JPH01159466A - Knocking controller for internal combustion engine - Google Patents

Abstract

PURPOSE:To make speedy learning control performable by storing a knock correction value just before the transition in memory as a learning correction value at one driving area when an engine driving state makes its transition from the said one driving are to another driving one. CONSTITUTION:An engine driving state is divided into plural driving areas, and a learning correction value memory means 62, storing a learning correction value, is installed at each of these driving areas, and on the basis of the learning correction value, a knock correction value being increased or decreased according to a signal out of a knock sensor 17 by a knock correction value operational means 63. Then, ignition timing is compensated at an ignition timing compensating means 64 by the knock correction value operated. In a controller like this, when the engine driving state is made into transition from one driving area to another driving one, the knock correction value just before the transition is processed so as to be stored as the learning correction value in the said one driving area by a learning means 65.

Description

[Detailed description of the invention]

"Industrial Application Field" The present invention relates to an ignition timing control device for an internal combustion engine, and in particular:
The present invention relates to a knocking control device that detects knocking and controls the ignition timing to be optimal using learning control. ``Conventional technology'' Engine performance such as fuel efficiency and output is strongly influenced by ignition timing. If the ignition timing is controlled to optimize performance such as fuel efficiency and output, knocking may occur.0 Strong knocking may cause unpleasant knocking noises and may damage the engine, which is undesirable1. Regarding ignition timing, it is well known that the earlier the ignition timing, the stronger the knocking, and it is desirable to correct the ignition timing so as to suppress the knocking to a state just before it occurs or to a very slight state. Therefore, knocking control is used in which a vibration detector for detecting mechanical vibrations is attached to the engine's cylinder block or the like to detect knocking and correct the ignition timing. Even when looking at an engine, the knocking phenomenon varies greatly depending on various conditions such as the load state, rotational speed, air-fuel ratio, intake air humidity, and cooling water temperature. In order to flexibly respond to changes in these conditions and perform appropriate knock control,
Various methods have been proposed that perform learning control rather than mere feedback control. For example, JP-A-55-781
No. 68, JP-A-58-222976, JP-A-59-1
No. 36574, etc. All of these conventional devices are equipped with means for dividing the operating state of the engine into a large number of operating regions (sections) based on load, rotation speed, etc., and storing learning values corresponding to each operating region (for each division point). By using the ignition timing determined from the learned value as an initial value and performing feedback control based on knocking detection, the ignition timing is controlled to be optimal without time delay. However, a problem arises in the method of learning learning values. The former two devices repeatedly update the learning value according to the knocking condition every ignition cycle. This kind of learning control is performed during processing.

[This is possible in systems with ample margin for j, but there is a problem in that it is difficult to apply in systems where processing time is limited. In addition, the latter device forms a two-dimensional map with driving conditions as coordinates for each of the knock correction amount and the learned value. This involved performing a complex averaging process of changing the . This device has the advantage of being able to obtain more appropriate learning values, but
There is a problem in that it requires a large memory capacity and cannot be used in systems where the memory capacity of RAM, ROM, etc. is limited. "Problems to be Solved by the Invention" The present invention has been made to solve the above problems, and provides highly reliable learning values using a simple learning method that does not burden the processing time or memory capacity of the system. It is an object of the present invention to provide a knocking control device that can obtain the following characteristics and perform quick learning control. "Means for Solving Problems" The control concept of the present invention will be explained with reference to corresponding embodiment drawings. In the present invention, the operating state of the engine is divided into several operating regions. For example, in the example shown in Fig. 3, the parameter Q/N representing the engine load is determined from the intake air amount Q and the rotational speed N of the engine, and the operating state is divided into four modes based on this load Q/N and the rotational speed N. It is divided into areas. Operating region 0 is when the rotation speed N is 80 Q rps or less or the load Q/N is 0.
, 617 rev or less, and is a non-learning area in which knocking control is not performed. Driving area! , ■, and ■ are all medium-high load areas where the load Q/N is 0, 61/rev or more, and are learning areas in which knocking control is performed. Operating region I is when the rotational speed N is 800 rpm or higher2
Less than 000 rpm-, operating region ■ is when the rotation speed N is 2000 r.
pm or more and less than 4000 rpm, and the operating region (2) is a region where the rotational speed N is 4000 rpm or more. Here, we will consider the movement of the knock correction amount θ due to knocking control from when the vehicle enters one operating region within the learning region from another region until it leaves the other region again. Knocking control gradually increases or decreases the knock correction value θ so that the ignition timing reaches the knock limit, so as shown in Figure 4, knocking occurs when the knock correction value θ1 is too small when entering a certain operating range. If this occurs frequently, the knock correction value is gradually increased. On the other hand, as shown in FIG. The knock correction value θ approaches the correction value θ, which is the knock limit, and is therefore on the verge of leaving the area at the longest elapsed time in that area.
2. It is considered that θ is closest to the knock limit θ. Therefore, as shown in FIG. 2, each operating area 1. Learning value θ, (1) for II and ■, respectively. A means for storing θL(2), θL, and (3) is provided, and the knock correction value θ just before leaving the area i is set as the learned value θ+
, (1). The institution is
Driving area 0, which is a non-learning area. Driving area 1, which is a learning area. The engine is operated while changing between four operating ranges, It and (2), as needed. When operating in the operating range ■, ■, or Update and store the correction value θ8 as a learned value θL(i) representing the operating range,
Knock correction value θ8 when entering the relevant operation region i next time
It is intended to be used as the initial value of . Immediately after entering a new driving range, the knock correction value θ is:
Although it may still not be able to follow changes in the knock limit and indicate an inaccurate value, the knock correction value θ1 just before leaving the driving range is the correction value at the time when the time spent in that driving range is the longest. This is because tracking has been completed and has settled down, and it is considered that it shows the most stable and reliable value. Note that the control concept of the present invention described above is based on the control concept of the present invention.
JP-A-59-1 in terms of storing the knock correction amount of T.
Although it may seem similar to the control concept disclosed in Japanese Patent No. 36574, it is different for the following reasons. The learning retardation amount in the above-mentioned conventional technology is calculated based on the magnitude of the knocking correction retardation amount, which is increased or decreased depending on the presence or absence of knocking. Therefore, when moving from a knocking non-controlled area to a knocking controlled area,
If you start the knocking correction retard amount from 0, the learning retard amount will become small. To prevent this problem,
In the above-mentioned conventional technology, when the knocking non-control area is shifted to the knocking control area, the knock correction retardation amount is started at the time when the knocking correction area was shifted from the knocking non-control area to the knocking non-control area the previous time, in order to prevent the above-mentioned problem. It is. - The learning value in the power tree invention is used as the initial value of the knock correction amount in the learning area, so the learning value is stored not only when moving from the knocking non-control area to the knocking control area, but also when a certain learning This is also done when moving from one learning area to another. The difference is noticeable from driving area I to driving area ■ within the learning area.
This is the case when the transition occurs. In such a case, no learning is performed in the prior art, whereas in the control concept of the present invention, the learned value θL(1) of the operating region ■ is updated,
You can even get closer to the knock limit. FIG. 1 is a drawing clearly showing the configuration of the present invention. In order to realize the control concept described above, the present invention uses a knock sensor 7 that detects mechanical vibrations of the internal combustion engine, and an operating state detection means 61 that detects operating states such as engine rotation speed.
and a learning correction value storage means 62 that divides the operating state of the engine into a plurality of operating regions and stores a learning correction value θL(i) for each operating region, and a learning correction value θL(i) based on the learning correction value θ5(1). a knock correction value calculating means 63 for calculating a knock correction value θ, which is increased or decreased according to the presence or absence of knocking according to the signal from the knock sensor 17;
an ignition timing correction means 64 for correcting the ignition timing by;
When the operating state of the engine transits from one operating region to another, the revised knock correction value θ immediately before the transition,
There is provided a knocking control device for an internal combustion engine characterized by comprising a learning means 65 for storing the learned correction value θL(i) in the one operating region. "Operation" According to the above configuration, the opportunity to learn the learning correction value θL(i) is only at the time of transition between operating regions, so there is no need to perform the learning process for each ignition cycle as in the conventional case. On the other hand, the value learned as the learning correction value θL(i) is the knock correction value θ at the longest elapsed time in the relevant operating region, so the knock limit There is an extremely high probability that the accurate knock correction value θ that has completed tracking will be stored as the learned correction value θL(i).Therefore, it is possible to obtain the accurate learned value θL(i) with infrequent learning opportunities. This allows accurate and quick learning control to be performed while reducing processing time. "Embodiments" Examples of the present invention will be specifically described with reference to the drawings. FIG. 6 is an overall configuration diagram, where 1 is the engine body and 2 is an electronic control device. An electromagnetically actuated fuel injection valve 4 is disposed near each cylinder intake port of the intake manifold 3. Fuel adjusted to a constant pressure is fed under pressure to each fuel injection valve 4 from a fuel cylinder 1 (not shown), and the injection wall is controlled by the time during which the fuel injection valve 4 is operated and opened. As is well known, the distributor 5 that distributes power to the spark plugs installed in each cylinder rotates once every two revolutions of the crankshaft, and has a device inside that detects engine rotation and generates an output signal at every fixed crank angle. It includes a rotation angle sensor 7 and a reference angle sensor 6 that detects a specific position of the crankshaft. Ignition coil 8 connected to distributor 5
is constructed integrally with an igniter 9 that cuts off electricity to the ignition coil 8. In order to detect the operating state of the engine, an air flow meter 1 that detects the amount of intake air is installed in the intake passage communicating from the air cleaner 10 to the intake manifold 3.
2. An intake air temperature sensor 13 that detects the intake air temperature, and a throttle valve opening sensor 14 that detects the opening of the throttle valve 11 are installed. An 02 sensor 16 for detecting oxygen concentration is provided. Further, a knock sensor 17, which is a piezoelectric type vibration detector, is attached to the engine block and detects the vibration waveform of the engine. Of course, the knock sensor 17 may be of any type, such as an electrodynamic type, a piezoresistive type, or a strain gauge type, in addition to the piezoelectric type. FIG. 7 is a block diagram showing connections between the electronic control device 2 and each device. The reference angle sensor 6 and the rotation angle sensor 7 are electromagnetic pickup type detectors, and their respective output signals are shaped into pulses by waveform shaping circuits 21 and 22, and are sent to the microcomputer as a reference angle signal G and a rotation angle signal Ne, respectively. 23. Further, signals from the sensors 12 to 16 that detect the operating state are input to the digital input or analog input terminal of the microcomputer 23. The output signal of knock sensor 17 is input to amplitude detection circuit 24 . The amplitude detection circuit 24 detects a knocking-specific frequency component (for example, 8') from the output signal of the knock sensor 17.
A band-pass filter 25 that extracts only KH2), an amplifier 26 that amplifies its output, and a peak value of the vibration signal 26a that is the output of the amplifier 26 is held by, for example, a capacitor, and is reset by the cylinder switching signal R from the microcomputer 23. A peak hold circuit 27 is provided. The peak hold circuit 27 holds the peak value of the vibration signal 26a for each explosion stroke of each cylinder, and transmits the peak hold value 27a to the microcomputer 23 as the output voltage signal of the amplitude detection circuit 24. The microcomputer 23 receives rotation angle information from the reference angle sensor 6 and rotation angle sensor 7, and various operating state sensors 12.
The optimum ignition timing and energization time to the ignition coil for the best performance such as fuel efficiency and output are calculated based on the driving state information from 16 to 16 and the knocking information from the knock sensor 17, and an ignition signal is output. . And output buffer 28
The igniter 9 is driven through the ignition coil 8, and the ignition coil 8 is energized.
By cutting off the current at the calculated ignition timing, the high voltage generated when the current is cut off is guided to the spark plug 30 of a predetermined cylinder via the distributor 5, and each cylinder is ignited in sequence. Furthermore, the microcomputer 23 operates the injection valve drive circuit 29 to a predetermined crank angle position with a pulse width corresponding to the calculated fuel injection amount, and opens the fuel injection valve 4 for each 'A cylinder to inject fuel. Inject. FIG. 8 is a block diagram showing the internal configuration of the microcomputer 23. 8-bit configuration central processing unit (C
A read-only memory (ROM) 43 for storing control programs and constants necessary for calculations, a temporary memory (RAM) 44 for temporarily storing calculation data, and a CPU 41 are connected to the CPU 41 via a CPU bus 42. an interrupt control section 45 for controlling interrupts; a 16-bit timer 46 configured to be the basic cycle of CPU operation, and incremented by one every lock cycle; An A-D converter 48 for converting an analog signal into a digital signal, an input port 49 for the digital signal, and an output port 50 for outputting the digital signal are connected. Even when the ignition switch is turned off, battery power is supplied to a part of the time storage memory (RAM) 44 so that the stored contents are not erased. This portion is called the backup RAM area. The back-up RAM area is used as an area for storing the learning correction value θL(i) due to knocking. Signals from the air flow meter 12, the intake air temperature sensor 13, the coolant temperature sensor 15, and the peak hold circuit 27 are input to the multiplexer 47 to which analog signals are manually input. The input boat 49 includes a throttle valve opening sensor 1.
A 2-bit contact signal from 02 sensor 16 and a lean signal from 02 sensor 16 are input. The output boat 50 outputs an ignition signal to the igniter 9, an injection signal to the fuel injection valve 4, a cylinder switching signal to the peak hold circuit 27, and a control signal to the multiplexer 47.
The interrupt control unit 45 includes a reference angle sensor 6 and a rotation angle sensor 7.
A reference angle signal and a rotation angle signal are input. Based on the above hardware configuration, the processing performed by the microcomputer 23 that performs knocking control and controls the ignition timing will be described. FIG. 9 is a flowchart showing ignition timing control processing. When the internal combustion engine is started and an interrupt for ignition timing calculation is performed, an interrupt process 100 is started. The rotation speed N and the load Q/N (= intake air amount Q/rotation speed) are calculated in step 101. The rotation speed N is calculated from the generation interval time of the signal from the reference angle sensor 6, and is read directly by the signal from the air flow meter 12. The reference angle sensor 6, the air flow meter 12, and the processing in step 101 constitute the operating state detection means 61 of the present invention. In step 102, a pre-stored basic ignition timing map is searched and interpolated based on the rotational speed N and negative load/N calculated in step 101, and basic ignition timing θ@A! ! is calculated. In step 103, the intake air temperature sensor 13. Throttle valve opening sensor 14. The corrected ignition timing θ is based on information from the cooling water temperature sensor 15 and the like. is calculated. and,
In step 104, as shown in FIG. 10, a knock control process 200 is performed to calculate a knock correction value θ. When the knock control process 200 is started, step 201
Then, the current operating fII range is determined from the rotational speed N and the load Q/N. As shown in Figure 3, there are four operating regions, 0 and 1, consisting of 3 learning regions and 1 non-learning region.
．． It is divided into n and ■, and these operating areas o, r, n
It is determined which of the groups (1) and (2) it currently belongs to. In step 202, it is checked whether the jointly determined driving region is the same as the previously determined region. If they are the same, the learning process from step 203 onwards is omitted and the process jumps to step 207. On the other hand, if the previous and current driving ranges are different, the process moves to step 203 to perform the learning process since the driving range has changed. In step 203, the learning correction value θ1.corresponding to the previous driving range is determined. The value of (1-1) is updated to the current knock correction value θ, and is stored. Learning correction value θ. (1) is stored in a backup RAM area in the RAM 44 that does not disappear even if the ignition switch is turned off,
The knock correction value θ is stored in a normal area. Also, if the previous driving area was a non-learning area (area O),
No storage is performed because the corresponding memory address does not exist. In this way, in step 203, the learning correction value θL(i
) learning is executed. Next, in steps 204 to 206, the knock correction value θ
An initial value of 1 is set. That is, it is checked whether the current driving area is a learning area (area I, ■ or ■) (step 204), and if it is a learning area, the current driving area i is checked.
The learned correction value θL(i) stored corresponding to It is set to 0 (step 206). This completes the learning process, and from step 207 onwards, a feedback process is executed to detect knocking and vary the knock correction value θ1. In step 207, the current operating region is the region in which knocking control is performed, that is, in this embodiment, the learning region (region 1.
or ■). If it is the operating range 0 in which knocking control is not performed, nothing is executed, that is, the process ends with the knocking correction value θ remaining at 0. If it is an operating range in which knocking control is to be performed, the process advances to step 208 and a knocking determination is performed. In the Stella 7208, knocking is determined by averaging the peak hold values 27a of the vibration signals 26a from the knock sensor 17 in the past and comparing the magnitude of the current peak hold value 27a with the knocking determination level obtained by multiplying the peak hold value 27a by a predetermined value. Determine whether or not it has occurred. If it is determined that there is a knock, the process proceeds to step 209, where the value of the knock correction value θ, is increased by a predetermined amount Δθ to θ,=θ1+Δθ, and if it is determined that there is no knock, the value of the knock correction value θ, is increased in step 210. by a predetermined amount, θ, = θ, −Δ
Let it be θ. The knock correction value θ8 is set to 9. The knocking control process 200 is finished, and the process returns to step 105 shown in FIG. 9 again. In step 105, the basic ignition timing θBAIK is changed to the corrected ignition timing θ. and the final ignition timing θF+11AL corrected based on the knock correction value θ, is θjllll^L=θB
AI+θ. −θ, is calculated. Then, in step 106, the calculated time until the final ignition timing θFIIIAL is set in a predetermined output conveyor register, and when the set time elapses, the energization to the ignition coil 8 is cut off, and ignition is performed. "Effects of the Invention" As explained above, the present invention has the above configuration and learns and stores the knock correction value itself just before transitioning to another driving range as a learning correction value, so that simple learning is possible. Accurate learning values can be learned using relatively infrequent learning opportunities, and the load on system processing time and memory capacity can be significantly reduced, which is an excellent effect.

[Brief explanation of the drawing]

The drawings show one embodiment of the present invention, FIG. 1 is a corresponding diagram clearly showing the configuration of the present invention, FIG. 211 is a graph diagram explaining the control concept, FIG. 3 is a graph diagram showing the operating range, and FIG. and Fig. 5 is a schematic diagram showing the behavior of the knock correction value, Fig. 6 is an overall configuration diagram of the device, Fig. 7 is a block diagram, and Fig. 8 is a detailed diagram.
FIGS. 9 and 10 are block diagrams showing the m configuration, and flowcharts showing processing in the microcomputer. 110. Engine body, 211. electronic control device, 50
0. Distributor, 690. Reference angle sensor, 7
860 rotation angle sensor, 8182 ignition coil, 12°1. Air flow meter, 17, Knock sensor, 23, Microcomputer, 24, G
Width detection circuit, 27, peak hold times i? 8゜Figure 2Figure 3rpm Engine speedFigure 4Elapsed timeFigure 5Elapsed timeFigure 9

Claims (1)

[Scope of Claims] A knock sensor that detects mechanical vibrations of an internal combustion engine, an operating state detection means that detects operating states such as engine rotational speed, and a system that divides the operating state of the engine into a plurality of operating regions and a learning correction value storage means for storing a learning correction value for each driving region; and a knock calculating means for calculating a knock correction value that is increased or decreased depending on whether or not knocking occurs in accordance with a signal from the knock sensor based on the learning correction value. In a knocking control device comprising a correction value calculation means and an ignition timing correction means for correcting ignition timing using the knock correction value, when the operating state of the engine transitions from one operating range to another operating range, A knocking control device for an internal combustion engine, comprising learning means for storing the knock correction value immediately before transition as a learning correction value in the one operating region.