JP2015098823A - Exhaust particulate concentration detecting device for internal combustion engine - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an exhaust particulate concentration detecting device which is suitable to detect the concentration of exhaust particulates exhausted from a diesel engine, for accurately detecting the concentration of PMs in exhaust gas by using a PM sensor which issues an output value appropriate to the attachment amount of the PMs.SOLUTION: A PM sensor 22, including a sensor element 30 which issues an output value appropriate to the attachment amount of the PMs, is provided in an exhaust passage 16. The PM sensor 22 further includes a heater 42 for causing the disappearance of the PMs attached to the sensor element 30. An ECU 24 calculates a concentration detection value for the PMs on the basis of an output rising time (T, T, T) taken after the PM sensor 22 is reset with the disappearance of the PMs but before the output value rises. The concentration detection value is calculated to be a greater value as the output rising time (T, T, T) is shorter.

Description

  The present invention relates to an exhaust particulate concentration detection device for an internal combustion engine, and more particularly to an exhaust particulate concentration detection device suitable for detecting the concentration of exhaust particulate discharged from a diesel internal combustion engine.

  Patent Document 1 below discloses a system that detects the concentration of exhaust particulates (hereinafter referred to as “PM” (Particulate Matter)) discharged from an internal combustion engine using a PM sensor. The PM sensor used in this system includes an insulator and a pair of electrodes disposed thereon. A power supply circuit for applying a voltage between them is connected to the electrode pair.

  The PM sensor insulator is exposed in the exhaust passage. For this reason, PM contained in the exhaust gas is deposited on the insulator surface over time. Since PM is a conductor, the electrical resistance between the electrode pair decreases as the amount of PM deposited increases. The above conventional system utilizes this property to detect the PM concentration by the following method.

  That is, in the above conventional system, a voltage is applied between the electrodes, and a current flowing between the electrode pairs in that state is detected as an output value. This system detects the amount of change in the output value before and after a predetermined measurement time, and calculates the PM concentration based on the amount of change.

  As the PM concentration increases, more PM adheres to the insulator of the PM sensor. And the output value of PM sensor shows the change according to the quantity of PM adhering during measurement time. For this reason, the change amount of the output value has a correlation with the PM concentration, and according to the above method, the PM concentration in the exhaust gas can be detected to some extent accurately.

JP 2012-150028 A JP 2008-190502 A

  However, the amount of change that appears in the output value of the PM sensor does not reflect only the amount of PM that has newly adhered during the measurement time, but is also affected by the amount of PM that has already adhered at the start of the measurement time. Yes. For this reason, when the above measurement time is started at an arbitrary timing, the PM concentration cannot be accurately detected from the change amount of the output value generated before and after the measurement time.

  The PM sensor can burn off PM adhering to the insulator by heating the heater. If the PM is burned out by this function before the start of the measurement time, it is possible to align the PM adhesion state at the start time. However, even if such processing is performed, the amount of change in the output value occurring before and after the measurement time does not necessarily correspond to the PM concentration in the exhaust gas.

  That is, when the PM on the surface of the insulator is removed by the above-mentioned burning, there is no current flow path between the electrode pairs until a certain amount of PM adheres thereafter, and no output value is generated during that period. . The time during which no output value is generated varies greatly depending on the PM concentration in the exhaust gas. For this reason, in the method of starting the measurement time at an appropriate timing, it is not possible to determine at which point in the measurement time the change begins to appear, so in the end, the amount of change that appears in the output value before and after the measurement time. Does not represent the correct PM concentration.

  For the reasons described above, the conventional system is not necessarily optimal in the sense that PM concentration is detected with high accuracy. The present invention has been made to solve the above-described problems, and accurately detects the PM concentration in the exhaust gas using a PM sensor that emits an output value corresponding to the amount of adhesion of PM. It is an object of the present invention to provide a PM concentration detection device for an internal combustion engine capable of performing the same.

  In order to achieve the above object, a first invention is an exhaust particulate concentration detection device for an internal combustion engine, which includes a sensor element that emits an output value according to the amount of exhaust particulate attached, and exhaust particulate attached to the sensor element. A particulate detection sensor that includes a particulate removal mechanism that eliminates exhaust gas, and a concentration detection value that represents the concentration of exhaust particulates based on the output rise time required until the output value rises after reset due to the disappearance of the exhaust particulates The density detection value is calculated to be larger as the output rise time is shorter.

  According to a second invention, in the first invention, the sensor element includes an insulator exposed in an exhaust passage of the internal combustion engine, a pair of electrodes disposed in the insulator, and a gap between the pair of electrodes. A power supply mechanism for applying a voltage, and the particulate removal mechanism includes a heater that heats the insulator and burns exhaust particulates.

  According to a third aspect, in the first or second aspect, the output rise time is a time required for the output value to exceed a determination threshold after the reset.

  According to a fourth aspect of the present invention, in the first or second aspect, the output rise time is a time required for the output value increase rate to exceed a determination threshold after the reset. To do.

  According to a fifth aspect of the present invention, in any one of the first to fourth aspects, the electronic device corrects the concentration detection value based on an exhaust flow velocity of the internal combustion engine, and the correction includes at least a part of the exhaust gas. In the flow velocity range, the lower the exhaust flow velocity, the larger the concentration detection value becomes, and the lower the exhaust flow velocity, the larger the correction rate.

  According to a sixth aspect of the present invention, in the fifth aspect of the present invention, the slope of the change in the correction factor with respect to the change in the exhaust flow rate is greater in the low exhaust flow rate region where the exhaust flow rate is lower than the variable value. It is characterized by being gentle in a high exhaust flow velocity range higher than the value.

  According to a seventh aspect, in any one of the first to sixth aspects, the electronic device corrects the concentration detection value based on a temperature difference between the exhaust gas and the sensor element, and the correction is performed on the exhaust gas. Is performed such that the detected density value becomes smaller as the temperature is higher than that of the sensor element.

  Further, an eighth invention according to any one of the first to seventh inventions, the exhaust particulate concentration estimating means for calculating the estimated value of the exhaust particulate discharged from the internal combustion engine using the exhaust particulate concentration estimation model, Parameter learning means for learning parameters of the exhaust particulate concentration estimation model so that the concentration estimated value approaches the detected concentration value.

  According to the first aspect, it is possible to calculate the exhaust gas particulate concentration detection value based on the output rise time after reset. This output rise time has a high correlation with the PM concentration in the exhaust gas. Therefore, according to the present invention, it is possible to obtain a concentration detection value that accurately represents the PM concentration in the exhaust gas.

  According to the second invention, the sensor element according to the first invention can be realized by an insulator, a pair of electrodes, and a power supply mechanism that applies a voltage therebetween. Further, according to the present invention, the fine particle removing mechanism in the first invention can be realized by the heater.

  According to the third invention, after the reset, the output rise time can be detected by counting the time until the output value of the sensor element exceeds the determination threshold.

  According to the fourth invention, after the reset, the output rise time can be detected by counting the time until the increase rate of the output value of the sensor element exceeds the determination threshold.

  According to the fifth aspect of the present invention, the concentration detection value can be corrected to a larger value as the exhaust flow rate is lower in at least a part of the exhaust flow rate region. In a region where the exhaust flow rate is low, a situation where PM cannot reach the sensor element is likely to occur, and the concentration detection value tends to be too small relative to the actual PM concentration. According to the present invention, it is possible to obtain such a density detection value that cancels out such characteristics and accurately matches the actual PM density.

  According to the sixth aspect of the invention, the gradient of the change in the correction factor with respect to the change in the exhaust flow rate can be made gentler in the high exhaust flow rate region and in the low exhaust flow rate region. In this case, the concentration detection value is corrected with high sensitivity to changes in the exhaust flow velocity in the low exhaust flow velocity region, while the correction sensitivity of the concentration detection value with respect to changes in the exhaust flow velocity is kept low in the high exhaust flow velocity region. Become. In the low exhaust flow rate region, the PM becomes less likely to reach the sensor element as the exhaust flow rate decreases, so the concentration detection value needs to be corrected with high sensitivity in order to offset the influence. On the other hand, in the high exhaust flow rate region, PM tends to reach the sensor element as the exhaust flow rate becomes faster, while PM adhering to the sensor element is easily blown away. For this reason, in the high exhaust flow velocity region, the PM adhesion amount is not easily affected by the exhaust flow velocity, and the need for correction for changes in the exhaust flow velocity is reduced. According to the present invention, the difference between the high exhaust flow velocity region and the low exhaust flow velocity region described above can be appropriately reflected in correction, and a concentration detection value that accurately matches the actual PM concentration in all exhaust flow velocity regions can be obtained. Can do.

  According to the seventh aspect of the invention, the detected concentration value can be corrected to a smaller value as the exhaust gas has a higher temperature than the sensor element. If the exhaust gas has a higher temperature than the sensor element, a phenomenon occurs in which PM in the exhaust gas moves toward the sensor element due to thermophoresis caused by the temperature difference. For this reason, the higher the temperature of the exhaust gas compared to the sensor element, the more easily the amount of PM adhering to the sensor element becomes larger than the actual PM concentration, and therefore the detected concentration value tends to be too large. According to the present invention, it is possible to obtain such a density detection value that cancels out such characteristics and accurately matches the actual PM density.

  According to the eighth invention, it is possible to calculate a concentration estimation value representing the PM concentration using the exhaust particulate concentration estimation model. Furthermore, according to the present invention, it is possible to learn the parameters of the exhaust particulate concentration estimation model so that the estimated concentration value approaches the detected concentration value. In the present invention, a detected density value that accurately matches the actual PM density is obtained. For this reason, if the parameter is learned so that the concentration estimated value matches the detected concentration value, the exhaust particulate concentration estimation model can be maintained in a state where a concentration estimated value that accurately matches the actual PM concentration can be calculated. .

It is a figure for demonstrating the system configuration | structure of Embodiment 1 of this invention. It is sectional drawing of PM sensor with which the system shown in FIG. 1 is provided. It is a figure for demonstrating the output characteristic of PM sensor shown in FIG. It is an example of the map for converting the output rise time detected by PM sensor shown in FIG. 2 into a density | concentration detection value. It is an example of the experimental result which shows the relationship between the output rise time detected by PM sensor, and an exhaust flow velocity. It is an example of the map of the correction coefficient for eliminating the influence which an exhaust gas flow velocity has on a concentration detection value. It is an example of the experimental result which shows the relationship between the output rise time detected by PM sensor, and the temperature difference of exhaust gas and a sensor element. It is an example of the map of the correction coefficient for eliminating the influence which the temperature difference of exhaust gas and a sensor element has on the concentration detection value. FIG. 9A is a diagram showing a temporal change in the exhaust gas flow velocity. FIG. 9B is a diagram showing a temporal change in a concentration detection value that does not consider the influence of the exhaust flow velocity and a concentration estimation value estimated by a PM concentration estimation model learned based on the concentration detection value. . FIG. 10A is a diagram showing a temporal change in the exhaust gas flow velocity. FIG. 10B is a diagram showing a temporal change of the concentration estimation value estimated by the PM concentration estimation model learned based on the concentration detection value excluding the influence of the exhaust flow velocity and the concentration detection value. FIG. 2 is a block diagram of PM concentration detection control executed in the system shown in FIG. 1. It is a flowchart of PM density | concentration detection control performed in the system shown in FIG.

Embodiment 1 FIG.
[System Configuration of Embodiment 1]
FIG. 1 is a diagram for explaining a hardware configuration of a system according to the first embodiment of the present invention. The system of this embodiment includes a diesel-type internal combustion engine 10. The internal combustion engine 10 includes an air flow meter 14 in the intake passage 12. The air flow meter 14 generates an output corresponding to the flow rate of air flowing through the intake passage 12.

  A diesel particulate filter (DPF) 18 is disposed in the exhaust passage 16 of the internal combustion engine 10. The DPF 18 is a filter for removing PM contained in the exhaust gas discharged from the internal combustion engine 10.

  An exhaust gas temperature sensor 20 is disposed upstream of the DPF 18. A PM sensor 22 is disposed downstream of the DPF 18. In addition to these sensors, the internal combustion engine 10 includes an NE sensor for detecting the engine speed NE, an accelerator opening sensor for detecting the accelerator opening AA, and a water temperature for detecting the cooling water temperature THW. A sensor or the like is mounted (both not shown). These sensors, including the air flow meter 14 described above, are all electrically connected to the ECU 24.

  The ECU 24 is an electronic device that includes a microcomputer and a memory therein. In the present embodiment, the ECU 24 stores a program for detecting the PM concentration in the exhaust gas based on the output of the PM sensor 22 and the like. In executing this program, the ECU 24 determines the exhaust flow velocity based on the output of the air flow meter 14, the temperature of the exhaust gas based on the output of the exhaust temperature sensor 20, and the engine speed based on the outputs of the various sensors described above. It is possible to detect physical quantities such as

  FIG. 2 is a cross-sectional view for explaining the configuration of the PM sensor 22 described above. The PM sensor 22 has a housing 26 that is fixed to the exhaust passage 16. A sensor element 30 is held inside the housing 26. The sensor element 30 includes an insulator 32 that is mounted so as to isolate the inside and outside of the exhaust passage 16. Electrode pairs 34 and 36 are arranged on the surface of the insulator 32, that is, the surface exposed to the inside of the exhaust passage 16.

  A power supply mechanism 38 and a current detection mechanism 40 are connected to the electrode pairs 34 and 36. The power supply mechanism 38 can apply a voltage between the electrode pair in accordance with a command from the ECU 24. Further, the current detection mechanism 40 can detect a current flowing between the electrode pairs 34 and 36 and provide the result to the ECU 24 as an output value of the PM sensor 22.

  The sensor element 30 also includes a heater 42 attached to the back surface of the insulator 32 and a heater control mechanism 44 for controlling the heater 42. The heater control mechanism 44 can control the driving of the heater 42 in accordance with a command from the ECU 24.

  The sensor element 30 is covered with a cover 46 inside the exhaust passage 16. The cover 46 is provided with a large number of ventilation holes for guiding exhaust gas therein. An element temperature sensor 48 is attached to the cover 46 so as to be positioned in the vicinity of the sensor element 30. The element temperature sensor 48 can detect the temperature of the sensor element 30 and provide the result to the ECU 24.

  As shown in FIG. 2, PM is contained in the exhaust gas discharged from the internal combustion engine 10. The PM in the exhaust gas can flow into the cover 46 through the vent hole. Part of the inflow PM adheres to the surface of the insulator 32. In the present embodiment, the PM sensor 22 can burn off the PM adhered in this manner by heating the insulator 32 with the heater 42.

  Under the situation where PM is not present on the surface of the insulator 32, the two electrodes 34 and 36 are electrically insulated. In this case, since no current flows between the electrode pairs, no output value is generated from the PM sensor 22 (current detection mechanism 40). On the other hand, when the gap between the electrode pairs is filled with PM as shown in FIG. 2, since PM has conductivity, a current can flow between the electrodes 34 and 36. In this case, the PM sensor 22 emits an output value corresponding to the voltage applied by the power supply mechanism 38 and the electrical resistance between the electrode pair.

[Calculation principle of PM concentration detection value]
FIG. 3 is a diagram for explaining the output characteristics of the PM sensor 22. The inventor of the present invention has found that the PM sensor 22 has an output characteristic as shown in FIG. In the present embodiment, a concentration detection value representing the PM concentration in the exhaust gas is calculated using the output characteristics.

I _TH shown in FIG. 3 is a determination threshold used in the present embodiment. The determination threshold value I _TH is a value experimentally determined in advance as a current generated at an early stage in which the output value of the PM sensor 22 rapidly increases in the process in which the space between the electrode pairs 34 and 36 of the PM sensor 22 is filled with PM. It is.

The waveform (i) shown in FIG. 3 shows that the output value of the PM sensor 22 rises after T ₁ second from the start of measurement, that is, after the measurement starts, until the output value exceeds the determination threshold value I _TH. T shows a situation where ₁ second is required. Further, the waveform (ii) and the waveform (iii) respectively show a state where it takes T ₂ seconds or T ₃ seconds for the output value to exceed the determination threshold value I _TH . Hereinafter, this time is referred to as “output rise time”.

  “Measurement start” shown in FIG. 3 means a point in time when the voltage applied to the electrode pairs 34 and 36 is started while the PM adhering to the insulator 32 is burned out during the operation of the internal combustion engine 10. After the start of measurement, the output value is maintained at substantially zero until a conductive path is formed between the electrodes 34 and 36 by the adhered PM. Then, an output value starts to be generated when a PM conductive path is formed between the electrodes 34 and 36, and then the output value changes to a large value due to an increase in PM adhesion amount and a decrease in the electrical resistance of the conductive path. To do.

At this time, the output rise time becomes shorter as the increase rate of the PM adhesion amount is faster. And the increase rate of PM adhesion amount becomes so high that PM concentration in exhaust gas is high. Therefore, as shown in FIG. 3, the output rise time (T ₁ , T ₂ , T ₃ ) becomes shorter as the PM concentration in the exhaust gas is higher (T ₁ <T ₂ <T ₃ ).

FIG. 4 is a map showing the relationship between the output rise time [sec] of the PM sensor 22 and the PM concentration [mg / m ³ ] in the exhaust gas. The map shown in FIG. 4 is obtained from the result of an experiment in which the output rise time is measured for each PM concentration by changing the PM concentration in the exhaust gas. With reference to this map, it is possible to roughly detect the PM concentration in the exhaust gas based on the output rise time of the PM sensor 22. In this specification, the value detected in this way is referred to as a “density detection value”.

[Correction of concentration detection value based on exhaust flow velocity]
FIG. 5 shows the relationship between the output rise time and the exhaust flow velocity. This relationship is obtained from the result of an experiment in which the exhaust flow rate is changed while the PM concentration is kept constant, and the output rise time is measured for each exhaust flow rate. The results shown in FIG. 5 show the following two phenomena.
(1) In the low exhaust flow velocity range where the exhaust flow velocity is less than the variable value of 30 [m / s], the lower the exhaust flow velocity, the longer the output rise time.
(2) In a high exhaust flow rate range where the exhaust flow rate exceeds the above-mentioned inflection value, the exhaust flow rate has little influence on the output rise time.

The phenomenon (1) is considered to be caused by the fact that PM in the exhaust gas does not easily enter the cover 46 as the exhaust flow rate decreases, and the PM does not easily adhere to the insulator 32 in the low exhaust region. On the other hand, the phenomenon of (2) is considered to be caused by the following two events canceling out in the high exhaust region.
(A) As the exhaust gas flow rate increases, the PM easily enters the cover 46, and the PM easily adheres to the insulator 32.
(B) As the exhaust gas flow rate increases, the PM adhering to the insulator 32 is easily blown away, and the amount of PM adhesion is less likely to increase.

  The results shown in FIG. 5 indicate that, in a physical sense, the output rise time becomes excessive with respect to the actual PM concentration in the low exhaust flow velocity region. More specifically, the results shown in FIG. 5 indicate that in order to obtain an accurate concentration detection value based on the output rise time, the output rise time must be corrected to be shorter as the exhaust flow rate is lower in the low exhaust flow rate region. It shows that there is.

  FIG. 6 is an example of an exhaust flow velocity correction coefficient map used by the ECU 24 to enable the above correction. According to the map shown in FIG. 6, the exhaust flow velocity correction coefficient is a constant value “1” in the high exhaust flow velocity region where the exhaust flow velocity exceeds 30 [m / s]. Further, in the low exhaust flow velocity region where the exhaust flow velocity is lower than 30 [m / s], an exhaust flow velocity correction coefficient smaller than 1 is set in proportion to the exhaust flow velocity. In the present embodiment, the ECU 24 corrects the output rise time by multiplying the correction coefficient set with reference to this map. As a result, the ECU 24 can calculate the output rise time corresponding to the actual PM concentration without being affected by the level of the exhaust gas flow rate.

[Correction of density detection value based on temperature difference]
FIG. 7 shows the relationship between the temperature difference obtained by subtracting the temperature of the sensor element 30 from the exhaust temperature and the output rise time. This relationship is obtained from the result of an experiment in which the temperature difference is changed while the PM concentration is kept constant, and the output rise time is measured for each temperature difference. This relationship indicates that when the PM concentration is constant, the output rise time is shortened as the exhaust temperature is higher than the temperature of the sensor element 30.

  It is known that the amount of PM adhering to the insulator 32 is affected by thermophoresis. That is, when the temperature of the exhaust gas is higher than the temperature of the insulator 32, the PM floating in the vicinity of the insulator 32 receives a force toward the insulator 32 due to the temperature difference. The force becomes larger as the temperature of the exhaust gas is higher than the temperature of the insulator 32.

  The result shown in FIG. 7 is that the higher the exhaust gas is, the higher the temperature of the insulator 32 (sensor element 30), the more easily PM adheres to the insulator 32 due to the influence of thermophoresis, and as a result, the output rise time It represents a shortening. This result physically indicates that the output rise time becomes too short with respect to the actual PM concentration as the exhaust gas is at a higher temperature than the sensor element 30. Therefore, in order to obtain an accurate density detection value based on the output rise time, it is necessary to correct the output rise time as the temperature difference between them increases.

  FIG. 8 is an example of a temperature difference correction coefficient map used by the ECU 24 to enable the above correction. According to the map shown in FIG. 8, it is possible to obtain a temperature difference correction coefficient that changes in proportion to the temperature difference obtained by subtracting the sensor element temperature from the exhaust gas temperature. The ECU 24 corrects the output rise time by multiplying the correction coefficient set with reference to this map. As a result, the ECU 24 can calculate the output rise time corresponding to the actual PM concentration without being affected by the temperature difference.

[PM concentration estimation model and its learning]
The system of the present embodiment performs a PM concentration detection process using a known PM concentration estimation model while performing a PM concentration detection process using the PM sensor 22. Hereinafter, this estimation process will be described.

  The ECU 24 is equipped with a PM concentration estimation model similar to that disclosed in Japanese Patent Application Laid-Open No. 2007-46477. According to this model, the ECU 24 can estimate the PM concentration in the exhaust gas discharged from the internal combustion engine 10 based on the engine speed NE, the accelerator opening AA, the coolant temperature THW, and the like.

  As described above, the ECU 24 can calculate a concentration detection value corresponding to the PM concentration in the exhaust gas based on the output of the PM sensor 22. Since the PM sensor 22 is located downstream of the DPF 18, the detected concentration value corresponds to the concentration of PM that has passed through the DPF 18. Here, the ratio of PM that passes through the DPF 18 (hereinafter referred to as “DPF slip-through rate”) is known or can be detected. If the DPF slip-through rate is used, the concentration detection value downstream of the DPF 18 can be converted into the concentration detection value upstream of the DPF 18.

  In the present embodiment, the ECU 24 estimates the PM concentration upstream of the DPF 18 using the above-described model (hereinafter, this value is referred to as “concentration estimated value”), and the PM concentration upstream of the DPF 18 using the above method. Calculate the concentration detection value to represent. While the DPF 18 is normal, the density estimation value and the density detection value are substantially the same value. On the other hand, when an abnormality occurs in the DPF 18 and the amount of PM passing through downstream increases, the detected concentration value suddenly becomes a large value with respect to the estimated concentration value. The ECU 24 can detect an abnormality of the DPF 18 using this characteristic (first advantage by using a model and actual measurement in combination).

  The model for estimating the PM concentration is adapted under the assumption that various parameters are at the median tolerance. For this reason, in order to obtain an estimated value corresponding to a condition change such as an individual difference of the internal combustion engine 10, a change with time, or a change in operating state, the model is appropriately set so that the estimated concentration value matches the actual PM concentration. It is beneficial to learn.

  The output value of the PM sensor 22 reflects the influence of a change in conditions related to the internal combustion engine 10. For this reason, the concentration detection value calculated based on the output value can be regarded as the actual PM concentration. Therefore, in the present embodiment, the ECU 24 learns the model so that the estimated concentration value based on the model matches the detected concentration value based on the output value of the PM sensor 22.

  FIG. 9 and FIG. 10 are examples of timing charts for explaining the effect of using the density detection value obtained by the method of this embodiment as the basis of the above-described model learning. Specifically, these figures show effects obtained when model learning is performed based on the detected concentration value, and the detected concentration value always accurately represents the actual PM concentration regardless of the influence of the exhaust flow velocity. Is explained.

The waveform indicated by (i) in FIG. 9 (a) shows the actual change in the exhaust gas flow velocity. Further, the broken line shown in the figure indicates the value of the exhaust velocity used for model adaptation (hereinafter referred to as “model adaptation value”). 9 (a) is an exhaust flow rate (i) is, until time t ₁ is was equivalent to model adaptive value indicates the then gradually how decreased.

The waveform indicated by (ii) in FIG. 9B indicates the actual PM concentration upstream of the DPF 18, and the waveform indicated by (iii) is obtained when correction based on the exhaust flow velocity is not performed. The concentration detection value is shown. Further, a waveform indicated by (iv) in the figure shows a concentration estimation value reflecting a learning result based on the concentration detection value (iii). In FIG. 9 (b), the time t ₂ is the time when the model learning is initiated density estimate.

If correction is performed based on the exhaust flow rate, the time t ₁ after, as the exhaust flow rate (i) is decreased, the concentration detection value (iii), as a value smaller than the actual PM (ii) Calculated. Time t ₂ after, by the model learning is initiated, concentration estimated value (iv) is gradually converged to a value equal inevitably its concentration detected value (iii). In this case, similarly to the concentration detection value (iii), the concentration estimation value (iv) is also a value deviating from the actual PM concentration (ii).

FIG. 10 (a) shows the exhaust gas flow velocity model adaptation value (broken line) and the actual exhaust gas flow velocity (i), as in FIG. 9 (a). Again, the exhaust flow rate (i) represents a time t ₁ after reduction. The waveform indicated by (v) in FIG. 10 (b) indicates the concentration detection value when correction based on the exhaust flow rate is performed. In addition, the waveform indicated by (vi) in the figure indicates a concentration estimation value that reflects the result of model learning based on the concentration detection value (v).

If performed correction based on the exhaust flow rate, the time t ₁ after, also the exhaust flow rate (i) is reduced, the concentration detection value (v) is calculated to a value that is consistent with actual PM (ii). In this case, if the time t ₂ subsequent model learning starts, concentration estimated value (vi) is inevitably converges to a value that matches the actual PM concentration (ii).

  As described above, the system of the present embodiment corrects the detected concentration value based on the exhaust flow rate. For this reason, according to this system, as shown in FIG. 10, model learning can be correctly performed even in a situation where the exhaust flow velocity (i) deviates from the model fit value, and it matches with the actual PM concentration with high accuracy. A concentration estimate (vi) can be calculated.

  In addition, in the present embodiment, the concentration detection value is corrected based on the difference between the exhaust temperature and the sensor element temperature. Further, in the present embodiment, the concentration detection value is detected as a value that accurately represents the actual PM concentration based on the output rise time (see FIG. 3). For this reason, according to the system of the present embodiment, model learning can be correctly performed under any circumstances, and a concentration estimation value that accurately matches the actual PM concentration can be calculated.

[Flow of control related to PM concentration]
FIG. 11 shows a block diagram of control related to PM concentration executed by the ECU 24. The control of the block diagram shown in FIG. 11 is specifically realized by the ECU 24 executing a routine according to FIG.

  In the present embodiment, the ECU 24 repeatedly executes the routine shown in FIG. 12 every predetermined time after the internal combustion engine 10 is started. Here, it is first determined whether or not the PM sensor 22 is activated as a premise for detecting the PM concentration (step 100). Specifically, it is determined whether the temperature of the sensor element 30 has reached a predetermined activation temperature and whether the incineration removal of PM adhering to the sensor element 30 has been completed. The former determination is made based on the output of the element temperature sensor 48. The latter determination is made based on the energization time to the heater 42 for PM incineration.

If the activation of the PM sensor 22 is recognized, then the output rise time of the PM sensor 22 is acquired (step 102). Here, first, voltage application to the electrode pairs 34 and 36 is started. The voltage application is continued until the current value detected by the current detection mechanism 40 exceeds the determination threshold value _ITH . ECU24 after the start of voltage application to acquire the time the current value required until exceeding the judgment threshold I _TH as the output rise time.

  Next, correction based on the exhaust flow velocity and correction based on the temperature difference are executed (step 104). As for the former correction, first, the exhaust flow velocity around the PM sensor 22 is detected based on the output of the air flow meter 14. Next, the exhaust flow velocity correction coefficient is specified by fitting the exhaust flow velocity to the map shown in FIG. Regarding the correction based on the temperature difference, first, the temperature difference (= exhaust temperature−sensor element temperature) is calculated based on the output of the exhaust temperature sensor 20 and the output of the element temperature sensor 48. Next, the temperature difference is applied to the map shown in FIG. 8 to specify the temperature difference correction coefficient. Here, the correction based on the exhaust flow velocity and the correction based on the temperature difference are executed by multiplying these two correction coefficients by the output rise time acquired in step 102.

  When the above processing is completed, a PM concentration detection value is acquired according to the map shown in FIG. 4 (step 106). Specifically, by applying the output rise time corrected in step 104 to the PM concentration map shown in FIG. 4, the concentration detection value corresponding to the rise time is specified.

  As described above, in the present embodiment, the PM concentration detection value is specified based on the output rise time of the PM sensor 22. Since the output rise time is always measured from a constant state, that is, from the state where PM is not attached to the sensor element 30, the output rise time exhibits excellent reproducibility with respect to the actual PM concentration. Furthermore, the output rise time reflecting the influence of the temperature difference between the exhaust gas and the sensor element and the influence of the exhaust flow rate is used for specifying the concentration detection value. For this reason, according to the present embodiment, it is possible to obtain a concentration detection value that accurately matches the actual PM concentration regardless of the exhaust flow velocity or the temperature difference.

The concentration detection value acquired in step 106 is a value representing the PM concentration downstream of the DPF 18. Here, a conversion process is performed in which the detected concentration value is converted into a detected PM concentration value upstream of the DPF 18 by the following equation (1) (step 108).
(DPF upstream PM concentration detection value)
= (DPF downstream PM concentration detection value) / (DPF slip-through rate) (1)
However, the DPF slip-through rate is known or can be detected by the ECU 24 as described above.

  In the routine shown in FIG. 12, next, an estimated PM concentration value upstream of the DPF 18 is acquired based on the above-described model (step 110).

Next, the ratio between the estimated value and the detected value of the PM concentration is calculated as a correction coefficient for performing final correction on the estimated value by the following equation (2) (step 112).
(Correction coefficient) = (DPF upstream PM concentration detection value) / (DPF upstream PM concentration estimated value)
... (2)

  The correction coefficient calculated in this way is used for correcting the density estimation value acquired in step 110 (step 114) and recorded in a learning map for model learning (step 116). The correction in step 114 and the correction coefficient recorded in the learning map are used so that the final density estimation value approaches the density detection value upstream of the DPF 18. For this reason, according to the above processing, it is possible to obtain a concentration estimated value that accurately matches the PM concentration detection value without being affected by a change with time of the internal combustion engine 10 or the like.

[Modification]
By the way, in the first embodiment described above, the PM concentration estimation using the model and the PM concentration detection using the PM sensor 22 are made to coexist in a single system. It is not limited to a kind of system. That is, a PM sensor may be added upstream of the DPF 18 and the PM concentration at that position may be actually measured by the sensor. Furthermore, when there is no need to know the PM concentration above and below the DPF 18, the PM sensor may be disposed only on one of them.

In the first embodiment described above, the end of the output rise time is the time when the output value of the PM sensor 22 reaches the determination threshold _ITH , but the end is not limited to this. That is, the end of the output rise time may be determined as a point in time when the increase rate of the output value of the PM sensor 22 exceeds a preset determination threshold.

  In the first embodiment described above, the output rise time of the PM sensor 22 is corrected based on the exhaust flow velocity [m / s]. However, the exhaust flow velocity may be replaced with the exhaust flow rate. In this regard, in this specification, the exhaust flow rate is a concept including the exhaust flow rate.

  In the first embodiment described above, the exhaust flow velocity correction coefficient is changed in proportion to the change in the exhaust flow velocity in the low exhaust flow velocity region as shown in FIG. 6, and is constant in the high exhaust flow velocity region. To keep on. However, the tendency of the exhaust flow velocity correction coefficient is not limited to this. For example, the exhaust flow velocity correction coefficient may be changed so that the gradient of the change with respect to the change in the exhaust flow velocity becomes gradually gentler as the exhaust flow velocity increases.

  Further, in the first embodiment described above, the PM sensor 22 is reset by burning off the adhered PM by the heater 42, but the resetting method is not limited to this. That is, any known method may be used as the resetting method as long as the PM adhering between the electrode pairs 34 and 36 can be eliminated.

  In the first embodiment described above, the heater 42 of the PM sensor 22 corresponds to the “particulate removal mechanism” in the first invention. In the first embodiment described above, the “exhaust particulate concentration estimating means” according to the eighth aspect of the present invention is realized by the ECU 24 executing the processing of step 110. Further, the “parameter learning means” according to the eighth aspect of the present invention is implemented when the ECU 24 executes the processing of step 116. Furthermore, in the above-described first embodiment, the ratio of the concentration detection value obtained under the execution of the correction based on the exhaust flow velocity and the concentration detection value obtained when the correction is not performed is the above-mentioned claim 5. Or it corresponds to the “correction rate” in 6.

10 Internal combustion engine 18 DPF
20 Exhaust temperature sensor 22 PM sensor 24 ECU
30 Sensor element 32 Insulator 34, 36 Electrode (electrode pair)
38 Power supply mechanism 42 Heater 48 Element temperature sensor

Claims (8)

An exhaust particulate sensor comprising a sensor element that emits an output value according to the amount of exhaust particulate attached, and a particulate removal mechanism that eliminates exhaust particulate attached to the sensor element;
An electronic device that calculates a concentration detection value representing a concentration of exhaust particulates based on an output rise time required until the output value rises after reset due to disappearance of the exhaust particulates;
The concentration detection value is calculated as a larger value as the output rise time is shorter. The sensor element includes an insulator exposed in an exhaust passage of an internal combustion engine, a pair of electrodes disposed in the insulator, and a power supply mechanism that applies a voltage between the pair of electrodes,
The exhaust particulate concentration detection apparatus according to claim 1, wherein the particulate removal mechanism includes a heater that heats the insulator to burn away exhaust particulates.   3. The exhaust particulate concentration detection apparatus according to claim 1, wherein the output rise time is a time required for the output value to exceed a determination threshold after the reset.   3. The exhaust particulate concentration detection apparatus according to claim 1, wherein the output rise time is a time required for the rate of increase of the output value to exceed a determination threshold after the reset. The electronic device corrects the concentration detection value based on the exhaust flow velocity of the internal combustion engine,
5. The correction according to claim 1, wherein the correction is performed at a large correction rate so that the concentration detection value becomes larger as the exhaust flow rate is lower in at least a part of the exhaust flow rate region. The exhaust particulate concentration detecting device according to the item.   The inclination of the change in the correction factor with respect to the change in the exhaust flow rate is more gradual in the high exhaust flow rate region where the exhaust flow rate is higher than the variable value than in the low exhaust flow rate region where the exhaust flow rate is lower than the variable value. The exhaust particulate concentration detector according to claim 5. The electronic device corrects the concentration detection value based on a temperature difference between the exhaust gas and the sensor element,
The exhaust gas according to any one of claims 1 to 6, wherein the correction is performed such that the concentration detection value becomes smaller as the exhaust gas temperature is higher than that of the sensor element. Particulate concentration detector. Exhaust particulate concentration estimating means for calculating an estimated value of exhaust particulate discharged from the internal combustion engine using an exhaust particulate concentration estimation model;
Parameter learning means for learning parameters of the exhaust particulate concentration estimation model so that the concentration estimation value approaches the concentration detection value;
The exhaust particulate concentration detector according to any one of claims 1 to 7, further comprising:

Images