JP2013104316A - Relief valve open judgment device and purge system leakage diagnostic device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To highly precisely determine the opening of a relief valve provided in a blockage valve between a fuel tank and a canister, and thereby adequately execute leakage diagnosis.SOLUTION: If tank internal pressure Ptf equal to or larger than reference pressure value Pvo is satisfied (YES in S193), whether or not both tank internal pressure Ptf reduction and canister internal pressure Pc rise are generated at the same time is judged (S195) after sealing the canister (S194). By this judgment, real opening of a relief valve is judged highly precisely. If the relief valve is really opened (YES in S195), leakge diagnosis frequency limitation is strengthened (S197). Therefore, also for leakage diagnosis in the case of internal combustion engine stoppage for a long time, frequency of the leakage diagnosis is made higher by taking advantage of opportunities of leakage diagnosis while certainly preventing fuel vapor leakage from the canister.

Description

  The present invention relates to a relief valve opening determination device for determining whether or not the relief valve is open in a blocking valve that is provided with a solenoid valve and a relief valve in parallel to cut off between a fuel tank and a canister of an internal combustion engine. The present invention relates to a purge system leak diagnosis apparatus.

  As a leak diagnosis for the evaporative fuel processing mechanism of the internal combustion engine including the fuel tank and the canister, the gas is discharged from the evaporative fuel processing mechanism through the canister while the internal combustion engine is stopped, and the internal pressure of the evaporative fuel processing mechanism is reduced. There is known a method for diagnosing the presence or absence of a leak based on a change in internal pressure during the pressure reduction. At the time of pump pressure reduction for such leak diagnosis, the canister needs to be in a state where it can sufficiently adsorb fuel vapor. If the canister is saturated or close to the adsorbed amount, fuel vapor may leak through the canister when the pump is depressurized.

  In order to prevent such a fuel vapor leak at the time of the leak diagnosis, the fuel vapor adsorbed by the canister during refueling is diagnosed on the condition that the purge is sufficiently discharged into the intake air during the subsequent operation of the internal combustion engine. A technique for executing the technique has been proposed (see, for example, Patent Document 1).

  In particular, in an internal combustion engine mounted on a hybrid vehicle, the internal combustion engine is frequently stopped. Therefore, the purge process itself is repeatedly stopped in a short period of time. For this reason, if the purge amount in one internal combustion engine operation is detected, it cannot be determined whether or not the fuel vapor adsorbed by the canister has been sufficiently purged.

  In order to cope with this, Patent Document 1 integrates the purge amount for each operation in an internal combustion engine that is repeatedly operated and stopped, and determines the fuel vapor adsorption state of the canister based on the accumulated purge amount. Based on this determination, the leak diagnosis is executed in a state where the fuel vapor is sufficiently removed from the canister.

  An evaporative fuel processing mechanism is known in which a fuel tank can be sealed by disposing a blocking valve in which an electromagnetic valve and a relief valve are provided in parallel between a canister and a fuel tank (for example, Patent Document 2). , 3). This shut-off valve is closed when the internal combustion engine is stopped, but when the fuel tank internal pressure reaches a predetermined valve opening pressure, the solenoid valve is opened to discharge the gas in the fuel tank to the canister side. When the internal pressure of the fuel tank exceeds a predetermined pressure range in the state where the solenoid valve is not controlled, the relief valve is mechanically opened to discharge the gas in the fuel tank to the canister side. This prevents the fuel tank internal pressure from becoming excessive.

Therefore, a large amount of fuel vapor is introduced into the canister not only during refueling, but also when the block valve is opened, a large amount of fuel vapor is introduced into the canister.
For this reason, in Patent Document 2, if the relief valve is opened at the time of detecting the purge flow (fuel vapor purge amount), the purge flow cannot be accurately detected due to the influence of fluctuations in the pressure of the gas passage. Purge flow detection is prohibited when the valve may open.

  In Patent Document 3, in a situation where the evaporated fuel in the fuel tank is directly introduced into the intake system due to the electromagnetic valve being forcibly opened, the purge valve is closed to stabilize the operation of the internal combustion engine. I am trying.

JP 2010-216287 A (pages 8 to 10, FIG. 2) Japanese Patent Laying-Open No. 2005-256624 (page 8, FIG. 2) JP 2006-118473 A (page 11, FIG. 3)

  In Patent Document 1, a state in which the fuel vapor is sufficiently detached from the canister is determined based on the integrated purge amount as a precondition for leak diagnosis. This integrated purge amount is determined based on the fuel vapor released to the intake system during the operation of the internal combustion engine. It takes into account only the amount. For this reason, the increased amount of adsorption introduced into the canister by opening the solenoid valve or the relief valve in the block valve is not taken into consideration.

  In Patent Document 2, it is detected from the internal pressure of the fuel tank that the possibility of opening the relief valve has increased. When the fuel tank internal pressure becomes high, the relief valve may be opened, so that detection of the fuel vapor purge amount is prohibited. However, this Patent Document 2 does not consider the amount of fuel vapor introduced from the fuel tank into the canister by opening the relief valve. In addition, the fact that the internal pressure of the fuel tank has reached the range of the relief valve opening pressure captures the high possibility of opening the relief valve. Therefore, it cannot be accurately determined whether or not the relief valve has actually opened, and it cannot be determined whether or not fuel vapor has actually been introduced from the fuel tank into the canister.

  In Patent Document 3, the purge valve is only closed when the solenoid valve of the block valve is opened. Since the relief valve opening is unknown, it cannot be determined whether or not fuel vapor has been introduced from the fuel tank into the canister based on the presence or absence of the relief valve opening.

  Even if all of Patent Documents 1 to 3 are combined, it is impossible to determine the relief valve opening with high accuracy. This makes it impossible to accurately determine whether or not fuel vapor has actually been introduced from the fuel tank to the canister when the relief valve is opened.

  Therefore, even if a large amount of fuel vapor is introduced into the canister due to the actual opening of the relief valve, it is judged that the canister has sufficient adsorption capacity with only the integrated purge amount, and the leak diagnosis is performed. There is a risk of it.

  On the other hand, there is a chance of leak diagnosis because the relief valve may be open even though the canister has sufficient adsorption capacity because the relief valve has not actually opened yet. There is a risk of missing.

  It is an object of the present invention to accurately determine the opening of a relief valve provided in a block valve, and to appropriately execute a leak diagnosis based on the presence or absence of the relief valve that has been determined with high accuracy. It is the purpose.

In the following, means for achieving the above-mentioned purpose, and its operation and effect are described.
In the relief valve opening determination device according to claim 1, it is determined whether the relief valve is open in a block valve that includes a solenoid valve and a relief valve in parallel to shut off the fuel tank and the canister of the internal combustion engine. A relief valve opening determination device comprising: tank internal pressure detection means for detecting pressure in the fuel tank; canister internal pressure detection means for detecting pressure in the canister; and tank internal pressure when the solenoid valve is closed. When the pressure in the fuel tank detected by the detection means changes to a higher pressure side than the atmospheric pressure, the canister sealing means for sealing the canister, and the tank internal pressure after the canister is sealed by the canister sealing means The pressure in the fuel tank detected by the detecting means decreases and the canister internal pressure detected by the canister internal pressure detecting means When the force is increased, wherein said relief valve is a and determining the relief valve opening determining means is opened.

The canister sealing means seals the canister when the pressure in the fuel tank is changed to a higher pressure side than the atmospheric pressure and the relief valve may be opened.
After the canister is sealed, the relief valve opening determining means determines whether the relief valve opens when the pressure in the fuel tank detected by the tank internal pressure detecting means decreases and the pressure in the canister detected by the canister internal pressure detecting means increases. Is determined to have opened.

  When the relief valve is actually opened due to an increase in pressure on the fuel tank side, both the canister and the fuel tank are sealed, so the fuel tank side is moved from the fuel tank side to the canister side. Gas is introduced. For this reason, the pressure in the fuel tank detected by the tank internal pressure detecting means decreases, and conversely, the pressure in the canister detected by the canister internal pressure detecting means increases.

  If the relief valve remains closed, such a pressure change does not occur. That is, the influence of the change in the external air pressure and the influence of the outside air temperature are simultaneously generated in the fuel tank and the canister, and the pressure increases or decreases together. Even if the influence of the outside air temperature is biased, only one pressure changes and the other pressure does not change.

  As a result, the relief valve opening determining means determines that the relief valve is opened when both the pressure decrease in the fuel tank and the pressure increase in the canister occur, so that the relief valve provided in the block valve is provided. The actual valve opening at can be determined with high accuracy.

This makes it possible to accurately determine whether fuel vapor has actually been introduced from the fuel tank to the canister when the relief valve is opened.
In the relief valve open determination device according to claim 2, in the relief valve open determination device according to claim 1, the canister sealing means provides a reference pressure value on a higher pressure side than the atmospheric pressure, and this reference The canister is sealed when the pressure in the fuel tank detected by the tank internal pressure detecting means becomes higher than the pressure value.

By providing the reference pressure value on the higher pressure side than the atmospheric pressure, the canister can be sealed in a pressure region in the fuel tank where the possibility of opening the relief valve is high.
Accordingly, the canister can be sealed at a more appropriate timing, and the canister sealed state is not continued more than necessary.

  The relief valve opening determination device according to claim 3, wherein the canister sealing means is provided in an atmosphere opening provided in a passage connecting the canister to the atmosphere side. The canister is hermetically sealed by closing both the valve and a purge valve provided in a purge passage that discharges fuel vapor in the canister to the intake passage side of the internal combustion engine.

Thus, the canister sealing means can seal the canister by closing both the air release valve and the purge valve.
In the purge system leak diagnosis apparatus according to claim 4, when the internal combustion engine is stopped, an evaporative fuel processing mechanism of the internal combustion engine including the fuel tank and the canister is formed with an airtight compartment, and gas is passed from the compartment through the canister. A purge system leak diagnosis device that diagnoses a leak based on an internal pressure state of the section by discharging the relief valve opening determination device according to any one of claims 1 to 3, and the canister A canister fuel adsorption state detection means for detecting a physical quantity reflecting the fuel adsorption state; and when the physical quantity is detected by the canister fuel adsorption state detection means, a count threshold value for limiting the number of leak diagnoses while the internal combustion engine is stopped, Leakage diagnosis count threshold setting means for setting based on the physical quantity, and the count threshold by the leak diagnosis count threshold setting means When set, the relief valve is opened by the leak diagnosis repeating means for repeating the leak diagnosis within the number of times limited based on the count threshold and the relief valve opening determining means of the relief valve opening determining device. If it is determined, there is provided a limit increasing means for strengthening a limit on the number of times of leak diagnosis.

  Since the physical quantity detected by the canister fuel adsorption state detection unit reflects the fuel adsorption state of the canister, the leak diagnosis count threshold setting unit calculates the number of times of leak diagnosis while the internal combustion engine is stopped based on the physical quantity. The count threshold to be limited can be set. The leak diagnosis repeater can repeat the leak diagnosis without limiting the canister until it is limited to the count threshold.

Therefore, even when the internal combustion engine is stopped for a long time, the frequency of leak diagnosis can be increased without saturating the canister.
When the relief valve opening determining means determines that the relief valve has been opened with such processing being performed, the limit increasing means is increasing the number of leak diagnosis.

  In other words, when the relief valve is actually opened due to the pressure increase in the fuel tank, a large amount of fuel vapor is introduced into the canister, so the canister is saturated within the number of times that is limited based on the count threshold that does not consider the relief valve opening. The risk of getting into a state increases. Therefore, the limit increasing means increases the limit on the number of leak diagnosis for the canister in consideration of the actual relief valve opening.

  As described above, the relief valve opening determination means can accurately determine the relief valve opening provided in the blocking valve, and thus can appropriately limit the number of leak diagnosis. Therefore, also for leak diagnosis when the internal combustion engine is stopped for a long period, the frequency of leak diagnosis can be increased by taking advantage of the chance of leak diagnosis while reliably preventing leakage of fuel vapor from the canister.

  6. The purge system leak diagnosis apparatus according to claim 5, wherein the leak diagnosis repeater clears a diagnosis execution counter and sets the count threshold every time the count threshold is set. Count processing for adding the integrated value to the diagnosis execution counter for each leak diagnosis executed after setting of the above, and while the diagnosis execution counter is smaller than the count threshold, the leak diagnosis is allowed to be repeatedly executed. And a leak diagnosis limiting process for prohibiting repeated execution of the valve, and when the relief valve opening determining means determines that the relief valve has been opened, the limit increasing means By adding the integrated value at the time of valve opening, the number of leak diagnosis is restricted more frequently.

  In the leak diagnosis repeating means, the diagnosis execution counter calculated by the counter process and the count threshold set by the leak diagnosis count threshold setting means are compared in the leak diagnosis limiting process as described above. By such a comparison, the leak diagnosis can be repeated until it is limited to the count threshold.

  When the relief increasing valve determining unit determines that the relief valve has been opened, the limit increasing unit adds the integrated value when the relief valve is opened to the diagnosis execution counter. As a result, it is possible to appropriately increase the limit on the number of leak diagnosis.

  Therefore, also for leak diagnosis when the internal combustion engine is stopped for a long period of time, it is possible to increase the frequency of leak diagnosis by taking advantage of the chance of leak diagnosis while reliably preventing leakage of fuel vapor from the canister.

  6. The purge system leak diagnosis apparatus according to claim 6, wherein the limit increasing means determines that the relief valve has been opened by the relief valve opening determination means. When the relief valve is opened based on one or both of the pressure decrease amount in the fuel tank detected by the tank internal pressure detection means and the pressure increase amount in the canister detected by the canister internal pressure detection means An integrated value is set.

  The amount of fuel vapor introduced into the canister when the relief valve is opened corresponds to the amount of gas flowing from the fuel tank side to the canister side via the relief valve. Therefore, the restriction increasing means is based on one or both of the pressure decrease amount in the fuel tank and the pressure increase amount in the canister detected when it is determined that the relief valve is opened. By setting the integrated value, the number of leak diagnosis can be limited with high accuracy.

FIG. 2 is a block diagram showing a drive system in the hybrid vehicle of the first embodiment. 6 is a flowchart of leak diagnosis preliminary processing executed by the ECU according to the first embodiment. 5 is a flowchart of leak diagnosis processing executed by the ECU according to the first embodiment. 5 is a flowchart of leak diagnosis processing executed by the ECU according to the first embodiment. The flowchart of the relief valve opening determination process which ECU of Embodiment 1 performs. The timing chart which shows an example of the leak diagnosis process in case a tank internal pressure is near atmospheric pressure. The timing chart which shows an example of the leak diagnosis process in case the tank internal pressure is away from atmospheric pressure. The block diagram of map MAPcm which sets count threshold value Cm based on the purge fuel concentration memory | storage value fp used by the said leak diagnostic process. 4 is a timing chart illustrating an example of control according to the first embodiment. 4 is a timing chart illustrating an example of control according to the first embodiment. 4 is a timing chart illustrating an example of control according to the first embodiment. The block diagram of map MAPdv which sets the relief valve opening time integration value Dv based on the pressure difference (DELTA) P used in Embodiment 2. FIG. (A), (B) The block diagram of map MAPa, MAPb which sets integrated value A, B based on the tank internal pressure Ptf used by other embodiment.

[Embodiment 1]
<Configuration of Embodiment 1> FIG. 1 is a block diagram of a drive system in a hybrid vehicle to which the above-described invention is applied. This drive system includes an internal combustion engine 2 and an electric motor (motor generators MG1, MG2 to be described later). The internal combustion engine 2 is a gasoline engine. The internal combustion engine 2 includes a fuel supply system 4 and a control system 6.

  This hybrid vehicle is a plug-in hybrid vehicle. Therefore, the battery 12 can be charged from the external power source 8 through the charging mechanism 10. When the electric power of the battery 12 is supplied to the motor generator MG2 by the power control unit 14, a rotational driving force is output from the motor generator MG2.

The rotational driving force from the internal combustion engine 2 and the motor generator MG2 is decelerated by the speed reduction mechanism 16 and transmitted to the drive wheels 18.
A power split mechanism 20 is disposed between the internal combustion engine 2 and the speed reduction mechanism 16, and the rotational driving force of the internal combustion engine 2 is transmitted to the speed reduction mechanism 16 side and another motor generator MG1 side as a generator. It is possible to supply it divided into two.

The two motor generators MG1 and MG2 function as a generator and an electric motor, respectively, and the functions between them can be switched as necessary.
A fuel injection valve 24 is disposed in each intake port 22 corresponding to each cylinder of the internal combustion engine 2. The fuel stored in the fuel tank 26 is pumped to these fuel injection valves 24 by the fuel pump module 28 via the fuel path 28b. By fuel injection control, fuel is injected from the fuel injection valve 24 during intake at a predetermined timing, and is sucked into each cylinder and burned. As a result, the internal combustion engine 2 is operated.

  Further, a fuel temperature sensor 28 a is arranged in a form attached to the fuel pump module 28. The fuel temperature sensor 28 a detects the fuel temperature of the fuel supply system 4, particularly the fuel temperature Tf in the fuel tank 26 here.

The fuel supply system 4 also functions as an evaporative fuel processing mechanism, and includes a fuel tank 26, a canister 36, various passages attached thereto, various valves, various pumps, and the like.
In the fuel tank 26, a fuel sender gauge 30 for detecting the fuel level SGL in the fuel tank 26 by the float 30a is provided. A tank internal pressure sensor 32 is provided above the fuel tank 26 to detect the pressure in the upper space 26a of the fuel tank 26 (tank internal pressure Ptf). The tank internal pressure Ptf (Pa) is actually a differential pressure between the atmospheric pressure and the upper space 26a, but may be an absolute pressure of the upper space 26a.

  The fuel is introduced into the fuel tank 26 at the time of refueling from the fuel inlet pipe 34. The upper space 26 a of the fuel tank 26 is connected to a canister 36 by an evaporated fuel passage 35. In the middle of the evaporative fuel passage 35, there is provided a blocking valve 38 provided in parallel with an electromagnetic valve 38a for blocking the fuel tank 26 and a relief valve 38b.

  The solenoid valve 38a is a solenoid valve that is controlled to open by energization, and the solenoid valve 38a is controlled to be in a valve-open state during refueling. As a result, the upper space 26 a of the fuel tank 26 and the inside of the canister 36 communicate with each other through the evaporated fuel passage 35. For this reason, at the time of refueling, the fuel vapor generated in the upper space 26a of the fuel tank 26 is discharged to the canister 36 side. In the canister 36, the fuel vapor is adsorbed by an adsorbent such as activated carbon housed therein. This prevents the fuel vapor from leaking outside.

  When the electromagnetic valve 38a is closed, that is, when the evaporated fuel passage 35 is blocked and the fuel tank 26 is sealed, the fuel vapor generated in the upper space 26a of the fuel tank 26 is released from the relief valve 38b. Is not discharged to the canister 36 side unless the valve is opened.

  Connected to the canister 36 is an air passage 40 communicating with a fuel inlet box 34 a provided in the fuel inlet pipe 34. The air passage 40 is provided with an air filter 40a on the way. Further, a leak diagnosis pump module 42 is provided in the atmospheric passage 40 at a position closer to the canister 36 than the air filter 40a. An air release valve 42a, which is attached to the leak diagnosis pump module 42 and is configured as a normally open solenoid valve and opens the canister 36 through the air passage 40, and an internal pressure Pc on the canister 36 side are provided. A pressure sensor 42b for detection is provided.

  The canister 36 is connected to an intake passage 46 of the internal combustion engine 2 by a purge passage 44. In particular, it is connected at a position downstream of the throttle valve 48 for adjusting the intake air amount. In the middle of the purge passage 44, a purge control valve 50 as a normally closed electromagnetic valve is arranged.

  The purge is executed by opening the purge control valve 50 and the air release valve 42a when the internal combustion engine 2 is in operation. That is, when the intake negative pressure in the intake passage 46 is introduced into the canister 36 from the purge passage 44 side, the fuel vapor is released from the adsorbent in the canister 36 and the air flow introduced from the atmospheric passage 40 side To be released. The fuel vapor is released from the purge passage 44 through the purge control valve 50 and into the intake air flowing through the intake passage 46 along the airflow. At this time, the purge rate into the intake air is adjusted by the opening degree of the purge control valve 50. The intake air including the purged fuel vapor flowing into the surge tank 52 is distributed to the intake port 22 of each cylinder, and flows into the combustion chamber of each cylinder together with the fuel injected from the fuel injection valve 24 to be combusted. Become.

  In the intake passage 46, an air flow meter 56 is provided between the air filter 54 and the throttle valve 48 to detect an intake air amount GA (g / sec) supplied to the internal combustion engine 2.

  An air-fuel ratio sensor (or oxygen sensor) 60 is provided in the exhaust passage 58 for discharging exhaust gas after combustion from the internal combustion engine 2, and detects the air-fuel ratio or oxygen concentration from the exhaust component for air-fuel ratio feedback control. .

  In addition, an accelerator opening sensor 62 that is provided on an accelerator pedal operated by the vehicle driver and detects the accelerator opening ACCP, an engine speed sensor 64 that detects the crank speed NE of the internal combustion engine 2, an IGSW (ignition switch) 66) Other sensors and switches are provided to output signals, respectively. Examples of other signals include cooling water temperature, intake air temperature, and vehicle speed.

  Detection signals from the fuel temperature sensor 28a, the fuel sender gauge 30, the throttle opening sensor 48a, the air flow meter 56, the air-fuel ratio sensor 60, the accelerator opening sensor 62, the engine speed sensor 64, the IGSW 66, and the like are configured around a microcomputer. Is input to an ECU (electronic control circuit) 70.

  Then, based on such signal data and data stored or calculated in advance, the ECU 70 executes arithmetic processing to control the fuel injection amount from the fuel injection valve 24, the opening degree TA of the throttle valve 48, and the like. To do.

  Further, the ECU 70 performs a purge control process during a period in which the internal combustion engine 2 is operating. In this purge control process, the fuel valve adsorbed from the fuel tank 26 side through the evaporated fuel passage 35 to the canister 36 side by opening the electromagnetic valve 38a with the refueling is taken into the intake passage during the operation of the internal combustion engine. 46 is a process to be released.

In this purge control process, the purge rate is adjusted by duty-controlling the opening of the purge control valve 50, and the fuel vapor adsorbed by the canister 36 is discharged to the intake passage 46 via the purge passage 44. Note that the concentration of the fuel vapor purged at this time (purge fuel concentration) is obtained as a learning value by a calculation performed periodically based on the control deviation amount of the air-fuel ratio in the air-fuel ratio feedback control executed by the ECU 70. ing.
<Operation of Embodiment 1> Next, with respect to the operation of the present embodiment, leak diagnosis preliminary processing (FIG. 2), leak diagnosis processing (FIGS. 3 and 4) and relief valve opening determination processing (FIG. 5) executed by ECU 70. ). Each process is repeatedly executed at regular time intervals. The steps in the flowchart corresponding to the individual processing contents are represented by “S˜”.

  When the leak diagnosis preliminary processing (FIG. 2) is started, it is first determined whether or not it is ready on (S102). The ready-on state indicates that the key-on is performed and the hybrid vehicle is ready for traveling, including during traveling.

If it is not ready-on (NO in S102), the process is terminated as it is. Accordingly, no substantial processing is performed.
When ready-on (YES in S102), it is next determined whether or not it is the first process in the current ready-on state (S104). If it is the first (YES in S104), the purge fuel concentration update history flag is then set to OFF (S106).

  Next, it is determined whether or not purge control is being performed (S108). That is, the internal combustion engine 2 is operated, and it is determined whether or not the fuel vapor from the canister 36 is purged in the intake passage 46 by controlling the opening degree of the purge control valve 50.

If the internal combustion engine 2 is not in an operating state, or if the purge control itself has not been executed even if the internal combustion engine 2 is in operation (NO in S108), the present process is exited.
If the ready-on state continues, YES in step S102 in the next execution cycle, but not the first (NO in S104), it is immediately determined whether purge control is being performed (S108). If the state in which the purge control is not performed continues (NO in S108), the process is exited as it is.

Even if the ready-on state continues, if the purge control is not performed during that time and the ready-off state is set, the state returns to the state determined as NO in step S102.
When the purge control is started in the ready-on state (YES in S108), it is next determined whether or not there is a purge fuel concentration update history (S110).

  As described above, the purge fuel concentration is obtained as a learned value by a calculation that is periodically performed based on the control deviation amount of the air-fuel ratio in the air-fuel ratio feedback control executed by the ECU 70. Thus, since the purge fuel concentration is obtained as a learned value, it takes a certain number of times of learning to obtain a new purge fuel concentration as a highly accurate value in the purge control started after the current ready-on. .

  Therefore, even if the purge control is being performed (YES in S108) in the ready-on state (YES in S102), the purge fuel concentration update history is not generated until a highly accurate learning value is calculated (in S110). NO), the process after step S112 is not performed. Accordingly, when the ready-off state is maintained, the state returns to NO in step S102.

  When learning in the air-fuel ratio feedback control is repeated and a highly accurate learning value is calculated, the purge fuel concentration is updated. Since this results in an update history (YES in S110), the purge fuel concentration update history flag is set to ON (S112). Then, the purge fuel concentration at the time of this update is stored as a purge fuel concentration storage value fp (S114). Then, the diagnosis execution counter Cx is cleared (S116), and this process is exited.

  In the subsequent execution cycle, when the purge control is performed in the ready-on state and the purge fuel concentration update is repeated (YES in S110), steps S112 to S116 are executed every time the purge fuel concentration is updated. The stored value fp always reflects the latest purge fuel concentration.

  Next, the leak diagnosis process (FIGS. 3 and 4) will be described. When this process is started, it is first determined whether or not a leak diagnosis condition is satisfied (S136). As the leak diagnosis condition, for example, a time of about several hours has passed after the ready-off. In addition to this, the fact that leakage diagnosis prohibition (to be described later) is not performed is added as a logical product condition. Furthermore, a condition such that the outside air pressure and the outside air temperature are within a predetermined range may be added as a logical product condition.

Here, if the leak diagnosis condition is not satisfied (NO in S136), the present process is exited as it is, so that the substantial process is not performed.
If the leak diagnosis condition is satisfied (YES in S136), then the state of the purge fuel concentration update history flag is determined (S138). This is to determine whether or not the current leakage diagnosis condition is satisfied immediately after the purge fuel concentration is updated.

  Here, when the purge fuel concentration update history flag is ON, that is, when the current leak diagnosis condition is established immediately after the purge fuel concentration is updated, the purge fuel stored as the purge fuel concentration storage value fp is next stored. It is determined whether or not the density is higher than the reference density (S140). This reference concentration is set as a reference indicating that the canister 36 is in a fuel adsorption state in which there is sufficient fuel vapor adsorption margin below this concentration and that a leak diagnosis is possible.

  Here, if the purge fuel concentration stored value fp is higher than the reference concentration (YES in S140), the leak diagnosis cannot be performed, so the leak diagnosis is prohibited until the next purge fuel concentration update (S142). Therefore, thereafter, until the operation of the internal combustion engine 2 is started and the purge fuel concentration is newly updated by air-fuel ratio feedback control, the leak diagnosis condition in step S136 is not satisfied.

  If the purged fuel concentration stored value fp is equal to or lower than the reference concentration (NO in S140), is the tank internal pressure Ptf detected next as a relative pressure to the atmospheric pressure by the tank internal pressure sensor 32 within a predetermined range? It is determined whether or not (S148). This predetermined range indicates a range where the differential pressure is small with respect to the atmospheric pressure. More specifically, as shown in Expression 1, it is determined whether or not it exists between the lower limit value L (Pa) (<0) and the upper limit value H (Pa) (> 0).

[Formula 1] L ≦ Ptf ≦ H
The solenoid valve 38a is closed in the ready-off state. In such a state, the tank internal pressure Ptf is lower than the lower limit value L or higher than the upper limit value H, that is, the state in which the above equation 1 is not satisfied, the fuel tank 26 that is a compartment sealed by the electromagnetic valve 38a. Indicates that no leak occurred.

  However, in the state where the above equation 1 is satisfied, there is no leak in the fuel tank 26, but there is a leak in the fuel tank 26 when the internal fuel vapor pressure satisfies the above equation 1 due to the temperature of the fuel tank 26. In some cases, the internal fuel vapor pressure cannot be maintained.

  If Equation 1 is satisfied (YES in S148), there is a possibility that there is a leak in the fuel tank 26 as described above. Therefore, the leak diagnosis process (S150) on the canister 36 side and the fuel tank 26 side The leak diagnosis process (S152) is executed. These two leak diagnosis processes (S150, S152) are executed as shown in FIG. The timing chart of FIG. 6 shows an example of a pressure change detected by the pressure sensor 42b at the time of leak diagnosis.

  First, a leak diagnosis process on the canister 36 side is executed (S150). Since the atmospheric release valve 42a is in the open state when the leak diagnosis condition is satisfied, the pressure sensor 42b initially detects the internal pressure of the canister 36 released to the atmosphere. Before the leak diagnosis start timing t0, the atmospheric pressure is Since it is the same, it is 0 (kPa).

  First, as a calibration by the pump in the leak diagnosis pump module 42, the atmosphere is sucked through a reference orifice having a diameter of 0.5 mm in the pump. In this suction state, the pressure between the pump and the reference orifice is detected by the pressure sensor 42b (t0 to t1). Thus, an internal pressure Pc equivalent to the case where a 0.5 mm leak hole is present on the side of the canister 36 to be sucked from now is detected as the φ0.5 hole determination value Pref1.

  Next, by switching the switching valve (actually, the atmospheric release valve 42a itself also functions as a switching valve), the gas is discharged from the canister 36 to the atmosphere without passing through the reference orifice (t1 to t2). In this state, the internal pressure Pc on the canister 36 side is detected by the pressure sensor 42b as shown by the solid line.

  Next, the switching valve is returned to the original position, and the air is sucked through the reference orifice as calibration again by the pump. In this state, the pressure between the pump and the reference orifice is detected by the pressure sensor 42b (t3 to t4). As a result, the second φ0.5 hole determination value Pref2 is detected.

  When the difference between the φ0.5 hole determination value Pref2 measured for the second time and the φ0.5 hole determination value Pref1 measured for the first time is sufficiently small, the accuracy of the φ0.5 hole determination values Pref1 and Pref2 is sufficiently high It can be judged that it is secured. Therefore, the φ0.5 hole determination value Pref2 measured for the second time is compared with the reached value (t2) of the internal pressure Pc measured by reducing the pressure on the canister 36 immediately before.

As indicated by the solid line, if the reached value (t2) of the internal pressure Pc is lower than the second φ0.5 hole determination value Pref2, it is determined that there is no leak in the canister 36.
As indicated by the broken line (t1 to t2), even if the pump drive time elapses to some extent, the internal pressure Pc on the canister 36 side is the second φ0.5 hole determination value Pref2 (or the first φ0.5 hole determination value Pref1). If it is above, it is determined that there is a leak. When the leak is abnormal as described above, the process is terminated, and the next leak diagnosis (S152) on the fuel tank 26 side is not executed.

  Even when the difference between the φ0.5 hole determination value Pref2 measured for the second time and the φ0.5 hole determination value Pref1 measured for the first time is not sufficiently small, the processing is terminated and the next fuel is determined. The leak diagnosis (S152) on the tank 26 side is not executed.

If it is determined by the above-described leak diagnosis on the canister 36 side that there is no leak on the canister 36 side, then a leak diagnosis process on the fuel tank 26 side is executed (S152).
First, the switching valve is switched, the pump in the pump module 42 for leak diagnosis is connected to the canister 36 without passing through the reference orifice, and the electromagnetic valve 38a is further opened (t4 to t5). Then, gas discharge is started from the fuel tank 26 side by the pump in the pump module 42 via the canister 36 and the electromagnetic valve 38a (t5).

In this state, as indicated by the solid line, the internal pressure Pc is detected by the pressure sensor 42b (t5 to t6).
Next, the switching valve is returned to the original position, and the air is sucked through the reference orifice by the pump as calibration as described above, and the pressure between the pump and the reference orifice is detected by the pressure sensor 42b (t6 to t7). As a result, the third φ0.5 hole determination value Pref3 is detected.

A comparison is made between the φ0.5 hole determination value Pref3 measured for the third time and the reached value (t6) of the internal pressure Pc measured immediately before the fuel tank 26 is depressurized.
As indicated by the solid line, if the reached value (t6) of the internal pressure Pc is lower than the third φ0.5 hole determination value Pref3, it is determined that there is no leak in the fuel tank 26.

  As indicated by the broken line (t5 to t6), if the internal pressure Pc on the fuel tank 26 side is equal to or greater than the third φ0.5 hole determination value Pref3 even after a certain amount of pump drive time has elapsed, the fuel tank 26 leaks. Judge that it exists.

  When the canister-side leak diagnosis process (S150) and the fuel tank-side leak diagnosis process (S152) are thus completed, the state of the purge fuel concentration update history flag is next determined (S154). That is, it is determined whether or not the current leak diagnosis process is the first leak diagnosis after the purge fuel concentration is updated.

  Since this is the first leak diagnosis process and the purge fuel concentration update history flag is ON, this process is exited as it is. Thereafter, until a new leak diagnosis condition is established, it is determined as NO in step S136.

  If the tank internal pressure Ptf <L or the tank internal pressure Ptf> H is satisfied in step S148 (NO in S148), there is no possibility that the fuel tank 26 has a leak as described above. It is determined that there is no leak (S172).

  Next, leak diagnosis processing on the canister 36 side is executed (S174). As shown in FIG. 7, the canister side leak diagnosis process executes the same process as the canister side leak diagnosis process in step S150 described above.

  That is, first, calibration is performed using a reference orifice having a diameter of 0.5 mm by a pump in the pump module 42 for leak diagnosis, and the pressure between the pump and the reference orifice is set to φ0.5 hole determination value Pref1 as a pressure sensor. Detection is performed at 42b (t10 to t11).

  Next, by switching the switching valve, the gas is discharged from the canister 36 to the atmosphere without passing through the reference orifice (t11 to t12). In this state, the internal pressure Pc on the canister 36 side is detected by the pressure sensor 42b as shown by the solid line.

  Next, the switching valve is returned to the original position, and the air is sucked through the reference orifice as calibration again by the pump. In this state, the pressure between the pump and the reference orifice is detected by the pressure sensor 42b (t13 to t14). As a result, the second φ0.5 hole determination value Pref2 is detected.

  When the difference between the φ0.5 hole determination value Pref2 measured at the second time and the φ0.5 hole determination value Pref1 measured at the first time is sufficiently small, the φ0.5 hole determination value Pref2 measured at the second time A comparison is made with the reached value (t12) of the internal pressure Pc measured immediately before reducing the canister 36 side.

As indicated by the solid line, if the reached value (t12) of the internal pressure Pc is lower than the second φ0.5 hole determination value Pref2, it is determined that there is no leak in the canister 36.
As indicated by broken lines (t11 to t12), it is determined that there is a leak if the internal pressure Pc on the canister 36 side is equal to or larger than the second φ0.5 hole determination value Pref2 even after a certain amount of pump drive time has elapsed. When the leak is abnormal as described above, the process is terminated, and the next blockade valve sticking determination process (S176) is not executed.

  Even when the difference between the φ0.5 hole determination value Pref2 measured at the second time and the φ0.5 hole determination value Pref1 measured at the first time is not sufficiently small, the process is terminated and the next block valve The sticking determination process (S176) is not executed.

  If it is determined that there is no leak on the canister 36 side, next, a blocking valve adhering determination process is executed (S176). First, after the pump is stopped, the atmosphere opening valve 42a is closed after setting the inside of the canister 36 to atmospheric pressure (t14 to t15). Then, the solenoid valve 38a that has been in the closed state is opened by a valve opening signal (t15 to t16). In this state, the internal pressure Pc is detected by the pressure sensor 42b (t15 to t16).

  Since the situation is determined as NO in step S148, the tank internal pressure Ptf has a certain difference from the atmospheric pressure. For this reason, if the solenoid valve 38a is normally opened by the valve opening signal from the ECU 70, the internal pressure Pc of the canister 36 is clearly changed such as rising as indicated by a solid line or decreasing as indicated by a broken line.

If no clear fluctuation has occurred, it is determined that the solenoid valve 38a is abnormally stuck, the process is terminated, and the next purge fuel concentration update history flag state determination (S178) is not executed.
If there is a clear fluctuation, it is determined that the block valve is not stuck abnormally, and then the state of the purge fuel concentration update history flag is determined (S178). That is, it is determined whether or not the current leak diagnosis process is the first leak diagnosis after the purge fuel concentration is updated.

  Here, since the purged fuel concentration update history flag = ON, that is, the first leak diagnosis process, this process is left as it is. Therefore, thereafter, NO is determined in step S136 until a new leak diagnosis condition is established.

  In the first leak diagnosis after the purge fuel concentration update described above, it is assumed that the leak abnormality is not detected and the ready-on state is entered again. In the ready-on state, only EV traveling is executed, and it is assumed that the internal combustion engine 2 is not operated and is ready-off.

  In this case, at the time of ready-on, the purge fuel concentration update history flag is turned off in the leak diagnosis preliminary process (FIG. 2) (S106), but purge control is not performed because it is only EV travel. Accordingly, the purge fuel concentration is not updated, and no history is generated. Therefore, the purge fuel concentration update history flag remains OFF and the ready-off state is set.

  Therefore, when the leak diagnosis condition is satisfied, in the leak diagnosis process (FIGS. 3 and 4), after YES is determined in step S136, the purge fuel concentration update history flag is OFF in step S138. A count threshold Cm is set based on the fuel concentration storage value fp (S144).

  Specifically, it is set using a map MAPcm as shown by a solid line in FIG. In this map MAPcm, the lower the purged fuel concentration stored value fp (g / l), the larger the count threshold Cm, and the higher the purged fuel concentration stored value fp, the smaller the counted threshold Cm. Note that the count threshold Cm is the maximum value Cmmax when the purge fuel concentration storage value fp = 0 (g / l), and the count threshold Cm = 0 when the purge fuel concentration storage value fp is higher than the limit value fpmax.

  Next, it is determined whether or not the diagnosis execution counter Cx is smaller than the count threshold Cm (S146). The diagnosis execution counter Cx is a counter that is increased as leak diagnosis is executed and the relief valve 38b is opened as will be described later.

  When the purge fuel concentration is updated at the previous ready-on, the diagnosis execution counter Cx is cleared (S116) in the above-described leak diagnosis preliminary process (FIG. 2). For this reason, since the diagnosis execution counter Cx is smaller than the count threshold Cm, YES is determined in the step S146, and the leak diagnosis process in the step S148 and subsequent steps is executed as described in the timing charts of FIGS.

  If YES is determined in step S148, the purged fuel concentration update history flag is determined after the leak diagnosis process (S150, S152) (S154). At this time, since the purge fuel concentration update history flag = OFF, that is, the leak diagnosis processing for the second and subsequent times after the purge fuel concentration is updated, step S156 is executed next. Here, as shown in Expression 2, the integrated value A is added to the diagnosis execution counter Cx.

[Formula 2] Cx ← Cx + A
This integrated value A is the fuel vapor present in the upper space 26a of the fuel tank 26 when the gas is discharged from the fuel tank 26 side through the canister 36 by the fuel tank side leak diagnosis process (S152) executed immediately before. Is a value corresponding to the amount of adsorbent adsorbed in the canister 36. When the internal combustion engine 2 is stopped for several hours in a steady state, the fuel vapor pressure existing in the upper space 26a of the fuel tank 26 and the time taken for the fuel tank side leak diagnosis process (S152) (FIG. 6: t5 Based on t6), the value is set in advance.

Thereafter, until a new leak diagnosis condition is established, it is determined as NO in step S136.
When it is determined NO in step S148, the determination in step S178 is performed after the determination that there is no fuel tank leak (S172), the canister side leak diagnosis process (S174), and the blocking valve adhering determination process (S176). In this determination, since the purge fuel concentration update history flag = OFF, that is, the second or later leak diagnosis process after the purge fuel concentration is updated, the tank internal pressure Ptf> the upper limit value is determined in the next determination in step S148. It is determined whether or not it is H (S180). In this state, in step S148, either tank internal pressure Ptf <lower limit value L or tank internal pressure Ptf> upper limit value H is satisfied.

  If the tank internal pressure Ptf <the lower limit value L, the gas moves from the canister 36 side to the fuel tank 26 side by opening the solenoid valve 38a, so that the amount of fuel adsorbed in the canister 36 increases. There is nothing. However, when the tank internal pressure Ptf> the upper limit value H, high-pressure fuel vapor is introduced from the fuel tank 26 side to the canister 36 side, so that the amount of fuel adsorbed in the canister 36 increases.

  Therefore, if the tank internal pressure Ptf> the upper limit value H is not satisfied (NO in S180), the present process is left as it is, and thereafter, it is determined as NO in step S136 until a new leak diagnosis condition is satisfied. become.

If the tank internal pressure Ptf> the upper limit value H (YES in S180), the integrated value B is added to the diagnosis execution counter Cx as shown in Expression 3 (S182).
[Formula 3] Cx ← Cx + B
This integrated value B is set in advance corresponding to the amount of fuel vapor introduced from the fuel tank 26 side to the canister 36 side by the sealing valve adhering determination process (S176) executed immediately before. Note that the integrated value A described above is set to be larger than the integrated value B.

Thereafter, until a new leak diagnosis condition is established, it is determined as NO in step S136.
Next, the relief valve opening determination process shown in FIG. 5 will be described. When this process is started, it is first determined whether or not it is ready-on (S191). If it is not ready-on (NO in S191), this process is exited.

When ready-on (YES in S191), it is next determined whether or not a pressure change (tank pressure Ptf decrease and canister pressure Pc increase), which will be described later, is on standby (S192).
If the pressure change is not waiting (YES in S192), it is next determined whether or not the tank internal pressure Ptf is equal to or higher than the reference pressure value Pvo (S193).

  The reference pressure value Pvo is a reference pressure value set on the higher pressure side than the atmospheric pressure. This reference pressure value Pvo is set based on the lowest level of the pressure region in which the relief valve 38b may open due to the high tank internal pressure Ptf, taking into account variations in the opening pressure of the relief valve 38b. Thus, the pressure value is set by providing a certain margin on the low pressure side from the lowest level. Therefore, if the pressure is less than the reference pressure value Pvo, the relief valve 38b is not likely to open in a state where gas is discharged from the fuel tank 26 side to the canister 36 side. There is a possibility of opening in a state where gas is discharged from the side to the canister 36 side.

  If tank internal pressure Ptf <reference pressure value Pvo (NO in S193), the canister 36 is opened to the atmosphere (S198), and the process is exited. The canister 36 is opened to the atmosphere by opening the air release valve 42a. If the canister 36 is already open to the atmosphere, this open state is maintained in step S198.

  In the ready-on state (YES in S191), if the tank internal pressure Ptf ≧ reference pressure value Pvo is reached due to the temperature rise of the fuel tank 26 or a decrease in atmospheric pressure (YES in S193), the canister 36 is then sealed. The process is executed (S194). Until just before, the air release valve 42a is in the open state and is open to the atmosphere, so the air release valve 42a is closed. If the purge control valve 50 is not closed, it is completely closed. This seals the canister 36.

  Then, it is determined whether or not the pressure change after the canister 36 is sealed, that is, whether the tank internal pressure Ptf is decreased and the canister internal pressure Pc is increased in the sealed state (S195). The decrease in the tank internal pressure Ptf and the increase in the canister internal pressure Pc are each provided with a change amount reference value to determine whether or not a further change has occurred.

If the pressure change has not yet occurred (NO in S195), the present process is exited.
Assuming that the ready-on state continues (YES in S191), in the next execution cycle, since the canister 36 is sealed and the pressure change standby is made (NO in S192), the tank internal pressure Ptf decreases and the canister internal pressure It is determined whether Pc has increased (S195).

At this time, if the pressure change has not yet occurred (NO in S195), the present process is exited.
After such processing is repeated, if the pressure change, that is, the tank internal pressure Ptf decreases and the canister internal pressure Pc increases (YES in S195), the pressure is increased with the canister 36 in the atmospheric pressure state. It is determined that the relief valve 38b has been opened due to the pressure difference between the fuel tank 26 and the fuel tank 26 (S196).

  By opening the relief valve 38b, the fuel vapor in the upper space 26a of the fuel tank 26 is introduced into the canister 36 through the relief valve 38b and the evaporated fuel passage 35, and the amount of fuel vapor adsorbed in the canister 36 is increased. Will increase. That is, the fuel vapor adsorption capacity of the canister 36 is reduced.

Therefore, as shown in Equation 4, the relief valve opening integrated value Dv is added to the diagnosis execution counter Cx (S197).
[Formula 4] Cx ← Cx + Dv
The integrated value Dv when the relief valve is opened is set in advance corresponding to the amount of fuel vapor introduced from the fuel tank 26 side to the canister 36 side by opening the relief valve 38b. In this way, this processing is exited.

  Assuming that the ready-on state continues (YES in S191), in the next execution cycle, the pressure change standby is not in progress (YES in S192), and therefore processing for determining whether the tank internal pressure Ptf is equal to or higher than the reference pressure value Pvo ( S193) is executed. The subsequent processing is as described above.

  FIG. 9 shows an example of processing according to the present embodiment. The vehicle is in a ready-on state (t20), and then the operation of the internal combustion engine 2 is started (t21). Purge control is started during operation of the internal combustion engine (t22), and learning by air-fuel ratio feedback control is started.

  When the learning value reaches a high-accuracy state by repeating this learning (t23), the purge fuel concentration is updated and the purge fuel concentration update history exists (FIG. 2: YES in S110). Accordingly, the purge fuel concentration update history flag is set to ON (S112), the newly calculated purge fuel concentration is set to the purge fuel concentration storage value fp (S114), and the diagnosis execution counter Cx is cleared (Cx = 0) ( S116).

  Thereafter, the engine is ready-off and the internal combustion engine is also stopped (t24). If the leak diagnosis condition is satisfied when the internal combustion engine is stopped (t25), the purge fuel concentration update history flag is ON at this time. Therefore, in the leak diagnosis processing (FIGS. 3 and 4), the purge fuel concentration stored value fp is greater than the reference concentration. It is determined whether it is high (S140). Assuming that the purged fuel concentration stored value fp ≦ reference concentration (NO in S140), the leak diagnosis process as described with reference to FIG. 6 or FIG. 7 is performed in the processes after step S148 (t25 to t26).

  Since this leak diagnosis process is a process with the purge fuel concentration update history flag = ON, step S156 and step S182 are not executed. Therefore, at the end of the leak diagnosis process (t26), the diagnosis execution counter Cx is not counted up and Cx = 0 remains.

  Next, the ready-on state is entered (t27), and the purged fuel concentration update history flag returns to OFF (S106). During this ready-on period (t27 to t29), the internal combustion engine is stopped only by EV traveling, and purge control is not performed. Therefore, the purge fuel concentration update history flag remains OFF.

  In the ready-on period (t27 to t29), in the relief valve opening determination process (FIG. 5), it is determined that the relief valve 38b is opened by determining YES in step S195 (S196). Accordingly, in response to the introduction of fuel vapor from the upper space 26a of the fuel tank 26 into the canister 36, the count up for the integrated value Dv when the relief valve is opened is executed with respect to the diagnosis execution counter Cx according to the above equation 4. (S197, t28).

  Next, it is ready-off (t29), and then the leak diagnosis condition is satisfied (t30). At this time, since the purge fuel concentration update history flag is OFF, in the leak diagnosis process (t30 to t31), a count threshold Cm is newly calculated from the purge fuel concentration storage value fp stored at timing t23 (S144). Then, the count threshold Cm is compared with the value of the diagnosis execution counter Cx (S146).

  At the previous timing t23, a purge fuel concentration storage value fp that is higher than the purge fuel concentration storage value fp stored previously is stored. Therefore, based on the relationship of the map MAPcm (FIG. 8), the count threshold Cm set at the timing t30 is slightly smaller than before that.

  Here, the diagnosis execution counter Cx = Dv, but since Cx <Cm still (YES in S146), the leak diagnosis process as described in FIG. 6 or FIG. (T30-t31). At this time, the above equation 1 is established (YES in S148), and the leak diagnosis process (S150, S152) described in FIG. 6 is executed.

  After this leak diagnosis processing, the purged fuel concentration update history flag is OFF in step S154, so the count up for the integrated value A is executed with respect to the diagnosis execution counter Cx by the above equation 2 (S156, t31).

  Even in the next ready-on period (t32 to t33), the EV engine is stopped only by EV traveling, and purge control is not performed. For this reason, the purge fuel concentration update history flag remains OFF and is not set to ON.

  Next, the system is ready-off (t33), and the leak diagnosis condition is satisfied (t34). At this time, since the purge fuel concentration update history flag is OFF, in the leak diagnosis process (t34 to t35), the count threshold Cm is calculated from the purge fuel concentration storage value fp stored at timing t23 (S144). Then, the count threshold Cm is compared with the value of the diagnosis execution counter Cx (S146).

  Here, the diagnosis execution counter Cx = Dv + A, but since Cx <Cm (YES in S146), the leak diagnosis process as described in FIG. 6 or FIG. (T34 to t35). At this time, if the tank internal pressure Ptf> the upper limit value H, Expression 1 is not satisfied (NO in S148). Therefore, the leak diagnosis process (S174) and the blocking valve adhering determination process (S176) described in FIG. 7 are executed.

  After this processing, the purged fuel concentration update history flag is OFF (S178), and the tank internal pressure Ptf> the upper limit value H (YES in S180). Therefore, the integrated value B with respect to the diagnosis execution counter Cx according to Equation 3 above. Minutes are counted up (S182, t35).

  Thereafter, the ready-on state is reached (t36 to t40). At this ready time, the internal combustion engine 2 is operated by HV traveling (t37 to t40). Purge control starts during operation of the internal combustion engine (t38), and learning by air-fuel ratio feedback control starts.

  When the learning value reaches a highly accurate state by repeating this learning (t39), the purge fuel concentration is updated and the purge fuel concentration update history exists (YES in S110). Accordingly, the purge fuel concentration update history flag is set to ON (S112), the newly calculated purge fuel concentration is set to the purge fuel concentration storage value fp (S114), and the diagnosis execution counter Cx is cleared (Cx = 0) ( S116).

  Thereafter, the engine is ready-off and the internal combustion engine is also stopped (t40). If the leak diagnosis condition is satisfied when the internal combustion engine is stopped (t41), the purge fuel concentration update history flag is ON at this time. Therefore, in the leak diagnosis processing (FIGS. 3 and 4), the purge fuel concentration stored value fp is greater than the reference concentration. It is determined whether it is high (S140). Assuming that the purged fuel concentration stored value fp ≦ reference concentration (NO in S140), the leak diagnosis process as described in FIG. 6 or FIG. 7 is executed in the processes after step S148 (t41 to t42).

  In this leak diagnosis process, the purged fuel concentration update history flag is ON, so step S156 and step S182 are not executed. Therefore, at the end of the leak diagnosis process (t42), the diagnosis execution counter Cx is not counted up and Cx = 0 remains.

  Next, a ready-on state is entered (t43), and the purged fuel concentration update history flag returns to OFF (S106). During this ready-on period (t43 to t44), only the EV running is performed and the internal combustion engine is stopped and the purge control is not performed. Therefore, the purge fuel concentration update history flag remains OFF.

  Next, the system is ready-off (t44), and the leak diagnosis condition is satisfied (t45). At this time, since the purge fuel concentration update history flag is OFF, in the leak diagnosis process (t45 to t46), the count threshold Cm is calculated from the purge fuel concentration storage value fp stored at timing t39 (S144). Then, the count threshold value Cm is compared with the value of the diagnosis execution counter Cx (S146), and it is determined whether or not to actually execute the leak diagnosis process (the process after S148).

  In the example of FIG. 9, at timing t39, a value lower than the purge fuel concentration storage value fp stored at the previous timing t23 is stored in the purge fuel concentration storage value fp. For this reason, the count threshold Cm set at the timing t45 based on the relationship of the map MAPcm (FIG. 8) is slightly larger than before.

  Thereafter, every time the purged fuel concentration is updated, the count threshold Cm changes according to the stored value fp. Except for the first leak diagnosis immediately after the purge fuel concentration update, at the time of the leak diagnosis, the diagnosis execution counter Cx and the count threshold value Cm, which are added as shown in the above formulas 2 to 4, are compared, and the actual leak diagnosis is performed. Determine whether to execute. In the example of FIG. 9, since Cx ≧ Cm is not satisfied, the leak diagnosis is repeatedly executed.

FIG. 10 shows an example in which Cx ≧ Cm. From timing t50 to t66, the state is the same as that from timing t20 to t36 in FIG.
At timing t66, the vehicle is in a ready-on state. During this ready-on period (t66 to t67), only EV traveling is performed and the internal combustion engine 2 remains stopped. Accordingly, the purge control is not performed, so the purge fuel concentration is not updated. Therefore, the purge fuel concentration update history flag remains OFF, and the value of the diagnosis execution counter Cx is maintained.

  Next, the system is ready-off (t67), and the leak diagnosis condition is satisfied (t68). In the leak diagnosis process (t68 to t69) at this time, the count threshold Cm is calculated from the purged fuel concentration stored value fp stored at the timing t53 (S144), and the count threshold Cm and the value of the diagnosis execution counter Cx are calculated. The comparison is made (S146). Here, since the diagnosis execution counter Cx <Cm is still satisfied (YES in S146), the leak diagnosis process is actually executed in the processes after step S148.

  Since the leak diagnosis process (t68 to t69) at this time is a process with the purge fuel concentration update history flag = OFF, the diagnosis execution counter Cx is incremented unless the tank internal pressure Ptf is smaller than the lower limit value L ( S156 or S182) is executed.

In the example of FIG. 10, the count-up process in step S156 is executed, and the integrated value A is added to the diagnosis execution counter Cx (t69). As a result, Cx ≧ Cm.
Thereafter, the vehicle is in a ready-on state (t70). Even during this ready-on period (t70 to t71), only EV traveling is performed, and the internal combustion engine remains stopped. Accordingly, the purge control is not performed, so the purge fuel concentration is not updated. Therefore, the purge fuel concentration update history flag remains OFF, and the value of the diagnosis execution counter Cx is maintained.

  Next, the system is ready-off (t71), and the leak diagnosis condition is satisfied (t72). If Cx <Cm, the leak diagnosis process (t72 to t73) is executed as indicated by the broken line. However, since Cx ≧ Cm here (NO in S146), the leak diagnosis process after step S148 is not performed. .

  Thereafter, as shown in FIG. 11, as long as the state where the internal combustion engine operation is not performed only by EV traveling at the time of ready-on continues (t80 to t85), there is no change in the state of Cx ≧ Cm. Even the leak diagnosis process is not executed.

  After such a state of Cx ≧ Cm continues, a ready-on state is reached (t86), and the internal combustion engine operation (t87-t90) is performed during this ready-on period (t86-t90). During the operation of the internal combustion engine, the purge fuel concentration is updated (t89) by learning the air-fuel ratio feedback control (from t88). Therefore, the purge fuel concentration update history flag is set to ON (S112), the updated purge fuel concentration is stored as the purge fuel concentration storage value fp (S114), and the diagnosis execution counter Cx is cleared (S116).

  If the new purged fuel concentration stored value fp is equal to or lower than the reference concentration (NO in S140) in the ready-off period (t90 to t93) immediately after that, the processing described in FIG. 6 or FIG. Such leak diagnosis processing is executed (t91 to t92).

  In the next ready-on period (t93 to t95), only EV traveling is performed. During this period, in the relief valve opening determination process (FIG. 5), it is determined that the relief valve 38b is opened (S196) by determining YES in step S195 (S196). Then, the count-up is performed for the integrated value Dv when the relief valve is opened (S197, t94).

  In the leak diagnosis process (t96 to t97) at the next ready-off, the count threshold Cm is calculated based on the purged fuel concentration stored value fp updated at the timing t89 (t96, S144).

In the example of FIG. 11, the purged fuel concentration stored value fp is updated to a considerably high level, and accordingly, the count threshold Cm is set to a small value (t96).
Since Cx <Cm even in this state (YES in S146), the leak diagnosis shown in FIG. 6 or FIG. 7 is executed (t96 to t97).

Thereafter, when the purge fuel concentration storage value fp is not updated and the leak diagnosis condition is satisfied, the purge fuel concentration storage value fp is the same, so that the addition is performed according to the above formulas 2 to 4 based on the same count threshold Cm. By determining the size of the diagnostic execution counter Cx to be executed, it is determined whether or not the leak diagnosis process is executed.
<Relationship between Embodiment 1 and Claims> In the configuration described above, the tank internal pressure sensor 32, the pressure sensor 42b, and the ECU 70 correspond to a relief valve opening determination device and a purge system leak diagnosis device. The tank internal pressure sensor 32 corresponds to tank internal pressure detection means, and the pressure sensor 42b corresponds to canister internal pressure detection means. Steps S193 and S194 of the relief valve opening determination processing (FIG. 5) executed by the ECU 70 are processing as canister sealing means, steps S195 and S196 are processing as relief valve opening determination means, and step S197 is limit increasing means. It corresponds to the processing as. The process of calculating the purge fuel concentration from the control deviation amount of the air-fuel ratio in the air-fuel ratio feedback control executed by the ECU 70 includes the leak diagnosis preliminary process (FIG. 2) and the leak diagnosis process (FIG. 3) as the process as the canister fuel adsorption state detecting means. 4) corresponds to processing as a leak diagnosis count threshold setting means and a leak diagnosis repetition means.
<Effects of Embodiment 1> (1) In the relief valve opening determination process (FIG. 5), the fact that the tank internal pressure Ptf has changed from the atmospheric pressure to the higher pressure side is established as follows: tank internal pressure Ptf ≧ reference pressure value Pvo Whether or not is determined (S193). When the tank internal pressure Ptf ≧ the reference pressure value Pvo is established (YES in S193), the relief valve 38b provided in the blocking valve 38 may be opened, so that the actual opening of the relief valve 38b is performed with high accuracy. First, both the atmosphere release valve 42a and the purge control valve 50 are closed to seal the canister 36 (S194).

  After such sealing of the canister 36, it is determined whether or not the tank internal pressure Ptf detected by the tank internal pressure sensor 32 has decreased and the canister internal pressure Pc detected by the pressure sensor 42b provided in the canister 36 has increased (see FIG. S195).

  When the tank internal pressure Ptf is equal to or higher than the opening pressure of the relief valve 38b and the relief valve 38b is actually opened, the canister 36 and the fuel tank 26 are both sealed, so that the relief valve 38b existing between them. Then, a gas containing fuel vapor is introduced from the fuel tank 26 side to the canister 36 side. For this reason, the tank internal pressure Ptf decreases, and conversely, the canister internal pressure Pc increases.

  If the relief valve 38b remains closed, such a pressure change does not occur. That is, the influence of the change in the outside air pressure and the influence of the outside air temperature are simultaneously generated in the fuel tank 26 and the canister 36, and both the pressures are increased or decreased. Even if the influence of the outside air temperature is biased, only one pressure changes and the other pressure does not change.

  Therefore, when both the decrease in the tank internal pressure Ptf and the increase in the canister internal pressure Pc occur, it is determined that the relief valve 38b has been opened, so that the actual valve opening in the relief valve 38b provided in the blockade valve 38 is increased. The accuracy can be determined.

Thus, it is possible to accurately determine whether or not fuel vapor has actually been introduced from the fuel tank 26 to the canister 36 by opening the relief valve 38b.
(2) The state in which the tank internal pressure Ptf has changed to the higher pressure side than the atmospheric pressure is determined by whether or not the tank internal pressure Ptf ≧ the reference pressure value Pvo is satisfied (S193). As described above, the reference pressure value Pvo is provided on the higher pressure side than the atmospheric pressure and the tank internal pressure Ptf is determined, so that the canister 36 is sealed in the tank internal pressure Ptf region where the relief valve 38b is highly likely to be opened. Can be.

Thus, the canister 36 can be sealed at a more appropriate timing, and the sealed state of the canister 36 is not continued more than necessary.
(3) The internal combustion engine is stopped based on a physical quantity (here, purge fuel concentration) reflecting the fuel adsorption state of the canister by the leak diagnosis preliminary process (FIG. 2) and the leak diagnosis process (FIGS. 3 and 4). The count threshold value Cm that limits the number of leak diagnoses at the time is set. As a result, while the limit condition of diagnosis execution counter Cx <count threshold Cm is satisfied, the leak diagnosis can be repeated without bringing the canister 36 into a saturated state.

  When the relief valve 38b is opened due to the increase in the tank internal pressure Ptf, the diagnosis execution counter Cx is added with the relief valve opening integrated value Dv according to the equation 4 (S197). This increases the limit on the number of leak diagnosis.

  That is, when the relief valve 38b is opened with the tank internal pressure Ptf ≧ the reference pressure value Pvo, a large amount of fuel vapor is introduced into the canister 36. For this reason, there is a possibility that the canister 36 may be saturated within the number of leak diagnosis limited based on the diagnosis execution counter Cx and the count threshold Cm calculated only by the expressions 2 and 3. Therefore, by adding the integrated value Dv when the relief valve is opened to the diagnosis execution counter Cx according to the above equation 4, the limit of the number of leak diagnosis for the canister 36 is strengthened.

  In addition, as described above, in the relief valve opening determination process (FIG. 5), the actual valve opening of the relief valve 38b can be determined with high accuracy, so that the limit of the number of leak diagnosis can be appropriately increased. Therefore, also in the leak diagnosis when the internal combustion engine 2 is stopped for a long period of time, it is possible to increase the frequency of the leak diagnosis by taking advantage of the chance of the leak diagnosis while reliably preventing the fuel vapor leakage from the canister 36.

[Embodiment 2]
<Configuration of Embodiment 2> In this embodiment, the configuration is the same as that of Embodiment 1 except for the relief valve opening determination process (FIG. 5). In the relief valve opening determination processing (FIG. 5), the value of the integrated value Dv at the time of relief valve opening that is added to the diagnosis execution counter Cx as shown in the equation 4 in step S197 is set in advance. The difference is that it is obtained from the map MAPdv shown in FIG. The others are as described with reference to FIG.

  A map MAPdv in FIG. 12 shows the relationship between the pressure difference ΔP and the integrated value Dv when the relief valve is opened. Here, the pressure difference ΔP is a value indicating the amount of pressure increase of the canister internal pressure Pc when the tank internal pressure Ptf decreases and the canister internal pressure Pc increases after the canister 36 is sealed (S194).

  Note that instead of the map MAPdv indicating the relationship between the increase in the canister internal pressure Pc and the relief valve opening integrated value Dv, the decrease in the tank internal pressure Ptf is used as the pressure difference ΔP, and the pressure decrease amount and the relief of the tank internal pressure Ptf. You may use the map showing the relationship with the valve opening integration value Dv.

  Alternatively, both a map MAPdv indicating the relationship between the pressure increase amount of the canister internal pressure Pc and the integrated value Dv when the relief valve is opened, and a map indicating the relationship between the pressure decrease amount of the tank internal pressure Ptf and the integrated value Dv when the relief valve is opened. By using the integrated value Dv when the relief valve is opened, the sum or average value of the two may be added to the diagnosis execution counter Cx.

Alternatively, the relief valve opening integrated value Dv may be obtained using a map showing the relationship between both the pressure increase amount of the canister internal pressure Pc and the pressure decrease amount of the tank internal pressure Ptf and the relief valve opening integrated value Dv. .
<Operation of Embodiment 2> By using such a map, the integration at the time of opening of the relief valve reflects the amount of fuel vapor introduced from the fuel tank 26 into the canister 36 at the time of opening of the relief valve 38b with higher accuracy. The value Dv can be obtained. Then, a more appropriate diagnosis execution counter Cx can be obtained by adding the relief valve opening integration value Dv to the diagnosis execution counter Cx in step S197.
<Relationship between Embodiment 2 and Claims> The same as in Embodiment 1 above.
<Effects of the Second Embodiment> (1) As described in the effects of the first embodiment, the diagnosis execution counter Cx has a higher accuracy, so that the limit on the number of leak diagnosis is more appropriately strengthened. be able to. For this reason, also in leak diagnosis when the internal combustion engine 2 is stopped for a long time, it is possible to increase the frequency of leak diagnosis by taking advantage of the chance of leak diagnosis while reliably preventing leakage of fuel vapor from the canister 36.

[Other embodiments]
In the second embodiment, the integrated value Dv when the relief valve is opened is set from the pressure difference ΔP. However, since the fuel vapor pressure is related to the temperature, the relief is based on the fuel temperature Tf detected by the fuel temperature sensor 28a. The valve opening integrated value Dv may be set. Further, the relief valve opening integrated value Dv may be set based on the pressure difference ΔP and the fuel temperature Tf.

  In each of the above embodiments, the integrated value A in step S156 and the integrated value B in step S182 of the leak diagnosis process (FIGS. 3 and 4) may be obtained from a map as shown in FIG. A map MAPa shown in FIG. 13A is a map for obtaining the integrated value A based on the tank internal pressure Ptf immediately before the fuel tank side leak diagnosis process, and is set in the range from the lower limit value L to the upper limit value H. ing. A map MAPb shown in FIG. 13B is a map for obtaining the integrated value B based on the tank internal pressure Ptf immediately before the blocking valve adhering determination process, and is set in a range higher than the upper limit value H.

  As a result, the effects of the above-described embodiments are produced, and the integrated values A and B are set in more detail by the maps MAPa and MAPb, so that the number of leak diagnosis can be limited with higher accuracy.

  DESCRIPTION OF SYMBOLS 2 ... Internal combustion engine, 4 ... Fuel supply system, 6 ... Control system, 8 ... External power supply, 10 ... Charging mechanism, 12 ... Battery, 14 ... Electric power control unit, 16 ... Deceleration mechanism, 18 ... Drive wheel, 20 ... Power split Mechanism: 22 ... intake port, 24 ... fuel injection valve, 26 ... fuel tank, 26a ... upper space, 28 ... fuel pump module, 28a ... fuel temperature sensor, 28b ... fuel path, 30 ... fuel sender gauge, 30a ... float, 32 ... Tank internal pressure sensor, 34 ... Fuel inlet pipe, 34a ... Fuel inlet box, 35 ... Evaporative fuel passage, 36 ... Canister, 38 ... Sealing valve, 38a ... Solenoid valve, 38b ... Relief valve, 40 ... Atmospheric passage, 40a ... Air filter 42 ... Pump module 42a ... Air release valve 42b ... Pressure sensor 44 ... Purge passage 46 ... Suction Passage, 48 ... throttle valve, 48a ... throttle opening sensor, 50 ... purge control valve, 52 ... surge tank, 54 ... air filter, 56 ... air flow meter, 58 ... exhaust passage, 60 ... air-fuel ratio sensor, 62 ... accelerator open Degree sensor, 64 ... engine speed sensor, 66 ... IGSW, 70 ... ECU, MG1, MG2 ... motor generator.

Claims (6)

A relief valve opening determination device that determines whether the relief valve is open in a block valve that includes a solenoid valve and a relief valve in parallel to cut off between a fuel tank and a canister of an internal combustion engine,
Tank internal pressure detection means for detecting the pressure in the fuel tank;
Canister internal pressure detecting means for detecting the pressure in the canister;
A canister sealing means for sealing the canister when the pressure in the fuel tank detected by the tank internal pressure detecting means when the solenoid valve is closed changes to a higher pressure side than the atmospheric pressure;
After the canister is sealed by the canister sealing means, when the pressure in the fuel tank detected by the tank internal pressure detecting means decreases and the pressure in the canister detected by the canister internal pressure detecting means increases, A relief valve opening determining means for determining that the relief valve has been opened;
A relief valve opening determination device characterized by comprising: 2. The relief valve opening determining device according to claim 1, wherein the canister sealing means provides a reference pressure value on a higher pressure side than the atmospheric pressure, and the tank internal pressure detecting means detects the reference pressure value from the reference pressure value. A relief valve open determination device which seals the canister when the internal pressure becomes high. 3. The relief valve opening determination device according to claim 1, wherein the canister sealing means includes an air release valve provided in a passage connecting the canister to the atmosphere side, and fuel vapor in the canister is taken into the intake air of the internal combustion engine. A relief valve opening determination device characterized in that the canister is hermetically closed by closing a purge valve provided in a purge passage that discharges to the passage side. When the internal combustion engine is stopped, an airtight compartment is formed in the evaporative fuel processing mechanism of the internal combustion engine including the fuel tank and the canister, and the gas is discharged from the compartment through the canister, so that the internal pressure state of the compartment is determined. A purge system leak diagnosis device for diagnosing leaks,
Relief valve opening determination device according to any one of claims 1 to 3,
Canister fuel adsorption state detection means for detecting a physical quantity reflecting the fuel adsorption state of the canister;
When the physical quantity is detected by the canister fuel adsorption state detection means, a leak diagnosis count threshold setting means for setting a count threshold for limiting the number of leak diagnoses while the internal combustion engine is stopped based on the physical quantity;
When the count threshold is set by the leak diagnosis count threshold setting means, the leak diagnosis repeat means for repeating the leak diagnosis within the number of times limited based on the count threshold;
When it is determined by the relief valve opening determining means of the relief valve opening determining device that the relief valve has been opened, a limit increasing means for increasing the number of leak diagnosis,
A purge system leak diagnosis apparatus comprising: 5. The purge system leak diagnosis apparatus according to claim 4, wherein the leak diagnosis repeater clears the diagnosis execution counter every time the count threshold is set and performs the diagnosis every leak diagnosis executed after the count threshold is set. A count process for adding the integrated value to the counter, and a leak diagnosis limit process for permitting repeated execution of leak diagnosis while the diagnosis execution counter is smaller than the count threshold but prohibiting repeated execution of leak diagnosis when the counter becomes smaller And
When the relief increasing valve determining unit determines that the relief valve has been opened, the limit increasing unit adds the integrated value at the time of opening the relief valve to the diagnosis execution counter so that the number of times of leak diagnosis is increased. A purge system leak diagnosis device characterized by strengthening the limit. 6. The purge system leak diagnosis apparatus according to claim 5, wherein the limit increasing means is detected by the tank internal pressure detecting means when the relief valve opening determining means determines that the relief valve is opened. A purge system leak characterized in that the integrated value when the relief valve is opened is set based on one or both of a pressure decrease amount in the fuel tank and a pressure increase amount in the canister detected by the canister internal pressure detecting means. Diagnostic device.