data between the injection control unit 20 and external circuits.

[0019] As shown in FIG. 4, the injection control unit 20 includes an initial-stage learning block 22, a subsequent-stage learning block 23, and a Ti calculation block 24, and generates a command value to be given to the drive circuit 12. As will be described in detail later, the initial-stage learning block 22 is a functional block that learns the initial-stage injection during multi-injection in a conventional manner and performs the calculations and corrections required for the initial-stage injection. The subsequent-stage learning block 23 is a functional block that learns the subsequent-stage injection during multi-injection and performs the calculations and corrections required for the subsequent-stage injection. The Ti calculation block 24 is a functional block that calculates the energization time (Ti), which is the command value to be given to the drive circuit 12, based on the output from the initial-stage learning block 22 or the subsequent-stage learning block 23.

[0020] Next, the operation and effects of the above-described configuration will be described. First, learning in the first stage learning block 22 will be described. The first stage learning block 22 includes a PL correction coefficient learning unit 221 and a first stage required Q correction unit 222. The PL correction coefficient learning unit 221 is a functional block that learns the correction coefficient for partial lift injection. Partial lift injection is a minute injection in which current supply is stopped and the valve closing operation is initiated before the valve disc 303 rises to the maximum valve open position, thereby shortening the valve opening time and injecting a minute amount of fuel.

[0021] In partial lift injection, the linearity between the energization time (Ti) and the actual injection amount tends to be low. This is because the variation in the amount of lift of the valve element 303 increases due to the mechanical characteristics of the fuel injection valve 3. In full lift injection, in which fuel is injected with the valve element 303 raised to the maximum open position, the linearity between the energization time (Ti) and the actual injection amount is high, and the injection amount per unit stroke increases in a manner roughly proportional to the energization time.

[0022] The PL correction coefficient learning unit 221 learns a PL correction coefficient for correcting the injection amount during partial lift injection based on the valve closing time (Tc) of the fuel injector 3 during the first injection and parameters. Here, the parameters include, for example, fuel pressure, fuel temperature, current application time (Ti), and current application pulse interval. However, the parameters are not limited to those exemplified here, and other parameters can be used as needed. Hereinafter, the fuel injection amount required for one injection will be referred to as required Q.

[0023] The PL correction coefficient is provided as a learned value to a first-stage required Q correction unit 222. When the injection is the first stage during multi-injection and is a partial lift injection, the first-stage required Q correction unit 222 corrects the required Q based on the PL correction coefficient and outputs the corrected required Q as a provisional required Q to the Ti calculation block 24. Note that when the injection is the first stage but is not a partial lift injection, the first-stage required Q correction unit 222 outputs the required Q as a provisional required Q without correcting it.

[0024] 4, the Ti calculation block 24 calculates the energization time (Ti) for injecting fuel of the tentative demand Q based on the input tentative demand Q. At this time, the relationship between the energization time (Ti) and the fuel injection amount (Q) is stored in advance in the storage unit 11 as a Ti-Q map 31 schematically shown in graph G1. Hereinafter, the difference between the demand Q and the tentative demand Q is referred to as the injection correction amount (ΔQ), and the absolute value of the injection correction amount (ΔQ) is referred to as the difference value (q).

[0025] For example, when the provisional demand Q is greater than the demand Q, the relationship between the energization time (Ti) and the actual injection amount is offset from the graph G1 by the difference value (q) in the direction of increasing the injection amount, as shown by the dashed line graph G2. In this case, if the fuel injector is controlled using the energization time Ti1 corresponding to the demand Q obtained from the graph G1, the actual relationship will be as shown in the graph G2, and therefore an amount of fuel greater than the demand Q will be injected. Therefore, the Ti calculation block 24 obtains the energization time to be actually output as Ti2, which corresponds to the intersection of the demand Q and the graph G2, based on the demand Q and the injection correction amount (ΔQ).

[0026] Note that Ti2 corresponds to the energization time corresponding to the intersection with the demand Qq when there is no offset in graph G1. Therefore, the actual calculation is to find the demand Q+ΔQ and find the energization time corresponding to the intersection of the demand Q+ΔQ and graph G1. Similarly, when the provisional demand Q is greater than the demand Q, the graph G1 is offset by the difference value (q) in the direction that reduces the injection amount, but by finding the demand Q+ΔQ, the energization time that should be output can be found from graph G1.

[0027] In multi-injection, minute injection is performed in which the total required injection amount of fuel required in one cycle is injected in multiple instalments. Since it is expected that the injection amount of this minute injection will vary depending on the mechanical characteristics of the fuel injection valve 3, as described above, the first-stage injection is learned and the learned value is used to correct the injection amount.

[0028] However, it was found that in the case of minute injections during one cycle, the behavior of the fuel injector 3 differs between the first-stage injection and the second-stage injection. This is thought to be because the second-stage injection is affected by factors such as the electromotive force from the first-stage injection. Furthermore, in the case of multiple injections, the effects are thought to be more pronounced because minute injections are repeated in a short period of time. Furthermore, since changes in the mechanical characteristics of the fuel injector 3 are thought to affect the valve opening and closing operations, it is expected that the fuel injector 3 will also be affected by aging.

[0029] Therefore, the injection control device 1 improves the accuracy of correction in multi-injection by learning about the latter injection, as will be explained below. Also, the injection control device 1 aims to reduce the processing load related to learning, as will be explained below.

[0030] First, the reduction in processing load will be described. As shown in Fig. 4, the subsequent learning block 23 includes a Tc map learning unit 231 and a subsequent required Q correction unit 232. In the injection control device 1, in order to reduce the processing load during learning, the Tc map learning unit 231 executes the processing shown in Fig. 6. The Tc map learning unit 231 determines whether N≧2, that is, whether the injection is the second or subsequent injection, whether it is a minute injection, and whether the learning completion flag is off (S1). That is, in step S1, the Tc map learning unit 231 determines whether the conditions for requiring learning are met.

[0031] The learning flag is a flag that indicates whether learning has been completed or not, and is turned on if learning has been completed, and is turned off if learning has not been completed. In this embodiment, the learning flag is turned off when the injection control device 1 is started. Therefore, learning is performed once after the ignition is turned on. This makes it possible to perform appropriate corrections in response to aging even if the fuel injector 3 has aged.

[0032] If the Tc map learning unit 231 determines in step S1 that any of the conditions is not satisfied (S1: NO), it terminates the process. On the other hand, if the Tc map learning unit 231 determines in step S1 that all of the conditions are satisfied (S1: YES), it acquires injection parameters (S2). At this time, the Tc map learning unit 231 acquires the valve closing time (Tc), fuel pressure, fuel temperature, injection interval, and the like as injection parameters. Note that the Tc map learning unit 231 also acquires other parameters, such as current application time, if these parameters are required.

[0033] Next, the Tc map learning unit 231 uses the acquired parameters to update the learned Tc map 30 (S3). The learned Tc map 30 is a data group that stores the correspondence between the valve closing time (Tc) and each parameter. Specifically, as shown in FIG. 7, the learned Tc map 30 stores the valve closing time corresponding to, for example, the fuel pressure and temperature as learned Tc maps 30A to 30C, which are associated with, for example, each injection interval or each current application time. However, the learned Tc map 30 shown in FIG. 7 is just an example.

[0034] For example, it is assumed that the shorter the interval, the greater the effect of the earlier injection on the later injection. Therefore, by storing the valve closing time corresponding to the interval in the learned Tc map 30, appropriate correction can be made. Furthermore, it is assumed that the injection amount changes depending on the fuel pressure even if the interval is the same. Therefore, by storing the valve closing time corresponding to the fuel pressure in the learned Tc map 30, appropriate correction can be made. In other words, by learning for each of multiple parameters, the Tc map learning unit 231 is able to make appropriate corrections for the next and subsequent injections.

[0035] If this is the first learning, the Tc map learning unit 231 does not store the correspondence between the valve closing time and each parameter, so it associates the currently acquired valve closing time with each parameter and records it in the corresponding region (P0) as an initial learning point.The Tc map learning unit 231 then weights the value of the initial learning point to estimate and store the valve closing times in other regions, as partially indicated by the dashed arrow, thereby updating the learned Tc map 30.

[0036] On the other hand, if it is the second or subsequent learning, the Tc map learning unit 231 smooths the valve closing time corresponding to the acquired parameters to remove noise components and stores the result. That is, the Tc map learning unit 231 learns the valve closing time and updates the learned Tc map 30. As a result, even if the behavior of the fuel injector 3 differs in the second or subsequent injections due to the influence of the electromotive force of the preceding injection, it is possible to obtain the learned Tc map 30 in a state where the difference in behavior has been absorbed.

[0037] Next, the Tc map learning unit 231 calculates a learning update difference, which is the difference between the stored value and the updated value (S4), and determines whether the learning update difference is equal to or smaller than a predetermined value (S5). In this case, the Tc map learning unit 231 determines whether the learning update difference is equal to or smaller than a predetermined value based on, for example, the absolute value of the learning update difference or the ratio to the previous value.

[0038] If the Tc map learning unit 231 determines that the learning update difference is not equal to or less than a predetermined value (S4: NO), the process ends without changing the learning completion flag. Therefore, during the first learning or when the learning value is not stable, the learning process is repeated.

[0039] On the other hand, if the Tc map learning unit 231 determines that the learning update difference is equal to or smaller than a predetermined value (S4: YES), it turns on the learning completion flag (S5) and then terminates the processing. In this embodiment, learning is completed between the end of injection and the start of the next injection, which increases the processing load. Therefore, if the learning is stable, the Tc map learning unit 231 turns on the learning flag and does not perform the next learning. In other words, the injection control device 1 is configured to stop learning when the learned value is stable.

[0040] The learned Tc map 30 is referred to by the subsequent-stage requirement Q correction unit 232 as shown in Fig. 4 and is used to correct the requirement Q. Specifically, the subsequent-stage requirement Q correction unit 232 executes the process shown in Fig. 8 and corrects the requirement Q in the procedure shown in Fig. 9 to obtain the tentative requirement Q.

[0041] The subsequent-stage required Q correction unit 232 determines whether N≧2, that is, whether the injection is the second or subsequent injection and whether it is a minute injection (S11). If N≧2 is not satisfied or if it is determined that the injection is not a minute injection (S11: NO), the subsequent-stage required Q correction unit 232 outputs the required Q for this injection as a tentative required Q to the Ti calculation block 24 without correcting it (S17), and ends the process.

[0042] On the other hand, if it is determined that N≧2 and the injection is minute (S11: YES), the subsequent-stage required Q correction unit 232 acquires parameters necessary for correction (S12). In this step S2, parameters such as the current application time, interval, fuel pressure, and fuel temperature are acquired or calculated. If other parameters are necessary, they are also acquired or calculated.

[0043] Once the parameters are acquired, the subsequent-stage requirement Q correction unit 232 obtains an estimated Tc from the learned Tc map 30 (S13). This estimated Tc is the valve closing time corresponding to the requirement Q, and as shown in FIG. 9, is a learned value stored in an area (P1) corresponding to the fuel temperature and pressure in the learned Tc map 30A corresponding to the current supply time obtained from the requirement Q. In this way, the subsequent-stage requirement Q correction unit 232 obtains an appropriate estimated Tc according to, for example, the current supply time or the interval, even if, for example, the fuel pressure or temperature is the same. Furthermore, the subsequent-stage requirement Q correction unit 232 obtains an appropriate estimated Tc according to, for example, the fuel pressure or temperature, even if, for example, the interval is the same.

[0044] Next, the subsequent stage required Q correction unit 232 obtains an estimated Q from the Tc-Q map 32 as shown in Fig. 8 (S14). This Tc-Q map 32 stores a relationship between the valve closing time (Tc) and the injection amount (Q) in advance, as schematically shown as graph G11 in Fig. 9. Therefore, the subsequent stage required Q correction unit 232 obtains the injection amount corresponding to the estimated Tc as the estimated Q by referring to the Tc-Q map 32.

[0045] Further, the subsequent-stage required Q correction unit 232 calculates a q deviation correction value as shown in Fig. 8 (S15). This q deviation correction value is calculated based on a correction map 33 in which the relationship between the deviation in the opening of the fuel injector 3 and the correction value is stored in advance as shown in Fig. 9. Then, the subsequent-stage required Q correction unit 232 adds the estimated Q and the q deviation correction value as shown in Fig. 8 to calculate a tentative required Q for the subsequent injection (S16), outputs the calculated tentative required Q to the Ti calculation block 24 (S17), and ends the processing. Then, in the Ti calculation block 24, the energization time (Ti) is calculated based on the tentative required Q as shown in Fig. 5.

[0046] In this way, the injection control device 1 learns the second and subsequent injections in a multi-injection and enables appropriate corrections according to the behavior of the fuel injection valve 3 from the second injection onwards, thereby improving the accuracy of corrections in multi-injection.

[0047] According to the injection control device 1 described above, the following effects can be obtained. The injection control device 1 includes an injection control unit 20 that controls the fuel injector 3, which is driven by a charging voltage obtained by boosting the battery voltage, by multi-injection, which involves injecting fuel multiple times during one cycle in each cylinder. The injection control unit 20 then learns about the minute injections performed from the second injection onwards during multi-injection and corrects the injection amount. This makes it possible to correct the difference in behavior through learning, even if the behavior of the fuel injector 3 during the second or subsequent injections differs from the behavior during the first injection due to the influence of the electromotive force of the previous injection, thereby improving the accuracy of correction during multi-injection. Furthermore, by learning all injections as in the embodiment, the accuracy of correction during multi-injection can be improved even if the behavior of the third or subsequent injection differs from the behavior during the previous injections.

[0048] Because the fuel injection valve 3 repeats fuel injection at a relatively short interval during microinjection, it is thought to be susceptible to the influence of factors such as the electromotive force of the previous injection. It is also thought that the behavior of the fuel injection valve 3 itself changes over time. Therefore, the injection control device 1 performs learning and correction on the fuel injection valve 3, which has a valve element 303 pressed against an injection hole 306 by a spring member, and which injects fuel from the injection hole 306 when a drive signal based on a command value output from the injection control unit 20 is given, driving the valve element 303. This makes it possible to correct the injection amount while absorbing differences in behavior during the second and subsequent microinjections and also absorbing changes in behavior due to aging, thereby improving the accuracy of correction in multiple injections.

[0049] The injection control device 1 learns and corrects the injection amount for each parameter used to calculate the injection amount. This improves the accuracy of correction in multi-injection, for example, by making it possible to calculate an appropriate valve closing time (Tc) according to a different energization time even if the fuel pressure or temperature is the same.

[0050] The injection control device 1 stops learning when the learned value becomes stable. The injection control device 1 needs to complete learning between the end of injection and the start of the next injection, which tends to increase the processing load. On the other hand, the mechanical characteristics of the spring member of the fuel injection valve 3 and the like are not thought to deteriorate significantly with each cycle, so the load can be reduced by stopping learning when the learned value is stable.

[0051] Furthermore, although the embodiment has been described as an example of a configuration in which learning is stopped when the learning value is stable, a configuration can be adopted in which the load is reduced by limiting the number of times valve-close detection is performed to detect the valve-closing time (Tc) of the fuel injector 3. For example, a configuration can be adopted in which valve-close detection is performed during the second injection in one cycle shown in FIG. 3, but not during the third or subsequent injections. Furthermore, for example, a process for determining a flag indicating whether learning was performed during the previous injection can be added to the process of FIG. 6, so that valve-close detection is not performed in the cycle following the cycle in which valve-close detection was performed. Furthermore, for example, a process for counting injections to be learned can be added to the process of FIG. 6, so that valve-close detection is performed every predetermined number of cycles. Furthermore, a configuration can be adopted in which limiting the number of times valve-close detection is performed and stopping learning when the learning value is stable are combined.

[0052] This allows the injection amount to be corrected while absorbing differences in the behavior of the fuel injection valve 3 during minute injection and also absorbing changes in behavior due to aging, improving the accuracy of correction in multi-injection.In addition, since the load can be reduced, ICs for the control circuit 10 and drive circuit 12 with a wide range of specifications can be used.

[0053] When the learning value has stabilized, the injection control unit 20 can be configured to select whether or not to perform learning depending on the processing load. For example, when the rotation speed of the internal combustion engine 2 is high, one cycle shown in FIG. 3 becomes shorter, and the interval between microinjections also becomes shorter. In this case, performing learning for each injection increases the load, but by incorporating load determination into step S1 of FIG. 6, for example, it is possible to configure the control circuit 10 not to perform learning when the load is high. This allows learning and correction to be performed without making the control circuit 10 excessively high performance. In this case, the injection control unit 20 can be configured to perform learning when low-load processing is being executed, which has a long grace period until the processing is completed. This reduces the load on learning.

[0054] The injection control unit 20 can be configured to execute the learning itself as a low-load task with a long grace period. That is, the learning load itself can be reduced, and for example, a process for determining whether the load is high or low can be incorporated before or after step S1 in Fig. 6, so that the learning can be performed under a low-load condition. This is because, as described above, it is considered that the mechanical characteristics of the spring member of the fuel injector 3 and the like are not so significant that they deteriorate with each cycle.

[0055] The classification of the functional blocks and the calculation contents they are responsible for described in the embodiment are merely examples, and how to allocate the functional blocks and calculation contents can be selected as appropriate. In other words, the injection control device 1 as a whole should be configured to learn about the minute injections performed from the second onwards during multi-injection and correct the injection amount.

[0056] The control unit and the method described herein may be implemented by a special-purpose computer configured by configuring a processor and memory programmed to perform one or more functions embodied in a computer program. Alternatively, the control unit and the method described herein may be implemented by a special-purpose computer configured by configuring a processor with one or more dedicated hardware logic circuits. Alternatively, the control unit and the method described herein may be implemented by one or more special-purpose computers configured by combining a processor and memory programmed to perform one or more functions with a processor configured with one or more hardware logic circuits. Furthermore, the computer program may be stored as instructions executed by a computer on a computer-readable non-transitory tangible storage medium.

[0057] In addition to the inventions described in the claims, this case also includes the following inventions: [1] an injection control unit that controls a fuel injection valve driven by a charging voltage obtained by boosting a battery voltage by multi-injection that performs fuel injection in multiple steps during one cycle in each cylinder; The injection control unit is an injection control device that learns about minute injections performed from the second injection onwards during multi-injection and corrects the injection amount. [2] The injection control device according to [1], wherein the injection control unit corrects the injection amount by learning each parameter used to calculate the injection amount. [3] The injection control device according to [1] or [2], wherein the injection control unit stops learning when the learned value becomes stable. [4] The injection control device according to any one of [1] to [3], wherein the injection control unit limits the number of times that the closed valve of the fuel injection valve is detected when the learned value is stabilized. [5] The injection control device according to any one of [1] to [4], wherein the injection control unit selects whether to perform learning depending on a processing load when the learned value is stabilized. [Explanation of symbols]

[0058] In the drawing, 1 indicates an injection control device, 3 indicates a fuel injection valve, and 20 indicates an injection control section.

Claims

1. The fuel injection valve (3) is driven by a charging voltage obtained by boosting the battery voltage, and an injection control unit (20) controls the fuel injection valve (3) by multi-injection, which performs fuel injection in multiple steps during one cycle in each cylinder, The injection control unit is an injection control device that learns about the micro-injections performed from the second onwards among the multiple micro-injections performed during a multi-injection, and corrects the injection amount of the micro-injections performed from the second onwards during the next multi-injection.

2. 2. The injection control device according to claim 1, wherein the injection control unit corrects the injection amount by learning each parameter used to calculate the injection amount.

3. 2. The injection control device according to claim 1, wherein the injection control unit stops learning when a learned value learned about the second or subsequent minute injection becomes stable.

4. 2. The injection control device according to claim 1, wherein the injection control unit limits the number of times the valve is closed when a learned value learned about the second or subsequent minute injection becomes stable.

5. 2. The injection control device according to claim 1, wherein the injection control unit selects whether to perform learning depending on a processing load when a learned value learned about the second or subsequent minute injection becomes stable.

Images