EP0407406B1 - Learning control process and device for internal combustion engines - Google Patents

Abstract

In a learning process for regulating and presetting the lambda value of an air/fuel mixture to be supplied to an internal combustion engine (11), a large comparison value is compared with a small comparison value. The large comparison value is obtained by averaging adaptation factors for large pilot control values and the small comparison factors for small pilot control values. If the large comparison value is smaller than the small comparison value, a global summand is increased by a correction value, and decreased in other cases. The advantage of a device which operates according to this process is that disturbances having a cumulative effect on the injection time are compensated with a high degree of precision.

Description

The invention relates to a learning control method with pilot control for setting the lambda value for the air / force mixture to be supplied to an internal combustion engine. The invention also relates to a device for carrying out such a method (see in each case 1st part of claims 1, 5).

State of the art

Such a method and an associated device are known from DE 3505965A1 (US Ser. No. 831476/1986). The device has a pilot control means, a setpoint generator means, a control means and an adaptation factor memory. The method is used, for example, to set the injection time. The pilot control means outputs a pilot control value for the injection time as a function of values from operating variables other than the injection time. The setpoint generator provides a single controlled variable setpoint, namely the lambda value 1. This is compared with the respective actual lambda value, which is measured by a lambda probe. The control means depends on the difference between the above Both values are a manipulated variable, namely a control factor with which the respective pilot control value is corrected by multiplication. However, the pilot control value is also corrected in a controlling manner, with the aid of an adaptation factor which is respectively read out of the adaptation factor memory. The adaptation factor memory stores adaptation values in an addressable manner via values of addressing operating variables. To correct the precontrol value, it reads out the adaptation factor that belongs to the set of values of the addressing operating variables that is present at the time. The pilot control value is linked multiplicatively with this factor. The adaptation factors are determined again and again with the help of the control factor supplied by the control means. In predetermined larger time periods, the factors of the adaptation factor memory are evaluated in such a way that the average of all factors is formed and this average is incorporated into a so-called multiplicative global factor. This value then takes into account global corrections that are necessary both due to the multiplicative disturbing influences on the injection time and the additive acting disturbing influences.

Interfering influences that have an additive effect are better taken into account in a process as described in SAE paper no. 860594, 1986 is also known for adjusting the injection time. In addition to the above-mentioned functional levels, the associated device also has a summand determining means which determines a summand which is added to the pilot control value corrected by multiplicative factors. The summand is measured at idle, i.e. with short injection times. This is due to the consideration that with short injection times there is a multiplicative effect Interference is relatively weak, but an additive interference has a relatively strong effect.

The system just mentioned has the following disadvantage. It can easily happen that even with short injection times, an additive disturbing influence is compensated by an opposing multiplicative one. Then the pre-control time is not corrected additively (and counter-multiplicatively), although this would actually be necessary. This error, which results from the determination during idling, affects the entire load and speed range of the internal combustion engine.

The invention is based on the object of specifying a method for learning control with precontrol for setting the lambda value, which takes disturbing influences which have an additive effect on the metering of the fuel quantity into account better than known methods. The invention is also based on the object of specifying a device for carrying out such a method.

Advantages of the invention

The invention is given for the method by the features of claim 1 and for the device by the features of claim 5. Advantageous further developments and refinements are the subject of the dependent claims.

The device according to claim 5 has the means already described, that is to say a pilot control means, a setpoint generator means, a control means, an adaptation factor memory and a summand determination means. It also has a comparator means and a changing means. The comparator means compares a large comparison value with a small comparison value and outputs an increase or a decrease signal. The changing means increases the global summand in response to the increase signal by a correction value or decreases the summand in response to the decrease signal.

The method according to the invention compares a large comparison value with a small comparison value, the large comparison value being formed by averaging adaptation factors for large pilot control values, while the small comparison value is formed by averaging adaptation factors for small pilot control values. If the large comparison value is smaller than the small comparison value, the additive for additively correcting the pilot control value is increased by a correction value, otherwise it is decreased.

This measure is based on the following finding. With a short injection time, i.e. a small pilot control value, the additive error in the pilot control value is, for example, + 5% and the multiplicative error is also 5%. The total error is then 10% and the adaptation factor is 1.1 as long as no additive correction is carried out. If the injection time is five times longer, the fixed additive error is only 1%, while the multiplicative error is still 5%. The total deviation thus amounts to 6% and results in an adaptation factor of 1.06 as long as no additive correction is made. However, if the pilot control time is corrected not only by the adaptation factor, but also by a summand, the situation changes. It is assumed that the summand has been determined exactly correctly, that is to say add the short period of time to the pilot control time that is required to compensate for the additive error. Then only the multiplicative error remains, both for short and Even for long injection times, an error of 5% in the example case, i.e. an adaptation factor of 1.05. The example illustrates the knowledge that a smaller adaptation factor for long injection times in comparison to the adaptation factor for short injection times is a sign that an additive error is present and that a summand is expediently added to the respective pilot control time for correction.

This measure also remedies the above-described lack of the simulated unnecessary correction due to the effect of opposite influences. If there is an additive error, e.g. + 5% and a multiplicative error of -5%, this leads to an adaptation factor of 1.0 for the short injection time considered, but to a factor of 0.96 for a five times longer injection time (+ 1% additive, -5% multiplicative). In this case too, the adaptation factor for the long injection time is smaller than the adaptation factor for the short injection time, which, as explained, is the sign of the need to add a correction sum.

In the example just described to explain the principle of the invention, only one adaptation factor for a short and a long injection time was assumed in comparison. In practice, however, it is more advantageous to form a large comparison value from a plurality of adaptation factors for large input control values and to calculate a small comparison value accordingly for small input control values. Then not only two adaptation factors, but the two comparison values are compared with one another by the comparator stage. For the formation of these comparison values are for different system structures Different methods are advantageous, which will be explained in more detail below using exemplary embodiments.

With regard to the vibration stability of a system controlled by this method, it is particularly advantageous not to change the summand with every small deviation between the large and small comparison value, but rather to make a change only when a predetermined threshold value is exceeded. Smaller fluctuations then do not lead to changes in the system parameters.

It also helps to stabilize against tendencies to vibrate by changing the addend only by a small predetermined fixed correction value, regardless of the difference between the large and small comparison value. Deviating from this and using a correction value, the size of which is proportional to the difference between the large and small comparison value, is only advisable if the comparison values are formed by averaging a relatively large number of individual values, so that a large change in the summand quickly results in a Changes in the control parameters brings, but this leads only to weak feedback on the comparison values.

drawing

Fig. 1

ein als Blockschaltbild dargestelltes Funktionsdiagramm eines lernenden Vorsteuerungs/Regelungs-verfahrens zum Einstellen der Einspritzzeit unter anderem mit Hilfe eines globalen Summanden;

Fig. 2

ein als Blockschaltbild dargestelltes Funktionsdiagramm derjenigen Funktionsgruppe innerhalb von Fig. 1, die den globalen Summanden bestimmt; und

Fig. 3

ein als Blockschaltbild dargestelltes Funktionsdiagramm einer Variante einer Funktionsuntergruppe innerhalb von Fig. 2.

The invention is explained in more detail below on the basis of exemplary embodiments illustrated by figures. Show it:

Fig. 1

a functional diagram shown as a block diagram of a learning pre-control / regulating method for setting the injection time, inter alia with the aid of a global summand;

Fig. 2

a functional diagram shown as a block diagram of that functional group within FIG. 1 that determines the global summand; and

Fig. 3

3 shows a functional diagram of a variant of a functional subgroup within FIG. 2, shown as a block diagram.

Description of exemplary embodiments

1 and 2 relate to a single exemplary embodiment, FIG. 1 giving an overall overview of a precontrol / regulation method for setting the injection time for an injection valve of an internal combustion engine 10, while in FIG. 2 the function group most important for the invention within FIG. 1 is shown in detail.

An injection valve 12 is arranged in the intake manifold 11 of an internal combustion engine 10 and is controlled with a signal for the injection time TI. Depending on the amount of fuel injected and the amount of air drawn in, a lambda value is set, which is measured by a lambda probe 14 arranged in the exhaust duct 13 of the internal combustion engine 10. The measured actual lambda value is compared with a desired lambda value supplied by a setpoint generator 15 in a comparison step 16, and the control deviation value formed is fed to a control means 17 with integrating behavior, which outputs a control factor FR as a manipulated variable. With this control factor, a pilot control time TIV for the injection time is modified by multiplication in a multiplication step 18. The pilot control time TIV is provided in the exemplary embodiment shown by a pilot control memory 19 which addressable via values of the speed n and the position of an accelerator pedal FP stores pilot control times TIV.

The pilot control times TIV are defined for certain operating conditions and certain system properties. Now, however, the operating conditions, for example the air pressure or the system properties, for example leakage air properties or the closing time of the injection valve 12, change during operation of the internal combustion engine Adaptation factor FA (FP, n) modified. This adaptation factor is read from an adaptation factor memory 21, which has a corresponding number of support points as the pilot control memory 19 and, like this, can be addressed via sets of values of the speed n and the accelerator pedal position FP. For example, there are 64 support points each with k = 8 addresses for classes of accelerator pedal positions FP and l = 8 addresses for classes of speed values n. The respective adaptation factor FA is also multiplied by the multiplying step 18, as is a global factor FG. Strictly speaking, the following multiplicative correction should take place:
$$\text{TIV × (FG × FA (FP, n) × FR).}$$

However, since all correction factors in practice only deviate a few percent from 1.0, the value just mentioned corresponds approximately to the following value:
$$\text{TIV × (FG + FA (FP, n) + FR) = TIV × F.}$$

In the system according to the exemplary embodiment shown, the factor F formed by summing the correction factors is multiplicatively linked in multiplication step 18 with the respective pilot control time TIV. Instead, there could also be three multiplier levels.

In addition to the multiplicative correction, the pilot control time is also subjected to an additive correction by a global addend in an adding step 27. The injection time TI calculated as follows is thus supplied to the injection valve 12:
$$\text{TI = TIV × F + SG.}$$

The adaptation factors FA, the global factor FG and the global summand SG are formed in an adaptation means 22 which has three functional subgroups, namely an adaptation factor calculation means 23, a global summation calculation means 24 and a global factor calculation means 25. The function is of particular interest of global summand calculation means 24, which is explained in more detail below with reference to FIG. 2. First, however, the function of the adaptation factor calculation means 23 and the global factor calculation means 25 will be briefly discussed. The two calculation means just mentioned can work as described, for example, in DE 3505965A1 already mentioned at the beginning. This is because the control means FR is fed to the adaptation means 22 via an averaging step 26, and a new value is calculated from this on the basis of the old adaptation factor for a support point whenever the values of the addressing operating variables move in a range that corresponds to the support point under consideration heard, and then this area is left. The newly determined adaptation factor is transferred to the adaptation factor memory 21 after it has been determined, so that it is available as an improved value when an operating state occurs again with the same values of the addressing operating variables.

In predetermined larger periods of time, the average value is formed from all the adaptation factors in the adaptation factor memory 21 and the global factor FG, which previously applied, is modified with this. The adaptation factors of previously visited support points are corrected.

However, the adaptation factors FA and the global factor FG can be obtained in any way. The procedures according to the above-mentioned document serve only as an example. They have no influence on the acquisition of the global addend SG described below.

To obtain the global summand, the global summand calculation means 24, whose function is shown in detail in FIG. 2, has an average value calculation means 28, a large comparison value means 29.G, a small comparison value means 29.K, a comparator means 30, a correction value memory 31 , a switching step 32 with switch actuating means 33, a linking means step 34 and a sample / hold means (S / H) 35.

The mean value calculation means 28 calculates the mean value from all pilot times TIV, as they are stored for the k × l, that is to say 8 × 8 support points of the pilot control memory 20, and divides the sum by the value k × l. The mean so obtained TIV _{k, l} serves only to be able to distinguish for which values of the indices k and l feedforward times TIV _{k, l are} greater than the mean and for which values of the indices the pilot control times are smaller. This information is important for the two comparison means. The large comparison value mean 29.G namely forms the sum of all adaptation factors which are stored under those values of the reference point indices k and l for which the respective pilot control time is greater than the mean value of all pilot control times in the pilot control memory 20 of the same index. Correspondingly, the small comparison value mean 29.K forms the sum for all adaptation factors FA _{k, l} that belong to pilot control times that are smaller than the mean value of all pilot control times. The difference between the two sums is formed by the comparator means 30, which outputs a difference signal D. If the large comparison value supplied from the large comparison value mean 29.G is greater than the small comparison value supplied by the small comparison value mean 29.K, ie if the difference D is negative, the correction value memory 31 outputs a negative fixed correction value -ΔSG, otherwise one fixed positive correction value + ΔSG of the same size. The difference signal D is also fed to the switch actuating means 33, which executes the switching step 32 when the amount of the difference exceeds a threshold value D₀. The positive or negative correction value ΔSG is then added to the old global summand SG stored in the sample / hold means 35 in the linking step 34, as a result of which a new increased or decreased global summand SG is formed. As explained further above, a difference signal D occurs as long as the global summand SG acting additively on the pilot control time is not correctly determined and the adaptation factors for large injection times deviate from those for small injection times.

A variant of the function groups for obtaining the large comparison value and the small comparison value is shown in FIG. 3. Instead of the mean value calculation means 28 and the two comparison value means 29.G and 29.K, there are only the two comparison value means in a different mode of operation, namely a large comparison value means 29.G3 and a small comparison value means 29.K3, to which the adaptation factors FA _{k, l are} supplied. In the comparison value means itself is stored, the values for which k₉ L₉ and the indices k and l are valid and for which values k and l _k the indices _k small pilot control values are relatively large pilot control values. The summation takes place for adaptation factors with the corresponding indices.

The method according to FIG. 2 with the mean value calculation means 28 has the advantage of great flexibility, but the disadvantage of a certain computing effort. The flexibility is due to the fact that devices of the type described here are generally designed in microcomputer technology and that when a device is adapted to a particular engine type, essentially only the values stored in the pilot control memory 20 need to be changed. If the variant according to FIG. 3 is used, the values of those indices for which large or small pilot control times apply now generally have to be specified for the adaptation to a new motor type. However, if these values are stored, the system according to FIG. 3 has the advantage that the calculation effort for forming the mean value of the pilot control times is eliminated.

The computational effort can be reduced even further, for the fewer adaptation factors, the sum is formed by the comparison value means 29.x. In the borderline case, it would be sufficient compare the adaptation factor that belongs to a support point with a particularly long pilot control time with an adaptation factor that belongs to a support point with a particularly short pilot control time. However, this only works with a method that ensures that these nodes are regularly adapted, for example by a method for adapting distant nodes or by a method that works with a global multiplication factor. Such methods are described in DE 3505965A1, which has already been mentioned several times. However, it is safer to calculate the sum of the adaptation factors over as many support points as possible.

Forming the sum over many support points also has the advantage that a strong change in the adaptation factor of a support point has a relatively weak effect on the sum as a percentage. This reduces the tendency of the system to vibrate. Then the correction value can also be determined according to a variant as indicated in brackets in FIG. 2 in the symbol for the correction value memory 31, namely in that the value is obtained by multiplying the value of the difference signal D by a proportionality constant M. The global summand SG is then corrected the more the larger the value of the difference signal D is. This has the advantage that the method can react quickly to larger, additive-acting faults. The disadvantage, however, is that vibrations can occur due to the existing feedback. As already explained, this tendency to oscillate is reduced if the feedback is weak due to the fact that a changed adaptation value has only a weak effect on the value of the difference signal.

It has already been indicated at various points in the description that details of the exemplary embodiment are irrelevant to the invention. In addition to these statements, it should also be mentioned here that pilot control times TIV can also be obtained by dividing the signal supplied by an air mass sensor by the rotational speed, as is customary in commercially available devices. In this case, however, the variant according to FIG. 2 for obtaining the comparison values is ruled out, and only variants can be carried out in which it is determined in advance for which indices of support points adaptation factors are to be added. It should also be pointed out that the setpoint generator 16 does not have to be designed as a map, as shown in FIG. 1, but that the setpoint can also be determined differently, in particular that the only fixed lambda setpoint "1" can be specified.

In the exemplary embodiment, the condition for changing the global summand SG was that the magnitude of the difference signal D should be greater than a threshold value D₀. As already mentioned, this has the advantage that the global summand is not immediately changed with every small change in an adaptation factor, which would increase the tendency to oscillate. Depending on the tendency of the entire system to vibrate, other conditions can also be used, e.g. that the global summand is corrected after a predetermined time or that the correction takes place after a predetermined number of corrections of adaptation factors.

It is essential for the invention that a global summand is formed depending on the difference between adaptation factors for large pilot control values and adaptation factors for small pilot control values, the summand being increased, if the difference is negative and it is lowered if the difference is positive.

The correction values by which the global summand is increased or decreased can have different sizes. The concrete values are to be determined in such a way that the adaptation is as quick and good as possible with a low tendency to vibrate.

Claims (5)

1. Process for learning control with pilot control for setting the lambda value for the air/fuel mixture to be fed to an internal-combustion engine, in which a pilot control value is corrected by a control system correcting value, by adaptation factors and a global summand, characterised in that global summands are established as follows:

― a large comparison value is compared with a small comparison value, the large comparison value being formed by averaging adaptation factors for large pilot control values and the small comparison value being formed by averaging adaptation factors for small pilot control values, and

― the global summand is increased by a correction value if the large comparison value is less than the small comparison value, and in the converse case, the correction value is decreased (sic).

2. Process according to Claim 1, characterised in that an identical fixed value is used as correction value respectively for the decreasing and the increasing.

3. Process according to Claim 1, characterised in that the correction value is determined proportionally to the difference between the large comparison value and the small comparison value.

4. Process according to one of Claims 1-3, characterised in that the global summand is changed only when the amount of the difference between the large comparison value and the small comparison value exceeds a predetermined thereshold value.

5. Device for learning control with pilot control for setting the lambda value for the air/fuel mixture to be fed to an internal-combustion engine, having

― a pilot control means (19), which outputs, depending on values of operating variables, a pilot control value for a fuel-metering operating variable,

― a reference value generator means (15) for outputting the lambda reference value,

― an automatic control means (17), which forms a control factor as correcting value depending on the difference between the lambda reference value and the respectively measured lambda feedback value, with which correcting value the respective pilot control value is corrected under closed-loop control by multiplication,

― an adaptation factor memory (21), which stores adaptation factors addressably by means of values of addressing operating variables and in each case outputs that adaptation factor which belongs to the respectively applying set of values of the addressing operating variables, with which adaptation factor the pilot control value is additionally multiplied for correcting under open-loop control, and

― a global summand calculation means (24), which establishes a summand, which is added to the pilot control value corrected by the multiplicative factors,
characterised in that the global summand establishing means has the following functional means:

― a comparator means (28, 29.G, 29.K, 30), which compares a large comparison value with a small comparison value, the large comparison value being formed by averaging adaptation factors for large pilot control values and the small comparison value being formed by averaging adaptation factors for small pilot control values, and outputs an increase signal if the large comparison value is less than the small comparison value and, in the converse case, outputs a decrease signal, and

― a changing means (31-35), which increases the global summand by a correction value in response to the increase signal or decreases it by a correction value in response to the decrease signal.

Images