WO2015110893A1

WO2015110893A1 - Method and apparatus for controlling fuel injection of an internal combustion engine

Info

Publication number: WO2015110893A1
Application number: PCT/IB2015/000041
Authority: WO
Inventors: Yoshiyazu ITO; Naoyuki Yamada
Original assignee: Toyota Jidosha Kabushiki Kaisha; Denso Corporation
Priority date: 2014-01-22
Filing date: 2015-01-19
Publication date: 2015-07-30
Also published as: JP6215718B2; DE112015000457T5; JP2015137595A; WO2015110893A8

Abstract

A system is provided for an internal combustion engine including a fuel supply system that supplies pressurized fuel to a fuel injector, and a pressure sensor that detects a fuel pressure inside the fuel supply system. The system includes an ECU. The ECU is configured to detect characteristics of the fuel injector based on the fuel pressure detected by the pressure sensor, execute multi-stage injection in which fuel is injected from the fuel injector a plurality of times in one combustion cycle, and detects the operating characteristics associated with each injection of the multi-stage injection. The ECU is configured to execute a first combustion cycle in which the operating characteristics associated with at least one particular injections other than an initial-stage injection are not detected, and a second combustion cycle in which the operating characteristics associated with said at least one particular injections are detected.

Description

METHOD AND APPARATUS FOR CONTROLLING FUEL INJECTION

OF AN INTERNAL COMBUSTION ENGINE

BACKGROUND OF THE INVENTION

1. Field of the Invention

[0001] The invention relates to fuel injection characteristic detection system and method of detecting operating characteristics of fuel injectors, based on detection results of the fuel pressure in a fuel supply system.

2. Description of Related Art

[0002] A fuel supply system is mounted in an internal combustion engine. The fuel supply system principally consists of a pressure accumulator to which fuel that has been pressurized is supplied, fuel injectors, connection passages that connect the pressure accumulator and the fuel injectors, and so forth.

[0003] A system disclosed in Japanese Patent Application Publication No. 2012-167617 (JP 2012-167617 A) is provided with a pressure sensor for detecting the fuel pressure inside the fuel supply system as described above, and is operable to detect operating characteristics of each of the fuel injectors based on the fuel pressure detected by the pressure sensor. The fuel pressure in the fuel supply system is reduced when the valve of the fuel injector is driven to be opened, and is increased by the amount of the reduction when the valve of the fuel injector is subsequently driven to be closed. Thus, the fuel pressure in the fuel supply system is temporarily reduced with opening and closing of the fuel injector. In the system as described above, the fuel pressure is detected, and the operating characteristics of the fuel injector are detected based on the detection result of the fuel pressure.

[0004] In the system of JP 2012-167617 A, as arithmetic processing for detecting the operating characteristics of the fuel injector, an operation to execute detection of the fuel pressure by the pressure sensor in short cycles, and an operation to analyze the detection result of the fuel pressure and specify the operating characteristics, are performed. Since these operations require a large computation load, it takes some amount of time to perform these operations. Therefore, when a length of time for which the above-described arithmetic processing can be performed is short, such as when the engine speed is high and the time it takes to complete one combustion cycle is short, the execution time of the arithmetic processing may not be sufficiently ensured, and the operating characteristics of the fuel injector may not be appropriately detected.

[0005] To ensure a sufficient time of execution of the arithmetic processing, it may be considered to cancel execution of the arithmetic processing at given intervals. However, if execution of the arithmetic processing is simply cancelled, the computation load may be reduced, but the frequency of detection of the operating characteristics of the fuel injector is reduced; therefore, the operating characteristics of the fuel injector may not be detected at appropriate intervals.

SUMMARY OF THE INVENTION

[0006] The invention provides fuel injection characteristic detection system and method that make it possible to detect operating characteristics of a fuel injector at appropriate intervals, while reducing a computation load involved with an operation to detect the operating characteristics of the fuel injector.

[0007] According to a first aspect of the invention, a detection system for detecting fuel injection characteristics of an internal combustion engine, the internal combustion engine including a fuel supply system that supplies fuel that has been pressurized to a fuel injector, and a pressure sensor that detects a fuel pressure inside the fuel supply system, the detection system includes an electronic control unit. The electronic control unit is configured to: (i) detect operating characteristics of the fuel injector based on detection results of the fuel pressure detected by the pressure sensor; (ii) execute multi-stage injection in which fuel is injected from the fuel injector a plurality of times in one combustion cycle; (iii) detect the operating characteristics associated with each injection of the multi-stage injection, and (iv) execute a first combustion cycle in which the operating characteristics associated with an initial-stage injection of the multi-stage injection are detected and the operating characteristics associated with at least one particular injection other than the initial-stage injection are not detected, and a second combustion cycle in which the operating characteristics associated with the initial-stage injection and said at least one particular injection in the multi-stage injection are detected.

[0008] In the multi-stage injection, the interval between the second-stage or subsequent injection, and the previous fuel injection immediately before the second-stage or subsequent injection, is relatively short. Therefore, the second-stage or subsequent injection is likely to be influenced by pulsation of the fuel pressure in the fuel supply system resulting from the previous fuel injection. In comparison, the interval between the initial-stage injection in the multi-stage injection, and the previous injection, is relatively long; therefore, the initial-stage injection is less likely to be influenced by pulsation of the fuel pressure resulting from the previous fuel injection. Therefore, during execution of the initial-stage injection, the operating characteristics of the fuel injector can be accurately detected, in a condition where the influence of the fuel pressure pulsation caused by the previous injection is suppressed or reduced.

[0009] According to the detection system as described above, in the first combustion cycle in which detection of the operating characteristics is restricted, the operating characteristics associated with the particular injections other than the initial-stage injection, namely, the second-stage and subsequent injections in which the detection accuracy is somewhat reduced, are not detected. As a result, the computation load involved with detection of the operating characteristics of the fuel injector can be reduced. The operating characteristics associated with the initial-stage injection are detected in both the first and second combustion cycles. Therefore, detection of the operating characteristics associated with the initial-stage injection in which highly accurate detection is possible can be surely carried out, and the operating characteristics of the fuel injector can be detected with high accuracy.

[0010] In the detection system as described above, the respective injection stages of the multi-stage injection are weighted in terms of the level of importance. Namely, detection of the operating characteristics for the initial-stage injection having high importance is surely carried out, while detection of the operating characteristics for the particular injections (the second-stage and subsequent fuel injections) having low importance is restricted, so that the computation load can be appropriately reduced. Since detection of the operating characteristics for the particular injections having low importance is intermittently carried out, the chances of detection of the operating characteristics for the particular injections are ensured even though the frequency of detection is reduced.

[0011] In the detection system as described above, the internal combustion engine may include a plurality of cylinders, and the electronic control unit may be configured to detect the operating characteristics for each cylinder of the internal combustion engine. The electronic control unit may be configured to set an interval of the cylinders to which the second combustion cycle is applied, in the order of the cylinders in which the multi-stage injection is performed, to a positive integer that is smaller than a total number of the cylinders of the internal combustion engine, and is equal to a given number that is not an integral multiple of a factor of the total number of the cylinders.

[0012] In the system in which the operating characteristics of the fuel injector are detected for each cylinder in the internal combustion engine having a plurality of cylinders, the length of time for which the operation to detect the operating characteristics of the fuel injector can be performed is more likely to be short, as compared with a system used in an internal combustion engine having a single cylinder, or a system in which the operating characteristics are detected only with respect to a particular cylinder or cylinders as a part of the plurality of cylinders.

[0013] When the above-described detection system is used in the internal combustion engine having a plurality of cylinders, it is possible to detect the operating characteristics of the fuel injector at appropriate intervals, while reducing the computation load involved with the operation to detect the operating characteristics of the fuel injector. Further, the second combustion cycle in which detection of the operating characteristics is not restricted can be prevented from being applied only to the particular cylinder or cylinders, out of all of the cylinders of the internal combustion engine. Therefore, the first combustion cycle in which detection of the operating characteristics is restricted, and the second combustion cycle in which detection of the operating characteristics is not restricted, can be applied to each cylinder of the engine at respective given intervals.

[0014] In the detection system as described above, the electronic control unit may be configured to execute detection of the operating characteristics, under a condition that an engine speed is equal to or higher than a predetermined speed. When the engine speed is high, the length of time it takes to complete one combustion cycle is reduced. Therefore, the length of time for which the arithmetic processing for detection of the operating characteristics of the fuel injector can be performed is shortened, and the computation load involved with the arithmetic processing is likely to be large.

[0015] In the detection system as described above, when the engine speed is high, and the computation load is likely to be large, the operating characteristics associated with the initial-stage injection having high importance are surely detected, and detection of the operating characteristics associated with the particular injections having low importance is restricted, so that the computation load can be appropriately reduced. Further, when the engine speed is low, and the computation load is relatively small, the operating characteristics associated with all of the injection stages are detected, without restricting detection of the operating characteristics associated with the particular injections. Consequently, the operating characteristics associated with fuel injection of each injection stage can be detected with high accuracy.

[0016] In the detection system as described above, it is important to grasp the variation waveform of the fuel pressure in a period in which the fuel pressure is temporarily reduced in accordance with opening and closing of the valve of the fuel injector, so as to detect the operating characteristics of the fuel injector. On the other hand, the necessity of grasping the variation waveform of the fuel pressure before opening of the fuel injector or after the temporary reduction of the fuel pressure in accordance with opening and closing of the fuel injector is low.

[0017] In the detection system as described above, the electronic control unit may be configured to use the fuel pressure detected by the pressure sensor in a first period from a point in time at which a valve-opening signal associated with each injection of the multi-stage injection is generated to the fuel injector to a point in time at which completion of each injection of the multi-stage injection is determined based on the detection results of the fuel pressure, and the electronic control unit may be configured not to use the fuel pressure detected by the pressure sensor at points in time outside the first period, for detection of the operating characteristics.

[0018] In the detection system as described above, the electronic control unit may be configured to use the fuel pressure detected by the pressure sensor in a second period from a point in time at which a predetermined time elapses from output of a valve-opening signal associated with each injection of the multi-stage injection to a point in time at which completion of each injection of the multi-stage injection is determined based on the detection results of the fuel pressure, and the electronic control unit may be configured not to use the fuel pressure detected by the pressure sensor at points in time outside the second period, for detection of the operating characteristics.

[0019] In the detection system as described above, the electronic control unit may be configured to use the fuel pressure detected by the pressure sensor in a third period from a point in time at which a valve-opening signal associated with each injection of the multi-stage injection is generated to the fuel injector to a point in time at which a predetermined time elapses from output of a valve-closing signal associated with each injection of the multi-stage injection, and the electronic control unit may be configured not to use the fuel pressure detected by the pressure sensor at points in time outside the third period, for detection of the operating characteristics.

[0020] In the detection system as described above, the electronic control unit may be configured to use the fuel pressure detected by the pressure sensor in a fourth period from a point in time at which a predetermined time elapses from output of a valve-opening signal associated with each injection of the multi-stage injection to a point in time at which a predetermined time elapses from output of a valve-closing signal associated with each injection of the multi-stage injection, and the electronic control unit may be configured not to use the fuel pressure detected by the pressure sensor at points in time outside the fourth period, for detection of the operating characteristics.

[0021] In the detection system as described above, the fuel pressure detected in the period in which a temporary reduction of the fuel pressure in the fuel supply system may be induced by opening and closing of the valve of the fuel injector is used for detection of the operating characteristics of the fuel injector, while the fuel pressure detected before opening of the fuel injector, or the fuel pressure detected after the temporary reduction of the fuel pressure with opening and closing of the fuel injector is completed, is not used for detection of the operating characteristics. Thus, the fuel pressure detected when the necessity of grasping the variation waveform of the fuel pressure is considerably low is not used in detection of the operating characteristics; therefore, the computation load involved with the operation to detect the operating characteristics of the fuel injector can be reduced accordingly. Furthermore, the variation waveform of the fuel pressure in the fuel supply system during opening of the fuel injector can be grasped, based on the fuel pressure detected when the fuel pressure in the fuel supply system is temporarily reduced. Accordingly, even though the variation waveform of the fuel pressure is not grasped in some periods, the operating characteristics of the fuel injector can be detected based on the variation waveform of the fuel pressure, and the frequency of detection of the operating characteristics is less likely or unlikely to be reduced.

[0022] According to a second aspect of the invention, a method of detecting fuel injection characteristics of an internal combustion engine, the internal combustion engine including a fuel supply system that supplies fuel that has been pressurized to a fuel injector, and a pressure sensor that detects a fuel pressure inside the fuel supply system, is provided. The method includes the steps of: detecting operating characteristics of the fuel injector based on detection results of the fuel pressure detected by the pressure sensor, executing multi-stage injection in which fuel is injected from the fuel injector a plurality of times in one combustion cycle, detecting the operating characteristics associated with each injection of the multi-stage injection, and executing a first combustion cycle in which the operating characteristics associated with an initial-stage injection of the multi-stage injection are detected and the operating characteristics associated with at least one particular injection other than the initial-stage injection are not detected, and a second combustion cycle in which the operating characteristics associated with the initial-stage injection and said at least one particular injection in the multi-stage injection are detected.

BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Features, advantages, and technical and industrial significance of exemplary embodiments of the invention will be described below with reference to the accompanying drawings, in which like numerals denote like elements, and wherein:

FIG. 1 is a view schematically showing the construction of an internal combustion engine in which a fuel injection characteristic detection system according to one embodiment of the invention is used;

FIG. 2 is a cross-sectional view showing a cross-section structure of a fuel injector, according to the embodiment of FIG. 1;

FIG. 3 is a time chart indicating the relationship between a driving pulse and the fuel injection rate, along with characteristic parameters of the fuel injector, according to the embodiment of FIG. 1;

FIG. 4 is a time chart indicating the relationship between a time waveform of the fuel pressure and a detected waveform of the fuel injection rate, according to the embodiment of FIG. 1;

FIG. 5 is a time chart indicating the relationship between the detected waveform of the fuel injection rate and the basic waveform of the fuel injection rate, according to the embodiment of FIG. 1;

FIG. 6 is a conceptual diagram showing a map structure of an initial-stage learning map in which the relationship among the target injection quantity, target injection pressure, and learning terms, during execution of pilot injection, is stored, according to the embodiment of FIG. 1;

FIG. 7 is a conceptual diagram showing a map structure of a main learning map in which the relationship among the target injection quantity, target injection pressure, and learning terms, during execution of main injection, is stored, according to the embodiment of FIG. 1;

FIG. 8 is a table indicating the relationship between the injection stage of multi-stage injection, and difference correction terms for each stage, according to the embodiment of FIG. 1;

FIG. 9 is a conceptual diagram showing one example of the manner of reflecting the difference correction terms and the learning terms;

FIG. 10 is a time chart indicating one example of the relationship between an output pattern of driving pulses, and the detected waveform of the fuel injection rate, according to the embodiment of FIG. 1;

FIG. 11 is a time chart showing specific examples of the manner of setting a first combustion cycle and a second combustion cycle, according to the embodiment of FIG. 1;

FIG. 12 is a graph indicating the relationship among the number of injection stages of multi-stage injection, the engine speed, and the computation load factor, according to the embodiment of FIG. 1;

FIG. 13 is a flowchart illustrating the execution procedure of a cycle setting process, according to the embodiment of FIG. 1;

FIG. 14 is a flowchart illustrating the execution procedure of a detected waveform forming process, according to the embodiment of FIG. 1;

FIG. 15 is a flowchart illustrating the execution procedure of a detected waveform forming process of a modified example;

FIG. 16 is a flowchart illustrating the execution procedure of a detected waveform forming process of another modified example; and

FIG. 17 is a flowchart illustrating the execution procedure of a detected waveform forming process of a still another modified example.

DETAILED DESCRIPTION OF EMBODIMENTS

[0024] A system of detecting fuel injection characteristics as one embodiment of the invention will be described. As shown in FIG. 1, an intake passage 12 is connected to cylinders 11 of an internal combustion engine 10. In operation, air is drawn into the cylinders 11 of the engine 10 via the intake passage 12. A diesel engine having a plurality of cylinders 11 (four cylinders #1, #2, #3, #4 in this embodiment) is employed as the internal combustion engine 10. In the engine 10, a direct-injection-type fuel injector 20 is mounted for each of the cylinders 11 (#1 - #4), and the fuel injector 20 is operable to directly inject fuel into the corresponding cylinder 11. The fuel is injected from the fuel injector 20 when it is driven to open its valve. The fuel thus injected comes into contact with intake air that is compressed and heated in the cylinder 11 of the engine 10, so that the fuel is ignited and burned. Then, in the engine 10, a piston 13 received in each cylinder 11 is pushed down due to energy produced with combustion of the fuel in the cylinder 11, so that a crankshaft 14 is rotated. Combustion gas produced from combustion in the cylinder 11 of the engine 10 is discharged as exhaust gas into an exhaust passage 15 of the engine 10.

[0025] The fuel injectors 20 are individually connected to a common rail 34 via respective branch passages 31a. The common rail 34 is connected to a fuel tank 32 via a supply passage 31b. A fuel pump 33 that feeds the fuel under pressure is provided in the supply passage 31b. In this embodiment, the fuel pumped and pressurized by the fuel pump 33 is stored in the common rail 34 that serves as a pressure accumulator, and is supplied to the inside of each fuel injector 20. In this embodiment, the fuel injectors 20, branch passages 31a, supply passage 31b, fuel pump 33, and the common rail 34 constitute a fuel supply system.

[0026] A return passage 35 is connected to each of the fuel injectors 20. The return passages 35 for the respective fuel injectors 20 are connected to the fuel tank 32. A part of the fuel in the fuel injector 20 is returned to the fuel tank 32 via the corresponding return passage 35.

[0027] The internal structure of each fuel injector 20 will be described. As shown in FIG. 2, a needle valve 22 is provided within a housing 21 of the fuel injector 20. The needle valve 22 is mounted such that it can reciprocate (move in the vertical direction in FIG. 2) within the housing 21. A spring 24 that normally urges the needle valve 22 toward injection holes 23 (downward in FIG. 2) is provided within the housing 21. In the housing 21, a nozzle chamber 25 is formed on one side of the needle valve 22 (downside in FIG. 2), and a pressure chamber 26 is formed on the other side of the needle valve 22 (upside in FIG. 2).

[0028] The injection holes 23 that communicate the nozzle chamber 25 with the outside of the housing 21 are formed through a bottom wall of the housing 21 which defines the nozzle chamber 25. The fuel is supplied from the branch passage 31a (the common rail 34) to the nozzle chamber 25 via an inlet passage 27. The nozzle chamber

25 and the branch passage 31a are connected to the pressure chamber 26 via a communication channel 28. The pressure chamber 26 is also connected to the return passage 35 (the fuel tank 32) via a discharge channel 30.

[0029] An electrically driven fuel injector is employed as the fuel injector 20. More specifically, a piezoelectric actuator 29 is provided within the housing 21 of the fuel injector 20. The piezoelectric actuator 29 consists of laminated piezoelectric elements that expand or contract in response to an input signal in the form of a driving pulse (a valve-opening signal or a valve-closing signal). A valve body 29a is mounted to the piezoelectric actuator 29. The valve body 29a is provided within the pressure chamber 26. As the piezoelectric actuator 29 operates to move the valve body 29a, the pressure chamber

26 is brought into communication with a selected one of the communication channel 28 (the nozzle chamber 25) and the discharge channel 30 (the return passage 35).

[0030] In the fuel injector 20 as described above, when a valve-closing signal is generated to the piezoelectric actuator 29, the piezoelectric actuator 29 contracts and moves the valve body 29a. With the valve body 29a thus moved, the communication channel 28 and the pressure chamber 26 are brought into communication with each other, and communication between the return passage 35 and the pressure chamber 26 is shut off. In this manner, the nozzle chamber 25 communicates with the pressure chamber 26 in a condition where the fuel in the pressure chamber 26 is inhibited from being discharged to the return passage 35 (the fuel tank 32). As a result, a pressure difference between the nozzle chamber 25 and the pressure chamber 26 is significantly reduced, and the needle valve 22 moves, under the bias force of the spring 24, to a position at which it blocks or closes the injection holes 23. Thus, the fuel injector 20 is placed in a state (valve-closed state) in which no fuel is injected from the injector 20.

[0031] On the other hand, when a valve-opening signal is generated to the piezoelectric actuator 29, the piezoelectric actuator 29 expands and moves the valve body 29a. With the valve body 29a thus moved, communication between the communication channel 28 and the pressure chamber 26 is shut off, and the return passage 35 and the pressure chamber 26 are brought into communication with each other. In this manner, a part of the fuel in the pressure chamber 26 is returned to the fuel tank 32 via the return passage 35, in a condition where the fuel is inhibited from flowing from the nozzle chamber 25 into the pressure chamber 26. As a result, the pressure of the fuel in the pressure chamber 26 is reduced, and a pressure difference between the pressure chamber

26 and the nozzle chamber 25 is increased. Due to the thus increased pressure difference, the needle valve 22 moves away from the injection holes 23 against the bias force of the spring 24. Thus, the fuel injector 20 is placed in a state (valve-open state) in which the fuel is injected from the injector 20.

[0032] A pressure sensor 51 that detects the fuel pressure PQ in the inlet passage

27 is integrally mounted on the fuel injector 20. Thus, the pressure sensor 51 can detect the fuel pressure at a location closer to the injection holes 23 of the fuel injector 20, as compared with a device that detects the fuel pressure at a position remote from the fuel injector 20, for example, the fuel pressure within the common rail 34 (see FIG. 1). With the pressure sensor 51 thus integrally mounted on the fuel injector 20, it is possible to accurately detect changes in the fuel pressure inside the fuel injector 20 when the injector 20 opens its valve. The pressure sensor 51 is provided for each fuel injector 20, namely, provided for each of the cylinders 11 (#1 - #4) of the engine 10.

[0033] As shown in FIG. 1, the internal combustion engine 10 is provided with various sensors for detecting operating conditions, as its peripheral equipment. The sensors include an intake air amount sensor 52, crank position sensor 53, and an accelerator position sensor 54, in addition to the pressure sensor 51 as described above. The intake air amount sensor 52 detects the amount of air that passes through the intake passage 12 (passage air amount GA). The crank position sensor 53 detects the rotational speed of the crankshaft 14 (engine speed NE). The accelerator position sensor 54 detects the operation amount (accelerator operation amount ACC) of an accelerating member (such as an accelerator pedal).

[0034] An electronic control unit 40 including an arithmetic processing unit is also provided as peripheral equipment of the engine 10. The electronic control unit 40 receives output signals of the above-indicated various sensors, and performs various computations based on the output signals. Then, the electronic control unit 40 performs various controls in connection with operation of the engine 10, including control of operation (injection quantity control) of the fuel injectors 20, and control of operation (injection pressure control) of the fuel pump 33, based on the result of computations. In this embodiment, the fuel pressure PQ is detected by the pressure sensor 51 at extremely short intervals (in this embodiment, 10 microseconds), and the fuel pressure PQ thus detected is stored in the electronic control unit 40 in association with the detection timing thereof.

[0035] In this embodiment, the injection pressure control is performed in the following manner. Initially, a control target value (target injection pressure) for the fuel pressure inside the common rail 34 is calculated based on the passage air amount GA and the engine speed NE. Also, the operating amount (fuel feed amount or fuel return amount) of the fuel pump 33 is adjusted so that the actual fuel pressure becomes equal to the target injection pressure. Through the adjustment of the operating amount of the fuel pump 33, the fuel pressure inside the common rail 34 is adjusted. As a result, the fuel injection pressure of the fuel injectors 20 is controlled to a pressure level suitable for the engine operating conditions.

[0036] In this embodiment, the injection quantity control is basically performed in the following manner. Initially, a control target value (target injection quantity TQ) for the fuel injection quantity is calculated, and an injection pattern is selected, based on operating conditions (more specifically, the accelerator operation amount ACC and the engine speed NE) of the engine 10. Then, various control target values for respective injections of the selected injection pattern are calculated, based on the target injection quantity TQ and the engine speed NE. Then, the fuel injectors 20 are individually driven to be opened according to the control target values.

[0037] In this embodiment, a plurality of multi-stage injection patterns as combinations of different numbers of pilot injections and a main injection are set in advance, and these injection patterns are stored in the electronic control unit 40. When the injection quantity control is performed, one of the injection patterns is selected. As various control target values, a control target value (target injection quantity) for the fuel injection quantity of each injection of the main injection, pilot injections, etc., and control target values for the execution timing of each injection are calculated. The control target values for the execution timing of each injection include the start time of the main injection, an interval between the pilot injections, and an interval between the pilot injection and the main injection, for example.

[0038] Then, a control target value (target injection duration TAU) for the valve-opening duration of the fuel injector 20, with respect to each fuel injection, is set according to a model formula, based on the target injection quantity and the fuel pressure PQ. In this embodiment, a physical model is constructed as a model of fuel supply system consisting of the common rail 34, each branch passage 31a, each fuel injector 20, etc., and the target injection duration TAU is calculated by use of the physical model. More specifically, the model formula is determined using the target injection quantity, fuel pressure PQ, learning terms, difference correction terms, etc., as parameters, and is stored in advance in the electronic control unit 40. The learning terms and difference correction terms will be described later. The target injection duration TAU is calculated according to the model formula as described above.

[0039] For each fuel injection, a driving pulse is generated from the electronic control unit 40 according to the control target values for the execution timing and the target injection duration TAU. Each of the fuel injectors 20 is driven to open its valve for each injection in response to the driving pulse. In this manner, the fuel is injected from each fuel injector 20 in an amount commensurate with the current engine operating conditions, in the injection pattern suitable for the current engine operating conditions, and supplied into the corresponding cylinder 11 of the engine 10. As a result, rotary torque commensurate with the engine operating conditions is applied to the crankshaft 14. Thus, in this embodiment, the fuel is injected from each fuel injector 20 at two or more times in one combustion cycle, to achieve multi-stage injections.

[0040] In this embodiment, a learning process of learning a plurality of characteristic parameters as operating characteristics of the fuel injector 20, based on the fuel pressure PQ detected by the pressure sensor 51, is performed. The learning process is executed provided that execution conditions for determining whether the engine 10 is in a stable operating state that is less likely or unlikely to change are satisfied. It is determined that the execution conditions are satisfied, when the amount of change of the target fuel injection quantity per unit period (e.g., a period for which the crankshaft 14 rotates several times, for example) is small, and the amount of change of the target injection pressure per unit period is small, for example.

[0041] FIG. 3 shows one example of characteristic parameters learned by the learning process. As shown in FIG. 3, the characteristic parameters used in this embodiment include a valve-opening delay time xd, the rate of increase of the injection rate Qup, the maximum injection rate Qmax, the valve-closing delay time xe, and the rate of reduction of the injection rate Qdn. More specifically, the valve-opening delay time xd is a period of time from a point in time at which a valve-opening signal (FIG. 3 (a)) is generated from the electronic control unit 40 to the fuel injector 20, to a point in time at which fuel injection from the fuel injector 20 is actually started. The rate of increase of the injection rate Qup is the rate at which the fuel injection rate (FIG. 3 (b)) increases after the valve-opening operation of the fuel injector 20 is started. The maximum injection rate Qmax is the maximum value of the fuel injection rate. The valve-closing delay time τε is a period of time from a point in time at which a valve-closing signal is generated from the electronic control unit 40 to the fuel injector 20, to a point in time at which a valve-closing operation of the fuel injector 20 (more specifically, movement of the needle valve 22 toward the valve-closed position) is started. The rate of reduction of the injection rate Qdn is the rate at which the fuel injection rate decreases after the valve-closing operation of the fuel injector 20 is started.

[0042] In the learning process, a time waveform (detected waveform) of the actual fuel injection rate is initially formed based on the fuel pressure PQ detected by the pressure sensor 51. The fuel pressure inside the fuel injector 20 (more specifically, the nozzle chamber 25) decreases with increase of a lift amount of the needle valve 22 when the fuel injector 20 is driven to open the valve. Then, the fuel pressure increases with reduction of the lift amount of the needle valve 22 when the fuel injector 20 is driven to close the valve. In this embodiment, the above-indicated valve-opening delay time xd, the rate of increase of the injection rate Qup, the maximum injection rate Qmax, the valve-closing delay time xe, and the rate of reduction of the injection rate Qdn are specified, based on changes in the fuel pressure inside the fuel injector 20 (more specifically, fuel pressure PQ) with time. Then, the time waveform (detected waveform) of the actual fuel injection rate is formed based on the specified values. The time waveform of the fuel pressure PQ may be formed based on values smoothed using a low-pass filter, or based on values corrected based on the fuel pressure PQ detected by the pressure sensor 51 corresponding to a non-injection cylinder.

[0043] FIG. 4 shows the relationship between the time waveform of the fuel pressure PQ and the detected waveform of the fuel injection rate. As shown in FIG. 4, an average value of the fuel pressure PQ (FIG. 4 (c)) in a given period Tl immediately before a valve-opening operation of the fuel injector 20 is started is initially calculated. The average value thus calculated is stored as a reference pressure Pbs. The reference pressure Pbs is used as a pressure corresponding to a fuel pressure inside the fuel injector 20 at the time of closing of the valve in the fuel injector 20.

[0044] Then, an operating pressure Pac (=Pbs-Pl) is calculated by subtracting a given pressure PI from the reference pressure Pb. The given pressure PI corresponds to an amount by which the fuel pressure PQ changes even though the needle valve 22 is held in the closed position when the fuel injector 20 is driven to open the valve or close the valve. Namely, the given pressure PI corresponds to an amount of change of the fuel pressure PQ which does not contribute to movement of the needle valve 22.

[0045] Thereafter, a straight line LI (in FIG. 4, a linear function on Cartesian coordinates, of which the vertical axis indicates the fuel injection rate and the horizontal axis indicates time) of which a difference from the fuel pressure PQ is minimized, in a period of decrease of the fuel pressure PQ immediately after the fuel injection is started, is obtained by the least-square method. Also, an intersection point A of the straight line LI and the operating pressure Pac is calculated. Then, a point in time corresponding to point AA that is earlier in time than the intersection point A by an amount of delay in detection of the fuel pressure PQ is specified as the time (injection start time Tos, FIG. 4 (b)) at which the fuel starts being injected from the fuel injector 20. The amount of delay in detection of the fuel pressure PQ is a period corresponding to a delay in the timing of change of the fuel pressure PQ relative to the timing of pressure change in the nozzle chamber 25 (see FIG. 2) of the fuel injector 20. Also, the delay is produced depending on a distance between the nozzle chamber 25 and the pressure sensor 51, for example. In this embodiment, a period of time from the time when the valve-opening signal (FIG. 4 (a)) is generated from the electronic control unit 40 to the above-mentioned injection start time Tos is specified as the valve-opening delay time xd.

[0046] A straight line L2 (in FIG. 4, a linear function on Cartesian coordinates, of which the vertical axis indicates the fuel injection rate (FIG. 4 (b)) and the horizontal axis indicates time) of which a difference from the fuel pressure PQ is minimized, in a period of increase of the fuel pressure PQ that increases after once decreasing with the start of the fuel injection, is obtained by the least-square method. Also, an intersection point B of the straight line L2 and the operating pressure Pac is calculated. Then, a point in time corresponding to point BB that is earlier in time than the intersection point B by the amount of delay in detection is specified as the time (injection stop time Tee) at which the fuel injection by the fuel injector 20 is stopped.

[0047] Further, an intersection point C of the straight line LI and the straight line L2 is calculated. Also, a difference between the fuel pressure PQ and the operating pressure Pac (a hypothetical pressure reduction ΔΡ [=Pac-PQ]) at the intersection point C is obtained. Also, a hypothetical maximum injection rate VRt (=APxGl) is calculated by multiplying the hypothetical pressure reduction ΔΡ by a gain Gl. Further, the maximum injection rate Qmax (=VRtxG2) is calculated by multiplying the hypothetical maximum injection rate VRt by a gain G2. Each of the gains Gl, G2 is set based on the target injection quantity and the target injection pressure. In this embodiment, as the target injection quantity and target injection pressure used for setting each of the gains Gl, G2, values set at the time when the pressure sensor 51 detects the fuel pressure PQ used for forming the detected waveform are employed.

[0048] Thereafter, a point CC that is earlier in time than the intersection point C by the amount of delay in detection is calculated. Also, a point D at which the fuel injection rate becomes equal to the hypothetical maximum injection rate VRt at the time CC is specified. Then, a point in time corresponding to the specified point D is specified as the time (valve-closing start time Tcs) at which the valve-closing operation of the fuel injector 20 is started. In this embodiment, a period of time from the time when the valve-closing signal is generated from the electronic control unit 40 to the fuel injector 20 to the above-mentioned valve-closing start time Tcs is specified as the valve-closing delay time xe.

[0049] A straight line L3 that connects the above-indicated point D with the injection start time Tos (more specifically, a point at which the fuel injection rate is equal to "0" at the time Tos) is obtained. Also, the slope of the straight line L3 (more specifically, the amount of increase of the fuel injection rate per unit time) is specified as the rate of increase of the injection rate Qup.

[0050] Further, a straight line L4 that connects the point D with the injection stop time Tee (more specifically, a point at which the fuel injection rate becomes equal to "0" at the time Tee) is obtained. Also, the slope of the straight line L4 (more specifically, the amount of reduction of the fuel injection rate per unit time) is specified as the rate of reduction of the injection rate Qdn.

[0051] In this embodiment, the time waveform having a trapezoidal shape is formed based on the thus specified valve-opening delay time xd, the rate of increase of the injection rate Qup, the maximum injection rate Qmax, the rate of reduction of the injection rate Qdn, and the valve-closing delay time xe. The trapezoidal time waveform thus formed is used as the detected waveform of the fuel injection rate.

[0052] In the learning process of this embodiment, a basic waveform of the fuel injection rate is calculated, based on various calculation parameters, such as the target injection quantity, control target values for the execution timing, and the target injection pressure. In this embodiment, the relationship between an engine operating region determined by these calculation parameters, and the basic waveform suitable for the operating region, is obtained in advance based on results of various experiments and simulations, and stored in the electronic control unit 40. The electronic control unit 40 calculates the basic waveform from the above relationship, based on various calculation parameters.

[0053] FIG. 5 shows one example of the basic waveform as described above. As shown in FIG. 5 (a) and (b), a trapezoidal waveform specified by the valve-opening delay time xdb, the rate of increase of the injection rate Qupb, the maximum injection rate Qmaxb, the valve-closing delay time xeb, and the rate of reduction of the injection rate Qdnb is set as the basic waveform.

[0054] In the learning process of this embodiment, learning terms for a plurality of characteristic parameters of the fuel injector 20 are learned based on the relationship between the detected waveform and the basic waveform. Initially, during operation of the engine 10, the detected waveform and the basic waveform are compared with each other, and a difference in each characteristic parameter between these waveforms is sequentially calculated. More specifically, a difference Axd (=xdb-xd) in the valve-opening delay time, a difference AQup (=Qupb-Qup) in the rate of increase of the injection rate, a difference AQmax (=Qmaxb-Qmax) in the maximum injection rate, a difference AQdn (=Qdnb-Qdn) in the rate of reduction of the injection rate, and a difference Axe (=xeb-xe) in the valve-closing delay time are calculated as the differences in the respective characteristic parameters. Then, weighted averages of these differences Axd, AQup, AQmax, AQdn, Δτε are calculated, and the weighted averages are stored in the electronic control unit 40 as learning terms Gtd, GQup, GQmax, GQdn, Gxe used for compensating for variations in operating characteristics of the fuel injector 20.

[0055] Also, in this embodiment, a plurality of learning regions defined by the fuel injection pressure (more specifically, the target injection pressure) and the fuel injection quantity (more specifically, the target injection quantity) are determined. Then, the learning terms are learned and stored for each of these regions.

[0056] As shown in FIG. 6, the electronic control unit 40 stores a map (initial-stage learning map) that specifies the relationship among the target injection quantity, target injection pressure, and the learning terms, in a learning region used in pilot injection, namely, in a learning region in which the target injection quantity is small. When the target injection duration TAU of the pilot injection is calculated, the learning terms are calculated from the initial-stage learning map as shown in FIG. 6, based on the target injection quantity and target injection pressure of the fuel injection for which the target injection duration TAU is to be calculated, and used.

[0057] Also, as shown in FIG. 7, the electronic control unit 40 stores a map (main learning map) that specifies the relationship among the target injection quantity, target injection pressure, and the learning terms, in a learning region used in main injection, namely, in a learning region including a region in which the target injection quantity is small, and a region in which the target injection quantity is large. When the target injection duration TAU of the main injection is calculated, the learning terms are calculated from the main learning map as shown in FIG. 7, based on the target injection quantity and target injection pressure of the main injection, and used.

[0058] The pressure fluctuations in the fuel supply system during execution of the second-stage or subsequent fuel injection of the multi-stage injection include pulsation of the fuel pressure which arises from a previous fuel injection effected immediately before the fuel injection in question. Therefore, if the above-mentioned learning terms are learned during execution of the main injection, simply based on the fuel pressure PQ detected by the pressure sensor 51, the accuracy with which the learning terms are learned may be reduced due to an influence of the pulsation of the fuel pressure.

[0059] Thus, in this embodiment, in a region in which the fuel injection quantity is small (more specifically, in a region (indicated by hatched lines in FIG. 7) in which pilot injection is executed), the learning terms are learned, based on the fuel pressure PQ detected by the pressure sensor 51 during execution of the initial stage of pilot injection. Also, the learning terms stored in the initial-stage learning map (see FIG. 6) are learned, based on the fuel pressure PQ detected by the pressure sensor 51 during execution of the initial stage of pilot injection. Thus, the learning terms are accurately learned while the influence of the fuel pressure pulsation caused by another injection is significantly reduced or minimized. Further, in the main learning map, in a region in which the fuel injection quantity is large (more specifically, in a region in which pilot injection is not executed), the learning terms are learned, based on the fuel pressure PQ detected by the pressure sensor 51 during execution of main injection.

[0060] If the learning terms thus learned are reflected as they are by calculation of the target injection duration TAU, an error may appear in the fuel injection quantity, due to fuel pressure pulsation that arises from the previous injection. In this embodiment, difference correction terms Kxd, KQup, KQmax, KQdn, Rte for correcting the error are calculated. Initially, differences Axd, AQup, AQmax, AQdn, Axe in the respective parameters for the fuel injection for which the difference correction terms are to be calculated are detected. Also, the learning terms Gxd, GQup, GQmax, GQdn, Gxe reflected in the calculation of the target injection duration TAU of the fuel injection are read. Then, differences (=Axd-Gxd, AQup-GQup, AQmax-GQmax, AQdn-GQdn, Δτε-Gte) between the differences in the respective parameters and the learning terms are calculated. The differences thus calculated are temporarily stored as the difference correction terms Kxd, KQup, KQmax, KQdn, Kte. The process of calculating the difference correction terms in this manner is carried out for each stage of the second-stage and subsequent fuel injections in the multi-stage injection.

[0061] The pulsation of the fuel pressure caused by the previous injection effected immediately before the injection in question is not constant, but varies depending on the interval between the injection events, fuel injection pressure, and the fuel injection quantity of the previous injection. Therefore, if the above-mentioned learning terms are learned and the difference correction terms are calculated, without taking account of the fuel pressure pulsation caused by the previous injection as described above, the time waveform of the fuel pressure PQ and the detected waveform suffer from unnecessary changes, resulting in reduction of the accuracy with which the learning terms are learned and the accuracy with which the difference correction terms are calculated.

[0062] In this embodiment, when the detected waveform is formed with respect to the second-stage or subsequent injection, a pressure waveform (correction waveform) that can cancel out the pressure pulsation produced by the previous injection is superimposed on the time waveform of the fuel pressure PQ based on which the detected waveform is formed, so as to curb reduction of the accuracies as described above. Through the above-described operation, an influence of the fuel pressure pulsation arising from the previous injection is removed from the detected waveform. Therefore, appropriate values are detected as differences in the above-indicated respective parameters, so that appropriate values are learned as the learning terms.

[0063] The correction waveform is calculated for each of the second-stage and subsequent injections of the multi-stage injection, based on the injection pattern of the combustion cycle including the fuel injection on which the correction is performed, target injection quantity of each injection, intervals between successive injection events, and the target injection pressure. In this embodiment, the relationship among the injection pattern, target injection quantity of each injection, intervals between successive injection events, target injection pressure, and the correction waveform suitable for each of the second-stage and subsequent injections of the multi-stage injection is obtained in advance, based on results of various experiments and simulations, and is stored in the electronic control unit 40. Then, the correction waveform for the second-stage or subsequent injection of the multi-stage injection is calculated based on the above relationship, and used.

[0064] FIG. 8 indicates the relationship between the injection stage of the multi-stage injection and the difference correction terms. As shown in FIG. 8, through execution of the process of calculating the difference correction terms, values calculated based on the (N+l)-stage fuel injection are stored as difference correction terms K(N+1) corresponding to the (N+l)-stage fuel injection, where "N" denotes a natural number. For example, values calculated based on the second-stage fuel injection are stored as difference correction terms K2 corresponding to the second-stage injection, and values calculated based on the third-stage injection are stored as difference correction terms K3 corresponding to the third-stage injection. The initial value ("0" in this embodiment) is set as difference correction terms corresponding to an injection stage that was not carried out in the multi-stage injection.

[0065] Thus, in this embodiment, the difference correction terms are not calculated in association with the injection position, like correction terms corresponding to main injection, or correction terms corresponding to pilot injection executed immediately before the main injection. Rather, the difference correction terms are calculated in association with the order of injection, like difference correction terms K2 corresponding to the second-stage injection, or difference correction terms K3 corresponding to the third-stage injection.

[0066] In this embodiment, the learning terms Grd, GQup, GQmax, GQdn, Gxe, and the difference correction terms Ktd, KQup, KQmax, KQdn, Kxe are respectively used as calculation parameters for use in calculation of the target injection duration TAU based on the model formula as described above. By calculating the target injection duration TAU for fuel injection of each stage of the multi-stage injection, an influence of variations in the operating characteristics due to chronological changes of the fuel injector 20, and an influence of fuel pressure pulsation caused by the previous injection are respectively compensated for. The difference correction terms Ktd, KQup, KQmax, KQdn, Rte are reflected in the above-indicated model formula, in calculation of the target injection duration TAU of fuel injection in a combustion cycle (reflection combustion cycle) following the combustion cycle (calculation combustion cycle) including the fuel injection for which the difference correction terms are calculated. In this embodiment, the process of calculating the learning terms based on the fuel pressure PQ and the process of calculating the difference correction terms are performed, based on the output signal of the pressure sensor 51 corresponding to each of the cylinders 11 (#1 - #4) of the internal combustion engine 10.

[0067] FIG 9 shows one example of the manner of reflecting the learning terms and the difference correction terms in the reflection combustion cycle. In the example shown in FIG. 9, three-stage fuel injection consisting of two-stage pilot injection and main injection is carried out, in the calculation combustion cycle and the reflection combustion cycle. Therefore, the difference correction terms K2 corresponding to the second-stage injection are calculated based on the second-stage pilot injection in the calculation combustion cycle, and the difference correction terms K3 corresponding to the third-stage injection are calculated based on the main injection.

[0068] Then, as shown in FIG. 9, the difference correction terms K2 are reflected in calculation of the target injection duration TAU for the second-stage pilot injection in the reflection combustion cycle, and the difference correction terms K3 are reflected in calculation of the target injection duration TAU for the main injection in the reflection combustion cycle. Also, the learning terms reflected in calculation of the target injection duration TAU for each pilot injection are calculated based on the initial-stage learning map (FIG. 6), and the learning terms reflected in calculation of the target injection duration TAU for the main injection are calculated based on the main learning map (FIG. 7).

[0069] In the system of this embodiment, as arithmetic processing for detecting characteristic parameters of the fuel injector 20, an operation to detect the fuel pressure PQ with the pressure sensor 51 in short cycles, and an operation to analyze the time waveform of the detected fuel pressure PQ and specify the characteristic parameters are performed. Since a large computation load is applied to the electronic control unit 40 in these operations, it takes a certain amount of time to perform the operations. Therefore, when the length of time for which the above arithmetic processing can be performed is short, such as when the engine speed NE is high and the time it takes to complete one combustion cycle is short, for example, a sufficient amount of execution time for the arithmetic processing cannot be ensured, and learning of the learned terms and calculation of the difference correction terms may not be adequately accomplished.

[0070] In the system of this embodiment, the operation to detect the characteristic parameters of the fuel injector 20 is performed for each of the cylinders 11 of the internal combustion engine 10. Therefore, the length of time for which the arithmetic processing for detecting the characteristics parameters can be performed is likely to be short, as compared with the system in which the same processing is performed only for a particular cylinder or cylinders..

[0071] In the system of this embodiment, a period (detection period TA) is determined which starts when a valve-opening signal (driving pulse) is generated to the fuel injector 20, and ends (at the above-indicated injection stop time Tee) when completion of fuel injection by the fuel injector 20 is determined based on the detected waveform of the fuel injection rate. Then, in the process (detected waveform forming process) of forming the detected waveform of the fuel injection rate, the fuel pressure PQ detected by the pressure sensor 51 in the above-mentioned detection period TA is used, while the fuel pressure PQ detected at points in time outside the detection period TA is not used.

[0072] In the following, the effects provided by executing the detected waveform forming process in the above manner will be described. In FIG. 10 (a) and (b), one example of the relationship between an output pattern of driving pulses and the detected waveform of the fuel injection rate is shown.

[0073] As shown in FIG. 10, when the detected waveform of the fuel injection rate is formed, fuel pressures PQ detected by the pressure sensor 51 upon and after output of a valve-opening signal (FIG. 10 (a)) to the fuel injector 20, out of the fuel pressures PQ stored in the electronic control unit 40 in association with the detection timing, start being read (time til, tl3, tl5). At this time, the fuel pressures PQ are read in the order of detection, and reading of the fuel pressures PQ is repeated until the time when the injection stop time Tee (see FIG. 4) is specified based on the fuel pressure PQ (time til - tl2, tl3 - tl4, tl5 - tl6). Then, if the injection stop time Tee (see FIG. 4) is specified (time tl2, tl4, tl6) based on the fuel pressure PQ, reading of the fuel pressures PQ is stopped, and the detected waveform of the fuel injection rate is formed based on the fuel pressures PQ thus read.

[0074] In order to detect respective characteristic parameters of the fuel injector 20, it is important to grasp the variation waveform of the fuel pressure PQ in a period in which the fuel pressure is temporarily reduced with opening and closing of the fuel injector 20. On the other hand, the necessity of grasping the variation waveform of the fuel pressure PQ before opening of the fuel injector 20 or after the temporary reduction of the fuel pressure with opening and closing of the fuel injector 20 is considerably low.

[0075] According to the system of this embodiment, the fuel pressures PQ detected in the detection periods TA (time til - tl2, tl3 - tl4, tl5 - tl6) in which the fuel pressure in the fuel supply system may be temporarily reduced due to opening and closing of the fuel injector 20 are used in the detected waveform forming process. On the other hand, the fuel pressures PQ detected before opening of the fuel injector 20 or after temporary reduction of the fuel pressure with opening and closing of the fuel injector 20 (before time til, tl2 - tl3, tl4 - tl5, after tl6) are not used in the detected waveform forming process. Thus, the fuel pressure PQ detected when the necessity of grasping the variation waveform of the fuel pressure PQ is considerably low is not used in the detected waveform forming process; therefore, the computation load involved with the operation to read the fuel pressure PQ and the computation load involved with the operation to form the detected waveform can be reduced accordingly. Further, the detected waveform of the fuel injection rate can be formed based on the fuel pressure PQ detected in the detection period TA in which the fuel pressure in the fuel supply system is temporarily reduced, and the characteristic parameters of the fuel injector 20 can be detected based on the detected waveform. Accordingly, even though the variation waveform of the fuel pressure PQ is not grasped in some periods, the variation waveform of the fuel pressure PQ is grasped in the periods needed for detection of the characteristic parameters; therefore, reduction of the frequency of detection of the characteristic parameters can be curbed.

[0076] In the system of this embodiment, when the number of injection stages of the multi-stage injection is a predetermined number of stages, and the engine speed NE is equal to or higher than a predetermined speed, two types of combustion cycles, namely, a first combustion cycle and a second combustion cycle, are set in the same cylinder 11. In the first combustion cycle, the characteristic parameters for particular injections other than the initial-stage injection in the multi-stage injection are not detected, namely, detection of the characteristic parameters for the particular injections is restricted. In the second combustion cycle, detection of the characteristic parameters for the particular injections is not restricted.

[0077] More specifically, the first combustion cycle and the second combustion cycle are set, when the engine speed NE is equal to or higher than a criterial speed Jl (e.g., 2200 rpm) during execution of four-stage injection. In this case, as shown in FIG. 11 (a), the second combustion cycle is applied to one cylinder 11 each time the first combustion cycle is successively applied to two cylinders 11, in the order of the cylinders (in this embodiment, cylinder 11 (#1)→ (#3)→ (#4)→ (#2)) in which the multi-stage injection is performed. Thus, the second combustion cycle is applied every three cylinders, or at intervals of three cylinders. In the example shown in FIG. 11 (a), in the order of the cylinders in which the multi-stage injection is performed, "the first combustion cycle" is applied to the cylinder 11 (#1), and "the second combustion cycle" is applied to the subsequent cylinder 11 (#3). Then, "the first combustion cycle" is applied to the subsequent cylinder 11 (#2), and "the second combustion cycle" is applied again to the subsequent cylinder 11 (#1).

[0078] Thus, in the system of this embodiment, the interval of the cylinders 11 to which the second combustion cycle is applied, in the order of the cylinders in which multi-stage injection is performed, is set to a positive integer that is smaller than the total number of cylinders ("4" in this embodiment) of the engine 10, and is equal to a given number ("3" in this embodiment) that is not an integral multiple of a factor of the total number of cylinders. With this arrangement, the second combustion cycle is prevented from being applied to only a particular cylinder or cylinder 11, as a part of all cylinders 11 of the engine 10, and the first combustion cycle and the second combustion cycle can be applied to each cylinder 11 of the engine 10 at respective given intervals.

[0079] In the first combustion cycle, the characteristic parameters are detected, and the detected waveform is formed, only with respect to the initial-stage injection in the multi-stage injection. In the second combustion cycle, the characteristic parameters are detected with respect to the particular injections (the second-stage injection through the fourth-stage injection) in addition to the initial-stage injection, and the detected waveforms associated with all of the injection stages are formed.

[0080] During execution of three-stage injection, the first combustion cycle and the second combustion cycle are set when the engine speed NE is equal to or higher than a criterial speed J2 (e.g., 3800 rpm). In this case, as shown in FIG. 11 (b), the second combustion cycle is applied every three cylinders, or at intervals of three cylinders, in the order of the cylinders in which the multi-stage injection is performed, and the first combustion cycle is applied to the other cylinders 11. In the first combustion cycle, the characteristic parameters are detected only with respect to the initial-stage injection in the multi-stage injection, and the detected waveform associated with the initial-stage injection is formed. In the second combustion cycle, the characteristic parameters are detected with respect to the particular injections (the second-stage injection and the third-stage injection) in addition to the initial-stage injection, and the detected waveforms associated with all of the injection stages are formed.

[0081] The effects provided by setting the first combustion cycle and the second combustion cycle in the above manner will be described. If the engine speed NE is increased as described above, the time allocated to one combustion cycle is reduced, and the period of time for which the arithmetic processing for operation control of the fuel injector 20 can be performed is shortened. Also, if the number of injection stages in the multi-stage injection is increased, the total period of time for which the fuel pressure is temporarily reduced with opening and closing of the fuel injector 20 is increased, and the period required to form the detected waveforms is also increased with the increase of the total period of pressure reduction. Accordingly, when the number of injection stages of the multi-stage injection is large, and the engine speed NE is high, the computation load involved with the arithmetic processing for detecting the characteristic parameters of the fuel injector 20 is likely to be large. [0082] Since each interval between the second-stage or subsequent injection and the previous fuel injection immediately before the second-stage or subsequent injection in the multi-stage injection is short, the second-stage or subsequent injection is likely to be influenced by pulsation of the fuel pressure PQ in the fuel supply system resulting from the previous fuel injection. On the other hand, an interval between the initial-stage injection in the multi-stage injection, and the previous injection, is large; therefore, the influence of the pulsation of the fuel pressure PQ resulting from the previous injection is small. Therefore, during execution of the initial-stage injection, it is possible to accurately detect the characteristic parameters of the fuel injector 20 in a condition where the influence of the fuel pressure pulsation due to the previous injection is suppressed, as compared with the time when the second-stage or subsequent- stage injection is executed.

[0083] In the system of this embodiment, when the computation load is increased with increase of the engine speed NE, detection of the characteristic parameters is restricted with respect to the particular injections other than the initial-stage injection, namely, the second-stage and subsequent injections in which the detection accuracy is somewhat reduced, so that the computation load involved with detection of the characteristic parameters of the fuel injector 20 can be reduced. Further, even when the engine speed NE is high, the characteristic parameters for the initial-stage injection are detected in all of the combustion cycles. Therefore, the characteristic parameters can be surely detected with respect to the initial-stage injection for which the parameters can be detected with high accuracy, and the characteristic parameters of the fuel injector 20 can be detected with high accuracy.

[0084] Thus, the respective injection stages in the multi-stage injection are weighted in terms of the level of importance. Namely, detection of the characteristic parameters for the initial-stage injection having high importance is surely carried out, while detection of the characteristic parameters for the particular injections (the second-stage and subsequent fuel injections) having low importance is restricted, so that the computation load can be appropriately reduced. Further, even when the engine speed NE is high, detection of the characteristic parameters for the particular injections having low importance in the same cylinder 11 is not totally inhibited from being carried out, but is intermittently carried out each time the cylinder 11 in which the multi-stage injection is performed switches from one to another twelve times (namely, each time the crankshaft 14 makes three revolutions). Therefore, the chances of detection of the characteristic parameters for the particular injections can be ensured even though the frequency of the detection is reduced.

[0085] Since the initial-stage injection in the multi-stage injection is a pilot injection, and how the pilot injection is performed has a great influence on the ignition performance of the fuel, it is important to keep high accuracy in adjustment of the injection quantity in the pilot injection. Also, the quantity of fuel injected in the pilot injection is considerably small, and, in the fuel injection of such a minute quantity of fuel, an error is likely to arise in the injection quantity, in view of the structure of the fuel injector 20. Therefore, in order to keep high accuracy in adjustment of the injection quantity in the initial-stage injection, it is desirable to detect the characteristic parameters for the initial-stage injection at a high frequency. In the system of this embodiment, the characteristic parameters for the initial-stage injection are detected at a high frequency.

[0086] On the other hand, the main injection is one of the second-stage and subsequent fuel injections in the multi-stage injection. If an error arises in the injection quantity in the main injection, the output performance and exhaust performance of the internal combustion engine 10 are largely influenced by the error; therefore, it is desirable to keep high accuracy in adjustment of the injection quantity. The injection quantity is relatively large in the most part of injection quantity region of the main injection, and chronological changes in characteristics of the fuel supply system including the fuel injector 20 are a main factor of the error in the injection quantity, in the region in which the injection quantity is large. Therefore, the error in the injection quantity changes relatively slowly. Thus, it is possible to keep high accuracy in adjustment of the fuel quantity even if the frequency at which the characteristic parameters for the main injection are detected is reduced.

[0087] Also, the detection results of the characteristic parameters for the initial-stage injection are reflected, in the injection stages (the second-stage and subsequent injections in the multi-stage pilot injection) other than the initial-stage injection and the main injection in the multi-stage injection, so that the accuracy in adjustment of the injection quantity is less likely or unlikely to be reduced. Thus, it is less important to detect the characteristic parameters for the injection stages other than the initial-stage injection and the main injection in the multi-stage injection, and the frequency of carrying out the detection can be reduced.

[0088] In the system of this embodiment, the characteristic parameters for the particular injections other than the initial-stage injection in the multi-stage injection, i.e., the second-stage and subsequent fuel injections, are detected at a reduced frequency. FIG. 12 indicates the relationship among the number of injection stages of the multi-stage injection in the system in which the detected waveform is formed only in the predetermined period TA, the engine speed NE, and the computation load factor (the ratio of the actual computation load to the maximum value of the computation power) of the electronic control unit 40. In FIG. 12, one-dot chain lines indicate the above-described relationship in the system in which only the second combustion cycle is set, and solid lines indicate the above-described relationship in the system in which the first combustion cycle and the second combustion cycle are set. Also, broken lines in FIG. 12 indicate an execution limit of multi-stage injection, which is determined by the performance of a drive circuit of the fuel injector 20.

[0089] In the system of this embodiment, the fuel pressure PQ used in formation of the detected waveform of the fuel injection rate is limited to values detected in the detection periods TA (see FIG. 10). Therefore, the computation load factor is reduced as compared with the system in which the detected waveform of the fuel injection rate is formed over the whole period. It makes it possible to expand the operating region (in this embodiment, a region in which the computation load factor is equal to or lower than a given percentage (e.g., 80%)) in which the learning process and the process of calculating the difference correction terms can be performed. In the system of this embodiment, even when the two-stage injection is executed at the execution limit (C2 in FIG. 12) of the two-stage injection which is determined by the performance of the drive circuit of the fuel injector 20, the detected waveform associated with all of the injection stages can be formed even if the first combustion cycle and the second combustion cycle are not set.

[0090] Since the first combustion cycle and the second combustion cycle are set when the three-stage injection is executed, the operation to read the fuel pressure PQ and the operation to form the detected waveform, with respect to the particular injections (the second-stage injection, the third-stage injection), are not performed in the first combustion cycle. As a result, the computation load is reduced as indicated by arrow Al in FIG. 12; therefore, a region in which the injection waveform of the three-stage injection can be detected is expanded by an amount indicated by arrow A2 in FIG. 12. Accordingly, a situation where the injection waveform of the three-stage injection cannot be detected under the operation control of the fuel injector 20 is less likely or unlikely to occur, for the reason that the detected waveform of any of the three injection stages, such as that of the main injection, cannot be formed. In the system of this embodiment, the first combustion cycle and the second combustion cycle are set when the engine speed NE is equal to or higher than the criterial speed J2, so that the injection waveform of the three-stage injection can be detected up to the execution limit (C3 in FIG. 12) of the injection waveform detection of the three-stage injection determined by the performance of the drive circuit of the fuel injector 20.

[0091] Since the first combustion cycle and the second combustion cycle are set when the four-stage injection is executed, the operation to read the fuel pressure PQ and the operation to form the detected waveform, with respect to the particular injections (the second-stage injection through the fourth-stage injection), are not performed in the first combustion cycle. As a result, the computation load is reduced as indicated by arrow A3 in FIG. 12; therefore, a region in which the injection waveform of the four-stage injection can be detected is expanded by an amount indicated by arrow A4 in FIG. 12. Accordingly, a situation where the injection waveform of the four-stage injection cannot be detected under the operation control of the fuel injector 20 is less likely or unlikely to occur, for the reason that the detected waveform of any of the four injection stages, such as that of the main injection, cannot be formed. In the system of this embodiment, if it is assumed to set only the second combustion cycle without setting the first combustion cycle, the criterial speed Jl becomes the execution limit of the injection waveform detection of the four-stage injection. In the system of this embodiment, the first combustion cycle and the second combustion cycle are set when the engine speed NE is equal to or higher than the criterial speed Jl, so that the injection waveform of the four-stage injection can be detected up to the execution limit (C4 in FIG. 12) of the injection waveform detection of the four-stage injection determined by the performance of the drive circuit of the fuel injector 20.

[0092] In the following, the process (cycle setting process) of setting the second combustion cycle and the first combustion cycle, and the process of forming the detected waveform will be described in detail. Initially, the procedure of the cycle setting process will be described with reference to FIG. 13.

[0093] A series of steps indicated in the flowchart of FIG. 13 are executed by the electronic control unit 40, as an interrupt routine executed at given intervals, under a condition that Condition I or Condition II as follows is satisfied. Condition I is that the engine is in an operating region in which three-stage injection is performed, and the engine speed NE is equal to or higher than the criterial speed J2. Condition II is that the engine is in an operating region in which four-stage injection is performed, and the engine speed NE is equal to or higher than the criterial speed Jl.

[0094] As shown in FIG. 13, in this process, at the time when the ignition cylinder (specifically, the cylinder 11 in which the fuel is injected from the fuel injector 20) is switched from one to another (step Sll: YES), "1" is added to a count value of an injection process mode counter (step S12).

[0095] If the count value is "1" or "2" (step S13: NO), an initial-stage detection mode is set (step S14), and then the process of FIG. 13 is finished. If the initial-stage detection mode is set, the combustion cycle of the ignition cylinder after switching becomes the first combustion cycle.

[0096] If, on the other hand, the count value of the injection process mode counter is "3" (step S13: YES), the all-stage detection mode is set (step S15), and the count value of the injection process mode counter is reset to "0" (step SI 6). Thereafter, the process of FIG. 13 is finished. If the all-stage detection mode is set, the combustion cycle of the ignition cylinder after switching becomes the second combustion cycle.

[0097] Referring next to FIG. 14, the execution procedure of the detected waveform forming process will be described. A series of steps indicated in the flowchart of FIG. 14 conceptually illustrates the execution procedure of the detected waveform forming process, and is executed by the electronic control unit 40 each time the fuel injection of each stage of multi-stage injection is carried out.

[0098] As shown in FIG. 14, in this process, it is initially determined whether the all-stage detection mode is set (step S21). If the all-stage detection mode is set (step S21: YES), the detected waveform is formed (step S22 through step S24), without depending on the injection stage for which the detected waveform is formed. More specifically, the fuel pressures PQ detected after a valve-opening signal associated with a fuel injection for which the detected waveform is to be formed is generated to the fuel injector 20 are read in the order of detection thereof (step S22). Then, the fuel pressures PQ are repeatedly read (step S22), until the injection stop time Tee (see FIG. 4) is specified based on the fuel pressures PQ thus read. If the injection stop time Tee is specified based on the fuel pressures PQ thus read (step S23: YES), the detected waveform is formed (step S24) based on the fuel pressures PQ detected in the detection period TA from the time when the valve-opening signal is generated to the time when the injection stop time Tee is determined based on the detection results of the fuel pressure PQ. Then, the process of FIG. 14 is finished

[0099] On the other hand, when the initial-stage detection mode is set (step S21: NO), the detected waveform is formed (step S22 through step S24), only in the case where the injection stage for which the detected waveform is to be formed is the initial-stage injection (step S25: YES). Accordingly, when the initial-stage detection mode is set (step S21: NO), and the injection stage for which the detected waveform is to be formed is not the initial-stage injection (step S25: NO), the detected waveform is not formed (step S22 through step S24 are skipped), and the process of FIG. 14 is finished.

[0100] In the system of this embodiment, leaning of the learning terms and calculation of the difference correction terms are performed based on the detected waveform. As explained above, the effects as described below are provided according to this embodiment.

[0101] In the same cylinder 11, the second combustion cycle and the first combustion cycle are set. Accordingly, the characteristic parameters associated with the initial-stage injection having high importance are surely detected, while detection of the characteristic parameters associated with the second-stage and subsequent injections having low importance is restricted, so that the computation load can be appropriately reduced. Although the frequency of detection of the characteristic parameters associated with the second-stage and subsequent fuel injections is reduced, the chances of the detection can be ensured.

[0102] The interval of the cylinders 11 to which the second combustion cycle is applied, in the order of the cylinders in which the multi-stage injection is performed, is set to the predetermined number "3" that is a positive integer smaller than the total number "4" of the cylinders of the engine 10 and is not an integral multiple of a factor of the total number of the cylinders. Therefore, the first combustion cycle and the second combustion cycle can be applied to each of the cylinders 11 of the engine 10 at respective given intervals.

[0103] In the same cylinder 11, the second combustion cycle and the first combustion cycle are set, under the condition that the engine speed NE is equal to or higher than the criterial speed Jl during execution of the four-stage injection, or the condition that the engine speed NE is equal to or higher than the criterial speed J2 during execution of the three-stage injection. Therefore, when the computation load is likely to be large because the engine speed NE is high, the characteristic parameters associated with the initial-stage injection having high importance are surely detected, while detection of the characteristic parameters associated with the particular injections having low importance is restricted, so that the computation load can be appropriately reduced. Further, when the engine speed NE is low, and the computation load is relatively small, the characteristic parameters associated with all of the injection stages are detected, without skipping detection of the characteristic parameters associated with the particular injections, so that the characteristic parameters associated with the fuel injection of each injection stage can be detected with high accuracy.

[0104] The fuel pressures PQ detected by the pressure sensor 51 in the detection period TA from the time when the valve-opening signal is generated to the fuel injector 20 to the time when the injection stop time Tee is determined based on the detection results of the fuel pressure PQ are used in the process of forming the detected waveform, while the fuel pressures PQ detected at points in time outside the detection period TA are not used in the process of forming the detected waveform. With this arrangement, the fuel pressures PQ detected before opening of the valve of the fuel injector 20 and the fuel pressures PQ detected after the temporary reduction of the fuel pressure with opening and closing of the fuel injector 20 is completed are not used in the process of forming the detected waveform; therefore, the computation load involved with the operation to read the fuel pressure PQ and the computation load involved with the operation to form the detected waveform are reduced accordingly. Further, while the variation waveform of the fuel pressure PQ is not grasped in some periods, the variation waveform of the fuel pressure PQ can be grasped in periods required for detection of the characteristic parameters; therefore, reduction of the frequency of detection of the characteristic parameters can be curbed.

[0105] The invention may be embodied by modifying the illustrated embodiment as follows. The detection period TA may be replaced with a detection period TB that starts when a predetermined time T2 elapses from output of a valve-opening signal to the fuel injector 20, and ends when the injection stop time Tee is determined based on the detection results of the fuel pressure PQ. The fuel pressures PQ detected by the pressure sensor 51 in the detection period TB may be used in the process of forming the detected waveform of the fuel injection rate, and the fuel pressures PQ detected at points in time outside the detection period TB may not be used in the same process.

[0106] In the system of the illustrated embodiment, if the output timing of the valve-opening signal to the fuel injector 20 is found, the time at which the valve of the fuel injector 20 starts being opened can be estimated, based on the output timing and the operating characteristics (such as the valve-opening delay time) of the fuel injector 20. Therefore, the system in which the above-mentioned detection period TB is employed, in place of the detection period TA, can provide effects equivalent to those provided by the system of the illustrated embodiment. Further, since the period in which the fuel pressures PQ are read can be shortened, as compared with the system of the illustrated embodiment, the total number of fuel pressures PQ used in formation of the detected waveform can be reduced. Accordingly, the computation load of the arithmetic processing associated with detection of the characteristic parameters of the fuel injector 20 can be further reduced. In the system as described above, a given length of time by which the detection period is shortened so that a part of the fuel pressures PQ detected in the valve-opening delay time is not read while making it possible to appropriately form the detected waveform may be obtained in advance based on the result of various experiments or simulations, and may be stored as the predetermined time T2 in the electronic control unit 40.

[0107] FIG. 15 shows the execution procedure of the detected waveform forming process in the system as described above. FIG. 15 mainly shows a portion of the detected waveform forming process which is different from a corresponding portion of the same process shown in FIG. 14. In the following description, the same reference numerals or step numbers are assigned to the same steps as those of the detected waveform forming process shown in FIG. 14, and these steps will not be described in detail. As shown in FIG. 15, in this process, when the detected waveform is formed (step S21: YES, or step S25: YES), the fuel pressures PQ detected after the predetermined time T2 elapses from output of the valve-opening signal to the fuel injector 20 are read in the order in which they are detected (step S32). Then, the fuel pressures PQ are repeatedly read (step S32), until the injection stop time Tee is specified based on the fuel pressures PQ thus read (step S23: NO). Then, if the injection stop time Tee is specified based on the fuel pressures PQ thus read (step S23: YES), the detected waveform is formed (step S34), based on the fuel pressures PQ detected in the detection period TB from the time when the predetermined time T2 elapses from the output of the valve-opening signal, to the time when the injection stop time Tee is determined based on the detection results of the fuel pressures PQ. Thereafter, the process of FIG. 16 is finished.

[0108] The detection period TA may be replaced with a detection period TC that starts when a valve-opening signal is generated to the fuel injector 20, and ends when a predetermined time T3 elapses from output of a valve-closing signal to the fuel injector 20. Then, the fuel pressures PQ detected by the pressure sensor 51 in the detection period TC may be used in the process of forming the detected waveform of the fuel injection rate, and the fuel pressures PQ detected at points in time outside the detection period TC may not be used in the same process.

[0109] In the system of the illustrated embodiment, if the output timing of the valve-closing signal to the fuel injector 20 is found, the time at which the valve of the fuel injector 20 is placed in the closed state can be estimated, based on the output timing and the operating characteristics (such as the valve-closing delay time, and the rate of reduction of the injection rate) of the fuel injector 20. Therefore, the system in which the above-mentioned detection period TC is employed, in place of the detection period TA, can provide effects equivalent to those provided by the system of the illustrated embodiment. In the system as described above, a given length of time by which the detection period is shortened so that the fuel pressures PQ detected after the fuel injector 20 is placed in the closed state are not read while making it possible to appropriately form the detected waveform may be obtained in advance based on the result of various experiments or simulations, and may be stored as the predetermined time T3 in the electronic control unit 40.

[0110] FIG. 16 shows the execution procedure of the detected waveform forming process in the system as described above. FIG. 16 mainly shows a portion of the detected waveform forming process which is different from a corresponding portion of the same process shown in FIG. 14. In the following description, the same reference numerals or step numbers are assigned to the same steps as those of the detected waveform forming process shown in FIG. 14, and these steps will not be described in detail. As shown in FIG. 16, in this process, when the detected waveform is formed (step S21: YES, or step S25: YES), the fuel pressures PQ detected in the detection period TC are read (step S42). Then, the detected waveform is formed (step S43), based on the fuel pressures PQ thus read, namely, the fuel pressures PQ detected in the detection period TC from the time when the valve-opening signal associated with the fuel injection for which the detected waveform is to be formed is generated, to the time when the predetermined time T3 elapses from the output of the valve-closing signal associated with the same fuel injection. Thereafter, the process of FIG. 16 is finished.

[0111] The detection period TA may be replaced with a detection period TD that starts when a predetermined time T2 elapses from output of a valve-opening signal to the fuel injector 20, and ends when a predetermined time T3 elapses from output of a valve-closing signal to the fuel injector 20. Then, the fuel pressures PQ detected by the pressure sensor 51 in the detection period TD may be used in the process of forming the detected waveform of the fuel injection rate, and the fuel pressures PQ detected at points in time outside the detection period TD may not be used in the same process.

[0112] FIG. 17 shows the execution procedure of the detected waveform forming process in the system as described above. FIG. 17 mainly shows a portion of the detected waveform forming process which is different from a corresponding portion of the same process shown in FIG. 14. In the following description, the same reference numerals or step numbers are assigned to the same steps as those of the detected waveform forming process shown in FIG. 14, and these steps will not be described in detail. As shown in FIG. 17, in this process, when the detected waveform is formed (step S21: YES, or step S25: YES), the fuel pressures PQ detected in the detection period TD are read (step S52). Then, the detected waveform is formed (step S53), based on the fuel pressures PQ thus read, namely, the fuel pressures PQ detected in the detection period TD from the time when the predetermined time T2 elapses from the output of the valve-opening signal associated with the fuel injection for which the detected waveform is to be formed, to the time when the predetermined time T3 elapses from the output of the valve-closing signal associated with the same fuel injection. Thereafter, the process of FIG. 17 is finished.

[0113] If the computation load is appropriately suppressed or reduced, the arrangement of using only the fuel pressures PQ detected in the detection period TA, TB, TC, TD in the process of forming the detected waveform of the fuel injection rate may be omitted. Namely, the fuel pressures PQ detected before opening of the valve of the fuel injector 20 or after the temporary reduction of the fuel pressure caused by opening and closing of the fuel injector 20 is completed may be used in the detected waveform forming process.

[0114] In the system of the illustrated embodiment, after injection or post injection may be carried out after the main injection. In this case, the characteristic parameters for only the initial-stage injection may be detected in the first combustion cycle, and the characteristic parameters for all of the injection stages may be detected in the second combustion cycle.

[0115] In the system as described above, when five-stage injection consisting of three-stage pilot injections, main injection and after injection (or post injection) is carried out, the characteristic parameters may be detected in the following manner. Namely, when the engine speed NE is lower than a criterial speed J3, the characteristic parameters for all of the five injection stages are detected. On the other hand, when the engine speed NE is equal to or higher than the criterial speed J3, the first combustion cycle and the second combustion cycle are set in the same cylinder 11. Then, the characteristic parameters for only the initial-stage injection are detected in the first combustion cycle, and the characteristic parameters for the first-stage injection through the fourth-stage injection (main injection) except the last-stage injection (after injection) are detected in the second combustion cycle. As the criterial speed J3, the lower-limit speed in the range of the engine speed NE in which the computation load factor of the electronic control unit 40 exceeds a given percentage (e.g., 80%), in the case where only the second combustion cycle is set but the first combustion cycle is not set during execution of the five-stage injection, is set. In this system, when the engine speed NE becomes high, the characteristic parameters for the fifth-stage injection (after injection) cannot be detected, but the characteristic parameters for the initial-stage injection are surely detected. Also, the computation load can be appropriately reduced, by restricting detection of the characteristic parameters for the second-stage and subsequent fuel injections having lower importance. Although the frequency of detection of the characteristic parameters associated with the second-stage through fourth-stage fuel injections including the main injection is reduced, the chances of the detection can be ensured.

[0116] The first combustion cycle and the second combustion cycle may be set when the engine speed NE is equal to or higher than a certain criterial speed, without depending on the number of injection stages. Also, the first combustion cycle and the second combustion cycle may be set, without depending on the engine speed NE.

[0117] In the learning process, the weighted averages of the respective differences Δτά, AQup, AQmax, AQdn, Δτε in the characteristic parameters may not be calculated, but the differences in the characteristic parameters themselves may be stored as learning terms (red, GQup, GQmax, GQdn, Gxe.

[0118] The learning regions may be defined by only one of the target injection quantity and the target injection pressure. As the parameters based on which the learning regions are defined, the target injection quantity and the target injection pressure are not necessarily used, but the engine speed NE, passage air amount GA, accelerator operation amount ACC, intake air amount, etc., may be used.

[0119] The correction waveform may not be calculated and reflected. The characteristic parameters indicative of the operating characteristics of the fuel injector 20 may be changed as desired. For example, only one, two, three, or four of the valve-opening delay time xd, the rate of increase of the injection rate Qup, the maximum injection rate Qmax, the rate of reduction of the injection rate Qdn, and the valve-closing delay time xe may be used as a characteristic parameter or parameters. Also, the time at which the fuel injection rate reaches the maximum injection rate, the time at which the fuel injection rate starts decreasing from the maximum injection rate, the time at which the fuel injection rate becomes equal to "0", etc. may be newly employed as characteristic parameters. [0120] The pressure sensor 51 may not be directly mounted to the fuel injector 20, but the manner of mounting the pressure sensor 51 may be changed as desired, provided that the pressure indicative of the fuel pressure inside the fuel injector 20 (specifically, inside the nozzle chamber 25), in other words, the fuel pressure that changes in accordance with change of the fuel pressure inside the fuel injector 20, can be properly detected. More specifically, the pressure sensor 51 may be mounted in a portion (branch passage 31a) of the fuel supply passage between the common rail 34 and the fuel injector 20 in the fuel supply passage, or may be mounted in the common rail 34.

[0121] The fuel injector 20 of the type driven by the piezoelectric actuator 29 may be replaced with a fuel injector of a type driven by an electromagnetic actuator including a solenoid coil, for example.

[0122] The fuel injection characteristic detection system as described above is not limitedly used in the internal combustion engine having four cylinders, but may be used in engines having one to three cylinders, or five or more cylinders. When the invention is applied to a single-cylinder engine, the first combustion cycle and the second combustion cycle may be set at given intervals; for example, the first combustion cycle and the second combustion cycle may be alternately set. When the invention is applied to a multi-cylinder engine, the interval of the cylinders to which the combustion cycle (the second combustion cycle) in which detection of the characteristic parameters is not restricted is applied, in the order of the cylinders in which multi-stage injection is performed, may be set to a positive integer that is smaller than the total number of cylinders of the engine, and is equal to a given number that is not an integral multiple of a factor of the total number of cylinders. For example, the given number may be set to "5" when the invention is applied to a six-cylinder engine, and the given number may be set to any one of "3", "5", and "7" when the invention is applied to an eight-cylinder engine, while the given number may be set to any one of "5", "7", and "11" when the invention is applied to a twelve-cylinder engine.

[0123] The fuel injection characteristic detection system as described above is not limitedly used in the diesel engine, but may be used in gasoline engines using gasoline fuel, and natural gas engines using natural gas fuel.

Claims

CLAIMS:

1. A detection system for detecting fuel injection characteristics of an internal combustion engine (10), the internal combustion engine including a fuel supply system that supplies fuel that has been pressurized to a fuel injector (20), and a pressure sensor (51) that detects a fuel pressure inside the fuel supply system, the detection system characterized by comprising:

an electronic control unit (40) configured to:

(i) detect operating characteristics of the fuel injector based on detection results of the fuel pressure detected by the pressure sensor;

(ii) execute multi-stage injection in which fuel is injected from the fuel injector a plurality of times in one combustion cycle;

(iii) detect the operating characteristics associated with each injection of the multi-stage injection; and

(iv) execute a first combustion cycle in which the operating characteristics associated with an initial-stage injection of the multi-stage injection are detected and the operating characteristics associated with at least one particular injection other than the initial-stage injection are not detected, and a second combustion cycle in which the operating characteristics associated with the initial-stage injection and said at least one particular injection in the multi-stage injection are detected.

2. The detection system according to claim 1, characterized in that:

the internal combustion engine includes a plurality of cylinders (11);

the electronic control unit is configured to detect the operating characteristics for each cylinder of the internal combustion engine; and

the electronic control unit is configured to set an interval of the cylinders to which the second combustion cycle is applied, in the order of the cylinders in which the multi-stage injection is performed, to a positive integer that is smaller than a total number of the cylinders of the internal combustion engine, and is equal to a given number that is not an integral multiple of a factor of the total number of the cylinders.

3. The detection system according to claim 1 or 2, characterized in that the electronic control unit is configured to execute detection of the operating characteristics, under a condition that an engine speed (NE) is equal to or higher than a predetermined speed.

4. The detection system according to any one of claims 1 to 3, characterized in that: the electronic control unit is configured to use the fuel pressure detected by the pressure sensor in a first period (TA) from a point in time at which a valve-opening signal associated with each injection of the multi-stage injection is generated to the fuel injector to a point in time at which completion of each injection of the multi-stage injection is determined based on the detection results of the fuel pressure; and

the electronic control unit is configured not to use the fuel pressure detected by the pressure sensor at points in time outside the first period, for detection of the operating characteristics.

5. The detection system according to any one of claims 1 to 3, characterized in that: the electronic control unit is configured to use the fuel pressure detected by the pressure sensor in a second period (TB) from a point in time at which a predetermined time (T2) elapses from output of a valve-opening signal associated with each injection of the multi-stage injection to a point in time at which completion of each injection of the multi-stage injection is determined based on the detection results of the fuel pressure; and the electronic control unit is configured not to use the fuel pressure detected by the pressure sensor at points in time outside the second period, for detection of the operating characteristics.

6. The detection system according to any one of claims 1 to 3, characterized in that: the electronic control unit is configured to use the fuel pressure detected by the pressure sensor in a third period (TC) from a point in time at which a valve-opening signal associated with each injection of the multi-stage injection is generated to the fuel injector to a point in time at which a predetermined time (T3) elapses from output of a valve-closing signal associated with each injection of the multi-stage injection; and

the electronic control unit is configured not to use the fuel pressure detected by the pressure sensor at points in time outside the third period, for detection of the operating characteristics.

7. The detection system according to any one of claims 1 to 3, characterized in that: the electronic control unit uses the fuel pressure detected by the pressure sensor in a fourth period (TD) from a point in time at which a predetermined time (T2) elapses from output of a valve-opening signal associated with each injection of the multi-stage injection to a point in time at which a predetermined time (T3) elapses from output of a valve-closing signal associated with each injection of the multi-stage injection; and

the electronic control unit is configured not to use the fuel pressure detected by the pressure sensor at points in time outside the fourth period, for detection of the operating characteristics.

8. A method of detecting fuel injection characteristics of an internal combustion engine (10), the internal combustion engine including a fuel supply system that supplies fuel that has been pressurized to a fuel injector (20), and a pressure sensor (51) that detects a fuel pressure inside the fuel supply system, characterized by comprising:

detecting operating characteristics of the fuel injector based on detection results of the fuel pressure detected by the pressure sensor;

executing multi-stage injection in which fuel is injected from the fuel injector a plurality of times in one combustion cycle;

detecting the operating characteristics associated with each injection of the multi-stage injection; and

executing a first combustion cycle in which the operating characteristics associated with an initial-stage injection of the multi-stage injection are detected and the operating characteristics associated with at least one particular injection other than the initial-stage injection are not detected, and a second combustion cycle in which the operating characteristics associated with the initial-stage injection and said at least one particular injection in the multi-stage injection are detected.