JP2013079611A - Fuel injection control device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To prevent stability or smoothness of operations of an engine from being impaired due to large variation in correction values of fuel injection amounts.SOLUTION: A feedback learning correction value for correcting the fuel injection amount is calculated for each of learning areas R1-R4 set by an intake air pressure and an engine speed, using a feedback learning processing. A first smoothing processing is performed when ignition is turned ON, such that high variability is removed which is caused among the feedback learning correction values due to an unexpected error of the feedback learning processing or the like. A second smoothing processing is performed, if, after the feedback learning processing is performed to any of the learning areas, high variability is caused between a feedback learning correction value for a learning area in which the feedback learning processing has already been completed and another feedback learning correction value for a learning area in which the feedback learning processing has not yet been completed, such that the high variability is removed.

Description

  The present invention relates to a fuel injection control device that controls a fuel injection amount of a fuel injection device that supplies fuel to a combustion chamber of an engine.

  The fuel injection control device that controls the fuel injection amount of the fuel injection device that supplies fuel to the combustion chamber of the engine calculates a correction value for each of a plurality of operation zones divided by the intake air pressure and the engine speed. One having a function of finely correcting the fuel injection amount using the correction value is known.

  For example, an engine control device described in Patent Document 1 below calculates a fuel injection amount based on a basic fuel injection amount, a feedback correction value, and a feedback learning correction value. The basic fuel injection amount is calculated based on, for example, fuel temperature, fuel pressure, intake air temperature, intake air pressure, throttle opening, engine speed, engine temperature, and the like. The feedback correction value is a value for correcting the fuel injection amount according to the difference between the ideal air-fuel ratio and the actual air-fuel ratio, and is calculated based on the oxygen concentration in the exhaust gas. The feedback learning correction value is a value for finely correcting the fuel injection amount in accordance with the intake air pressure and the engine speed, and is calculated for each of the nine operation zones divided by the intake air pressure and the engine speed.

  In this engine control apparatus, nine operation zones are set as follows. That is, assuming an XY plane composed of an X axis indicating the engine speed and a Y axis indicating the intake air pressure, threshold values X1 and X2 are set for the engine speed, and the engine speed is less than X1. , A section where the engine speed is X1 or more and less than X2, and a section where the engine speed is X2 or more. Further, threshold values Y1 and Y2 are set for the suction air pressure, and the suction air pressure is divided into a section where the suction air pressure is less than Y1, a section where the suction air pressure is Y1 or more and less than Y2, and a section where the suction air pressure is Y2 or more. As a result, the XY plane is divided into nine operation zones by three sections for engine speed and three sections for intake air pressure.

  The CPU that performs arithmetic processing in the engine control device performs the processing when the correction of the fuel injection amount by the feedback correction value becomes too small while the state in which the intake air pressure and the engine speed do not change greatly continues for a predetermined time. The feedback learning correction value corresponding to the operating zone to which the intake air pressure and the engine speed belong is increased, and the correction of the fuel injection amount by the feedback correction value is enhanced using the increased feedback learning correction value. On the other hand, if the correction of the fuel injection amount by the feedback correction value becomes over-air while the state where the intake air pressure and the engine speed do not change significantly continues for a predetermined time, the intake air pressure and the engine speed belong. The feedback learning correction value corresponding to the operation zone is decreased, and the correction of the fuel injection amount by the feedback correction value is suppressed using the reduced feedback learning correction value.

  By the way, since the feedback learning correction value is calculated after a state in which the intake air pressure and the engine speed do not change significantly as described above continues for a predetermined time, it takes time to calculate the feedback learning correction value. Furthermore, since the feedback learning correction value is calculated for each of the nine operation zones, it may take a long time to complete the calculation of the feedback learning correction value for all the operation zones.

  When the fuel property changes due to refueling of the engine, the CPU of the engine control device calculates and updates the feedback learning correction value so that the fuel injection amount is adapted to the changed fuel property. At this time, it may take a long time to calculate and update the feedback learning correction value due to the above-described circumstances.

  Therefore, the engine control device detects refueling, and after refueling is detected, the feedback learning correction value corresponding to all the operation zones is simultaneously updated using the feedback learning correction value calculated first. A batch update function is provided to reduce the time for calculating and updating feedback learning correction values.

JP 2002-309978 A

  For example, when the engine control device of Patent Document 1 is applied to an engine of an automobile or a motorcycle, a section where the actual suction air pressure belongs and a section where the actual suction air pressure belongs are generated in a plurality of sections regarding the suction air pressure. Similarly, in a plurality of sections concerning the engine speed, a section where the actual engine speed belongs and a section where the actual engine speed belongs are high. As a result, among the plurality of operation zones, there are an operation zone in which the feedback learning correction value is frequently updated and an operation zone in which the feedback learning correction value is not updated. For this reason, it is conceivable that the feedback learning correction value varies greatly for each operation zone. In addition, some feedback learning correction values may become excessive or small due to an unexpected error in the calculation process, and this may cause large variations in the feedback learning correction values.

  For example, when a change in intake air pressure or engine speed is repeated with the operation of an automobile or motorcycle, the feedback learning correction value corresponding to each of a plurality of operation zones is used to calculate the fuel injection amount instead of If the feedback learning correction value varies greatly for each operation zone, the fuel injection amount may fluctuate excessively, and the stability and smoothness of engine operation may be impaired.

  The present invention has been made in view of, for example, the above-described problems, and an object of the present invention is to prevent the stability or smoothness of engine operation from being impaired due to large variations in the correction value of the fuel injection amount. An object of the present invention is to provide a fuel injection control device that can be used.

  In order to solve the above problems, a first fuel injection control device of the present invention is a fuel injection control device that controls a fuel injection amount of a fuel injection device that supplies fuel to a combustion chamber of an engine, and includes intake air, Fuel injection amount calculating means for calculating the fuel injection amount based on the state of the fuel and the engine, and correction value storage means for storing a plurality of correction values respectively corresponding to a plurality of ranges specified by the intake air pressure and the engine speed. Correction value calculating means for performing learning by recognizing a change in the state of fuel injection or air-fuel ratio within a predetermined time for each of the plurality of ranges, and calculating the plurality of correction values based on the learning result; Correction value leveling means for performing leveling processing for reducing the deviation of each correction value in a plurality of correction values, and the fuel injection amount calculated by the fuel injection amount calculation means, Characterized in that it comprises a fuel injection amount correction means for correcting using serial least one of the correction values of the plurality of correction values.

  According to the first fuel injection control device of the present invention, it is possible to prevent the variation (deviation) of the correction value of the fuel injection amount from increasing, and to maintain or improve the stability or smoothness of engine operation. be able to.

  In order to solve the above-described problem, the second fuel injection control device of the present invention is the above-described first fuel injection control device of the present invention, wherein the correction value leveling means includes the plurality of correction values in the leveling process. An average value of the maximum value and the minimum value of the correction values is calculated, a correction value that is greater than or equal to a predetermined value among the plurality of correction values is specified, and the specified correction value is Increasing or decreasing to approach the average value.

  According to the second fuel injection control device of the present invention, by increasing / decreasing only the correction value having a large variation among the plurality of correction values, the correction value has a large correction value while suppressing the accuracy of the correction value from decreasing. Variations can be eliminated.

  In order to solve the above-described problem, the third fuel injection control device of the present invention is the above-described first or second fuel injection control device of the present invention, wherein the correction value leveling means is configured to start the engine operation. The leveling process is performed.

  According to the third fuel injection control device of the present invention, since a large variation in the correction value can be removed when starting use (reuse) of the engine, for example, at the time of ignition ON of an automobile or a motorcycle, While the engine is actually used, the stability and smoothness of the engine can be maintained or increased.

  In order to solve the above problems, a fourth fuel injection control device according to the present invention is the above-described first fuel injection control device according to the present invention, wherein the correction value leveling means is any one of the plurality of ranges. When the learning is completed for the range, in the leveling process, the learning is not completed so as to approach the correction value corresponding to the range for which the learning is completed among the plurality of ranges. The corresponding correction value is increased or decreased.

  According to the fourth fuel injection control device of the present invention, it is possible to prevent a large variation in the correction value during learning. For example, the fuel property may change due to fuel replacement or refueling, and the correction value may vary greatly due to learning based on the changed fuel property. In this case, in the fuel injection control device to which the present invention is not applied, the difference between the correction value corresponding to the range in which learning is completed and the correction value corresponding to the range in which learning is not completed becomes excessive during learning. When a plurality of correction values are viewed as a whole, it is considered that a large variation occurs in the correction values. According to the fourth fuel injection control device of the present invention, it is possible to prevent such a large variation in the correction value.

  In order to solve the above-described problems, a fifth fuel injection control device of the present invention is the above-described first fuel injection control device of the present invention, wherein the plurality of ranges include at least both intake air pressure and engine speed. The first range is relatively small, the second range is relatively large and the engine speed is relatively small, the third range is relatively small and the engine speed is relatively large, and the intake air pressure and the engine The correction value leveling means corresponds to the first range in the leveling process when the learning is completed for the first range, including a fourth range in which both the rotational speeds are relatively large. The correction value corresponding to the range in which the learning is not completed among the second, third, and fourth ranges is increased or decreased so as to approach the correction value.

  According to the fifth fuel injection control device of the present invention, it is possible to suppress a decrease in accuracy of each correction value when removing a large variation in the correction value during learning. For example, during steady running in an automobile or motorcycle, both the suction air pressure and the engine speed are relatively small, and the suction air pressure and the engine speed do not change significantly. In a situation where the intake air pressure and the engine speed do not change significantly, learning is continuously performed, so that a stable or highly accurate correction value is calculated. In other words, in the first range where both the intake air pressure and the engine speed are relatively small, it is easier to calculate a correction value that is more stable or more accurate than the other ranges. Therefore, the correction value corresponding to the first range in which learning has been performed is brought close to the correction value corresponding to another range in which learning has not been performed, so that a large variation in correction values can be removed and each correction value can be corrected. Can be suppressed.

  According to the present invention, it is possible to prevent the stability or smoothness of operation of the engine from being impaired due to the large variation in the correction value of the fuel injection amount.

It is explanatory drawing which shows the engine with which the fuel-injection control apparatus by embodiment of this invention was applied. 1 is a block diagram illustrating a fuel injection control device, various sensors, and a fuel injection device according to an embodiment of the present invention. It is explanatory drawing which shows the learning table in the fuel-injection control apparatus by embodiment of this invention. It is a flowchart which shows the feedback correction value calculation process in the fuel-injection control apparatus by embodiment of this invention. It is a flowchart which shows the feedback learning process in the fuel-injection control apparatus by embodiment of this invention. It is a flowchart which shows the 1st leveling process in the fuel-injection control apparatus by embodiment of this invention. It is a flowchart which shows the 2nd leveling process in the fuel-injection control apparatus by embodiment of this invention.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

(Engine to which the fuel injection control device is applied)
FIG. 1 shows an engine of a motorcycle, for example, to which a fuel injection control device according to an embodiment of the present invention is applied. In FIG. 1, an engine 1 includes a cylinder block 2, a cylinder head 3, a cylinder head cover 4, a piston 5, and the like. The engine 1 also drives the intake port 7 and the exhaust port 8 communicating with the combustion chamber 6, the intake valve 9 and the exhaust valve 10 for opening and closing the intake port 7 and the exhaust port 8, and the intake valve 9 and the exhaust valve 10 for opening and closing. An intake camshaft 11 and an exhaust camshaft 12 are provided.

  An air cleaner 13, an intake pipe 14, a throttle body 15, an intake manifold 16, and the like are provided on the intake side of the engine 1, and an intake passage 17 that communicates with the intake port 7 is formed by these. The throttle body 15 is provided with a throttle valve 18. On the other hand, an exhaust manifold 20, a first catalyst 21, an exhaust pipe 22, a second catalyst 23, and the like are provided on the exhaust side of the engine 1, thereby forming an exhaust passage 24 that communicates with the exhaust port 8. Yes.

  A spark plug 31 is attached to the cylinder head 3 so as to face the combustion chamber 6, and an ignition coil 32 is attached to the cylinder head cover 4 for causing the spark plug 31 to ignite.

  Further, a fuel tank 34 that stores fuel such as gasoline is connected to the engine 1 via a fuel supply pipe 36 and the like. A fuel pump 35 and the like are provided in the fuel tank 34. A fuel injection device 37 that supplies fuel toward the combustion chamber 6 is attached to the cylinder head 3 so as to face the intake port 7.

  An intake air temperature sensor 41 for detecting the intake air temperature and an air flow sensor 42 for detecting the intake air amount are attached to the downstream side of the air cleaner 13, and the throttle opening is set near the throttle valve 18. A throttle opening sensor 43 for detection is attached, and an intake air pressure sensor 44 for detecting intake air pressure is provided in the middle of the intake passage 17. On the other hand, an oxygen concentration sensor 45 that detects the oxygen concentration in the exhaust gas is provided in the middle of the exhaust passage 24. On the other hand, a fuel temperature sensor 46 for detecting the temperature of the fuel and a fuel pressure sensor 47 for detecting the pressure of the fuel are provided in the middle of the fuel passage formed by the fuel supply piping 36 (see FIG. 2). . Further, the engine 1 is provided with an engine rotation sensor 48 for detecting the engine speed (see FIG. 2).

  The sensors 41 to 48, the fuel injection device 37, the ignition coil 32, and the like are electrically connected to a fuel injection control device 51 that controls the fuel injection amount of the fuel injection device 37. The fuel injection control device 51 is mounted on a motorcycle as a part of an engine control unit, for example.

(Fuel injection control device)
FIG. 2 shows the internal configuration of the fuel injection control device 51, and the ignition coil 32, various sensors 41 to 48, and the fuel injection device 37 that are electrically connected to the fuel injection control device 51. The fuel injection control device 51 calculates the basic fuel injection amount, and calculates the final fuel injection amount by correcting the basic fuel injection amount with the feedback correction value and the feedback learning correction value. Further, the fuel injection control device 51 performs two kinds of leveling processes on the feedback learning correction value.

  As shown in FIG. 2, the fuel injection control device 51 includes a CPU (Central Processing Unit) 52 and a memory 53 that is, for example, a nonvolatile semiconductor memory element. The CPU 52 reads a computer program stored in the memory 53 and executes it to thereby execute a basic fuel injection amount calculation unit 54, a feedback correction value calculation unit 55, a feedback learning correction value calculation unit 56, and a first correction value. It functions as a leveling processing unit 57, a second correction value leveling processing unit 58, and a final fuel injection amount calculation unit 59.

  The basic fuel injection amount calculation unit 54 calculates the basic fuel injection amount based on, for example, fuel temperature, fuel pressure, intake air temperature, intake air pressure, throttle opening, engine speed, engine temperature, and the like. When calculating the basic fuel injection amount, the basic fuel injection amount calculation unit 54 includes a fuel temperature sensor 46, a fuel pressure sensor 47, an intake air temperature sensor 41, and an intake air pressure sensor electrically connected to the fuel injection control device 51. 44. Detects fuel temperature, fuel pressure, intake air temperature, intake air pressure, throttle opening, engine speed, etc. using throttle opening sensor 43, engine speed sensor 48, engine temperature sensor (not shown), etc. To do.

  The feedback correction value calculation unit 55 calculates a feedback correction value based on the oxygen concentration in the exhaust gas. The feedback correction value is a value for correcting the fuel injection amount in accordance with the difference between the ideal air fuel ratio and the actual air fuel ratio. The feedback correction value is a feedback correction set value (for example, a proportional value for eliminating the deviation between the ideal air-fuel ratio and the actual air-fuel ratio and an integrated integral value that is an integrated value of the deviation from the theoretical air-fuel ratio). It is calculated by adding to 1.0). The feedback correction value calculation unit 55 uses the oxygen concentration sensor 45 to detect the oxygen concentration in the exhaust gas when calculating the feedback correction value. Hereinafter, the process of calculating the feedback correction value is referred to as “feedback correction value calculation process”. The feedback correction value calculation process will be described later with reference to FIG.

  The feedback learning correction value calculation unit 56 calculates a feedback learning correction value. The feedback learning correction value is a value for further finely correcting the fuel injection amount in accordance with the operating state of the engine, specifically, the engine speed and the intake air pressure. The feedback learning correction value is calculated for each of a plurality of learning areas (a plurality of ranges) divided by the intake air pressure and the engine speed. The feedback learning correction value calculation unit 56 performs learning by recognizing the change in the state of fuel injection or air-fuel ratio within a predetermined time for each learning area, and based on the learning result, the feedback learning correction value for each learning area. Calculate. Hereinafter, the process of calculating the feedback learning correction value is referred to as “feedback learning process”. The feedback learning process will be described later with reference to FIG.

  The first correction value leveling processing unit 57 equalizes the variation (deviation) of the feedback learning correction value corresponding to each learning area when the ignition is ON. Hereinafter, the leveling process performed by the first correction value leveling processing unit 57 is referred to as a “first leveling process”. The first leveling process will be described later with reference to FIG.

  The second correction value leveling processing unit 58 performs the feedback learning process for any learning area, and as a result, the feedback learning correction value and the feedback learning process corresponding to the learning area for which the feedback learning process is completed are completed. When a variation (deviation) occurs between the feedback learning correction values corresponding to the non-learning areas, this variation is leveled. Hereinafter, the leveling process performed by the second correction value leveling processing unit 58 is referred to as “second leveling process”. The second leveling process will be described later with reference to FIG.

  The final fuel injection amount calculation unit 59 calculates the fuel basic injection amount calculated by the fuel basic injection amount calculation unit 54 by the feedback correction value calculated by the feedback correction value calculation unit 55 and the feedback learning correction value calculation unit 56. The final fuel injection amount is calculated by correcting the feedback learning correction value.

  The basic fuel injection amount calculation unit 54 is a specific example of the fuel injection amount calculation unit, the feedback learning correction value calculation unit 56 is a specific example of the correction value calculation unit, and the first correction value leveling processing unit 57 and The second correction value leveling processing unit 58 is a specific example of the correction value leveling means, and the final fuel injection amount calculating unit 59 is a specific example of the fuel injection amount correction means. The memory 53 is a specific example of the correction value storage means.

(Learning table)
FIG. 3 shows the learning table 61. The learning table 61 is a table in which feedback learning correction values calculated in the feedback learning process are arranged, and is stored in the memory 53 in a rewritable state.

  As shown in FIG. 3, the learning table 61 has, for example, an XY plane composed of an X axis (horizontal axis) indicating the intake air pressure and a Y axis (vertical axis) indicating the engine speed, as a predetermined value for the intake air pressure. Are divided into four learning areas R1, R2, R3, and R4 based on a predetermined threshold value Xth and a predetermined threshold value Yth for the engine speed.

  The learning area R1 is an area where the intake air pressure is not less than the minimum value Xmin and less than the threshold value Xth, and the engine speed is not less than the minimum value Ymin and less than the threshold value Yth. The learning area R2 is an area where the intake air pressure is not less than the threshold value Xth and not more than the maximum value Xmax, and the engine speed is not less than the minimum value Ymin and less than the threshold value Yth. The learning area R3 is an area where the intake air pressure is not less than the minimum value Xmin and less than the threshold value Xth, and the engine speed is not less than the threshold value Yth and not more than the maximum value Ymax. The learning area R4 is an area where the intake air pressure is not less than the threshold value Xth and not more than the maximum value Xmax, and the engine speed is not less than the threshold value Yth and not more than the maximum value Ymax.

  The feedback learning correction value is calculated for each of the learning areas R1, R2, R3, and R4, arranged in the learning table 61 for each learning area R1, R2, R3, and R4 and stored in the memory 53. Each learning area R1, R2, R3, R4 is assigned a unique number (for example, a serial number such as 1, 2, 3, 4).

(Feedback correction value calculation processing)
FIG. 4 shows a flow of feedback correction value calculation processing performed by the feedback correction value calculation unit 55. The feedback correction value calculation process is started after the temperature of the oxygen concentration sensor 45 is increased to a predetermined temperature by the exhaust gas and the detection signal from the oxygen concentration sensor 45 is stabilized, and then repeatedly performed at a cycle of, for example, 10 milliseconds. Is called.

  As shown in FIG. 4, in the feedback correction value calculation process, the feedback correction value calculation unit 55 first determines whether the current air-fuel ratio state is on the rich side based on the detection signal output from the oxygen concentration sensor 45. It is determined whether or not (step S1). When the current air-fuel ratio state is on the rich side (step S1: YES), the feedback correction value calculation unit 55 determines the feedback correction value required to shift the air-fuel ratio from the rich side to the lean side. The proportional value is calculated (step S2), the integral correction value is further calculated, and the integrated integral value is updated by adding the integral correction value to the current value of the integrated integral value (step S3). Since the proportional value of the feedback correction value required to shift the air-fuel ratio from the rich side to the lean side is a negative value and the integral correction value is an absolute value, the integration correction value is integrated into the integrated integral value. Is a process of subtracting the integral correction value from the integral integral value.

  Subsequently, the feedback correction value calculation unit 55 adds the proportional value calculated in step S2 and the integrated integral value updated in step S3 to the feedback correction set value (for example, 1.0) to obtain a final feedback correction value. Calculate (step S6).

  On the other hand, if the result of determination in step S1 is that the current air-fuel ratio is on the lean side (step S1: NO), the feedback correction value calculation unit 55 shifts the air-fuel ratio from the lean side to the rich side. The proportional value of the feedback correction value required for the calculation is calculated (step S4), the integral correction value is calculated, and the integrated integral value is updated by integrating the integral correction value with the current integrated integral value ( Step S5). Since the proportional value of the feedback correction value required to shift the air-fuel ratio from the lean side to the rich side is a positive value and the integral correction value is an absolute value, the integration correction value is integrated into the integrated integral value. Is a process of adding the integral correction value to the integral integral value.

  Subsequently, the feedback correction value calculation unit 55 calculates the final feedback correction value by adding the proportional value calculated in step S4 and the integrated integral value updated in step S5 to the feedback correction set value (step S6). .

(Feedback learning process)
FIG. 5 shows the flow of feedback learning processing performed by the feedback learning correction value calculation unit 56. The feedback learning process is started after the feedback correction value calculation process is started, and thereafter, for example, is performed at a cycle of 10 milliseconds.

  In the feedback learning process, the feedback learning correction value calculator 56 first learns the learning areas R1, R2, and R2 of the learning table 61 based on the detection signal output from the intake air pressure sensor 44 and the detection signal output from the engine rotation sensor 48. A learning area to which the current intake air pressure and the current engine speed belong (hereinafter referred to as “current learning area”) is recognized from R3 and R4 (step S11).

  Subsequently, in the feedback learning correction value calculation unit 56, the current learning area is the same as the learning area recognized in step S1 of the feedback learning process in the previous cycle (hereinafter referred to as “previous learning area”). Whether or not (step S12).

  If the current learning area and the previous learning area are not the same (step S12: NO), the current learning area number is stored in the memory 53, the timer is reset, and the timer is immediately started (step S13). In this case, the feedback learning process for the current cycle is completed.

  On the other hand, when the current learning area and the previous learning area are the same (step S12: YES), the feedback learning correction value calculation unit 56 subsequently determines whether or not the timer value has reached a predetermined value. Is determined (step S14). If the timer value has not reached the predetermined value (step S14: NO), the feedback learning process for the current cycle is immediately terminated.

  On the other hand, if the current learning area and the previous learning area are the same (step S12: YES) and the timer value has reached a predetermined value (step S14: YES), the feedback learning correction value The calculation unit 56 starts calculating the feedback learning correction value corresponding to the current learning area.

  That is, the state in which the intake air pressure is in the range of Xmin to less than Xth or Xth to Xmax and the engine speed is in the range of Ymin to less than Yth or Yth to Ymax, that is, the intake air pressure and the engine speed are When the state belonging to the same learning area continues for a predetermined time, and as a result, the same learning area is continuously recognized a predetermined number of times in step S1 of the feedback learning process repeatedly performed several times during that time, the timer The value reaches a predetermined value. In such a case, the correction of the basic fuel injection amount by the feedback correction value is under or under atmosphere, and it is necessary to calculate a feedback learning correction value for appropriately correcting this.

  When the timer value reaches a predetermined value, the feedback learning correction value calculation unit 56 determines whether or not the integrated value of the feedback correction value exceeds a predetermined determination reference value (step S15).

  If the integrated integral value of the feedback correction value exceeds the predetermined judgment reference value (step S15: YES), the basic fuel injection amount is understated because the correction of the basic fuel injection amount by the feedback correction value is too small. In order to increase the correction, the feedback learning correction value calculation unit 56 increases the feedback learning correction value corresponding to the current learning area, and updates the feedback learning correction value in the learning table 61 (step S16).

  On the other hand, if the integrated value of the feedback correction value does not exceed the predetermined judgment reference value (step S15: NO), the fuel basic injection amount correction based on the feedback correction value is over-atmosphere. In order to suppress the correction of the injection amount, the feedback learning correction value calculation unit 56 decreases the feedback learning correction value corresponding to the current learning area, and updates the feedback learning correction value in the learning table 61 (step S17).

  In this way, in the feedback learning process, since the feedback learning correction value is not updated unless the state where the intake air pressure and the engine speed belong to the same learning area continues for a predetermined time, one feedback learning process is not performed. It takes some time for the feedback learning correction values corresponding to the learning areas to be updated, and furthermore, the feedback learning correction values corresponding to all the learning areas R1, R2, R3, R4 belonging to the learning table 61 are updated. Takes even longer.

(First leveling process)
FIG. 6 shows the flow of the first leveling process performed by the first correction value leveling processing unit 57. The first leveling process is performed when the ignition is ON. The first leveling process is mainly caused by the occurrence of a learning area in which the feedback learning correction value is frequently updated and a learning area in which the feedback learning correction value is not updated among the plurality of learning areas R1, R2, R3, and R4. This is performed for the purpose of leveling the variation (deviation) of the learning correction value and the variation (deviation) of the feedback learning correction value caused by an unexpected error in the feedback learning process.

  As shown in FIG. 6, in the first leveling process, the first correction value leveling unit 57 recognizes that the ignition is turned on (step S21).

  Subsequently, the first correction value leveling processing unit 57 selects the maximum value Lmax and the minimum value Lmin of the feedback learning correction values from the feedback learning correction values corresponding to all the learning areas R1, R2, R3, and R4. Is specified (step S22).

  Subsequently, the first correction value leveling processing unit 57 determines whether or not the difference between the maximum value Lmax and the minimum value Lmin of the feedback learning correction value is greater than or equal to a predetermined reference value Lr (step S23). When there is a large variation in the feedback learning correction value between the learning areas R1, R2, R3, and R4, the difference between the maximum value Lmax and the minimum value Lmin of the feedback learning correction value is equal to or greater than the reference value Lr.

  When the difference between the maximum value Lmax and the minimum value Lmin of the feedback learning correction value is not greater than or equal to the reference value Lr (step S23: NO), the first correction value leveling processing unit 57 performs the first leveling process of this time. End processing immediately.

  On the other hand, in the learning area Rp, when the difference between the maximum value Lmax and the minimum value Lmin of the feedback learning correction value is greater than or equal to the reference value Lr (step S23: YES), the first correction value leveling processing unit 57 The count value P is initialized to 0 (step S24). Subsequently, the first correction value leveling processing unit 57 increments the count value P by 1 (step S25), and then stores the feedback learning correction value Lp corresponding to the learning area Rp in the learning table stored in the memory 53. Read from 61 (step S26). When the count value P is 1, the learning area R1 is selected as the learning area Rp, when the count value P is 2, the learning area R2 is selected as the learning area Rp, and when the count value P is 3, the learning area Rp is learned. When the area R3 is selected and the count value P is 4, the learning area R4 is selected as the learning area Rp.

  Subsequently, the first correction value leveling processing unit 57 calculates an average value of the maximum value Lmax and the minimum value Lmin of the feedback learning correction value. Then, the first correction value leveling processing unit 57 determines whether or not the difference between the feedback learning correction value Lp corresponding to the learning area Rp and the average value is half or more of the reference value Lr. (Step S27). That is, in step S27, the first correction value leveling processing unit 57 determines whether or not the following mathematical formula (F1) is established.

| Lp − {(Lmax + Lmin) / 2} | ≧ Lr / 2 (F1)
When the feedback learning correction value Lp corresponding to the learning area Rp is too large or too small, Formula (F1) is established.

  If the mathematical formula (F1) is not satisfied (step S27: NO), the first correction value leveling processing unit 57 immediately shifts the process to step S31.

  On the other hand, when Formula (F1) is established (step S27: YES), the first correction value leveling processing unit 57 uses the feedback learning correction value Lp corresponding to the learning area Rp as feedback learning in the learning area Rp. The correction value is increased or decreased so as to approach the average value of the maximum value Lmax and the minimum value Lmin.

  Specifically, when the feedback learning correction value Lp corresponding to the learning area Rp is larger than the average value of the maximum value Lmax and the minimum value Lmin of the feedback learning correction value (step S28: YES), the first The correction value leveling processing unit 57 performs the calculation shown in the following mathematical formula (F2), and uses the current value of the feedback learning correction value Lp corresponding to the learning area Rp in the learning table 61 stored in the memory 53 as the mathematical formula. The feedback learning correction value Lp corresponding to the learning area Rp is updated by replacing with the value obtained by the calculation shown in (F2) (step S29).

{(Lmax + Lmin) / 2} + (Lr / 2) -α (F2)
Note that α in the formula (F2) is a predetermined set value.

  As a result, the feedback learning correction value Lp corresponding to the learning area Rp decreases and approaches the average value of the maximum value Lmax and the minimum value Lmin of the feedback learning correction value. As a result, the feedback learning correction value Lp is not excessive.

  On the other hand, when the feedback learning correction value Lp corresponding to the learning area Rp is equal to or less than the average value of the maximum value Lmax and the minimum value Lmin of the feedback learning correction value (step S28: NO), the first correction value leveling is performed. The conversion processing unit 57 performs the calculation shown in the following mathematical formula (F3), and the current value of the feedback learning correction value Lp corresponding to the learning area Rp in the learning table 61 stored in the memory 53 is expressed in the mathematical formula (F3). The feedback learning correction value Lp corresponding to the learning area Rp is updated by replacing with the value obtained by the calculation shown (step S30).

{(Lmax + Lmin) / 2}-(Lr / 2) + α (F3)
As a result, the feedback learning correction value Lp corresponding to the learning area Rp increases and approaches the average value of the maximum value Lmax and the minimum value Lmin of the feedback learning correction value. As a result, the feedback learning correction value Lp is not too small.

  Subsequently, the first correction value leveling processing unit 57 determines whether or not the count value P is equal to or greater than Pn (step S31). Pn is a numerical value indicating the total number of learning areas, that is, 4 in this embodiment. If the count value P is not greater than or equal to Pn (step S31: NO), the first correction value leveling processing unit 57 returns the process to step S25. On the other hand, when the count value P is greater than or equal to Pn (step S31: YES), the first correction value leveling processing unit 57 finishes the first leveling process.

  According to the first leveling process described above, each time the ignition is turned on, the feedback learning correction value corresponding to each of the learning areas R1, R2, R3, and R4 is excessive or too small. Is present, the excessive or small feedback learning correction value approaches the average value of the maximum value and the minimum value among the feedback learning correction values corresponding to the learning areas R1, R2, R3, and R4, respectively. Replaced with a value. As a result, large variations in the feedback learning correction values respectively corresponding to the learning areas R1, R2, R3, and R4 are removed.

(Second leveling process)
FIG. 7 shows the flow of the second leveling process performed by the second correction value leveling processing unit 58. Instead of the first leveling process described above, the second leveling process can be applied. The second leveling process is performed, for example, after the ignition is turned on. In the second leveling process, the property of the fuel is changed mainly because the fuel is replaced (for example, normal regular gasoline is replaced with gasohol) or the fuel is replenished. A feedback learning process is performed for any learning area. As a result, a feedback learning correction value corresponding to the learning area for which the feedback learning process has been completed, and a feedback learning correction value for a learning area for which the feedback learning process has not been completed, This is performed for the purpose of leveling the variation when a variation (deviation) occurs. In the second leveling process, the learning area where the feedback learning process is completed is the learning area R1, and the learning area where the feedback learning process is completed is a learning area other than the learning area R1, that is, the learning area R2, A leveling process with different processing contents is performed when the learning area is either R3 or R4.

  First, the second leveling process when the learning area where the feedback learning process is completed is the learning area R1 will be described. As shown in FIG. 7, the second correction value leveling processing unit 58 calculates a difference D1 between the previous value and the current value of the feedback learning correction value L1 corresponding to the learning area R1 (step S71). The current value of the feedback learning correction value corresponding to each learning area R 1, R 2, R 3, R 4, that is, the latest value is arranged in the learning table 61 and stored in the memory 53. In addition to this, the feedback learning in the previous cycle of the feedback learning process in which the previous value of the feedback learning correction value corresponding to each learning area R1, R2, R3, R4, that is, the latest value of the feedback learning correction value is calculated. The feedback learning correction value calculated by the processing is also stored in the memory 53.

  Subsequently, the second correction value leveling processing unit 58 determines whether or not the difference D1 between the previous value and the current value of the feedback learning correction value L1 corresponding to the learning area R1 is greater than or equal to the reference value Ds (step). S72). When the feedback learning process is completed for the learning area R1 and the feedback learning correction value is updated after the fuel property is changed due to fuel replacement or the like, the difference D1 is equal to or greater than the reference value Ds. On the other hand, if the fuel property has not changed, or if the fuel property has changed but the feedback learning process has not yet been completed for the learning area R1, the difference D1 does not exceed the reference value Ds. That is, the second correction value leveling processing unit 58 recognizes whether or not the feedback learning process is completed for the learning area R1 after the change in the fuel property, based on the determination in step S72.

  When the difference D1 is greater than or equal to the reference value Ds (step S72: YES), the second correction value leveling processing unit 58 temporarily stores the difference D1 in the memory 53 (step S73). Subsequently, after setting the count value Q to 1, the second correction value leveling processing unit 58 immediately increases the count value Q by 1 (steps S74 and S75), and feedback learning corresponding to the learning area R1. Based on the correction values, the feedback learning correction values corresponding to the learning areas R2, R3, and R4 are leveled.

  That is, the second correction value leveling processing unit 58 first calculates the difference Dq between the previous value and the current value of the feedback learning correction value Lq corresponding to the learning area Rq (step S76). When the count value Q is 2, the learning area Rq and the feedback learning correction value Lq mean the learning area R2 and the feedback learning correction value L2, respectively. When the count value Q is 3, the learning area Rq and the feedback learning correction value Lq are The learning area R3 and the feedback learning correction value L3 are meant respectively. When the count value Q is 4, the learning area Rq and the feedback learning correction value Lq mean the learning area R4 and the feedback learning correction value L4, respectively.

  Subsequently, the second correction value leveling processing unit 58 determines whether or not the difference Dq is greater than or equal to the reference value Dk (step S77). When the feedback learning process is completed for the learning area Rq after the fuel property is changed and the feedback learning correction value corresponding to the learning area Rq is updated, the difference Dq is equal to or greater than the reference value Dk. The reference value Dk is, for example, the same value as the reference value Ds. If the difference Dq is greater than or equal to the reference value Dk (step S77: YES), the process returns to step S75.

  On the other hand, when the difference Dq is not equal to or greater than the reference value Dk (step S77: NO), the second correction value leveling processing unit 58 sets a value obtained by multiplying the difference D1 by a predetermined set value β in the learning area R1. The leveling value Lav of the feedback learning correction value is calculated by adding the corresponding value to the previous value of the feedback learning correction value L1. Then, the second correction value leveling processing unit 58 replaces the feedback learning correction value Lq corresponding to the learning area Rq arranged in the learning table and stored in the memory 53 with the leveling value Lav, and performs the feedback learning. The correction value Lq is updated (step S78).

  Subsequently, the second correction value leveling processing unit 58 determines whether or not the count value Q is equal to or greater than Qn (step S79). Qn is a numerical value indicating the total number of learning areas, that is, 4 in this embodiment. If the count value Q is not equal to or greater than Qn (step S79: NO), the second correction value leveling processing unit 58 returns the process to step S75. On the other hand, when the count value Q is equal to or greater than Qn (step S79: YES), the second correction value leveling processing unit 58 ends the second leveling process.

  By the processing in steps S71 to S79, the degree of variation between the feedback learning correction value L1 corresponding to the learning area R1 and the feedback learning correction values L2, L3, and L4 corresponding to the learning areas R2, R3, and R4 is reduced. .

  Next, the second leveling process when the learning area where the feedback learning process is completed is a learning area other than the learning area R1 will be described.

  In steps S71 and S72 in FIG. 7, when the difference D1 between the previous value and the current value of the feedback learning correction value L1 corresponding to the learning area R1 is not equal to or greater than the reference value Ds (step S72: NO), the second The correction value leveling process 58 searches for the learning area in which the feedback learning correction value is updated after the feedback learning process is completed after the fuel property is changed, among the learning areas R2, R3, and R4.

  That is, the second correction value leveling processing unit 58 immediately increases the count value X by 1 after setting the count value X to 1, and the previous value and current value of the feedback learning correction value Lx corresponding to the learning area Rx. A difference Dx is calculated, and it is determined whether or not the difference Dx is greater than or equal to the reference value Ds (steps S80 to S83). When the count value X is 2, the learning area Rx and the feedback learning correction value Lx mean the learning area R2 and the feedback learning correction value L2, respectively. When the count value X is 3, the learning area Rx and the feedback learning correction value Lx are The learning area R3 and the feedback learning correction value L3 are meant respectively. When the count value X is 4, the learning area Rx and the feedback learning correction value Lx mean the learning area R4 and the feedback learning correction value L4, respectively.

  If the difference Dx is not greater than or equal to the reference value Ds and the count value X is not greater than or equal to Xn (step S83: NO, step S84: NO), the second correction value leveling processing unit 58 returns the process to step S81. . Xn is a numerical value indicating the total number of learning areas, that is, 4 in this embodiment.

  On the other hand, when the difference Dx is greater than or equal to the reference value Ds (step S83: YES), the second correction value leveling processing unit 58 temporarily stores the feedback learning correction value Lx in the memory 53 (step S85). Subsequently, after setting the count value Q to 1, the second correction value leveling processing unit 58 immediately increases the count value Q by 1 (steps S86 and S87), and the feedback temporarily stored in the memory 53 Based on the learning correction value Lx, feedback learning correction values corresponding to learning areas other than the learning area Rx among the learning areas R2, R3, and R4 are leveled.

  That is, the second correction value leveling processing unit 58 first calculates the difference Dq between the previous value and the current value of the feedback learning correction value Lq corresponding to the learning area Rq, and the difference Dq is greater than or equal to the reference value Dk. Is determined (steps S88, S89). If the difference Dq is greater than or equal to the reference value Dk (step S89: YES), the second correction value leveling processing unit 58 returns the process to step S87.

  On the other hand, when the difference Dq is not greater than or equal to the reference value Dk (step S89: NO), the second correction value leveling processing unit 58 corresponds to the learning area Rq arranged in the learning table and stored in the memory 53. The feedback learning correction value Lq is replaced with the feedback learning correction value Lx temporarily stored in the memory 53 in step S85, and the feedback learning correction value Lq is updated (step S90).

  Subsequently, the second correction value leveling processing unit 58 determines whether or not the count value Q is equal to or greater than Qn (step S91). If the count value Q is not equal to or greater than Qn (step S91: NO), the second correction value leveling processing unit 58 returns the process to step S87. On the other hand, when the count value Q is equal to or greater than Qn (step S91: YES), the second correction value leveling processing unit 58 ends the second leveling process.

  By the processes in steps S80 to S91, large variations in the feedback learning correction values corresponding to the learning areas R2, R3, and R4 are removed.

  On the other hand, if the fuel property has not changed, or if the feedback learning process has not been completed in any of the learning areas R1, R2, R3, and R4 after the fuel property has changed, the count value X is determined to be in step S84. It is determined that it is greater than or equal to Xn, and the second leveling process ends. In this case, the feedback learning correction values corresponding to the learning areas R1, R2, R3, and R4 are considered not to vary greatly. Therefore, the feedback learning correction values are leveled in any of the learning areas R1, R2, R3, and R4. There is no conversion.

  As described above, according to the first leveling process in the fuel injection control device 51 according to the embodiment of the present invention, a large variation in the feedback learning correction value among the learning areas R1, R2, R3, and R4 is removed when the ignition is turned on. can do. In addition, according to the second leveling process, a large variation in the feedback learning correction value generated as a result of the feedback learning process being performed for any one of the learning areas in accordance with a change in fuel properties due to fuel replacement or refueling is removed. can do. By removing the large variation in the feedback learning correction value in this way, for example, the change in the intake air pressure or the engine speed is repeated with the operation of the motorcycle, and the feedback learning correction value corresponding to each of the plurality of learning areas is replaced. Even when the fuel injection amount is used to calculate the alternative fuel injection amount, it is possible to prevent the fuel injection amount from fluctuating excessively, and to maintain or improve the stability and smoothness of engine operation.

  Further, according to the second leveling process, the feedback learning correction value corresponding to the learning area where the feedback learning process is not completed approaches the feedback learning correction value corresponding to the learning area where the feedback learning process is completed. After the fuel properties have changed due to fuel replacement or refueling, the feedback learning process can be completed in a short time for each learning area, and appropriate fuel injection according to the changed fuel properties can be realized at an early stage. .

  In the above-described embodiment, the case where the first leveling process or the second leveling process is performed when the ignition is turned on is described as an example. However, the present invention is not limited to this. The first leveling process or the second leveling process may be performed at another timing.

  In the above-described embodiment, the learning area where the feedback learning process is completed is the learning area R1, and the learning area where the feedback learning process is completed is a learning area other than the learning area R1. Although the case where the leveling process of 2 is performed is taken as an example, the second leveling process may be performed only when the learning area where the feedback learning process is completed is the learning area R1. That is, if the difference D1 is not greater than or equal to the reference value Ds in step S72 in FIG. 7, the process returns to step S71, whereby the difference D1 between the previous value and the current value of the feedback learning correction value L1 corresponding to the learning area R1. Waits for the reference value Ds to be exceeded. When the difference D1 between the previous value and the current value of the feedback learning correction value L1 corresponding to the learning area R1 exceeds the reference value Ds, the processes of steps S74 to S79 are performed.

  The fuel injection control device of the present invention can be applied not only to an engine using liquid fuel such as ordinary gasoline and gasohol but also to an engine using gaseous fuel such as compressed natural gas. Moreover, the use of the engine to which the fuel injection control device of the present invention is applied is not limited to a vehicle such as a motorcycle or an automobile.

  Further, the present invention can be appropriately changed without departing from the gist or concept of the invention that can be read from the claims and the entire specification, and a fuel injection control apparatus that includes such a change is also a technique of the present invention. Included in thought.

1 Engine 6 Combustion Chamber 37 Fuel Injection Device 44 Suction Air Pressure Sensor 48 Engine Rotation Sensor 51 Fuel Injection Control Device 52 CPU
53 Memory (correction value storage means)
54 Fuel basic injection amount calculation unit (fuel injection amount calculation means)
55 Feedback Correction Value Calculation Unit 56 Feedback Learning Correction Value Calculation Unit (Correction Value Calculation Means)
57 First correction value leveling processing unit (correction value leveling means)
58 Second correction value leveling processing unit (correction value leveling means)
59 Final fuel injection amount calculation section (fuel injection amount correction means)
61 Learning table

Claims (5)

A fuel injection control device that controls a fuel injection amount of a fuel injection device that supplies fuel to a combustion chamber of an engine,
Fuel injection amount calculating means for calculating the fuel injection amount based on intake air, fuel and engine state;
Correction value storage means for storing a plurality of correction values respectively corresponding to a plurality of ranges specified by the intake air pressure and the engine speed;
Correction value calculating means for performing learning by recognizing the transition of the state of fuel injection or air-fuel ratio within a predetermined time for each of the plurality of ranges, and calculating the plurality of correction values based on the learning result;
Correction value leveling means for performing leveling processing to reduce the deviation of each correction value in the plurality of correction values;
And a fuel injection amount correction unit that corrects the fuel injection amount calculated by the fuel injection amount calculation unit using at least one correction value of the plurality of correction values. Fuel injection control device.   The correction value leveling means calculates an average value of a maximum value and a minimum value among the plurality of correction values in the leveling process, and a difference between the average value among the plurality of correction values is predetermined. 2. The fuel injection control device according to claim 1, wherein a correction value that is equal to or greater than a value is specified, and the specified correction value is increased or decreased so as to approach the average value.   The fuel injection control apparatus according to claim 1 or 2, wherein the correction value leveling means performs the leveling process at the start of engine operation.   When the learning is completed for any one of the plurality of ranges, the correction value leveling means corrects the correction corresponding to the range in which the learning has been completed among the plurality of ranges in the leveling process. The fuel injection control device according to claim 1, wherein a correction value corresponding to a range in which the learning is not completed is increased or decreased so as to approach a value. The plurality of ranges are at least a first range in which both the intake air pressure and the engine speed are relatively small, a second range in which the intake air pressure is relatively large and the engine speed is relatively small, and an intake air pressure is relatively small. A third range in which the rotational speed is relatively large, and a fourth range in which both the intake air pressure and the engine rotational speed are relatively large,
The correction value leveling means, when the learning for the first range is completed, in the leveling process, the correction value leveling means approaches the correction value corresponding to the first range. 2. The fuel injection control device according to claim 1, wherein a correction value corresponding to a range in which the learning is not completed in the fourth range is increased or decreased.

Images