Error estimation device, post-processing system, error estimation method, and computer program
The error estimation device improves the accuracy of observation error estimation in exhaust gas components by using state estimation theory and constraint conditions, enhancing the precision of abnormality detection and system efficiency.
Patent Information
- Application Number
- JP2023074464
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2023-04-28
- Publication Date
- 2024-11-08
AI Technical Summary
Existing systems for estimating observation errors in exhaust gas components, such as NOx sensors and air-fuel ratio sensors, are inaccurate due to hardware variations and deterioration over time, leading to ineffective control and diagnosis in exhaust treatment systems.
An error estimation device that utilizes state estimation theory to estimate observation errors in acquisition units by comparing estimated state quantities with actual sensor readings, incorporating constraint conditions to improve accuracy and correct observed values.
Enhances the accuracy of observation error estimation in exhaust gas components, enabling more frequent and precise abnormality detection and improving the efficiency of exhaust treatment systems, including fuel efficiency.
Smart Images

Figure 2024158869000001_ABST
Abstract
Description
[Technical field]
[0001] The present invention relates to a technique for estimating an observation error in an acquisition section provided in a post-processing system. [Background technology]
[0002] There are known technologies for purifying harmful substances contained in exhaust gas discharged from internal combustion engines, etc. For example, Patent Document 1 discloses an aftertreatment system using an SCR catalyst (Selective Catalytic Reduction catalyst) that uses NH3 (ammonia) as an additive and purifies NOx (nitrogen oxides) as a harmful substance. For example, Patent Documents 2 and 3 disclose an aftertreatment system using an NSR catalyst (NOx Storage Reduction catalyst, also called NOx catalyst) that uses CO (carbon monoxide), H2 (hydrogen), and HC (hydrocarbon) generated by rich spike processing as additives and purifies NOx as a harmful substance.
[0003] Non-Patent Document 1 discloses the use of pseudo observations in the problem of tracking flying objects. In the problem of tracking flying objects, the target speed is high, so that a normal Kalman filter cannot provide sufficient estimation accuracy, and attempts have been made to generate auxiliary observation data to improve the estimation accuracy. For example, physical constraints such as the target moving in a uniform linear motion are considered separately from the dynamics of the target, and the fact that the constraints are satisfied is considered as if it were observed, and these are pseudo-observed and added to the original observation equation to form an extended observation equation. Observations generated in this auxiliary manner are called "pseudo observations." Non-Patent Document 2 also discloses the introduction of pseudo observations to improve the estimation accuracy of the position, speed, and acceleration of a target in the problem of tracking a ship using a radar. Meanwhile, Non-Patent Document 3 discloses a technology for estimating the time delay of an oxygen sensor in an engine system. [Prior art documents] [Patent documents]
[0004] [Patent Document 1] JP 2008-190383 A [Patent Document 2] JP 2004-308455 A [Patent Document 3] JP 2009-180086 A [Non-patent literature]
[0005] [Non-Patent Document 1] M. Tahk, JL Speyer, Target tracking problems subject to kinematic constraints, IEEE Transactions on Automatic Control, 3 March 1990, vol. 35, No. 3, pp. 324-326. [Non-Patent Document 2] Akira Osumi, Seiichi Yasuki, Junji Hirata, Yoshihisa Ijiri, Takehisa Nagayama, Hiroto Sasaki, Tracking of Moving Objects with Irregular Course and Speed Changes Using an Extended Kalman Filter, Transactions of the Institute of Systems, Control and Information Engineers, Vol. 14, No. 10, 2001, pp. 490-498. [Non-Patent Document 3] Michael M. Moser, Christopher H. Onder, Lino Guzzella, Recursive parameter estimation of exhaust gas oxygen sensors with input-dependent time delay and linear parameters, Control Engineering Practice, 28 May 2015, vol. 41, pp. 149-163. Summary of the Invention [Problem to be solved by the invention]
[0006] Incidentally, in the above-mentioned aftertreatment system, an acquisition unit that detects the physical quantities of a predetermined component in the exhaust gas is provided in the main flow passage through which the exhaust gas flows, and the aftertreatment system is controlled using the observed values of the sensor. For example, in the technology described in Patent Document 2, NOx sensors are provided upstream and downstream of the NSR catalyst, and the deterioration state of the NSR catalyst is diagnosed from changes in the observed values of these NOx sensors. For example, in the technology described in Patent Document 3, an air-fuel ratio sensor that detects the O2 (oxygen) concentration in the exhaust gas is provided, and the execution time of the reduction process in the NSR catalyst is determined using the observed value of this air-fuel ratio sensor. However, these acquisition units may not be able to detect accurate physical quantities due to hardware variations (for example, manufacturing errors in the detection element, etc.), deterioration over time, etc.
[0007] In this regard, the technology described in Patent Document 1 realizes a fault diagnosis of the NOx sensor by comparing a reference pattern of the NOx flow rate discharged from the internal combustion engine with a tracking pattern of the NOx concentration based on the observation value of the NOx sensor. However, the technology described in Patent Document 1 has a problem that, depending on the operating state of the internal combustion engine, the NOx flow rate discharged from the internal combustion engine may not match the reference pattern for a long time, and in such a case, the fault diagnosis of the NOx sensor cannot be performed. In addition, Non-Patent Documents 1 and 2 do not discuss at all the estimation of the observation error in the NOx sensor or the air-fuel ratio sensor, or the constraint conditions in the aftertreatment system. Furthermore, the technologies described in Patent Documents 2, 3, and Non-Patent Document 3 do not consider at all the estimation of the observation error in the NOx sensor or the air-fuel ratio sensor.
[0008] In addition, such a problem is not limited to the case where the NOx sensor or the air-fuel ratio sensor described in Patent Documents 1 to 3 is used as the acquisition unit, but is common to all "acquisition units" that detect or acquire by calculation a physical quantity of a predetermined component (any component) in the exhaust gas. In other words, the "acquisition unit" includes both a sensor that directly acquires a physical quantity and a functional unit that calculates the physical quantity. In addition, the physical quantity is not limited to the NOx concentration, and the concentration, flow rate, temperature, pressure, etc. of a predetermined component other than NOx may be used. Furthermore, such a problem is not limited to SCR catalysts that use NH3 as an additive, and NSR catalysts that use CO, H2, and HC generated by rich spike processing as additives, but is also common to three-way catalysts that use O2 generated by oxidation-reduction reaction as an additive.
[0009] The present invention has been made to solve at least part of the above-mentioned problems, and aims to improve the technology for estimating observation errors in an acquisition unit that acquires physical quantities for specified components in exhaust. [Means for solving the problem]
[0010] The present invention has been made to solve at least part of the above-mentioned problems, and can be realized in the following forms.
[0011] (1) According to one aspect of the present invention, there is provided an error estimation device. The error estimation device includes an exhaust purification device including a main flow path through which exhaust gas flows and a catalyst for purifying harmful substances in the exhaust gas, a sensor for acquiring a first physical quantity for a predetermined component in the exhaust gas, an acquisition unit for acquiring a second physical quantity for the predetermined component in the exhaust gas, a state estimation unit for estimating a state quantity that represents an internal state of the exhaust purification device and affects the first physical quantity acquired by the sensor, and an error estimation unit for estimating an observation error in the acquisition unit using a state estimation theory from the state quantity estimated by the state estimation unit and the first physical quantity acquired by the sensor. The error estimation unit estimates the observation error using the state quantity estimated by the state estimation unit, using a constraint condition that the estimated value is equal to the calculated value calculated from the first physical quantity acquired by the sensor, the second physical quantity acquired by the acquisition unit, and the observation error estimated by the error estimation unit.
[0012] According to this configuration, the error estimation unit estimates the observation error in the acquisition unit using the state estimation theory from the state quantity estimated by the state estimation unit and the first physical quantity acquired by the sensor. Therefore, the observation error in the acquisition unit can be estimated without being influenced by the operating state of the internal combustion engine or the flow pattern of the predetermined component, so that the frequency of abnormality diagnosis of the acquisition unit using the observation error can be improved. In addition, the error estimation unit estimates the observation error using a constraint condition (A=B) that an estimated value A estimated using the state estimation unit is equal to a calculated value B calculated from the results of acquisition and estimation by the sensor, the acquisition unit, and the error estimation unit. Here, generally, the observation error in the acquisition unit is reflected in the observation data only through the state variable. However, according to this configuration, the constraint condition includes the observation error estimated by the error estimation unit (in the past), and thus the observation error estimated by the error estimation unit (in the past) is added as information to be considered when the error estimation unit estimates the observation error. Therefore, compared to a case where the error estimating unit is configured with the observation error as an unknown parameter, the estimation accuracy of the observation error in the acquiring unit can be improved without increasing costs. As a result, the error estimating device of this configuration can improve the technology for estimating the observation error in the acquiring unit that acquires the physical quantity of a predetermined component in the exhaust.
[0013] (2) In the error estimation device of the above embodiment, the catalyst may be an NSR catalyst that purifies NOx as a harmful substance, the sensor may be an outlet NOx sensor that acquires the NOx concentration in the exhaust downstream of the catalyst as the first physical quantity, the acquisition unit may be an inlet NOx sensor that acquires the NOx concentration in the exhaust upstream of the catalyst as the second physical quantity, and the state estimation unit may estimate the amount of NOx occlusion in the catalyst as the state quantity. According to this configuration, the error estimation unit can estimate the measurement error of the inlet NOx sensor that acquires the NOx concentration in the exhaust gas upstream of the catalyst.
[0014] (3) In the error estimation device of the above embodiment, the constraint condition y p is defined by the following formula: TIFF2024158869000002.tif21170In the above formula, R NOx,st is the estimated NOx storage rate, and NO obs x,in is the second physical quantity acquired by the acquisition unit, and NO x,out is the first physical quantity acquired by the sensor, and θ in may be the measurement error estimated by the error estimating unit. According to this configuration, the NOx absorption rate estimated value A(R NOx,st ) is equal to the calculated value B calculated from the results of acquisition and estimation by the sensor, acquisition unit, and error estimator. p =AB=0 can be defined.
[0015] (4) In the error estimation device of the above aspect, the error estimator may use an algorithm for obtaining an estimate that minimizes a mean square error of a state quantity when a mean square error is used as an evaluation function as the state estimation theory. According to this configuration, the error estimation unit can estimate a state quantity that is difficult to measure (i.e., the observation error in the acquisition unit) by using an algorithm that finds an estimate that minimizes the mean square error of the state quantity when the mean square error is used as the evaluation function as the state estimation theory.
[0016] (5) In the error estimation device according to the above aspect, the acquisition unit may correct the acquired second physical quantity by using the observation error estimated by the error estimator. According to this configuration, the acquisition unit corrects the acquired second physical quantity using the measurement error estimated by the error estimation unit, and therefore the accuracy of the second physical quantity can be improved. As a result, the accuracy of various controls (e.g., estimation control of state quantity in the state estimation unit and additive supply control for purifying harmful substances in the catalyst) performed using the corrected second physical quantity can be improved.
[0017] (6) The error estimation device according to the above aspect may further include a control unit that detects an abnormality in the acquisition unit by using the observation error estimated by the error estimating unit. According to this configuration, the control unit detects an abnormality in the acquisition unit by using the observation error estimated by the error estimation unit. Therefore, abnormality detection can be performed without requiring a special operation mode for the purpose of abnormality detection (fault diagnosis), and the frequency of abnormality detection can be improved.
[0018] (7) According to one aspect of the present invention, there is provided an aftertreatment system comprising: the error estimation device of the above aspect; a catalyst provided in a main flow path through which exhaust gas flows and purifying harmful substances in the exhaust gas; a supply unit that supplies an additive for purifying the harmful substances to the catalyst; and a supply control unit that controls the supply of the additive in the supply unit. According to this configuration, it is possible to provide an aftertreatment system capable of supplying an additive taking into account measurement errors in the acquisition unit, thereby improving purification efficiency in the aftertreatment system and contributing to improved fuel economy.
[0019] The present invention can be realized in various forms, for example, in the form of an error estimation device, a sensor state estimation device, a control device that controls a post-processing system using an estimated observation error, a post-processing system including these devices, a control method for these devices and systems, a computer program executed in these devices and systems, a server device for distributing the computer program, a non-transitory storage medium on which the computer program is stored, etc. [Brief description of the drawings]
[0020] [Figure 1] 1 is a block diagram of a post-processing system according to an embodiment of the present invention; [Diagram 2] FIG. 2 is a diagram illustrating each function in the error estimation device. [Diagram 3] FIG. 13 is a diagram illustrating a fourth model. [Figure 4]4 is a flowchart showing a procedure of an error estimation process in the error estimation device. [Diagram 5] This shows the simulation results when the correct value for the observation error of the inlet NOx concentration acquisition section was set to +20%. [Figure 6] This shows the simulation results when the correct value for the observation error of the inlet NOx concentration acquisition unit was set to -20%. [Figure 7] 4 is a flowchart showing a procedure of an abnormality detection process in the error estimation device. [Figure 8] FIG. 11 is a block diagram of a post-processing system according to a second embodiment. [Figure 9] FIG. 11 is a block diagram of a post-processing system according to a third embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0021] First Embodiment 1 is a block diagram of an aftertreatment system 1 according to one embodiment of the present invention. The aftertreatment system 1 of this embodiment includes an error estimation device 10 and an exhaust purification device 20. The error estimation device 10 of this embodiment can estimate an observation error in an acquisition unit that acquires a physical quantity of a predetermined component in the exhaust gas, by the process described below.
[0022] Here, the "acquisition unit" includes both a sensor that directly acquires the physical quantity of a predetermined component in the exhaust gas, and a functional unit that determines the physical quantity by calculation. In addition to NOx (nitrogen oxides), CO (carbon monoxide), H2 (hydrogen), HC (hydrocarbon), O2 (oxygen), NH3 (ammonia), etc. can be used as the predetermined component. In the following examples, NOx is exemplified as the predetermined component in the exhaust gas. In addition, the acquisition unit is exemplified by a sensor (inlet NOx sensor) that is provided upstream of the catalyst and directly acquires the NOx concentration in the exhaust gas.
[0023] The exhaust purification device 20 is a device that purifies NOx, a harmful substance, in the exhaust gas of the internal combustion engine 92, and is equipped with an NSR catalyst 70 (NOx Storage Reduction catalyst, also called a NOx catalyst) as a catalyst, and uses CO, H2, and HC generated by rich spike processing as additives.
[0024] The internal combustion engine 92 is, for example, a diesel engine or a gasoline engine with a lean burn operation method. The combustion state control unit 91 controls the injection of air and fuel into the internal combustion engine 92 to control the air-fuel ratio in the internal combustion engine 92 to each of the lean, stoichiometric, and rich states. The combustion state control unit 91 is implemented, for example, by an ECU (Electronic Control Unit). In the following description, the side of the exhaust purification device 20 that is closer to the internal combustion engine 92 is called the "upstream side," and the side that is farther from the internal combustion engine 92 is called the "downstream side." In the case of FIG. 1, the left side corresponds to the upstream side, and the right side corresponds to the downstream side. The exhaust purification device 20 may be connected to another device (for example, a factory, various types of combustion devices, etc.) that emits exhaust containing harmful substances, instead of the internal combustion engine 92.
[0025] The exhaust purification device 20 includes an exhaust pipe 30 connected to an internal combustion engine 92, an NSR catalyst 70 provided on the exhaust pipe 30, and a rich spike control unit 93. The exhaust pipe 30 forms a main flow path through which exhaust from the internal combustion engine 92 flows. The exhaust from the internal combustion engine 92 passes through the main flow path in the exhaust pipe 30, passes through the NSR catalyst 70, and is released into the outside air. The NSR catalyst 70 purifies the NOx in the exhaust by storing and accumulating the NOx in the exhaust inside the NSR catalyst 70. The NSR catalyst 70 corresponds to a "catalyst."
[0026] The rich spike control unit 93 performs rich spike control (also called "reduction control") to purify NOx stored in the NSR catalyst 70. In the rich spike control, the rich spike control unit 93 makes the air-fuel ratio in the internal combustion engine 92 rich for a short time, thereby discharging CO, H2, and other unburned gases such as HC (hereinafter also called "CO, H2, HC, etc.") from the internal combustion engine 92. Then, the discharged CO, H2, HC, etc. reduce the NOx stored in the NSR catalyst 70 to nitrogen gas (N2). That is, CO, H2, HC, etc. function as additives (more specifically, reducing agents) used for the reduction reaction of NOx stored in the NSR catalyst 70. In this embodiment, at least one or more elements (which may be one or more) of CO, H2, and HC correspond to the "additive". The rich spike control unit 93 corresponds to the "supply control unit", and the internal combustion engine 92 corresponds to the "supply unit". The rich spike control unit 93, like the combustion state control unit 91, can be implemented by an ECU.
[0027] The error estimation device 10 includes a CPU 11, a memory unit 12, a display unit 13, a flow rate acquisition unit 52, an exhaust temperature acquisition unit 53, an inlet NOx concentration acquisition unit 54, a temperature acquisition unit 78, and an outlet NOx sensor 99. The CPU 11 and the memory unit 12 are implemented by, for example, an ECU. The display unit 13 can be any display device, such as an LED (Light Emitting Diode), a liquid crystal panel, a touch panel, or the like.
[0028] The CPU 11 loads a computer program stored in the ROM into the RAM and executes it to control each unit of the error estimating device 10. In addition, the CPU 11 also functions as a state estimation unit 110, an error estimating unit 111, and a control unit 112.
[0029] 2 is a diagram for explaining each function in the error estimation device 10. A state estimation unit 110 estimates the amount of NOx occlusion in the NSR catalyst 70 by using a catalyst state estimation model 120 stored in the storage unit 12. The amount of NOx occlusion estimated by the state estimation unit 110 is a state quantity that represents the internal state of the exhaust purification device 20. The amount of NOx occlusion affects the observation value of the outlet NOx sensor 99 (hereinafter also referred to as the "first physical quantity").
[0030] The error estimation unit 111 estimates the observation error in the inlet NOx concentration acquisition unit 54 by an error estimation process described later in FIG. 4. Here, the "observation error" is a value that indicates how far the value obtained by the inlet NOx concentration acquisition unit 54 is from the median (true value), in other words, a value that indicates the difference from the median. When the inlet NOx concentration acquisition unit 54 is configured as a sensor that directly acquires a physical quantity of a predetermined component in the exhaust, the observation error in the inlet NOx concentration acquisition unit 54 occurs due to a manufacturing error (hardware variation) of the built-in detection element or aging deterioration of the detection element. When the inlet NOx concentration acquisition unit 54 is configured as a functional unit that calculates a physical quantity, the observation error occurs due to a modeling error (error existing between a model and an actual object) or the like. Hereinafter, the observation error in the inlet NOx concentration acquisition unit 54 is also simply called the "observation error in the acquisition unit" or the "observation error".
[0031] As shown in FIG. 2, the error estimation unit 111 uses the state quantity estimated by the state estimation unit 110 and the first physical quantity acquired by the outlet NOx sensor 99 to estimate the observation error in the inlet NOx concentration acquisition unit 54 (the observation error in the acquisition unit) using the state estimation theory. At this time, the error estimation unit 111 estimates the observation error using the constraint condition 121 stored in the storage unit 12. Details will be described later. The error estimation unit 111 of this embodiment illustrates a case where an unscented Kalman filter (UKF) is used as a state estimator that realizes the state estimation theory. The unscented Kalman filter is an algorithm that obtains an estimated value that minimizes the mean square error of the state quantity when the mean square error is used as the evaluation function. The error estimation unit 111 may use other similar algorithms, for example, an extended Kalman filter (EKF), a sigma point Kalman filter, a particle filter, etc.
[0032] The control unit 112 performs abnormality detection for the inlet NOx concentration acquisition unit 54, using the measurement error estimated by the error estimation unit 111. Details will be described with reference to FIG.
[0033] The flow rate acquiring unit 52 acquires the flow rate of exhaust from the internal combustion engine 92. The flow rate acquiring unit 52 may be realized, for example, by acquiring a measurement signal measured by a Pitot tube flow meter provided in the exhaust pipe 30. The flow rate acquiring unit 52 may also estimate the flow rate of exhaust from an intake air amount signal to the internal combustion engine 92 or a fuel injection amount signal. The exhaust temperature acquiring unit 53 is a sensor that measures the temperature of exhaust from the internal combustion engine 92.
[0034] The inlet NOx concentration acquisition unit 54 is a sensor (inlet NOx sensor) that measures the NOx concentration in the exhaust gas upstream of the NSR catalyst 70, in other words, the NOx concentration in the exhaust gas flowing into the NSR catalyst 70. The inlet NOx concentration acquisition unit 54 corresponds to an "acquisition unit", and the NOx concentration measured by the inlet NOx concentration acquisition unit 54 corresponds to a "second physical quantity". Note that the inlet NOx concentration acquisition unit 54 may be configured as a functional unit that determines the NOx concentration in the exhaust gas by calculation from the combustion state (lean, stoichiometric, rich) of the internal combustion engine 92, instead of measurement by a sensor.
[0035] The front end temperature acquisition unit 76 is a sensor that measures the temperature near the inlet (front end) of the NSR catalyst 70. The temperature acquisition unit 78 is a sensor that measures the bed temperature of the NSR catalyst 70.
[0036] The outlet NOx sensor 99 is a sensor (outlet NOx sensor) that measures the NOx concentration in the exhaust downstream of the NSR catalyst 70, in other words, the NOx concentration in the exhaust discharged from the NSR catalyst 70. The outlet NOx sensor 99 corresponds to a "sensor", and the NOx concentration measured by the outlet NOx sensor 99 corresponds to a "first physical quantity". Note that the outlet NOx sensor 99 may be configured as a functional unit that determines the NOx concentration in the exhaust by calculation from the combustion state (lean, stoichiometric, rich) of the internal combustion engine 92, instead of measuring by a sensor.
[0037] The storage unit 12 is composed of a flash memory, a memory card, a hard disk, etc. The storage unit 12 stores in advance a catalyst state estimation model 120 and a constraint condition 121. The catalyst state estimation model 120 includes a first model, a second model, and a fourth model.
[0038] The first model is a model for estimating the "NOx purification rate" of the NSR catalyst 70, and is configured by a machine learning model, specifically, an NN (Neural Network). For example, at least one of the parameters shown in the following items a1 and a2 can be used as input variables of the first model. The output variable of the first model is an estimated value of the NOx purification rate of the NSR catalyst 70 under the conditions represented by the input variables. (a1) Catalyst information: at least one of the following: the temperature of the front end of the NSR catalyst 70; the temperature of the NSR catalyst 70; the amount of NOx stored in the NSR catalyst 70 one time ago; and the ratio of the amount of NOx stored (actual amount of storage) in the NSR catalyst 70 to the saturated amount of NOx stored. (a2) Exhaust information: at least one of the temperature of the exhaust gas from the internal combustion engine 92, the flow rate of the exhaust gas, and the amount of NOx contained in the exhaust gas.
[0039] The second model is a model for estimating the "amount of NOx reduced" in the NSR catalyst 70, and is configured by a machine learning model, specifically, a neural network. For example, at least one of the parameters shown in the following items b1 to b3 can be used as the input variables of the second model. The output variable of the second model is an estimate of the amount of NOx reduced by rich spike control under the conditions represented by the input variables. (b1) Catalyst information: at least one of the temperature at the front end of the NSR catalyst 70, the temperature of the NSR catalyst 70, and the amount of NOx stored in the NSR catalyst 70 one time ago (actual amount of NOx stored). (b2) Exhaust information: at least one of the temperature of the exhaust from the internal combustion engine 92 and the flow rate of the exhaust (b3) Information on additives: at least one of the amount of additives (CO, H2, HC, etc.) flowing into the NSR catalyst 70 and the amount of additives that do not contribute to the reduction reaction in rich spike control calculated by a fourth model described later.
[0040] In the first and second models, the relationship between the input and output variables is learned by the NN so that the output variable coincides with the physical quantity to be estimated, and the weighting constants in the NN are determined in advance. In learning the first and second models, it is preferable to use data acquired during transient operation of the internal combustion engine 92 as training data.
[0041] FIG. 3 is a diagram for explaining the fourth model. The fourth model is a model for estimating "the amount of additives (amounts of CO, H2, HC, etc.) that do not contribute to the reduction reaction" in the above-mentioned rich spike control, and is configured by a physical model. Depending on the temperature of the exhaust gas from the internal combustion engine 92 and the temperature of the NSR catalyst 70, the reduction reaction between the NOx stored in the NSR catalyst 70 and the additives may not proceed sufficiently. In the fourth model, the amount of additives that do not contribute to such a reduction reaction is estimated. In FIG. 3, as an example of the fourth model, the Arrhenius equation for estimating the amount of CO that does not contribute to the reduction reaction is shown. When estimating the amount of H2 or HC that does not contribute to the reduction reaction, CO in FIG. 3 can be read as H2 or HC. In the Arrhenius equation, the frequency factor (A) and the activation energy (E) can use values obtained in advance by experiments or the like. For example, at least one or more of the parameters shown in the following items c1 to c3 can be adopted as input variables of the fourth model. The output variable Z of the fourth model is an estimate of the amount of additive that does not contribute to the reduction reaction under rich spike control under the conditions represented by the input variable (vector U). (c1) Catalyst information: at least one of the temperature of the front end of the NSR catalyst 70 and the temperature of the NSR catalyst 70 (c2) Exhaust information: at least one of the temperature of the exhaust gas from the internal combustion engine 92 and the air-fuel ratio of the exhaust gas flowing into the NSR catalyst 70. (c3) Information on additives: Amount of additives (CO, H2, HC, etc.) flowing into the NSR catalyst 70
[0042] The catalyst state estimation model 120 is a model for obtaining an estimated value of the amount of NOx stored in the NSR catalyst 70 one time later. The catalyst state estimation model 120 is configured by the following mathematical formula 1.
[0043]
number
[0044] The values in the formula 1 are as shown in the following items d1 to d5. (d1) NOxSt[k+1]: an estimate of the amount of NOx stored in the NSR catalyst 70 one time later (d2) NOxSt[k]: NOx storage amount (actual storage amount) obtained as a result of the previous estimation process (d3) NOx_NSR in [k]: Amount of NOx flowing into the NSR catalyst 70 (d4) NOx_NSR out [k]: the amount of outflowing NOx calculated from the NOx purification rate of the NSR catalyst 70 estimated by the first model and the amount of NOx flowing into the NSR catalyst 70 (d3 above), which is the amount of outflowing NOx flowing out from the NSR catalyst 70 at the current time (d5) NOx_rdct[k]: The amount of NOx reduced by rich spike control estimated by the second model
[0045] In addition, in the formula 1, the time t=kΔt means the current time, the time t=(k+1)Δt means one time later (one unit time later), and the time t=(k-1)Δt means one time ago (one unit time ago). k is an integer. Δt is a unit time in the error estimation device 10 (for example, the calculation period of the CPU 11, or the sampling period in each of the above-mentioned acquisition units 52 to 78). These points are the same in the explanations of FIG. 4 and subsequent figures.
[0046] The state estimation unit 110 estimates the NOx purification rate NOxConv[k] using the first model, the amount of additive Ad_thrmlytc[k] that does not contribute to the reduction reaction using the fourth model, and the amount of reduced NOx NOx_rdct[k] using the second model. After that, the state estimation unit 110 inputs these estimation results and the acquired values of the flow rate acquisition unit 52 and the like into the catalyst state estimation model 120 (Equation 1), thereby being able to obtain an estimated value NOxSt[k+1] of the amount of NOx stored in the NSR catalyst 70 one time later.
[0047] In this way, the catalyst state estimation model 120 is configured by a mathematical model. The catalyst state estimation model 120 is a model that estimates the amount of NOx occluded in the NSR catalyst 70 at one time later (time t=(k+1)Δt) by using the estimation results (NOx purification rate, NOx reduction amount) of the first model, which is an NN model (machine learning model), and the second model, and the estimation result (amount of additive that does not contribute to the reduction reaction) of the fourth model, which is a physical model. Therefore, by using the catalyst state estimation model 120, the state estimation unit 110 can estimate the amount of NOx occluded NOxSt[k+1], which is influenced by many factors (for example, the factors listed in items a1 to a2, b1 to b3, c1 to c3), with high accuracy and high speed while satisfying the laws of physics.
[0048] Furthermore, the amount of NOx occlusion in the NSR catalyst 70 varies depending on the amount of occlusion at the previous time (in other words, it varies depending on the time history). According to the catalyst state estimation model 120, the NOx occlusion amount NOxSt[k+1] at the next time k+1 is estimated using the estimation result NOxSt[k] at the current time k by the first, second, and fourth models. Therefore, the state estimation unit 110 can estimate the NOx occlusion amount at the next time with high accuracy based on the NOx occlusion amount of the NSR catalyst 70 at the previous time.
[0049] 4 is a flowchart showing the procedure of the error estimation process in the error estimation device 10. The error estimation process is a process for estimating an observation error in an acquisition unit that acquires a physical quantity of a predetermined component in the exhaust gas, specifically, an observation error of the inlet NOx concentration acquisition unit 54 (inlet NOx sensor). The error estimation process can be executed at any timing, and is executed repeatedly after the internal combustion engine 92 is started, for example. The error estimation process may be executed in response to a request from a user of the aftertreatment system 1, or in response to a request from another control unit in the vehicle or the like in which the aftertreatment system 1 is mounted.
[0050] The preconditions for the error estimation process will be described. In the error estimation process, a vector x shown in Equation 2 is used as a state quantity representing the internal state of the exhaust purification device 20. Here, NO in Equation 2 x,st represents the amount of NOx stored in the NSR catalyst 70, and θ in represents the observation error of the inlet NOx concentration acquisition unit 54, in other words, the observation error in the acquisition unit. Also, y[k] in Equation 3 represents the observation value of the outlet NOx sensor 99, in other words, the first physical quantity acquired by the sensor.
[0051]
number
number
[0052] In the error estimation process, the constraint 121 described in FIG. 1 and FIG. 2 is expressed as y p Here, R in Equation 4 is used. NOx,st is the estimated NOx storage rate. R NOx,st can be estimated by the state estimation unit 110 using the estimated value (state quantity) of the NOx occlusion amount estimated by the state estimation unit 110 using the catalyst state estimation model 120. NOx,st corresponds to "estimated value". obs x,in is the observation value of the inlet NOx concentration acquisition unit 54, in other words, the second physical quantity acquired by the acquisition unit. NO x,outis the observation value of the outlet NOx sensor 99, in other words, the first physical quantity acquired by the sensor. in is the observation error of the inlet NOx concentration acquisition unit 54, in other words, the observation error in the acquisition unit. in is estimated by the error estimator 111 through an error estimation process.
[0053] The constraint y shown in Equation 4 p The next estimated value A is equal to the calculated value B (y p =AB=0). In the following processing, the constraint y p We consider the fact that the condition holds true as if it were observed, and add it to the original observation equation as a pseudo-observation value to construct an extended observation equation. p In the following processing, the pseudo-measurement y p Also called. (A) Estimated value R NOx,st (B)NO x,out And NO obs x,in and θ in and the calculated value calculated from the above as shown in Equation 4
[0054]
number
[0055] In the error estimation process, Equation 5 is used as a discrete-time nonlinear state space expression of the state quantity representing the internal state of the exhaust purification device 20. Equation 6 is used as a discrete-time nonlinear state space expression of the observation value of the outlet NOx sensor 99 (first physical quantity acquired by the sensor). Here, f1(·) represents the catalyst state estimation model 120. w[k] represents normal white noise with mean 0 and covariance matrix Q[k]. v1[k] represents normal white noise with mean 0 and variance R1[k]. v2[k] represents normal white noise with mean 0 and variance R2[k].
[0056]
number
number
[0057] The processing procedure of the error estimation process will be described. In step S10, the error estimation unit 111 determines whether or not a processing start condition is satisfied. Specifically, for example, the error estimation unit 111 can determine that the processing start condition is satisfied when the temperature acquisition unit 78 is normal, the rich spike control unit 93 is normal, and the inlet NOx concentration acquisition unit 54 and the outlet NOx sensor 99 are active. If the start condition is satisfied (step S10: YES), the error estimation unit 111 transitions the processing to step S12. If the start condition is not satisfied (step S10: NO), the error estimation unit 111 ends the processing.
[0058] In step S12, the error estimation unit 111 acquires the current flow rate Q[k] inside the exhaust pipe 30 (at time t=kΔt) from the flow rate acquisition unit 52.
[0059] In step S14, the error estimation unit 111 calculates the amount of NOx NOx_NSR in Specifically, the state estimation unit 110 acquires the current NOx concentration [k] in the exhaust gas flowing into the NSR catalyst 70 from the inlet NOx concentration acquisition unit 54. Next, the error estimation unit 111 calculates the NOx amount NOx_NSR from the acquired NOx concentration [k] in the exhaust gas and the exhaust flow rate Q[k] acquired in step S12. in Calculate [k].
[0060] In step S16, the error estimation unit 111 acquires the current (time t=kΔt) bed temperature T[k] of the NSR catalyst 70 from the temperature acquisition unit 78. Instead of acquiring the bed temperature T[k] from the temperature acquisition unit 78, the error estimation unit 111 may estimate the bed temperature T[k] using a state estimator such as a nonlinear Kalman filter.
[0061] In step S18, the error estimation unit 111 acquires an input variable u[k-1] of the catalyst state estimation model 120 in the estimation process executed by the state estimation unit 110 one time period ago (time t=(k-1)Δt). Specifically, the error estimation unit 111 acquires input variables d2 to d5 of the catalyst state estimation model 120 one time period ago, which are shown below. (d2) NOxSt[k-1]: NOx storage amount (actual storage amount) obtained as a result of the previous estimation process (d3) NOx_NSR in [k-1]: Amount of NOx flowing into the NSR catalyst 70 (d4) NOx_NSR out [k-1]: Amount of NOx flowing out from the NSR catalyst 70 (d5) NOx_rdct[k-1]: The amount of NOx reduced by rich spike control estimated by the second model
[0062] In step S20, the error estimator 111 calculates 2n+1 sigma points using the following items e1 and e2. The 2n+1 sigma points correspond to sample points corresponding to the average value and standard deviation, and are expressed by Equation 7. Here, (√P) i means the i-th column of the square root matrix of the covariance matrix P. (e1) The state quantity x[k-1] obtained one time before, that is, the amount of NOx stored in the NSR catalyst 70, NO x,st [k-1] and the observation error θ of the inlet NOx concentration acquisition unit 54 in [k-1] (e2) Covariance matrix P[k-1] obtained one time before For items e1 and e2, the values obtained one time before are used from the second processing onwards, but any initial value can be used for the first processing.
[0063]
number
[0064] In step S22, the error estimation unit 111 calculates f=[f1f2] as shown in Equation 8. TUsing Sigma Point X i  ̄[k] is updated. Note that f1 is the NOx storage amount NO x,st f2 is a time update equation for the observation error of the inlet NOx concentration acquisition unit 54 in the state quantity vector x (Equation 2).
[0065]
number
[0066] In step S24, the error estimation unit 111 calculates the prior state estimates of the NOx occlusion amount and the observation error at the current time k. Specifically, the error estimation unit 111 calculates the prior state estimates of the sigma point X i By inputting ̂ k into the following formula 9, the prior state estimated value x̂ k of the NOx occlusion amount and the variation amount at the present time (time t=kΔt) is obtained (the left term of formula 9). Here, the error estimator 111 calculates each sigma point X i For  ̄, the weighting coefficient w i is calculated using Equation 10.
[0067]
number
number
[0068] In step S26, the error estimation unit 111 uses Equation 11 to calculate the prior error covariance matrix P̂[k] of the amount of NOx occlusion and the amount of variation at the current time k.
[0069]
number
[0070] In step S28, the error estimator 111 calculates 2n+1 sigma points at the current time k by using Equation 12 and the following items f1 and f2. (f1) State quantity x[k] at the current time, i.e., the amount of NOx stored in the NSR catalyst 70, NO x,st [k] and the measurement error θ of the inlet NOx concentration acquisition unit 54 in [k] (f2) Covariance matrix P[k] of the current time
[0071] Thereafter, the error estimation unit 111 uses the observation value y[k] of the outlet NOx sensor 99 (the first physical quantity acquired by the sensor) and the observation equation y[k]=h1(x[k]) to calculate the sigma point Y i  ̄[k] is calculated (Equation 13, upper part). Here, the observation equation y[k]=h1(x[k]) is the relationship between the observation value of the outlet NOx sensor 99 and the state quantity of NOx adsorption NO x,st and the measurement error θ of the inlet NOx concentration acquisition unit 54 in This is the relational expression.
[0072]
number
number
[0073] In step S29, the error estimator 111 calculates the pseudo observation amount y p [k] and the observation equation y p [k] = h2(x[k]) and the pseudo-observable y p Sigma point Y with respect to [k] p,i Calculate  ̄[k] (lower part of Equation 13). Here, the observation equation y p [k]=h2(x[k]) is the pseudo-observable y p and the state quantity (NOx storage amount NO x,st , the measurement error θ of the inlet NOx concentration acquisition unit 54 in ) is the relation:
[0074] In step S30, the error estimation unit 111 calculates the preliminary output estimate ŷ  ̄[k],y p The preliminary output estimate y^ ̄[k] is an estimate of the observation value of the outlet NOx sensor 99 at the current time k, prior to the acquisition of the observation value of the outlet NOx sensor 99 (the first physical quantity acquired by the sensor). Similarly, the preliminary output estimate y p ̂ ̂[k] is an estimate of the pseudo observation. Specifically, the error estimator 111 estimates the sigma point Y i  ̄[k] into the upper part of Equation 14 to calculate the preliminary output estimate ŷ ̄[k]. In addition, the error estimator 111 calculates the sigma point Y p,i By inputting  ̄[k] into the lower part of Equation 14, the preliminary output estimate y p Calculate the weighting factor w i is as explained in Equation 10.
[0075]
number
[0076] In step S32, the error estimator 111 calculates the pre-output error covariance matrix P yy  ̄[k],P ypyp Specifically, the error estimator 111 calculates the sigma point Y i  ̄[k] and the prior output estimate y^ ̄[k] obtained in step S30 are input into the following formula 15 to obtain the prior output error covariance matrix P yy In addition, the error estimator 111 obtains the sigma point Y p,i  ̄[k] and the prior output estimate y p By inputting ^ ̄[k] and the above into Equation 15 below, the a priori output error covariance matrix P ypyp Obtain the weighting coefficient w i is as explained in Equation 10.
[0077]
number
[0078] In step S34, the error estimator 111 calculates the prior state-output error covariance matrix P xy  ̄[k],P xyp Specifically, the error estimator 111 calculates the sigma point X i  ̄[k], the prior state estimate x^ ̄[k] calculated in step S24, and the sigma point Y i  ̄[k] and the preliminary output estimate y^ ̄[k] calculated in step S30 are input to the upper part of the following formula 16, and the observed value of the outlet NOx sensor 99 and the state quantity (NOx storage amount NO x,st , the measurement error θ of the inlet NOx concentration acquisition unit 54 in ) and the prior state-output error covariance matrix P xy Calculate  ̄[k]. Weight coefficient w i is as explained in Equation 10.
[0079] In step S34, the error estimator 111 calculates the sigma point X i  ̄[k], the prior state estimate x^ ̄[k] calculated in step S24, and the sigma point Y p,i  ̄[k] and the prior output estimate y calculated in step S30 p By inputting ^ ̄[k] into the lower part of the following formula 16, the pseudo observation quantity and the state quantity (NOx storage amount NO x,st , the measurement error θ of the inlet NOx concentration acquisition unit 54 in ) and the prior state-output error covariance matrix P xyp Calculate  ̄[k]. Weight coefficient w i is as explained in Equation 10.
[0080]
number
[0081] In step S36, the error estimator 111 calculates the Kalman gain. Specifically, the error estimator 111 calculates the Kalman gain using the prior state-output error covariance matrix P xy  ̄[k],P xyp  ̄[k] and the pre-output error covariance matrix P obtained in step S32 yy  ̄[k],P ypyp The Kalman gain G[k] is calculated by inputting  ̄[k] into the following formula 17.
[0082]
number
[0083] In step S38, the error estimation unit 111 acquires the observation value y[k] of the outlet NOx sensor 99. In other words, the observation value y[k] of the outlet NOx sensor 99 is an actual measurement value of the NOx concentration downstream of the NSR catalyst 70 (in other words, the outlet of the exhaust pipe 30).
[0084] In step S40, the error estimator 111 calculates the state estimates x1^[k] and x2^[k]. Specifically, the error estimator 111 calculates the prior state estimate x^ ̄[k] calculated in step S24, the Kalman gain G[k] calculated in step S36, the observed value y[k] of the outlet NOx sensor 99 acquired in step S38, and the pseudo observed quantity y p and the prior output estimate y^ ̄[k],y calculated in step S30. p By inputting ^ ̄[k] and x^[k] into the following equation 18, the state estimates x1^[k] and x2^[k] are obtained.
[0085]
number
[0086] Here, as explained in Equation 2, the vector x represented by x^ represents the amount of NOx stored in the NSR catalyst 70, NO x,st and the measurement error θ of the inlet NOx concentration acquisition unit 54 inIn this way, the NOx storage amount NO x,st In order to estimate the NOx occlusion amount NO , the error estimation unit 111 estimates the NOx occlusion amount NO , which is more accurate than the NOx occlusion amount estimated by the state estimation unit 110. x,st Further, the error estimation unit 111 uses the state estimation theory to obtain the observation error θ of the inlet NOx concentration acquisition unit 54 from the amount of NOx occlusion estimated by the state estimation unit 110 and the observation value (first physical quantity) of the outlet NOx sensor 99. in That is, the error estimation unit 111 can estimate the measurement error θ in can be estimated.
[0087] In step S42, the error estimator 111 calculates the a posteriori error covariance matrix P[k] using Equation 19, and ends the process. The a posteriori error covariance matrix P[k] is used in the next process.
[0088]
number
[0089] 5 shows the simulation results when the correct value of the measurement error of the inlet NOx concentration acquisition unit 54 is set to +20%. in FIG. 5B shows the time-series change in the estimation result of the NOx storage amount NO x,st 5A and 5B show the time-series change in the estimation result of the method described in the above embodiment (specifically, the pseudo observation quantity y p The dashed line shows the estimation results obtained by the method using the pseudo-observation y p The dashed line shows the estimation result obtained without using the observation error θ in and NOx storage amount NO x,st For both the pseudo-observables y p The method using the pseudo-measurement y pIt can be seen that the method is closer to the correct value shown by the solid line and has higher accuracy than the method that does not use the method (dash line).
[0090] 6 shows the simulation results when the correct value of the measurement error of the inlet NOx concentration acquisition unit 54 is set to −20%. in FIG. 6B shows the time-series change in the estimation result of the NOx storage amount NO x,st The figure shows the change over time in the estimated results of . Even if the correct value of the observation error is on the negative side, the observation error θ in and NOx storage amount NO x,st For both the pseudo-observables y p The method using the pseudo-measurement y p It can be seen that the method is closer to the correct value shown by the solid line and has higher accuracy than the method that does not use the method (dash line).
[0091] 7 is a flowchart showing the procedure of the abnormality detection process in the error estimation device 10. The abnormality detection process is a process for detecting an abnormality (deterioration, failure, etc.) in the inlet NOx concentration acquisition unit 54. The abnormality detection process can be executed at any timing, and is executed automatically after the error estimation process (FIG. 4) is completed, for example. The abnormality detection process may be executed in response to a request from a user of the aftertreatment system 1, or in response to a request from another control unit in the vehicle or the like in which the aftertreatment system 1 is mounted.
[0092] In step S100, the control unit 112 determines whether the calculation of the state quantity x in the abnormality detection process (FIG. 4) has ended normally. If the calculation has not ended normally (step S100: NO), the control unit 112 ends the process. If the calculation has ended normally (step S100: YES), the control unit 112 transitions the process to step S102.
[0093] In step S102, the control unit 112 calculates the measurement error θ in In step S104, the control unit 112 obtains the observation error θ inis equal to or less than a predetermined threshold. The predetermined threshold is an allowable upper limit of the measurement error of the inlet NOx concentration acquisition unit 54, and is determined in advance and stored in the storage unit 12.
[0094] Observation error θ in If the measurement error θ is equal to or smaller than the predetermined threshold (step S104: YES), the control unit 112 determines that the inlet NOx concentration acquisition unit 54 is normal (step S106) and ends the process. in is not equal to or less than the predetermined threshold (step S104: NO), the control unit 112 determines that there is an abnormality in the inlet NOx concentration acquisition unit 54 (step S108). In step S108, the control unit 112 may display an alert on the display unit 13, or may leave an abnormality determination log in the storage unit 12. At this time, the control unit 112 may in The value of the measurement error θ itself may be output by displaying an alert or leaving a log. The alert can be displayed in any form, such as turning on an LED, displaying a different color of an LED, or displaying a predetermined character or icon. in The "degree of deterioration" estimated from the size of the defect may be displayed.
[0095] Note that at least one of the following processes g1 and g2 may be executed instead of or together with the abnormality detection process described with reference to FIG. (g1) Correction of the observed value of the inlet NOx concentration acquisition unit 54 The inlet NOx concentration acquisition unit 54 calculates the observed value (actual observed value: second physical quantity) with an observation error θ in The corrected second physical amount may be used in the rich spike control of the process g2. The corrected second physical amount may be transmitted to another ECU and used for control in that ECU. (g2) Rich spike control (additive supply control) The control unit 112 detects the amount of stored NOx NO obtained by the abnormality detection process (FIG. 4). x,stand the second physical amount corrected by process g1 are transmitted to the rich spike control unit 93. The rich spike control unit 93 uses these values to perform the rich spike control unit 93, thereby controlling the supply of additives (CO, H2, HC, etc.).
[0096] As described above, according to the error estimation device 10 of the first embodiment, the error estimation unit 111 uses the state estimation theory to estimate the state quantity estimated by the state estimation unit 110 and the first physical quantity acquired by the outlet NOx sensor 99 (sensor), and calculates the observation error θ in Therefore, the measurement error θ in the inlet NOx concentration acquisition unit 54 (acquisition unit) is not influenced by the operating state of the internal combustion engine 92 or the flow rate pattern of the predetermined component. in Since we can estimate the measurement error θ in The error estimation unit 111 can improve the frequency of abnormality diagnosis of the acquisition unit using the above (FIG. 7). In addition, the error estimation unit 111 satisfies a constraint condition (Equation 4, pseudo observation quantity y p ) to obtain the observation error θ in Here, the measurement error θ in the inlet NOx concentration acquisition unit 54 (acquisition unit) is generally estimated. in is reflected in the observation data only through the state variables. However, according to the error estimation device 10, the constraint condition y p , the observation error θ estimated (in the past) by the error estimation unit 111 in By including in As information to be taken into account when estimating, the measurement error θ estimated (in the past) by the error estimator 111 in Therefore, the observation error θ in In the case where the error estimator 111 is configured only with the unknown parameters (pseudo observation quantity y p In comparison with the case where the measurement error θ inAs a result, the error estimation device 10 can improve the estimation accuracy of the measurement error θ in It is possible to improve the technology for estimating the above.
[0097] According to the error estimation device 10 of the first embodiment, the error estimation unit 111 estimates the measurement error θ in can be estimated.
[0098] Furthermore, according to the error estimation device 10 of the first embodiment (Equation 4), the NOx occlusion rate estimated value A(R NOx,st ) is equal to the calculated value B calculated from the results of acquisition and estimation by the outlet NOx sensor 99 (sensor), the inlet NOx concentration acquisition unit 54 (acquisition unit), and the error estimation unit 111. p =AB=0 can be defined.
[0099] Furthermore, according to the error estimation device 10 of the first embodiment, the error estimation unit 111 uses an algorithm for obtaining an estimated value that minimizes the mean square error of the state quantity when the mean square error is used as the evaluation function as the state estimation theory, thereby making it possible to estimate the state quantity x (i.e., the observation error θ in ) can be estimated.
[0100] Furthermore, according to the error estimation device 10 of the first embodiment, the inlet NOx concentration acquisition unit 54 (acquisition unit) estimates the observation error θ in Since the acquired second physical quantity is corrected using the second physical quantity, the accuracy of the second physical quantity can be improved (process g1). As a result, the accuracy of various controls (for example, additive supply control for purifying harmful substances in the NSR catalyst 70: process g2) performed using the corrected second physical quantity can be improved.
[0101] Furthermore, according to the error estimating device 10 of the first embodiment, the control unit 112 calculates the measurement error θ in is used to detect an abnormality in the inlet NOx concentration acquisition unit 54 (acquisition unit) (FIG. 7). Therefore, abnormality detection can be performed without requiring a special operation mode for the purpose of abnormality detection (fault diagnosis), and the frequency of abnormality detection can be improved.
[0102] Furthermore, according to the aftertreatment system 1 of the first embodiment, the measurement error θ in Therefore, it is possible to provide an aftertreatment system 1 capable of supplying an additive taking the above into consideration (process g2). This makes it possible to improve the purification efficiency in the aftertreatment system 1, which also contributes to improving fuel efficiency.
[0103] <Second embodiment> 8 is a block diagram of an aftertreatment system 1a according to a second embodiment. In the second embodiment, a configuration for estimating the total amount of NOx occlusion in a plurality of NSR catalysts when a plurality of NSR catalysts are installed will be described. The aftertreatment system 1a according to the second embodiment includes an error estimation device 10a instead of the error estimation device 10, and an exhaust purification device 20a instead of the exhaust purification device 20.
[0104] In addition to the configuration described in the first embodiment, the exhaust purification device 20a includes a second NSR catalyst 71 disposed downstream of the NSR catalyst 70. Hereinafter, the NSR catalyst 70 is also referred to as the first NSR catalyst 70 for the sake of distinction. The second NSR catalyst 71 is the same type of storage reduction catalyst as the first NSR catalyst 70. In this embodiment, "same type catalyst" means a catalyst having the same or similar exhaust purification mechanism. The second NSR catalyst 71 purifies the NOx in the exhaust by storing and storing in the second NSR catalyst 71 the harmful substances (NOx) that cannot be stored in the first NSR catalyst 70 and leak downstream. Hereinafter, the multiple catalysts of the same type (first and second NSR catalysts 70, 71) mounted in the exhaust purification device 20a are also collectively referred to as the "catalyst group CG." The first NSR catalyst 70 is also referred to as the "catalyst located at the most upstream position," and the second NSR catalyst 71 is also referred to as the "catalyst located at the most downstream position."
[0105] In this embodiment, the rich spike control (reduction control) described in the first embodiment is performed in each of the multiple catalysts. Specifically, the NOx stored in the first NSR catalyst 70 is reduced to nitrogen gas (N2) by CO, H2, HC, etc. discharged from the internal combustion engine 92. In addition, the NOx stored in the second NSR catalyst 71 is reduced to nitrogen gas (N2) by CO, H2, HC, etc. that are not used in the reduction control in the first NSR catalyst 70 and flow out from the first NSR catalyst 70.
[0106] The error estimation device 10a further includes a second temperature acquisition unit 79 consisting of a sensor that measures the bed temperature of the second NSR catalyst 71. The error estimation device 10a also includes a state estimation unit 110a instead of the state estimation unit 110, and an error estimation unit 111a instead of the error estimation unit 111. The error estimation device 10a also includes a catalyst state estimation model 120a instead of the catalyst state estimation model 120, and a constraint condition 121a instead of the constraint condition 121. The error estimation device 10a may also include a temperature estimation unit that estimates the temperature of the downstream catalyst (the second NSR catalyst 71 in the example of FIG. 5) instead of the second temperature acquisition unit 79. The temperature estimation unit calculates the temperature of the second NSR catalyst 71 from the temperature of the first NSR catalyst 70 acquired by the temperature acquisition unit 78, using a calculation formula, a map, or the like prepared in advance. The temperature estimation unit may calculate the temperature of the second NSR catalyst 71 by taking into consideration the temperature of the exhaust gas from the internal combustion engine 92, the flow rate of the exhaust gas, the heat of reaction of the first NSR catalyst 70, and any other parameters.
[0107] Among the catalyst state estimation models 120a, the first, second, and fourth models are configured to output the NOx purification rate, the amount of NOx reduction, and the amount of additives that do not contribute to the reduction reaction (amounts of CO, H2, HC, etc.) of the catalyst group CG as a whole. The state estimation unit 110a uses such catalyst state estimation models 120a to estimate the amount of NOx occlusion of the catalyst group CG as a whole.
[0108] Hereinafter, with reference to FIG. 4, only the parts of the process that are different from the first embodiment will be described. In step S14, the error estimation unit 111a acquires the amount of NOx flowing into the most upstream catalyst (first NSR catalyst 70) of the catalyst group CG. In step S16, the error estimation unit 111a acquires the bed temperature of the first NSR catalyst 70 from the first temperature acquisition unit 78, and acquires the bed temperature of the second NSR catalyst 71 from the second temperature acquisition unit 79. In steps S20 and S22, the error estimation unit 111a calculates and updates the sigma point using the total NOx occlusion amount of the entire catalyst group CG. From this point on, it is sufficient to replace the "NOx occlusion amount" described in FIG. 4 with the "total NOx occlusion amount", and the other process contents are the same as those of the first embodiment described in FIG. 4. That is, in the error estimation device 10a, the NOx occlusion amount (NOx occlusion amount NOxSt) estimated by the state estimation unit 110a using the catalyst state estimation model 120a and the corrected occlusion amount (NOx occlusion amount NOxSt) estimated by the error estimation unit 111a through the process of FIG. x,st ) are the total amounts of NOx stored in the catalyst group CG as a whole.
[0109] Thus, according to the error estimation device 10a of the second embodiment, when a plurality of first and second NSR catalysts 70, 71 of the same type are provided in the exhaust pipe 30 (main flow passage), the state estimation unit 110a regards the plurality of catalysts 70, 71 as one catalyst (catalyst group CG) and estimates the total value of the NOx occlusion amount of the entire catalyst group CG. Therefore, by incorporating the influence of the exhaust pipe 30 between the first NSR catalyst 70 and the second NSR catalyst 71, etc., in advance into the mathematical model (catalyst state estimation model 120a), the total value of the NOx occlusion amount can be estimated with high accuracy without acquiring information about the exhaust pipe 30, etc. In other words, since the number of sensors, etc. can be reduced compared to the case where the NOx occlusion amount of each catalyst 70, 71 is estimated individually, the number of parts of the aftertreatment system 1a can be reduced, and the calculation load of the state estimation unit 110a can be reduced, thereby speeding up the processing. Furthermore, when the multiple catalysts 70, 71 provided in the exhaust pipe 30 are of the same type, the state estimation unit 110a regards them as one catalyst (catalyst group CG) and estimates the total value of the NOx occlusion amount. If the catalysts are of the same type, there is no difference in the catalyst information items that affect the total value of the NOx occlusion amount (for example, the catalyst temperature, the amount of additive occluded in the catalyst), so the state estimation unit 110a can accurately estimate the total value of the NOx occlusion amount. Furthermore, the sensor is provided upstream of the most upstream catalyst of the multiple catalysts 70, 71 (specifically, the inlet NOx concentration acquisition unit 54 upstream of the first NSR catalyst 70) or downstream of the most downstream catalyst (specifically, the outlet NOx sensor 99 downstream of the second NSR catalyst 71). Therefore, the error estimation unit 111a can handle the sensor (the outlet NOx sensor 99 or the inlet NOx concentration acquisition unit 54) as a sensor provided upstream or downstream of the catalyst group CG consisting of the multiple catalysts 70, 71.
[0110] <Third embodiment> FIG. 9 is a block diagram of an aftertreatment system 1b in the third embodiment. In the aftertreatment system 1b in the third embodiment, a three-way catalyst is used instead of the NSR catalyst described in the first embodiment. The three-way catalyst purifies CO, HC, and NOx (harmful substances) in the exhaust gas by using O2 (oxygen) generated by an oxidation-reduction reaction as an additive. In this embodiment, O2 is exemplified as the predetermined component in the exhaust gas. The aftertreatment system 1b includes an error estimation device 10b instead of the error estimation device 10, and an exhaust purification device 20b instead of the exhaust purification device 20.
[0111] The exhaust purification device 20b is provided with a three-way catalyst 80 instead of the NSR catalyst 70, and is not provided with a rich spike control unit 93. The three-way catalyst 80 removes CO, HC, and NOx contained in the exhaust gas by oxidizing or reducing them, respectively, and purifies these harmful substances. The three-way catalyst 80 has an oxygen storage capacity (OSC) and can store (store) O2 in the exhaust gas flowing into the three-way catalyst 80. In this embodiment, the "additive" means O2 used in the oxidation reaction. In this embodiment, the three-way catalyst 80 corresponds to the "catalyst". The combustion state control unit 91 corresponds to the "supply control unit", and the internal combustion engine 92 corresponds to the "supply unit".
[0112] The error estimation device 10b includes an inlet oxygen concentration acquisition unit 84 in place of the inlet NOx concentration acquisition unit 54, a front end temperature acquisition unit 86 in place of the front end temperature acquisition unit 76, a temperature acquisition unit 88 in place of the temperature acquisition unit 78, an outlet oxygen sensor 99d in place of the outlet NOx sensor 99, a state estimation unit 110b in place of the state estimation unit 110, an error estimation unit 111b in place of the error estimation unit 111, a catalyst state estimation model 120b in place of the catalyst state estimation model 120, and a constraint condition 121b in place of the constraint condition 121.
[0113] The inlet oxygen concentration acquisition unit 84 acquires the O2 concentration in the exhaust gas from the internal combustion engine 92. The inlet oxygen concentration acquisition unit 84 may be realized, for example, by acquiring a measurement signal measured by an A / F sensor provided in the exhaust pipe 30, or may be realized by acquiring a measurement signal measured by an oxygen sensor. The inlet oxygen concentration acquisition unit 84 may estimate the O2 concentration in the exhaust gas from an intake air amount signal to the internal combustion engine 92 or a fuel injection amount signal. The front end temperature acquisition unit 86 is a sensor that measures the temperature near the inlet (front end) of the three-way catalyst 80. The temperature acquisition unit 88 is a sensor that measures the bed temperature of the three-way catalyst 80. The outlet oxygen sensor 99d is a sensor that acquires the O2 concentration in the exhaust gas downstream of the three-way catalyst 80 (in other words, at the outlet of the exhaust pipe 30), and can be realized, for example, by an A / F sensor or an oxygen sensor.
[0114] In this embodiment, the predetermined component in the exhaust gas is O2, the oxygen concentration acquisition unit 84 corresponds to the "acquisition unit", and the observation value of the oxygen concentration acquisition unit 84 corresponds to the "second physical quantity". Additionally, the outlet oxygen sensor 99d corresponds to the "sensor", and the observation value of the outlet oxygen sensor 99d corresponds to the "first physical quantity".
[0115] The catalyst state estimation model 120b includes a first model, which is configured by a machine learning model and outputs the CO purification rate of the three-way catalyst 80, a second model, which is configured by a machine learning model and outputs the NOx purification rate in the three-way catalyst 80, a third model, which is configured by a machine learning model and outputs the THC purification rate in the three-way catalyst 80, a fourth model, which is configured by a machine learning model and outputs the outflow amount of an additive (O2) flowing out from the three-way catalyst 80, and a fifth model, which is configured by a physical model and outputs the ratio of NO (nitric oxide) and NO2 (nitrogen dioxide) in NOx flowing into the three-way catalyst 80. The state estimation unit 110b uses such catalyst state estimation model 120b to obtain the oxygen storage amount OSA of the three-way catalyst 80 using physical laws such as the law of conservation of mass, the law of material balance, the law of heat balance, and the law of energy balance.
[0116] The error estimation unit 111b performs the same process as in FIG. 4 (error estimation process) of the first embodiment, using the oxygen storage amount OSA (the amount of storage of a predetermined component) of the three-way catalyst 80 and the observation error of the oxygen concentration acquisition unit 84 as vector x in Formula 2, to estimate the oxygen storage amount OSA and the observation error of the oxygen concentration acquisition unit 84. Note that the constraint condition 121b may be designed in the same way as Formula 4.
[0117] The error estimation device 10b may use the oxygen storage amount OSA estimated by the error estimation unit 111b and the observation error to perform an abnormality detection process (similar to the process in FIG. 7) for the oxygen concentration acquisition unit 84 or to correct the observed value by the oxygen concentration acquisition unit 84 (similar to the process g1). Furthermore, the error estimation device 10b uses at least one of the oxygen storage amount OSA estimated by the error estimation unit 111b and the value corrected by the process g1 to control the injection of air or fuel by the combustion state control unit 91, thereby controlling the air-fuel ratio in the internal combustion engine 92 to each of the lean, stoichiometric, and rich states, thereby controlling the additive amount (i.e., the amount of O2). In this way, the third embodiment using the three-way catalyst 80 can also achieve the same effects as the first embodiment described above.
[0118] <Modifications of this embodiment> The present invention is not limited to the above-described embodiment, and can be embodied in various forms without departing from the spirit of the present invention. For example, a part of the configuration realized by hardware may be replaced by software, and conversely, a part of the configuration realized by software may be replaced by hardware. In addition, for example, the following modifications are also possible.
[0119] [Variation 1] In the above embodiment, an example of the configuration of the aftertreatment system is shown. However, the configuration of the aftertreatment system can be modified in various ways. For example, the exhaust purification device of the aftertreatment system is equipped with a combination of multiple catalysts selected from an SCR catalyst, an NSR catalyst, and a three-way catalyst, and the error estimation device may estimate the amount of storage of different predetermined components (e.g., NOx, NH3, O2, A / F) for each of the multiple catalysts and the measurement error in an acquisition unit provided upstream or downstream of the catalyst using state estimation theory. In addition, the exhaust purification device may be equipped with other components such as a particulate matter removal filter (DPF: Diesel Particulate Filter) that removes particulate matter (PM) and a hydrolysis catalyst.
[0120] For example, the temperature acquisition unit that acquires the catalyst bed temperature may be provided in front of (near the inlet) or behind (near the outlet) the catalyst. For example, at least one of the flow rate, NOx concentration, front end temperature, and catalyst temperature acquired by the flow rate acquisition unit, NOx concentration acquisition unit, front end temperature acquisition unit, and temperature acquisition unit, respectively, may be an estimated value obtained using a neural network, a map, or the like, instead of a measured value by a sensor. For example, the error estimation device may not be equipped with a display unit. Furthermore, the error estimation device may use another device (such as a smartphone or an in-vehicle device) connected via a network as the display unit.
[0121] [Variation 2] In the above embodiment, an example of the error estimation process (FIG. 4) and the anomaly detection process (FIG. 7) has been shown. However, the contents of these processes can be modified in various ways. For example, the order of execution of each step may be changed, some steps may be omitted, and other steps not described may be executed. For example, the anomaly detection process (FIG. 7) may not be executed.
[0122] For example, in the error estimation process, the observation error θ in Instead, the observation error θ of the outlet NOx sensor outIn this case, in the error estimation process described in FIG. 4, the "inlet NOx concentration acquisition unit 54" and the "outlet NOx sensor 99" may be read in reverse and the process may be executed. Furthermore, the error estimation device may estimate the observation error θ out and the observation error θ of the inlet NOx concentration acquisition part in Both may be estimated.
[0123] For example, the state estimation unit may estimate the NOx occlusion amount, the NOx purification rate, the NOx reduction amount, and the amount of additive that does not contribute to the reduction reaction using any means such as a NN, a map, a physical formula, etc., different from the above-mentioned configuration (catalyst state estimation model 120, the first, second, and fourth models). For example, in the configuration of the first embodiment, the state estimation unit may estimate the NOx occlusion amount using a fifth model, instead of the second model, that estimates the outflow additive amount, which is the amount of additive that flows out from the NSR catalyst.
[0124] Although the present aspect has been described above based on the embodiment and modified examples, the above-mentioned embodiment of the aspect is intended to facilitate understanding of the present aspect and does not limit the present aspect. The present aspect may be modified or improved without departing from the spirit and scope of the claims, and equivalents are included in the present aspect. Furthermore, if a technical feature is not described as essential in this specification, it may be deleted as appropriate.
[0125] The present invention can also be realized in the following forms. [Application example 1] An error estimation device, comprising: An exhaust purification device including a main flow path through which exhaust gas flows and a catalyst that purifies harmful substances in the exhaust gas, a sensor that acquires a first physical quantity of a predetermined component in the exhaust gas; an acquisition unit that acquires a second physical quantity for the predetermined component in the exhaust gas; a state estimation unit that estimates a state quantity that represents an internal state of the exhaust purification device and that affects the first physical quantity acquired by the sensor; an error estimating unit that estimates an observation error in the acquiring unit from the state quantity estimated by the state estimating unit and the first physical quantity acquired by the sensor, using a state estimation theory; Equipped with The error estimator, an estimated value estimated by the state estimation unit using the state quantity estimated by the state estimation unit; a calculated value calculated from the first physical quantity acquired by the sensor, the second physical quantity acquired by the acquisition unit, and the observation error estimated by the error estimating unit; and estimating the measurement error using a constraint that the estimated value and the calculated value are equal. [Application example 2] The error estimation device according to Application Example 1, The catalyst is an NSR catalyst that purifies NOx as a harmful substance, the sensor is an outlet NOx sensor that acquires a NOx concentration in the exhaust gas downstream of the catalyst as the first physical quantity, the acquisition unit is an inlet NOx sensor that acquires a NOx concentration in the exhaust gas upstream of the catalyst as the second physical quantity, The state estimation unit estimates an amount of NOx occlusion in the catalyst as the state quantity. [Application example 3] The error estimating device according to Application Example 1 or 2, The constraint y p is defined by the following formula: TIFF2024158869000022.tif21170In the above formula, R NOx,st is the estimated NOx storage rate, NO obs x,in is the second physical quantity acquired by the acquisition unit, NO x,out is the first physical quantity acquired by the sensor, θ inis the observation error estimated by the error estimating unit. [Application example 4] The error estimating device according to any one of Application Examples 1 to 3, an error estimation device, wherein the error estimating unit uses an algorithm for obtaining an estimate that minimizes a mean square error of a state quantity when a mean square error is used as an evaluation function as the state estimation theory. [Application example 5] The error estimating device according to any one of Application Examples 1 to 4, The acquisition unit corrects the acquired second physical quantity by using the observation error estimated by the error estimating unit. [Application Example 6] The error estimating device according to any one of Application Examples 1 to 5, further comprising: and a control unit that detects an abnormality in the acquisition unit by using the observation error estimated by the error estimating unit. [Application Example 7] 1. An aftertreatment system comprising: An error estimating device according to any one of Application Examples 1 to 6; A catalyst that is provided in a main flow passage through which the exhaust gas flows and purifies harmful substances in the exhaust gas; a supply unit that supplies an additive for purifying harmful substances to the catalyst; A supply control unit that controls the supply of the additive in the supply unit; An aftertreatment system comprising: [Application Example 8] An error estimation method, comprising: acquiring a first physical quantity of a predetermined component in exhaust gas using a sensor for an exhaust gas purification device including a main flow path through which exhaust gas flows and a catalyst that purifies harmful substances in the exhaust gas; acquiring a second physical quantity for the predetermined component in the exhaust gas using an acquisition unit; a state estimating step of estimating a state quantity that represents an internal state of the exhaust purification device and that affects the first physical quantity acquired by the sensor; an error estimating step of estimating an observation error in the acquisition unit from the estimated state quantity and the acquired first physical quantity using a state estimation theory; Run In the error estimation step, an estimated value estimated by the state estimation step using the estimated state quantity; a calculated value calculated from the acquired first physical quantity, the acquired second physical quantity, and the measurement error estimated in the error estimating step; The error estimation method estimates the measurement error using a constraint that the estimated value and the calculated value are equal for [Application Example 9] A computer program comprising: A function of acquiring a first physical quantity of a predetermined component in exhaust gas by using a sensor, for an exhaust gas purification device including a main flow path through which exhaust gas flows and a catalyst that purifies harmful substances in the exhaust gas; a function of acquiring a second physical quantity for the predetermined component in the exhaust gas using an acquisition unit; a state estimation function that estimates a state quantity that represents an internal state of the exhaust purification device and that affects the first physical quantity acquired by the sensor; an error estimation function that estimates an observation error in the acquisition unit from the estimated state quantity and the acquired first physical quantity using a state estimation theory; Run the command, The error estimation function an estimated value estimated by the state estimation function using the estimated state quantity; a calculated value calculated from the acquired first physical quantity, the acquired second physical quantity, and the observation error estimated by the error estimation function; and estimating the measurement error using a constraint that the estimated value and the calculated value are equal for [Explanation of symbols]
[0126] 1, 1a, 1b...Aftertreatment system 10,10a,10b...error estimation device 11...CPU 12...Storage section 13…Display section 20, 20a, 20b...Exhaust gas purification device 30…Exhaust pipe 52...Flow rate acquisition section 53…Exhaust temperature acquisition unit 54…Inlet NOx concentration acquisition unit (acquisition unit) 70...Catalyst, 1st NSR catalyst 71…2nd NSR catalyst 76...Front end temperature acquisition section 78...Temperature acquisition section, first temperature acquisition section 79...Second temperature acquisition section 80…Three-way catalyst 84...Oxygen concentration acquisition unit (acquisition unit) 86…Front end temperature acquisition section 88…Temperature acquisition section 91...Combustion state control unit 92…Internal combustion engine 93…Rich spike control section 99...Outlet NOx sensor (sensor) 99d…Outlet oxygen sensor (sensor) 110, 110a, 110b... State estimation unit 111,111a,111b...error estimation section 112...Control unit 120, 120a, 120b...Catalyst state estimation model 121,121a,121b…constraint conditions
Claims
1. An error estimation device, comprising: An exhaust purification device including a main flow path through which exhaust gas flows and a catalyst that purifies harmful substances in the exhaust gas, a sensor that acquires a first physical quantity of a predetermined component in the exhaust gas; an acquisition unit that acquires a second physical quantity for the predetermined component in the exhaust gas; a state estimation unit that estimates a state quantity that represents an internal state of the exhaust purification device and that affects the first physical quantity acquired by the sensor; an error estimator that estimates an observation error in the acquisition unit from the state quantity estimated by the state estimator and the first physical quantity acquired by the sensor using a state estimation theory; Equipped with The error estimator, an estimated value estimated by the state estimation unit using the state quantity estimated by the state estimation unit; a calculated value calculated from the first physical quantity acquired by the sensor, the second physical quantity acquired by the acquisition unit, and the observation error estimated by the error estimating unit; and estimating the measurement error using a constraint that the estimated value and the calculated value are equal.
2. 2. The error estimation device according to claim 1, The catalyst is an NSR catalyst that purifies NOx as a harmful substance, the sensor is an outlet NOx sensor that acquires a NOx concentration in the exhaust gas downstream of the catalyst as the first physical quantity, the acquisition unit is an inlet NOx sensor that acquires a NOx concentration in the exhaust gas upstream of the catalyst as the second physical quantity, The state estimation unit estimates an amount of NOx occluded in the catalyst as the state quantity.
3. 3. The error estimation device according to claim 2, The constraint y p is defined by the following formula: In the above formula, R NOx,st is the estimated NOx storage rate, NO obs x,in is the second physical quantity acquired by the acquisition unit, NO x,out is the first physical quantity acquired by the sensor, θ in is the observation error estimated by the error estimating unit.
4. 4. The error estimation device according to claim 3, an error estimation device, wherein the error estimating unit uses an algorithm for obtaining an estimate that minimizes a mean square error of a state quantity when a mean square error is used as an evaluation function as the state estimation theory.
5. The error estimation device according to any one of claims 1 to 4, The acquisition unit corrects the acquired second physical quantity by using the observation error estimated by the error estimating unit.
6. The error estimation device according to any one of claims 1 to 4, further comprising: and a control unit that detects an abnormality in the acquisition unit by using the observation error estimated by the error estimating unit.
7. 1. An aftertreatment system comprising: An error estimation device according to any one of claims 1 to 4; A catalyst that is provided in a main flow passage through which the exhaust gas flows and purifies harmful substances in the exhaust gas; a supply unit that supplies an additive for purifying harmful substances to the catalyst; A supply control unit that controls the supply of the additive in the supply unit; An aftertreatment system comprising:
8. An error estimation method, comprising: acquiring, by using a sensor, a first physical quantity of a predetermined component in exhaust gas from an exhaust gas purification device including a main flow passage through which exhaust gas flows and a catalyst that purifies harmful substances in the exhaust gas; acquiring a second physical quantity for the predetermined component in the exhaust gas using an acquisition unit; a state estimating step of estimating a state quantity that represents an internal state of the exhaust purification device and that affects the first physical quantity acquired by the sensor; an error estimating step of estimating an observation error in the acquisition unit from the estimated state quantity and the acquired first physical quantity using a state estimation theory; Run In the error estimation step, an estimated value estimated by the state estimation step using the estimated state quantity; a calculated value calculated from the acquired first physical quantity, the acquired second physical quantity, and the measurement error estimated in the error estimating step; The error estimation method estimates the measurement error using a constraint that the estimated value and the calculated value are equal for
9. A computer program comprising: a function of acquiring a first physical quantity of a predetermined component in exhaust gas by using a sensor, the first physical quantity being obtained from an exhaust gas purification device that includes a main flow path through which exhaust gas flows and a catalyst that purifies harmful substances in the exhaust gas; a function of acquiring a second physical quantity for the predetermined component in the exhaust gas using an acquisition unit; a state estimation function that estimates a state quantity that represents an internal state of the exhaust purification device and that affects the first physical quantity acquired by the sensor; an error estimation function that estimates an observation error in the acquisition unit from the estimated state quantity and the acquired first physical quantity by using a state estimation theory; Run the command, The error estimation function an estimated value estimated by the state estimation function using the estimated state quantity; a calculated value calculated from the acquired first physical quantity, the acquired second physical quantity, and the observation error estimated by the error estimation function; and estimating the measurement error using a constraint that the estimated value and the calculated value are equal for
Citation Information
Patent Citations
Conjecture sensor that can be constituted for vehicle control system
JP2016089837A
Control device of engine
JP2018178870A
Control device for post-treatment system, post-treatment system, method for controlling post-treatment system, and computer program
JP2022075531A
Model-based monitoring for selective catalytic reduction device in aftertreatment assembly
US20190024563A1
Diagnostic device of exhaust gas post-treatment device for internal combustion engine
JP2004308455A
