JP5582459B2 - Particulate matter detection device and particulate filter failure detection device - Google Patents

Description

  The present invention relates to a particulate matter detection device that detects particulate matter in exhaust gas from an internal combustion engine and a particulate filter failure detection device that collects particulate matter.

  Conventionally, in order to suppress soot discharged from an internal combustion engine such as a diesel engine, a diesel particulate filter (DPF, particulate filter) that collects particulate matter (PM) constituting the soot is provided in the exhaust passage. Sometimes. The DPF can be used repeatedly by periodically performing a regeneration process for removing the PM deposited on the DPF by combustion. However, the DPF may overheat due to the regeneration process and the like, and the DPF may be melted or cracked due to the excessive temperature rise (DPF failure). When a DPF breaks down, the amount of PM that passes through the DPF increases, so there is a risk that the exhaust gas regulations will not be satisfied. In recent years, in response to a request for on-board-diagnostics (OBD) performed by a computer mounted on a vehicle, development of a failure detection device that detects a failure of a DPF is desired. Conventionally, as the failure detection device, a failure detection device using an electrode-type PM sensor (see Patent Literatures 1 and 2) that detects PM has been proposed (see Patent Literature 1).

  An electrode-type PM sensor has an insulator attachment portion on which a pair of electrodes are formed. The PM sensor is provided in the exhaust passage, and is used with a voltage applied between a pair of electrodes. PM contained in the exhaust gas adheres to the adhesion portion of the PM sensor. Since PM is composed of carbon particles and has electrical conductivity, a current flows between the electrodes (energizes) when PM adheres to the adhering portion more than a certain amount. The value of the current is a value according to the amount of PM adhering, that is, a value according to the amount of PM contained in the exhaust gas. Therefore, by reading the current value (resistance value between the electrodes corresponding to the current value) The amount of PM can be detected.

JP 2009-1444577 A Japanese Unexamined Patent Publication No. Sho 62-35252

  By the way, components other than PM, such as an ash component derived from engine oil and a metal component constituting the exhaust passage, may adhere to the adhesion portion of the PM sensor. When a component other than PM adheres, the electrical characteristics between the electrodes change due to the component other than PM, and therefore the timing at which the PM sensor is energized and the detection value output from the PM sensor change. That is, the detection accuracy of the PM amount is lowered. In addition, when trying to determine the presence or absence of a failure of a particulate filter using a PM sensor to which components other than PM are attached, it is determined that the particulate filter is normal despite being a failure, or conversely, although it is normal. There is a risk of misjudgment that it is determined that there is a failure.

  The present invention has been made in view of the above problems, and it is a failure of a particulate matter detection device and a particulate filter that can accurately detect PM even when a component other than PM adheres to the attachment portion of the PM sensor. It is an object of the present invention to provide a particulate filter failure detection apparatus capable of accurately determining the presence or absence of a particulate filter.

In order to solve the above problems, a particulate matter detection device of the present invention is provided in an exhaust passage of an internal combustion engine, and has an insulator adhering portion on which a pair of electrodes are formed. This is equivalent to a current flowing between the pair of electrodes that is energized between the pair of electrodes when it adheres to the adhering portion more than a certain amount and changes according to the amount of particulate matter adhering to the adhering portion during the energization. A PM sensor that outputs a detection value;
Removing means for removing particulate matter adhering to the adhering portion;
Based on a detection value after removal that is a detection value output from the PM sensor after removal of the particulate matter by the removing means, a first determination is made as to whether a conductive component other than the particulate matter is attached to the attachment portion. Adhesion judgment means,
And a first correction unit that corrects the detected value to a value that is smaller by a predetermined value when the first adhesion determination unit determines that the conduction component is adhered to the adhesion part. And

  According to the present invention, the particulate matter attached to the attachment portion is removed by the removing means. Here, when a conduction component other than the particulate matter adheres to the adhering portion, the conduction component works in a direction of energizing between the electrodes, so the output (detected value after removal) of the PM sensor is the initial value. It does not return to (value when no particulate matter is attached). Therefore, the 1st adhesion judgment means can judge whether conduction components other than particulate matter have adhered to the adhesion part by checking the detection value after the removal. In addition, when a conductive component is attached to the attachment portion, it becomes easy to energize between the electrodes due to the conductive component, so that the PM sensor energization timing is advanced and the amount of particulate matter attached is the same. The detection value of the PM sensor increases. In the present invention, when the conduction component adheres to the adhesion portion, the first correction means corrects the detection value to a value that is smaller by a predetermined value. Can be returned to the size side. Therefore, particulate matter can be detected with high accuracy even when a conductive component adheres to the attachment portion.

In addition, the particulate matter detection device of the present invention has an insulating attachment portion provided in an exhaust passage of an internal combustion engine and formed with a pair of electrodes, and particulate matter in the exhaust gas is more than a certain amount in the attachment portion. The PM that energizes between the pair of electrodes when attached, and outputs a detection value corresponding to the current flowing between the pair of electrodes that changes in accordance with the amount of particulate matter attached to the attachment portion during the energization. A sensor,
An emission amount estimating means for estimating an emission amount of a predetermined non-conductive component discharged from the internal combustion engine ;
First temperature acquisition means for acquiring the temperature of the adhering portion;
Second temperature acquisition means for acquiring the exhaust temperature;
A second method for estimating a non-conductive component adhesion amount that is an amount of the non-conductive component adhering to the adhesion portion based on a discharge amount of the non-conduction component and a temperature difference between the temperature of the adhesion portion and the exhaust temperature. Means for estimating the amount of adhesion of
Second correction means for correcting the detected value to a large value by an amount corresponding to the non-conductive component adhesion amount estimated by the second adhesion amount estimation means ;
It is characterized by providing.
In addition, the particulate matter detection device of the present invention has an insulating attachment portion provided in an exhaust passage of an internal combustion engine and formed with a pair of electrodes, and particulate matter in the exhaust gas is more than a certain amount in the attachment portion. The PM that energizes between the pair of electrodes when attached, and outputs a detection value corresponding to the current flowing between the pair of electrodes that changes in accordance with the amount of particulate matter attached to the attachment portion during the energization. A sensor,
An emission amount estimating means for estimating an emission amount of a predetermined non-conductive component discharged from the internal combustion engine;
An exhaust flow rate acquisition means for acquiring an exhaust flow rate;
A second adhesion amount estimating means for estimating a non-conductive component adhesion amount that is an amount of the non-conductive component adhering to the adhesion portion based on the discharge amount of the non-conduction component and the exhaust flow rate;
Second correction means for correcting the detected value to a large value by an amount corresponding to the non-conductive component adhesion amount estimated by the second adhesion amount estimation means;
It is characterized by providing.

A predetermined non-conductive component discharged from the internal combustion engine may adhere to the attachment portion of the PM sensor, and the non-conductive component adhesion amount (non-conductive component adhesion amount) is a non-conductive component discharged from the internal combustion engine. It is thought that there is a correlation with the amount of emissions. According to the present invention, the discharge amount estimating means estimates the discharge amount of the predetermined non-conducting component discharged from the internal combustion engine, so the second adhesion amount judging means determines the discharge amount of the non-conductive component and the adhesion amount . The amount of non-conductive component adhesion can be estimated based on the temperature difference between the temperature of the part and the exhaust gas temperature and the exhaust gas flow rate . In addition, when a non-conductive component is attached to the attachment portion, the non-conductive component makes it difficult for the electrodes to be energized. Therefore, the PM sensor energization timing is delayed, and the amount of particulate matter adhered is the same. However, the detection value of the PM sensor becomes small. In the present invention, if the non-conductive component attachment portion is attached, since the second correction means corrects the detection value to the amount corresponding large value in accordance with the non-conductive component adhesion quantity, the non-conductive component The detection value that has become smaller due to adhesion can be returned to the original size. Therefore, particulate matter can be detected with high accuracy even when a non-conductive component adheres to the attachment portion.

Further, the first adhesion determination means in the present invention includes a first adhesion amount estimation means for estimating a conduction component adhesion amount that is an amount of the conduction component adhered to the adhesion portion based on the detection value after removal,
The first correction means corrects the detected value to a smaller value by an amount corresponding to the conduction component adhesion amount estimated by the first adhesion amount estimation means.

  According to this, since it is considered that there is a correlation between the amount of conduction component adhering to the adhesion part (conduction component adhesion amount) and the detection value of the PM sensor (detection value after removal), the first adhesion amount estimation The means can estimate the conduction component adhesion amount based on the detection value after removal. And since the 1st correction | amendment means makes a detection value small by the part according to the conduction | electrical_connection component adhesion amount, it can return the detection value which became large by the conduction | electrical_connection component adhesion amount to the original magnitude | size.

In addition, the PM sensor in the present invention has a heater for heating the adhesion portion,
The removing means controls the heater to heat the adhering portion to a temperature at which the particulate matter is combusted. Thereby, the particulate matter adhering to the adhering portion can be removed by combustion.

  The first adhesion determination means in the present invention includes a detection value acquisition means for acquiring a detection value after removal by applying an AC voltage between the pair of electrodes after removal of the particulate matter by the removal means. And

  When a conductive component is attached between a pair of electrodes of the PM sensor, a capacitance corresponding to the amount of conductive component attached is generated between the electrodes. Since the detection value acquisition unit applies an alternating current between the pair of electrodes, the generated capacitance can be acquired as a detection value after removal. As a result, even if the conduction component adhesion amount is less than a certain amount that is not energized between the electrodes, it is possible to obtain a detection value after removal that reflects the conduction component adhesion amount.

The exhaust passage in the present invention is provided with a particulate filter that collects particulate matter in the exhaust,
The PM sensor is provided downstream of the particulate filter. Thereby, the particulate matter that has passed through the particulate filter can be detected.

In the present invention, the non-conductive component is ash,
The exhaust passage is provided with a particulate filter that can collect particulate matter in the exhaust and also collect ash,
The PM sensor is provided downstream of the particulate filter,
A collection rate estimating means for estimating the collection rate of the ash by Pas tee particulate filter,
Based on the ash emission amount estimated by the emission amount estimation means and the ash collection rate estimated by the collection rate estimation means, the slip-through amount estimation means for estimating the slip-through amount that is the amount of ash that has passed through the particulate filter and, with a,
The second adhesion amount estimation means estimates the adhesion amount of ash adhering to the adhesion portion as the non-conductive component adhesion amount based on the slipping amount estimated by the slipping amount estimation means.

  According to this, since the PM sensor is provided downstream of the particulate filter, it is possible to detect the particulate matter that has passed through the particulate filter. In addition, since the second adhesion amount estimation means estimates the adhesion amount of the ash adhered to the adhesion portion as the non-conductive component adhesion amount, the particulate matter can be accurately detected even when the ash adheres to the adhesion portion. it can.

In the present invention, a first correspondence storage means for storing a correspondence between the operating state of the internal combustion engine and the amount of ash discharged from the internal combustion engine;
Operating condition acquisition means for acquiring the operating state of the internal combustion engine,
The emission amount estimation means estimates the emission amount of ash based on the operation state of the internal combustion engine acquired by the operation state acquisition means and the correspondence relationship stored in the first correspondence relationship storage means.

  Since ash is a component contained in engine oil that is discharged without being burned by the internal combustion engine, there is a correlation between the amount of ash discharged and the operating state of the internal combustion engine. According to the present invention, since the correspondence relationship between the operating state of the internal combustion engine and the ash emission amount is stored in the first correspondence relationship storage means, the operating state of the internal combustion engine can be determined by referring to the correspondence relationship. The amount of ash emissions considered can be estimated. That is, the ash discharge amount can be accurately estimated.

In the present invention, the second correspondence storage means for storing the correspondence between the travel distance of the vehicle on which the internal combustion engine is mounted and the amount of ash discharged from the internal combustion engine before traveling each travel distance. When,
Mileage acquisition means for acquiring the mileage of the vehicle,
The discharge amount estimation means estimates the discharge amount of ash based on the travel distance of the vehicle acquired by the travel distance acquisition means and the correspondence relationship stored in the second correspondence relationship storage means.

  Since ash is derived from engine oil, it is considered that as mileage of the vehicle increases, the amount of accumulated ash emissions increases. That is, it is considered that the travel distance of the vehicle and the ash discharge amount have a correlation. According to the present invention, since the correspondence relationship between the travel distance of the vehicle and the ash discharge amount is stored in the second correspondence relationship storage means, the travel distance of the vehicle is taken into account by referring to the correspondence relationship. Ash emissions can be estimated. That is, the ash discharge amount can be accurately estimated.

The particulate filter failure detection apparatus of the present invention is
A particulate matter detection device of the present invention;
When the timing at which the output of the PM sensor corrected by the first correction unit or the second correction unit is generated is earlier than the predetermined timing, it is determined that the particulate filter has failed, and when the timing is later than the predetermined timing Comprises a failure presence / absence determining means for determining that the particulate filter is normal.

  According to the failure detection apparatus of the present invention, since the particulate matter detection apparatus of the present invention is used, even if a conductive component or non-conductive component other than the particulate matter adheres to the PM sensor, the particulates Presence or absence of filter failure can be accurately determined. Further, the failure presence / absence determining means determines whether or not there is a failure of the particulate filter based on the timing when the output of the corrected PM sensor is generated, that is, the energization timing of the PM sensor. Even when the detection value of the sensor fluctuates, the determination can be made accurately.

1 is a diagram showing a configuration of an engine system 1. FIG. It is a figure explaining the structure and function of PM sensor 41. FIG. It is a figure explaining how the output of PM sensor 41 changes when conduction components other than PM adhere. It is a figure explaining how the output of PM sensor 41 changes when non-conducting components other than PM adhere. It is the flowchart which showed the correction process 1 which correct | amends the output of PM sensor 41 changed by conduction | electrical_connection component adhesion. It is the figure which showed that the output of PM sensor 41 after reproduction | regeneration does not return to an initial value. It is the figure which illustrated the relationship between the output y0 after reproduction | regeneration, and the conduction component adhesion amount z. It is the figure which illustrated the relationship between conduction component adhesion amount z and the correction coefficient k1. It is the flowchart which showed the correction process 2 which correct | amends the output of PM sensor 41 changed by the ash component adhesion. It is the figure which showed the map of the discharge amount of the ash with respect to the driving | running state of the engine 10, and a travel distance. It is the figure which showed the relationship between the deposition amount to DPF30, and the collection rate of ash. It is the figure which showed the relationship between the adhesion amount w of ash, and the correction coefficient k2. 5 is a flowchart showing a failure detection process for detecting a failure of a DPF 30. It is the figure which showed the output of PM sensor 41 at the time of failure of DPF30, and normal.

  Next, embodiments of a particulate matter detection device and a particulate filter failure detection device according to the present invention will be described with reference to the drawings. FIG. 1 is a diagram showing a configuration of a vehicle engine system 1 in which the present invention is embodied. The engine system 1 includes a diesel engine 10 (hereinafter referred to as an engine) as an internal combustion engine. The engine 10 is provided with an injector 11 for injecting fuel into the combustion chamber. The engine 10 generates power by the fuel injected from the injector 11 being self-ignited in the combustion chamber.

  A diesel particulate filter (DPF) 30 is installed in the exhaust passage 21 of the engine 10. The DPF 30 is a ceramic filter having a known structure. For example, a heat-resistant ceramic such as cordierite is formed into a honeycomb structure so that a large number of cells serving as gas flow paths are staggered on the inlet side or the outlet side. Contained and configured. The exhaust discharged from the engine 10 flows downstream while passing through the porous partition wall of the DPF 30, and PM (particulate matter) contained in the exhaust is collected in the DPF 30 and gradually accumulates. The PM constitutes soot and is composed of carbon particles.

  Since the DPF 30 cannot collect PM indefinitely, when the amount of PM deposited on the DPF 30 (PM deposition amount) increases, a regeneration process is performed to regenerate the DPF 30 by removing the deposited PM by combustion. It is like that. The regeneration process is performed, for example, once or a multi-stage post-injection at a time delayed by a predetermined time from the main fuel injection (main injection) that is performed to obtain the power of the engine 10 (generate output torque). Is done by. Specifically, this post-injection raises the exhaust gas temperature and adds unburned fuel (hydrocarbon, HC) to the oxidation catalyst (DOC, not shown) provided upstream of the DPF 30 to react with the heat of reaction. To burn PM.

  The DPF 30 may overheat due to the regeneration process of the DPF 30. Specifically, when the regeneration process is changed to the idle state, the intake air is throttled in the idle state, so that heat generated due to PM combustion is trapped in the DPF 30, and thereby the DPF 30 is excessively heated. There is. Then, due to the excessive temperature rise, the DPF 30 may be melted down, or the DPF 30 may be cracked by a thermal stress due to a temperature difference in the DPF 30 (failure of the DPF 30). When the DPF 30 breaks down, the PM collection capability is reduced, and the amount of PM discharged outside the vehicle is increased. Therefore, in the engine system 1 according to the present embodiment, the failure of the DPF 30 is detected when the amount of PM discharged outside the vehicle is equal to or greater than a predetermined reference, and the failure of the DPF 30 is detected. The detection method will be described later.

  A PM sensor 41 for detecting the PM amount is provided on the downstream side 21 a of the exhaust passage 21 from the DPF 30. Here, FIG. 2 is a diagram for explaining the structure and function of the PM sensor 41. FIG. 2A shows an enlarged view of a region A in the vicinity of the PM sensor 41 in FIG. In FIG. 2A, the substrate 412 provided in the cover 411 is shown through. FIG. 2A shows the substrate 412 as viewed from the side. As shown in FIG. 2A, the PM sensor 41 includes a cover 411 having a hollow inside, and the cover 411 is provided so as to be exposed in the exhaust passage 21a. The cover 411 is formed with a plurality of holes 411a communicating between the inside and outside of the cover 411, and a part of the exhaust can enter the cover 411 through the holes 411a. Further, the cover 411 is formed with a discharge hole 411b through which exhaust gas that has entered the cover 411 is discharged. 2A shows an example in which a discharge hole 411b is formed at the tip of the cover 411. FIG.

  A substrate 412 is provided in the cover 411. The substrate 412 is a substrate made of an insulator such as alumina. FIG. 2B shows a view of one substrate surface 412a (hereinafter referred to as a surface) of the substrate 412 as viewed from above. As shown in FIG. 2B, a pair of electrodes 413 (electrodes 413a and 413b) are provided on the surface 412a of the substrate 412 so as to be separated from each other and face each other. A constant voltage Vd is applied between the electrodes 413a and 413b by a voltage source 415 (see FIG. 2C). Part of the PM contained in the exhaust gas that has entered the cover 411 adheres to the surface 412a of the substrate 412 (strictly speaking, between the electrodes 413a and 413b). PM that has not adhered to the substrate 412 is discharged from a discharge hole 411 b formed in the cover 411.

  Further, a heater 414 composed of a heating wire such as platinum Pt for heating the substrate 412 is provided on the other substrate surface 412b (hereinafter referred to as the back surface) side of the substrate 412 (see FIG. 2A). . The heater 414 heats the substrate 412 and burns and removes PM adhering to the substrate 412. As a result, the PM amount can be repeatedly detected by the PM sensor 41.

  The electrodes 413a and 413b and the heater 414 are connected to a control circuit 416 (see FIG. 2A). The control circuit 416 measures the current flowing between the electrodes 413a and 413b to detect the amount of PM in the exhaust, and controls the temperature of the heater 414. Here, FIG. 2C is a diagram illustrating a measurement circuit that measures the output of the PM sensor 41. Note that the measurement circuit is configured in the control circuit 416. FIG. 2D is a diagram illustrating the output of the PM sensor 41. Specifically, the change in the output of the PM sensor 41 with respect to the amount of PM attached to the substrate 412 (PM attached amount) (PM sensor 41). A line 101 indicating that the output is also a time change. As shown in FIG. 2C, a voltage source 415 for applying a constant DC voltage Vd is provided between the pair of electrodes 413a and 413b of the PM sensor 41. Since the substrate 412 is made of an insulator and the two electrodes 413a and 413b are separated from each other, the electrodes 413a and 413b are insulated when no PM is attached thereto.

  Even if PM adheres between the two electrodes 413a and 413b, it is not immediately energized. That is, while the amount of PM attached is small (insensitive mass range in FIG. 2D), the electrodes 413a and 413b are not energized. In this case, the output of the PM sensor 41 is not generated. Thereafter, when a certain amount or more of PM adheres with the passage of time, since the PM is composed of carbon particles and has conductivity, the resistance between the two electrodes 413a and 413b decreases, and the electrodes 413a, 413a, The space between 413b is energized. Assuming that the current flowing between the electrodes 413a and 413b during energization is I1, a current detection resistor R1 (shunt resistor) is provided on the line through which the current I1 flows, as shown in FIG. And the output (detection value) according to the current I1 can be obtained by measuring the voltage across the shunt resistor R1.

  Further, since the current I1 flowing between the electrodes 413a and 413b increases as the amount of PM attached to the substrate 412 increases, the output of the PM sensor 41 increases as the amount of PM attached increases as shown by the line 101 in FIG. Is getting bigger. Since the PM adhesion amount and the PM amount in the exhaust gas are considered to be correlated, the PM amount in the exhaust gas can be detected from the output of the PM sensor 41. In FIG. 2D, for convenience of explanation, the line 101 whose output changes linearly with respect to the amount of PM adhesion is shown, but in reality, the output does not always change linearly.

  Further, since the PM that can adhere to the substrate 412 is limited, the output of the PM sensor 41 is saturated when the amount of the PM attached to the substrate 412 exceeds a certain level. Therefore, the control circuit 416 periodically heats the substrate 412 with the heater 414 to burn and remove PM adhering to the substrate 412 (regeneration of the PM sensor 41). Note that the control circuit 416 heats the substrate 412 at, for example, about 700 ° C. as a temperature at which PM is burned and removed. In this case, if the regeneration of the PM sensor 41 is performed at a point P1 shown in FIG. 2D, the output of the PM sensor 41 is reset from the point P1 to an initial value (zero).

  Further, the PM sensor 41 can detect a change in capacitance between the electrodes 413a and 413b. Specifically, an AC voltage can also be applied by the voltage source 415 (see FIG. 2C), and a detection value corresponding to the capacitance between the electrodes 413a and 413b is output by the AC voltage. .

  In the above description, it is assumed that components other than PM do not adhere to the substrate 412 of the PM sensor 41, but in fact, components other than PM may adhere to the substrate 412. Specifically, as components other than PM, there are, for example, conductive components such as metal particles constituting the exhaust passage 21 and ash (Ash) as a non-conductive component. The ash is a component derived from engine oil. Specifically, the component contained in the engine oil is discharged without being burned by the engine 10. If these conductive components and non-conductive components adhere to the substrate 412, the output of the PM sensor 41 is affected depending on the amount of adhesion. Here, FIG. 3 is a diagram for explaining how the output of the PM sensor 41 changes when a conductive component other than PM adheres. Specifically, FIG. 3A is a diagram schematically showing a state in which the conductive component 51 is attached between the electrodes 413a and 413b. 3A corresponds to the BB cross section of FIG. 2B. FIG. 3B shows the output of the PM sensor 41 when the conductive component is adhered (line 120 indicating the output change with respect to the PM adhesion amount). FIG. 3B also shows the output of the PM sensor 41 when the conductive component is not attached (normal) (the line 110 showing the change in the output of the PM sensor 41 with respect to the amount of PM attached).

  As shown in FIG. 3A, when the conductive component 51 adheres between the electrodes 413a and 413b, the conductive component 51 facilitates the conduction between the electrodes 413a and 413b. Therefore, the electrodes 413a and 413b are energized with a smaller amount of PM than when the conductive component 51 is not attached. That is, as shown in FIG. 3B, when the conduction component adheres, the output of the PM sensor 41 is generated with a PM adhesion amount x2 smaller than the normal PM adhesion amount x1. In other words, the timing at which the output of the PM sensor 41 is generated (the timing at which the electrodes 413a and 413b are energized) is earlier than normal. Further, during energization, the current I2 flowing between the electrodes 413a and 413b is larger than the current I1 flowing during normal operation. That is, as shown in FIG. 3B, even when the PM adhesion amount is the same, the output of the PM sensor 41 when the conductive component is adhered is larger than the normal output. Note that how much the timing at which the output of the PM sensor 41 is generated (energization timing) and how much the output of the PM sensor 41 is increased are the amount of conduction component adhering to the substrate 412 (conduction component adhesion amount). It is thought that it is decided by.

  Next, the influence when ash as a non-conductive component adheres to the substrate 412 will be described. Here, FIG. 4 is a diagram for explaining how the output of the PM sensor 41 changes when ash adheres. Specifically, FIG. 4A is a diagram schematically showing a state where the ash 52 is adhered between the electrodes 413a and 413b. FIG. 4B shows the output of the PM sensor 41 when the ash is attached (line 130 showing the output change with respect to the amount of PM attached). FIG. 4B also shows the output of the PM sensor 41 when the ash is not attached (normal time) (the line 110 showing the output change with respect to the amount of PM attached).

  As shown in FIG. 4A, when the ash 52 adheres between the electrodes 413a and 413b, the ash 52 makes it difficult to energize between the electrodes 413a and 413b. Therefore, unless a larger amount of PM is attached than when the ash 52 is not attached, the electrodes 413a and 413b are not energized. That is, as shown in FIG. 4B, when the ash is adhered, the output of the PM sensor 41 is generated with a PM adhesion amount x3 larger than the normal PM adhesion amount x1. In other words, the timing at which the output of the PM sensor 41 is generated (the timing at which the electrodes 413a and 413b are energized) is later than normal. In addition, during energization, the current I3 flowing between the electrodes 413a and 413b is smaller than the current I1 flowing during normal operation. That is, as shown in FIG. 4B, even if the PM adhesion amount is the same, the output of the PM sensor 41 at the time of ash adhesion is smaller than the output at the normal time. Note that how late the timing (energization timing) at which the output of the PM sensor 41 is generated and how small the output of the PM sensor 41 is are the amount of ash adhering to the substrate 412 (non-conducting component adhesion amount). It is thought that it is decided by.

  As described above, when a component other than PM (conducting component, ash) adheres to the substrate 412, the output of the PM sensor 41 changes. Therefore, the PM amount in the exhaust gas is calculated based on the output, or the DPF 30. If it is determined whether or not there is a failure, accurate calculation and determination cannot be performed. Therefore, in the present invention, when a component other than PM adheres, the output of the PM sensor 41 is corrected. The correction method will be described in detail later.

  Returning to the description of FIG. 1, the engine system 1 is provided with a differential pressure sensor 42 that detects a differential pressure across the DPF 30. The differential pressure sensor 42 has one end connected to the exhaust passage 21 upstream of the DPF 30 and the other end connected to the exhaust passage 21 downstream of the DPF 30. An exhaust temperature sensor 43 that detects the exhaust temperature is provided on the upstream side of the DPF 30 in the exhaust passage 21. The intake passage 22 of the engine system 1 is provided with an air flow meter 44 that detects the amount of fresh air and a throttle valve 45 (intake throttle valve) that adjusts the amount of fresh air taken into the engine 10. The engine system 1 is provided with a rotation speed sensor 46 that detects the rotation speed of the engine 10. The rotation speed sensor 46 may be a crank angle sensor that measures the rotation angle of the crank 12 connected from the engine 10, for example. Further, the engine system 1 is provided with a travel distance memory 47 in which information on the travel distance (total travel distance up to the present time) of the vehicle on which the engine system 1 is mounted is stored. The travel distance of the vehicle is calculated based on, for example, a measurement value of a rotary encoder that measures wheel rotation, and information on the calculated travel distance is sequentially stored in the travel distance memory 47. Each of the sensors 41 to 46 and the memory 47 are connected to an ECU 60 described later.

  The engine system 1 includes an ECU 60 that controls the entire engine system 1. The ECU 60 has a normal computer structure, and includes a CPU (not shown) for performing various calculations and a memory 61 in which various information is stored. For example, the ECU 60 detects an operating state based on detection signals from the various sensors, calculates an optimal fuel injection amount, injection timing, injection pressure, and the like according to the operating state, and injects fuel into the engine 10. Control. Further, the ECU 60 executes a correction process for correcting the output of the PM sensor 41, and further executes a failure detection process for detecting a failure of the DPF 30 based on the corrected output of the PM sensor 41. Below, the detail of the correction | amendment process which correct | amends the output of PM sensor 41 first is demonstrated.

  FIG. 5 is a flowchart showing the procedure of the correction process 1 for correcting the output of the PM sensor 41 when a conductive component other than PM adheres to the PM sensor 41 (substrate 412). Hereinafter, with reference to FIG. 5, correction when the conductive component is attached will be described. 5 is executed when the output of the PM sensor 41 is necessary, such as when a failure of the DPF 30 is detected. First, an instruction is given to the control circuit 416 (FIG. 2A) of the PM sensor 41, and the substrate 412 is heated to about 700 ° C. by the heater 414 (S11), that is, the PM adhering to the substrate 412 is burned and removed. Then, the PM sensor 41 is regenerated (S11).

  Here, FIG. 6 is a diagram for explaining how the output of the PM sensor 41 is changed by the process of S11, and specifically shows a line 120 of the output of the PM sensor 41 with respect to the PM adhesion amount. ing. When regeneration of the PM sensor 41 is performed, PM is removed from the substrate 412, while conductive components other than PM that have not been burned by heating at about 700 ° C. remain on the substrate 412. Therefore, for example, if the regeneration of the PM sensor 41 is performed at the point P2 in FIG. 6, the output (line 121) of the PM sensor 41 after the regeneration does not return to the initial value (zero). Then, the output y0 (detected value after removal) of the PM sensor 41 after reproduction is confirmed (S12). Specifically, since a capacitance is generated between the electrodes 413a and 413b due to the conduction component attached to the substrate 412, the capacitance is acquired as the output y0 of the PM sensor 41 (S12). More specifically, an AC voltage is applied between the electrodes 413a and 413b with a voltage source 415 (see FIG. 2C). Then, since a current corresponding to the capacitance flows between the electrodes 413a and 413b, a value corresponding to the current is acquired as the output y0 of the regenerated PM sensor 41 (S12). Thus, by detecting the electrostatic capacitance, even when the resistance between the electrodes 413a and 413b is large (when there is no conductive component to the extent that the electrodes 413a and 413b are energized), the conductive component Information indicating that is attached can be obtained.

  Next, based on the output y0 after reproduction confirmed in S12, the amount z (conduction component adhesion amount) of the conduction component adhering to the PM sensor 41 (substrate 412) is estimated (S13). There is a correlation between the output y0 after reproduction and the conduction component adhesion amount z. Specifically, it is considered that the conduction component adhesion amount z increases as the output y0 after reproduction increases. Therefore, in S13, for example, a map 210 of the conduction component adhesion amount z with respect to the output y0 after reproduction shown in FIG. 7 is stored in the memory 61 (see FIG. 1) in advance. Then, referring to the map 210, the conduction component adhesion amount z corresponding to the current output y0 is obtained. Note that FIG. 7 shows a map 210 in which the conduction component adhesion amount z changes linearly with respect to the output y0 after reproduction. However, in practice, the map 210 does not always change linearly.

  Next, a correction coefficient k1 indicating how much the output of the PM sensor 41 is corrected is calculated based on the conduction component adhesion amount z estimated in S13 (S14). As the conduction component adhesion amount z increases, the electrodes 413a and 413b are more easily energized and the output of the PM sensor 41 increases. That is, it is considered that the amount to be corrected (correction coefficient k1) increases. Therefore, in S14, for example, a map 220 of the correction coefficient k1 for the conduction component adhesion amount z shown in FIG. 8 is stored in the memory 61 (see FIG. 1) in advance. Then, with reference to the map 220, a correction coefficient k1 corresponding to the current conduction component adhesion amount z is obtained. Note that FIG. 8 shows a map 220 in which the correction coefficient k1 changes linearly with respect to the conduction component adhesion amount z, but in reality, it does not always change linearly.

  Next, the output of the PM sensor 41 (line 120 in FIG. 3B) is multiplied by the correction coefficient k1, and the output is corrected in the direction of decreasing (S15). Thereby, it is possible to correct the output when the conduction component is not attached (line 110 in FIG. 3B). Thereafter, the process of the flowchart of FIG. In the above, an example of correcting with the correction coefficient k1 has been described. However, the correction amount of the output of the PM sensor 41 may be calculated in S14, and the output may be subtracted by that correction amount in S15.

  Next, correction of the output of the PM sensor 41 when ash as a non-conductive component adheres to the PM sensor 41 (substrate 412) will be described. FIG. 9 is a flowchart showing the procedure of the correction process 2 for correcting the output. 9 is executed when the output of the PM sensor 41 is necessary, such as when a failure of the DPF 30 is detected. First, the amount of ash discharged from the engine 10 is estimated (S21). Ash is a component that is contained in the engine oil and discharged without being burned by the engine 10, and is largely dependent on the operating state of the engine 10. Therefore, in S21, the ash discharge amount is estimated based on the operating state of the engine 10, for example. Specifically, as shown in FIG. 10 (a), a map 301 (corresponding relationship) of the ash discharge amount with respect to the rotational speed NE and the fuel injection amount Q of the engine 10 as the operating state of the engine 10 is stored in advance in the memory 61 ( (See FIG. 1). Then, referring to the map 301, the ash discharge amount corresponding to the current rotational speed NE and the fuel injection amount Q is obtained. The rotational speed NE of the engine 10 is calculated based on the detection value from the rotational speed sensor 46 (see FIG. 1). The combustion injection amount Q uses a command value for the injector 11.

  In addition, the cumulative discharge amount of ash discharged from the engine 10 increases as the travel distance of the vehicle increases. Therefore, in S21, the ash discharge amount (integrated value) may be calculated based on the travel distance of the vehicle. Specifically, as shown in FIG. 10B, a map 302 (corresponding relationship) of the ash discharge amount with respect to the travel distance is stored in the memory 61 in advance. Then, with reference to the map 302, the ash discharge amount corresponding to the current travel distance is obtained. The travel distance information stored in the travel distance memory 47 is used as the travel distance of the vehicle. FIG. 10B shows a map 302 in which the ash discharge amount changes linearly with respect to the travel distance.

  Next, the ash collection rate by the DPF 30 when the DPF 30 is not broken (normal) is estimated (S22). The DPF 30 collects PM, but can also collect ash. In S22, for example, the ash collection rate is estimated based on the amount of PM or ash deposited on the DPF 30. Here, FIG. 11 shows a line 310 showing the relationship between the PM and ash deposition amount and the ash collection rate. As indicated by the line 310, in the region B where the accumulation amount on the DPF 30 is small (see FIG. 11), the collection rate of ash is not so high, and the collection rate increases as the accumulation amount increases. It is considered a trend. This is based on the property that when PM or ash is collected in the DPF 30, the collected PM and ash collect another PM and ash. That is, if the accumulation amount is small, the collected PM and ash are less likely to be collected by another PM and ash, so the collection rate is considered to be low. Beyond that region B, the collection rate is maintained at a certain level (for example, 80%). In S22, the line 310 of FIG. 11 (a map of the collection rate of ash with respect to the accumulation amount) is stored in advance. Then, referring to the map 310, the ash collection rate corresponding to the current accumulation amount is obtained. Since the PM and ash deposition amounts accumulated in the DPF 30 are considered to have a correlation with the differential pressure across the DPF 30, the deposition amount may be calculated based on the differential pressure across the DPF 30. The detected value of the differential pressure sensor 42 (see FIG. 1) is used as the differential pressure across the DPF 30.

  Next, the amount of ash that has passed through the DPF 30 is estimated based on the amount of ash discharged in S21 and the collection rate estimated in S22 (S23). Next, the attached amount w of ash attached to the PM sensor 41 (substrate 412) is estimated based on the ash slipping amount estimated in S23 (S24). Specifically, the ash adhesion amount w is estimated by, for example, multiplying the ash slipping amount by a predetermined coefficient. Note that the degree of adhesion to the substrate 412 is considered to change depending on the temperature of the substrate 412 and the exhaust temperature (strictly, the temperature difference). Specifically, if the temperature of the substrate 412 is higher than the exhaust temperature, a force due to thermophoresis acts in a direction away from the substrate 412, so that the adhesion rate to the substrate 412 is considered to be small. On the contrary, if the exhaust temperature is higher, the adhesion rate to the substrate 412 is considered to increase. Therefore, in S24, the ash adhesion amount w may be estimated in consideration of the influence of the thermophoresis. What is necessary is just to estimate the temperature of the board | substrate 412 based on the resistance value of the heater 414 (refer Fig.2 (a)), for example. Further, the exhaust temperature may be estimated based on the detection value of the exhaust temperature sensor 43 (see FIG. 1).

  Further, it is considered that the degree of adhesion to the substrate 412 varies depending on the exhaust flow rate. Specifically, it is considered that when the exhaust gas flow rate is large, the flow rate of the exhaust gas becomes high, and as a result, the adhesion rate to the substrate 412 becomes small. Therefore, in S24, the ash adhesion amount w may be estimated in consideration of the exhaust flow rate. In order to calculate the exhaust flow rate, first, the intake air amount is calculated by the air flow meter 44. Then, by correcting the intake air amount by the amount of exhaust expansion corresponding to the exhaust temperature detected by the exhaust temperature sensor 43 or the amount of exhaust compression corresponding to the pressure detected by the pressure sensor (not shown), The exhaust flow rate (volume flow rate) may be calculated.

  Next, a correction coefficient k2 indicating how much the output of the PM sensor 41 is corrected is calculated based on the ash adhesion amount w estimated in S24 (S25). As the ash adhesion amount w increases, the electrode 413a and 413b are less likely to be energized and the output of the PM sensor 41 decreases. That is, the amount to be corrected (correction coefficient k2) is considered to be large. Therefore, in S25, for example, a map 320 of the correction coefficient k2 for the ash adhesion amount w shown in FIG. 12 is stored in the memory 61 (see FIG. 1) in advance. Then, with reference to the map 320, a correction coefficient k2 corresponding to the current ash adhesion amount w is obtained. Note that FIG. 12 shows a map 320 in which the correction coefficient k2 changes linearly with respect to the ash adhesion amount w, but actually, the map 320 does not always change linearly.

  Next, the output of the PM sensor 41 (line 130 in FIG. 4B) is multiplied by the correction coefficient k2, and the output is corrected in the direction of increasing (S26). As a result, the output when the ash is not attached (line 110 in FIG. 4B) can be corrected. Thereafter, the processing of the flowchart of FIG. 9 ends. In the above description, an example in which correction is performed using the correction coefficient k2 has been described. However, the correction amount of the output of the PM sensor 41 may be calculated in S25, and the output may be added by that correction amount in S26.

  As described above, by performing the correction process 1 in FIG. 5 and the correction process 2 in FIG. 9, even if the conductive component and the non-conductive component (ash) are attached to the substrate 412, these components are attached. It can be corrected to the output when not. Therefore, the PM amount in the exhaust gas can be accurately estimated based on the corrected output of the PM sensor 41, and the presence / absence of a failure of the DPF 30 can be accurately determined. Next, a method for determining whether or not the DPF 30 has failed will be described. FIG. 13 is a flowchart illustrating a procedure of failure detection processing that is executed by the ECU 60 and detects a failure of the DPF 30. Note that the processing of the flowchart of FIG. 13 is started, for example, when the ECU 60 is started, and thereafter periodically executed.

  First, the correction process 1 (see FIG. 5) and the correction process 2 (see FIG. 9) described above are executed to correct the output of the PM sensor 41 (S31). Next, based on the corrected output of the PM sensor 41, it is determined whether or not the DPF 30 has failed (S32). Specifically, the presence or absence of a failure of the DPF 30 is determined based on the timing at which the output of the PM sensor 41 is generated (the energization timing at which the electrodes 413a and 413b are energized). Here, FIG. 14 is a diagram for explaining the concept of the determination, and specifically shows the change in the output of the PM sensor 41 with respect to time. The time on the horizontal axis in FIG. 14 is the time after the PM sensor 41 is regenerated. When the DPF 30 breaks down, the amount of PM passing through the DPF 30 increases, so that more PM adheres to the substrate 412 of the PM sensor 41 earlier. That is, when the DPF 30 fails, the energization timing for energizing the electrodes 413a and 413b is advanced, and as a result, the output of the PM sensor 41 is generated earlier. Therefore, in S32, it is determined that the DPF 30 has failed when output occurs earlier than a predetermined timing, and it is determined that the DPF 30 has not failed (normal) when output occurs later.

  Note that “the failure of the DPF 30” specifically refers to a case where the collection rate of the DPF 30 is remarkably reduced due to the failure and the regulation value of OBD (On-board-diagnostics) cannot be satisfied. . The OBD regulation value is set to be stricter than the regulation value by a law such as EURO6. For example, in a specific travel mode, when the legal regulation value is PM = 4.5 mg / km, the OBD regulation value is set to, for example, twice PM = 9.0 mg / km. Therefore, as shown in FIG. 14, a line 701 of an OBD regulation value (PM = 9.0 mg / km) is set as a threshold value for determining whether or not there is a failure. When the output of the PM sensor 41 is indicated by a line 702 on the left side of the threshold line 701, that is, when the energization timing t1 is earlier than the threshold energization timing t0, it is determined that the DPF 30 has failed (S32). ). In this case, for example, the user is notified that a failure has occurred. On the other hand, when the output of the PM sensor 41 is indicated by a line 703 on the right side of the threshold line 701, that is, when the energization timing t2 is later than the threshold energization timing t0, the DPF 30 is not faulty (normally (S32).

  If the correction process of S31 is not performed, the output of the PM sensor 41 changes from the normal line 703 to the fault line 702 because a conductive component other than PM adheres to the PM sensor 41. May end up. In this case, it is erroneously determined that the DPF 30 is malfunctioning despite being normal. On the contrary, when the ash component adheres to the PM sensor 41, the output of the PM sensor 41 may change from the line 702 at the time of failure to the line 703 at the time of normal operation. In this case, it is erroneously determined that the DPF 30 is normal even though the DPF 30 is out of order. In the present invention, since the correction process is performed in S31, such erroneous determination can be prevented. After the process of S32, the process of the flowchart of FIG.

  As described above, by determining whether or not the DPF 30 has failed at the timing when the output of the PM sensor 41 is generated, the determination can be made with high accuracy. On the other hand, if the presence / absence of a failure of the DPF 30 is determined based on the absolute value of the output of the PM sensor 41, the electrical resistance of the PM adhering between the electrodes 413a and 413b varies greatly due to the influence of the temperature. There is a case.

  The particulate matter detection device and the particulate filter failure detection device according to the present invention are not limited to the above-described embodiments, and can be variously modified without departing from the scope of the claims. For example, in the above embodiment, a PM sensor is used to detect a failure of the DPF. However, a PM sensor is provided on the upstream side of the DPF, and the PM accumulated in the DPF based on the output of the PM sensor. You may make it estimate the amount of accumulation of. According to this, since the output of the PM sensor is corrected, the amount of accumulated PM can be accurately estimated.

  In the above embodiment, the ECU 60 and the heater 414 that execute the processing of S11 in FIG. 5 correspond to the “removing means” of the present invention. The ECU 60 that executes the processes of S12 and S13 in FIG. 5 corresponds to the “first adhesion determination means” of the present invention. The ECU 60 that executes the processes of S14 and S15 in FIG. 5 corresponds to the “first correcting means” of the present invention. The ECU 60 that executes the process of S12 of FIG. 5 corresponds to the “detected value acquisition means” of the present invention. The ECU 60 that executes the process of S13 in FIG. 5 corresponds to the “first adhesion amount estimating means” of the present invention. The ECU 60 that executes the process of S21 in FIG. 9 corresponds to the “emission amount estimation means”, “driving state acquisition means”, and “travel distance acquisition means” of the present invention. The ECU 60 that executes the processes of S21 to S24 in FIG. 9 corresponds to the “second adhesion determination unit” of the present invention. The ECU 60 that executes S25 and S26 of FIG. 9 corresponds to the “second correcting means” of the present invention. The ECU 60 that executes the process of S22 of FIG. 9 corresponds to the “collection rate estimating means” of the present invention. The ECU 60 that executes the process of S23 in FIG. 9 corresponds to “a slip-through amount estimating means” of the present invention. The ECU 60 that executes the process of S24 in FIG. 9 corresponds to the “second adhesion amount estimating means” of the present invention. The memory 61 corresponds to the “first correspondence storage unit” and the “second correspondence storage unit” of the present invention. The ECU 60 that executes the process of S32 in FIG. 13 corresponds to the “failure presence / absence determination means” of the present invention.

1 Engine system 10 Diesel engine (internal combustion engine)
21 Exhaust passage 30 Diesel particulate filter (DPF)
41 PM sensor 411 cover 411a hole 412 substrate 413, 413a, 413b electrode 414 heater 415 voltage source 416 control circuit 60 ECU
61 memory

Claims (11)

An insulator is provided in an exhaust passage of an internal combustion engine, and a pair of electrodes is formed. When particulate matter in the exhaust adheres to the adhesion portion more than a certain amount, the pair of electrodes is energized. A PM sensor that outputs a detection value corresponding to a current flowing between the pair of electrodes that changes in accordance with the amount of particulate matter adhering to the adhering portion during energization;
Removing means for removing particulate matter adhering to the adhering portion;
Based on a detection value after removal that is a detection value output from the PM sensor after removal of the particulate matter by the removing means, a first determination is made as to whether a conductive component other than the particulate matter is attached to the attachment portion. Adhesion judgment means,
And a first correction unit that corrects the detected value to a value that is smaller by a predetermined value when the first adhesion determination unit determines that the conduction component is adhered to the adhesion part. A particulate matter detection device. An insulator is provided in an exhaust passage of an internal combustion engine, and a pair of electrodes is formed. When particulate matter in the exhaust adheres to the adhesion portion more than a certain amount, the pair of electrodes is energized. A PM sensor that outputs a detection value corresponding to a current flowing between the pair of electrodes that changes in accordance with the amount of particulate matter adhering to the adhering portion during energization;
An emission amount estimating means for estimating an emission amount of a predetermined non-conductive component discharged from the internal combustion engine ;
First temperature acquisition means for acquiring the temperature of the adhering portion;
Second temperature acquisition means for acquiring the exhaust temperature;
A second method for estimating a non-conductive component adhesion amount that is an amount of the non-conductive component adhering to the adhesion portion based on a discharge amount of the non-conduction component and a temperature difference between the temperature of the adhesion portion and the exhaust temperature. Means for estimating the amount of adhesion of
Second correction means for correcting the detected value to a large value by an amount corresponding to the non-conductive component adhesion amount estimated by the second adhesion amount estimation means ;
A particulate matter detection device comprising: An insulator is provided in an exhaust passage of an internal combustion engine, and a pair of electrodes is formed. When particulate matter in the exhaust adheres to the adhesion portion more than a certain amount, the pair of electrodes is energized. A PM sensor that outputs a detection value corresponding to a current flowing between the pair of electrodes that changes in accordance with the amount of particulate matter adhering to the adhering portion during energization;
An emission amount estimating means for estimating an emission amount of a predetermined non-conductive component discharged from the internal combustion engine;
An exhaust flow rate acquisition means for acquiring an exhaust flow rate;
A second adhesion amount estimating means for estimating a non-conductive component adhesion amount that is an amount of the non-conductive component adhering to the adhesion portion based on the discharge amount of the non-conduction component and the exhaust flow rate;
Second correction means for correcting the detected value to a large value by an amount corresponding to the non-conductive component adhesion amount estimated by the second adhesion amount estimation means;
A particulate matter detection device comprising: The first adhesion determination means includes first adhesion amount estimation means for estimating a conduction component adhesion amount that is an amount of the conduction component adhered to the adhesion portion based on the detection value after removal,
2. The particle according to claim 1 , wherein the first correction unit corrects the detection value to a smaller value by an amount corresponding to the conduction component adhesion amount estimated by the first adhesion amount estimation unit. A substance detection device. The PM sensor has a heater for heating the adhering portion,
The particulate matter detection device according to claim 1, wherein the removing unit controls the heater to heat the adhering portion to a temperature at which particulate matter is combusted . The first adhesion determination means includes a detection value acquisition means for acquiring the detection value after removal by applying an alternating voltage between the pair of electrodes after removal of the particulate matter by the removal means. The particulate matter detection device according to any one of claims 1, 4, and 5 . The exhaust passage is provided with a particulate filter that collects particulate matter in the exhaust,
The particulate matter detection device according to any one of claims 1 to 6, wherein the PM sensor is provided downstream of the particulate filter . The non-conductive component is ash;
The exhaust passage is provided with a particulate filter capable of collecting particulate matter in the exhaust and collecting ash.
The PM sensor is provided downstream of the particulate filter,
A collection rate estimating means for estimating the collection rate of the ash by pre Symbol particulate filter,
Based on the ash discharge amount estimated by the discharge amount estimation means and the ash collection rate estimated by the collection rate estimation means, a slip-through that estimates the slip-through amount that is the amount of ash that has passed through the particulate filter. comprising a quantity estimating means, and
The second adhesion amount estimation means estimates the adhesion amount of ash adhered to the adhesion portion as the non-conducting component adhesion amount based on the slipping amount estimated by the slipping amount estimation means. The particulate matter detection device according to claim 2 or 3 . First correspondence storage means for storing the correspondence between the operating state of the internal combustion engine and the amount of ash discharged from the internal combustion engine;
Operating state acquisition means for acquiring the operating state of the internal combustion engine,
The emission amount estimation means estimates the emission amount of the ash based on the operation state of the internal combustion engine acquired by the operation state acquisition means and the correspondence relationship stored in the first correspondence relationship storage means. The particulate matter detection device according to claim 8. A second correspondence storage means for storing a correspondence between a travel distance of a vehicle on which the internal combustion engine is mounted and an amount of ash discharged from the internal combustion engine before traveling each travel distance;
Mileage acquisition means for acquiring the mileage of the vehicle,
The discharge amount estimation means estimates the discharge amount of the ash based on the travel distance of the vehicle acquired by the travel distance acquisition means and the correspondence relationship stored in the second correspondence relationship storage means. The particulate matter detection device according to claim 8 or 9. The particulate matter detection device according to any one of claims 7 to 10,
When the timing at which the output of the PM sensor corrected by the first correction unit or the second correction unit is generated is earlier than a predetermined timing, it is determined that the particulate filter has failed, and the predetermined A failure detection device for a particulate filter, comprising: failure presence / absence determination means for determining that the particulate filter is normal when the timing is later than the timing.