JP3287984B2 - Idling control system for internal combustion engine - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a fuel injection control type gasoline engine for an internal combustion engine, and more particularly to a control device for controlling idling by applying a neural network (hereinafter abbreviated as "NN"). .

[0002]

2. Description of the Related Art Conventionally, in idling control, a flow path of air that bypasses a throttle valve is provided, and a linear solenoid valve whose opening time varies according to a command current value is attached to the flow path to control an air flow rate. There is a method of changing the suction negative pressure to perform optimal fuel injection. In this method, PID feedback control is performed on a command current value so as to reach a target rotation speed determined by a group of data set in advance according to a load condition of an engine such as a water temperature or an air conditioner on / off, and the command current value is controlled. The air flow is controlled by changing the opening time of the linear solenoid valve according to the value, and the optimal fuel injection is performed by changing the suction negative pressure.

As a measure against exhaust gas regulations, idling is a useless rotation of the engine as its name implies, so the number of revolutions is reduced to the limit. If it is lowered, the engine speed will fluctuate greatly due to engine load fluctuations that do not pose a problem during cruising, for example, turning on / off the air conditioner switch, turning on / off the light switch, etc. Resulting in. Although in order to avoid such as this, it is necessary to finely control the current value to command the solenoid valve on the anticipation of the engine speed fluctuation amount when given various loads, depending on various loads optimal to set the a parameter, it is necessary a lot of time.

[0004]

In such a conventional approach to the idling control device for an internal combustion engine, a target idling speed must first be determined. Further, the data of the target rotational speed is determined in advance by experiments, simulations, and the like, and it is necessary to change the data depending on operating conditions and the like. In order to perform current control with high accuracy, experiments and simulations must be repeated to increase the amount of data.

[0005] Other than the on / off of the air conditioner switch, the factors that cause the idling speed to fluctuate are: (a) switching of the position of the automatic transmission (neutral or drive); on/
OFF) (c) Change in steering angle (operation or non-operation of power steering) (d) State of air conditioner clutch (connection / disconnection) (e) Operation state of AC generator (switching charging mode) In any case, it is not easy to describe the command current in a simple linear form with a complicated structure, and it is necessary to have a large number of data.

The present invention has been made in view of the above-mentioned conventional problems, and an object of the present invention is to provide an idling control device for an internal combustion engine that performs a simple and highly accurate idling speed control using a neural network. is there.

[0007]

SUMMARY OF THE INVENTION In order to achieve the above object, the present invention provides an idling control device for an internal combustion engine, in which a neural network learns a nonlinear structure of an actual current value of a linear solenoid valve related to an idling target rotation speed. Also, by using the previously learned neural network to calculate an estimated current value for maintaining idling at the target rotation speed, and issuing a command to the linear solenoid valve based on the estimated current value, The gist is that the idling is controlled to the target rotation speed.

With this configuration, according to the present invention, the idling control can be performed with high accuracy even when the load state of the engine fluctuates variously without setting complicated parameters.

[0009]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An invention according to claim 1 of the present invention comprises: a sensor group for detecting an operation state of an internal combustion engine; an accessory device state detecting means for detecting a state of an accessory device which is a load on the internal combustion engine; The opening time changes according to the actual current value.
Therefore, by controlling the air flow rate and changing the suction load, the internal combustion engine
A linear solenoid valve for controlling the fuel injection amount to the gin, the output of the and the actual current value detecting means for detecting the actual current value of the linear solenoid valves, each of the detection value and the accessory equipment state detecting means of the sensor group and from the actual current value, a neuro-operation means the current value predicted estimate of the linear solenoid valve required to maintain the target speed calculated by the neural network computation, the current value predicted estimated value to the linear solenoid valve and a current command means for commanding, the current command current value is controlled to <br/> target rotation amount the fuel injection by varying the air flow drives the linear solenoid valve according to from the command means When the idling control is performed so as to obtain the numerical value, the element of the neuro operation is made more precise and the estimated current value is obtained. Cormorant having an effect.

[0010] The invention according to claim 2 of the present invention is the invention of claim 1, wherein the actual current value detecting means detects the actual current value of the linear solenoid valve in the air flow rate change operation, the detection obtains a current value predicted estimate the actual current value as an operation element in the neural operation unit, before
The estimated current value is fed back as an operation control signal for the linear solenoid valve, so that the neuro operation is more accurately performed.

[0011] The invention described in claim 3 of the present invention includes a sensor group for detecting the operating state of the internal combustion engine, and accessory equipment state detecting means for detecting the state of the attachment device as a load to the internal combustion engine, the actual current The opening time changes according to the value.
Control the air flow and change the suction load to
Solenoid valve that controls the amount of fuel injected into the
When, from the output of the detection values of the sensors and the attachment device state detection means, a current value predicted estimate of the re <br/> near solenoid valve required to maintain the target speed in neural network calculations a neuro-operation means for calculating, and a current command means for commanding said current predicted estimated value to the linear solenoid valve, and drives the linear solenoid valve according to a command current value from the current command means the <br / > The fuel injection amount is controlled by changing the air flow rate.
The idling control is performed so as to control the engine speed to the target rotational speed, and the idling control of the engine is performed with a simple configuration.

According to a fourth aspect of the present invention, in the first or the third aspect of the present invention, the neuro-calculating means comprises:
The neural network that outputs a predicted estimated value of the current value of the linear solenoid valve required to maintain the target rotational speed is output in real time, and the neuro operation is performed at a higher speed.

An embodiment of the present invention will be described below with reference to the drawings.

(Embodiment 1) FIG. 1 is a block diagram showing a structure of an internal combustion engine idling control device according to a first embodiment of the present invention.
In FIG. 1, reference numeral 112 denotes an internal combustion engine (hereinafter, simply referred to as an engine), 100 denotes a group of sensors for detecting an operation state of the engine 112, and 17 denotes a state of an accessory device such as an air conditioner, which is a load on the engine 112. The attached device state detecting means 20 changes the opening time according to the actual current value.
To control the air flow and change the suction load
A linear solenoid valve for controlling the amount of fuel injected into the engine 112 ; 21 is an actual current value detecting means for detecting an actual current value of the linear solenoid valve 20;
The data storage means 110 stores the detected value of 0 and the detection data from the attached device state detecting means 17. Each of the detected values of the sensor group 100 stored in the data storing means 50 and the output and actual values of the attached device state detecting means 17 are provided. Current value detection 2
A neural network means for performing a neural network operation which outputs a predicted estimated value of the current value of the linear solenoid valve 20 based on the actual current value of 1; Current command means for issuing a command to the linear solenoid valve 20 based on the current command.

As shown in FIG. 2, for example, as shown in FIG. 2, the sensor group 100 for detecting the operating state of the engine 112 includes a rotational speed detecting sensor 10 for detecting an engine rotational speed, an intake pressure sensor 11 for detecting an intake pipe internal pressure, and the like. a throttle opening sensor 12 for detecting a throttle opening, a water temperature sensor 13 for detecting a coolant temperature, an intake air temperature sensor 14 for detecting an intake air temperature, an atmospheric pressure sensor 15 for detecting the atmospheric pressure, f Tsu Doraito etc. And a current consumption sensor 16 for detecting the current consumption of the electrical component.

The actual current value detection means 21 drives the linear solenoid valve 20 in accordance with the command current value generated by the current command means 111 , and performs idling control of the engine 112 by changing the air flow rate. The actual current value of the linear solenoid valve 20 is detected. The data storage means 50 stores each detection data obtained by the sensor group 100, the detected value of the actual current value detection means 21, and the output of the accessory device state detection means 17. In addition, neuro operation means 1
Reference numeral 10 designates each of the detected values of the sensor group 100 stored in the data storage means 50, the output of the accessory device state detecting means 17 and the real current value which is the output from the real current value detecting means 21 as inputs. The operation of the neural network that outputs a predicted estimated value of the current value of the linear solenoid valve 20 necessary for maintaining the idling rotational speed (that is, the target rotational speed) is performed in real time.

The operation of the idling control device having such a configuration will be described below. FIG. 3 is a flowchart illustrating a processing operation procedure of idling control in the above embodiment. This process is performed every engine rotation step (18
One cycle is performed at every 0 degree crank angle).

The the first starting of the engine 112 is performed, in process step (hereinafter simply referred to as steps) 4 1, the actual current value obtained by the detection output value and the actual current value detection unit 21 of the various sensors 100, and accessories The detection output of the device state detection means 17 is stored in the data storage means 50. Next, in step 42, data conversion, that is, normalization of each stored value is performed so as to become an input value to the input layer of the neural network, and in step 43, a neuro operation is performed using the estimated current value as an output. Next, in step 44, the estimated current value is used as the final control value, and a command is sent to the linear solenoid valve 20. Next, at step 45, it is checked whether or not the engine 112 is stopped. If the engine 112 is not stopped, the process returns to step 41 to continue the idling control. The control processing operation ends.

FIG. 4 shows an input layer in a four-cylinder engine.
Fig. 3 shows a neural configuration provided by three layers, an intermediate layer and an output layer. Here, k is a value updated at every engine rotation step (180 degree crank angle), Pb is the pressure in the intake pipe, ne is the rotation speed, θ is the throttle opening, Tw is the coolant temperature, and Ta is the intake air. Temperature, Pa is atmospheric pressure, EL is current consumption, Load is the state of attached equipment, and Iact is the actual current value of the linear solenoid valve 20.

Further, a neuro construction for outputting a current value is constructed by the neuro operation means 110, and the result is sent to the linear solenoid valve 20 via the current command means 111.
, It is possible to calculate the optimum current value Icmd according to the operation state so as to reach the target rotation speed.

(Embodiment 2) FIGS. 5 and 6 are block diagrams of an idling control device for an internal combustion engine according to a second embodiment of the present invention. In these figures, FIG. 5 shows a schematic configuration of the second embodiment, and FIG. 6 shows the configuration of the sensor group in FIG. 5 in more detail. In FIG. 5, reference numeral 112 denotes an internal combustion engine, 100 denotes a sensor group for detecting an operation state of the internal combustion engine 112, 17 denotes an accessory device state detecting means for detecting a status of an accessory device such as an air conditioner, which is a load on the engine 112, 20 changes the opening time according to the actual current value
To control the air flow and change the suction load
50 is a linear solenoid valve for controlling the amount of fuel injected into the engine 112 , 50 is data storage means for storing detection data obtained by the sensor group 100 and data from the auxiliary equipment state detection means 17, 110 is data storage means 5
A neuro operation means for performing an operation of a neural network which outputs a predicted estimated value of the current value of the linear solenoid valve 20 based on each detected value of the sensor group 100 stored in 0 and the output of the attached device state detecting means 17;
Reference numeral 1 denotes current command means for issuing a command to the linear solenoid valve 20 based on the estimated current value estimated by the neuro operation means 110.

As shown in FIG. 6, for example, a sensor group 100 for detecting the operating state of the engine 112 includes a rotational speed detecting sensor 10 for detecting an engine rotational speed, an intake pressure sensor 11 for detecting an intake pipe internal pressure, and the like. a throttle opening sensor 12 for detecting a throttle opening, a water temperature sensor 13 for detecting a coolant temperature, an intake air temperature sensor 14 for detecting an intake air temperature, an atmospheric pressure sensor 15 for detecting the atmospheric pressure, f Tsu Doraito etc. And a current consumption sensor 16 for detecting the current consumption of the electrical component.

The data storage unit 50 stores the output of the auxiliary device condition detecting means 17 and contact each detection data obtained by the sensor group 100. In addition, neuro operation means 1
Reference numeral 10 denotes each detected value of the sensor group 100 stored in the data storage unit 50 and the output of the attached device state detecting unit 17.
A neural network is operated in real time, in which a force is input and a predicted estimated value of a current value of the linear solenoid valve 20 necessary for maintaining a set idling rotational speed (that is, a target rotational speed) is output.

The operation of the idling control device according to the second embodiment having such a configuration will be described below. FIG. 7 is a flowchart for explaining a processing operation procedure of idling control in the second embodiment. This process is performed every engine rotation step (each 180 degree crank angle).
In one round.

First, when the engine 112 is started, in step 31, the detected output values of the various sensor groups 100 and the detected outputs of the attached device state detecting means 17 are stored in the data storage means 50. Next, in step 32, data conversion, that is, normalization of each stored value is performed so as to become an input value to the input layer of the neural network, and in step 33, a neuro operation is performed using the estimated current value as an output. Next, in step 34, the estimated current value is used as the final control value, and a command is issued to the linear solenoid valve 20. Next, in step 35, it is checked whether or not the engine 112 is stopped. If not, the process returns to step 31 to continue the idling control, while if the engine 112 is stopped, idling is performed based on the engine stop. The control processing operation ends.

[0026]

As described above, according to the present invention,
Each detection value of the sensor group that detects the operating state of the internal combustion engine
And the state of accessories that load the internal combustion engine
Output of the accessory device status detection means, and if necessary,
Actual current value detection that detects the actual current value of the near solenoid valve
Necessary to maintain the target speed based on the output of the
The predicted value of the current value of the linear solenoid valve is
Of the neural network by the
In real time, and the estimated current value
Because the solenoid valve is commanded, complicated
Engine without parameter setting
Control even when the load condition of the machine fluctuates in various ways
It can be performed with high precision.

[Brief description of the drawings]

FIG. 1 is a block diagram schematically showing a configuration of an idling control device according to a first embodiment of the present invention.

FIG. 2 is a block diagram more specifically showing a sensor group in the first embodiment.

FIG. 3 is a flowchart illustrating a processing operation procedure of idling control according to the first embodiment.

FIG. 4 is a diagram showing a neuro structure provided by three layers of an input layer, a middle layer, and an output layer in the four-cylinder engine used in the first embodiment.

FIG. 5 is a block diagram schematically showing a configuration of an idling control device according to a second embodiment of the present invention;

FIG. 6 is a block diagram more specifically showing a sensor group in the second embodiment.

FIG. 7 is a flowchart illustrating a processing operation procedure of idling control according to the second embodiment.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 10 Rotation speed detection sensor 11 Intake pressure sensor 12 Throttle opening sensor 13 Water temperature sensor 14 Intake temperature sensor 15 Atmospheric pressure sensor 16 Current consumption sensor 17 Attached equipment state detection means 20 Linear solenoid valve 21 Actual current value detection means 50 Data storage means 100 Sensor group 110 Neuro calculation means 111 Current command means 112 (internal combustion) engine

────────────────────────────────────────────────── ─── of the front page continued (51) Int.Cl. ⁷ identification mark FI F02D 45/00 314 F02D 45/00 314S 340 340C 360 360Z 390 390Z (72) inventor Ishida Akira, Osaka Prefecture Kadoma Oaza Kadoma 1006 address Matsushita Electric Industrial Co., Ltd. (56) References JP-A-4-231647 (JP, A) JP-A-3-96636 (JP, A) (58) Fields investigated (Int. Cl. ⁷ , DB name) F02D 41/16 F02D 9/02 F02D 9/02 305 F02D 41/20 315 F02D 45/00 314 F02D 45/00 340 F02D 45/00 360 F02D 45/00 390

Claims (4)

(57) [Claims] 1. A and sensor group for detecting the operating state of the internal combustion engine, and accessory equipment state detecting means for detecting the state of the attachment device as a load to the internal combustion engine, opened according to the actual current value
The air flow rate is controlled by changing
A linear solenoid valve for controlling the fuel injection amount to the internal combustion engine by changing the linear solenoid valve;
And the actual current value detecting means for detecting the actual current value, required from the output and the actual current value of the detection value and the accessory equipment state detecting means of the sensors, to maintain the target speed
Wherein comprising a neural operation unit for calculating a current value prediction estimate of the linear solenoid valve in the neural network calculation, and a current command means for commanding said current predicted estimated value to the linear solenoid valve command from said current command means the combustion by varying the air flow drives the linear solenoid valve according to the current value
An idling control device for an internal combustion engine, wherein the idling control is performed to control a fuel injection amount to a target rotation speed. Wherein said actual current value detecting means detects the actual current value of the linear solenoid valve in the air flow rate change operation, the current value predicted estimate the actual current value the detected as an operation element in the neuro-operation means The current
2. The idling control device for an internal combustion engine according to claim 1, wherein the estimated value is fed back as an operation control signal of the linear solenoid valve. 3. A sensor group for detecting an operation state of the internal combustion engine, an auxiliary device state detecting means for detecting a state of an auxiliary device serving as a load on the internal combustion engine, and an open state according to an actual current value.
The air flow rate is controlled by changing
Lini to control fuel injection amount to internal combustion engine by changing
And A solenoid valve, from the output of the detection values of the sensors and the attachment device state detection means, calculates a current value predicted estimate of the linear solenoid valve required to maintain the target speed in neural network calculations a neuro-operation unit, and a current command means for commanding said current predicted estimate the linear solenoid valve, changing the air flow drives the linear solenoid valve in accordance with a command current value from the current command means By
An idling control device for an internal combustion engine, wherein the idling control is performed so that the fuel injection amount is controlled to a target rotation speed. 4. A neural network according to claim 1, wherein said neural operation means performs a real-time operation of a neural network which outputs a predicted estimated value of a current value of said linear solenoid valve necessary for maintaining said target rotational speed. The idling control device for an internal combustion engine according to claim 1 or 3.