JPH08284713A - Evaporative fuel processing device for engine - Google Patents

Abstract

PURPOSE: To avoid emission of fuel and influence to the vehicle operation property when there is a divergence between the actual fuel concentration and a learning value at the time of starting an engine by suppressing the purge gas flow to smaller value than normal until the divergence is corrected. CONSTITUTION: When an engine 1 is started and a purge ON granting terms is concluded while driving thereafter, a purge control valve 18 is opened by a control unit 7 and the engine 1 inlet negative pressure works on a canister 14. Therefore, evaporative fuel sucked in a suction material 14a of the canister 14 by air from a new air induction opening 16 come off and purge gas containing evaporative fuel passes through a purge passage 17 and is fed into an inlet manifold 4. At this time, processing degree of fuel density learning is judged and such control as controlling the purge gas flow to initial purge gas flow of a quantity less than the normal purge gas flow until the fuel density learning processes after starting the engine and switching over to a normal purge gas flow after processing the fuel density learning is carried out.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention collects vaporized fuel in a canister so that the vaporized fuel generated in a fuel tank is not released into the atmosphere, and the vaporized fuel is sucked into the engine during engine operation for combustion processing. The present invention relates to an evaporated fuel processing device for an engine.

[0002]

2. Description of the Related Art Conventionally, as an evaporated fuel processing apparatus for an engine, a canister for adsorbing evaporated fuel from a fuel tank and a purge gas containing evaporated fuel desorbed by introducing fresh air into the canister are used for the engine. A purge control valve provided in a purge passage leading downstream of the throttle valve in the intake passage is provided, and the opening of the purge control valve is set to, for example, a purge rate (ratio of the purge gas flow rate to the intake air flow rate) according to the engine operating conditions. It is known that the control is performed so as to be constant.

Further, as in Japanese Patent Laid-Open No. 63-297757, the amount of evaporated fuel generated during the stop is predicted based on the internal pressure of the fuel tank when the engine is stopped, and accordingly, the purge control valve is increased accordingly at the next start. There is a correction to reduce the opening degree of. Further, as in Japanese Patent Laid-Open No. 63-111277, it is assumed that the canister filling rate is high when refueling is performed, and the purge rate is reduced for a while after the start. Further, there is a method in which the fuel injection amount is subtracted instead of the purge rate (JP-A 1-211661).

[0004]

However, such a conventional evaporated fuel processing apparatus for an engine has the following problems. Considering the prior art of Japanese Patent Laid-Open No. 63-297757, the fuel tank is not hermetically sealed, but a pipe is provided and connected to the canister, and the generated fuel vapor is adsorbed by the canister through the pipe. Therefore, the internal pressure does not actually rise to the vapor pressure of gasoline at that temperature. If a check valve is provided in the line, the valve opening pressure will rise, but it is usually about several mmHg.

Therefore, it is practically impossible to predict the amount of evaporated fuel generated from the internal pressure of the fuel tank when the engine is stopped. Even if the fuel tank is sealed and the vapor pressure at that temperature is measured, if it is sealed, the vaporized fuel does not come out, so the amount generated is zero,
It contradicts the original meaning. Even if the fresh air inlet of the canister is closed and the purge control valve is closed, the evaporated fuel from the fuel tank is adsorbed by the canister until the canister is full, so the internal pressure of the fuel tank also rises. However, even if the generated amount is predicted from the internal pressure that rises after it becomes full, the canister is already full (that is, the adsorption amount is constant) at that time, and the amount generated after that can not be further adsorbed There is no point in doing it.

In addition to JP-A-63-297757,
Conventional controls such as Japanese Patent Laid-Open No. 63-111277 reduce the purge rate in order to prevent a rich error on the assumption that the purge gas is rich. Because the purge amount is limited, the required purge amount cannot be earned when the cister is full and it is desired to purge a lot in a short time, or when the purge frequency is low due to repeated short-time operation (start → stop). However, there is a problem that the evaporated fuel easily leaks to the outside of the fuel tank because it is accumulated in the canister.

In the actual use environment, when a large amount of evaporated fuel is generated due to midsummer or continuous parking for a long time, when the purge concentration is changed to 0 or thin due to disconnection or clogging of the purge pipe, or there is a failure. It is conceivable that the canister is replaced with a new one and the purge concentration becomes 0 due to the above circumstances, and the conventional technology is not sufficient to deal with any of these.

Further, this point has a learning value of the fuel concentration in the purge gas, updates the concentration learning value from the change of the air-fuel ratio feedback correction coefficient before and after the purge, and actually inhales from the obtained concentration learning value. The same applies to the control method in which the purged fuel flow rate and the purged air flow rate are calculated, subtracted from the fuel injection amount, and added to the intake air flow rate.

This is because the concentration learning value that was updated and held when the engine was stopped last time is corrected at the time of the first purge after the next start even if the concentration learning value is backed up, even if the concentration change value is stopped. Must be,
This is because the same problem remains with respect to "how many initial values should be started?".

The present invention has been made by paying attention to such a conventional problem. In the first purge after starting, the purge rate is reduced, and even if there is a deviation, the air-fuel ratio error is small in the purge gas. By re-learning the fuel concentration of and re-learning (correction of the change while the engine is stopped), it is possible to quickly shift to a large amount of purge, thereby minimizing the air-fuel ratio error after engine start. The goal is to enable large volume purges as soon as possible. .

[0011]

Therefore, in the invention according to claim 1, a fuel injection valve for injecting and supplying fuel to the engine is provided, and a canister for adsorbing the evaporated fuel from the fuel tank, and a new canister for this canister. The purge control valve for controlling the purge gas flow rate, which is interposed in the purge passage that guides the purge gas containing the evaporated fuel desorbed by introducing air to the downstream side of the throttle valve of the engine intake passage, and the opening degree of this purge control valve In an evaporative fuel treatment system for an engine, which is provided with a purge control means for controlling the purge gas flow rate, the following means is provided as shown in FIG.

That is, basic fuel injection amount calculation means for calculating the basic fuel injection amount based on the engine operating conditions,
Air-fuel ratio feedback correction coefficient setting means for setting an air-fuel ratio feedback correction coefficient for correcting the fuel injection amount based on a signal from an air-fuel ratio sensor provided in the exhaust passage, and a purge gas based on a change in the air-fuel ratio feedback correction coefficient. Purge fuel concentration learning means for learning the inside fuel concentration, purge fuel flow rate calculation means for calculating the purge fuel flow rate based on the purge gas flow rate and the learned value of the fuel concentration by the purge control means, and purging from the basic fuel injection amount A purge fuel amount subtraction basic fuel injection amount calculation means for calculating the basic fuel injection amount by subtracting the fuel flow amount
Fuel injection amount calculation means for calculating the fuel injection amount by the fuel injection valve based on the purge fuel amount subtraction basic fuel injection amount and the air-fuel ratio feedback correction coefficient.

Further, a normal purge gas flow rate setting means for setting a normal purge gas flow rate based on the operating conditions of the engine, and an initial purge gas flow rate for setting the initial purge gas flow rate smaller than the normal purge gas flow rate for the first time after the engine is started. A flow rate setting means, a learning progress degree determining means for determining a progress degree of the fuel concentration learning by the purge fuel concentration learning means, and a purge gas flow rate controlled by the purge control means from the engine start until the fuel concentration learning progresses. The initial purge gas flow rate set by the initial purge gas flow rate setting means is provided, and a switching means for switching to the normal purge gas flow rate set by the normal purge gas flow rate setting means after the progress of the fuel concentration learning is provided.

In the invention according to claim 2, the normal purge gas flow rate setting means sets a target purge rate as a ratio of the purge gas flow rate to the intake air flow rate so that the purge gas flow rate corresponds to the intake air flow rate. It is a rate setting means, and the initial purge gas flow rate setting means is an initial purge rate setting means for setting an initial purge rate smaller than the target purge rate.

According to the third aspect of the present invention, the learning progress degree determining means determines the degree of progress of learning by whether or not the update increasing / decreasing direction of the learning value of the fuel concentration is reversed a predetermined number of times or more. Characterize. The invention according to claim 4 is characterized in that learning initializing means for initializing the learned value of the fuel concentration by the purged fuel concentration learning means each time the engine is started (shown by a dotted line in FIG. 1).

[0016]

In the invention according to claim 1, the basic fuel injection amount calculation means calculates the basic fuel injection amount based on the operating conditions of the engine. Further, the air-fuel ratio feedback correction coefficient setting means sets the air-fuel ratio feedback correction coefficient based on the signal from the air-fuel ratio sensor.

Here, by the purge fuel concentration learning means,
The fuel concentration in the purge gas is learned based on the change in the air-fuel ratio feedback correction coefficient. Specifically, the concentration is learned by updating the learning value so that the air-fuel ratio feedback correction coefficient before and after the purge becomes the same. Then, the purge fuel flow rate calculation means calculates the purge fuel flow rate based on the purge gas flow rate and the learned value of the fuel concentration by the purge control means, and the purge fuel amount subtraction basic fuel injection amount calculation means calculates from the basic fuel injection amount. The purge fuel flow rate is subtracted to calculate the purge fuel subtraction basic fuel injection amount.

Then, the final fuel injection amount is calculated by the fuel injection amount calculation means based on the purge fuel amount subtraction basic fuel injection amount and the air-fuel ratio feedback correction coefficient.
On the other hand, the normal purge gas flow rate setting means sets the normal purge gas flow rate based on the operating conditions of the engine, but the initial purge gas flow rate setting means sets the initial purge gas flow rate smaller than the normal purge gas flow rate for the first time after the engine is started. To set. The initial purge gas flow rate is such a value that even if there is a deviation between the actual fuel concentration and the learned value, the effect on emissions and drivability will be small.

Thus, at the first purge after the engine is started, the purge gas flow rate controlled by the purge control means is set to the initial purge gas flow rate. That is, at the first purge after engine start, the purge gas flow rate is reduced, and even if there is a deviation between the actual fuel concentration and the learned value, the purge is performed under the condition that the air-fuel ratio error is small. The fuel concentration in the purge gas is learned by the learning means.

Then, the degree of progress of the fuel concentration learning is determined by the learning progress degree determining means, and when the fuel concentration learning progresses, the switching means switches the purge gas flow rate controlled by the purge control means to the normal purge gas flow rate. . As described above, as soon as the learning of the fuel concentration (correction of the change while the engine is stopped) is completed, the flow rate is switched to the normal purge gas flow rate so that a large amount of purge can be performed quickly.

Therefore, even if there is a deviation between the actual fuel concentration and the learned value learned last time when the engine is started, the purge gas flow rate can be kept smaller than the normal value until the deviation is corrected, so that emissions and drivability are improved. The effect on fuel consumption can be suppressed, and if the deviation of the fuel concentration is corrected,
Since the normal mass purge is rapidly performed, it is possible to prevent emission and deterioration of drivability from deteriorating and to mass purge at the same time.

According to the second aspect of the present invention, the purge rate is controlled so that the purge gas flow rate corresponds to the intake air flow rate, and the initial purge rate smaller than the normal target purge rate is set at the first purge after starting. By doing so, the purge gas flow rate can be made to correspond to the intake air flow rate also at this time. According to the third aspect of the present invention, the convergence state of the learning can be accurately grasped by determining the degree of progress of the learning depending on whether or not the update increasing / decreasing direction of the learning value of the fuel concentration has been reversed a predetermined number of times or more.

In the invention according to claim 4, the learning initializing means initializes the learning value of the fuel concentration every time the engine is started, and the re-learning of the fuel concentration always starts from a predetermined value.
It is also possible to suppress variations in learning speed and the like, although memory backup is unnecessary.

[0024]

EXAMPLE An example of the present invention will be described below. FIG. 2 shows the system configuration. The air introduced from the air cleaner 2 and controlled by the throttle valve 3 is sucked into the engine 1 through the intake manifold 4. The intake manifold 4 is provided with a fuel injection valve 5 for each cylinder,
An amount of fuel corresponding to the intake air is injected and supplied. As a result, the air-fuel mixture is generated and burned in each cylinder of the engine 1, and the exhaust gas is introduced into the three-way catalyst (not shown) through the exhaust passage 6.

The fuel injection valve 5 is energized and opened by a drive pulse signal output from the control unit 7 having a built-in microcomputer in synchronization with the rotation of the engine 1 at a predetermined timing, and deenergized and closed. The fuel injection amount is controlled by the pulse width of the drive pulse signal. Control unit 7
Signals are input to the air flow meter 8, the crank angle sensor 9, and the air-fuel ratio sensor (O ₂ sensor) 10.
The air flow meter 8 is provided upstream of the throttle valve 3 in the intake passage and detects the intake air flow rate QA (intake air flow rate detecting means). The crank angle sensor 9 has a predetermined crank angle (720 ° / number of cylinders) in synchronization with the rotation of the engine 1.
The reference signal REF is output for each
The engine speed NE can be detected from the cycle. O ₂
The sensor 10 is provided in the exhaust passage 6, and detects the air-fuel ratio (rich / lean) of the air-fuel mixture supplied to the engine 1 from the composition of the exhaust (O ₂ concentration). In addition to these, signals such as the throttle valve opening TVO and the vehicle speed VSP are input to the control unit 7 from various sensors, and ON / OFF signals of various switches are also input.

An auxiliary air passage 11 for bypassing the throttle valve 3 is provided in the intake passage.
An idle control valve 12 for controlling the idle speed of which duty is controlled is interposed in the valve 11. This idle control valve
12 is also controlled by the control unit 7. on the other hand,
A canister 14 is provided to process the evaporated fuel generated from the fuel tank 13.

The canister 14 is a container filled with an adsorbent 14a such as activated carbon, and a fuel tank is provided on the upper space side.
Evaporative fuel introduction pipe 15 from 13 is connected. Therefore,
The evaporated fuel generated in the fuel tank 13 while the engine 1 is stopped is guided to the canister 14 through the evaporated fuel introducing pipe 15 and adsorbed there. The canister 14 also has a fresh air inlet 16 formed on the lower space side and a purge passage 17 extending from the upper space side. The purge passage 17 is connected via a purge control valve 18 to the intake passage downstream of the throttle valve 3 (intake manifold 4).

The purge control valve 18 is opened by a signal output from the control unit 7 under a predetermined condition while the engine 1 is operating, and the opening degree is duty-controlled. Therefore, when the engine 1 is started and the purge ON permission condition is satisfied during the subsequent operation, the purge control valve 18 is opened and the suction negative pressure of the engine 1 is reduced to the canister 14.
As a result, the evaporated fuel adsorbed on the adsorbent 14a of the canister 14 is desorbed by the air introduced from the fresh air introduction port 16, and the purge gas containing the desorbed evaporated fuel passes through the purge passage 17. It is sucked into the intake manifold 4 and thereafter burned in the cylinder of the engine 1.

The control of the purge control valve 18 by the control unit 7, the fuel injection valve 5 and the idle control valve 12
The control is performed according to the flowcharts shown in FIGS. 3 to 12, and will be described below with reference to these flowcharts. FIG. 3 is a job for calculating the purge rate (ratio of the flow rate of the purge gas passing through the purge control valve to the intake air flow rate QA) and determining the purge ON permission condition, and is executed as a background job.

First, in step 101, various sensors (air flow meter, O ₂ sensor, throttle sensor, etc.) are set to N.
If there is an NG sensor, the purge rate PAGERT = NGPGRT # is checked in step 102.
Determined as (constant rate). If the sensor is not NG, the routine proceeds to step 103, where the subroutine branches to FIG. 5 for calculating the purge rate.

In step 201 of FIG. 5, the target purge rate T
Calculate PAGERT. Target purge rate TPAGERT
Is a ratio of the target purge gas flow rate to the intake air flow rate QA and is set to a fixed value, or as shown in FIG. 13, the higher the concentration becomes in accordance with the concentration learning value (learning value of the fuel concentration in the purge gas) WC. To set. Next step
In 202, it is determined whether the fuel concentration learning after the engine has started, that is, whether the correction of the difference between the actual fuel concentration in the purge gas after the engine start and the concentration learned value (deviation of the initial concentration) has been completed. If #FSTWCOK is checked and the correction is not completed (when # FSTWCOK = 0), the process proceeds to step 203.

In step 203, the purge rate PAGERT
Is not the target purge rate TPAGERT but an initial purge rate (fixed value) smaller than the target purge rate set for the initial purge after the engine is started. This is because we want to correct (re-learn) the concentration change (deviation from the learning hold value) that occurs due to the generation of evaporated fuel while the engine is stopped, the replacement of the canister, etc. This is because the value is set to a value lower than the target purge rate that is originally desired to be flown in order to carry out under the condition that the fuel ratio and drivability are not affected.

The initial purge rate may not be a fixed value but may be calculated as the target purge rate * GAIN #. Here, GAIN # is a correction factor of 1 or less. Even during this time,
The density learning value WC is learned / updated by a rotation synchronization job of FIGS. 6 to 7 described later. Then step 204
Then, it is checked whether the update increase / decrease direction of the density learning value WC has been inverted three times or more. This is to determine whether the correction of the deviation between the actual fuel concentration and the concentration learning value has been completed. If the update and increase / decrease directions of the concentration learning value are reversed in the purge and the concentration learning performed after the start, It is judged that it has caught up with the actual value. The reason that the number of determinations is 3 here is originally 1
Although it may be repeated, considering the overshoot of control,
This is to increase the accuracy.

If it has not been inverted three times or more, this subroutine is terminated as it is, but if it has been inverted three times or more,
It is judged that the fuel concentration learning has progressed and the correction of the "deviation of the initial concentration" has been completed (the learned value has caught up with the actual concentration),
Go to step 205. In step 205, the purge rate PA
GERT = target purge rate TPAGERT. That is, the purge rate, which has been kept low until now, is switched to the target purge rate.

Next, at step 206, the fuel concentration learning progresses and a flag #FSTWCOK indicating that the "deviation of the initial concentration" has been corrected is set (# FSTWCOK =
1) End this subroutine. Go back and step
If the correction of “deviation in initial density” has been completed in the determination of 202 (when # FSTWCOK = 1), step 207
Purging rate PAGERT = target purge rate TPA
This subroutine is ended as GERT. After that, this operation is repeated.

As described above, in this subroutine, the actual fuel concentration in the purge gas and the concentration learning value deviate due to the evaporated fuel generated while the engine is stopped, so that the emission and operability are improved at the first purge after the engine is started. In order to avoid the influence of the above, the first purge after starting is always kept to a low purge rate (initial purge rate), the deviation of the concentration is corrected (re-learning) under the condition that the influence is small, and it is prompt after the correction is completed. Since it is configured to shift to the large-scale purge that should be flowed originally, even if there is a deviation between the actual concentration and the learning value at the first barge after starting, the purge gas flow rate will be normal until the deviation is corrected. Since the value is suppressed to be smaller than the value, it is possible to prevent deterioration of drivability and emission and to secure a large amount of purge at the same time.

Here, this subroutine corresponds to the purge control means, and particularly the steps 201 and 207 correspond to the normal purge gas flow rate setting means (target purge rate setting means),
The step 203 corresponds to the initial purge gas flow rate setting means (initial purge rate setting means), the steps 204 and 206 correspond to the learning progress degree determining means, and the step 202 corresponds to the switching means.

Returning to FIG. 3, the description will be continued. Step 104
At 112, the purge ON permission condition is determined. That is, in steps 104 to 108, it is determined whether the ignition switch is OFF, the engine is stalled, the starter switch is ON, the engine is idle (idle switch ON), and the vehicle speed VSP is low.
Proceeding to FIG. 4 (step 125), purging is prohibited.

Further, in step 109, referring to the characteristic table shown in FIG. 14, the engine speed NE is used to determine the purge permission lower limit basic fuel injection amount (purge permission lower limit TP) TPC.
Ask for a PC. Then, in step 110, the actual basic fuel injection amount TP and the purge permission lower limit basic fuel injection amount TPCPC
When TP <TPCPC (the load is too small), purging is similarly prohibited.

Further, in step 111, the intake air flow rate QH0 obtained from the throttle valve opening TVO and the engine speed NE is compared with a predetermined upper limit value EVPCGH to obtain Q
If H0> EVPCGH (the load is too large), the purging is similarly prohibited. Further, in step 112, the air-fuel ratio feedback control (Ramcon) checks whether or not clamping is in progress. If clamping is in progress, purging is similarly prohibited.

If the determinations at steps 104 to 112 are all NO, it is determined that the purge ON permission condition is satisfied, and the process proceeds to FIG. 4 (step 113) to permit purge ON. FIG. 4 is a job executed subsequently to FIG. 3, in which the purge gas flow rate passing through the purge control valve (herein referred to as the purge valve flow rate).
Further, the ON duty of the purge control valve (herein referred to as the purge valve duty) is calculated and the concentration learning is permitted.

The case where the purge ON permission condition is satisfied and the case where the purge ON permission condition is not satisfied will be separately described. [When the purge ON permission condition is satisfied] Step 11
At 3, allow purge ON. Then in step 114, the diagram
The differential pressure correction factor KPVQH of the purge valve flow rate is obtained from the intake air flow rate QH0 obtained from the throttle valve opening TVO and the engine speed NE by referring to the characteristic table shown in FIG. The differential pressure correction rate KPVQH is a correction rate for changing the flow rate due to the front-back differential pressure even if the purge valve flow passage area is constant.

It should be noted that, in terms of phase during transition, the basic fuel injection amount TP is closer to the purge valve differential pressure than QH0, and it is better to obtain it using TP. However, TP is the atmospheric pressure, the intake valve temperature, and the purge valve differential pressure. QH0 is used here because the relationship with Next, at step 115, the voltage correction rate KPVVB of the purge valve flow rate is obtained from the battery voltage VB by referring to the characteristic table shown in FIG. This voltage correction factor KPVVB is a correction factor for the purge valve duty / flow rate relationship that varies depending on the purge valve applied voltage, and of course varies depending on the valve type.

Next, at step 116, the target purge valve flow rate QPV is calculated by the following equation. QPV = (QA * PAGERT) / (KPVQH * KP
VQH) That is, the intake air flow rate QA is multiplied by the purge rate PAGERT, and the differential pressure correction rate KPVQH and the voltage correction rate KPV are added.
It is corrected by QH to obtain the purge valve flow rate QPV. Next, at step 117, the purge valve duty (basic DUTY) EVAP required to actually flow the purge valve flow rate QPV is
It is obtained by referring to the characteristic table shown in FIG.

Then, the calculated purge valve duty E
VAP is compared with a predetermined upper limit value EVPMAX # (step 118), and when EVAP> EVPMAX #, the purge valve duty EVAP is limited to the upper limit value EVPMAX # (step 119). Next, at step 120, the purge fuel flow rate QEF is calculated by the following equation. QEF = QPV * WC That is, the purge valve flow rate (purge gas flow rate) QPV is multiplied by the concentration learning value (learning value of the fuel concentration in the purge gas) WC to obtain the purge fuel flow rate QEF. This portion corresponds to the purged fuel flow rate calculation means.

Next, at step 121, the purge air flow rate QE
A is calculated by the following formula. QEA = QPV−QEF Purge air flow rate QE is the remainder after subtracting the purge fuel flow rate QEF from the purge valve flow rate (purge gas flow rate) QPV.
This is because it corresponds to A. This part corresponds to the purge air flow rate calculation means.

Next, at step 122, it is checked whether or not this routine has passed the previous time. If it is the first time, at step 123,
A flag that has passed once is set (step 123), and then concentration learning (WC learning) is permitted in step 124 (#FWC).
GKOK = 1). If it is not the first time, this job is ended as it is. [When the purge ON permission condition is not satisfied] Step
At 125, ban purging.

Further, at step 126, the purge valve duty EVAP = 0 is set. As a result, the purge valve duty EVAP output from the control unit also becomes 0 in the processing of FIG. 9 described later, and the purge control valve is controlled to close. Accordingly, in steps 127 and 128, the purge fuel flow rate QEF and the purge air flow rate QEA are both set to
Set to 0. When the purge control valve is closed, the purge fuel flowing into the intake manifold becomes zero. However, this purge fuel is diffused in the intake manifold and the phase is delayed, so that the purge fuel is sucked into the cylinder. Fuel does not go to zero immediately. Therefore, the calculation for this point is performed in FIG. 9 as described later.

Further, in step 129, the permission of the concentration learning is canceled with the purge OFF (# FWCGKOK =
0), this job is ended. FIG. 6 is a job for determining the start and end of density learning (WC learning), which is executed as a rotation synchronization job. First, in step 301, it is checked whether or not various sensors are NG, and if there is an NG sensor, in step 302 the concentration learning value WC = NGWC #
After that, the flag for WC learning and the RAM are initialized (step 310), and this job is ended.

If the sensor is not NG, the routine proceeds to step 303, where WC learning is permitted (# FWCGKOK =
Check 1) and if WC learning is not permitted, end this job as it is. If the WC learning is permitted, the learning start permission condition is determined in steps 304 to 306.

That is, at step 304, it is checked whether the air-fuel ratio feedback control (Ramcon) is under clamping. In step 305, the purge valve duty E
VAP is compared with a predetermined lower limit value WCGDTTY #, and E
Check if VAP <WCGDTY # (purge valve duty too low). In step 306, the intake air flow rate QH0 obtained from the throttle valve opening TVO and the engine speed NE is compared with a predetermined upper limit value WCGQH #, and it is checked whether QH0> WCGQH # (the load is too high). .

If YES in any of these determinations, it is determined that WC learning is not possible, and a flag for WC learning or a RAM.
After initializing (step 310), this job ends. If all the judgments in steps 304 to 306 are NO,
Learning starts as WC learning condition OK, but first, step 30
At 7, the "end judgment" of WC learning is performed. That is, the step amount ± PWC of the update amount of the density learning value WC is NSWCGK
# Check if added more than once.

If NSWCCGK is added # times or more, a flag (# FWC1KAI) indicating that WC learning has been performed once after starting is set in step 308, and the step is executed.
At 309, the density learning value WC is updated and clamped to the average value of WC immediately before the addition of the latest two steps by the following equation. WC = (OLDWC1 + OLDWC2) / 2 Subsequently, after the flags for WC learning and the RAM are initialized (step 310), this job is ended.

In the determination at step 307 described above, the step amount ± PWC of the update amount of the density learning value WC is NSWCCGK #.
If it has not been added more than once, it branches to FIG. 7. Figure 7
6 is a job that is executed subsequently to FIG. 6, is a main job of concentration learning (WC learning), and corresponds to purge fuel concentration learning means. What branches to this job is purge ON, WC learning condition OK, and concentration learning value WC.
The number of times the update amount is incremented by ± PWC is NSWCG
When it is less than K #.

First, in step 311, it is checked whether or not it is the first time to enter WC learning, and if it is the first time, step 312
Then, the average value ALPAV of the air-fuel ratio feedback correction coefficient α calculated by the job of FIG. 8 described later is stored in the learning start holding RAM = ALPST to prepare for learning. Next, at step 313, referring to the characteristic table shown in FIG. 18, the density learning value W is changed from the current density learning value WC.
The integral IWC of the update amount of C is obtained.

Here, the use of the concentration learning value WC for the calculation of the integrated amount IWC of the update amount is related to the fuel concentration because when the fuel concentration in the purge gas is high, the actual concentration rapidly changes. If a constant update amount (here, the case of setting an intermediate value) is given, the update of the density learning value cannot catch up, and the density learning value remains deviated to a value larger than the actual density. This is to prevent a lean error from occurring due to an excessive amount of fuel injection amount reduction correction. On the contrary, in the thinned region, the update amount is prevented from being too large with respect to the change in the actual concentration and the control amount overshoots, thereby preventing hunting in the air-fuel ratio control.

In other words, the update amount of the concentration learning value WC is solved by referring to the concentration learning value WC capable of predicting the change rate of the fuel concentration. As a result, the update amount of the concentration learning value WC corresponding to the actual changing speed of the fuel concentration accompanying the progress of purging of the canister can be obtained over the entire area from the initial stage of purging.

Next, at step 314, the change amount of the air-fuel ratio feedback correction coefficient α (change amount from the start of learning) Δα
Is calculated by the following formula. Δα = ALPAV−ALPST That is, the current average value ALPAV of the air-fuel ratio feedback correction coefficient α is read, and the average value ALPST at the start of learning is subtracted from this to obtain the amount of change Δα from the start of learning.

Next, at step 315, ALPAV = ALP
In ST, it is checked whether there is no change (Δα = 0). If there is no change, it is determined that the actual fuel concentration and the learning value match, and in step 316, the update amount ΔWC = 0 of the concentration learning value.
And If Δα is not 0, then in step 317, it is checked whether Δα is positive or negative. If Δα is negative,
When purge is turned on, the air-fuel ratio changes to the rich side compared to before the purge, and α is reduced to correct it. Therefore, the actual fuel concentration is "richer" than the current holding value ( It can be determined that the injection amount reduction correction is still insufficient with the current concentration learning value).

Therefore, if Δα is negative, the density learning value WC is increased. At this time, in step 318, it is checked if the learning update direction is reversed with respect to the previous update, and if it is reversed. In step 319, as the update amount ΔWC, + PWC, which is a large update amount step for accelerating the convergence of learning, is added (ΔWC = + P
WC). If it is not inversion, in step 322, + IWC, which is an integral part of the small update amount obtained according to the density change based on the density learning value WC, is added (ΔWC = + I).
WC).

If Δα is positive, it means that the air-fuel ratio changes to the lean side as compared with that before the purge when the purge is ON, and α becomes large to correct it, so the actual fuel concentration is It can be judged as "thinner" than the current holding value. Therefore, if Δα is positive, the density learning value WC is reduced. At this time, in step 323, it is determined whether the learning update direction is reversed with respect to the previous update. At 324, the update amount ΔWC is a large update amount step for accelerating the convergence of learning −
PWC is added (ΔWC = −PWC). If it is not inversion, it is the integral of the small update amount obtained in step 327 according to the density change based on the density learning value WC-
IWC will be added (ΔWC = −IWC).

At the time of adding ± PWC for the step, in step 320 or step 325, the WC immediately before the addition of the step is saved from the newest to the last two (OLDWC).
1 to OLDWC2, WC to OLDWC1). Then, in step 321 or step 326, the step amount ± P
The counter indicating the number of times WC is added is incremented by one. This counter is used to determine the end of WC learning in step 307 of FIG.

After that, this operation is repeated so that the value of the air-fuel ratio feedback correction coefficient α before and after the purge becomes the same, and the fuel accompanying the progress of the purge is maintained so that the air-fuel ratio does not shift even during the purge. The change in density is corrected sequentially. Next, at step 328, the update amount ΔWC is added to the current value (WC ₋₁ ) of the density learning value WC as in the following equation,
The density learning value WC is updated.

WC = WC _-1 + ΔWC Next, in step 329, the updated density learning value WC is compared with predetermined upper and lower limit values, and when these values are exceeded, they are limited to the upper and lower limit values. , This job ends. FIG. 8 is a job for calculating the average value ALPAV of the air-fuel ratio feedback correction coefficient α, which is executed as a REF synchronization job in synchronization with the generation of the reference signal REF from the crank angle sensor.

In the air-fuel ratio feedback control using the O ₂ sensor, the air-fuel ratio feedback correction coefficient α is controlled by the step portion (P portion) and the integral portion (I minute), and since the variation of α is large, P ALPA is the simple average of α when minutes are added (immediately before addition) and α when the previous P minutes are added (immediately before addition)
Seeking V. This is a method often used in basic air-fuel ratio learning control.

Therefore, in step 401, it is determined whether the air-fuel ratio feedback control (ramcon) is in the clamp state. If it is in the clamp state, in step 402, ALPAV = 1.0.
, ALPOLD = 1.0. If it is not clamped, it is determined in step 403 whether P minutes are added. If the P component is being added, in step 404, immediately before the P component is added, the peak-peak average value ALPAV of the air-fuel ratio feedback correction coefficient α is calculated by the following equation.

ALPAV = (α + ALPOLD) / 2 Next, the current α is held in ALPOLD to prepare for the next time. Thereafter, in step 405, a normal P component calculation is performed to add the P component to the air-fuel ratio feedback correction coefficient α. If it is not when P minutes are added, in step 406 the normal I
The air-fuel ratio feedback correction coefficient α is calculated as I
Add minutes.

Here, the steps 403 to 406 correspond to the air-fuel ratio feedback correction coefficient setting means. Figure 9
Is a job for performing duty output to the purge control valve and calculating the cylinder intake purge fuel flow rate in consideration of the delay, and is executed as a REF synchronization job. First step
At 501, output the purge valve duty EVAP to the control unit / output port. As a result, the purge control valve is turned on and off. The output to the purge control valve is a pulse signal with a 6.4 ms cycle, and the ON time ratio is controlled by its duty.

Next, at step 502, the purge fuel flow rate QE
F flows from the purge control valve into the intake manifold,
The calculation for approximating the amount sucked into the cylinder due to diffusion and transfer delay in the intake manifold by control is performed as follows. First, assuming that the diffusion of the fuel gas in the intake manifold is due to the first-order lag, the QEF is multiplied by the weighted average to be approximated by the following equation.

QEF1 = QEF * EDMP # + QEF1
_-1 * (1-EDMP #) Next, in order to approximate the dead time (simple time delay) in the intake manifold, bring in QEF1 obtained in this job a few REF before with the memory operation as shown in the following formula. The final cylinder intake purge fuel flow rate QEFC. QWFC = value before number REF of QEF1 Of course, since QEF1 obtained this time is used after several REF, it is stored in a considerable memory.

FIG. 10 shows a job for calculating the cylinder intake air flow rate Q, which is executed, for example, every 4 ms. In step 601, the purge air flow rate QEA is added to the intake air flow rate QA detected by the air flow meter to obtain the cylinder intake air flow rate Q used for the fuel injection amount calculation as in the following equation. This portion corresponds to the cylinder intake air flow rate calculation means.

Q = QA + QEA In this way, by correcting the purge air amount from the canister which is sucked into the engine without being measured by the air flow meter, the lean error of the air-fuel ratio which becomes large especially as the desorption progresses is prevented. can do. Figure 11
Idle control valve opening aiming at correction of fuel injection amount based on WC learning result and prevention of deterioration of drivability due to change in torque due to increase in intake air amount to the engine by the purge air amount of the canister Is a job for performing correction of, and is executed, for example, every 10 ms.

First, at step 701, a normal basic fuel injection amount (basic injection pulse width) TP is obtained from the following formula as usual from the cylinder intake air flow rate Q and the engine speed NE. This portion corresponds to the basic fuel injection amount calculation means. TP0 = Q * KCONST # / NE TP = TP0 * FLOAD + TP- ₁ * (1-FLOA
D) Here, KCONST # is a proportional constant that has been conventionally used, and FLOAD is a weighted average coefficient for adjusting the phase of TP to the cylinder intake amount.

Next, at step 702, the cylinder intake purge fuel flow rate QEFC obtained in FIG. 9 is converted into a unit corresponding to the injection pulse width by the following equation, and the purge fuel amount (reduction correction amount for the cylinder intake purge fuel flow amount) TEFC To get TEFC = QEFC * KFQ # * KCONST # / NE Here, KFQ # is a correction coefficient for purge gas specific gravity.

Next, in step 703, the purge fuel amount (reduction correction amount for the cylinder intake purge fuel flow rate) TEFC is subtracted from the basic fuel injection amount TP as in the following equation to remove the purge fuel amount, and the intake port portion is removed. The purged fuel subtraction basic fuel injection amount TPE, which is the required basic injection amount of This portion corresponds to the purge fuel subtraction basic fuel injection amount calculation means.

TPE = TP-TEFC Next, at step 704, the target fuel air ratio TFBYA, based on the purge fuel amount subtracting basic fuel injection amount TPE, is calculated as follows.
The final fuel injection amount TI (sequential injection) is obtained by making a correction with the air-fuel ratio feedback correction coefficient α or the like.
This portion corresponds to the fuel injection amount calculation means.

TI = TPE * TFBYA * α * 2 + TS Here, TS is an invalid pulse width. As described above, by using TPE instead of the conventional TP, it is possible to reduce the amount of purge fuel from the fuel injection amount, so that a large amount of purge can be performed without deteriorating the air-fuel ratio at the time of purge ON including transients. It will be possible.

Next, the calculation part (steps 705 and 706) of the idle speed control specification will be described. This is the duty to the normal idle control valve (ISC duty)
On the other hand, during the purge, the duty correction amount ISCEVP corresponding to the purge air flow rate QEA is subtracted to eliminate the step difference in the intake air amount at the time of purge ON and OFF and prevent the deterioration of the drivability due to the sudden change of the torque. Is aimed at.

In step 705, the purge air flow rate QEA is converted into the ISC duty for the idle control valve as shown in the following equation to obtain the duty correction amount ISCEVP.
To calculate. ISCEVP = QEA * ISCEVG # / (KPVQH
* KPVVB) where ISCEVG # is the purge air flow rate QEA
It is a coefficient for converting into C duty. Also, K
PVQH and KPVVB are the above-mentioned differential pressure correction rate and voltage correction rate.

Next, at step 706, the duty correction amount ISCEVP is subtracted from the normal ISC duty to obtain a calculated duty (ON duty) ISCON to the idle control valve, which is output. ISCON = (normal ISC duty) -ISCEV
P FIG. 12 shows an initialization (initialization) job, which corresponds to learning initialization means and is executed only once when the engine is started.

In step 801, an initial value is given to the density learning value WC to initialize it. The initial value here is 0% or a predetermined value. It is conceivable that the predetermined value is the lower limit value of the fuel concentration in the purge gas in the market or the intermediate value in the usage range. Next, the effect of this embodiment is shown in FIG.
Will be explained. When the fuel concentration becomes high due to the evaporated fuel generated while the engine is stopped, there is a deviation from the previously learned concentration learning value. Therefore, if a large amount of purge is performed during the first purge at the next engine start,
Since the deviation of the density cannot be corrected, the emission and drivability are adversely affected. That is, when the target purge rate is given from the beginning (conventional example 1), the concentration deviation is large * the high purge rate, and the air-fuel ratio error is large.

Therefore, in the present invention, the purge rate at the first time after starting is suppressed to a low purge rate (initial purge rate), and the concentration deviation is corrected (relearned) under the condition of little influence. And
After completion of the correction, that is, after the learned value catches up with the actual concentration, the purge rate is switched to the target purge rate, and the mass purge is started. Therefore, it is possible to quickly shift to the large-scale purge, and it is possible to prevent deterioration of operability and emission and to secure the large-scale purge at the same time.

On the other hand, when the purge rate is first reduced and then increased stepwise (conventional example 2), the air-fuel ratio error is small, but the purge amount is also small.

[0084]

As described above, according to the first aspect of the present invention, even if there is a deviation between the actual fuel concentration and the learning value at the time of engine start, the purge gas flow rate remains until the deviation is corrected. Since it can be kept smaller than the normal value, the effect on emission and drivability can be avoided, and if the deviation of the fuel concentration is corrected, it is possible to quickly shift to the normal large-volume purge, which deteriorates emission and drivability. The effect that both prevention and large-scale purging can be achieved is obtained.

According to the second aspect of the present invention, at the time of the first purge after starting, the initial purge rate smaller than the normal target purge rate is set so that the purge gas flow rate also corresponds to the intake air flow rate. The effect that can be obtained is obtained. According to the third aspect of the present invention, the convergence state of learning can be accurately grasped by determining the degree of progress of learning depending on whether or not the update increasing / decreasing direction of the learning value of the fuel concentration is reversed a predetermined number of times or more. The effect is obtained.

According to the invention of claim 4, the learning value of the fuel concentration is initialized every time the engine is started, and the re-learning of the fuel concentration is always started from a predetermined value. Therefore, the memory backup is not necessary, but the learning is performed. It is possible to obtain an effect that variation in speed and the like can be suppressed.

[Brief description of drawings]

FIG. 1 is a functional block diagram showing the configuration of the present invention.

FIG. 2 is a system diagram showing an embodiment of the present invention.

FIG. 3 is a flowchart of a job for calculating a purge rate, etc.

FIG. 4 is a flowchart of a job following FIG.

FIG. 5 is a flowchart of a purge rate calculation subroutine.

FIG. 6 is a flowchart of a job for determining the start and end of density learning.

FIG. 7 is a flowchart of a density learning main job following FIG.

FIG. 8 is a flowchart of a job for air-fuel ratio feedback control.

FIG. 9 is a flowchart of a cylinder intake purge fuel flow rate calculation job.

[Fig. 10] Flow chart of a cylinder intake air flow rate calculation job

FIG. 11 is a flowchart of a job for calculating a fuel injection amount and the like.

FIG. 12 is a flowchart of a job for calculating a weight reduction correction rate.

FIG. 13 is a diagram showing a characteristic example of a target purge rate calculation table.

FIG. 14 is a diagram showing a characteristic example of a purge permission lower limit TP calculation table.

FIG. 15 is a diagram showing a characteristic example of a differential pressure correction factor calculation table.

FIG. 16 is a diagram showing a characteristic example of a voltage correction factor calculation table.

FIG. 17 is a diagram showing a purge valve duty calculation table.

FIG. 18 is a diagram showing a characteristic example of an IWC calculation table.

FIG. 19 is a diagram showing the function and effect of the embodiment.

[Explanation of symbols]

1 engine 2 air cleaner 3 throttle valve 4 intake manifold 5 fuel injection valve 6 exhaust passage 7 control unit 8 air flow meter 9 crank angle sensor 10 O ₂ sensor 11 bypass passage 12 idle control valve 13 fuel tank 14 canister 15 evaporative fuel introduction pipe 16 new Air inlet 17 Purge passage 18 Purge control valve

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification code Internal reference number FI technical display area F02M 25/08 301 F02M 25/08 301U

Claims (4)

[Claims] 1. A canister for adsorbing evaporated fuel from a fuel tank, which comprises a fuel injection valve for injecting fuel into an engine, and a purge gas containing evaporated fuel desorbed by introducing fresh air into the canister. A purge control valve for controlling the purge gas flow rate, which is interposed in a purge passage that guides the engine downstream of the throttle valve in the engine intake passage; and a purge control means for controlling the opening of the purge control valve to control the purge gas flow rate. In an evaporative fuel processing system for an engine, a basic fuel injection amount calculation means for calculating a basic fuel injection amount based on engine operating conditions and a fuel injection amount is corrected based on a signal from an air-fuel ratio sensor provided in an exhaust passage. And an air-fuel ratio feedback correction coefficient setting means for setting an air-fuel ratio feedback correction coefficient for Purge fuel concentration learning means for learning the fuel concentration in the purge gas based on the change of the coefficient, and purge fuel flow rate calculation means for calculating the purge fuel flow rate based on the purge gas flow rate and the fuel concentration learning value by the purge control means. Purge fuel amount subtraction by subtracting the purge fuel flow amount from the basic fuel injection amount Purge fuel amount subtraction for calculating the basic fuel injection amount Basic fuel injection amount calculation means, Purge fuel amount subtraction Basic fuel injection amount and air-fuel ratio feedback correction A fuel injection amount calculation means for calculating the fuel injection amount by the fuel injection valve based on the coefficient, a normal purge gas flow rate setting means for setting a normal purge gas flow rate based on the operating condition of the engine, and a first time after the engine is started. Initial purge gas flow rate setting means for setting a smaller initial purge gas flow rate than the normal purge gas flow rate for The learning progress degree determining means for determining the degree of progress of the fuel concentration learning by the purge fuel concentration learning means, and the purge gas flow rate controlled by the purge control means until the fuel concentration learning progresses after the engine is started, the initial purge gas flow rate setting The initial purge gas flow rate set by the means, and switching means for switching to the normal purge gas flow rate set by the normal purge gas flow rate setting means after the progress of the fuel concentration learning; apparatus. 2. The normal purge gas flow rate setting means is a target purge rate setting means for setting a target purge rate as a ratio of the purge gas flow rate to the intake air flow rate so that the purge gas flow rate corresponds to the intake air flow rate. The evaporated fuel processing apparatus for an engine according to claim 1, wherein the initial purge gas flow rate setting means is an initial purge rate setting means for setting an initial purge rate smaller than the target purge rate. 3. The learning progress degree determining means determines the learning progress degree based on whether or not the update increasing / decreasing direction of the learning value of the fuel concentration is reversed a predetermined number of times or more. Alternatively, the evaporated fuel processing device for an engine according to claim 2. 4. The learning initialization means for initializing the learning value of the fuel concentration by the purge fuel concentration learning means each time the engine is started, according to any one of claims 1 to 3. Evaporative fuel treatment system for an engine according to item 1.