JPH02238150A - Fuel injection control device of internal combustion engine - Google Patents

Abstract

PURPOSE:To correct a control error and to improve the control accuracy by adding a variation component to the fuel injection amount, inferring a model parameter from the variation amount of the air-fuel ratio derived following this addition, and correcting an operation formula for computing the fuel feeding amount thereby. CONSTITUTION:To the fuel feeding amount computed by a device M3 depending on the operating condition detected by a device M1 and the condition variable inferred by a device M2, a specific variation component is superposed, and the fuel injection amount from a fuel injection valve controlled by a device 4 is varied by a device M5. And the air-fuel ratio of the fuel mixture gas is detected by a device M6 depending on the exhaust gas component of the internal combustion engine E/G. Furthermore, while the variation amount of the air-fuel ratio is computed by a device M7, a specific parameter in a physical model is inferred by a device 8 depending on this variation amount. And, depending on the parameter, an operation formula used when the fuel feeding amount is computed by the device M3 is corrected by a device M9.

Description

[Detailed Description of the Invention] [Field of Industrial Application] The present invention relates to an internal combustion engine in which the amount of fuel injected from a fuel injection valve is controlled in accordance with a physical model representing the behavior of fuel flowing into a cylinder of an internal combustion engine. The present invention relates to a fuel injection amount control device.

[Prior art 1] Conventionally, as one type of fuel injection amount control device that controls the amount of fuel injection from a fuel injection valve so that the air-fuel ratio of the fuel mixture supplied to the internal combustion engine becomes a target air-fuel ratio, for example, As described in Japanese Unexamined Patent Publication No. 59-196930, a correction value for correcting the basic fuel injection amount determined from the rotational speed of the internal combustion engine and the amount of intake air is controlled. The actual measured value of is the control output, and identification is performed assuming that a linear approximation holds between the control input and the control output, a mathematical model that describes the dynamic behavior of the internal combustion engine is obtained, and the design is based on this. A control device based on so-called linear control theory, which controls the fuel injection amount using a control law, is known. However, the relationship between the above-mentioned control input and control output is inherently non-linear, and obtaining a mathematical model using a single-bilinear approximation as described above can only describe the dynamic behavior of an internal combustion engine under extremely narrow operating conditions. For example, as described in Japanese Unexamined Patent Publication No. 59-7751, a mathematical model is set up for each of multiple operating regions where linear approximation can be considered to hold, and based on this, each Therefore, in the past, the control law had to be changed for each operating region of the internal combustion engine, which caused problems such as complicated control. There is also the problem that control becomes unstable at the boundary point due to switching of the control law.

Therefore, applicant 1; l. , Patent application 1986-1898
No. 89, Japanese Patent Application No. 62-189891, etc., by determining a nonlinearly compensated control law based on a physical model that describes fuel behavior in an internal combustion engine, it is possible to avoid switching the control law as described above (i.e. with one control law)
[Problem to be solved by the invention] However, even with the above physical model, it is difficult to accurately describe the fuel behavior of an internal combustion engine. The model parameters fluctuate due to changes in the characteristics of the intake system, etc. The control amount error due to these fluctuations is 1tll. Although the control law can be extended to the well-known servo system and compensated by integral operation, it is necessary to increase the ratio of integral compensation. This may result in poor response, and in order to further improve controllability, it is desirable to find the model parameters in some way and correct the control law.

Therefore, the applicant of the present application (Friend) proposed the above-mentioned fuel injection amount control device 1:. We proposed a device that controls the fuel injection amount by installing an identifier that estimates model parameters from the fuel injection amount, air-fuel ratio, etc., and correcting the control law based on the identification results. However, the proposed device uses least squares estimation. Because the model parameters were estimated based on mathematical principles known as equations, the calculation equations for estimation were complex, and calculations took a long time with the inexpensive 8-bit microcomputers used in conventional internal combustion engine control systems. In addition, since the air-fuel ratio detection results using the air-fuel ratio sensor are used as they are for estimating the model parameters, there are no changes in the characteristics of the air-fuel ratio sensor, disturbance noise, or changes in the exhaust flow velocity. Due to the accompanying delay in detecting the air-fuel ratio, etc., errors tend to occur in the estimation of model parameters, and smoothing processing of the air-fuel ratio detection results to prevent this problem also complicates stability determination.

Therefore, the present invention (1) controls the fuel injection amount using a control law based on a physical model that describes the fuel behavior of an internal combustion engine, and also estimates model parameters as necessary and compensates for control errors caused by fluctuations in these parameters. In the apparatus configured as shown in FIG. 1, model parameters can be estimated easily and accurately. The above operation is performed using an operation state detection means M1 that detects a predetermined operation state, and a first calculation formula set based on a physical model that describes the behavior of fuel flowing into the cylinder of the internal combustion engine E/G. a state variable estimating means M2 that estimates a state variable of the physical model based on the detection result of the state detecting means M1 and the previously calculated fuel supply amount; and a second calculation formula set based on the physical model. , a fuel supply amount calculation means M3 that calculates a fuel supply amount based on the detection result of the driving state detection means M1 and the estimation result of the state variable estimation means M2; The fuel injection valve installed in the intake passage of the E/G is opened, and the internal combustion engine
A fuel injection amount control device for an internal combustion engine, comprising: a fuel supply means M4 for supplying fuel to /G; A fuel injection amount variation control means M5 that varies the fuel injection amount from the fuel injection valve controlled by M4; An air-fuel ratio detecting means M6 that detects the air-fuel ratio; an air-fuel ratio variation calculating means M7 that calculates the amount of variation in the air-fuel ratio based on the detection result of the air-fuel ratio detecting means M6; and the calculated amount of air-fuel ratio variation. model parameter estimating means M8 for estimating predetermined model parameters of the physical model based on the model parameters;
The gist of the present invention is a fuel injection amount control device for an internal combustion engine, which is characterized in that it is provided with a correction means M9 for correcting the arithmetic expression.

[Operation] In the fuel injection amount control device of the present invention configured as described above, first, the state variable estimating means M2 compares the predetermined operating state of the internal combustion engine E/G detected by the operating state detecting means M1 with the previous operating state. Based on the calculated fuel supply amount, the internal combustion engine E/
The state variables of the physical model describing the fuel behavior in E/G are estimated, and the fuel supply amount calculation means M3 calculates the fuel supply amount based on this estimation result and a predetermined operating state of the internal combustion engine E/G. Then, the fuel injection amount fluctuation control means M5 superimposes a predetermined fluctuation component on the calculated fuel supply amount, and the fuel supply means M4 controls the fuel injection valve according to the fuel supply amount on which this fluctuation component is superimposed. The valve is opened to supply fuel to the internal combustion engine E/G.

In addition, in the present invention, the air-fuel ratio fluctuation amount calculation means M7 calculates the amount of fluctuation in the air-fuel ratio detected by the air-fuel ratio detection means M6, and the model parameter estimating means M8 performs the above-mentioned operation based on the calculated amount of fluctuation in the air-fuel ratio. Estimate model parameters. Then, the correction means M9 corrects the second arithmetic expression used by the fuel supply amount calculation means M3 to calculate the fuel supply amount based on the estimated model parameters.

That is, in the present invention, the fuel injection amount fluctuation control means M512 gives a predetermined fluctuation component to the fuel injection amount to the internal combustion engine, estimates model parameters from the resulting fluctuation amount of the air-fuel ratio, and calculates the control law. It is designed to be corrected.

[Example] Embodiments of the present invention will be described below with reference to the drawings.

First, FIG. 2 is a schematic configuration diagram showing the configuration of an internal combustion engine 2 and its peripheral devices to which the present invention is applied. As shown in the figure, in the intake pipe 4 of the internal combustion engine 2 (L), from its upstream side, an air cleaner 6 that purifies the intake air, a throttle valve 8 that controls the amount of intake air, a surge tank 1o that suppresses the pulsation of intake air,
An intake pressure sensor 12 that detects the pressure (intake pipe pressure) PM in the surge tank 10 and an intake air temperature sensor 14 that detects the intake air temperature THA are provided. , and an air-fuel ratio sensor 19 that detects the air-fuel ratio of the fuel mixture supplied to the internal combustion engine 2 based on the oxygen concentration in the exhaust gas upstream of the three-way catalyst 18. 1. The intake pressure sensor 12, the intake temperature sensor 14, and the air-fuel ratio sensor 19 are included as sensors for detecting the operating state.A rotation speed for detecting the rotation speed ω of the internal combustion engine 2 from the rotation of the distributor 20. A sensor 22, a crank angle sensor 24 for detecting the fuel injection timing to the internal combustion engine 2 based on the rotation of the distributor 20, and a water temperature sensor 26 for detecting the cooling wood temperature THW attached to the water jacket of the internal combustion engine 2 are provided. The distributor 20 is for applying the high voltage from the igniter 28 to the ignition plug 29 at a predetermined ignition timing.

The detection signals from each of the above sensors are then input to an electronic control circuit 30 configured as a logical operation circuit centered on a microcomputer to drive the fuel injection valve 32 and control the amount of fuel injected from the fuel injection valve 32. used to do.

That is, the electronic control circuit 301L is a CPU 40 that executes arithmetic processing for fuel injection amount control according to a preset control program, and an R in which control programs and initial data necessary for executing arithmetic processing in the CPU 40 are recorded in advance.
OM42, same < RAM44 where data used to execute arithmetic processing by CPU40 is temporarily read and written
, an input port 46 for inputting detection signals from each of the sensors, and an output port 48 for outputting a drive signal to the fuel injection valve 32 in accordance with the calculation result of the CPU 40.
The fuel injection amount from the fuel injection valve 32 is controlled so that the air-fuel ratio of the fuel mixture flowing into the cylinder 2a of the internal combustion engine 2 becomes a preset target air-fuel ratio.

Next, the control law for fuel injection amount control executed by this electronic control circuit 30 will be explained based on the block diagram shown in FIG.

FIG. 3 is a diagram showing the control law of the embodiment,
This does not indicate a hardware configuration, and actual fuel injection control is realized by executing a series of control programs shown in flowcharts in FIGS. 4 and 5, which will be described later.

In addition, as will be described later, the following equation (1
) and (2) (hereinafter also referred to as fuel behavior model).

+ (I R S) fi (k) ... (2) (
However, in the above formula, fw: amount of fuel adhering to the intake pipe wall, fv
: Amount of evaporated fuel in the intake pipe, fI = Fuel injection amount, Vf
w: amount of fuel evaporated from the intake pipe wall surface, fc: amount of fuel flowing into the cylinder, P, Q, R, S, D: constants. ) As shown in FIG.
First, the intake pipe pressure PM detected by the intake pressure sensor 12 and the cooling water temperature THW detected by the water temperature sensor 26 are input to the fuel evaporation rate calculation section B1.

Fuel evaporation rate calculation unit All friends This is for calculating the amount of fuel evaporation per unit time (fuel evaporation rate) Vf from the intake pipe wall surface from the intake pipe pressure PM and cooling water temperature THW.
First, the saturated vapor pressure Ps of the fuel in the intake pipe 4 is determined from the cooling water temperature THW, and the fuel evaporation rate Vf is calculated from this saturated vapor pressure Ps and the intake pipe pressure PM.

In other words, the fuel evaporation rate from the intake pipe wall surface Vff1 intake pipe 4
It can be determined as a function of the saturated vapor pressure Ps of the fuel inside the air and the pressure inside the intake pipe 4 (intake pipe pressure) PM, and the saturated vapor pressure Ps is a function of the temperature Tq of the fuel adhering to the wall of the intake pipe. , the adhering fuel temperature Tq can be represented by the cooling water temperature of the internal combustion engine 2 or the cylinder head temperature near the intake boat. 3) to find the saturated vapor pressure Ps, Ps=β1-T}-IW2-β2・THW×β3・
-.(3) (where β1, β2, β3 are constants) Thereafter, the fuel evaporation rate Vf is calculated from the calculated saturated vapor pressure Ps and intake pipe pressure PM.

Next, the fuel evaporation rate Vf calculated by the fuel evaporation rate calculation unit B1 is input to the fuel evaporation amount calculation unit B2. This fuel evaporation amount calculation unit M21 calculates the amount of evaporation from the intake pipe wall surface per one rotation of the internal combustion engine 2 by dividing the fuel evaporation rate Vf by the rotational speed ω of the internal combustion engine 2 detected using the rotational speed sensor 22. Fuel amount V fw (= V
f/ω), and the calculation result (fuel evaporation amount) Vfw is input to the coefficient f4 multiplier B3 and multiplied by a preset grandchild number f4.

Next, the intake pipe pressure PM and the rotational speed ω1 are also input to the cylinder inflow air amount calculation unit P4 together with the intake air temperature THA detected by the intake air temperature sensor 14.

Cylinder intake air amount calculation part P41 Based on these input intake pipe pressure PM, rotational speed ω, and intake air temperature THA, the following formula (4) mc= (βX((Ll)・PM−βy(ω))
/ T }-I A-(4) (However, βX(ω), βy
(ω): a function of rotational speed ω) to calculate the amount of air mc flowing into the cylinder 2a during the intake stroke of the internal combustion engine 2, and the calculation result mc is sent to the target fuel amount calculation unit BS. It is input. Then, the target fuel amount calculation unit B5 multiplies the calculated air amount mc by a preset target fuel-air ratio (reciprocal of the target air-fuel ratio) λ', and calculates the cylinder 2a.
Calculate the amount of fuel that should flow into the tank (that is, the target fuel amount) fcr. Further, this target fuel amount fcrl is multiplied by a preset coefficient f3 input to the coefficient f3 multiplier B6. On the other hand, the fuel evaporation amount VfwJ calculated by the fuel evaporation amount calculation section B2 is also input to the state variable estimation section P7. State variable estimation section P71 table Using a preset calculation formula, calculate the input fuel evaporation amount Vfw and the temporary delay section B8.
The previously calculated fuel injection amount fi(k
-1) and a model parameter variation calculation unit 813, which will be described later.
The variation amount ΔP of the model parameter P calculated in and the state variable amount f w (
k-1) and? From v(k-1), state variables fw and fv for calculating the next fuel injection amount fi(k) are estimated. And this estimated result? W and? V is multiplied by coefficients fl and f2 in a coefficient ft multiplication unit B9 and a coefficient f2 multiplication unit BIO, respectively.

Next, the air-fuel ratio detection result A by the air-fuel ratio sensor 19
/ F [. t, A/F fluctuation amount calculation unit B11. A/F fluctuation amount calculation part B 1 1 1 friend
This is for calculating the fluctuation width (A/F fluctuation amount) ΔA/F from the time series data of the air-fuel ratio detected by the air-fuel ratio sensor 19, and the calculation result ΔA / F t& Cylinder inflow air amount calculation unit It is input to the in-cylinder fuel fluctuation amount calculation unit B12 together with the air amount mε calculated in B4. Then, the cylinder inflow fuel fluctuation amount calculation unit 812 calculates Lt. Based on the input data ΔA/F and mc, a fluctuation width fcD of the amount of fuel flowing into the cylinder 2a fc is calculated, and the calculation result is output to the model parameter fluctuation amount calculating section 813. Model parameter fluctuation amount calculation unit 8131L Based on the input fluctuation range fcD of the inflow fuel amount fc, the model parameters in the fuel behavior model described by the above equations (1) and (2) are calculated using a preset map. This is to calculate the fluctuation range ΔP of P, and the calculation result ΔP
(& {N coefficient f5 of the state variable estimator B7
The multiplication unit 814 multiplies the output card by a coefficient f5. Also, this multiplication result f5·ΔP is input to the state variable multiplier B15.The state variable estimated by the state variable estimator B7? W is multiplied. Then, the multiplication result I-L by this state variable multiplier 815
The adders 816 to 819 add the multiplication results in the other coefficient multipliers 83, 86, 89, and 810. Thereby, the fuel injection amount f1 from the fuel injection valve 32 is calculated.

In addition, a sine wave sin2χfT output from the sine wave generator B20 is hidden in the calculated fuel injection amount f1 by the amplifying unit 8.
Fluctuation component Δ obtained by amplifying A times in 21
fi (=As i n2πfT) is added to the adder B22
This determines the amount of fuel TALJ (=fi+Δfi) to be actually injected from the fuel injection valve 32. Next, a fuel behavior model that is the basis of the above control law and a method of designing the above control law based on the fuel behavior model will be explained. In addition, as a design method for this kind of control law, {Table 1f, Katsuhisa Furuta, [Digital Control of Actual Systems] System and Synonyms Vol. 2B, No. 12.
1984 Style I am familiar with the Society of Instrument and Control Engineers, etc., so I will briefly explain it here.

First, the amount of fuel fc flowing into the cylinder 2a of the internal combustion engine 2
lj It can be written as in the following equation (5) using the fuel injection amount f from the fuel injection valve 32, the fuel amount fw adhering to the wall surface of the intake pipe 4, and the evaporated fuel amount fv inside the intake pipe 4. Can be done. fc =al-f i +cr2・fw+cr3-fv
-(5) That is, the above fuel amount fcl & the direct inflow amount αl-fI of the injected fuel from the fuel injection valve 32 and the indirect inflow amount α2・fw from the intake pipe 4 to which the injected fuel adheres.
and the inflow amount α3·fv of evaporated fuel existing inside the intake pipe 4 due to the evaporation of the injected fuel or fuel adhering to the wall surface. Therefore, as shown in equation (5) above, The amount of fuel flowing in fc can be described.

In the above formula (5), the fuel injection amount fi is the fuel injection amount fi of the fuel injection valve 32.
Therefore, the amount of fuel adhering to the wall of the intake pipe 4 fw and the amount of evaporated fuel fv within the intake pipe 4 can be known (the amount of fuel fc can be predicted). Consider the adhering fuel amount fw and the evaporated fuel position fv.

First, the amount of fuel adhering to the wall of the intake pipe 4, α2, decreases in part during each intake stroke due to the inflow into the cylinder 2a during the intake stroke. The amount increases due to the adhesion of a portion α4 of the fuel injection amount f1 injected from the fuel injection valve 32 in synchronization with the cycle.

In addition, from the amount of fuel evaporation per intake cycle (.t, the amount of fuel evaporation per unit time (i.e. fuel evaporation rate) Vf and the rotational speed ω of the internal combustion engine 2, α5・Vf/ω (2 α5・V fw
, α5: constant of proportionality). Therefore, the amount fw of fuel adhering to the wall surface of the intake pipe 4 can be expressed as shown in the following equation (6). f w (k+1) = (iα2) − f w (k) +
a4・fi(k)−a5・Vfw(k)
- (6) (where k: intake cycle) On the other hand, the amount of evaporated fuel fv inside the intake pipe 4 [y] 4N fuel whose part α3 decreases every intake cycle due to the inflow into the cylinder 2a during the intake stroke The injection amount fi increases as part of α6 evaporates, and further increases as the adhering fuel evaporates. Therefore, the amount of evaporated fuel fv in the intake pipe 4 can be expressed as shown in the following equation (7).

f v(k+1)=(t-a 3)・f v(k)+
a 6・f i (k)+ a 5・V fw(k)
- (7) Therefore, in the above equations (5) to (7), by rearranging (l-α2) as P, (l-α3) as Q, α4 as R, α6 as S, and α5 as D, the intake air The above-mentioned (1) and (
2) Equation can be obtained, and a fuel behavior model expressed in a discrete system using the intake cycle of the internal combustion engine 2 as a sampling period is determined. In such a fuel behavior model, each model parameter P,
By determining Q, R, S, and D using a well-known identification method, it becomes possible to accurately describe the fuel behavior of a 1fS internal combustion engine during steady operation over the entire operating range.

However, since the characteristics of the internal combustion engine 2 change over time, the fuel behavior model described above may not be able to accurately describe the fuel behavior of the internal combustion engine 2. Furthermore, even if the internal combustion engine (tS) is the same model, there are variations in the characteristics of each engine, so it is difficult to describe the fuel behavior of multiple internal combustion engines with a single fuel behavior model. Even if a control law is designed based on a specific fuel behavior model, it is difficult to ensure control accuracy for each engine.

Therefore, in this embodiment, in order to automatically correct the control law and execute fuel injection control in accordance with such characteristic fluctuations and dispersion of the internal combustion engine 2, the model of the above fuel behavior model is introduced. The following equation (8) and (
The control law is determined by a fuel behavior model such as 9). I will explain.

Since the above fuel behavior model is nonlinear, in order to apply linear control theory, the fuel behavior model is first linearly approximated. In the above equations (8) and (9), x(k) = [f
w(k) fv(k)] ” −(10)
+( 1 −R−S) f i(k)−ΔP−fw(k
) ... (9) Next, regarding the design procedure of the control law of this embodiment based on the fuel behavior model described by the above equations (8) and (9), y(k) = f c(k) - ( 1 -R-S)f
i(k)+8P−f w(k)−(14)u
(k)” f i (k)
--- (15)C=[1-P 1
-Ql ... (I6), then the above equations (8) and (9) are x (k + 1) = A - x (k) + Bu (k) + w (k)
−(17)y (k)=C −x (k)
−(18). Here, when y(k) = y r (target value) and becomes stationary, u (k) = ur, x (k) = x
If r, the above equations (l7) and (18) become the following equation (19)
, as shown in (2b). x r =A−x r 1B −u r 1w(k) −(19)y r
=C-xr-(20)
From the above equations (17) to (20), x(k+1)-x r =A (x(k)-x r )
+B (u(k) -u r ) −(21)y (
k)-y r =C (x(k)-x r )
= {22) order 1:, in the above equations (21) and (22), X (k) = x (k) − x r
− (23) U (k)= u (k)
- ur -(24)Y (k
)=y(k)−y r−(2
5), equations (21) and (22) become the following equations (26) and (
27). X (k+1)= A X (k)
+ B u (k) ... (26) y (
k)=C x (k)...
(27) In (26) and (27), X(k)
-0 and 1 Y (k) = O. 11(k)-
Even if it is ur, it becomes l'Ly(k)-yr. Therefore, the above formula (
It is sufficient to design the optimal regulator of 26). That is, by solving the discrete Riccati equation, optimal control can be achieved using the following equation (2
8).

U (k) = F X (k) Also, this formula (28) (friend (29).

u(k)=F-x(k)-F-x r +ur Therefore, in the above equations (l9) and (20), the above (23)
And from equation (24), the following equation...(28)...(29) If solved for X', ur, the above equation (29) is confirmed,
Now it is possible to find u (k).

In the case of this example, the above equation (30) is
From equation 6), the following equation (3l) is obtained, and xr, ur (ie, fwr, fvr, fir) are determined as shown in the following equations (32) to (34), respectively.

fwr=β11・Vfw(k)+β12・(f cr
(k)-(1-R-S)f i(k)+△p-t
w(k)l ++・(32)fvr=β'21・V
fw(k) ten β22・(f cr(k)−(1−R
−S) f i (k) 1△P−f w (k) l −
(33) fir=β21−V fw(k)+β23・
(f cr(k) (1-R-S) f i(k) 10△
P-f w(k)) ”・(34) (However, β11
~β23 is a constant) Therefore, by substituting these equations (32), (33), and (34) into equation (29) above, the control input u
(k), that is, fi(k), is found as shown in the following equation (35).

f i (k) = f 1 - f w (k) + f 2 - f
v(k) tenf 3-f cr(k) tenf i V f
w(k)+f 5・ΔP−f w(k)・−(35
) Furthermore, this formula (35) (friend) The above-mentioned fuel supply amount calculation 1,000 stages M
3, and describes the various multipliers B3, 86, 89, BIO, B14, B15 and adders 816 to 819 in FIG.

In addition, the state variable estimator B71t is for estimating the state variable in the above equation (35), that is, the adhering fuel amount fw and the evaporated fuel amount fv, but in this embodiment, the state variable estimator B7 ) is used as is to estimate the state variables fw and fv. Therefore, in this embodiment, the above equation (8) corresponds to the first calculation equation used in the state variable estimating means M2 described above.

Next, various calculation units 811 to B13 for calculating the variation amount ΔP of the state variable P in the above equation (35), and the above equation (
The fluctuation component Δf is added to the fuel injection amount f calculated based on 35).
The sine wave generating section B20, the amplifying section 821, and the adding section 822 that set the actual fuel injection amount to the internal combustion engine 2 as the amount obtained by adding i will be explained. First, in the fuel behavior model described by equations (1) and (2) above, the amount of evaporated fuel fv and the amount of fuel evaporation VfwLL are significantly smaller than the adhering fuel amount fw and the fuel injection amount fi, and the model parameters R, S
, D have little effect on control accuracy, so these values are ignored. Then, the fuel behavior model can be simplified as shown in the following equations (36) and (37).

f w (k+1) = P - f w (k) + R f i (
k) − (36) f c(k)=(1 −
P)f w(k)+(1 −R)f i(k:l{3
7) Next, in the above equation, the control input is f1, and the control output is f1.
As c, perform two transformations and derive the transfer function of the above model. 2. The conversion method etc. is explained in detail in Masaaki Mitani's Digital Filter Design (Shokodo), so a detailed explanation will be omitted.

First, by converting the above equations (36) and (37) by 2, these equations become as shown in the following equations (38) and (39), and Z F
w(Z) = PFw(Z) +RF i(Z)
-(3B)Fc(Z)”(1-P)Fw(Z)
+(IR)Fi(Z)...(39)(However, Fw(
Z), Fi(Z), Fc(Z)It. (2 converted values of fw, fi, and fc, respectively) From the above equation (38), F w (Z) can be written as shown in the following equation (40).

Fw(Z)=R-Fi(Z)/(Z p)
-(4o) Then, by substituting the above equation (40) into equation (39), F c (Z) = l-1 (Z) - F i (Z
) − (41) Then, the transfer function H (Z) is obtained.

Therefore, as the fuel injection amount fi for actually injecting fuel from the fuel injection valve 32, the normal fuel injection amount fIO calculated by the above equation (35) as shown in the following equation (43)
Considering the known variation fiD, f i= f i
o+f iD −F i(Z) = F io(Z)
+F iD(Z)...(43) On the fuel injection amount fiol Normally, the air-fuel ratio is the stoichiometric air-fuel ratio (1
4.7), so the above calculation formula (3
Assuming that 5) corresponds to the actual characteristics of the internal combustion engine 2, the amount of fuel flowing into the cylinder 2a fc is determined by the known variation fi between the fuel amount fco corresponding to the stoichiometric air-fuel ratio and the fuel injection amount fi.
Due to D, the following equation (44) is obtained, f c= f cO
+ f cD - (44) Formula (41) above (can be written as Tomoji's formula).

F c (Z) = F cO (Z) + F cD (Z) =
H (Z)・F io(Z)+ H (Z)・F i
D(Z) − (45) Therefore, the amount of variation in the amount of fuel flowing into the cylinder fc can be written as the following equation (46), F cD(Z)” H (Z)・F iD(Z)
- (46) The fluctuation amount fcD of the inflow fuel is divided by the known fluctuation amount fiD added to the fuel injection amount {
L The transfer function (Z) of the system can be found as shown in the following equation (47).

H (Z) = F cD (Z) / F iD (Z)
・= (47) On the other hand, the fluctuation amount fcDIt of the fuel flowing into the cylinder 2a, the deviation of the air-fuel ratio A/F from the stoichiometric air-fuel ratio detected by the air-fuel ratio sensor 19, and the amount of air flowing into the cylinder 2a mc. The division (f?) can be calculated as shown in the following equation (48).

(However, Mc(Z) and A/F(7) are two-converted values of the inflow air amount mc and the air-fuel ratio detection result A/F, respectively.) Therefore, in this example, one friend was obtained using the above equation (35). As the known disturbance fiD added to the fuel injection amount fi, the amplitude A
, a sine wave of frequency f is considered, and the fuel injection amount f1 has a predetermined period T [sec. ], the model parameter P is estimated from the resulting variation in the air-fuel ratio, and the variation ΔP of the model parameter P is calculated.

That is, the Z-transformed value F io(z)1 of the known disturbance fiD
When the input signal of the above equation (49) is added to the system of the transfer function H (Z) as shown in the equation (49), the output amplitude, that is, the amount of fluctuation in the inflow fuel, becomes as shown in the following equation (5b). .

l F cD (Z) l= l H (Z) ・F i
D(Z) l:AIH(j 2πfT)1 Therefore, in this embodiment, sine wave generating section B20, amplifying section B
21 and adder B22, a predetermined sinusoidal fluctuation component Δf1 with fixed amplitude A, frequency f, and period T is superimposed on the fuel injection amount fi determined by the above equation (35), and the resulting cylinder 2a is The amplitude of the sinusoidal fluctuation component fcD of the amount of fuel flowing into the cylinder is determined by the fluctuation amount ΔA/ of the air-fuel ratio detection result A/F calculated by the A/F fluctuation amount calculation section 811 in the cylinder inflow fuel fluctuation amount calculation section 812. F and the inflow air amount calculation unit 84
The model parameter variation calculation unit 813 estimates the model parameter P from this value and calculates the variation ΔP of the model parameter P. There is.

Note that this calculation can be performed using an arithmetic expression that is a modification of the above equation (50).
) A map for estimating model parameters P is set up based on the formula, and this map is used to estimate model parameters P. In addition, the above-mentioned amplitude A and frequency f
, the period T is set to a value that allows the model parameter P to be estimated with the highest accuracy, taking into account the model parameters known at the time of design and the characteristics of the air-fuel ratio sensor. (Z) is a low-pass filter, and for example, in a region where the frequency f is low (12), changes in the transfer function H(Z) caused by changes in the model parameter P do not clearly appear as a difference in response, and the estimated value This is because there is a risk that sufficient accuracy may not be obtained.Next, the fuel injection control executed by the electronic control circuit 30 is
This will be explained based on the flowchart shown in FIG. In the following explanation, the quantity handled in the current process is indicated with a subscript (k), and the value obtained in the previous process (i.e., one cycle before internal combustion engine 2) is indicated with a subscript (k-1).
It is expressed with . First, FIG. 4 is a flowchart showing a main routine of fuel injection control that starts when the internal combustion engine 2 is started and is repeatedly executed while the internal combustion engine 2 is in operation.

As shown in the figure, when this process starts, first step 1
Execute 00 and check the amount of attached fuel? w(k-1), amount of evaporated fuel? v(k-1) and the fuel injection amount fi(k-1), and proceed to step 110, where the intake pipe pressure p is determined based on the output signals from each of the sensors.
v(k), intake air temperature THA(k), rotational speed ω(k) of the internal combustion engine 2, and cooling water temperature THW(k).

Next, in step 120 {above} Based on the intake pipe pressure P M (k) obtained in step 110 and the rotational speed ω(k) of the internal combustion engine 2, the target fuel-air ratio λr according to the load of the internal combustion engine 2 is determined. Calculate. Normally, the target fuel-air ratio λr is set so that the excess air ratio is 1 (that is, the stoichiometric air-fuel ratio). In order to increase the output of the internal combustion engine, the target fuel-air ratio λ is set to the rich side.
During light load operation, etc., the target fuel-air ratio λr is set to the lean side in order to reduce fuel consumption compared to normal and improve fuel efficiency. When the target fuel-air ratio λ'(k) is set in step 120, the process moves to step 130, and the step 120
Intake pipe pressure P M (k) and intake air temperature T H determined by
Based on A (k) and the internal combustion engine 20 rotational speed ω(k), the amount of air flowing into the cylinder 2a m c (k) is calculated using the above-mentioned formula (4) or the data map. It executes processing as the amount calculation unit B4.

In addition, in the following step 140, the cooling water temperature THW (k) obtained in the above step 110 and the intake pipe pressure PM (k) are
The amount of fuel that evaporates from the wall surface of the intake pipe 4 during one cycle of the internal combustion engine 2 can be determined by calculating the evaporation rate Vf of the fuel adhering to the wall surface based on the following, and dividing that value by the rotational speed ω(k) of the internal combustion engine 2. (That is, the amount of fuel evaporation) Executes processing as a fuel evaporation rate calculation section B1 and a fuel evaporation amount calculation section B2 that calculate V fw (k). Then, in the next step 150, the fuel evaporation amount Vfw(k) obtained in step 140, the variation amount 8P of the model parameter P obtained in the model parameter variation calculation process described later, and the previous fuel injection amount f i (k-1) and the state variable quantity 7 w(k-1) obtained in the previous step 150
as well as? v(k-1), state variable 1, i.e., the amount of attached fuel? State variable estimation unit B7 that estimates W(k”) and the amount of evaporated fuel?v(k)
Execute processing as .

Next, in step 160, the target fuel air ratio λ'(k) set in step 120 is multiplied by the air amount m(k) obtained in step 130, and the target fuel amount f flowing into the cylinder 2a is calculated. It executes processing as the target fuel amount calculation unit B5 to calculate cr(k) (=λr-mc).

Then, in the following step 170 (above step 140)
~Fuel evaporation amount Vfw(k) obtained in step 160, adhering fuel amount? W(k), amount of evaporated fuel? v(k), the target fuel amount λ"・mc(k), and the variation amount ΔP of the model parameter P calculated in the model parameter variation amount calculation process described later, using the above-mentioned formula (35). Calculate the fuel injection amount fi(k).

Further, in the subsequent step 180, a fluctuation component Δfi set in the model parameter fluctuation amount calculation process described later is added to the IL thus obtained fuel injection amount f i (k), and the fuel is actually injected from the fuel injection valve 32. Fuel injection amount T for
Determine AU.

Then, in step 190, the above crank angle sensor 2
At the fuel injection timing determined based on the detection signal from step 4, the fuel injection valve 32 is opened in accordance with the injection amount TAU determined in step 180, and fuel injection is performed.

In this step 190, fuel injection is performed.
Once the fuel supply to is finished, the process moves to step 200, and the state variable quantity ? w(k),
9 v(k) and fuel injection amount f i(k) are set to 1:, respectively, for the next processing. w(k-1), ? v
(k-1), f i (k-1), and the process returns to step 110.

Next, FIG. 5 is a flowchart showing a model parameter variation calculation process executed as an interrupt process at predetermined time intervals to the main routine shown in FIG. 4 above.

As shown in the figure, in this model parameter variation calculation process, 1
, t. First, execute step 210, and then perform step 8 above.
0, the fluctuation component Δf1 of the fuel injection amount f used when setting the fuel injection amount TAU from the fuel injection valve 32 is expressed by the following formula (
5l) as the sine wave generating section B20 and the amplifying section 821.

Δfi = As i n2πfnT − (51
) (where n: a constant that is incremented each time the model parameter variation calculation process is executed). Then, in the following step 220, the air-fuel ratio A/F(n) is detected based on the output signal from the air-fuel ratio sensor 19. Then, in step 230, the constant n is incremented, and the process moves to step 240.

In step 240, it is determined from the value of the incremented constant n whether a predetermined sampling time has elapsed since the previous step 250 and subsequent steps were executed, that is, in step 230, the air-fuel ratio A/F detection data is determined. Determine whether a predetermined number of samples have been sampled. If it is determined in this step 240 that the predetermined sampling time has not elapsed, the process is immediately terminated, and conversely, if it is determined that the predetermined sampling time has elapsed, the process proceeds to the following step 250, From the plurality of A/F detection data detected in step 230, the A/F fluctuation amount calculation unit 811 calculates the A/F fluctuation amount ΔA/F representing the fluctuation width of the air-fuel ratio. In the next step 260, the cylinder 2a
The inflow fuel fluctuation amount calculation unit B12 calculates the fluctuation amount fcD of the fuel flowing into the interior of the engine, and the process proceeds to step 270, where a preset map is created based on the calculated fluctuation amount fcD. Using the model parameter P
After executing the process as the model parameter variation calculation unit B13 that calculates the variation ΔP of , the process is temporarily terminated.

In addition, calculation of this amount of variation ΔP {Friend As mentioned above 1: Estimate the current model parameter P using a preset map based on the above amount of variation fcD, and compare this estimated value with the value at the time of design. This is done by taking the deviation.

As explained above, in the fuel injection control device of this embodiment (above), the calculation formula (3
5) Perform fuel injection by adding the sinusoidal fluctuation component Δfi to the fuel injection amount fi calculated by 5), and calculate the fluctuation amount fcD of the fuel flowing into the internal combustion engine from the resulting fluctuation amount ΔA/F of the air-fuel ratio. A variation amount ΔP of the model parameter P is calculated from the variation amount fcD, and this variation amount ΔP is fed back to the fuel injection amount calculation system. As a result, even if the fuel behavior model at the time of design no longer corresponds to the actual internal combustion engine due to characteristic fluctuations or variations in the characteristics of the internal combustion engine, it is possible to satisfactorily correct the control errors caused by this, and the The accuracy of fuel ratio control can be improved. In addition, when estimating model parameters, there is no need to use complicated arithmetic expressions as in the devices described in the prior art.
Since the model parameters can be easily estimated using the map, the calculation time can be shortened, and the microcomputer conventionally used in internal combustion engine control devices can be used as is.

Furthermore, since the variation amount ΔA/F of the air-fuel ratio detected by the air-fuel ratio sensor is used to estimate the model parameter P, the model parameter P can be estimated with high accuracy. In short, even if there is an error in the detected value of the air-fuel ratio, the amount of variation can be detected without any problem, making it possible to estimate model parameters with high accuracy, and improving the accuracy of air-fuel ratio control. .

Here, in the above example, the model parameter to be estimated in Table 1 is the parameter P related to the state variable fw in the fuel behavior model, but it may also be the parameter R related to the fuel injection amount fi, and both parameters P and R can be estimated. In this way, control errors due to parameter fluctuations may be corrected.

In addition, when estimating the model parameter R, please refer to Table 1 above (50
) expression or a map based on it,
Although it can be estimated in the same way as in the above example, P,
Table 1 to estimate both model parameters of R (50
) cannot be estimated using the formula alone. Therefore, in this case, the phase delay between the fluctuation component Δlf1 added to the fuel injection amount fi and the air-fuel ratio fluctuation component may be determined and estimated from this phase difference and the fluctuation amount fcD of the fuel flowing into the cylinder.

In other words, when the input of equation (49) above is added to the system of transfer function H (Z), the phase of the output becomes as shown in equation (52), so /FcD(Z)=t a n-' fH (j 2 y
r f T))=tan-' [(R-P+(
1 -R) cos 2yr f Tj(1 -R
)s i n2πfT)−(P −cos2πfT+j
s in2πfTl -- (52) Based on this equation (52) and the above-mentioned equation (50), 2 is calculated using the phase delay of the air-fuel ratio fluctuation component and the amount of fluctuation of the incoming fuel based on the air-fuel ratio fluctuation component as parameters. A dimensional map may be created in advance and the model parameters P and R may be estimated using this map.

In addition, when estimating the model parameters P and R, the phase lag of the air-fuel ratio fluctuation component changes depending on the flow velocity of the exhaust gas up to the air-fuel ratio sensor. For example, the phase shift occurring in the exhaust system may be determined based on the rotational speed and the intake pipe pressure, and the phase delay of the air-fuel ratio fluctuation component may be calculated by taking this calculation result into consideration.

Next, in the above example, the model parameter R in the fuel behavior model is used as a basic equation to derive equation (50) above.
Equations (36) and (37) were used, which ignored , S, and D, but each of these values was fixed, two transformations were performed, and the above equation (50) was derived using the same procedure as above, and the model parameters were estimated. You can also design the system. [Effects of the invention] As explained above, 1:. In the fuel injection amount control device for an internal combustion engine of the present invention, a fluctuation component is given to the fuel injection amount to the internal combustion engine, model parameters are estimated from the resulting fluctuation amount of the air-fuel ratio, and the fuel supply amount is calculated based on the estimation result. It is designed to correct the calculation formula for . Therefore, even if the physical model at the time of design no longer corresponds to the actual internal combustion engine due to changes in the characteristics of the internal combustion engine or variations in characteristics, the control error caused by this can be corrected, and the control accuracy of the air-fuel ratio can be improved. can be improved. Also, the model parameters {
飄 Unlike the devices described in the prior art, it is possible to easily estimate from the amount of fluctuation in the air-fuel ratio without performing complex calculations such as least squares estimation formulas, so the control device is equipped with an expensive, high-speed processing device. The present invention can be fully realized with the microcomputer conventionally used in internal combustion engine control devices without using a microcomputer. Furthermore, model parameter 1 does not directly use the air-fuel ratio detection result, but estimates it based on its variation, so even if an error occurs in the air-fuel ratio detection result due to deterioration of the sensor that detects the air-fuel ratio, the model parameter can be estimated with high accuracy, which also improves the control accuracy of the air-fuel ratio.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing the configuration of the present invention. FIG. 2 is a block diagram showing the internal combustion engine and its peripheral devices according to the embodiment.
Figure 4 is a flowchart representing the main routine for fuel injection control executed by the electronic control circuit, and Figure 5 is a block diagram representing the control law for fuel injection control by the electronic control circuit. This is a flowchart showing the model parameter variation calculation process that is executed every time. M1... Operating state detection means 2... State variable estimation means 3... Fuel supply amount calculation means 4... Fuel supply means 5... Fuel injection amount fluctuation control means 6... Air-fuel ratio detection means 7... Air-fuel ratio variation calculation means 8... Model parameter estimation means 9... Correction means E/G, 2... Internal combustion engine 2.
...Intake pressure sensor 14...Intake temperature sensor 9...
Air-fuel ratio sensor 20...Rotational speed sensor 6...Water temperature sensor 30...Electronic control circuit 2...Fuel injection valve agent Patent attorney Tsutomu Adachi (and 2 others) Drawing part≠

Claims (1)

[Claims] Uses an operating state detection means for detecting a predetermined operating state of the internal combustion engine, and a first calculation formula set based on a physical model that describes the behavior of fuel flowing into the cylinders of the internal combustion engine. and a state variable estimating means for estimating a state variable of the physical model based on the detection result of the driving state detecting means and the previously calculated fuel supply amount, and a second calculation formula set based on the physical model. a fuel supply amount calculation means for calculating a fuel supply amount based on a detection result of the driving state detection means and an estimation result of the state variable estimation means; A fuel injection amount control device for an internal combustion engine, comprising: a fuel supply means for opening a fuel injection valve provided in an intake passage of the fuel injection valve and supplying fuel to the internal combustion engine; a fuel injection amount variation control means for superimposing a predetermined variation component to vary the amount of fuel injected from the fuel injection valve controlled by the fuel supply means; air-fuel ratio detection means for detecting the air-fuel ratio of the fuel mixture; air-fuel ratio variation calculation means for calculating the variation amount of the air-fuel ratio based on the detection result of the air-fuel ratio detection means; a model parameter estimation means for estimating a predetermined model parameter of the physical model based on the amount of variation; and a second calculation used when calculating the fuel supply amount by the fuel supply amount calculation means based on the estimated model parameter. A fuel injection amount control device for an internal combustion engine, comprising: a correction means for correcting the equation.