JPH03217638A - Method for detecting fuel blend ratio - Google Patents

Abstract

PURPOSE:To calculate accurate blend ratio in the case of failure of a blend ratio sensor by calculating control blend ratio through compensating the assumed blend ratio on the basis of the assumed blend ratio and the knock learning value taken in from a memory means. CONSTITUTION:A knock sensor 2 generates electric voltage having a level corresponding to engine vibration as an information of knocking. A knock learning value calculating means 3 obtains frequency of knocking in every periods of renewing a required learning value. When this value is over a criterion value, the criterion value is added to the preceding knock learning value. When this value is under the criterion value, the criterion value is subtracted from the preceding knock learning value. Thus, the succeeding learning value is renewed. A memory means 4 stores the blend ratio and the knock learning value. A failure judging means 8 of the control means 5 takes in the blend ratio, and judges the failure of the blend ratio sensor 1. While failure signal is being input, the blend ratio just before the failure is taken in from the memory means 4. When the knock learning value is over the upper limit of the criterion value, the blend ratio just before the failure is reduced by a definite amount. When the knock learning value is under the lower limit of the criterion value, the blend ratio just before the failure is increased by a definite amount. Consequently, the blend ratio can be calculated accurately.

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to a fuel blend ratio detection method for checking the blend ratio of mixed fuel supplied to an internal combustion engine.

(Prior Art) Methanol has recently attracted attention as a low-pollution fuel, and methanol engines are being developed.

However, it is almost impossible to immediately switch the fuel used in all automobiles from gasoline to methanol, and it is expected that methanol fuel and gasoline fuel will coexist, at least temporarily, during the switching period.

In order to deal with such a situation, it has been proposed to introduce a vehicle (hereinafter simply referred to as FF) which can use either gasoline fuel or methanol fuel, that is, has flexibility in the fuel used.

By the way, this kind of FF is the ignition timing of the engine,
In order to accurately control the amount of fuel injection, etc., the blend ratio, which is the mixture ratio of gasoline and methanol, must always be checked before executing various engine controls. As a blend ratio detection means used in this case, there is a blend ratio sensor that is installed directly opposite to the fuel supply system and can directly detect the blend ratio, and this has been researched and developed and is in use. ) However, if the blend rate sensor is broken and the blend rate deviates significantly from the actual value, if this value is used to control the engine, the engine will malfunction. In particular, ignition timing control in an FFV engine has a wider control range than a normal gasoline engine, which can cause damage to the engine, which is a problem.

An object of the present invention is to provide a fuel blend ratio detection method that can accurately calculate the friend ratio even when the blend ratio sensor is out of order.

(Means for Solving the Problems) In order to achieve the above-mentioned object, the present invention includes a blend ratio sensor that outputs blend ratio information of gasoline and methanol in fuel supplied to an internal combustion engine, and a knock sensor configured to read knock information; a knock learned value calculation means to calculate a knock learned value based on the knock information; a storage device in which the blend rate and the knock learned value are sequentially stored; and the blend rate. and a control means for calculating a blend rate from the knock learning value, and when the blend rate sensor fails, the control means calculates the knock learning value and the assumed blend rate detected immediately before the failure from the above-mentioned general means. In addition, if the above-mentioned knock learning value is above the specified upper limit, the above-mentioned assumed blending rate is lowered by a specified amount,
If the knock learning value is less than a specified lower limit value, the assumed blend rate is increased by a specified amount to calculate the control blend rate.

(Function) When the blend rate sensor is normal, the output blend rate is used as the control blend rate, and in the event of a failure, the assumed blend rate and knock learning value from the storage means are taken in, and when the knock learning value is greater than the specified upper limit, the blend rate is assumed. Reduce the blending rate by a specified amount,
If the knock learning value is less than the specified lower limit value, the assumed blend rate is increased by a specified amount to become the control blend rate, so even in the event of a failure, the blend rate can be calculated accurately based on the knock information.

(Example) Hereinafter, a fuel blend ratio checking method according to the present invention will be explained.

In this method, as shown in FIG. 1, a blend ratio sensor 1 outputs a blend ratio B5 of a mixed fuel of gasoline and methanol, a knock sensor 2 outputs knock information N3 of the internal combustion engine, and a knock sensor 2 that outputs knock information N3 of the internal combustion engine. , the knock learning value K. 1, a knock learning value calculation means 3 for calculating the blend ratio B3 and the knock learning constant K, . A storage means 4 for sequentially storing and processing 1 and a blending rate? ．． 3 and knock learning value K8,
Control means 5 for calculating and outputting the control blend ratio from 1
It is used.

The blend ratio sensor 1 here has a well-known configuration in which refractive index information that changes depending on the blend ratio of fuel is checked by an optical system, and the change in light amount is photoelectrically converted and output.

The knock sensor 2 includes a weight that presses a piezoelectric element in accordance with engine vibration, and is configured to generate a voltage as knock information N, at a level corresponding to the vibration.

The knock learning value calculation means 3 performs learning when the engine operating range is in a set learning range (the shaded area in FIG. 2, where a high load range where knocking occurs frequently is selected). Here, the frequency of knocking is calculated every update period of the predetermined learned value, and if the same frequency exceeds the specified value, the specified value 10G is added to the previous knocking learned value K1(t-1) and the learned value. , if the same frequency is lower than the specified value, the previous learning value K, .
．． ．． , (t-1) minus the specified value -61 to get the current learning f+I! K. ．． 、． (t) is updated 7. The storage means 4 stores the blend rate BS and knock learning value KK,,1
Equipped with an area to store.

The failure determination means 6 of the control means 5 first takes in the blend ratio B5, determines whether the blend ratio sensor 1 is malfunctioning based on this value, the previous value, etc., and outputs a failure signal a. Further, the blend rate calculation means 7 selects the blend rate B5 from the blend rate sensor as the control blend rate B and outputs it while the failure signal a from the failure determination means 6 is not input.

On the other hand, while the failure signal a is being input, the blend rate B, (t-1) stored in the storage means 4 immediately before the failure is taken in instead of the output from the blend rate sensor, and the value is Knock learning value K. , , , and outputs it as the control blend rate B. here,
If the knock learning value K11 is equal to or higher than the specified upper limit, that is, knocking occurs frequently, so to lower this value, the blend ratio B9 (t-1) immediately before the failure is lowered by a specified amount, and the knock learning value is set to the specified lower limit. In the following, the blend width p, (t-1) immediately before the failure is increased by a specified amount in order to correct the occurrence of knock failure to the side where knock occurs frequently.Next, the fuel blend ratio detection method of the present invention will be described. An engine control system for an FFV vehicle employing the following will be explained with reference to FIG.

Here, the combustion chamber 11 of the engine 10 is communicated with the intake passage 12 and the exhaust passage 13 in a timely manner. The intake passage 12 includes an air cleaner l4, a first intake pipe 15, an expansion pipe 16, and a second intake pipe l7.
The exhaust path 13 is formed by a first exhaust pipe 18, a catalyst l
9, a second exhaust pipe 20, and a muffler 21.

Inside the air cleaner 14, an air flow sensor 22 that outputs passing air amount information, an atmospheric pressure sensor 23 that outputs atmospheric pressure information, and an atmospheric temperature sensor 24 that outputs air temperature information are arranged, and these are connected to an engine control unit (hereinafter simply referred to as controller) 25.

A throttle valve 26 is installed in the expansion pipe l6, and a throttle position sensor 27 is installed opposite to the valve.Moreover, this throttle valve 26 has its idle position determined by an idle speed control motor (ISC motor) 28.
It is configured to be controlled by the controller 251 via the controller 251.

A water jacket is provided opposite to a part of the second intake pipe l7, and a water temperature sensor 29 is attached thereto.

An 02 sensor 30 is installed in the middle of the first exhaust pipe 18 to output information on oxygen concentration in the exhaust gas.

Furthermore, a fuel injection valve 31 is attached to the end of the intake passage 12. This fuel injection valve 31 is connected to a fuel pipe 33 via a branch pipe 32. This fuel pipe 33 connects a fuel pump 34 and a fuel tank 35, and a blend rate sensor 43 and a fuel pressure regulator 36 for adjusting fuel pressure are attached to the middle of the pipe. Furthermore, the ejector 36 here is configured to be able to increase or decrease the fuel pressure by 5g depending on the boost pressure.

An ignition plug 46 is attached to the combustion chamber 11 of the carrot 10, and the plug 46 is connected to an ignition circuit 45 comprising a power transistor (not shown) and an ignition coil driven to open and close by the power transistor. This ignition circuit 45 is in contact with an ignition drive circuit 44, which will be described later, and is omitted. Furthermore, a knock sensor 47 is attached near the combustion chamber 11 of the engine, and its knock information N, is output to the controller 25.

In FIG. 4, reference numeral 37 indicates a crank angle sensor that outputs crank information, and reference numeral 38 indicates a top dead center sensor that outputs top dead center information of the first cylinder.

The controller 25 includes a control circuit 39, a memory circuit 40, a bath input circuit 41, and drive circuits 42 and 44.

Here, the control circuit 39 receives each input signal from each sensor, processes these in accordance with a predetermined control program, and outputs a control signal.

The memory circuit 40 stores various control programs such as the engine control main routine shown in FIG. 5, the blend ratio calculation routine shown in FIG. 6, the ignition timing calculation routine (not shown), and the fuel injection routine, and each control value calculation map. It is configured to store and process the data, and also includes an area for capturing correction coefficients, calculated data, and other values used during control.

The Ohata force circuit 41 operates to appropriately receive the output signals of the above-mentioned sensors, and outputs a valve drive signal to open the fuel injection valve 31 at a predetermined time via the valve drive circuit 42, or outputs a valve drive signal to open the fuel injection valve 31 at a predetermined time. The ignition signal is outputted via the drive circuit 44, and the other control signals are outputted via a drive circuit (not shown).

Here, the operation of the controller 25 will be explained together with the control programs shown in FIGS. 5 and 6.

When a key switch (not shown) of the engine is turned on, the controller and each sensor start driving.

First, the controller 25 maintains each set value, fixed value, etc. at the initial value and enters a blend ratio calculation routine in step a2.

In the blend rate calculation routine, if the failure flag of the blend rate sensor 43 is on or not, +P1 is determined. B (t
) is stored in the memo JNO as the previous control blend rate B(t-1) and return.If it is determined in step bl that the blend rate sensor 43 is malfunctioning, the process advances to step b4. Here, the previous control blend rate B (t-1) is loaded from the memory, and the knock learning value X-N+(t) is loaded from the memory.

This knock learning value K K N I is sequentially determined by executing a knock learning value calculation routine (not shown) each time the engine operating range enters a predetermined learning range (see FIG. 2). In this knock learning value calculation process, for example, if the knock retard control amount θ,(t) is in the dead band from 1.1 to 1.8 as shown in FIG. If you use it without changing it and it stays at 1.8 or more for more than a long time, the knock learning value K. , 1 by a negative specified value -68, and if it is less than 1.1 for more than a time, it is configured to repeat the calculation of adding the knock learning value KKll+ by a positive specified value 10 GK. .

In step b6, it is determined whether the current knock learning value K,+(1) is greater than or equal to the prescribed maximum knock learning permissible 1 shift K to 8, and if the value is above, the process proceeds to step b7, and if it is below, the process proceeds to step b9.

In step b7, assume that there is one frequent occurrence area, and lower the blend rate to suppress this. That is, the previous control blend rate B (t-1) is corrected by subtracting the decreasing side blend rate correction gain Δ8-, and the process proceeds to step b8, where the knock learning value K. 1. is reflected in the blend ratio, the knock learning value KKN+(j) is cleared, and the process proceeds to step b3.

On the other hand, in step b9, the current knock learning value KKNI(t)
is less than or equal to a specified minimum knock learning allowable value KMIN.
Nine steps forward.

At step bll, the current knock learning value K. 、． l (t
) is in the dead zone, the previous control blend rate B (
t-1) is used as the current blending ratio B (t), and the process proceeds to step b3.

On the other hand, step because it is a knock failure area! : Proceeding from l9 to blO, here the blending rate is increased in order to move to the knock generation side. That is, the blend rate correction gain ΔB is increased from the previous control blend rate B (t-1).
+ is added to correct it, and the process proceeds to step h8. Here, assuming that the knock learning value K K 11 1 is reflected in the end of the blend, the knock learning value K -= + (t) is added and the process proceeds to step b3. move on.

When the blend ratio calculation routine is finished and the process returns to step a3 of the main routine, the engine rotational speed etc. are taken in, and this is the engine operation determination rotational speed N! ST. It is determined whether or not the value exceeds 2.

When step a4 is reached during engine rotation, the control blend ratio B (t) and various correction coefficients are appropriately taken in, fuel injection amount control processing, ignition timing control processing, and other various controls are performed, and step a5 is completed.

Here, for example, the fuel injection valve drive time T which is the fuel injection amount
1. In calculating 1, first calculate the basic drive time T, (=A/N(n)xK,) per intake air amount. The blend rate correction coefficient K8 here is the predetermined intake air amount A/N (
n) It is used to convert the basic driving time T, (basic amount of fuel) into an equivalent amount of gasoline. Furthermore, the fuel injection valve drive time T + *I is defined as the basic drive time T. , feedback correction coefficient KFF and atmospheric temperature correction coefficient Kt
．． Atmospheric pressure correction section: rT, . ．．
,=r,xK. . XKtXKbXKvtXKac):

When step a5 is reached, it is determined whether the key is off or not, and unless the key is off, the process returns to step a2, and when the key is off, each main process at the time of the key off, such as storing each data in a non-volatile memory, is performed. and exit.

When the engine is stopped from step a3 and step a7 is reached, the process waits for the starter switch to be turned on, and while the starter switch is off, step a8 is reached. Here, a predetermined process associated with stopping the engine is performed, and when the switch is turned on, the process proceeds to step a9. In step a9, various processes associated with starting are performed, and the process proceeds to step a5.

(Effects of the Invention) As described above, in the method of the present invention, when the blend rate sensor is normal, the output blend rate is used as the control blend rate, and when the blend rate sensor is malfunctioning, the assumed blend rate and the knock learning value from the storage means are taken in, and the knock Since the control blend rate is calculated by correcting the assumed blend position according to the learned value, an accurate blend rate can be calculated even in the event of a failure.

[Brief explanation of drawings]

Fig. 1 is a block diagram explaining the method of the present invention, and Fig. 2 is an explanatory diagram of the knock learning area used in the method of the present invention. Figure 3 (a)
, (b) are characteristic diagrams of the update time scale and update prescribed values used in the calculation of the knock learning value used in the method of the present invention, and FIG. 4 is a schematic configuration diagram of an engine control device adopting the method of the present invention. 5 and 6 show a flowchart of a control program used in the carrot control process performed by the engine control device of FIG. 4. 1,43...Blend rate sensor, 2,47...Knock sensor, 3...Knock learning value calculation means, 4,40 Storage means. 5.25... Control means, B (t)... Control blending, K. , I(t)...Knock learning value, κ
'- I N...Minimum knock learning tolerance value, K. 88・
...Maximum knock learning allowable value K, 1■,, ΔB+...Increasing blend rate correction gain, ΔB-...Decreasing blend rate correction gain. Most Z-eye 1 Phantom ink 0 Mouth

Claims (1)

[Claims] A blend ratio sensor that outputs blend ratio information of gasoline and methanol in fuel supplied to the internal combustion engine, a knock sensor that outputs knock information of the internal combustion engine, and a knock learning that calculates a knock learning value based on the knock information. using a value calculation means, a storage means for sequentially storing and processing the blend ratio and the knock learning value, and a control means for calculating the blend ratio from the blend ratio and the knock learning value,
When the blend rate sensor fails, the control means takes in the assumed blend rate detected immediately before the failure and the knock learned value from the storage means, and if the knock learned value is equal to or higher than the specified upper limit, A method for detecting a fuel blend rate, characterized in that the assumed blend rate is decreased by a prescribed amount, and when the knock learning value is below a prescribed lower limit value, the assumed blend rate is increased by a prescribed amount and calculated as a control blend rate.