The present invention relates to a fuel injection method for internal combustion engines according to the preamble of claim 1. Such a control method is known from the EP-A-106 366.

The Japanese Laid Open Publication No. 5841/1984 discloses a conventional acceleration adjustment in fuel injection devices of the internal combustion engine for automobiles wherein an incremental valve opening pulse for a fuel increment is generated independently of the basic valve opening pulse for causing the injectors to inject fuel normally during the normal condition. The pulse width of the incremental valve opening pulse is determined in dependence on the water temperature of an internal combustion engine.

However, this known acceleration adjustment in fuel injection devices has not provided fine adjustment of incremental valve opening pulse depending on the state of a running engine excepting the water temperature.

The above electronic fuel injector (MPI system) employs a vane air flow meter. Therefore, it can not abruptly increase the rate of air intake into an engine during acceleration. From this reason, a relatively rich mixture is supplied at the beginning of acceleration.

However, because an inherent delay exists in air intake into the engine, it has been impossible to achieve abrupt acceleration. On the rapid acceleration from idling to the fully open state under no load, the speeded up revolution time (To) is about 200 ms (confer Fig. 4).

Besides, a hot wire air flow meter employed in the EP-A-106 366 requires fine adjustment during acceleration, because it includes no member interrupting air intake as seen in the vane type airflow meter; hence, it causes an instantaneous flow of air into the engine.

The hot wire air flow meter which detects the volume of air is advantageous in that it allows air to be supplied to the engine more rapidly during acceleration because it has no throttle component. It does, however, have the problem of delayed fuel feed, that is, a lag in response time. This delay in fuel feed response time requires fine adjustment based on acceleration and other parameters.

The EP-A-106 366 describes a fuel injection control system for internal combustion engines wherein a basic fuel injector valve opening pulse is generated taking into account engine parameters. A hot-wire air- flow meter is present. An additional fuel amount is calculated in dependence on a magnitude of acceleration of the engine. The throttle-opening is fetched by interruption at 10 ms intervals. The difference of the throttle position is used to define engine acceleration and its magnitude. The additional fuel amount is injected at intervals of 10 ms after the basic injection pulse so that the additional injection pulses are generated in synchronism with the basic injection pulse.

According to Fig. 19 and claim 2 of the EP-A-106 366 additional acceleration fuel increments are (alternatively) calculated taking into account the position of the idle-switch during preacceleration condition.

According to the US-A-4 363 307 (especially Fig. 5, e.g. step 254) synchronous acceleration fuel injection pulses depending on the normal injection pulses are determined on the basis of the normalized variation of the intake-air flow during a predetermined time on the basis of a formula

The object of the present invention is to provide a fuel injection control method for internal combustion engines wherein the pulse duration of an incremental fuel injection pulse can be determined in dependence on the state of a running engine excepting the water temperature and delayed fuel feed can be adjusted, acceleration matching between the state after deceleration and the state after normal operation becomes possible, speed up revolution time of acceleration from idling to the fully open state under no load can be shortened, and inadequate acceleration during rapid acceleration which is to be effected after fuel reduction for deceleration can be improved.

The above object is solved by the features of claim 1.

The claims 2 to 5 characterize advantageous developments thereof.

By applying the judgement of conditions for adjustment after fuel reduction for deceleration, it has become possible to match acceleration between the state after deceleration and the state after normal operation.

• Fig. 1 is a schematic diagram of an engine system embodying the present invention.
• Fig. 2 is a block diagram of a relationship between input and output in the system of Fig. 1.
• Fig. 3 is a graphical illustration of engine injection timing.
• Fig. 4 is an explanatory graph of experimental results of respective speed up revolution times according to both the present invention and the prior art.
• Fig. 5 is a flow chart for depicting the acceleration adjusting method of the present invention.

One embodiment of the present invention will be described (in the description 6. Qa is used as an abbreviation of the normalized value ΔQa/Qa). Fig. 1 shows an engine system to which the present invention is applied. An engine 1 has intake pipes 2 which are provided with the fuel injection valves (injectors) 3 corresponding to the number of cylinders. These intake pipes 2 are brought into a single pipe at a collector 4 in the upstream side, and having a throttle valve 5 for determining the amount of intake for the engine further upstream.

An amount of intake for the engine 1 is measured by a hot wire air flow meter 6 provided still further upstream with engine rotational speed or revolutions being counted by a rotational speed sensor 13. In addition to the above, a control unit CU 12 also receives a signal 11 from engine temperature detector, an exhaust gas signal from an oxygen concentration measuring sensor 10, and a signal from an idle switch 14, etc.

Fuel is supplied to the engine 1 by opening a valve on each fuel injector 3, and the amount of fuel is measured based on valve opening time. Fuel is pressurized and regulated by a fuel pump 8 and a regulator 9.

Fig. 2 is a block diagram showing the relationship between input and output depicted in Fig. 1. With respect to the control unit CU, the left side includes sensors and the right side includes actuators.

In the control unit 12, as shown in Fig. 2 wave shaping circuits and AD- converters are arranged on the left side, an I/O LSI section for implementing therein exchange of input/output and arithmetic processing and a CPU for giving instructions to the I/O LSI section are at the center, and a circuit for driving the output actuators on the right side.

Next, injection timing, for example in a 4-cylinder, 4-cycle engine, will be measured with reference to Fig. 3. In the 4-cylinder, 4-cycle engine, fuel is usually injected one per revolution, i.e., at the timing of A and C, for all four cylinders.

The sections A and C are respectively the basic valve opening pulses for causing the injector to inject fuel normally during the normal condition.

The incremental fuel injection pulse described in the present invention is an additional fuel injection pulse shown in the section of B which is generated immediately upon detection of acceleration, so as to inject fuel at that time.

In the present invention, acceleration may be detected based on, for example, an ON ­ OFF signal of the idle switch 14 or a closing signal of the idle switch 14 indicating that a throttle valve 5 has been opened from an idle state, and the rate of variation ΔQa of the air intake volume Qa per unit time. The magnitude of acceleration can be detected based on the normalized value ΔQa/Oa of the rate of variation of the air intake volume Qa during a predetermined time (At).

Table 1 shows an example of setting the incremental fuel injection pulse for a running vehicle in response to the magnitude of acceleration.

By way of example, the process after acceleration is divided depending on the condition before acceleration. More specifically, after reducing fuel feed during deceleration, a large amount of fuel must be supplied to compensate for drying-up in the intake system during fuel reduction. Also after the normal condition, adjusting acceleration requires changes to correspond to the fully closed state of the throttle valve and other states thereof.

Further, acceleration is detected as a combination of the rate of variation ΔQa in the air intake volume per unit time and the idle switch setting. In Table 1, the legend "after idling" indicates an incremental fuel injection pulse based on detection of acceleration using the idle switch.

The reason for detecting acceleration based on both the idle switch and the rate of variation ΔQa in the air intake volume per unit time is in that detection with the idle switch always occurs earlier than detection with the rate of variation ΔQa in the air intake volume per unit time. Experiments have shown that this is attributable to a delay in the intake system accompanying detection of the rate of variation OQa in the air intake volume per unit time. Though this delay is relatively short, it is still longer than that accompanying the opening of the throttle valve.

Description will now be made by referring to typical examples of the respective cases. The incremental fuel injection pulse based on the rate of variation ΔQa in the air intake volume per unit time after fuel cutdown for deceleration will be described.

First, after fuel reduction for deceleration, the process is divided into slow acceleration (under light load. 5%<AQa<37.5%) and rapid acceleration (ΔQa≧37.5%). For slow acceleration, an incremental pulse width of 3 ms is generated; for rapid acceleration, an incremental pulse width of 9 ms is generated, (one for each detected acceleration rate).

The pulse widths of the above incremental fuel injection pulses are set preferably to have about 1.5-5 ms duration for slow acceleration and to have about 6-12 ms duration for rapid acceleration, respectively.

An incremental fuel injection pulse based on the idle switch is set to have, by way of example, a relatively long duration of 9 ms. The width of the above incremental fuel injection pulse after idling is set preferably to have about 6-12 ms duration.

An incremental fuel injection pulse based on the rate of variation ΔQa in the air take volume per unit time in normal condition is set to have 3 ms duration for slow acceleration (5%<AQa<10%) and is set to have a 6 ms duration for rapid acceleration (ΔQa≧10%) respectively, and the fuel is injected. The widths of these incremental fuel injection pulses are set preferably to have about 1.5-5 ms duration for slow acceleration and to have above 4-8 ms duration for rapid acceleration, respectively.

After the normal condition the idle switch is turned on, because the intake pipe does not dry as much as before, an incremental fuel injection pulse based on the idle switch may be set to a shorter duration, e.g., 3 ms, than that during deceleration. The width of this incremental fuel injection pulse is set preferably to have 1.5-5 ms duration.

Moreover, in acceleration from the state where the idle switch is turned off, i.e., the throttle valve is in an open state, better conditions exist in comparison with acceleration from a closed state of the throttle valve, in that the mixture in the intake pipe is more uniform, whereby acceleration adjustment based on the idle switch is not required.

Fig. 4 shows the experimental results obtained from application of the above-mentioned adjustment. The graph of Fig. 4 represents an example of rapid acceleration from idling to the fully open state under no load. It is evident from Fig. 4 that, as a result of adjusting the duration of the incremental fuel injection pulses during acceleration, revolution time is speeded up from To of about 200 ms in the prior art to T of about 120 ms in the present invention.

Fig. 5 shows a flow chart sheet for realizing the adjustment described in connection with Table 1. First, in acceleration after the normal condition, the process flow passes the right-hand loop. When the rate of variation of the airtake volume per unit time (ΔQa) is smaller than A1, no incremental fuel injection pulse is generated; when ΔQa is greater than A1, an incremental fuel injection pulse TA1 is issued. Note that, although inherently a decision must be made based on the inequality of Ai > ΔQa > Ai' so that the incremental fuel injection pulse may be divided into pulses TA1' (of 6 ms in Table 1)-and TA1" (of 3 ms in Table 1), this process has been omitted from the flow chart in all the following associated flow.

Next, in the central loop representing acceleration from idling, an incremental fuel injection pulse TA2 (3 ms) due to the condition after idling is set and thereafter an incremental fuel injection pulse TA3, also herein inherently divided into two types of pulses TA3', 3 ms in Table 1, and TA3", 6 ms in Table 1 - is set based on ΔQa.

The left-hand loop represents acceleration from the decelerated state. In this case, an incremental fuel injection pulse TA4 (9 ms) is issued due to the state after idling and acceleration is then determined on the basis of ΔQa. If A5 > AQa > A6, an incremental fuel injection pulse TA6 (3 ms) is issued and adjustment is completed therewith.

In all these adjustments the incremental fuel injection pulse based on ΔQa is generated only once for each acceleration; therefore, tests on a running engine have revealed the problem of inadequate acceleration during rapid acceleration which is to be effected after fuel reduction for deceleration.

The present invention is to overcome this problem. Namely, when ΔQa > A5, an incremental fuel injection pulse TA5 is set. The fact that the pulse TA5 has been issued is stored in memory, thus making it possible to repeat the adjustment with the incremental fuel injection pulse TA5 based on ΔQa > A5 only on the above condition.