JP2000008974A - Trouble diagnosing device for evaporated fuel treatment equipment - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a trouble diagnosing device for evaporated fuel treatment equipment capable of early detecting failure of the evaporated fuel treatment equipment by increasing the executive frequencies of a trouble diagnosis. SOLUTION: A tank pressure control valve 52 attached to a housing 21 of a canister 20 is connected to a fuel tank 10 via a vapor passage 64. A surge tank 16 is connected to the housing 21 via a purge passage 62. A pressure sensor 46 is connected to the vapor passage 64 and the housing 21 respectively via a three-way selector valve 44 and each of pressure passages 72 and 74. An electronic control unit 30 of an engine learns the valve opening pressure of the tank pressure control valve 52 on the basis of internal pressure in the fuel tank 10 when the specified learning condition is materialized. The electronic control unit 30 renews the learned value even in time of the learned condition being yet unsatisfied when the learned value is below a judging pressure set below the designed valve-opening pressure of the tank pressure control valve 52 and the internal pressure in the fuel tank 10 is larger than the learned value.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a failure diagnosis apparatus for an evaporative fuel treatment apparatus which moves and collects evaporative fuel generated in a fuel tank through a vapor passage to a canister, and more particularly, opens and closes the vapor passage. The present invention relates to a failure diagnosis device for an evaporative fuel treatment device that learns the valve opening pressure of a pressure valve and prohibits failure diagnosis when the internal pressure of a fuel tank is equal to or higher than this learning value.

[0002]

2. Description of the Related Art FIG. 9 shows a general configuration of an evaporative fuel processing apparatus provided in an internal combustion engine. In this fuel vapor processing apparatus, the canister 110 is connected to the vapor passage 120.
Is connected to the fuel tank 100 via a purge passage 140 and to the intake passage 130 via a purge passage 140. The canister 110 has an internal pressure (hereinafter, referred to as an internal pressure).
An atmosphere introduction valve 114 is provided which opens when the pressure (hereinafter referred to as “canister pressure”) becomes equal to or lower than a predetermined pressure to introduce the atmosphere into the canister 110.

The vapor passage 120 is provided with a pressure valve 112 that opens when the internal pressure of the fuel tank 100 (hereinafter, referred to as “tank pressure”) rises to a predetermined pressure or higher. When valved, fuel tank 100
(Hereinafter, referred to as “vapor”) is introduced into the canister 110 through the vapor passage 120.
An adsorbent (not shown) is built in the canister 110, and the fuel vapor introduced through the vapor passage 120 is adsorbed by the adsorbent.

[0004] The vapor thus adsorbed by the adsorbent is:
The intake passage 1 through the purge passage 140 together with the atmosphere introduced into the canister 110 from the atmosphere introduction valve 114.
It is supplied (purged) into 30. The purge passage 140 is provided with a purge amount control valve 150,
By 0, the amount of vapor purged from the canister 110 to the intake passage 130 is adjusted.

If the pressure passage system such as the canister 110 and the purge passage 140 connected to the canister 110 is damaged by corrosion or the like, there is a possibility that the vapor leaks from the damaged portion to the outside. For this reason, in the conventional evaporative fuel processing apparatus, a pressure sensor 170 connected to both the fuel tank 100 and the canister 110 via the three-way switching valve 180 is provided, and the tank pressure detected by the pressure sensor 170 and the canister Diagnosis of such a failure is made based on the pressure.

That is, when performing the failure diagnosis, first, the pressure sensor 170 is connected to the canister 110 by switching the three-way switching valve 180, and then the purge amount control valve 150 is opened to open the inside of the intake passage 130. A negative pressure is introduced into canister 110. Then, the purge amount control valve 1
After closing the valve 50 to cut off the connection between the canister 110 and the intake passage 130, the change in the canister pressure is detected by the pressure sensor 1.
70 to detect. At this time, if the rising speed of the canister pressure is higher than a predetermined value, it is assumed that the air has entered the pressure passage system from the damaged portion of the pressure passage system,
It can be diagnosed that a failure has occurred.

As described above, when diagnosing a failure in the pressure passage system based on the rising speed of the canister pressure, the pressure valve 11
2 must be in a closed state and the movement of the vapor from the fuel tank 100 to the canister 110 must be stopped.
If the failure diagnosis is performed while the vapor is flowing into the canister 110, if the rising speed of the canister pressure becomes larger than a predetermined value due to the inflow of the vapor, it is erroneously diagnosed as a failure of the pressure passage system. This is because there is a danger that they will be lost.

Therefore, in the conventional failure diagnosis apparatus, the valve opening pressure of the pressure valve 112 is learned and stored in the memory at every predetermined timing, and the tank pressure is less than the learning value stored in the memory. The failure diagnosis is executed only in the case (for example, “Failure diagnosis device for evaporative purge system” described in JP-A-10-54307).

For example, in the case of learning the valve opening pressure in the evaporative fuel processing apparatus shown in FIG. 9, first, the purge amount control valve 150 is closed and the purge passage 140 is shut off. Next, the canister pressure is detected by the pressure sensor 170, and the tank pressure after a lapse of a predetermined time after the rate of increase of the canister pressure becomes smaller than a predetermined value is stored in the memory as a learning value of the valve opening pressure.

By learning the valve opening pressure of the pressure valve 112 in this way, even if the valve opening pressures are different due to individual differences or change due to aging, etc. The failure diagnosis can be performed in a situation where the pressure valve 112 is reliably closed.

Further, when learning the valve opening pressure of the pressure valve 112, a learning condition is that a predetermined time has elapsed since the rate of increase of the canister pressure became smaller than a predetermined value.
Since there is almost no movement of the vapor from the fuel tank 100 to the canister 110, and when the tank pressure and the canister pressure are both in a stable equilibrium state, the learning of the valve opening pressure is performed. Is also accurate.

[0012]

By the way, in the conventional failure diagnosis apparatus, the supply of power to the memory is stopped,
When the contents stored in the memory are initialized, the learning value is set to “0” or an initial value sufficiently low to prevent erroneous diagnosis.

However, after the initialization of the memory, for example, the movement of the vapor between the fuel tank 100 and the canister 110 is frequently performed, and it takes a long time until the learning condition as described above is satisfied. If so, the execution timing of the failure diagnosis is determined based on the initial value, so that the frequency of execution of the failure diagnosis is reduced, and there is a possibility that the discovery of the failure may be delayed.

SUMMARY OF THE INVENTION The present invention has been made in view of the above circumstances, and has as its object to increase the frequency of execution of a failure diagnosis so that a failure in an evaporative fuel processing apparatus can be found at an early stage. An object of the present invention is to provide a device failure diagnosis device.

[0015]

In order to achieve the above object, according to the present invention, there is provided a vapor passage for connecting a fuel tank and a canister to move evaporated fuel generated in the fuel tank to the canister, and an internal pressure of the fuel tank. A pressure valve that opens when the pressure becomes equal to or higher than a predetermined valve opening pressure to allow movement of evaporated fuel, pressure detection means for detecting the internal pressure of the fuel tank and the canister, and the canister based on the internal pressure of the canister. Failure determination means for determining a failure of the pressure passage system including: learning means for learning the opening pressure of the pressure valve based on the internal pressure of the fuel tank detected by the pressure detection means when a predetermined learning condition is satisfied, A storage unit for storing a learned value of the valve opening pressure; and a failure determination unit when the internal pressure of the fuel tank detected by the pressure detection unit is equal to or greater than the learning value stored in the storage unit. And a prohibiting means for prohibiting the failure determination of the evaporative fuel treatment apparatus, wherein the learning value is less than a determination pressure set lower than a designed valve opening pressure of the pressure valve, and When the internal pressure of the fuel tank detected by the detecting means is larger than the learning value, the learning of the valve opening pressure is performed even when the learning condition is not satisfied.

According to the above configuration, even if it takes a long time until the learning condition is satisfied after the learning value stored in the storage means is initialized, the initialized learning value is maintained. Until the pressure rises and reaches a judgment pressure set lower than the designed valve opening pressure of the pressure valve, learning relating to the valve opening pressure of the pressure valve is performed, and the learning value is updated to a larger value. Become so.

[0017]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment in which the present invention is applied to a failure diagnosis device for an evaporative fuel treatment device provided in a gasoline engine for a vehicle will be described below.

FIG. 1 shows a schematic configuration of an evaporative fuel processing apparatus and a failure diagnosis apparatus according to the present embodiment. The canister 20 has a housing 21 having an internal space 21a.
And a pair of filters 2 provided in the housing 21.
2 and an adsorbent 26 made of activated carbon or the like provided between the filters 22. This housing 2
A tank pressure control valve 52, which is, for example, a diaphragm-type pressure valve, is attached to one side of the fuel cell 10. The valve 52 is connected to a fuel tank 10 in which fuel (gasoline) is stored via a vapor passage 64. ing.

When vapor is generated inside the fuel tank 10 and its internal pressure (hereinafter, referred to as “tank pressure PTNK”) rises to the opening pressure of the tank pressure control valve 52 or more, the tank pressure control valve 52 opens. Give a valve. As a result, fuel tank 1
The vapor inside the housing 21 passes through the vapor passage 64 to the housing 21.
Is introduced into the internal space 21a and is adsorbed in the adsorbent 26 in a liquid state.

The surge tank 16 provided at a position downstream of the throttle valve 18 in the intake passage 14 of the engine is connected to the housing 21 via a purge passage 62. In the purge passage 62, the canister 2
A purge amount adjusting valve 42 is provided for adjusting the amount of purged vapor (purge amount) in the surge tank 16 from zero.

In the intake passage 14, the throttle valve 1
An air cleaner 12 is provided at a position upstream of the air cleaner 8. An air introduction valve 54 is attached to a side of the housing 21, and the air introduction valve 54 is connected to the air cleaner 12 via an air introduction passage 66.

With the opening of the purge amount control valve 42, a negative pressure of the surge tank 16 is introduced into the canister 20,
The internal pressure of the canister 20 (hereinafter referred to as “canister pressure PCA”).
N ”) drops below a predetermined pressure.
4 is in an open state. As a result, the atmosphere is introduced into the canister 20 through the atmosphere introduction passage 66. The liquid fuel adsorbed by the adsorbent 26 becomes gas (vapor) again by the atmosphere introduced into the canister 20, and is then purged into the surge tank 16 through the purge passage 62.

At the side of the housing 21, an air introduction valve 54 is provided.
And an air release valve 56 is attached adjacent to. The air release valve 56 opens when the canister pressure PCAN rises to or above the atmospheric pressure, and reduces the canister pressure PCAN to the atmospheric pressure.

Pressure passages 72 and 74 are connected to the vapor passage 64 and the housing 21, respectively.
Each of these pressure passages 72 and 74 is provided with a three-way switching valve 44.
It is connected to the. The three-way switching valve 44 has a pressure detection passage 78
Is connected to the pressure sensor 46 via the. The pressure sensor 46 outputs a detection signal based on the magnitude of the tank pressure PTNK or the canister pressure PCAN according to the switching position of the three-way switching valve 44.

A vehicle speed sensor 48 is provided near wheels (not shown) of the vehicle. The vehicle speed sensor 48 outputs a detection signal based on the magnitude of the vehicle speed SPD. Thus, the detection signals output from the pressure sensor 46 and the vehicle speed sensor 48 are transmitted to an engine electronic control unit (hereinafter referred to as “ECU”).
30). The ECU 30 controls the fuel injection control and the ignition timing, controls the opening of the purge amount control valve 42 to adjust the purge amount to an amount suitable for the operating state of the engine (purge control), and controls the fuel vaporization. The control related to the failure diagnosis of the processing device is executed.

The ECU 30 has a memory 31 for executing such various controls according to predetermined procedures.
The memory 31 stores a ROM 31a in which control programs and function data used for various controls are stored in advance, a RAM 31b for temporarily storing the results of various arithmetic processing, and a failure diagnosis result even after the engine is stopped. And a backup RAM 31c that can perform the operations.
The contents stored in the backup RAM 31c are lost when the power supply from the battery (not shown) is stopped.
At the time of the next power supply, initialization is performed by the ECU 30.

Next, a procedure for diagnosing a failure of the evaporated fuel processing apparatus according to the present embodiment will be described with reference to the flowcharts of FIGS. In the failure diagnosis processing according to the present embodiment, first, an average value PTNKW of the tank pressure PTNK is detected in a “tank pressure detection routine” shown in FIG. In the following description, the tank pressure PTNK actually detected by the pressure sensor 46 is referred to as “the actual tank pressure PTN”.
K ”and the smoothed value PTNKW are simply referred to as“ tank pressure PT
NKW ".

Next, in the "update time determination routine" shown in FIG. 3, the learning value P relating to the valve opening pressure of the tank pressure control valve 52 is obtained.
With respect to TNKWG, it is determined whether it is time to update the learning value PTNKWG. In the “update condition determination routine” shown in FIGS. 4 and 5, the learning value PTNWG is determined.
It is determined whether a condition for updating KWG is satisfied.

Further, a "valve opening pressure learning routine" shown in FIG.
In, when the current control cycle is the update time of the learning value PTNKWG and it is determined that the update condition is also satisfied, the learning value PTNKWG is updated. Then, in the “failure diagnosis routine” shown in FIG.
Only when TNKW is less than the learned value PTNKWG, the failure diagnosis of the fuel vapor processing apparatus is executed.

Hereinafter, each processing of these routines will be described in order. First, each process of the “tank pressure detection routine” shown in FIG. 2 will be described. This routine is executed by the ECU 30 as an interruption process for each predetermined time period.

When the processing shifts to this routine, the ECU
In step 110, the pressure sensor 46 is switched by switching the three-way switching valve 44 so that the fuel tank 10
To the side. Next, the ECU 30 determines in step 112,
At 114, after reading the actual tank pressure PTNK detected by the pressure sensor 46, the current tank pressure PTN is read.
KW is set as the tank pressure PTNKWB in the previous control cycle.

Then, at step 116, the ECU
30 calculates the tank pressure PTNKW in the current control cycle based on the following equation (1). PTNKW = PTNKWB + (PTNK−PTNKWB) / 4 (1) As is clear from the above equation (1), in the present embodiment, the tank pressure PTNKW is set as a so-called smoothed value for the actual tank pressure PTNK. Therefore, the change is slower than the change in the actual tank pressure PTNK. As described above, the internal pressure of the fuel tank 10 is not the actual tank pressure PTNK which is the actual detection value, but the smoothed value (PTNK).
NKW) is detected because, even if the actual tank pressure PTNK fluctuates at a relatively fast cycle irrespective of the amount of generated vapor due to, for example, vehicle vibration, the influence of such fluctuation is considered. This is because it can be removed.

Next, at step 118, the ECU 3
0 compares the tank pressure PTNKW calculated in this way with the tank maximum pressure PTNKWM. This tank maximum pressure PTNKWM is the tank pressure P after the engine starts.
This is the maximum value of TNKW, and is a value that is initialized each time the operation of the engine is stopped. If it is determined in this step 118 that the currently calculated tank pressure PTNKW is higher than the tank maximum pressure PTNKWM, the ECU 30
Sets the tank pressure PTNKW as the new tank maximum pressure PTNKWM in step 120. Then, in step 122, the maximum pressure duration CPTNK
Reset WM to "0". This maximum pressure duration CP
TNKWM indicates the elapsed time since the tank maximum pressure PTNKWM was updated.

On the other hand, at step 118, the tank pressure P
When it is determined that TNKW is equal to or less than the tank maximum pressure PTNKWM, the ECU 30 determines in step 130 that
A predetermined time ΔT is added to the current maximum pressure duration CPTNKWM, and the added value is added to a new maximum pressure duration CPTNK.
Set as WM. Then, steps 122 and 130
After executing the processing of (1), the ECU 30 once ends the processing of this routine.

Next, each process of the "update time determination routine" shown in FIG. 3 will be described. This routine is executed by the ECU 30 as an interruption process for each predetermined time period. When the processing shifts to this routine, the ECU 30
In step 210, it is determined whether or not the post-start elapsed time CAST has reached a predetermined time CAST1. The post-start elapsed time CAST is an elapsed time since the engine was started, and is a time measured in a processing routine different from this routine. The predetermined time CAST1 is
This is for determining whether or not the post-start elapsed time CAST has become a sufficient time for determining the stability of the tank pressure PTNKW.

If it is determined in step 210 that the post-start elapsed time CAST is less than the predetermined time CAST1, it cannot be determined whether or not the tank pressure PTNKW is stable, and accordingly, the learning value PTNKWG is updated. If it is not time to do so, the ECU 30 shifts the processing to step 218. In step 218, the ECU 30 updates the update timing determination flag XRNWTIM
Set E to "0". This update timing determination flag XRN
WTIME is a flag for determining whether it is time to update the learning value PTNKWG.

On the other hand, if it is determined in step 210 that the post-start elapsed time CAST is equal to or greater than the predetermined time CAST1, the ECU 30 proceeds to step 212
It is determined whether or not the elapsed time after learning CPPCT is equal to or longer than a predetermined time CPPCT1.

The post-learning elapsed time CPPCT is an elapsed time after the learning value PTNKWG is updated, and is a time measured in a routine different from this routine. In this other routine, the elapsed time after learning CPP
The CT is reset to “0” each time the learning value PTNKWG is updated. Also, a predetermined time CPPCT1
Is for setting the update frequency of the learning value PTNKWG. The shorter the predetermined time CPPCT1 is set, the more frequently the learning value PTNKWG is updated.

In step 212, the elapsed time after learning C
When it is determined that the PPCT is equal to or longer than the predetermined time CPPCT1, the ECU 30 determines in step 214 whether the post-switching elapsed time CTPC is equal to or longer than the predetermined time CTPC1.

The post-switching elapsed time CTPC is an elapsed time since the pressure sensor 46 was connected to the fuel tank 10 side, and is a time measured in a processing routine different from this routine.

In the present embodiment, as described above, the tank pressure PTNKW is calculated as a smoothed value of the actual tank pressure PTNK. Therefore, the elapsed time from when the pressure sensor 46 is connected to the fuel tank 10 is calculated. If it is short, the tank pressure PTNKW may not accurately reflect the internal pressure of the fuel tank 10. The predetermined time CT
The elapsed time from the connection of the pressure sensor 46 to the fuel tank 10 is sufficiently long, and the PC 1 has a tank pressure PTN.
This is for determining that the KW accurately reflects the internal pressure of the fuel tank 10.

In step 214, the elapsed time after switching C
When it is determined that the TPC is equal to or longer than the predetermined time CTPC1, the ECU 30 sets the update timing determination flag XRNWTIME to "1" in step 216, assuming that it is the time to update the learning value PTNKWG.

On the other hand, the ECU 30 determines in step 212 that the elapsed time after learning CPPCT is equal to the predetermined time CPPCT1.
If it is determined that the elapsed time is shorter than the predetermined time CTPC1, or if it is determined in step 214 that the post-switching elapsed time CTPC is less than the predetermined time CTPC1.
The update timing determination flag XRNWTIME is set to "0"
Set to. After executing the processing of steps 216 and 218, the ECU 30 once ends the processing of this routine.

Next, each process of the "update condition determination routine" shown in FIGS. 4 and 5 will be described. This routine is executed by the ECU 30 as an interruption process for each predetermined time period.

When the processing shifts to this routine, the ECU
30 determines in step 310 whether both of the following conditional expressions (2) and (3) are satisfied. PTNKW-PTNKWC1 <α (2) PTNKW-PTNKWC2 <α (3) Here, “PTNKWC1” and “PTNKWC2” are tank pressures PTN before predetermined times T1 and T2 (> T1).
KW, and “α” is a predetermined value. In step 310, the current tank pressure PTNKW and the predetermined time T
1, the deviation from each tank pressure PTNKW before T2 (P
TNKW-PTNKWC1, PTNKW-PTNKWC
2) is compared with the predetermined value α to obtain the tank pressure PT
It is determined whether the NKW is stable.

In step 310, the tank pressure PTNK
If it is determined that W is stable, the ECU
30 is the maximum tank pressure PTN in step 312.
The current tank pressure PTNKW is subtracted from KWM, and the subtracted value is compared with a predetermined value β. In this step 312, it is determined whether or not the tank pressure PTNKW has decreased significantly.

In step 312, the tank pressure PTNK
If it is determined that W has not decreased significantly, ECU 3
0 is the maximum tank pressure PTNK in step 314.
The starting tank pressure PTNKST is subtracted from WM, and the subtracted value is compared with a predetermined value γ. This starting tank pressure PT
NKST is the internal pressure of the fuel tank 10 at the time of starting the engine. In this step 314, the tank pressure PTNK
It is determined whether W has risen sufficiently since the engine was started.

If it is determined in step 314 that the tank pressure PTNKW has risen sufficiently since the start of the engine, the ECU 30 determines in step 316 that the maximum pressure continuation time CPTNKWM has exceeded the predetermined time CPTNKWM.
It is determined whether it is 1 or more. In this step 316, the tank maximum pressure PTNKWM is increased for a predetermined time CPTNKWM.
It is determined whether it has not been updated for one or more times.

In step 316, the tank maximum pressure PT
If it is determined that the NKWM has not been updated for the predetermined time CPTNKWM1 or more, the ECU 30 determines in step 318 whether the vehicle speed SPD is less than the predetermined speed SPD1. In this step 318,
Based on the vehicle speed SPD, the amount of fuel consumed per unit time, in other words, the amount of fuel flowing out of the fuel tank 10 per unit time is estimated. If the vehicle speed SPD is less than the predetermined speed SPD1, the fuel It is determined that the fuel consumption is small and that there is no sudden change in the tank pressure PTNKW due to fuel consumption. The fuel consumption may be estimated based on the pressure of the intake air (intake pressure) in the surge tank 16 instead of the vehicle speed SPD.

If it is determined in step 318 that there is no sudden change in the tank pressure PTNKW due to fuel consumption, EC
U30 determines in step 320 that the tank pressure PTNK
W is compared with the initial learning value EPTNNKWG1. This initial learning value EPTNKWG1 is stored in the backup RAM 31
This is the initial value of the learning value PTNKWG that is set when the value of the learning value PTNKWG stored in c is initialized. This initial learning value EPTNKWG1 is set to a value sufficiently smaller than the valve opening pressure of the tank pressure control valve 52.

In step 320, the tank pressure PTNK
If it is determined that W is equal to or greater than the initial learning value EPTNNKWG1, the ECU 30 proceeds to step 322 shown in FIG.
Move to Then, in step 322, the ECU 30 sets the temporary update process execution flag XRNWLOW to “0”.
After that, in step 340, the update condition satisfaction flag XRNWCOND is set to “1”. The update condition satisfaction flag XRNWCOND is set to the learning value PTNKWG.
Is a flag for determining whether or not the condition for executing the update of is satisfied.

Therefore, in steps 310 to 320, (a) the temporal change of the tank pressure PTNKW is small and stable (step 310). (B) The tank maximum pressure PTNKWM and the tank maximum pressure P
Pressure value lower than TNKWM by a predetermined value β (PTNKWM
-Β) is present (steps 312 and 316). (C) The tank maximum pressure PTNKWM has risen sufficiently after the engine has started (step 314). (D) The tank maximum pressure PTNKWM is The tank pressure PTNKW has not been updated for a predetermined period of time, and therefore the tank pressure PTNKW has not increased (step 316). (E) The vehicle speed SPD is small and the fuel consumption is small (step 318). Learning value EPTNK
When all the conditions (a) to (f) such as WG1 or more (step 320) are satisfied, the update condition satisfaction flag XRNWCOND is set to “1”.

On the other hand, if a negative determination is made in any one of the steps 310 to 320, the ECU 3
0 is the tank pressure P in step 330 shown in FIG.
The TNKW is compared with the current learning value PTNKWG. Here, when it is determined that the tank pressure PTNKW is equal to or greater than the current learning value PTNKWG, the ECU 30 shifts the processing to step 332.

In step 332, the ECU 30
It is determined whether the following conditional expression (4) is satisfied. PTNKWG <(POPEN-k) (4) Here, “POPEN” is a valve opening pressure in the design of the tank pressure control valve 52, and “k” is a constant. This constant k
Is the valve opening pressure POPEN even if the actual valve opening pressure of the tank pressure control valve 52 differs due to individual differences of the tank pressure control valve 52 or changes due to aging or the like. (= POPE)
N−k, hereinafter referred to as “determination pressure”) is set in advance based on experiments so as to be always lower than the valve opening pressure. Therefore, when the conditional expression (4) is satisfied, it can be reliably determined that the learning value PTNKWG has not risen to the actual valve opening pressure.

In step 332, the above conditional expression (4)
Is satisfied, the ECU 30 determines in step 334 that the temporary update process execution flag XRNW
After setting LOW to “1”, in step 340, the update condition satisfaction flag XRNWCOND is set to “1”. Here, the provisional update processing execution flag XRNWLOW
Is to determine whether or not the learning value PTNKWG should be updated when the above conditions (a) to (f) are not satisfied. This temporary update processing execution flag XRNW
When LOW is set to “1”, the learning value PTNKWG is updated even when the above conditions (a) to (f) are not satisfied.

On the other hand, at step 330, the tank pressure P
If it is determined that TNKW is less than the current learning value PTNKWG, the learning value PTNKW is determined in step 332.
In any case where it is determined that G is equal to or higher than the determination pressure (POPEN-k), the ECU 30 determines in steps 341 and 34
In 2, the update timing determination flag XRNWTIME and the update condition satisfaction flag XRNWCOND are both set to “0”. After executing the processing of step 342 or the above-described step 340, the ECU 30 once ends the processing of this routine.

Next, the "failure diagnosis routine" executed based on the learning value PTNKWG will be described with reference to FIG. This routine is executed by the ECU 30 as an interruption process for each predetermined time period.

When the process proceeds to this routine, in step 508, the ECU 30 determines whether or not the operating state of the engine is in a state suitable for starting failure diagnosis. Here, if it is determined that the operating state is not suitable for starting the failure diagnosis, the ECU 30 once ends the processing of this routine.

On the other hand, if it is determined in step 508 that the operating state is suitable for starting the failure diagnosis, in step 510, the ECU 30 compares the tank pressure PTNKW with the learned value PTNKWG. Here, when it is determined that the tank pressure PTNKW is equal to or higher than the learning value PTNKWG, the tank pressure control valve 52 is opened,
Since it is not possible to perform accurate failure diagnosis, ECU
30 ends the processing of this routine once.

On the other hand, in step 510,
If it is determined that the tank pressure PTNKW is less than the learned value PTNKWG, the ECU 30 switches the three-way switching valve 44 to connect the pressure sensor 46 to the canister 20 in step 512. Then, in step 514, the ECU 30 closes the purge amount adjustment valve 42 to stop the purge control.

Next, in step 516, the ECU 30 calculates the increase rate △ PCAN of the canister pressure PCAN per unit time, and then in step 518, compares the increase rate △ PCAN with the determination value △ PCAN1. Here, the rising rate △ PCAN is equal to the judgment value △ PCAN1.
If it is determined that the above is the case, the ECU 30 determines that there is a damaged portion in the pressure passage system such as the housing 21 of the canister 20 and the purge passage 62, and that the atmosphere has penetrated into the interior from the damaged portion. In step 520, the failure flag XWCAN is set to "1".

On the other hand, in step 518, the rate of increase △
When it is determined that PCAN is less than the determination value △ PCAN1, the ECU 30 determines that the pressure passage system has not been damaged, and sets the failure flag XWCAN to “0” in step 521.

After executing steps 520 and 521, the ECU 30 switches the three-way switching valve 44 to connect the pressure sensor 46 to the fuel tank 10 in step 522. Then, in step 524, after restarting the purge control, the ECU 30 once ends the processing of this routine.

As described above, the failure diagnosis of the evaporative fuel treatment system is performed when the tank pressure PTNKW is the learning value PTNKWG.
Only executed when less than. Therefore, when the learning value PTNKWG is initialized and becomes lower than the actual valve opening pressure of the tank pressure control valve 52,
The tank pressure P reaches a pressure value at which the tank pressure control valve 52 opens.
Even though TNKW has not risen, it is determined that failure diagnosis cannot be performed.

In this respect, in the present embodiment, the learning value PTNK
By executing the update of the WG at an earlier timing, the execution frequency of the failure diagnosis is prevented from decreasing as much as possible even when the learning value PTNKWG is initialized. The procedure for updating the learning value PTNKWG will be described below. FIG. 7 shows each process of the "valve opening pressure learning routine", and this routine is executed by the ECU 30 as an interruption process at every predetermined time period.

When the processing shifts to this routine, the ECU
30, at steps 410 and 412, an update time determination flag XRNWTIME and an update condition satisfaction flag XRN
It is determined whether or not WCOND is “1”.
Here, these flags XRNWTIME, XRNWCO
If it is determined that one of the NDs is “0”, the ECU 3
0 ends the processing of this routine once.

On the other hand, in steps 410 and 412,
When it is determined that each of the flags XRNWTIME and XRNWCOND is “1”, in step 414, the ECU 30 sets the current learning value PTNKWG as the learning value PTNKWGB in the previous control cycle.

Next, in step 416, the ECU 30 determines whether or not the temporary update process execution flag XRNWLOW is "0". If a positive determination is made here, E
In step 418, the CU 30 updates the learning value PTNKWG based on the following equation (5). PTNKWG = PTNKWGB + (PTNKW−PTNKWGB) / 4 (5) On the other hand, if a negative determination is made in step 416 (X
RNWLOW = "1"), that is, each of the above conditions (a) to
If any of (f) is not satisfied, but the tank pressure PTNKW is equal to or higher than the learning value PTNKWG and the learning value PTNKWG is lower than the determination pressure (POPEN-k), the ECU 30 proceeds to step 419.
The learning value PTNKWG is updated based on the following equation (6). PTNKWG = PTNKWGB + (PTNKW−PTNKWGB) / 2 (6) As described above, in the present embodiment, instead of setting the learning value PTNKWG equal to the tank pressure PTNKW, the tank pressure PTNKW is determined by a predetermined smoothing number. It is calculated by performing better processing. Therefore, this learning value PTNK
The change in WG is slower than the change in tank pressure PTNKW. As described above, the learning value PTNKWG is changed to the tank pressure P
By calculating the average value of TNKW, even when the tank pressure PTNKW fluctuates due to, for example, vehicle vibration, the influence of such fluctuation on the update of the learning value PTNKWG can be eliminated.

When the learning value PTNKWG is updated after a negative determination is made in step 416 (step 419), since the smoothing number is set to "1/2", the change in the tank pressure PTNKW is not changed. The learning value PTNKWG is updated so as to respond more sensitively. Therefore, the learning value PTNKWG is earlier than the tank pressure PTNK.
It converges to W.

On the other hand, if the learning value PTNKWG is updated after an affirmative determination is made in step 416 (step 418), the learning value PTNKWG is set to "1/4" because the smoothed number is set to "1/4". It is less susceptible to fluctuations in pressure PTNKW. Therefore, the learning value PTN
Since the KWG stably converges on the actual valve opening pressure of the tank pressure control valve 52, the learning accuracy can be improved.

After executing the processing in steps 418 and 419, the ECU 30 compares the updated learning value PTNKWG with the upper limit value PTNKWGHI which is a guard value for the learning value PTNKWG in step 420. Here, the learning value PTNKWG is equal to the upper limit value PTNKWGH.
If it is determined that the learning value PTNKWG is greater than the upper limit value P in step 422,
Set equal to TNKWGHI. This step 42
The processing of 0,422 is for preventing erroneous learning of the learning value PTNKWG.

After executing the processing of step 422, or when it is determined in step 420 that the learning value PTNKWG is equal to or less than the upper limit value PTNKWGHI,
U30 determines in step 424 that the learning value PTNKWG
Is stored in the backup RAM 31c, and then the processing of this routine is temporarily terminated.

The solid line and the one-dot chain line shown in FIG. 8 indicate the learning value PTNKWG and the tank pressure PT updated as described above.
NKW indicates a change in NKW, and particularly indicates a learning value PTNKW stored in the backup RAM 31c.
It shows changes when the learning value PTNKWG is updated after the value of G is initialized.

When the engine is started after the contents stored in the backup RAM 31c are initialized, the learning value PTNKWG is set to the initial learning value EPTNKGWG1 at the time of starting (timing t1). Then, after the start of the engine, when the tank pressure PTNKW gradually increases and becomes larger than the current learning value PTNKWG (= EPTNNKWG1) (timing t2), the updating of the learning value PTNKWG is started. As a result, the learning value PTNKWG becomes the tank pressure P
It will rise with TNKW.

Then, the learning value PTNKWG becomes equal to the judgment pressure (P
OPEN-k), then the learning value PTNKW
The update of G is temporarily suspended, and the learning value PTNKWG is set to be equal to the determination pressure (POPEN-k) (timing t3 to t4).

Thereafter, the update timing determination flag XR
NWTIME and update condition satisfaction flag XRNWCOND
Are set to "1", the learning value PTNK is again
The update of the WG is started (timing t4), and the learning value P
TNKWG gradually rises and converges on the tank pressure PTNKW (timing t5).

As described above, in the present embodiment, even when the above conditions (a) to (f) are not satisfied, if the predetermined condition is satisfied, the learning value PTNKWG is updated. Therefore, for example, when the update of the learning value PTNKWG is started only when the above-mentioned conditions (a) to (f) are satisfied (change mode of the learning value PTNKWG in this case) 8 is indicated by a two-dot chain line in FIG. 8 as a comparative example), and the learning value PTNKWG can be converged earlier by the actual valve opening pressure of the tank pressure control valve 52.

Therefore, for example, at timing t3 shown in FIG.
In the comparative example, when the operation of the engine is stopped at the point of time, the update of the learning value PTNKWG has not been executed yet, and the learning value PTNKWG is equal to the initial learning value EPTN.
KWG1 is maintained. Therefore, even if the failure diagnosis is performed when the operation of the engine is restarted, the diagnosis is performed by setting the tank pressure PTNKW to the initial learning value EPTN.
It will be interrupted when it exceeds KWG1. For this reason, there is a possibility that the frequency of executing the failure diagnosis is reduced and the discovery of the failure is delayed.

In this regard, according to the present embodiment, even when the above-described conditions (a) to (f) are not satisfied, the tank pressure PTNKW is equal to or higher than the learning value PTNKWG, and the learning value PTNKWG is the judgment pressure (POPEN-
If it is less than k), the tank pressure PTNKW is updated. Therefore, even when the operation of the engine is stopped at the timing t3 as described above, if the operation of the engine is restarted, the tank pressure PTNK is reduced.
Learning value P where W is greater than initial learning value EPTNKWG1
Failure diagnosis can be performed up to TNKWG.

As a result, according to the present embodiment, the frequency of executing the failure diagnosis can be increased, and the learning value PTNKW
Even if G is initialized, it is possible to detect a failure in the fuel vapor treatment device at an early stage.

The present embodiment described above can be implemented by changing the configuration as follows. In the “valve opening pressure learning routine” of the present embodiment, the learning value PTN is determined according to the provisional update process execution flag XRNWLOW.
The number of averaging when calculating KWG is changed, but the number of averaging is changed by the provisional update processing execution flag XRNWLO.
The learning value PTNKWG may be calculated as a fixed value regardless of the value of W.

The number of updates of the learning value PTNKWG after the backup RAM 31c is initialized is counted, and the greater the number of updates, that is, the learning value P
The smoothing number may be changed to a smaller value as the difference between TNKWG and the actual valve opening pressure becomes smaller.

In the present embodiment, the determination pressure (POPE
N−k) is set to a constant value. However, for example, when there is a correlation between the change tendency of the valve opening pressure of the tank pressure control valve 52 and the number of engine starts, this constant k May be set as a function that changes according to the number of times the engine is started.

In this embodiment, the tank pressure PTNKW is calculated by smoothing the actual tank pressure PTNK, and the learning value PTNK is calculated based on the tank pressure PTNKW.
Although the WG is learned, when a low response sensor is used as the pressure sensor 46, such a smoothing process is not performed, and the learning value PTNK is calculated from the actual tank pressure PTNK.
You can also learn WG.

Hereinafter, technical ideas which can be grasped from the above embodiment will be described together with their effects. (1) In the failure diagnosis device for an evaporative fuel treatment device according to claim 1, the learning means smoothes the detected internal pressure of the fuel tank by a predetermined smoothing number to open the pressure valve. A failure diagnosis device for an evaporative fuel treatment device, characterized by learning valve pressure.

According to the above configuration, even when the internal pressure of the fuel tank fluctuates irrespective of the opening and closing of the pressure valve, the valve opening pressure can be learned after removing the influence of such fluctuation. As a result, learning accuracy can be improved.

(2) In the failure diagnosis device for an evaporative fuel treatment device described in (1) above, the learning means learns the valve opening pressure of the pressure valve when the predetermined learning condition is not satisfied. Wherein the smoothing number in the smoothing process is set to be larger than when the opening pressure of the pressure valve is learned when the learning condition is satisfied. Diagnostic device.

According to the above configuration, when the deviation between the learning value and the actual valve opening pressure of the pressure valve is large, the learning value can be brought closer to the actual valve opening pressure earlier, and the deviation can be increased. Is small, the learning value can be stably converged to the actual valve opening pressure of the pressure valve.

[0089]

According to the present invention, when the learned value stored in the storage means is less than the judgment pressure set lower than the designed valve opening pressure of the pressure valve and the internal pressure of the fuel tank is larger than the learned value. The learning of the valve opening pressure is performed even when a predetermined learning condition is not satisfied. Therefore, even if it takes a long time until the learning condition is satisfied after the learning value of the valve opening pressure stored in the storage means is initialized, the initialized learning value increases. Until the judgment pressure set lower than the designed valve opening pressure of the pressure valve is reached, learning relating to the valve opening pressure of the pressure valve is performed, and the learning value is updated to a larger value. As a result, the failure diagnosis is performed based on the updated learning value, so that the frequency of performing the failure diagnosis can be increased, and the failure of the evaporative fuel treatment device can be found early.

[Brief description of the drawings]

FIG. 1 is a schematic configuration diagram showing an evaporated fuel processing device and a failure diagnosis device thereof.

FIG. 2 is a flowchart showing a procedure for detecting a tank pressure.

FIG. 3 is a flowchart showing a procedure for determining a valve opening pressure learning time.

FIG. 4 is a flowchart showing a procedure for determining a valve opening pressure learning condition.

FIG. 5 is a flowchart showing a procedure for determining a valve opening pressure learning condition.

FIG. 6 is a flowchart illustrating a procedure for diagnosing a failure of the fuel vapor processing apparatus.

FIG. 7 is a flowchart showing a procedure for learning a valve opening pressure.

FIG. 8 is a timing chart showing a learning value of a valve opening pressure and a change in a tank pressure.

FIG. 9 is a configuration diagram showing a general configuration of an evaporative fuel processing apparatus.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 10 ... Fuel tank, 12 ... Air cleaner, 14 ... Intake passage, 16 ... Surge tank, 18 ... Throttle valve, 2
0: Canister, 21: Housing, 21a: Internal space, 22: Filter, 26: Adsorbent, 30: ECU, 3
1 ... memory, 31a ... ROM, 31b ... RAM, 31c
... Backup RAM, 42 ... Purge amount control valve, 44 ...
Three-way switching valve, 46: pressure sensor, 48: vehicle speed sensor, 5
2 ... tank pressure control valve, 54 ... atmosphere introduction valve, 56 ... atmosphere release valve, 62 ... purge passage, 64 ... vapor passage, 66 ... atmosphere introduction passage, 78 ... pressure detection passage.

Continued on the front page (72) Inventor Mamoru Yoshioka 1st Toyota Town, Toyota City, Aichi Prefecture Toyota Motor Corporation F-term (reference) 3G084 BA27 DA33 EA11 EB02 EB06 EB16 EB17 EB20 EB22 EB24 FA00 3G301 HA01 HA14 JB02 JB09 LC01 NA01 NA06 NA08 NB02 NB05 NB11 NC01 NC08 ND22 ND25 ND30 ND40 NE17 NE23 PB08Z PB09Z PF01Z PF16Z

Claims (1)

[Claims] 1. A vapor passage for connecting a fuel tank and a canister to move evaporated fuel generated in the fuel tank to the canister, and opening when a pressure inside the fuel tank becomes equal to or higher than a predetermined valve opening pressure. A pressure valve for permitting the movement of the evaporated fuel as a valve, pressure detecting means for detecting the internal pressure of the fuel tank and the canister, and failure of the pressure passage system including the canister based on the detected internal pressure of the canister Failure determining means for determining the opening pressure of the pressure valve based on the detected internal pressure of the fuel tank when a predetermined learning condition is satisfied; and Storage means for storing a learned value; and prohibiting means for prohibiting the failure determination by the failure determination means when the detected internal pressure of the fuel tank is equal to or greater than the stored learning value. In the failure diagnosis device for an evaporative fuel treatment device, the learning means is configured such that the learning value is less than a determination pressure set lower than a designed valve opening pressure of the pressure valve, and the detected fuel tank When the internal pressure is larger than the learning value, the valve opening pressure is learned even when the learning condition is not satisfied.