DE68908891T2 - FUEL INJECTION SYSTEM FOR COMBUSTION ENGINE. - Google Patents

Description

The invention relates to a fuel injection system for an internal combustion engine of the type described in the characterizing part of claim 1 (cf. WO-A-88 02811).

In an internal combustion engine with a fuel injection system, sufficient fuel must be injected to achieve the stoichiometric air-fuel ratio for the engine to run satisfactorily. This is the state of the art under steady-state conditions. As the load on the engine increases, however, more air is sucked into each cylinder and therefore more fuel must be injected. If, on the other hand, the load on the engine decreases, the amount of fuel injected must be reduced. The question of mixture adjustment under such dynamic conditions has not yet been clarified.

In known systems, a sensor is usually provided in the engine's air intake system, which measures the engine load and supplies an electrical signal with which the amount of fuel to be injected can be calculated. Examples of sensors are the pressure sensor, the hot-film or hot-wire air flow meter and the air flow meter. The signals obtained with such sensors reflect the load very precisely at constant load and even at slowly changing loads. However, with rapid load changes, the signals obtained are inaccurate and result in a mismatch of the injection quantity. The signals obtained from a pressure sensor, for example, follow the actual load change only slowly with a certain delay, so that the signals derived from them for the injection quantity result in a mixture that is too lean. The signals from the other sensor types mentioned "overshoot" the value corresponding to the actual load considerably in the event of rapid load changes, because they are measurements for the air sucked into the intake manifold, which must first be filled with air so that the actual cylinder charge can increase. In certain cases, the initial "overshoot" of the signal is followed by an "undershoot", and this applies in particular to air flow meters with a baffle plate. As a result, the mixture can be either too rich or too lean. Another reason for mismatching of the mixture when the load changes is the variation in the fuel wall film caused by the change, which must be compensated for using special algorithms.

In the documents US-A-4 463 732 and US-A-4 753 210, the calculation of the basic injection quantity using an intake manifold pressure sensor is described. If the throttle valve signal changes beyond a specified value, an asynchronous injection quantity is calculated in accordance with the change in the throttle valve signal. According to the information in the document WO-A-88/02811, the values used to calculate the basic (synchronous) injection quantity are filtered with a different filter characteristic if the throttle valve signal changes by more than a specified value. The summary of the document JP-A-58144635 teaches a reduction in the intake manifold pressure-dependent basic fuel quantity if the change in the throttle valve signal is less than zero.

The abstract of the publication JP-A-6166825 teaches a higher sampling frequency of values determined to be erroneous during accelerations.

One purpose of the invention is to provide a fuel injection system for an internal combustion engine with which rapid load changes can be tracked more accurately and which also makes it possible to maintain the desired mixture.

The said purpose is achieved by means of the features specified in claim 1.

Further advantages are obtained by using the features according to claims 2 - 4.

A specific embodiment of the invention is described below by way of example, with reference to the accompanying drawings. They show:

Fig. 1 is a diagram of a 4-cylinder gasoline engine with gasoline injection with inventive control of the gasoline injection;

Fig. 2 a schematic representation of the so-called wall film effect;

Fig. 3 is a diagram showing the relationship between the throttle valve opening angle αDK and the corresponding basic injection pulse width t1 for different engine speeds;

Fig. 4a to 4d show the relationship between the throttle valve position (Fig. 4a), the calculated and actual fuel injection pulses (Fig. 4b), the engine cylinder working strokes (Fig. 4c) and the intake manifold pressure (Fig. 4d) during a change in engine load;

Fig. 5 is a block diagram showing the operation of a specific embodiment of the invention; and

Fig. 6 shows a timing diagram for the load signals obtained using different processing methods.

First, Fig. 1 shows a 4-cylinder engine 10 with fuel injection, which has an intake manifold (suction pipe) 12 and an exhaust manifold 14. Air is sucked into the intake manifold 12 due to the negative pressure in the engine cylinders, and a measured amount of fuel is injected into the air via one or more injection valves 16. The air flow to the engine is determined by a throttle valve (usually designed as a flap) that is connected to the accelerator pedal. If the engine sucks in a larger amount of air, the amount of fuel injected must be adjusted accordingly. An electronic control unit 20 is used to calculate the injection amount.

The injection quantity must be precisely matched to the intake air quantity in order to achieve the stoichiometric air-fuel ratio. If this is achieved, cleaner and more efficient combustion of the fuel can be expected.

The amount of air drawn in by the engine is usually measured by a main load sensor in the air intake system. In the embodiment shown with a pressure sensor 22 and a hot wire measuring device 24, the latter units can be connected alternately to the electronic control unit, which is represented by a switch 26. It should be noted that usually only one main load sensor will be used, whereas in the embodiment described two main load sensors are provided to make it clear that a main load sensor of any type can be used. Strictly speaking, not only Not only the main load sensors specified above can be used, but also, for example, a baffle plate or hot-film air flow meter.

An electrical signal corresponding to the angular position α of the throttle valve is also fed to the electronic control unit 20. A line from the motor 10 supplies a signal for the speed n of the motor, as well as reference signals corresponding to various angular positions in the 720º cycle of the motor.

The upper part of Fig. 5 shows how the normal, angle-synchronous fuel injection pulses are calculated in the control unit 20. In step 27, the pressure signal from the pressure main load sensor 22 is sampled by the electronic control unit 20 at each reference mark tR of the engine, as well as in a position between two adjacent reference marks tR. The pressure is thus assumed to be the average of the two values mentioned, as follows: segment

It should be noted that other pressure acquisition methods, for example, fast IMS sampling of the pressure signal over one or more segments of 720º/number of cylinders, are also possible.

Then, in stage 28, an angle-synchronous basic injection pulse width t1p according to Fig. 4d is obtained by mapping the pressure value thus obtained with the signal corresponding to the engine speed n. (When using a main load sensor of a different type, a corresponding angle-synchronous basic injection pulse width t1M or t1Q is obtained). The basic injection pulse width obtained from the main load sensor in the manner described is from now on referred to as t₁(k). The basic injection duration t₁(k) is then filtered with a filter 32 (connectable via a switch 30). In this way, any "trembling" of the signal is eliminated and a filtered, angle-synchronous basic injection pulse width t1f(k) is obtained. The filtered signal goes via an enable gate 34 controlled by the enable function 36 to a load change compensation stage 38. The load change compensation stage 38 changes the basic signal multiplicatively or additively to compensate for various engine parameters, including the wall film model shown in Fig. 2.

The enable function 36 enables load change compensation at step 38 only if it has been determined that the load change of a specific cylinder calculated for two strokes is above a certain threshold. For example, the pressure difference for a specific cylinder calculated for consecutive strokes can be used. Thus:

Release function effective for P(K)-P(K-Z) > threshold, where Z = number of engine cylinders.

The release function makes it possible to ensure that load change compensation is only carried out when necessary. It ensures that no load change compensation is carried out if cylinder-specific differences are the cause of an apparent load change.

As Fig. 2 shows, of the fuel quantity mKE injected into the intake manifold of the engine, only the part mKα is fed into the cylinder, while the other part mKFZ initially remains as a fuel film on the wall of the intake manifold. On the one hand, fuel is deposited on the wall, but on the other hand it is also drawn off the wall and fed into the engine cylinder. A fuel quantity mKF therefore remains on the wall, of which the portion mKFA is drawn off the wall and fed into the cylinder. So: The total fuel quantity mKA entering the engine cylinder is made up of a directly injected portion mKα and a portion mKFA caused by the reduction of the wall film. The proportion of fuel that is deposited as a wall film and the proportion that evaporates from the wall film depends largely on the pressure and temperature of the air in the intake manifold.

Since the method, equipment and algorithms for compensating for the wall film effects described are known, the multiplicative and/or additive load change compensation is shown as a single stage 38 and will not be discussed in more detail. It should be noted, however, that the load change compensation strategy can be calculated from either the main load signal or the throttle valve load signal, or from the latter only if the switch 30 has been switched.

The multiplicative and/or additive load change compensation is multiplied or added (as the case may be) to the basic signal t1f(k) in stage 40, and in this way the adjusted normal angle-synchronous fuel injection times teN are obtained. The fuel injection described so far is conventional in every way, apart from the aforementioned cylinder-specific release function 36.

The fuel injection method described works very well in a stationary state with a constant load, even when the load only increases or decreases slowly. Under the conditions mentioned, the pressure sensor (or other main load sensor) works very precisely and delivers the correct amount of fuel. On the other hand, the signals from the main load sensors are not particularly suitable for larger load changes ("dynamic" conditions). In particular, the evaluated signal from the pressure sensor follows a rapidly increasing load too slowly, resulting in a mixture that is too lean, while on the other hand, a mixture that is too rich is achieved when the load drops quickly. Main load sensors of other types have the disadvantage that the evaluated signals from these sensors "overshoot" considerably beyond the correct value when there is a rapid load change (hot wire sensors) or oscillate around the correct value (flap sensors), so that the mixture is sometimes too rich for the engine and sometimes too lean. This behavior is shown in Fig. 6(a), (c) and (d).

A corresponding behavior is avoided according to the invention by additionally detecting the angular position of the throttle valve 18, as already described earlier. The throttle valve is usually connected to a potentiometer for detecting the full load position of the valve. It has now been proposed to detect the engine load via the angular position of the throttle valve, but this is not a practical way, since the signal is too inaccurate for stationary or slowly changing loads (a simple potentiometer is used) and therefore non-stoichiometric mixtures are achieved. On the other hand, for rapid dynamic changes, the load signal calculated from the throttle valve opening is far more useful than the other line signals (see Fig. 6(b)).

Fig. 4 shows the relationships between various engine parameters when the load increases significantly. From Fig. 4d it can be seen that the pressure difference, as described above, "lags" considerably during the rapid load change, while the signal α for the throttle valve angle and in particular the signal tLW derived from it follows the load change much more quickly or even leads it. The corresponding case - but in the opposite sense - occurs when the load is reduced quickly.

Referring again to Fig. 5, the signal α corresponding to the throttle valve angle is sampled at regular intervals (eg every 10 ms corresponding to a CLOCK signal) and is combined with an input signal from the engine concerning the engine speed n in the stage 42 in order to obtain a separate angle-synchronous basic injection duration t1w(j). This process is shown in Fig. 3, in which the output signal t1w(j) is shown for each value α of the throttle valve angle. The representation is only for three engine speeds, but it is clear that the relationship t1w(j)/α is stored for many other engine speeds and interpolation of intermediate values takes place. The rate of change of this quantity is determined in step 44, and in step 46 a decision is made as to whether the rate of change exceeds a specified threshold value or not.

If this threshold is exceeded, there are three options that can be chosen by appropriately programming the electronic control unit:

1. The switch 30 can be switched so that the output of the stage 42 is connected to the filter 32. In addition, the parameters of the filter 32 can be changed (but this is not essential) to reduce or eliminate the filter damping, since during a sharp acceleration or deceleration it is more important that a signal as current as possible is used to determine the normal angle-synchronous fuel injection quantity, or a special filter value can be selected for example for a better transition from one signal to another with a difference in value. It can be seen from Figs. 4 and 6 that the throttle valve load signal leads the actual load change of the engine and can therefore be preferred when the engine load changes sharply, since the signal t1w(j) obtained via the throttle valve angle is more accurate than the signal t1(k) from the main load sensor and by leading the signal, a time delay in signal processing can be compensated. The signal t1w(j) is then treated in the same way as the previously used signal t1(k), apart from the fact that it is possible to change or cancel the filtering at stage 32.

2. The second option would be for switch 30 to remain in the normal position, but the sampling of the pressure from the pressure sensor is carried out in a different way and/or the filter function 32 is weakened or eliminated, as described for the first option. It can be seen that a value is used for the steady-state pressure signal which is "out of date" by half a segment. With slowly changing loads or under steady-state conditions this is meaningless, while with rapidly changing loads a mixture that is too lean is achieved. As a result, the sampling is modified so that the pressure measurement for the pressure value is carried out at the most current reference mark tR, without averaging. This enables the pressure signal and the load signal calculated from it to follow the actual engine load more quickly with rapid load changes, with a smaller deviation of the air-fuel ratio as a result.

3. The third possibility is that the switch 30 remains in the normal position and - especially in the case of a hot-wire air flow meter - the sampling is carried out normally or in a modified manner, but the filter parameters 32 are selected to be relatively high, so that a load signal is obtained that does not react to an overshoot of the signal from the hot-wire air flow meter as a result of the intake manifold filling, but responds very precisely to the intake manifold pressure, as shown in Fig. 6a with a dashed line. The described features can be achieved with a simple first order low-pass filter.

The first aforementioned method of modifying the signals from the main load sensor by monitoring the throttle valve angle signal allows for a much more up-to-date angle-synchronous fuel injection signal, especially in the case of a plate-type air flow meter or a very slow pressure signal. The second and third aforementioned methods are preferable for turbocharged engines or engines with a bypass idle speed control system where the signal tLW does not necessarily correspond to the engine load. In all three cases, the modified strategy would remain effective during a time period TTLFU after the difference tLW is no longer greater than the threshold TLFU, as shown in Fig. 6b.

In addition to the three possibilities mentioned above, stage 48 also determines whether the rate of change of the injection duration calculated from the throttle valve angle signal reaches a second threshold value or not (this threshold value can be smaller, equal to or larger than the threshold value in stage 46). Exceeding the second threshold value mentioned means a rapid increase in the load on the engine, and the electronic control unit then causes further asynchronous injection pulses to be carried out to enrich the mixture. The threshold is usually set at a point at which sufficient enrichment of the cylinder mixtures is not possible by normal angle-synchronous injection. Fig. 4b shows, for example, how the normal, angle-synchronous injection pulses become longer when the load increases, without preventing the mixture in cylinder 4 from becoming leaner, and only partially preventing it from becoming leaner in cylinder 2. In order to prevent leaner combustion, asynchronous intermediate injections are carried out. They can take place at different time intervals during each segment, e.g. at intervals of 10 ms. The fact that several asynchronous injections take place means that in all cases fuel can be injected when the intake valve is still just open. Injecting the entire intermediate injection quantity at once would be impractical, as this could lead to over-enrichment in certain cylinders, e.g. cylinder 4, which only requires a relatively small increase in filling.

The amount of fuel to be injected is calculated taking into account a factor at stage 50. This factor could, for example, be based on the temperature of the engine, an adaptive correction according to the control of the air-fuel ratio or something else. The signal obtained is the asynchronous basic injection time tezcal. If it is determined at stage 52 that the calculated width of the asynchronous pulse is below a certain threshold value, the intermediate injection in question is suppressed because the injection must be minimal in view of the opening and closing times of the injection valve. If the pulse width is above the threshold value, the signal is retained, but if it is below the threshold value, it is stored in memory 54 and added to the pulse width of the following asynchronous intermediate injection. This case is shown in Figs. 4b2 and 4b3 under A.

To ensure that following asynchronous intermediate injection, normal angle-synchronous fuel injection does not lead to an over-rich mixture, the duration of all intermediate injection pulses in each segment is summed at stage 56 and is partially or completely subtracted - as determined at stage 58 - from the following normal, angle-synchronous injection pulse duration. The asynchronous injection pulses themselves open the injection valve, unless they are emitted simultaneously with a normal, angle-synchronous injection, as shown at point b in Fig. 4b2. In this case, the control unit 20 actuates a switch 60 to disconnect the injection valve from the asynchronous pulse generating circuit, and this part of the circuit is instead connected to the normal, angle-synchronous pulse circuit part, so that the calculated duration of the asynchronous intermediate injections mentioned is added to the calculated normal, angle-synchronous injection duration.

If the electronic control system decides that the compensation achieved by the normal, angle-synchronous injection pulses is not sufficient, then one or more asynchronous intermediate injection pulses can be carried out in each segment so that the mixture is quickly enriched and the mixture better meets the load requirements, i.e. comes closer to the stoichiometric ideal case.

The asynchronous additional pulses are of course usually only required during a rapid load increase, but not during a rapid load reduction.

To use the lead characteristic of the signal tLW during delay, with a decision by the switch 60, negative intermediate injection pulse widths are added in a memory 61 and corrected by a factoring circuit 62. The resulting negative injection duration is added to the next synchronous injection pulse to reduce the pulse width.

A filter can be provided at the input of stage 48 to smooth out signal fluctuations and to cause a slow signal drop after triggering a difference tLW, so that intermediate injections occur not only during tLW, but also in the subsequent sampling intervals with decreasing pulse width. A corresponding representation in Fig. 5 has been omitted for the sake of simplicity.

The algorithm discussed has been described for a system with simultaneous injection. A specialist could easily adapt it to a sequential or basic injection system. If the main load sensor or the throttle valve angle sensor fails, the system can be switched to function with the remaining sensor.

It should be noted that other functions can also be carried out with the signals from the normal, angle-synchronous circuit part and the signals relating to the asynchronous intermediate injection. However, how an injection pulse signal that has been calculated can be treated is familiar to the expert and is not part of the present invention.

Figure labels in German/English Figure 1 (Legend)

a) Electronic Control Unit Electronic Control Unit

Figure 3 (Legend)

a) 3000/min 3000 r.p.m.

Figure 2 Figure 5 (Legend)

a) Sampling Sample

b) Filters Filters

c) Release Enable

d) Load change compensation Load Change Compensation

e) X and/or + X And/Or +

f) to fuel injector To Fuel Injector

g) Clock signal Clock

h) Factoring Factor

i) Factoring Factor

j) Factoring Factor

k) Trigger signal Trigger

l) Storage Store

m) To fuel injector To fuel injector

Figure 4 (Legend)

a) Inlet Inlet

b) Exhaust Exhaust

c) Inlet Manifold Pressure P Inlet Manifold Pressure P

Figure 6 (Legend)

a) tL without filter tL without filter

b) tL filtered filtered tL

c) Stationary tL filter Stationary tL filter

d) Effective time for dynamic tL filter or/and use of tLW effective time for dynamic tL filter or/and use of tLw

e) Stationary tL filter Stationary tL filter

f) tL with normal integration and filtering tL with normal integration and filter process

g) tL without filter/integration tL without filter/integration

Claims (4)

1. Fuel injection system for an internal combustion engine with an engine main load sensor (22 or 24) and means (20) for calculating a basic pulse width for angle-synchronous fuel injections (t₁(k)) from signals received from said main load sensor, a further sensor in the form of a throttle valve angle sensor (18) for monitoring the opening (a) of an engine throttle valve and means (20) for changing the calculation of the width of the fuel injection pulses when the rate of change of a signal derived from the throttle valve opening reaches a specified value, the calculation change comprising a change in the filter characteristics of a filter function which is usually carried out for the angle-synchronous basic pulse width, characterized in that during a determined deceleration under certain operating conditions intermediate injection pulses are generated between the angle-synchronous basic injection pulses be initiated and the signal from the engine main load sensor is sampled in a different way. 2. Fuel injection system according to claim 1, characterized in that the system initiates one or more intermediate fuel injection pulses between the angle-synchronous pulses, with calculation from the rate of change of a signal for the throttle valve opening, preferably a signal tLW which depends on the throttle valve opening and speed, after reaching a fixed value. 3. Fuel injection system according to claim 2, characterized in that a negative correction pulse width is derived from a negative rate of change of the signal tLW in order to correct the angle-synchronous pulses in the event of a rapid load reduction. 4. Fuel injection system according to claim 3, characterized in that the negative asynchronous pulse widths are summed following the preceding angle-synchronous pulse and the sum is subtracted in whole or in part from the following angle-synchronous pulse.