JPS6345438A - Engine controller - Google Patents

Abstract

PURPOSE:To improve transient response, by determining a feedback control quantity based on a state variable estimated from a detected output torque, an exhaust component density and a feedback control quantity, a gain being set on the basis of a dynamic model, an accumulated value of torque deviation and an accumulated value of density deviation. CONSTITUTION:A state variable quantity estimating means M7 employs parameters predetermined on the basis of a dynamic model of an engine for estimating a dynamic state variable quantity of the engine based on the values detected through a torque detection means M2 such as cylinder inner pressure sensor and an exhaust gas component density detection means M3 such as a lean sensor and a feedback control quantity. An accumulating means M8 accumulates deviation between the values detected through the torque detection means M2 and the density detection means M3 and target values. A feedback control quantity determining means M9 determines a feedback control quantity based on the outputs from the state variable quantity estimating means M7 and the accumulating means M8 and a feedback gain predetermined on the basis of a dynamic model of the engine thus controlling the engine through a control input quantity variable means M5.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an engine control device, and more particularly to an engine control device that suitably controls the output torque and exhaust composition of an engine.

[Prior Art] The amount of intake air, which is one of the basic control inputs for an engine, has traditionally been determined by the opening of a throttle valve linked to the accelerator, but in recent years, the throttle valve has been driven by an actuator. An engine control device has been proposed that determines the intake air amount of the engine by comprehensively considering various conditions related to engine operation (e.g., ``Engine Control Device'' in Japanese Patent Application Laid-Open No. 111046/1983). ). These control devices first determine the fuel injection amount based on the accelerator operation amount, engine load, etc., then determine the optimal intake air amount based on the air-fuel ratio or fuel consumption, etc., and then control the throttle valve opening. are doing.

[Problems to be Solved by the Invention] However, these engine control devices still have the following problems, and further improvements have been desired.

(1) An engine control device that controls the throttle valve opening according to the fuel injection amount can control the throttle valve opening well in a steady state, that is, when the engine is operating in a constant state, but when the fuel injection amount changes suddenly. In the so-called transient state, there is a problem in that the fuel injection amount changes in advance and the air-fuel ratio fluctuates greatly. That is, during acceleration, the air-fuel ratio decreases, producing a so-called rich spike, and during deceleration, a so-called lean spike occurs, and such fluctuations in the air-fuel ratio not only increase harmful components in the exhaust from the engine,
The problem is that if this variation occurs beyond a predetermined range, drivability may deteriorate. Conventionally, in order to suppress such rich or lean spikes within a certain range, it has been necessary to perform various adaptations such as asynchronous fuel injection, which has been a bottleneck in the development of engine control devices.

(2) Furthermore, in order to improve the above-mentioned transient characteristics, there is a problem in that an actuator that can be driven at high speed is required. For example, FIG. 15 shows an engine control III system that determines the fuel injection amount and then determines the throttle valve opening accordingly, with an air-fuel ratio of 19.
This is a graph that measures the change in the air-fuel ratio when the fuel injection amount is increased in response to an acceleration command from the lean burn state of 8. As shown in the figure, when the throttle valve is
g] It can be seen that if an actuator that takes 4 Q Q m5ec to move is used, the air-fuel ratio will be greatly disturbed.

At this time, the composition of the exhaust gas changes significantly, making it impractical from the viewpoint of exhaust gas purification. Therefore, at least 10
It is necessary to use an actuator with characteristics below 0 [m5ec/dec+], but such high-speed actuators require large output torque, which increases the size and size of the device.
There was a problem that JE< caused high costs.

(3) Since conventional engine control devices can only control the air-fuel ratio, there is a problem in that the output torque of the engine cannot be controlled to a required value during a transient period. Conventionally, unexpected fluctuations in the engine's output torque during transient periods are absorbed by the inertial weight (inertia) of the flywheel fixed to the engine's output shaft. However, providing a flywheel that is heavy enough to handle fluctuations in output torque will have a negative effect on engine responsiveness and fuel efficiency.

Therefore, the present invention can control the exhaust state to an appropriate state without using a highly responsive actuator, and
The object of this invention is to provide an engine control device that can control the output torque as well as the state of the exhaust gas of the engine, and that solves all of the above-mentioned problems.

Structure of the Invention [Means for Solving the Problems] In order to achieve the above object, the present invention has the following structure as a means for solving the problems. That is, as illustrated in FIG. 1, torque detection means M detecting the output torque of the engine M1.
2, a concentration detection means M3 for detecting a concentration of a predetermined component reflecting the exhaust composition of the engine M1, and a target torque of the engine M1 and a target concentration of the exhaust component based on a request for operation of the engine M1. target value output means M4 for determining and outputting a target value; control input amount variable means M5 for varying various quantities including at least the intake air amount and fuel injection amount of the engine M1 as control inputs for the engine M1; and the torque detection Feedback of various control input quantities of the engine M1 in which the output torque of the engine M1 detected by the means M2 is the target torque, and the concentration of the predetermined component detected by the concentration detection means M3 is the target fI degree. An engine control device comprising: a control means M6 for determining a control amount and controlling the control input variable means M5; controlling the output torque of the engine M1 and the concentration of the predetermined component detected by the torque detection means M2 and the concentration detection means M3, and the control input amount variable means M5 using predetermined parameters based on the From the feedback control amount used for the engine M
state variable amount estimating means M7 for estimating a state variable amount of a predetermined order representing a dynamic internal state of engine M1; and a torque that is the cumulative deviation between the detected output torque of engine M1 and the set target torque. an accumulating means M8 for calculating a cumulative deviation value and a cumulative concentration deviation value obtained by accumulating the deviation between the detected concentration of the predetermined component and the set target concentration; Based on the set feedback gain, the state variable quantity determined in the 11F, and the determined torque deviation cumulative value and concentration deviation cumulative value, control input quantities including at least the intake air amount and fuel injection of the engine M1 are determined. Feedback control amount determining means M9 for determining the feedback control amount
This is the configuration of an engine control device comprising:

Here, the torque detection means M2 is a means for detecting the output torque of the engine M1, and may be a means for calculating the torque from the torsion of the output shaft, for example, or may be a means for detecting the cylinder pressure in each cylinder of the engine and outputting from it. It may also be a method for determining torque.

The concentration detection means M3 is a means for detecting the concentration of a predetermined component that reflects the exhaust composition of the engine M1, and detects the concentration of oxygen, carbon monoxide, or nitrogen oxide (NOx) in the exhaust.
Any configuration may be used as long as it can detect the concentration, etc. of > within the air-fuel ratio range applied to the engine M1. For example, the 02 sensor using zirconia is well known as a sensor for detecting oxygen concentration, etc., and the sensor using a laser is well known as a sensor for detecting the NOx level.

The target value setting means M4 determines the target torque and target concentration of the engine M1 based on a request for the operation of the engine M1, such as the amount of depression of the accelerator pedal, the number of revolutions of the engine, or the amount requested from the automatic transmission. It is a means, and may be configured integrally with the control means M6. Note that the target value output means M4 may be configured to set the target torque and the target concentration separately, but may compensate for the target concentration based on the deviation between the target torque and the output torque of the engine M1 or the target torque itself. A concentration compensation value calculation unit is provided to calculate the value,
Furthermore, the configuration may include a target density setting section that compensates the target density using this compensation value.

In this case, a certain correlation can be established between the target torque and the target concentration, and the target concentration of a predetermined component in the exhaust gas, such as oxygen, can be varied during a transient period when the output torque is controlled to a predetermined value. Therefore, the output torque control can be completed in a shorter period of time.

The control input amount variable means M5 is a means for varying various m including at least the intake air amount and the fuel amount 1; Possible configurations include a fuel injection valve, a fuel injection pump, and the like. Further, various types of throttle actuators can be used, such as a forward/reverse reversible DC or AC motor, a stepping motor, or an electromagnetic solenoid.

Note that other quantities of control input for the engine M1 include the ignition timing of the engine M1, the amount of exhaust gas recirculation, and the amount of supercharging.

The idle intake air amount, etc. can be considered, but if these quantities are also controlled as control inputs for the engine M1, various means for varying these quantities, such as switching for opening and closing of the ignition coil-next circuit, are required. element, EGR valve,
wastegate valve.

The ISO valve and the like also correspond to the control input variable means M5.

The control means M6 includes state variable quantity estimation means M7. Accumulation means M
8 and a feedback control amount determination means M9, the engine is configured to match the output torque of the engine M1 and the concentration of a predetermined component in the exhaust gas to the target torque and target 'a degrees set by the target value output means M4. M1
The control input amount variable means M5 is controlled by determining the feedback control 1ffi of various control human input quantities. This control means uses so-called modern control theory, is configured as an additional integral type optimal regulator with at least two manpowers and two outputs, and is capable of independently and optimally controlling the output torque and the concentration of the predetermined component.

A method for configuring an additive integral type optimal regulator is, for example, "Linear System Control Theory" by Katsuhisa Furuta (1978).
Since Shokodo et al. are well-versed in the matter, a detailed explanation will be omitted here, but when configuring such a control means M6, first,
It is necessary to know the dynamic model of the controlled object. Normally, for a complex object such as engine M1, it is difficult to theoretically accurately obtain a dynamic model that describes the behavior of the output torque and a predetermined component in the exhaust gas, and so-called system identification Modeling is done by

The details of system identification can be found in Setsuo Sagara's ``System Identification'' (1981), the Society of Correction and Automatic Control Engineers, etc., so a detailed explanation will be omitted here. A widely known method is to obtain the transfer function of a control system using multiplication, online identification, etc., and then construct a dynamic model from it. When creating a model, it is possible to perform control over a wide range of operating ranges by linear approximation by performing modeling in several ranges that can be considered to have linearity around a steady point.

The state variable estimation means M7 provided in the control means M6 is a so-called observer, and
A state variable quantity X of a predetermined order representing a dynamic internal state of 1
(k).

The state variable ff1X(k) includes various quantities that are directly involved in the operation of the engine M1, such as the amount of air actually being taken in, the dynamic behavior of combustion, or the amount of air in the air-fuel mixture that is involved in combustion. You can consider a situation such as the amount of fuel.

The higher the order of the state variable i (k), the more theoretically it contributes to improving control elimination, but if the order of the state variable becomes too large, it actually takes a lot of calculation time, so Control accuracy decreases. Therefore, the condition change is ¥1ff.
The order of iX(k) may be determined depending on the capability of the hardware constituting the state variable amount estimating means M7 and the required control accuracy, and may be, for example, about 4th to 6th order.

The state variable amount estimating means M7 has various configurations and its design method is <%D.

These are detailed, for example, in ``Mechanical System Control'' by Katsuhisa Furuta et al. (1981) published by Ohm Publishing, etc., and are explained in detail in ``Mechanical System Control'' by Katsuhisa Furuta et al. It can be designed as a finitely settled observer.

The accumulation unit is a means for accumulating the deviation between the target torque and the output torque, and the deviation between the target concentration and the actual concentration of a predetermined component in the exhaust gas, respectively.

The target concentration is determined based on the requirements for the operation of the engine M1. In other words, in the present invention, the engine control device is configured as a servo system. In other words, in general, servo system control requires control that eliminates the steady-state deviation between the target value and the actual control value, and this is achieved by 1/SQ (Ω-order integral) in the transfer function.
It is considered necessary to include the following. Furthermore, in the case where the transfer function of the system is determined by system identification as described above and an equation of the form 5 is established from this, it is desirable to include such an integral quantity from the viewpoint of stability against noise. In the present invention, it is sufficient to consider Ω=1, that is, first-order integral.

Therefore, if we expand the system by adding this cumulative value to the state variable quantity X(k) mentioned above, and determine the feedback amount by both of them and the predetermined optimal feedback gain F, we can obtain Appendix 7J [+ integral type optimal As a regulator, various control input quantities to the engine M1, which is a controlled object, are determined.

The feedback control amount determining means M9 that determines the feedback value (feedback control amount) determines various human power amounts for controlling the engine M1 from the optimal feedback gain determined in advance and the above-mentioned state variables and each cumulative value. Here, the optimal feedback gain can be determined in advance by simulation using an evaluation function.

[Operation] The engine control device of the present invention having the above configuration uses at least the intake air amount and fuel injection amount of the engine M1, which are varied by the control input variable means M5, as control input amounts;
A system with at least two human power outputs whose control outputs are the output torque of the engine M1 detected by the torque detection means M2 and the concentration of a predetermined component in the exhaust gas detected by the concentration detection means M3 is configured as an additive integral type optimal regulator. The control means M6 performs feedback control to make the output torque of the engine M1 and the state of the exhaust gas coincide with each target value set by the target value output means M4.

[Example] Next, an example of the present invention will be described in detail based on the drawings. Fig. 2 is a schematic diagram showing the engine and its peripheral equipment in an embodiment of the present invention, Fig. 3 is a graph illustrating a method for measuring the output torque of the engine, and Fig. 4 shows the operating state of the engine. 5 and 6 are block diagrams used to explain system identification, respectively. FIG. 7 is a flowchart showing an example of control executed in the electronic control unit, and FIG. FIG. 8 is a graph explaining the effects of this example.

The explanation will be given below in this order.

Although Fig. 2 mainly shows one cylinder of a 4-cylinder, 4-cycle engine 1, the intake system 2 includes an air cleaner (not shown) from upstream, an intake temperature sensor 5 that detects the intake air temperature Tha, and controls the amount of intake air. A throttle valve 7, a surge tank 9, an electromagnetic fuel injection valve 11, and the like are provided. Further, the exhaust gas from the engine 1 is discharged to the outside through an exhaust pipe 14 via an exhaust purification device, a muffler, etc. (not shown).

The combustion chamber (cylinder) is a piston 15. It is composed of an intake valve 17, an exhaust valve 192, a spark plug 21, etc., but since their operations are well known, their explanation will be omitted. Incidentally, a spark plug 2 generates a spark for ignition using a high voltage supplied from an igniter 24 via a distributor 25.
1 has a semiconductor type pressure sensor 27 built in,
It is configured to detect the cylinder pressure P(i). This cylinder pressure P(i) is used to determine the output torque of the engine 1, which will be described later.

In addition, the engine 1 includes a cooling water temperature sensor 29 that detects the temperature T i w of the cooling water, and a rotation speed sensor 31 that is provided in the distributor 25 and outputs a pulse signal with a frequency corresponding to the rotation speed N of the engine 1. , engine 1 no 1
A cylinder discrimination sensor 33 outputs one pulse signal per rotation (7200 degrees of crank angle) and a lean sensor 34 is installed in the exhaust pipe 14 and continuously detects the oxygen concentration in the exhaust gas.
etc. are provided. In addition, the throttle valve 7 has
A throttle actuator 35 is attached to drive this, and the opening degree θ of the throttle valve 7, that is, the intake air amount, is controlled by this throttle actuator 35. Therefore, the driver's depression of the accelerator 38 detected by the accelerator sensor 37 ff1Acc is:
Once input to the electronic control device 4O, the electronic control device 40
This configuration is reflected in the opening degree θ of the throttle valve 7.

Next, the configuration and function of the electronic control unit @40 will be explained. The electronic control unit 40 is operated by receiving power from a battery 43 via a key switch 41, and is powered by a well-known microprocessor (MPU) 44. ROM45.
RAM46. Backup RAM47. input boat 49
．． It is composed of an output boat 51 and the like, and each of the above-mentioned elements and boats are connected to each other by a bus 53.

The input port 49 of the electronic control device 40 receives signals indicating the operating state of the engine 1 from each sensor. Specifically, the accelerator opening ACC is measured by the accelerator sensor 37, and the oxygen concentration EO in the exhaust gas of the engine 1 is measured by the lean sensor 34.
Then, after inputting the intake air temperature 1-ha from the intake air temperature sensor 5, the output torque T from the pressure sensor 27, and the cooling water temperature ThW from the cooling water temperature sensor 29 and A/D converting them, the MP
The input port 49 is connected to an analog input section (not shown) which is delivered as data to U44, and a pulse input section (not shown) which inputs the rotational speed N of the engine 1 from the rotational speed sensor 31 and the cylinder discrimination signal from the cylinder discrimination sensor 33, respectively. It is configured.

On the other hand, the output port 51 controls the opening degree θ of the throttle valve 7 via the actuator 35 and the opening degree θ of the fuel injection valve 11.
It closes the valve and outputs signals that control the fuel injection FRs and the ignition timing via the igniter 24, respectively. These electronic controls (i!'11 control by the MPU 44 of the l device 40, except for the ignition timing control) will be described in detail later with reference to the flowchart of FIG. 3.

Before explaining the functions realized by the internal software of the electronic control device 40, a method for calculating the output torque using the pressure sensor 27 will be briefly explained.

In a four-stroke engine that repeats the suction, compression, explosion, and exhaust strokes, the difference in work between the suction and explosion strokes, which produce effective work We, and the compression and exhaust strokes, which produce loss work WQ<We-WQ) is expressed as the output torque of the engine, and this output torque is expressed as the stroke volume ■ and the mean effective pressure P
It is known that it is equal to the product of m. As a method for determining the above average effective pressure Pm(P)li), a simple calculation method using the following equation is known.

PM=ΣSφP(i)・△Vi : “1 = Σ5−P(i)・(ΔVi/Δθi) Δθi bow Here, S is a coefficient depending on the crank angle, and the compression and exhaust strokes (O≦i<180,360≦i<540), 5=-1, and the explosion and inhalation stroke (180≦!
<360,540≦i <720>, S=1
It is. ΔVi/Δθi represents the volume change in a predetermined crank angle range, and Δθi is set in advance by, for example, 1 [de
If it is determined as gl, ΔVi/Δθi can also be determined in advance. k is the period for calculating the average effective pressure PM, and is usually 720 [degl].

As shown in FIG. 3, the cylinder pressure P(i) detected by the pressure sensor 27 changes with each stroke, but in a four-cylinder engine, each stroke is executed simultaneously in each of the four cylinders. Therefore, the average effective pressure PM can be found at 360°CA by using the output of the pressure sensor 27 provided in the cylinder that is executing a stroke that is shifted by at least 360°CA.

Here, the reason why it is not 180['CA] is that pressure sensor 4-2
This is to cancel the offset amount included in the detection of the cylinder pressure P(i) by 7.

By the method described above, the electronic control device 40 can easily know the output torque level of the engine.

Next, the functions of the electronic control device 40 will be explained. The electronic control device 40 is designed as an additive integral type optimal regulator that simultaneously controls the output torque T of the engine 1 and the oxygen concentration EO in the exhaust gas. FIG. 4 shows this in the form of a block diagram.

As shown in the figure, the additive integral type optimal regulator operates given a target torque T and a target oxygen concentration EO' (Pl, P2>, but since it linearly approximates the nonlinear behavior of the engine 1, the linear approximation The stationary point (Ta, EO
a, θa, FRa) from the perturbation valve (δd, δEO, δ
It is configured to handle each control amount as (θ, δFR) (P3°P4). Further, the additive integral type optimal regulator converts the state variable amount X(k) into the perturbation valve (δd, δEO).

δθ, δFR) (P5), and the deviation ST between the target torque T and the actual output torque T(k).
Cumulative ZT (k) and cumulative value ZE of the deviation SEO between the target oxygen concentration EO and the actual oxygen concentration EO (k) in the exhaust gas
Obtain O(k) (P6.P7) and find the pixel product value ZT(k)
, ZEO(k), the state variable ff1X(k
) and multiplying it by a predetermined optimal feedback gain, the throttle valve opening θ
and the feedback control amount of the fuel injection ff1FR (here, the perturbation valves δθ(k) and δFR(k) from the steady point) are determined (P8).

Therefore, the engine 1 has this perturbation valve (δθ(k)).

The control amount (θ(k), FR(k)) which is the sum of the steady point value (θa, FRa) to δFR(k)) is output (
P4).

Next, the design procedure of the above-mentioned additive integral type optimal regulator will be explained.

(1) Modeling of the control system The behavior of the control system, here the system that controls the output torque T of the engine 1 and the oxygen concentration EO in the exhaust gas, is expressed by the state equation,
Using the output equation, X(k)=A−X(k−1)+I3−u(k−1
)−= (1)y(k)=[F]・X,(k)
...Described as (2). Furthermore, the formula (1
), (2>, X(k) is the state variable quantity, u(k)
is the control input quantity of the engine 1 (in this example, the engine 1
The throttle valve opening θ and the fuel injection amount FR) are expressed as v(
k) represents the output torque as a control output of the engine 1 and the oxygen concentration EO in the exhaust gas, and the subscript represents a variable indicating the number of I11 operations since the start of the first sampling. Therefore, in this embodiment, the control system is treated as a multicomponent system with two human power and two outputs.

FIG. 5 is a diagram in which the system of the engine 1, which is operated steadily as a two-man power, two-output system, is expressed by transfer functions Gl(Z) to G4(2). It is assumed that Z indicates the 2-conversion of the sample values of the input and output signals, and Gl(z>~G4(z) has an appropriate order. Therefore, the entire transfer function matrix G(
z) is expressed as

As in the engine 1 of this embodiment, the control system is powered by two people.
It is an output system, and if there is interference between input and output quantities, it is extremely difficult to define a physical model. In such a case, the transfer function can be determined by a type of simulation called system identification, but here, the least squares method is used to identify the transfer function.

When the engine 1 is operated steadily in a predetermined operating state, and the change in throttle opening δθ is set to zero, the change in fuel supply amount δF
Add a suitable test signal as R, then input δFR
The data of the changes in the output torque, which is the output, of the six engines are sampled N times. This is the input data series (u
(i>)=(δFRi), output data series (y(i)
)−(δTi) (However, r = 1.2.3...
・Represented as N).

At this time, the system can be considered to have one output per person, and the transfer function Gl(Z) of the system is Gl (z ＞=B(z-')/A (z-1>・
(3) That is, Gl(Z) = (bO+bl −z +・+bn z =) /
(1+al -z +a2 -z-2+...+
an-z)...calculated by (4). Note that here, z-1 is a unit transition operator,
This means z-1-x (k)=x (k-1).

Input/output data series (u(i >), (y(i
>) to the parameters a1~art, bO~b of equation (4)
Once O is determined, the transfer function Gl(z) of the system can be found. In system identification using the least squares method, this parameter al
~an, bo ~bn @Jo=Σ[(y (
k) +a1 − y (k−1)+−...shin +an −y(k−n>)− (bO−u (k)+bl −u (k−1) ten・+
bn −u (k−n) ) ]2− (5) is determined to be the minimum. In this example, each parameter was determined with 0=2. In this case, the signal flow diagram of the system becomes as shown in Figure 6, and the state variable quantity is [xl (k
) x2 (k) ]”Wotto, that state
The output equation is expressed as...
, IBI-, CI-, then it becomes.

In this example, the parameters for Gl(Z) are [al a2 ] = [-14,770,52][
bObl b2] = [06,62X10' 7.23X 10-3]
I got it. Using a similar method, transfer functions 02(z) to G4(Z) and system parameters A2 for each
- to A4 '', fB2- to B4'', C2-'
, A to C4- are obtained. Therefore, from these system parameters, the system parameters of the original two-man power, two-output multicomponent system, the state equation (1), and the vectors Δ, B, and C of the output equation (2) can be determined.

In this way, the dynamic model of this example was obtained by system identification, and when the engine 1 is operated in a predetermined state, a linear approximation holds in the vicinity of this state. It is determined in the form. Therefore, for a plurality of steady operating conditions, the transfer function Gl
(z) to G4(Z) are each obtained, and the vector A in each state equation (1) and output equation (2).

B and C are obtained, and the relationship between the input and output is established between the perturbation force δ.

(2) Observer design) Since it is thought that a linear relationship holds between the perturbing force δ, δEO and δθ, 6R, the observer (P5 ) to design.

Observers include same-dimensional observers and minimum-dimensional observers, and their design methods are detailed in ``Mechanical System Control'' (cited above).

In this embodiment, it is assumed that the same dimension observer is designed.

The same-dimensional observer has the configuration shown in FIG. 7, and as shown in the figure, the estimated value X(k) of the state variable is
Let K be the feedback gain, 10K (C-X(k
-1>-y(k-1>)=(A-to-C)X(k
-1> 10th layer (k-1> ) 10 B -u (k-1)
−(11). Here, eight-[
If we choose K that makes F] stable and make sure that the absolute values of the eigenvalues of the matrix (△-'t-c> are all less than 1, we get X(k)→X(k) at k→■ Therefore, by defining K as such, and further AoΔA−L・
C LA) K...(12)
If IBo AIB, the observer has the following equation:
[3o −u(k−1) =−(13).

As described above, the observer P5 was established from the vectors A, [B, and C of the state equation (1>, etc.) obtained by system identification.

(3) Expansion of the system Since the object to be controlled in this embodiment is a servo system in which the target torque below* and the target oxygen concentration EO* change, the system is expanded using the cumulative value. That is, the state variable ff1X(k) estimated by the observer P5 and the cumulative value Z(k) = [
ZT(k) 7EO(k)]", and set this as the state variable ff1X(k) again. That is, X(k) = [X(k)2cho(k)ZEO(k)]
”...(14) becomes.

(4) Calculation of optimal feedback gain F A method for determining the optimal feedback gain for the expanded state variable ff1X(k) is, for example, linear system control theory.
(cited above), etc., so I will omit the detailed explanation here and only show the results.

The system expanded in (3) above is expressed as follows.

...(15) Here, if the state variable 1X(k) is quartic, then X(k
) = [xl(k) x2[k) x3(k)
x4(k)] ``u(k) = [δθ δFRY'' v(k) = [δT 6EO]T 楇* (k) = [δT book δEO*]. As mentioned above, δT, δEO, δθ, and δFR are
It represents the deviation (perturbation) from the steady point.

At this time, finding the optimal control input that minimizes the next evaluation function J, that is, the operating condition U'' (k), solves the control problem for the engine as an additive integral type optimal regulator.

J=Σ[δtsuT(k)・Q・δy(k)−O 十δ鑞”T(k)・R・δu(k)] ...(17) Here, Q and IR are weights In the parameter matrix, k indicates the number of samples with O as the control start time, respectively,
The right side of equation (17) is expressed in a so-called quadratic form with G and IR as diagonal matrices.

As a result, the optimal control inputs u (k) are: u'' (k)=-Fl ・X(k)-F2 ・Z(
k)...(18)

Here, il"l, "2 is [PI F2] = [I・F3"-P...
(19), and P is a positive definite solution of the following matrix Rikkatti equation.

A”-P+P-A+0-P-1t3-fR-I・[BT
-IP-〇...(
20) Here, the meaning of the evaluation function J in equation (17) is that the various operating conditions 1u(k
) = [δθ δFR] while constraining the movement of various operating state quantities y (k) as control outputs, here the target values v of various my (k) including output torque δ and oxygen hotness δEO. The intention is to minimize the deviation from (k). The weighting of the constraints on various quantities u (k ) of the operating conditions can be changed by the values of the weight parameter matrices Q and IR. Therefore, using the already determined dynamic model of engine 1, that is, the matrix A, [8, C, and selecting arbitrary weight parameter matrices G, IR, the equation (2
0) to find P, and then the optimal feedback gain "can be found using equation (19). Therefore, using this optimal feedback gain E, the control inputs Mu (k) (here, δO1δFR) of the engine 1 can be calculated. can be obtained as u(k)=F·X(k) (21).

Above, we have explained the modeling of the control system, the design of the observer, the expansion of the system, and the determination of the optimal feedback gain based on the configuration of the additive integral type optimal regulator (Figure 4), but these are determined and required in advance. Inside the electronic control device 40, actual control is performed using only the results.

Next, the control actually performed by the electronic control device 40 will be explained with reference to the flowchart of FIG. In the following explanation, the amount handled in actual processing will be expressed with a subscript (k), and the amount handled last time will be expressed with a subscript (k-1). After the engine 1 is started, the MPU 44 repeats the processing from step '100 onwards. First, in step 100, the operating state of the engine 1, that is, the output torque T(k-1) of the engine 1, the oxygen concentration EO
(k-1>, etc.) is calculated based on the output of each sensor 27, 34.

In the following step '110, the amount of accelerator pedal depression detected by the driver and the cooling water temperature Thw detected by the accelerator sensor 37 are determined.
Calculate the target torque T of engine 1 based on etc.,
In step 120, a target oxygen concentration of 60 lines for the engine 1 is calculated. These processes correspond to Pi and P2 in FIG.

In step 130, the deviation ST between this target torque T* and the actually detected output torque T (k-1) is calculated, and the 60 target oxygen concentrations and the actual oxygen concentration EO (k-1>
Processing is performed to obtain the deviation SEO from each. In the subsequent step 140, the process of accumulating each deviation obtained in step 130, that is, ZT (k > = ZT (k-1
> +ST (k-1> gives the cumulative value ZT (k), while ZEO(k)=ZEO(k-1> +SEO(
k-1>, the process of determining the cumulative value ZEO(k) is performed. This process is shown in P6 of FIG. Corresponds to P7.

In the following step 150, when a dynamic model of the engine 1 is constructed from the operating state of the engine 1 read in step 100, the closest state (hereinafter referred to as , this is called a stationary point (EOa, Ta)). In step 160, the operating state of the engine 1 read in step 100 is determined at this steady point (EOa, Ta).
) is calculated as the perturbation (δEO, δT).

This process corresponds to the perturbation output mechanism P3 in FIG.

In step 170, a predetermined A○ is determined.

[30,1 and the perturbation component (δE
O, δt) and the state variable amount X (k-1
＞= [xl (k-1) x2 (k-1>
... k
−1>, the new state variable quantity X is calculated by the following equation (23)
A process for calculating (k) is performed. This process corresponds to the observer P5 in FIG. 4, but in this embodiment, as already mentioned, the observer P5 is configured as a finitely stable observer. That is, X(k) = to o −X (k−1) +IBo
]T+ "・[δEO(k-1)δT(k-1)]"・
...(22) is calculated.

In the subsequent step 180, vector multiplication is performed by the state variable ff1X(k) obtained in step 170, the cumulative value ZT (k >, ZEO(k)) obtained in step 140, and the optimal feedback gain prepared in advance. Therefore, [δFR(k > δθ(k)]=F
・[X (k) ZT(k) ZEO(k)IT
Processing is performed to obtain the perturbation forces δFR(k 2 ) and δθ(k 2 ) of the manipulated variable. This corresponds to the feedback amount determination section P8 in FIG.

In step 190, the perturbation forces δFR(k) and δθ(k) of the manipulated variables obtained in step 180 are added to the respective operations ff1FRa and θa at the steady point, and the actual engine
The operation amount output to the fuel injection valve 11 and actuator 35, that is, the operating condition FR(k).

θ(k) is found (corresponding to P4 in FIG. 4).

In the subsequent step 200, the value k indicating the number of numbering is incremented by 1 from 1 to V, and then the process returns to step 100 to repeat the above series of processes, steps 100 to 200.

By periodically and continuously performing the above control, the electronic control circuit 40 operates as an additional integral type optimal regulator that controls the operating state of the engine 1 to a target torque of 91 cylinders and a target oxygen concentration of 60 cylinders.

Therefore, the engine control device of this embodiment can accurately and quickly control the output torque and the oxygen concentration in the exhaust gas, which was previously impossible to achieve.

For example, Figure 9 shows the actual output torque T (k) of the engine 1 and the oxygen concentration EO (k ) (However, in this figure, the 'ffi axis is a graph showing the relationship between the air-fuel ratio of the air-fuel mixture corresponding to the oxygen concentration in the exhaust gas), but as shown in the figure, the air-fuel ratio is approximately the stoichiometric air-fuel ratio (A/ It can be seen that the output torque (k) can be controlled to the target torque of 91 in a short period of time while keeping the value F→15). There was no phenomenon in which the output torque temporarily dropped during a transient period, and the output torque T(k) was approximately 0.6 [seconds], which was the target torque of 91. As a result, in a vehicle equipped with the engine control device of this embodiment, the weight of the flywheel attached to the output shaft of the engine 1 can be reduced to approximately 1/2 of the conventional weight, and the responsiveness of the engine 1 can be reduced by - This results in an improvement in G.

In addition, since the output torque can be stably controlled, the entire drive system including the automatic transmission can be easily controlled.
Another advantage is that the durability of the catalyst can be improved because the oxygen concentration (EO) and the exhaust gas composition itself can be stably controlled.

Furthermore, in this embodiment, the output torque T(k) is detected by the pressure sensor 27, so that the torque detection means can have a simple configuration.

Next, a second embodiment of the present invention will be described. The engine control device of the second embodiment has almost the same configuration as the first embodiment, and as shown in FIG.
It has a configuration determined from the reference oxygen concentration E base and the compensation value CEO(k) determined by the target torque T* (
Pl). Specifically, as shown in the flowchart of FIG. 11, step 125 is provided in place of step 120 in the first embodiment, and the target oxygen concentration EO' is set to a fixed value (one in this embodiment). A process is performed to find the difference between Q-oxygen permeability EObase and compensation value CEO.

The compensation value CEO is determined by the transfer function GCEO to the target torque T.
1, and the graph in FIG. 12 shows this.

As a result, in addition to the effects of the first embodiment, as the target torque 91 increases, the target oxygen concentration EO is set lower, the air-fuel mixture is controlled to the rich side, and the target output torque T
I find it easy to get books.

Figure 13 shows that the target torque T* is 2.0 to 3°5 [k
It is a graph which shows the change of output torque T(k) when it changes to c+fml. As shown in the figure, in this example, as a result of the change in the target torque T*, the target oxygen concentration 20 also changes, and the output torque T(k) matches the target torque even faster than in the first example. I understand that. Note that the compensation (direct CEO) may be determined from the torque deviation as shown in FIG.

Several embodiments of the present invention have been described above, but
The present invention is not limited to these embodiments, and may be modified without departing from the gist of the present invention according to the aspect of the engine, such as a configuration in which ignition timing, exhaust gas recirculation amount, etc. are added as control input quantities. Of course, it can be implemented in various ways.

Effects of the Invention As detailed above, the engine control device of the present invention has an extremely excellent effect of being able to control the engine output torque and exhaust gas composition to the target values at all times, including during transient times. Therefore, it is possible to sufficiently control the entire drive system including the automatic transmission, etc., and it is also possible to contribute to improving the durability of the catalyst provided in the exhaust system of the engine. Furthermore, since the engine's output torque can be controlled during transient periods, the inertia of the flywheel attached to the engine's output shaft can be significantly reduced, significantly improving response during acceleration and deceleration and fuel efficiency. be able to. As a result, the engine's response to the driver's operations is extremely high, and both drivability and drive feeling can be improved.

In addition, since both the output torque and the concentration of a predetermined component in the exhaust gas (i.e., the air-fuel ratio) can be controlled independently, when it is desired to preferentially lower the output torque to the target torque, the target concentration of the predetermined component in the exhaust gas can be compensated. Therefore, it is easy to obtain the effect of further improving the control characteristics (responsiveness) of the output torque.

[Brief explanation of drawings]

FIG. 1 is a block diagram illustrating the basic configuration of the present invention, FIG. 2 is a schematic configuration diagram showing the hardware of an engine control device as a first embodiment of the present invention, and FIG. 3 is a block diagram illustrating the basic configuration of the present invention. A graph explaining the method of output torque detection using a sensor, Figure 4 is a block diagram showing the configuration as an additive integral type optimal regulator, and Figure 5 is a block diagram used to identify the model of the system in the example. , FIG. 6 is a signal flow diagram for determining the transfer function, FIG. 7 is an explanatory diagram showing the configuration of a same-dimensional observer, FIG. 8 is a flowchart showing control as an additive integral type optimal regulator in the embodiment, and FIG. FIG. 9 is a graph showing a control example in the embodiment, FIG. 10 is a block diagram showing the configuration in the second embodiment of the present invention, FIG. 11 is a flowchart showing an example of the control, and FIG. 13 is a graph showing an example of the relationship between the output torque and the oxygen concentration EO in the second embodiment, and FIG. 14 is a graph showing the relationship between the compensation value CEO and the oxygen concentration. FIG. 15 is a graph showing an example of calculation, and is a graph showing the relationship between throttle actuator characteristics and air-fuel ratio changes in the prior art. 1... Engine 7... Throttle valve 11... Fuel injection valve 27... Pressure sensor 34... Lean sensor 35... Throttle actuator 37... Accelerator sensor 40... Electronic control circuit

Claims (1)

[Scope of Claims] 1. Torque detecting means for detecting the output torque of the engine; Concentration detecting means for detecting the concentration of a predetermined component reflecting the exhaust gas composition of the engine; target value output means for determining and outputting a target torque of the engine and a target concentration of the exhaust component; and a control input for varying various quantities including at least an intake air amount and a fuel injection amount of the engine as control inputs for the engine. control of the engine, wherein the output torque of the engine detected by the torque detection means is the target torque, and the concentration of the predetermined component detected by the concentration detection means is the target concentration. An engine control device comprising: control means for determining feedback control amounts of various input quantities and controlling the control input amount variable means, the control means comprising: a dynamic model of a system related to the operation of the engine; is used to control the output torque and predetermined concentration of the engine detected by the torque detection means and the concentration detection means, and the control input amount variable means using predetermined parameters based on the state variable amount estimating means for estimating a state variable amount of a predetermined order representative of the dynamic internal state of the engine from the feedback control amount obtained by the engine; and the detected engine output torque and the set target torque. Torque deviation cumulative value, which is the cumulative deviation from
an accumulating means for calculating a concentration deviation cumulative value obtained by accumulating the deviation between the detected concentration of the predetermined component and the set target concentration; a feedback gain preset based on a dynamic model of the engine; Feedback control that determines the feedback control amount of control input variables including at least the intake air amount and fuel injection amount of the engine from the estimated state variable amount and the determined torque deviation cumulative value and concentration deviation cumulative value. An engine control device comprising: quantity determining means. 2. The target value output means includes a concentration compensation value calculation unit that calculates a target concentration compensation value based on the deviation between a target torque determined based on a request for engine operation and the output torque of the engine, or the target torque; 2. The engine control device according to claim 1, further comprising a target concentration setting section for compensating a target concentration of a predetermined component determined based on a request for operation using the calculated compensation value. 3. The control input amount variable means is configured to vary at least one of the engine's intake air amount and fuel injection amount, as well as the engine's ignition timing, exhaust gas recirculation amount, supercharging amount, and idle intake air amount; Claims 1 or 2, wherein the feedback control amount determining means is configured to determine feedback control amounts of control input variables including ignition timing, exhaust gas recirculation amount, supercharging amount, and idle intake air amount. The engine control device described.