JP2528335B2 - Control device for internal combustion engine - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a control device for an internal combustion engine that performs feedback control, and more particularly to a control device that uses fuzzy reasoning to improve control accuracy.

<Prior Art> An example of conventional feedback control of an internal combustion engine is ignition timing control. An example will be described with reference to FIG. 7. First, the basic ignition timing θ ₀ is obtained based on the detected value of the engine operating state (rotational speed load) (step 1).

Next, the crank angle θ _{Pmax at} which the maximum value of the in-cylinder pressure measured by the in-cylinder pressure sensor is given is detected (step 2).

Next, the θ _Pmax is compared with a crank angle θ _MBT (for example, 15 °) that gives the optimum ignition timing MBT (step 3),
If θ _Pmax is smaller than θ _MBT , the ignition timing θ _I is delayed by a unit angle α (step 4), and if θ _Pmax is larger than θ _MBT , the ignition timing θ _I is advanced by a unit angle (step 5). (See JP-A-62-35065).

<Problems to be Solved by the Invention> However, in such a conventional feedback control method, the same control as that in the steady state is performed even in the transient state of engine operation. Control cannot follow transient changes well,
In the case of the ignition timing control does not generate the optimum torque,
In knocking control, there are problems such as knocking cannot be avoided quickly. If you try to finely tune to avoid this, a lot of man-hours will be required.

The present invention has been made in view of such conventional problems, and uses fuzzy reasoning so that optimum control characteristics can always be obtained, and determines the shape of the membership function used in the fuzzy reasoning in the engine operating state. It is an object of the present invention to provide a control device for an internal combustion engine, which solves the above-mentioned problems by making changes according to the above.

<Means for Solving Problems> Therefore, according to the present invention, in a control device for an internal combustion engine that feedback-controls a predetermined control target of the engine according to the operating state detected by the operating state detecting means, membership in fuzzy inference Membership function storing means for storing a function, membership function correcting means for correcting the membership function stored in the membership function storing means in accordance with the detected driving state, and depending on the detected driving state Control amount setting means for performing fuzzy inference using the membership function corrected by the membership function correcting means to set the control amount of the controlled object.

<Operation> The membership function stored in the membership function storage means is modified by the membership function modification means in accordance with the driving status detected by the driving status detection means, and the fuzzy inference is performed using the modified membership function. Thus, the control amount of the controlled object such as ignition timing and air-fuel ratio is set by the control amount setting means.

As a result, for example, even during a transient operation, even if the deviation amount of the detected value becomes large with respect to the control target value of the controlled object, the control for increasing the feedback correction amount of the control amount according to the amount of the deviation amount. Can be performed, and the transient drivability can be improved by promptly converging to the target value. Further, even when the operating state temporarily changes during steady operation, the original state can be promptly restored.

<Example> An example of the present invention will be described below with reference to the drawings.

 FIG. 2 shows the system configuration of an embodiment of the present invention.

In the drawing, an in-cylinder pressure sensor 2 for detecting the pressure (in-cylinder pressure) in the combustion chamber 1a is attached to the engine 1 as a washer of a spark plug 3, and an in-cylinder pressure signal from the in-cylinder pressure sensor 2 is sent to an amplifier 4 After being amplified by, it is input to the controller 5 that sets the ignition timing by fuzzy reasoning.

The ignition advance value signal set by the controller 5 is input to the ignition control circuit 6, and the control circuit 6 outputs the ignition coil 8 based on the ignition advance value signal and the crank angle signal input from the crank angle sensor 7. By controlling the energization of the transistor 9 connected to the primary side of the ignition coil and cutting off the energization at the set ignition timing, a high voltage is generated on the secondary side of the ignition coil 8 to spark the ignition plug 3 and ignite it. I do.

The signal from the crank angle sensor 7 is sent to the controller 5
Is also entered.

Further, the in-cylinder pressure sensor 2 and the crank angle sensor 7 correspond to operating state detecting means.

Next, a method of setting the ignition advance value by using the fuzzy inference by the controller 5 will be described with reference to the flowchart shown in FIG.

First, in step 11, based on the in-cylinder pressure P _C input from the in-cylinder pressure sensor 2 via the amplifier 4 and the crank angle signal θ input from the crank angle sensor 7, a crank angle θ _Pmax that maximizes P _C is calculated. .

Next, at step 12, the above-mentioned θ _Pmax and optimum ignition timing θ _MBT
The deviation Δθ from

In step 13, fuzzy inference is performed on the deviation Δθ to set the ignition advance value θ _ADV .

 The fuzzy inference is performed as follows.

Fig. 4 enumerates the linguistic control rules in this control. As shown in Fig. 5, for each of the antecedent and consequent propositions of each term, the probability, that is, the antecedent with the fuzzy quantity set. The membership function and consequent membership function are stored in the ROM of the microcomputer constituting the controller 5. That is, the ROM constitutes a membership function storage means.

Here, in contrast to the one-term control rule that "advances the ignition timing when Δθ is positive," and the two-term control rule that "retards the ignition timing when Δθ is negative." For each, the antecedent membership function is also made fuzzy, and ultra-fuzzy inference is performed.

That is, the antecedent membership functions of the first term and the second term are set as, for example, ψtan ⁻¹ Δθ + C, −ψtan ⁻¹ Δθ + C,
The shape changes depending on the value of the parameter ψ. The value of the parameter ψ is determined by the antecedent membership function and the antecedent membership function for the control rules of the fifth and sixth terms.

That is, as shown in FIG. 6, the area characteristic of A shown in FIG. 6 is obtained for the control law of the term 5 that “the parameter ψ is increased when the absolute value | Δθ | is large” of the deviation Δθ, and “| Δθ | When the parameter is small, the parameter ψ is reduced. ”Similarly, the area characteristic of B in the figure is obtained for the control rule of 6 terms, and when these are superposed, the value of Δθ located at the center of gravity of the areas of A and B is ψ. Calculate as (= ψ ₁ ) (S13
A).

By substituting ψ ₁ thus obtained for the parameter ψ of the membership function of the first and second terms, the shapes of these functions are uniquely determined (S13B). This part corresponds to the membership function correcting means. Using the antecedent membership function and consequential membership function of the first and second terms determined as described above, the amount of advancement value correction (change amount) is calculated in exactly the same manner as shown in FIG. ) Calculate Δθ _AVD (S13C).

Here, the parameter ψ increases as the | Δθ | increases due to the fuzzy inference using the control terms 5 and 6, and the gain of the antecedent membership function of the terms 1 and 2 increases accordingly. Therefore, the correction amount Δθ _ADV of the advance angle value is set to increase as the value of | Δθ | increases, so that even if | Δθ | increases during the transition, | Δθ | is decreased in a short time to obtain a new optimum ignition. It is possible to quickly converge at a certain time, obtain an optimum torque, improve transient operation performance, and quickly avoid knocking.

Further, with respect to the engine speed N obtained on the basis of the signal from the crank angle sensor 7, the control side of the item 3 "when the engine speed N is large, the basic advance value θ _BASE is increased", when the engine speed is less than the basic advance value theta _bASE exactly the same way as is also shown in Figure 6 with respect to the four terms control law that a smaller "obtaining a basic advance angle value theta _bASE ( S13D).

Then, by adding the correction amount Δθ _{ADV of} the basic advance value θ _BASE obtained as described above, the ignition advance value Δθ
Determine _ADV (S13E).

These portions of S13C to S13E correspond to the control amount setting means.

The signal of the ignition advance value θ _ADV thus determined is output to the ignition control circuit 6 as described above, and the ignition is performed at the ignition timing corresponding to θ _ADV .

In addition, although the above embodiment shows the one applied to the ignition timing control, it can also be applied to the control of the air-fuel ratio and the like, and in this case, the effect that the air-fuel ratio during the transient operation can be appropriately set is also obtained. can get.

<Effects of the Invention> As described above, according to the present invention, in the feedback control in the internal combustion engine, the membership function used in the fuzzy inference for the control amount setting is modified according to the operating state. Even if the deviation amount from the target value increases during the transition, control can be performed such that the feedback correction amount becomes larger as the deviation amount increases, and thus the target value can be quickly approached. Stability is maintained by avoiding a situation in which the state of changes significantly.

Therefore, the engine operating state can be significantly improved by applying the control to the ignition timing and the air-fuel ratio.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of the present invention, FIG. 2 is a diagram showing a system configuration of an embodiment of the present invention, FIG. 3 is a flowchart showing an ignition advance setting routine according to the same embodiment, FIG. Is a diagram showing a control law of fuzzy reasoning used for setting the ignition advance value of the same as above, FIG. 5 is a diagram showing a membership function corresponding to the control law of the same as above, and FIG. 6 is a modification of the membership function above. FIG. 7 is a flowchart showing a method for obtaining parameters by fuzzy inference, and FIG. 7 is a flowchart showing an example of conventional ignition timing control. 1 ... Engine, 2 ... In-cylinder pressure sensor, 5 ... Controller,
7: Crank angle sensor

Claims (1)

(57) [Claims] 1. A control device for an internal combustion engine, which feedback-controls a predetermined controlled object of an engine according to an operating state detected by an operating state detecting means, comprising membership function storing means for storing a membership function in fuzzy inference. A membership function modifying means for modifying the membership function stored in the membership function storing means in accordance with the detected driving state; and a membership function modifying means for modifying the membership function in accordance with the detected driving state. A control device for an internal combustion engine, comprising: a control amount setting means for performing fuzzy inference using a membership function to set a control amount of the controlled object.