JP7279687B2 - Catalyst state estimation device, method for estimating catalyst state, and computer program - Google Patents

Description

The present invention relates to technology for estimating the state of a catalyst.

Techniques for quantitatively predicting phenomena under various conditions using mathematical models are known. As one of such mathematical models, a neural network (NN) is known, in which a neural network in the human brain is represented by a mathematical model called an artificial neuron.

For example, in Patent Document 1, ambient temperature, manifold pressure and temperature, fuel consumption rate, and engine speed are input to NN, and the selective reduction catalyst (SCR catalyst: Selective Catalytic Reduction catalyst) is discharged from Predicting the amount of nitrogen oxides (NOx) is described. For example, in Patent Document 2, each value of the EGR valve lift amount command value, boost pressure, intake air temperature, exhaust pressure, fuel injection amount, intake air flow rate, and engine speed is input to the NN, and NOx absorption reduction is performed. Predicting the amount of NOx trapped in a catalyst (NSR catalyst: NOx Storage Reduction catalyst) is described. For example, in Patent Document 3, each value of engine speed, fuel injection amount, fuel injection timing, intake air amount, air-fuel ratio, exhaust temperature, boost pressure is input to NN, and NOx emitted from the engine is calculated. It is described to predict the amount of

Japanese Patent Application Laid-Open No. 2003-328732 JP 2009-180086 A JP 2011-132915 A

However, in each of the techniques described in Patent Documents 1 to 3, only parameters related to the internal combustion engine (engine) and its intake system are assumed as inputs to the NN. Here, the purification performance of the catalyst fluctuates depending on the temperature of the catalyst and the amount of harmful substances stored in the catalyst. , there is a problem that the purification performance of the catalyst cannot be estimated with high accuracy. In addition, the techniques described in Patent Documents 1 to 3 have a problem that it is impossible to estimate the purification performance of catalysts for purifying exhaust from sources other than internal combustion engines, such as catalysts for purifying exhaust from industrial facilities. rice field.

Furthermore, in the techniques described in Patent Documents 1 to 3, fuel consumption is reduced in reduction control that consumes fuel and reduces harmful substances stored in the catalyst, such as rich spike control in an NSR catalyst. No consideration is given to reducing Furthermore, in the techniques described in Patent Documents 1 to 3, in a configuration in which a plurality of catalysts are used to purify exhaust gas (for example, a configuration in which two or more NSR catalysts are provided), the purification performance of the plurality of catalysts is estimated. nothing is considered about it.

SUMMARY OF THE INVENTION The present invention has been made to solve at least a part of the above-described problems, and improves the accuracy of estimation in a technology for estimating the purification performance of a catalyst that purifies exhaust gas by absorbing harmful substances. for the purpose.

The present invention has been made to solve at least part of the above problems, and can be implemented as the following modes. A catalyst state estimating device, which is a catalyst provided in a main flow path through which exhaust gas flows, and which acquires information about a catalyst that absorbs and purifies harmful substances in the exhaust gas; and a catalyst state. a storage unit that stores an estimation model in advance; and an estimation unit that estimates the purification performance of the catalyst by applying the catalyst information acquired by the first acquisition unit to the catalyst state estimation model, The catalyst state estimation model includes a first model for estimating a purification rate of the harmful substance in the catalyst, and an estimated reduction amount of the harmful substance in reduction control for reducing the harmful substance occluded in the catalyst. a second model, and a fourth model for estimating the amount of additives that do not contribute to the reduction reaction in the reduction control, wherein the catalyst state estimation model is the current hazardous substance estimated by the first model. The purification rate, the current reduction amount of the hazardous substance estimated by the second model, and the current amount of the additive that does not contribute to the reduction reaction estimated by the fourth model are input. is a model that uses physical laws to obtain the amount of the toxic substance stored in the catalyst at the time of , and the estimating unit uses the estimation result of the fourth model as an input to the second model to determine the amount of the toxic substance By using it to estimate the reduction amount, it is used as an indirect input to the catalyst state estimation model , and the estimation result by the second model is used as a direct input to the catalyst state estimation model, and the catalyst state A catalyst state estimation device for estimating the storage amount of the harmful substance at the next time as the purification performance of the catalyst using an estimation model. In addition, the present invention can also be implemented as the following modes.

(1) According to one aspect of the present invention, a catalyst state estimation device is provided. This catalyst state estimating device includes a first acquisition unit that acquires information about a catalyst that is provided in a main flow path through which exhaust gas flows and that purifies harmful substances in the exhaust gas by absorbing them; a storage unit that stores a model in advance; and an estimation unit that estimates purification performance of the catalyst by applying the information of the catalyst acquired by the first acquisition unit to the catalyst state estimation model, The catalyst state estimation model includes a first model for estimating the purification rate of the harmful substance in the catalyst, and a second model for estimating the reduction amount of the harmful substance in reduction control for reducing the harmful substance occluded in the catalyst. 2 model, a third model for estimating the purification amount of the harmful substance purified by the gas phase reaction of the catalyst, and a fourth model for estimating the amount of additive that does not contribute to the reduction reaction in the reduction control. including at least one or more mathematical models of

The purification performance of the catalyst fluctuates under the influence of catalyst information (for example, the temperature of the catalyst and the amount of harmful substances stored in the catalyst). According to this configuration, the estimating unit can accurately estimate the purification performance of the catalyst by applying information about the catalyst that affects the purification performance of the catalyst to the catalyst state estimation model. That is, according to this configuration, it is possible to improve the accuracy of estimation in the technique of estimating the purification performance of the catalyst that purifies exhaust gas by absorbing harmful substances. Further, according to this configuration, by using the first model, it is possible to estimate the purification rate of the harmful substances in the catalyst, and by using the second model, it is possible to reduce the harmful substances occluded in the catalyst. The amount of reduction can be estimated, and by using the third model, the amount of harmful substances purified by the gas phase reaction of the catalyst can be estimated. The amount of additive that does not contribute to the reaction can be estimated. Thus, according to the catalyst state estimating device of this configuration, various purification performances of the catalyst can be estimated by using different mathematical models.

(2) The catalyst state estimating device of the above aspect further includes a second acquiring section that acquires information on the exhaust gas flowing into the catalyst, and the estimating section further acquires the information on the catalyst in addition to the information on the catalyst. 2 The purification performance of the catalyst may be estimated by applying the exhaust information acquired by the acquisition unit to the catalyst state estimation model.
The purification performance of the catalyst fluctuates under the influence of not only the information of the catalyst but also the information of the exhaust flowing into the catalyst (for example, the temperature of the exhaust, the flow rate of the exhaust, and the amount of harmful substances in the exhaust). According to this configuration, the estimating section applies both the information about the catalyst that affects the purification performance of the catalyst and the information about the exhaust gas flowing into the catalyst to the catalyst state estimation model, thereby estimating the purification performance of the catalyst. Furthermore, it can estimate with high precision.

(3) The catalyst state estimating device of the above aspect further includes a third acquiring unit that acquires information about the additive supplied to the catalyst, and the estimating unit further includes, in addition to the information about the catalyst, Purification performance of the catalyst may be estimated by applying the additive information acquired by the third acquisition unit to the catalyst state estimation model.
According to this configuration, the estimating unit estimates the purification performance of the catalyst by applying the additive information (for example, the amount of additive) in addition to the catalyst information to the catalyst state estimation model. For this reason, for example, when estimating the purification performance of a catalyst that purifies harmful substances using additives, such as a storage reduction catalyst (NSR catalyst: NOx Storage Reduction catalyst), it is possible to further improve the accuracy of estimation. can.

(4) In the catalyst state estimating device of the above aspect, the first model is configured by a machine learning model, and the information on the catalyst includes the temperature of the front end of the catalyst in the main flow path, the temperature of the catalyst, and 1 inputting at least one of a storage amount of nitrogen oxides as the harmful substance stored in the catalyst before the time and a ratio of the storage amount to the saturated storage amount of the nitrogen oxides in the catalyst; , the nitrogen oxide purification rate of the catalyst may be used as an output, and the estimation unit may estimate the nitrogen oxide purification rate as the purification performance of the catalyst using the first model.
As long as there is a causal relationship between input and output variables, machine learning models can be used when classification or regression is not possible without complex function approximation (for example, when phenomena that are difficult to express with physical formulas are involved, phenomena can be described). (When multiple physical formulas are required, or when there are many inputs to the physical formula due to the assumption of the influence of many factors), the output (estimation result) can be obtained with a low computational load. According to this configuration, the estimation unit uses the first model configured by the machine learning model to estimate the purification rate of nitrogen oxides as harmful substances. Rate estimation can be performed at low computational load and at high speed. Further, by using a sufficiently learned machine learning model as the first model, the estimation unit can estimate the nitrogen oxide purification rate with high accuracy.

(5) In the catalyst state estimating device of the above aspect, the second model is configured by a machine learning model, and the information on the catalyst includes the temperature of the front end of the catalyst in the main flow path, the temperature of the catalyst, and 1 At least one of an amount of nitrogen oxides as the harmful substance occluded by the catalyst before the time is input, and an amount of reduction of the nitrogen oxides occluded by the catalyst is output. The estimation unit may estimate the amount of reduction of the nitrogen oxides as purification performance of the catalyst using the second model.
According to this configuration, the estimation unit uses the second model configured by the machine learning model to estimate the reduction amount of nitrogen oxides as harmful substances in the reduction control, so that more factors affect The reduction amount of nitrogen oxides can be estimated at high speed with a low computational load. Further, by using a sufficiently learned machine learning model as the second model, the estimation unit can estimate the reduction amount of nitrogen oxides with high accuracy.

(6) In the catalyst state estimating device of the above aspect, the third model is configured by a machine learning model, and the information on the catalyst includes the temperature of the front end of the catalyst in the main flow path and the temperature of the catalyst. with at least one of them as an input, and with the amount of nitrogen oxides removed as the harmful substance purified by the gas-phase reaction of the catalyst as an output, the estimation unit uses the third model to determine the amount of the catalyst A purification amount of the nitrogen oxides may be estimated as the purification performance.
According to this configuration, the estimating unit uses the third model configured by the machine learning model to estimate the removal amount of nitrogen oxides as harmful substances that are removed by the gas phase reaction of the catalyst. It is possible to estimate the amount of nitrogen oxides to be purified, which is affected by the factors of (1), at low computational load and at high speed. Further, by using a sufficiently learned machine learning model as the third model, the estimation unit can estimate the amount of nitrogen oxides to be purified with high accuracy.

(7) In the catalyst state estimating device of the above aspect, the fourth model is configured by a physical model, and the catalyst information includes the temperature of the front end of the catalyst in the main flow path and the temperature of the catalyst. and the amount of the additive that does not contribute to the reduction reaction in the reduction control is used as the output, and the estimation unit uses the fourth model to calculate the reduction reaction as the purification performance of the catalyst. may estimate the amount of said additive that does not contribute to
Since the physical model is based on the actual physical law and is based on the law of similarity, the output (estimation result) obtained by the physical model satisfies the physical law. On the other hand, since the machine learning model is a model constructed as a result of learning a huge amount of data, it may not be possible to obtain an output (estimation result) that satisfies the laws of physics. According to this configuration, the estimating unit uses the fourth model configured by the physical model to estimate the amount of the additive that does not contribute to the reduction reaction in the reduction control. can be estimated with high accuracy in accordance with the laws of physics.

(8) In the catalyst state estimating device of the above aspect, the catalyst state estimating model includes a current purification rate of the hazardous substance estimated by the first model and a current rate of the hazardous substance estimated by the second model. , the current purification amount of the harmful substance estimated by the third model, and the current amount of the additive that does not contribute to the reduction reaction estimated by the fourth model, It is a model that uses physical laws to determine the amount of the harmful substance stored in the catalyst at the next time. The storage amount of the harmful substance may be estimated.
According to this configuration, the estimating unit calculates the estimation results (purification rate of nitrogen oxides, amount of reduction of nitrogen oxides as harmful substances in reduction control, Purification amount of nitrogen oxides as harmful substances purified by gas phase reaction) and the estimation result by the physical model 4 model (amount of additive that does not contribute to reduction reaction in reduction control) to estimate the amount of nitrogen oxides stored in the catalyst. Therefore, according to this configuration, the estimation of the nitrogen oxide storage amount, which is influenced by a greater number of factors, can be performed with high accuracy and at high speed while satisfying the laws of physics. In addition, the amount of nitrogen oxides stored in the catalyst fluctuates under the influence of the previous storage amount (in other words, fluctuates under the influence of the time history). According to this configuration, the estimation unit uses the current estimation results from the first to fourth models to estimate the storage amount at the next time. The purification performance of the catalyst at the time can be estimated with high accuracy.

(9) In the catalyst state estimating device of the above aspect, at least one of the first model and the third model further includes, as information on the exhaust gas, the temperature of the exhaust gas, the flow rate of the exhaust gas, and the temperature of the exhaust gas. and at least one of the amount of the harmful substance contained, and the second model further receives at least one of the temperature of the exhaust gas and the flow rate of the exhaust gas as information on the exhaust gas. and at least one of an inflow amount of the additive flowing into the catalyst and an amount of the additive not contributing to the reduction reaction is input as the additive information, and the second The four models further input at least one of the temperature of the exhaust gas and the air-fuel ratio of the exhaust gas as information on the exhaust gas, and the additive flowing into the catalyst as information on the additive. may be used as an input.
According to this configuration, in the first to fourth models, estimation is performed in consideration of various kinds of information, so the purification performance of the catalyst can be estimated with higher accuracy.

(10) In the catalyst state estimating device of the above aspect, at least some of the first model, the second model, and the third model use teacher data obtained from a plurality of catalysts with different degrees of deterioration. may be created using
According to this configuration, at least one of the first model, the second model, and the third model is created using teacher data obtained from a plurality of catalysts with different degrees of deterioration. A first model, a second model, and a third model that are optimal together can be adopted. As a result, according to this configuration, the purification performance of the catalyst can be estimated with higher accuracy.

(11) According to one aspect of the present invention, there is provided a method for estimating the state of a catalyst. According to this configuration, the accuracy of estimation can be improved in the technique of estimating the purification performance of the catalyst.

(12) According to one aspect of the invention, there is provided a computer program for estimating the state of a catalyst. According to this configuration, the accuracy of estimation can be improved in the technique of estimating the purification performance of the catalyst.

(13) In the catalyst state estimating device of the above aspect, the catalyst state estimating model is an outflow additive that is the amount of the additive that flows out of the catalyst without being consumed in the reduction control, instead of the second model. A fifth model for estimating the quantity may be included.
According to this configuration, by using the fifth model, it is possible to estimate the amount of additive that flows out of the catalyst without being consumed in the reduction control (flowing additive amount). Then, the catalyst state estimating device performs feedback control to the reduction control unit that performs reduction control, and the reduction control unit performs reduction control so as to reduce the amount of outflow additive, thereby purifying harmful substances in the catalyst. The amount of extra additives can be reduced while maintaining performance. As a result, according to this configuration, it is possible to reduce the amount of fuel consumption in the reduction control that consumes the fuel and reduces the harmful substances occluded in the catalyst.

(14) In the catalyst state estimating device of the above aspect, the fifth model is configured by a machine learning model, and the information on the catalyst includes the temperature of the front end of the catalyst in the main flow path, the temperature of the catalyst, and 1 at least one of an amount of nitrogen oxides as the harmful substance occluded in the catalyst before the time and an output of the outflow additive amount; A model may be used to estimate the outflow additive amount as a purification performance of the catalyst.
According to this configuration, the estimating unit estimates the outflow additive amount using the fifth model configured by the machine learning model. Can be implemented at high speed. Further, by using a sufficiently learned machine learning model as the fifth model, the estimation unit can estimate the outflow additive amount with high accuracy.

(15) The catalyst state estimating device of the above aspect further includes a second acquisition section that acquires information about the exhaust gas flowing into the catalyst, and the estimation section further includes the second By applying the exhaust information acquired by the second acquisition unit to the catalyst state estimation model, the purification performance of the catalyst is estimated, and the fifth model further calculates the exhaust temperature as the exhaust information. and the flow rate of the exhaust gas may be used as the input.
According to this configuration, the estimating unit applies both the information about the catalyst that affects the purification performance of the catalyst and the information about the exhaust gas flowing into the catalyst to the catalyst state estimation model (fifth model), so that the catalyst purification performance can be estimated with higher accuracy.

(16) The catalyst state estimating device of the above aspect further includes a third acquiring unit that acquires information about the additive supplied to the catalyst, and the estimating unit further includes, in addition to the information about the catalyst, The purification performance of the catalyst is estimated by applying the additive information acquired by the third acquisition unit to the catalyst state estimation model, and the fifth model further obtains, as the additive information, At least one of an inflow amount of the additive that flows into the catalyst and an amount of the additive that does not contribute to the reduction reaction estimated by the fourth model may be used as an input.
According to this configuration, the estimating unit applies both the information about the catalyst that affects the purification performance of the catalyst and the information about the additive to the catalyst state estimation model (fifth model) to obtain the purification performance of the catalyst. can be estimated more accurately.

(17) In the catalyst state estimating device of the above aspect, the catalyst state estimating model includes the current purification rate of the harmful substance estimated by the first model and the current outflow addition rate estimated by the fifth model. The reduction amount of the harmful substance in the reduction control obtained from the agent amount, the current purification amount of the harmful substance estimated by the third model, and the current reduction reaction estimated by the fourth model and the amount of the additive that does not contribute, and the amount of the harmful substance stored in the catalyst at the next time is obtained by using the physical law, and the estimation unit uses the catalyst state estimation model. Then, the storage amount of the harmful substance at the next time may be estimated as the purification performance of the catalyst.
According to this configuration, the estimator can obtain the reduction amount of nitrogen oxides as harmful substances in the reduction control from the estimation result (outflow additive amount) by the fifth model, which is a machine learning model. In addition, the estimating unit calculates the reduction amount of harmful substances obtained, and the estimation results by the first and third models, which are machine learning models (purification rate of nitrogen oxides, nitrogen oxides as harmful substances purified by gas phase reaction). Purification amount of substances) and the estimation result by the physical model 4 model (the amount of additives that do not contribute to the reduction reaction in reduction control) are used together to calculate the amount of nitrogen oxides stored in the catalyst using the laws of physics. to estimate Therefore, according to this configuration, the estimation of the nitrogen oxide storage amount, which is influenced by a greater number of factors, can be performed with high accuracy and at high speed while satisfying the laws of physics. Further, the amount of nitrogen oxides stored in the catalyst fluctuates under the influence of the amount of storage at the previous time. According to this configuration, the estimation unit uses the current estimation results from the first, third to fifth models to estimate the storage amount at the next time, based on the purification performance of the catalyst at the previous time. , the purification performance of the catalyst at the next time can be estimated with high accuracy.

(18) In the catalyst state estimating device of the above aspect, the fifth model may be created using teacher data obtained from a plurality of catalysts having different degrees of deterioration.
According to this configuration, since the fifth model is created using teacher data obtained from a plurality of catalysts with different degrees of deterioration, it is possible to adopt an optimum fifth model according to the degree of deterioration of the catalyst. can. As a result, according to this configuration, the purification performance of the catalyst can be estimated with higher accuracy.

(19) In the catalyst state estimating device of the above aspect, when a plurality of catalysts of the same type are provided in the main flow path, the first acquisition unit obtains information about the catalyst located most upstream in the main flow path. and the estimation unit applies the information of the catalyst acquired by the first acquisition unit to the catalyst state estimation model to estimate the purification performance of the plurality of catalysts as a whole, and the catalyst Among the state estimation models, the first model is a model for estimating the purification rate of the harmful substance as a whole of the plurality of catalysts, and the second model is a model for the reduction control for each of the plurality of catalysts, the a model for estimating the total amount of reduction of harmful substances, wherein the third model is a model for estimating the total amount of purification of the harmful substances in the gas phase reaction of each of the plurality of catalysts; The four models may be models for estimating the total amount of additives that do not contribute to the reduction reaction in the reduction control for each of the plurality of catalysts.
According to this configuration, the purification performance can be estimated by regarding a plurality of catalysts in the main flow path as one catalyst. etc.) and the number of catalyst state estimation models prepared in advance can be reduced, and the computational load in the catalyst state estimation device can be reduced. Further, when the plurality of catalysts provided in the main flow path are of the same type, the estimation unit estimates the purification performance by regarding them as one catalyst. If the catalysts are of the same type, there is no difference in the information items of the catalysts that affect the purification performance, so the estimating unit can accurately estimate the purification performance.

(20) In the catalyst state estimating device of the above aspect, when a plurality of catalysts of the same type are provided in the main flow path, the first acquisition unit obtains information about the catalyst positioned most upstream in the main flow path. and the estimation unit applies the information of the catalyst acquired by the first acquisition unit to the catalyst state estimation model to estimate the purification performance of the plurality of catalysts as a whole, and the catalyst Among the state estimation models, the first model is a model for estimating the purification rate of the harmful substance as a whole of the plurality of catalysts, and the fifth model is the most downstream model without being consumed in the reduction control. A model for estimating an outflow additive amount, which is the amount of the additive flowing out from the positioned catalyst. The fourth model may be a model for estimating the total amount of additives that do not contribute to the reduction reaction in the reduction control for each of the plurality of catalysts.
According to this configuration, the purification performance can be estimated by regarding a plurality of catalysts in the main flow path as one catalyst. etc.) and the number of catalyst state estimation models prepared in advance can be reduced, and the computational load in the catalyst state estimation device can be reduced. In addition, by using the fifth model, it is possible to estimate the amount of additive that flows out from the catalyst without being consumed in reduction control (the amount of additive that flows out). while reducing the amount of excess additives and reducing fuel consumption.

It should be noted that the present invention can be implemented in various aspects, including, for example, a catalyst state estimation device and system, an exhaust purification device and exhaust purification system including the catalyst state estimation device, a control method for these devices and systems, and a method for controlling these devices and systems. It can be implemented in the form of a computer program executed in a device or system, a server device for distributing the computer program, a non-temporary storage medium storing the computer program, or the like.

1 is a block diagram of an exhaust purification system as one embodiment of the present invention; FIG. It is a figure explaining a 1st model. It is a figure explaining a 4th model. 4 is a flowchart showing a procedure of estimation processing by an estimation unit; It is a figure explaining step S24 of an estimation process. It is a block diagram of an exhaust gas purification system in a second embodiment. FIG. 11 is a block diagram of an exhaust purification system in a third embodiment; FIG. FIG. 11 is a block diagram of an exhaust purification system in a fourth embodiment; FIG. FIG. 14 is a flowchart showing a procedure of estimation processing by an estimation unit according to the fourth embodiment; FIG. FIG. 12 is a block diagram of an exhaust purification system in a fifth embodiment; FIG. FIG. 16 is a flow chart showing a procedure of estimation processing by an estimation unit according to the fifth embodiment; FIG. FIG. 11 is a block diagram of an exhaust purification system in a sixth embodiment; FIG. 11 is a block diagram of an exhaust gas purification system of a seventh embodiment;

<First embodiment>
FIG. 1 is a block diagram of an exhaust purification system 1 as one embodiment of the present invention. The exhaust purification system 1 of this embodiment includes a combustion state control section 91 , an internal combustion engine 92 , an exhaust purification device 20 , and a catalyst state estimation device 10 . The exhaust purification device 20 of the present embodiment is a device that purifies harmful substances (nitrogen oxides: NOx) in the exhaust gas of the internal combustion engine 92, and uses a NOx storage reduction catalyst (NSR catalyst: NOx Storage Reduction catalyst) 70 as a catalyst. Prepare. The NSR catalyst 70 purifies NOx in the exhaust gas by absorbing and storing NOx emitted from the internal combustion engine 92 mainly during lean combustion. The catalyst state estimating device 10 of the present embodiment can estimate the state of the NSR catalyst 70 mounted in the exhaust purification device 20, specifically, the purification performance of the NSR catalyst 70. FIG.

The internal combustion engine 92 is, for example, a lean burn gasoline engine or a diesel engine. The combustion state control unit 91 controls the air-fuel ratio in the internal combustion engine 92 to be lean, stoichiometric, or rich by controlling injection of air or fuel into the internal combustion engine 92 . The combustion state control section 91 is implemented by, for example, an electronic control unit (ECU). In the following description, the side of the exhaust purification device 20 closer to the internal combustion engine 92 is called "upstream side", and the side farther from the internal combustion engine 92 is called "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 includes an exhaust pipe 30 extending from an internal combustion engine 92 , an NSR catalyst 70 provided on the exhaust pipe 30 , and a rich spike control section 93 . The exhaust pipe 30 forms a main flow path through which exhaust from the internal combustion engine 92 flows. 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 to the outside air. The NSR catalyst 70 purifies NOx in the exhaust by absorbing and storing the NOx in the exhaust in the NSR catalyst 70 . The NSR catalyst 70 corresponds to "catalyst".

A rich spike control unit 93 performs rich spike control for purifying 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 period of time to reduce carbon monoxide (CO), hydrogen molecules (H ₂ ), and hydrocarbons (HC). and other unburned gases are discharged from the internal combustion engine 92 . Then, the emitted CO, H ₂ and other unburned gases reduce the NOx stored in the NSR catalyst 70 to nitrogen gas (N ₂ ). That is, CO, H ₂ , and other unburned gases function as “additives (more specifically, reducing agents)” used for reducing NOx stored in the NSR catalyst 70 . Hereinafter, rich spike control will also be referred to as "reduction control", and CO, _H2 , and other unburned gases will be collectively referred to simply as "additives". Like the combustion state control unit 91, the rich spike control unit 93 can be implemented by an ECU.

Catalyst state estimation device 10 includes CPU 11 , storage unit 12 , flow rate acquisition unit 52 , exhaust temperature acquisition unit 53 , NOx concentration acquisition unit 54 , front end temperature acquisition unit 76 , and temperature acquisition unit 78 . The CPU 11 and the storage unit 12 are implemented by an ECU, for example.

The CPU 11 controls each part of the catalyst state estimating device 10 by loading a computer program stored in the ROM into the RAM and executing the program. In addition, the CPU 11 functions as an estimating unit 110, receives each acquired value acquired from the flow rate acquiring unit 52, the exhaust temperature acquiring unit 53, the NOx concentration acquiring unit 54, the front end temperature acquiring unit 76, and the temperature acquiring unit 78, An estimation process, which will be described later, is executed. The storage unit 12 is composed of a flash memory, a memory card, a hard disk, or the like. A catalyst state estimation model 120 is stored in advance in the storage unit 12 . The catalyst state estimation model 120 includes a first model 121 , a second model 122 , a third model 123 and a fourth model 124 .

FIG. 2 is a diagram illustrating the first model 121. As shown in FIG. The first model 121 is a model for estimating the NOx purification rate of the NSR catalyst 70, and is composed of a machine learning model, specifically a neural network (NN). As shown in FIG. 2, the first model 121 of this embodiment is composed of three layers: an input layer, an intermediate layer, and an output layer. Illustrate the case. Note that the sigmoid function may not be used for the elements of the intermediate layer.

A vector U, which is an input variable of the first model 121, is composed of n components (u ₁ , u ₂ , . . . , u _n , where n is a natural number). For each component of the vector U, parameters shown in items a1 and a2 below can be adopted.
(a1) Catalyst information: 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 hour before, and the saturated amount of NOx stored in the NSR catalyst 70 and at least one of (a2) exhaust information: the temperature of the exhaust from the internal combustion engine 92, the flow rate of the exhaust, and the amount of NOx contained in the exhaust. at least one of

Here, for the vector U, it is necessary to input at least one or more of the catalyst information shown in item a1, and at least one or more of the exhaust information shown in item a2 is input. may be omitted entirely. However, in order to improve the estimation accuracy of the NOx purification rate in the first model 121, it is preferable that the number of input parameters is large.

After each component of the vector U is input to the input layer, the product of each component u _i of the vector U and the weight constant W _ij is added in the intermediate layer (Equation 1). After that, as shown in Equation 2, the eigenvalue θ _j (bias) of each node in the intermediate layer is taken into consideration, and then passed through the sigmoid function and output. In the output layer, the product of the input value Y _j and the weight constant W _jk is added, and the sum with the eigenvalue θ _k (bias) is obtained and output (Formula 3). The output variable Z of the first model 121 is an estimated value of the NOx purification rate of the NSR catalyst 70 under various conditions represented by the input variable (vector U). Note that n represents the number of input variables, and m represents the number of intermediate layer nodes. The first model 121 corresponds to the "first model".

The first model 121 causes the NN to learn the relationship between the input variables and the output variables so that the output variable Z matches the physical quantity to be estimated (the NOx purification rate in the case of the first model 121). Values W _ij , θ _j , W _jk and θ _k are determined in advance. For the learning of the first model 121, it is preferable to use data acquired during transient operation of the internal combustion engine 92 as teacher data. Note that the number of nodes in the intermediate layer can be determined, for example, in consideration of the accuracy of teacher data and the accuracy of data not used as teacher data.

The second model 122 is based on the reduction amount of NOx stored in the NSR catalyst 70, more specifically, the NOx amount ( This is a model for estimating the reduction amount). The second model 122 is a machine learning model, specifically a NN. The second model 122 has the same configuration as the first model 121 described with reference to FIG. 2 except for points described later.

A vector U, which is an input variable of the second model 122, is composed of n components (u ₁ , u ₂ , . . . , u _n , n is a natural number). For each component of the vector U, parameters shown in items b1 to b3 below can be adopted.
(b1) Catalyst information: at least one of the temperature of the front end of the NSR catalyst 70, the temperature of the NSR catalyst 70, and the NOx storage amount (actual storage amount) stored in the NSR catalyst 70 one hour before One (b2) Exhaust information: At least one of the temperature of the exhaust gas from the internal combustion engine 92 and the flow rate of the exhaust gas (b3) Additive information: Additives (CO, H ₂ , and other unburned gases), and at least one of the amount of additives that do not contribute to the reduction reaction in rich spike control

For the vector U of the second model 122, as with the first model 121, it is necessary to input at least one or more of the catalyst information shown in item b1, and the exhaust gas shown in items b2 and b3. At least one or more of the information and the additive information may be input, or all of them may be omitted. However, in order to improve the estimation accuracy of the outflow amount in the second model 122, it is preferable that the number of input parameters is large. The output variable Z of the second model 122 is an estimated value of the NOx reduction amount reduced by the rich spike control under various conditions represented by the input variable (vector U). Note that the second model 122 corresponds to the "second model".

The third model 123 is a model for estimating the amount of NOx purified by the gas phase reaction of the NSR catalyst 70 . The third model 123 is a machine learning model, specifically a NN. The third model 123 has the same configuration as the first model 121 described with reference to FIG. 2 except for points described later.

A vector U, which is an input variable of the third model 123, consists of n components (u ₁ , u ₂ , . . . , u _n , where n is a natural number). For each component of the vector U, parameters shown in items c1 and c2 below can be adopted.
(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: the temperature of the exhaust from the internal combustion engine 92 and the flow rate of the exhaust , the amount of NOx contained in the exhaust, and

For the vector U of the third model 123, as with the first model 121, it is necessary to input at least one or more of the catalyst information shown in item c1, and the exhaust information shown in item c2. may be input with at least one or more, or may be omitted entirely. However, in order to improve the estimation accuracy of the purification amount in the third model 123, it is preferable that the number of input parameters is large. The output variable Z of the third model 123 is an estimated value of the amount of NOx purified by the gas phase reaction of the NSR catalyst 70 under the conditions represented by the input variable (vector U). Note that the third model 123 corresponds to the "third model".

FIG. 3 is a diagram illustrating the fourth model 124. As shown in FIG. The fourth model 124 is a model for estimating the amount of additives (CO, H ₂ , unburned gas) that do not contribute to the reduction reaction in the rich spike control described above, and is composed of 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 occluded in the NSR catalyst 70 and the additive (unburned gas such as CO, _H2 , HC, etc.) is not sufficient. It may not progress. The fourth model 124 estimates the amount of additives that do not contribute to such a reduction reaction. FIG. 3 shows, as an example of the fourth model 124, the Arrhenius equation for estimating the amount of CO that does not contribute to the reduction reaction. When estimating the amount of _H2 that does not contribute to the reduction reaction, CO in FIG. 3 can be replaced with _H2 , and when estimating the amount of HC that does not contribute to the reduction reaction, CO in FIG. good. In the Arrhenius equation, the frequency factor (A) and the activation energy (E) can be values previously determined by experiments or the like.

A vector U, which is an input variable of the fourth model 124, is composed of n components (u ₁ , u ₂ , . . . , u _n , where n is a natural number). For each component of the vector U, parameters shown in items d1 to d3 below can be adopted.
(d1) 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 (d2) exhaust information: the temperature of the exhaust from the internal combustion engine 92 and the temperature of the NSR catalyst 70 and at least one of (d3) additive information: the amount of additive (CO, _H2 , and other unburned gases) entering the NSR catalyst 70;

As with the first model 121, for the vector U of the fourth model 124, it is necessary to input at least one or more of the catalyst information shown in item d1, and the exhaust gas shown in items d2 and d3. At least one or more of the information and the additive information may be input, or all of them may be omitted. However, in order to improve the estimation accuracy of the amount of additive that does not contribute to the reduction reaction in the fourth model 124, it is preferable that the number of input parameters is large. The output variable Z of the fourth model 124 is an estimate of the amount of additive that does not contribute to the reduction reaction in rich spike control under the conditions represented by the input variable (vector U). Note that the fourth model 124 corresponds to the "fourth model".

Returning to FIG. 1, the description is continued. The flow rate acquisition unit 52 acquires the flow rate of exhaust gas from the internal combustion engine 92 . The flow rate acquisition unit 52 may be implemented by acquiring a measurement signal measured by a pitot tube flow meter provided in the exhaust pipe 30, for example. Further, the flow rate obtaining unit 52 may estimate the flow rate of the exhaust gas from an intake air amount signal to the internal combustion engine 92 or a fuel injection amount signal. The exhaust temperature acquisition unit 53 is a sensor that measures the temperature of the exhaust from the internal combustion engine 92 . The NOx concentration acquisition unit 54 is a sensor that measures the NOx concentration in exhaust gas flowing into the NSR catalyst 70 . Note that the NOx concentration acquisition unit 54 may estimate the NOx concentration in the exhaust from the combustion state (lean, stoichiometric, rich) of the internal combustion engine 92 instead of the sensor measurement. The front end temperature acquisition unit 76 is a sensor that measures the temperature near the inlet (front end) of the NSR catalyst 70 . A temperature acquisition unit 78 is a sensor that measures the bed temperature of the NSR catalyst 70 .

The flow rate acquisition unit 52, the exhaust temperature acquisition unit 53, and the NOx concentration acquisition unit 54 correspond to a “second acquisition unit” that acquires information on the exhaust flowing into the NSR catalyst 70. FIG. The front end temperature acquisition section 76 and the temperature acquisition section 78 correspond to a “first acquisition section” that acquires information on the NSR catalyst 70 .

FIG. 4 is a flowchart showing a procedure of estimation processing by the estimation unit 110. As shown in FIG. The estimation process is a process of estimating the purification performance of the NSR catalyst 70, and is executed at arbitrary timing. For example, the estimation process may be executed in response to a request from a user of the exhaust purification system 1 or the catalyst state estimation device 10, or may be executed in response to a request from another control unit in the vehicle equipped with the exhaust purification system 1. good. The estimation process shown in FIG. 4 is periodically executed.

In the following description, except for the temperature at the front end of the NSR catalyst 70 and the air-fuel ratio of the exhaust gas flowing into the NSR catalyst 70, items a1, a2, items b1-b3, items c1, c2, A case where all the parameters described in d3 are adopted is illustrated. Also, in the following description, Δt represents a unit time in the catalyst state estimation device 10 (for example, the calculation cycle of the CPU 11, the sampling cycle in the acquisition units 52 to 78 described above). Therefore, the time t=kΔt means the current time, the time t=(k+1)Δt means one hour later (one unit time later), and the time t=(k−1)Δt means one hour earlier (1 unit time ago). k is an integer.

In step S10, the estimation unit 110 determines whether or not a condition for starting the estimation process is satisfied. Specifically, for example, the estimation unit 110 performs the estimation process when the temperature acquisition unit 78 is normal, the rich spike control unit 93 is normal, and the NOx concentration acquisition unit 54 is in an active state. It can be determined that the start condition is satisfied. If the condition for starting the estimation process is satisfied (step S10: YES), the estimation unit 110 shifts the process to step S12. If the conditions for starting the estimation process are not satisfied (step S10: NO), the estimation unit 110 terminates the process.

In step S<b>12 , the estimation unit 110 acquires the current flow rate Q[k] inside the exhaust pipe 30 (time t=kΔt) from the flow rate acquisition unit 52 . In step S14, the estimation unit 110 acquires the NOx amount NOx_NSR _in flowing into the NSR catalyst 70 at present (time t=kΔt). Specifically, the estimation unit 110 acquires the current NOx concentration [k] in the exhaust flowing into the NSR catalyst 70 from the NOx concentration acquisition unit 54 . Next, the estimation unit 110 calculates the NOx amount NOx_NSR _in [k] from the obtained NOx concentration [k] in the exhaust gas and the flow rate Q[k] of the exhaust gas obtained in step S12.

In step S<b>16 , the estimation unit 110 acquires the bed temperature T[k] of the NSR catalyst 70 at present (time t=kΔt) from the temperature acquisition unit 78 . In step S18, the estimating unit 110 calculates the amount of additive flowing into the NSR catalyst 70 at the present (time t=kΔt) from the amount of fuel injected by the rich spike control using a calculation formula, map, etc. prepared in advance. Get the quantity RDCT_NSR _in [k] of . Note that, in step S18, the estimation unit 110 also functions as a "third acquisition unit" that acquires additive information.

In step S<b>20 , the estimation unit 110 acquires the current (time t=kΔt) temperature T_NSR _in [k] of the exhaust gas flowing into the NSR catalyst 70 from the exhaust temperature acquisition unit 53 . In step S20, the estimating unit 110 may acquire the temperature near the entrance (front end) of the NSR catalyst 70 instead of the temperature of the exhaust gas, and use it as T_NSR _in [k].

FIG. 5 is a diagram illustrating step S24 of the estimation process. In FIG. 5 , the horizontal axis represents the bed temperature of the NSR catalyst 70 and the vertical axis represents the saturated NOx storage amount in the NSR catalyst 70 . In the NSR catalyst 70, the saturated storage amount of NOx increases as the bed temperature rises up to a predetermined bed temperature _Tx , as shown in the figure, and reaches the maximum storage amount _Smax at the predetermined bed temperature _Tx . It has the characteristic NP that it decreases as . A relational expression representing such a characteristic NP is obtained in advance by experiments or the like and stored in the storage unit 12 .

In step S<b>24 of FIG. 4 , the estimation unit 110 calculates the ratio of the actual NOx storage amount to the NOx saturated storage amount in the NSR catalyst 70 . Specifically, the estimation unit 110 applies the current bed temperature T[k] of the NSR catalyst 70 acquired in step S16 to the relational expression representing the characteristic NP, and calculates the current (time t=kΔt) NSR catalyst 70 NOx saturated storage amount StrtNOxSt [k] is obtained (FIG. 5). Next, using the obtained saturated storage amount StrtNOxSt[k] and the actual NOx storage amount NOxSt[k] of the NSR catalyst 70 at the present time (time t=kΔt), the ratio r Calculate [k].

For the current (time t=kΔt) actual NOx storage amount NOxSt[k] of the NSR catalyst 70, the estimation unit 110 converts the estimated value obtained in the previously executed estimation process (FIG. 4) into the actual storage amount NOxSt It can be used as [k]. For example, when there is no previous estimated value, such as when the estimation process is executed for the first time, the estimation unit 110 can use a predetermined default value as the actual storage amount NOxSt[k]. Although the default value can be set arbitrarily, for example, 0 or the saturated storage amount can be used. Note that, in step S24, the estimation unit 110 also functions as a "first acquisition unit" that acquires information about the catalyst.

In step S<b>26 , the estimation unit 110 calculates an estimated value of the NOx purification rate of the NSR catalyst 70 . Specifically, estimating unit 110 adds the following current (time t=kΔt) obtained in steps S12 to S24 to input variable vector U of first model 121 (FIG. 2) for estimating the NOx purification rate. Set each value and obtain the output variable Z. The estimation unit 110 sets the output variable Z obtained from the first model 121 as the NOx purification rate NOxConv[k] of the NSR catalyst 70 at present (time t=kΔt).
(a1) catalyst information:
・Temperature (floor temperature) T[k] of the NSR catalyst 70 (step S16)
・NOx storage amount (actual storage amount) NOxSt [k] stored in the NSR catalyst 70 one time before (result of the previous estimation process)
The ratio r[k] of the NOx storage amount (actual storage amount) to the NOx saturated storage amount in the NSR catalyst 70 (step S24)
(a2) Exhaust information:
・Temperature T_NSR _in [k] of exhaust gas from internal combustion engine 92 (step S20)
・Exhaust flow rate Q[k] (step S12)
Amount of NOx contained in exhaust NOx_NSR _in [k] (step S14)

Thus, in step S26, the estimation unit 110 uses the first model 121 to estimate the NOx purification rate NOxConv[k] as the purification performance of the NSR catalyst 70. FIG. A machine learning model such as the first model 121 can classify or regress without complex function approximation as long as there is a causal relationship between input and output variables (for example, the vector U of the input variable and the Z of the output variable). (For example, if a phenomenon that is difficult to express with a physical formula is included, if multiple physical formulas are required to describe the phenomenon, or if there are many inputs to the physical formula because the influence of many factors is assumed. ), an output (estimation result) can be obtained with a low computational load. According to step S26, the estimating unit 110 uses the first model 121 configured by the machine learning model to estimate the NOx purification rate NOxConv[k]. can be performed at high speed with low computational load. Further, by using a sufficiently learned machine learning model as the first model 121, the estimation unit 110 can estimate the NOx purification rate with high accuracy. In step S26, the estimation unit 110 also functions as a "first acquisition unit" that acquires catalyst information.

In step S<b>27 , the estimation unit 110 calculates the amount of NOx purified by the gas phase reaction of the NSR catalyst 70 . Specifically, the estimating unit 110 adds the input variable vector U of the third model 123 (FIG. 2) for estimating the NOx purification amount to the current (time t=kΔt) obtained in steps S12 to S24. , and obtain the output variable Z. The estimation unit 110 sets the output variable Z obtained from the third model 123 as the NOx purification amount NOx_Gas[k] of the NSR catalyst 70 at present (time t=kΔt).
(c1) catalyst information:
・Temperature (floor temperature) T[k] of the NSR catalyst 70 (step S16)
(c2) exhaust information:
・Temperature T_NSR _in [k] of exhaust gas from internal combustion engine 92 (step S20)
・Exhaust flow rate Q[k] (step S12)
Amount of NOx contained in exhaust NOx_NSR _in [k] (step S14)

Thus, in step S27, the estimation unit 110 uses the third model 123 to estimate the NOx purification amount NOx_Gas[k] purified by the gas phase reaction of the NSR catalyst 70 as the purification performance of the NSR catalyst 70. . As with the first model 121 described above, the third model 123 is composed of a machine learning model, so the estimation of the amount of NOx purified by the gas phase reaction, which is affected by a larger number of factors, can be estimated with a low calculation. It can be carried out under load and at high speed. Further, by using a sufficiently learned machine learning model as the third model 123, the estimation unit 110 can estimate the NOx purification amount with high accuracy. In step S27, the estimation unit 110 also functions as a "first acquisition unit" that acquires information on the catalyst.

In step S28, estimation unit 110 calculates an estimated value of the amount of additives (CO, _H2 , unburned gas) that do not contribute to the reduction reaction in rich spike control. Specifically, the estimating unit 110 adds the input variable vector U of the fourth model 124 for estimating the amount of additive that does not contribute to the reduction reaction to the current (time t=kΔt) calculated in steps S16 to S20. Set the following values to obtain the output variable Z. The estimating unit 110 sets the output variable Z obtained from the fourth model 124 to the current (time t=kΔt) additive amount Ad_thrmlytc[k] that does not contribute to the reduction reaction.
(d1) catalyst information:
・Temperature (floor temperature) T[k] of the NSR catalyst 70 (step S16)
(d2) exhaust information:
・Temperature T_NSR _in [k] of exhaust gas from internal combustion engine 92 (step S20)
(d3) Additive information:
Amount of additive flowing into NSR catalyst 70 RDCT_NSR _in [k] (step S18)

Thus, in step S28, the estimating unit 110 uses the fourth model 124 to estimate the additive amount Ad_thrmlytc[k] that does not contribute to the reduction reaction as the purification performance of the NSR catalyst 70. FIG. A physical model such as the fourth model 124 is a law based on the law of similarity based on the actual physical law. Therefore, the output (estimation result) obtained by the fourth model 124 (physical model) satisfies the laws of physics. On the other hand, since the first model 121, the second model 122, and the third model 123 (machine learning models) are models constructed as a result of learning a large amount of data, outputs (estimation results) that satisfy the laws of physics are obtained. may not be available. In this regard, according to step S28, the estimation unit 110 uses the fourth model 124 configured by a physical model to estimate the amount of the additive that does not contribute to the reduction reaction in the rich spike control (reduction control). , the amount of the additive, which has few influencing factors, can be estimated with high accuracy in accordance with the laws of physics. Note that, in step S28, the estimation unit 110 also functions as a "third acquisition unit" that acquires additive information.

In step S30, the estimating unit 110 calculates an estimated value of the amount of NOx (reduction amount) of the NOx stored in the NSR catalyst 70 that is reduced in the rich spike control. Specifically, the estimating unit 110 stores the input variable vector U of the second model 122 for estimating the NOx reduction amount at the current time (time t=kΔt) obtained in steps S12 to S20 and S28. Set the value and find the output variable Z. The estimating unit 110 sets the output variable Z obtained from the second model 122 as the current (time t=kΔt) NOx reduction amount NOx_rdct[k] of NOx reduced by the rich spike control.
(b1) catalyst information:
・Temperature (floor temperature) T[k] of the NSR catalyst 70 (step S16)
・NOx storage amount (actual storage amount) NOxSt [k] stored in the NSR catalyst 70 one time before (result of the previous estimation process)
(b2) exhaust information:
・Temperature T_NSR _in [k] of exhaust gas from internal combustion engine 92 (step S20)
・Exhaust flow rate Q[k] (step S12)
(b3) Additive information:
Amount of additive flowing into NSR catalyst 70 RDCT_NSR _in [k] (step S18)
Amount of additive not contributing to reduction reaction Ad_thrmlytc[k] in rich spike control (step S28)

Thus, in step S30, the estimation unit 110 uses the second model 122 to estimate the NOx reduction amount NOx_rdct[k] of NOx reduced by the rich spike control as the purification performance of the NSR catalyst . As with the first model 121 described above, the second model 122 is composed of a machine learning model, so the NOx reduction amount, which is affected by a greater number of factors, can be estimated at high speed with a low computational load. Further, by using a sufficiently learned machine learning model as the second model 122, the estimation unit 110 can estimate the reduction amount of NOx with high accuracy. In step S30, the estimation unit 110 also functions as a "first acquisition unit" that acquires catalyst information.

In step S32, the estimation unit 110 calculates an estimated value of the amount of NOx stored in the NSR catalyst 70 after one time (time t=(k+1)Δt). Specifically, the estimating unit 110 first calculates the NOx amount NOx_NSR _out [k] flowing out from the NSR catalyst 70 at present (time t=kΔt) according to Equation 4 below.

Here, NOxConv[k] is the NOx purification rate of the NSR catalyst 70 estimated by the first model 121 . NOx_NSR _in [k] is the amount of NOx flowing into the NSR catalyst 70 obtained in step S14.

Next, the estimating unit 110 calculates an estimated value of the NOx storage amount NOxSt[k+1] of the NOx stored in the NSR catalyst 70 after one time (time t=(k+1)Δt) using the following Equation 5. . Equation 5 functions as "catalyst state estimation model 120".

Each value in Equation 5 is the purification performance (state) of the NSR catalyst 70 at the present (time t=kΔt) obtained or estimated by the estimation unit 110 in steps S12 to S30. Specifically, NOxSt[k] is the NOx storage amount (actual storage amount) stored in the NSR catalyst 70 obtained as a result of the previous estimation process. NOx_NSR _in [k] is the amount of NOx flowing into the NSR catalyst 70 obtained in step S14. NOx_NSR _out [k] is the amount of NOx flowing out from the NSR catalyst 70, which is obtained by Equation (4). NOx_Gas[k] is the amount of NOx purified by the gas phase reaction of the NSR catalyst 70 estimated by the third model 123 . NOx_rdct[k] is the reduction amount of NOx reduced by the rich spike control estimated by the second model 122 .

After step S32 is completed, estimation unit 110 outputs the NOx storage amount NOxSt[k+1] estimated in step S32. The output can be performed in any manner, for example, it may be displayed on a display unit (not shown) provided in the catalyst state estimation device 10, or it may be recorded in the estimation history in the storage unit 12. Signals may be sent to other controls in the vehicle. In addition to the NOx storage amount NOxSt[k+1] estimated in step S32, or in place of the NOx storage amount NOxSt[k+1], the estimation unit 110 obtains, calculates, or estimates the NOx storage amount NOxSt[k+1] obtained, calculated, or estimated in at least one of steps S12 to S30. You can also output the result. After that, the estimation unit 110 terminates the processing.

As described above, according to the catalyst state estimating device 10 of the first embodiment, the estimating unit 110 uses the catalyst state estimating model 120 to determine the NOx (nitrogen oxidation as a harmful substance) occluded in the NSR catalyst 70. Estimate the storage amount NOxSt[k+1] of the substance). As described in step S30, the second model 122 of the catalyst state estimation model 120 uses the estimation result (the amount of additive that does not contribute to the reduction reaction in rich spike control) by the fourth model 124, which is a physical model, to obtain a rich The amount of NOx reduced in spike control is estimated. Further, as is clear from Equation 4, the catalyst state estimation model 120 uses the estimation result (NOx purification rate) by the first model 121, which is a machine learning model (NN model), to determine the amount of NOx flowing out of the NSR catalyst 70. is estimated. As is clear from Equation 5, the catalyst state estimation model 120 uses these estimation results together, and uses the physical law to determine whether the NSR catalyst 70 will absorb after one hour (time t=(k+1)Δt). This is a model for estimating the amount of NOx (nitrogen oxide) stored NOxSt[k+1]. Therefore, according to the present embodiment, the estimating unit 110 calculates the NOx storage amount NOxSt [ k+1] can be performed with high accuracy and at high speed while satisfying the laws of physics.

Further, the amount NOxSt[k+1] of NOx (nitrogen oxide) stored in the NSR catalyst 70 fluctuates under the influence of the amount NOxSt[k] at the previous time (time t=kΔt) (in other words, will fluctuate under the influence of time history). As described in steps S30 and S32 (formulas 4 and 5), the estimation unit 110 of the first embodiment obtains the current estimation results by the first model 121, the second model 122, the third model 123, and the fourth model 124. is used to estimate the NOx storage amount NOxSt[k+1] after one time (time t=(k+1)Δt). Therefore, the estimation unit 110 can highly accurately estimate the purification performance of the NSR catalyst 70 at the next time based on the purification performance of the NSR catalyst 70 at the previous time (time t=kΔt).

Furthermore, the purification performance of the NSR catalyst 70 is based on catalyst information (for example, the temperature of the NSR catalyst 70, the amount of NOx stored in the NSR catalyst 70, the ratio of the actual amount of NOx to the saturated amount of NOx stored in the NSR catalyst 70). ). According to the catalyst state estimating device 10 of the first embodiment, the estimating unit 110 uses the catalyst information that affects the purification performance of the NSR catalyst 70 as the first model 121 (item a1), the second model 122 (item b1), the third model 123 (item c1), and the fourth model 124 (item d1) are applied to the catalyst state estimation model 120 by adopting the parameters of the vector U as input variables. As a result, the estimating unit 110 can accurately estimate the purification performance of the NSR catalyst 70. In the technique of estimating the purification performance of the catalyst NSR catalyst 70 that purifies NOx (hazardous substances) in the exhaust gas by absorbing them, , can improve the accuracy of the estimation.

Furthermore, according to the catalyst state estimation device 10 of the first embodiment, the estimation unit 110 estimates the NOx (hazardous substance) conversion rate NOxConv[k] in the NSR catalyst 70 by using the first model 121. can be done. In addition, by using the second model 122, the estimation unit 110 can estimate the reduction amount NOx_rdct[k] of the NOx stored in the NSR catalyst 70 in the rich spike control (reduction control). Also, the estimation unit 110 can estimate the NOx purification amount NOx_Gas[k] purified by the gas phase reaction of the NSR catalyst 70 by using the third model 123 . In addition, the estimation unit 110 can estimate the additive amount Ad_thrmlytc[k] that does not contribute to the reduction reaction in the rich spike control by using the fourth model 124 . Thus, according to the catalyst state estimation device 10, various purification performances of the NSR catalyst 70 can be estimated by using different mathematical models 121-124.

Furthermore, the purification performance of the NSR catalyst 70 is affected by information on the exhaust flowing into the catalyst (for example, exhaust temperature, exhaust flow rate, amount of NOx in the exhaust, exhaust air-fuel ratio) in addition to catalyst information. varies depending on According to the catalyst state estimating device 10 of the first embodiment, the estimating unit 110 obtains the information (items a1, b1, c1, d1) of these catalysts that affect the purification performance of the NSR catalyst 70, and the exhaust gas flowing into the catalyst. (items a2, b2, c2, d2) to the catalyst state estimation model 120, the purification performance of the NSR catalyst 70 can be estimated with even higher accuracy. In addition to the information on the catalyst (items b1 and d1), the estimation unit 110 further adds information on the additive (for example, the amount of the additive flowing into the NSR catalyst 70, the amount of the additive that does not contribute to the reduction reaction: item By applying b3, d3) to the catalyst state estimation model 120, the purification performance of the NSR catalyst 70 is estimated. For this reason, for example, the NSR catalyst 70 exemplified in this embodiment, and other selective reduction catalysts (SCR catalysts: Selective Catalytic Reduction catalyst) Estimate the purification performance of catalysts that purify harmful substances using additives. In some cases, the accuracy of estimation can be further improved.

Furthermore, if data acquired during transient operation of the internal combustion engine 92 is used as training data for generating at least a portion of the NN model among the first model 121, the second model 122, and the third model 123, The estimating unit 110 can also estimate the purification performance of the NSR catalyst 70 during transient operation of the internal combustion engine 92 in addition to the purification performance of the NSR catalyst 70 during steady operation of the internal combustion engine 92 .

<Second embodiment>
FIG. 6 is a block diagram of an exhaust purification system 1a in the second embodiment. In the first embodiment shown in FIG. 1, the estimation unit 110 uses the estimation results of the first to fourth models 121 to 124 to calculate the purification performance of the NSR catalyst 70 after one time (time t=(k+1)Δt ), the amount of NOx stored in the NSR catalyst 70 was estimated. However, the estimation unit 110a of the second embodiment uses the first model 121 to estimate the NOx purification rate NOxConv[k] of the NSR catalyst 70 at present (time t=kΔt) as the purification performance of the NSR catalyst 70. do.

A catalyst state estimation model 120a including only the first model 121 described with reference to FIG. 2 is stored in advance in the storage unit 12 of the second embodiment. Estimating unit 110a uses first model 121 of catalyst state estimating model 120a to perform steps S10 to S24 and step S26 of the estimating process described with reference to FIG. After executing step S26, the estimation unit 110a outputs the NOx purification rate NOxConv[k] of the NSR catalyst 70 estimated in step S26 as the purification performance of the NSR catalyst 70. FIG. Thus, the second embodiment can also achieve the same effects as the first embodiment described above.

Note that the exhaust purification system 1 a may include a third model 123 instead of the first model 121 and/or together with the first model 121 . In this case, the estimator 110a executes step S27 instead of step S26 and/or along with step S26, and outputs the estimation result as the purification performance of the NSR catalyst 70. FIG. Also, the exhaust purification system 1 a may include a fourth model 124 instead of the first model 121 and/or together with the first model 121 . In this case, the estimator 110a executes step S28 instead of/or together with step S26, and outputs the estimation result as the purification performance of the NSR catalyst 70. FIG. Further, the exhaust gas purification system 1a does not include the above-mentioned formulas 4 and 5, but includes at least a part of the first model 121, the second model 122, the third model 123, and the fourth model 124. may In this case, the estimation unit 110a executes steps S26 to S30 and outputs at least part of the estimation results as the purification performance of the NSR catalyst . Even if it does in this way, there can exist an effect similar to 1st Embodiment mentioned above.

<Third Embodiment>
FIG. 7 is a block diagram of an exhaust purification system 1b in the third embodiment. In the first embodiment shown in FIG. 1, the purification performance of the NSR catalyst 70 was estimated using a single first to third models 121-123. However, in the exhaust purification system 1b of the third embodiment, the plurality of first to third models 121b, 122b, 123b are selectively used to estimate the purification performance of the NSR catalyst .

The exhaust purification system 1b includes a catalyst state estimation device 10b instead of the catalyst state estimation device 10. FIG. The catalyst state estimating device 10b includes an estimating section 110b instead of the estimating section 110, and a catalyst state estimating model 120b instead of the catalyst state estimating model 120. FIG. The catalyst state estimation model 120b includes a plurality of first models 121b, a plurality of second models 122b, a plurality of third models 123b, and a fourth model . Each first model 121b is created using teacher data obtained from a plurality of NSR catalysts 70 with different degrees of deterioration. Specifically, for example, the first model 121b(1) is created from teacher data acquired from the NSR catalyst 70 with a low degree of deterioration, and the first model 121b(2) is created from training data with an intermediate degree of deterioration. The first model 121b(3) is created from teacher data acquired from the NSR catalyst 70, and the first model 121b(3) is created from teacher data acquired from the NSR catalyst 70 having a large degree of deterioration. The degree of deterioration (degree of deterioration) may be determined from the state of the NSR catalyst 70, the timing of replacement of the NSR catalyst 70, or the like, or may be determined from the NOx concentration in the exhaust emitted from the NSR catalyst 70. It may be determined from the running distance or the like. Each of the second models 122b and each of the third models 123b are created using teacher data obtained from a plurality of NSR catalysts 70 with different degrees of deterioration, similarly to the first model 121b.

Estimation unit 110b estimates the NOx purification rate of NSR catalyst 70 using first model 121b corresponding to the degree of deterioration of NSR catalyst 70 in step S26 of the estimation process (FIG. 4). In step S27, the estimating unit 110b estimates the amount of NOx purified by the gas phase reaction of the NSR catalyst 70 using the third model 123b corresponding to the degree of deterioration of the NSR catalyst 70. FIG. Furthermore, in step S30, the estimating unit 110b uses the second model 122b corresponding to the degree of deterioration of the NSR catalyst 70 to estimate the reduction amount of NOx reduced by the rich spike control.

In this way, the same effects as those of the above-described first embodiment can be obtained also in the third embodiment. Further, according to the third embodiment, the catalyst state estimation model 120b includes a plurality of first models 121b and second models 122b created using teacher data acquired from a plurality of NSR catalysts 70 with different degrees of deterioration. , and third model 123b, the optimum first model 121b, second model 122b, and third model 123b can be adopted according to the degree of deterioration of the NSR catalyst 70. FIG. As a result, according to the third embodiment, the purification performance of the catalyst can be estimated with higher accuracy.

<Fourth Embodiment>
FIG. 8 is a block diagram of an exhaust purification system 1c in the fourth embodiment. In the exhaust gas purification system 1c of the fourth embodiment, the amount of additive (more specifically, the amount of reducing agent, hereinafter simply referred to as "outflow additive amount") flowing out from the NSR catalyst 70 during rich spike control is estimated. can. The estimated outflow additive amount can be regarded as the amount of additive that was not consumed in the rich spike control (in other words, the amount of additive that was excessive). Therefore, in the exhaust gas purification system 1c of the present embodiment, the fuel consumption can be reduced by executing the rich spike control so as to reduce the outflow additive amount.

Instead of the catalyst state estimation model 120, a catalyst state estimation model 120c is stored in the storage unit 12 of the exhaust purification system 1c. The catalyst state estimation model 120c includes a fifth model 125 instead of the second model 122 described in the first embodiment. The fifth model 125 is a model for estimating the amount of additive that flows out of the NSR catalyst 70 without being used to reduce NOx (that is, the amount of additive that flows out) during rich spike control. The fifth model 125 is composed of a machine learning model, specifically a NN, and has the same configuration as the first model 121 described in the first embodiment except for points described later.

A vector U, which is an input variable of the fifth model 125, is composed of n components (u1, u2, . . . un, where n is a natural number). For each component of the vector U, parameters shown in items e1 to e3 below can be adopted.
(e1) Catalyst information: at least one of the temperature of the front end of the NSR catalyst 70, the temperature of the NSR catalyst 70, and the NOx storage amount (actual storage amount) stored in the NSR catalyst 70 one hour before One (e2) Exhaust information: At least one of the temperature of the exhaust gas from the internal combustion engine 92 and the flow rate of the exhaust gas (e3) Additive information: Additives (CO, H ₂ , and other unburned gases) and the amount of additives (CO, H ₂ , unburned gases) that do not contribute to the reduction reaction in rich spike control.

For the vector U of the fifth model 125, as with the first model 121, it is necessary to input at least one or more of the catalyst information shown in item e1, and the exhaust information shown in item e2. may be input with at least one or more, or may be omitted entirely. However, in order to improve the estimation accuracy of the outflow additive amount in the fifth model 125, it is preferable that the number of input parameters is large. The output variable Z of the fifth model 125 is an estimated value of the amount of additive flowing out of the NSR catalyst 70 (outflow additive amount) during rich spike control under the conditions represented by the input variable (vector U). Note that the fifth model 125 corresponds to the "fifth model".

FIG. 9 is a flowchart showing the procedure of estimation processing by the estimation unit 110c of the fourth embodiment. The difference from the first embodiment described with reference to FIG. 4 is that steps S40 and S42 are executed instead of step S30.

In step S40, the estimation unit 110c calculates an estimated value of the amount of additive flowing out of the NSR catalyst 70 (outflow additive amount) during rich spike control. Specifically, the estimating unit 110c adds the following current (time t=kΔt) values obtained in steps S12 to S28 to the input variable vector U of the fifth model 125 for estimating the outflow additive amount: and obtain the output variable Z. The estimator 110c sets the output variable Z obtained from the fifth model 125 to the current (time t=kΔt) outflow additive amount RDCT_NSR _out [k].
(e1) catalyst information:
・Temperature (floor temperature) T[k] of the NSR catalyst 70 (step S16)
・NOx storage amount (actual storage amount) NOxSt [k] stored in the NSR catalyst 70 one time before (result of the previous estimation process)
(e2) Exhaust information:
・Temperature T_NSR _in [k] of exhaust gas from internal combustion engine 92 (step S20)
・Exhaust flow rate Q[k] (step S12)
(e3) Additive information:
Amount of additive flowing into NSR catalyst 70 RDCT_NSR _in [k] (step S18)
- Amount Ad_thrmlytc[k] of additive that does not contribute to reduction reaction in rich spike control, obtained in step S28

In step S42, the estimator 110c calculates an estimated value of the amount of NOx (reduction amount) of the NOx stored in the NSR catalyst 70 that is reduced in the rich spike control. Specifically, the estimating unit 110c first calculates the change amount RDCT_ac of the additive in the NSR catalyst 70 (in other words, the amount of the additive expected to be used in the NSR catalyst 70) by the following Equation 6. .

Here, RDCT_NSR _in [k] is the amount of additive flowing into the NSR catalyst 70 acquired in step S18, and RDCT_NSR _out [k] is the amount of additive from the NSR catalyst 70 during the rich spike control estimated in step S40. It is the amount of additive that flows out (outflow additive amount). Next, the estimating unit 110c obtains the NOx reduction amount NOx_rdct[k] to be reduced by the rich spike control from the well-known NOx reduction reaction formula and the change amount RDCT_ac of the additive obtained by Equation (6). After that, in step S32, the estimating unit 110c uses the obtained NOx reduction amount NOx_rdct[k] to store the NOx stored in the NSR catalyst 70 one time later (time t=(k+1)Δt). Calculate volume estimates. The details are the same as those of the first embodiment described with reference to FIG.

In this way, the same effects as those of the above-described first embodiment can be obtained also in the fourth embodiment. Further, according to the fourth embodiment, the estimator 110c uses the fifth model 125 to determine the amount of additive flowing out of the NSR catalyst 70 without being consumed in the rich spike control (reduction control) RDCT_NSR _out [k ] (flowing additive amount) can be estimated. Then, the catalyst state estimation device 10c performs feedback control to the rich spike control unit 93 (reduction control unit that performs reduction control) so that the rich spike control unit 93 reduces the outflow additive amount RDCT_NSR _out [k]. By implementing the rich spike control at , it is possible to reduce the amount of excess additive while maintaining the purification performance of the harmful substances in the NSR catalyst 70 . As a result, according to this configuration, it is possible to reduce fuel consumption in reduction control, such as rich spike control in the NSR catalyst 70, in which fuel is consumed to reduce harmful substances occluded in the NSR catalyst 70. FIG.

Furthermore, according to the fourth embodiment, the estimating unit 110c estimates the outflow additive amount RDCT_NSR _out [k] using the fifth model 125 configured by a machine learning model. Estimation of the additive amount can be performed at high speed with a low computational load. Further, by using a sufficiently learned machine learning model as the fifth model 125, the estimation unit 110c can estimate the outflow additive amount with high accuracy.

Furthermore, according to the fourth embodiment, the estimation unit 110c _calculates the rich spike control ( reduction control), the reduction amount NOx_rdct[k] of NOx as a harmful substance can be obtained ( FIG. 9 : step S42). In addition, the estimating unit 110c calculates the NOx reduction amount NOx_rdct[k] and the estimation results (the nitrogen oxide purification rate NOxConv[k], the gas phase Purified amount of nitrogen oxides (NOx_Gas[k]) as a harmful substance purified in the reaction, and the result of estimation by the fourth model 124, which is a physical model (amount of additive that does not contribute to the reduction reaction in reduction control Ad_thrmlytc[k]) ), and using the laws of physics, the NOx storage amount NOxSt[k+1] in the NSR catalyst 70 is estimated. Therefore, according to the catalyst state estimating device 10c of the fourth embodiment, the NOx storage amount, which is influenced by a greater number of factors, can be estimated with high precision and high speed while satisfying the laws of physics. Further, the amount of NOx stored in the NSR catalyst 70 fluctuates under the influence of the amount of storage at the previous time. According to the catalyst state estimation device 10c of the fourth embodiment, the estimation unit 110c uses the current estimation results from the first model 121, the third model 123, the fourth model 124, and the fifth model 125 to obtain the following The storage amount NOxSt[k+1] at time (time t=(k+1)Δt) is estimated. Therefore, the estimation unit 110c can highly accurately estimate the purification performance of the NSR catalyst 70 at the next time based on the purification performance of the NSR catalyst 70 at the previous time (time t=kΔt).

Furthermore, according to the fourth embodiment, the estimating unit 110c uses both the information on the catalyst that affects the purification performance of the NSR catalyst 70 and the information on the exhaust that flows into the NSR catalyst 70 into the catalyst state estimation model 120c (the second model). 5 model 125), the purification performance of the NSR catalyst 70 can be estimated with higher accuracy (FIG. 9: step S40). In addition, the estimating unit 110c applies both the information about the catalyst that affects the purification performance of the NSR catalyst 70 and the information about the additive to the catalyst state estimation model 120c (fifth model 125) so that the NSR catalyst 70 purification performance can be estimated with higher accuracy ( FIG. 9 : step S40).

Note that the catalyst state estimation device 10c of the fourth embodiment does not include at least part of the first, third, and fourth models, as in the second embodiment described with reference to FIG. good. In this case, the estimating unit 110c executes the executable steps using the model included in the catalyst state estimating model 120c in the process described in FIG. Output as performance. Even if it does in this way, there can exist an effect similar to 1st Embodiment and 4th Embodiment which were mentioned above.

Note that the catalyst state estimation device 10c of the fourth embodiment is created using teacher data acquired from a plurality of NSR catalysts 70 with different degrees of deterioration, as in the third embodiment described with reference to FIG. A plurality of fifth models 125 may be included. In this case, the estimating unit 110c estimates the outflow additive amount of the NSR catalyst 70 using the fifth model 125 corresponding to the degree of deterioration of the NSR catalyst 70 in step S40 of the process described with reference to FIG. In this way, the estimation unit 110c can adopt the fifth model 125 that is most suitable for the degree of deterioration of the NSR catalyst 70. FIG. As a result, the purification performance of the NSR catalyst 70 can be estimated with higher accuracy.

<Fifth Embodiment>
FIG. 10 is a block diagram of an exhaust purification system 1d according to the fifth embodiment. The exhaust purification system 1d of the fifth embodiment is used in an exhaust purification device 20d equipped with a plurality of NSR catalysts, and can estimate the purification performance of the plurality of NSR catalysts as a whole. The exhaust purification system 1d includes a catalyst state estimation device 10d instead of the catalyst state estimation device 10 in the configuration of the first embodiment described in FIG.

The exhaust purification device 20 d further includes a second NSR catalyst 71 arranged downstream of the NSR catalyst 70 . Hereinafter, the NSR catalyst 70 is also called the first NSR catalyst 70 for distinction. The second NSR catalyst 71 is the same type of NOx storage reduction catalyst as the first NSR catalyst 70 . In the present embodiment, the term "homogeneous catalyst" means catalysts having the same or similar exhaust purification mechanism. The second NSR catalyst 71 purifies the NOx in the exhaust gas by absorbing and accumulating harmful substances (NOx) that have leaked downstream because the first NSR catalyst 70 could not completely occlude them. Hereinafter, a plurality of catalysts of the same type (first and second NSR catalysts 70, 71) mounted on the exhaust purification device 20d are also collectively referred to as "catalyst group CG". In the configuration shown in FIG. 10, the first NSR catalyst 70 corresponds to "the most upstream catalyst" and the second NSR catalyst 71 corresponds to the "most downstream catalyst".

Note that in the present embodiment, the rich spike control (reduction control) described in the first embodiment is performed for each of the plurality of catalysts. Specifically, CO, H ₂ and other unburned gases discharged from the internal combustion engine 92 reduce NOx stored in the NSR catalyst 70 to nitrogen gas (N ₂ ). In addition, CO, H ₂ and other unburned gases that have flowed out of the first NSR catalyst 70 without being used for reduction control in the first NSR catalyst 70 convert NOx stored in the second NSR catalyst 71 into nitrogen gas (N ₂ ).

The catalyst state estimation device 10 d further includes a second temperature acquisition section 79 that is a sensor that measures the bed temperature of the second NSR catalyst 71 . Further, the catalyst state estimation device 10d includes an estimation unit 110d instead of the estimation unit 110, and a catalyst state estimation model 120d instead of the catalyst state estimation model 120. FIG. The estimation unit 110d differs from the first embodiment (FIG. 4) in the content of estimation processing. The catalyst state estimation device 10d includes a first model 121d, a second model 122d, a third model 123d, and a fourth model 124d.

The first model 121d is used for estimating the NOx purification rate of the plurality of catalysts of the same type (that is, the first and second NSR catalysts 70 and 71) mounted in the exhaust purification device 20d. is a model. The first model 121d is, like the first embodiment, a machine learning model, specifically a NN. For each component of the input variable vector U of the first model 121d, parameters shown in items a11 and a12 below can be adopted.
(a11) Catalyst information: the temperature of the front end of the first NSR catalyst 70 located most upstream, the temperatures of the first and second NSR catalysts 70 and 71, and and a ratio of the total value of the actual NOx storage amount to the total value of the saturated NOx storage amount in the plurality of catalysts 70 and 71 (a12) Exhaust information: At least one of the temperature of the exhaust gas flowing from the internal combustion engine 92 into the first NSR catalyst 70 located most upstream, the flow rate of the exhaust gas, and the amount of NOx contained in the exhaust gas.

As in the first embodiment, for the vector U of the first model 121d, it is necessary to input at least one or more of the catalyst information shown in item a11, and the exhaust information shown in item a12. may be input with at least one or more, or may be omitted entirely. The processing after each component of the vector U is input to the input layer is the same as in the first embodiment. The output variable Z of the first model 121d is an estimated value of the NOx purification rate of the plurality of catalysts (the first and second NSR catalysts 70, 71) as a whole under the conditions represented by the input variables.

As in the first embodiment, the first model 121d causes the NN to learn the relationship between the input variables and the output variables so that the output variable Z matches the physical quantity to be estimated, and each value W _ij , θ _j , W _jk , θ _k are determined in advance. During this learning, in the present embodiment, the catalyst group CG consisting of a plurality of catalysts (first and second NSR catalysts) of the same type arranged in the model main flow path (exhaust pipe) is the first NSR located most upstream. It is preferable to use teacher data acquired from the catalyst group CG, regarding the catalyst inlet as the inlet of the first NSR catalyst 70 and the outlet of the second NSR catalyst positioned most downstream as the outlet of the catalyst group CG.

The second model 122d is a model for estimating the total reduction amount of NOx in rich spike control (reduction control) for each of a plurality of catalysts. The second model 122d is, like the first embodiment, a machine learning model, specifically a NN. Parameters shown in the following items b11 to b13 can be adopted for each component of the vector U that is the input variable of the second model 122d.
(b11) catalyst information: the temperature of the front end of the first NSR catalyst 70 located most upstream, the temperatures of the first and second NSR catalysts 70 and 71, and and at least one of (b12) exhaust information: the temperature of the exhaust gas flowing from the internal combustion engine 92 to the first NSR catalyst 70 located most upstream, and the flow rate of the exhaust gas. at least one of (b13) additive information: the amount of additives (CO, H ₂ and other unburned gases) flowing into the first NSR catalyst 70 located most upstream, and a plurality of catalysts 70 , 71, and at least one of the total amount of additives that do not contribute to the reduction reaction in the rich spike control for each

As in the first embodiment, for the vector U of the second model 122d, it is necessary to input at least one or more of the catalyst information shown in item b11, and the exhaust gas shown in items b12 and b13. At least one or more of the information and the additive information may be input, or all of them may be omitted. The output variable Z of the second model 122d is the total value (total estimated value). As with the first model 121d, the second model 122d preferably uses teacher data acquired from the catalyst group CG consisting of a plurality of catalysts of the same type (first and second NSR catalysts).

The third model 123d is a model for estimating the total amount of purified NOx in the gas phase reaction of each of the plurality of catalysts. The third model 123d is composed of a machine learning model, specifically a NN, as in the first embodiment. Parameters shown in the following items c11 and c12 can be adopted for each component of the vector U that is the input variable of the third model 123d.
(c11) Catalyst information: At least one of the temperature of the front end of the first NSR catalyst 70 located most upstream and the temperatures of the first and second NSR catalysts 70 and 71 (c12) Exhaust information: Internal combustion At least one of the temperature of the exhaust flowing into the first NSR catalyst 70 located most upstream from the engine 92, the flow rate of the exhaust, and the amount of NOx contained in the exhaust.

As in the first embodiment, for the vector U of the third model 123d, it is necessary to input at least one or more of the catalyst information shown in item c11, and the exhaust information shown in item c12. may be input with at least one or more, or may be omitted entirely. The output variable Z of the third model 123d is the total value of the NOx purification amount (total of estimated value). The third model 123d preferably uses teacher data acquired from the catalyst group CG consisting of a plurality of catalysts of the same type (first and second NSR catalysts), like the first model 121d.

The fourth model 124d is a model for estimating the total amount of additives that do not contribute to the reduction reaction in rich spike control (reduction control) for each of a plurality of catalysts. The fourth model 124d is composed of a machine learning model, specifically a NN, as in the first embodiment. Parameters shown in the following items d11 to d13 can be adopted for each component of the vector U, which is the input variable of the fourth model 124d.
(d11) Catalyst information: at least one of the temperature of the front end of the first NSR catalyst 70 located most upstream and the temperatures of the first and second NSR catalysts 70 and 71 (d12) Exhaust information: internal combustion At least one of the temperature of the exhaust gas flowing from the engine 92 into the first NSR catalyst 70 located most upstream and the air-fuel ratio of the exhaust gas (d13) Additive information: the first NSR catalyst 70 located most upstream amount of additive (CO, _H2 , and other unburned gases) flowing into

As in the first embodiment, for the vector U of the fourth model 124d, it is necessary to input at least one or more of the catalyst information shown in item d11, and the exhaust gas shown in items d12 and d13. At least one or more of the information and the additive information may be input, or all of them may be omitted. The output variable Z of the fourth model 124d is the amount of additive that does not contribute to the reduction reaction in the rich spike control for each of the multiple catalysts (first and second NSR catalysts 70, 71) under the conditions represented by the input variables. is the total value of (estimated value of the total). The fourth model 124d preferably uses teacher data acquired from the catalyst group CG consisting of a plurality of catalysts of the same type (first and second NSR catalysts), like the first model 121d.

FIG. 11 is a flowchart showing the procedure of estimation processing by the estimation unit 110d of the fifth embodiment. The estimation process of the fifth embodiment is a process of estimating the purification performance of a plurality of catalysts as a whole (as a catalyst group CG as a whole), and is executed at an arbitrary timing as in the first embodiment shown in FIG. . Definitions of Δt, time t=kΔt, time t=(k+1)Δt, and time t=(k−1)Δt used in the following description are the same as in the first embodiment. Also, in the following description, only processing different from that of the first embodiment shown in FIG. 4 will be described.

In step S14d, the estimating unit 110d acquires the NOx amount _{NOx_NSR in} flowing into the first NSR catalyst 70 positioned most upstream at the present time (time t=kΔt). In step S16d, the estimation unit 110d acquires the bed temperature T1[k] of the first NSR catalyst 70 at present (time t=kΔt) from the first temperature acquisition unit 78. FIG. The estimator 110d also acquires the bed temperature T2[k] of the second NSR catalyst 71 at present (time t=kΔt) from the second temperature acquirer 79 . In step S18d, the estimating unit 110d calculates the current (time t=kΔt) current (time t=kΔt) first NSR catalyst located most upstream from the amount of fuel injected by the rich spike control using a previously prepared calculation formula, map, or the like. Obtain the amount of additive flowing into RDCT_NSR _in [k]. In step S20d, the estimation unit 110d acquires the current (time t=kΔt) temperature T_NSR _in [k] of the exhaust flowing into the first NSR catalyst 70 positioned most upstream from the exhaust temperature acquisition unit 53 .

In step S24d, the estimation unit 110d calculates the ratio of the total value of the actual NOx storage amounts to the total value of the saturated NOx storage amounts of the plurality of catalysts (first and second NSR catalysts 70, 71). Specifically, the estimation unit 110d applies the bed temperature T1[k] of the first NSR catalyst 70 acquired in step S16d to the relational expression representing the characteristic NP (FIG. 5) to obtain the current (time t=kΔt) , the NOx saturated storage amount of the first NSR catalyst 70 is obtained. Similarly, the estimating unit 110d applies the bed temperature T2[k] of the second NSR catalyst 71 acquired in step S16d to the relational expression representing the characteristic NP (FIG. 5) to obtain the current (time t=kΔt) The NOx saturated storage amount of the 2NSR catalyst 71 is obtained. The estimating unit 110d sets the calculated sum of the NOx saturated storage amounts of the first and second NSR catalysts 70 and 71 as the total value StrtNOxSt[k] of the NOx saturated storage amounts of the plurality of catalysts.

Next, the estimating unit 110d uses the obtained total value StrtNOxSt[k] of the saturated NOx storage amounts and the total value NOxSt[k] of the actual NOx storage amounts of the plurality of catalysts to determine the saturated storage amount. A ratio r[k] of the total value of the actual storage amount to the total value is calculated. Here, for the total value NOxSt[k] of the actual NOx storage amounts in a plurality of catalysts, the estimated value obtained in the previously executed estimation process (FIG. 11) can be used, and if there is no estimated value, a predetermined is the same as in the first embodiment in that the default value of can be used.

In step S26d, the estimator 110d calculates an estimated value of the NOx purification rate of the plurality of catalysts (first and second NSR catalysts 70, 71) as a whole. Specifically, the estimating unit 110d stores the input variable vector U of the first model 121d in steps S12, S14d, S16d, S20d, and S24d, the result of the previous estimation process, and the current (time t=kΔt) Set each value and obtain the output variable Z. The estimation unit 110d converts the output variable Z obtained from the first model 121d into the NOx purification rate NOxConv[k ].

In step S27d, the estimation unit 110d calculates the total value (total estimated value) of the NOx purification amount in the gas phase reaction of each of the plurality of catalysts (first and second NSR catalysts 70, 71). Specifically, the estimation unit 110d sets the current (time t=kΔt) values obtained in steps S12, S14d, S16d, and S20d to the input variable vector U of the third model 123d, and sets the output variable Z to demand. The estimating unit 110d uses the output variable Z obtained from the third model 123d as the NOx purification Let the total amount (estimated total) be NOx_Gas[k].

In step S28d, the estimation unit 110d calculates the total value (estimated total value) of the amounts of additives that do not contribute to the reduction reaction in the rich spike control for each of the plurality of catalysts (first and second NSR catalysts 70, 71). do. Specifically, the estimation unit 110d sets the current (time t=kΔt) values obtained in steps S16d, S18d, and S20d to the input variable vector U of the fourth model 124d, and obtains the output variable Z. The estimator 110d applies the output variable Z obtained from the fourth model 124d to the reduction reaction in rich spike control for each of the plurality of catalysts (first and second NSR catalysts 70, 71) at present (time t=kΔt). Let Ad_thrmlytc[k] be the total value (estimated value of the total) of the amount of additives that do not contribute to the

In step S30d, the estimation unit 110d calculates the total value (estimated total value) of NOx reduction amounts in the rich spike control for each of the plurality of catalysts (first and second NSR catalysts 70, 71). Specifically, the estimation unit 110d stores the input variable vector U of the second model 122d in steps S12, S16d, S18d, S20d, and S28d, the result of the previous estimation process, and the current (time t=kΔt) Set each value and obtain the output variable Z. The estimating unit 110d uses the output variable Z obtained from the second model 122d as the NOx NOx Let the total value (estimated value of the total) of the amount of reduction be NOx_rdct[k].

In step S32d, the estimation unit 110d calculates the total value (total Estimated value) NOxSt[k+1] is calculated. Specifically, the estimating unit 110d estimates or calculates in steps S12 and S14d to S30d the current (time t=kΔt) multiple is the purification performance (or state) of the catalyst. After step S32d ends, the estimation unit 110d outputs the total NOx storage amount NOxSt[k+1] estimated in step S32d. Details are the same as in the first embodiment.

As described above, the catalyst state estimating device 10d of the fifth embodiment can also achieve the same effects as the above-described first embodiment. Further, according to the catalyst state estimation device 10d of the fifth embodiment, when a plurality of catalysts (first and second NSR catalysts 70, 71) are provided in the main flow path (exhaust pipe 30), the estimation unit 110d By applying the information of the first NSR catalyst 70 positioned most upstream to the catalyst state estimation model 120d, the purification performance of the plurality of catalysts as a whole can be estimated. That is, according to the catalyst state estimation device 10d of the fifth embodiment, the plurality of catalysts (the first and second NSR catalysts 70, 71) on the main flow path are regarded as one catalyst (that is, the catalyst group CG), and the catalyst group CG purification performance can be estimated. For this reason, compared to the case of estimating the purification performance of each catalyst individually, the catalyst information acquisition part (for example, the sensors constituting the flow rate acquisition part 52, the exhaust temperature acquisition part 53, the NOx concentration acquisition part 54, etc.) It is possible to reduce the number of models and the number of catalyst state estimation models 120d prepared in advance, and reduce the computational load on the catalyst state estimation device 10d. Further, when the plurality of catalysts provided in the main flow path are of the same type (NSR catalyst in the example of the fifth embodiment), the estimation unit 110d regards them as one catalyst and estimates the purification performance. . If the catalysts are of the same type, there is no difference in the catalyst information items (for example, the amount of additive that flows in) that affect the purification performance, so the estimating unit 110d can accurately estimate the purification performance.

Further, according to the catalyst state estimation device 10d of the fifth embodiment, a plurality of catalysts arranged in the model main flow path are treated as one catalyst (catalyst group CG), and teacher data acquired from the catalyst group CG is used as By using and learning, a first model 121d, a second model 122d, a third model 123d, and a fourth model 124d of the catalyst state estimation model 120d can be constructed. Therefore, in each of these models of the catalyst state estimation model 120d, the influence of the main flow path (exhaust pipe, etc.) between the catalysts belonging to the catalyst group CG can be taken into consideration. By using the catalyst state estimation model 120d constructed in this manner in the estimation process, the estimation unit 110d can estimate the exhaust gas It is possible to accurately estimate the purification performance of a plurality of catalysts as a whole without requiring information on pipes and the like.

In addition, in the fifth embodiment, two NSR catalysts (first and second NSR catalysts 70 and 71) are illustrated as specific examples of the plurality of catalysts. However, the exhaust purification system 1d may have three or more NSR catalysts.

<Sixth Embodiment>
FIG. 12 is a block diagram of an exhaust purification system 1e in the sixth embodiment. In the fifth embodiment described with reference to FIG. 10, the temperature (floor temperature) of each catalyst provided in the exhaust purification device 20d is obtained by the first and second temperature acquisition units 78 and 79 provided for each catalyst. obtained respectively. However, in the exhaust purification system 1 e of the sixth embodiment, the catalyst state estimation device 10 e includes a temperature estimation section 159 instead of the second temperature acquisition section 79 . A temperature estimator 159 estimates the temperature of other catalysts (the second NSR catalyst 71 in the example of FIG. 12) other than the first NSR catalyst 70 positioned most upstream.

The temperature estimating unit 159 calculates the temperature T2 of the second NSR catalyst 71 from the temperature T1 of the first NSR catalyst 70 acquired by the first temperature acquiring unit 78 using a previously prepared calculation formula, map, or the like. In addition to the temperature T1 of the first NSR catalyst 70 positioned most upstream, the temperature estimating unit 159 further calculates the temperature of the exhaust gas from the internal combustion engine 92, the flow rate of the exhaust gas, the reaction heat generated by the first NSR catalyst 70, and the like. The temperature T2 of the second NSR catalyst 71 may be calculated in consideration of parameters. Similarly, even when three or more catalysts are installed in the exhaust gas purification system 1e, the temperature estimator 159 calculates the temperature T1 of the first NSR catalyst 70 positioned most upstream from the temperature Tn (n is a natural number for distinguishing each catalyst) can be calculated. The estimation unit 110e acquires the temperature Tn of the other catalyst from the temperature estimation unit 159 in step S16d of the estimation process described with reference to FIG.

As described above, the catalyst state estimating device 10e of the sixth embodiment can also achieve the same effects as those of the above-described first and fifth embodiments. Further, according to the catalyst state estimation device 10e of the sixth embodiment, the temperature estimation unit 159 can estimate the temperature of the other catalysts from the temperature of the first NSR catalyst 70 located most upstream. An acquisition unit (sensor, etc.) for acquiring the temperature of the catalyst can be omitted.

<Seventh Embodiment>
FIG. 13 is a block diagram of the exhaust purification system 1f of the seventh embodiment. The exhaust purification system 1f of the seventh embodiment is used in an exhaust purification device 20d equipped with a plurality of NSR catalysts, and the amount of additive flowing out from the plurality of NSR catalysts during rich spike control (more specifically, reducing agent and the amount of additive runoff) can be estimated. The exhaust purification system 1f includes a catalyst state estimation device 10f instead of the catalyst state estimation device 10 described in the first embodiment, and an exhaust purification device 20d instead of the exhaust purification device 20. FIG.

The configuration of the exhaust purification device 20d is as described in the fifth embodiment (FIG. 10). Instead of the catalyst state estimation model 120, a catalyst state estimation model 120f is stored in the storage unit 12 of the catalyst state estimation device 10f. The catalyst state estimation model 120f includes a first model 121d, a fifth model 125f, a third model 123d and a fourth model 124d. The configurations of the first model 121d, the third model 123d, and the fourth model 124d are as described in the fifth embodiment (FIGS. 10 and 11).

The fifth model 125f is not consumed in the rich spike control (reduction control) in the multiple catalysts of the same type (that is, the first and second NSR catalysts 70, 71) mounted in the exhaust gas purification device 20d, and is the most downstream This is a model for estimating the amount of additive (outflow additive amount) that flows out from the positioned catalyst (ie, the second NSR catalyst 71). The fifth model 125f is, like the fourth embodiment, a machine learning model, specifically a NN. For each component of the input variable vector U of the fifth model 125f, parameters shown in items e11 to e13 below can be employed.
(e11) Catalyst information: the temperature at the front end of the first NSR catalyst 70 located most upstream, the temperatures of the first and second NSR catalysts 70 and 71, and the temperature stored in the catalysts 70 and 71 one hour before and at least one of (e12) exhaust information: the temperature of the exhaust gas flowing from the internal combustion engine 92 to the first NSR catalyst 70 located most upstream, and the flow rate of the exhaust gas. , at least one of (e13) additive information: the amount of additives (CO, H ₂ and other unburned gases) flowing into the first NSR catalyst 70 located most upstream, and the plurality of catalysts 70 , 71, and at least one of the total amount of additives that do not contribute to the reduction reaction in the rich spike control for each

As in the fourth embodiment, for the vector U of the fifth model 125f, it is necessary to input at least one or more of the catalyst information shown in item e11, and the exhaust gas shown in items e12 and e13. At least one or more of the information and the additive information may be input, or all of them may be omitted. The output variable Z of the fifth model 125f is located at the most downstream position without being consumed in rich spike control with multiple catalysts of the same type (first and second NSR catalysts 70, 71) under various conditions represented by the input variables. is the amount of additive (outflow additive amount) that flows out from the catalyst (second NSR catalyst 71). The fifth model 125f preferably uses teacher data acquired from the catalyst group CG consisting of a plurality of catalysts of the same type (first and second NSR catalysts), like the first model 121d.

The procedure of the estimation process executed by the estimation unit 110f is the same as that of the fifth embodiment (FIG. 11) except that the following steps S40f and S42f are executed instead of step S30d.
Step S40f: Using the fifth model 125f, the estimating unit 110f uses the fifth model 125f to determine the amount of additive (outflow additive amount ) Estimate RDCT_NSR _out [k].
Step S42f: The estimating unit 110f uses Equation 6 described in the fourth embodiment (FIG. 9) to calculate NOx reduced by rich spike control in a plurality of catalysts (first and second NSR catalysts 70, 71). , the total value NOx_rdct[k] of the amount of reduction of .

As described above, the catalyst state estimating device 10f of the seventh embodiment can also achieve the same effects as those of the fourth and fifth embodiments described above. Also in the exhaust gas purification system 1f of the seventh embodiment, the temperature estimator 159 may be used to estimate the temperature (floor temperature) of the NSR catalyst on the downstream side, as in the sixth embodiment.

<Modification of this embodiment>
The present invention is not limited to the above-described embodiments, and can be implemented in various aspects without departing from the scope of the invention. For example, the following modifications are possible.

[Modification 1]
In the above embodiment, an example of the configuration of the exhaust gas purification system is shown. However, various modifications are possible for the configuration of the exhaust purification system. For example, the exhaust purification system may be used for purposes other than purifying the exhaust of an internal combustion engine. For example, exhaust gas cleaning systems may be used to clean exhaust gas from industrial facilities, commercial facilities, homes, and the like.

For example, an exhaust purification device of an exhaust purification system is equipped with a combination of a plurality of catalysts selected from among an NSR catalyst, an SCR catalyst, and a three-way catalyst, and the catalyst state estimation device measures the purification performance of these catalyst groups. can be estimated respectively. In addition, the exhaust purification device is equipped with a particulate matter removal filter (DPF: Diesel Particulate Filter) that removes particulate matter (PM), and the catalyst state estimation device may estimate the purification performance in this DPF. .

For example, the temperature acquisition unit that acquires the bed temperature of the catalyst may be provided in front (near the inlet) or behind (near the outlet) of the catalyst.

For example, at least one of the flow rate, the exhaust temperature, the NOx concentration, the front end temperature, and the temperature of the catalyst obtained by the flow rate acquisition unit, the exhaust temperature acquisition unit, the NOx concentration acquisition unit, the front end temperature acquisition unit, and the temperature acquisition unit is , the temperature estimated using the catalyst state estimation model may be substituted for the measured value by the sensor. Specifically, for example, the temperature of the catalyst is affected by the time history in the same way as the NOx storage amount in the NSR catalyst described above. Therefore, a catalyst state estimation model capable of estimating the temperature of the catalyst may be created separately, and the temperature of the catalyst may be estimated using the model for estimating the catalyst state.

[Modification 2]
In the above embodiment, an example of estimation processing in the estimation unit has been shown (FIGS. 4, 9, and 11). However, the content of the estimation process can be modified in various ways. For example, in the estimation processes shown in FIGS. 4, 9, and 11, the execution order of steps S12-S24 may be changed, and the execution order of steps S26-S30, S40, and S42 may be changed. At least some of steps S12 to S32, S40, and S42 may be omitted, and other steps not described may be executed.

For example, the estimation unit may perform at least one of the following steps S100 and S102 in addition to each step described above. Steps S100 and S102 can be executed at arbitrary timing.
Step S100: The rich spike control unit sets a sufficiently long time for performing the rich spike control to sufficiently remove the NOx stored in the NSR catalyst (until the actual storage amount with respect to the saturated storage amount can be regarded as 0). give back. In this state, the estimation unit resets the estimated value NOxSt[k+1] of the amount of NOx stored in the NSR catalyst to 0 after one time (time t=(k+1)Δt).
Step S102: The rich spike control unit causes the NSR catalyst to store a sufficient amount of NOx (until the ratio of the actual storage amount to the saturated storage amount becomes 100%) while the rich spike control is stopped. In this state, the estimator resets the estimated NOx storage amount NOxSt[k+1] of the NOx storage amount stored in the NSR catalyst to the saturated storage amount after one time (time t=(k+1)Δt).
According to steps S100 and S102, errors are accumulated by repeating the estimation processing shown in FIGS. can be done.

The present aspect has been described above based on the embodiments and modifications, but the above-described embodiments are intended to facilitate understanding of the present aspect, and do not limit the present aspect. This aspect may be modified and modified without departing from the spirit and scope of the claims, and this aspect includes equivalents thereof. Also, if the technical features are not described as essential in this specification, they can be deleted as appropriate.

[Modification 3]
The configurations of the exhaust gas purification system and the catalyst state estimating device of the first to seventh embodiments and the configurations of the exhaust gas cleaning system and the catalyst state estimating device of the modified examples 1 and 2 may be appropriately combined. For example, in the catalyst state estimating devices described in the fourth, fifth, sixth, and seventh embodiments, the catalyst state estimating model having a part of the model described in the second embodiment may be used. Using a catalyst state estimation model including any one or more of the plurality of first models, the plurality of second models, the plurality of third models, the plurality of fourth models, and the plurality of fifth models described in the third embodiment may

1, 1a to 1f... Exhaust purification system 10, 10a to 10f... Catalyst state estimation device 11... CPU
DESCRIPTION OF SYMBOLS 12... Storage part 20... Exhaust purification device 30... Exhaust pipe 52... Flow rate acquisition part 53... Exhaust temperature acquisition part 54... NOx concentration acquisition part 70... NSR catalyst 76... Front end temperature acquisition part 78... Temperature acquisition part 91... Combustion state control Part 92... Internal combustion engine 93... Rich spike control part 110, 110a to 110f... Estimation part 120, 120a to 110d, 110f... Catalyst state estimation model 121, 121b, 121d, 121f... First model 122, 122b, 122d, 122f... Second model 123, 123b, 123d, 123f Third model 124, 124b, 124d, 124f Fourth model 125, 125f Fifth model 159 Temperature estimator

Claims (17)

A catalyst state estimation device,
a first acquisition unit that acquires information about a catalyst that is provided in a main flow path through which exhaust gas flows, and that purifies the exhaust gas by storing harmful substances in the exhaust gas;
a storage unit that stores a catalyst state estimation model in advance;
an estimation unit that estimates the purification performance of the catalyst by applying the information of the catalyst acquired by the first acquisition unit to the catalyst state estimation model;
with
The catalyst state estimation model is
a first model for estimating a purification rate of the harmful substance in the catalyst;
a second model for estimating the reduction amount of the harmful substance in reduction control for reducing the harmful substance stored in the catalyst;
a fourth model for estimating the amount of additives that do not contribute to the reduction reaction in the reduction control;
including
The catalyst state estimation model is
a current purification rate of the hazardous substance estimated by the first model;
the current reduction amount of the harmful substance estimated by the second model;
and the amount of the additive that does not contribute to the current reduction reaction estimated by the fourth model, and
A model for determining the amount of the harmful substance stored in the catalyst at the next time using physical laws,
The estimation unit
The result of estimation by the fourth model is used as an input to the second model to estimate the amount of reduction of the harmful substance , so that it is used as an indirect input to the catalyst state estimation model ,
The result of estimation by the second model is used as a direct input to the catalyst state estimation model,
A catalyst state estimating device that uses the catalyst state estimating model to estimate the storage amount of the harmful substance at the next time as the purification performance of the catalyst. The catalyst state estimation device according to claim 1, further comprising:
a second acquisition unit that acquires information about the exhaust gas flowing into the catalyst;
The estimating unit estimates the purification performance of the catalyst by applying the information on the exhaust gas acquired by the second acquiring unit to the model for estimating the catalyst state, in addition to the information on the catalyst. estimation device. The catalyst state estimation device according to claim 1 or claim 2, further comprising:
A third acquisition unit that acquires information about the additive supplied to the catalyst,
The estimating unit estimates the purification performance of the catalyst by applying the additive information acquired by the third acquiring unit to the catalyst state estimation model in addition to the information on the catalyst. State estimator. The catalyst state estimation device according to any one of claims 1 to 3,
The first model is
Consists of machine learning models,
The information on the catalyst includes the temperature of the front end of the catalyst in the main flow path, the temperature of the catalyst, the amount of nitrogen oxides stored in the catalyst as the harmful substance one hour ago, and the and at least one of the ratio of the storage amount to the saturated storage amount of the nitrogen oxide in the catalyst, and
The output is the conversion rate of the nitrogen oxides in the catalyst,
The catalyst state estimating device, wherein the estimating unit estimates the nitrogen oxide purification rate as the purification performance of the catalyst using the first model. The catalyst state estimation device according to any one of claims 1 to 4,
The second model is
Consists of machine learning models,
In addition to the result of estimation by the fourth model, the information on the catalyst includes the temperature of the front end of the catalyst in the main flow path, the temperature of the catalyst, and the harmful substance occluded in the catalyst one hour ago. and at least one of the storage amount of nitrogen oxide as an input,
outputting the reduction amount of the nitrogen oxides occluded in the catalyst,
The catalyst state estimating device, wherein the estimating unit estimates the reduction amount of the nitrogen oxides as purification performance of the catalyst using the second model. The catalyst state estimation device according to any one of claims 1 to 5,
The fourth model is
Consists of a physical model,
inputting at least one of the temperature of the front end of the catalyst in the main flow path and the temperature of the catalyst as information about the catalyst;
The output is the amount of the additive that does not contribute to the reduction reaction in the reduction control,
The catalyst state estimating device, wherein the estimating unit uses the fourth model to estimate the amount of the additive that does not contribute to the reduction reaction as purification performance of the catalyst. The catalyst state estimation device according to any one of claims 1 to 6,
The first model further includes:
inputting at least one of a temperature of the exhaust gas, a flow rate of the exhaust gas, and an amount of the harmful substance contained in the exhaust gas as information on the exhaust gas;
The second model further includes:
inputting at least one of a temperature of the exhaust gas and a flow rate of the exhaust gas as information on the exhaust gas;
inputting at least one of an inflow amount of the additive flowing into the catalyst and an amount of the additive not contributing to the reduction reaction as the additive information;
The fourth model further includes:
inputting at least one of a temperature of the exhaust gas and an air-fuel ratio of the exhaust gas as information on the exhaust gas;
A catalyst state estimating device, wherein an inflow amount of the additive flowing into the catalyst is input as the additive information. The catalyst state estimation device according to any one of claims 1 to 7,
The catalyst state estimation device, wherein at least part of the first model and the second model is created using teaching data obtained from a plurality of the catalysts having different degrees of deterioration. A method for estimating the state of a catalyst, comprising:
A model for estimating the state of a catalyst, comprising: a first model for estimating a purification rate of a harmful substance in a catalyst that absorbs and purifies the harmful substance in exhaust gas; and a reduction that reduces the harmful substance stored in the catalyst. A storage unit for pre-storing a catalyst state estimation model including a second model for estimating the reduction amount of the harmful substance in the control and a fourth model for estimating the amount of the additive that does not contribute to the reduction reaction in the reduction control. An information processing device comprising
obtaining information about the catalyst;
a step of estimating the purification performance of the catalyst by applying the acquired information of the catalyst to the catalyst state estimation model;
with
The catalyst state estimation model is
a current purification rate of the hazardous substance estimated by the first model;
the current reduction amount of the harmful substance estimated by the second model;
and the amount of the additive that does not contribute to the current reduction reaction estimated by the fourth model, and
A model for determining the amount of the harmful substance stored in the catalyst at the next time using physical laws,
In the estimating step,
The result of estimation by the fourth model is used as an input to the second model to estimate the amount of reduction of the harmful substance , so that it is used as an indirect input to the catalyst state estimation model ,
The result of estimation by the second model is used as a direct input to the catalyst state estimation model,
A method of estimating the storage amount of the harmful substance at the next time as purification performance of the catalyst using the catalyst state estimation model. A computer program,
A model for estimating the state of a catalyst, comprising: a first model for estimating a purification rate of a harmful substance in a catalyst that absorbs and purifies the harmful substance in exhaust gas; and a reduction that reduces the harmful substance stored in the catalyst. A storage unit for pre-storing a catalyst state estimation model including a second model for estimating the reduction amount of the harmful substance in the control and a fourth model for estimating the amount of the additive that does not contribute to the reduction reaction in the reduction control. to an information processing device comprising
a function of acquiring information about the catalyst;
a function of estimating the purification performance of the catalyst by applying the acquired information of the catalyst to the catalyst state estimation model;
and
The catalyst state estimation model is
a current purification rate of the hazardous substance estimated by the first model;
the current reduction amount of the harmful substance estimated by the second model;
and the amount of the additive that does not contribute to the current reduction reaction estimated by the fourth model, and
A model for determining the amount of the harmful substance stored in the catalyst at the next time using physical laws,
The function to estimate is
The result of estimation by the fourth model is used as an input to the second model to estimate the amount of reduction of the harmful substance , so that it is used as an indirect input to the catalyst state estimation model ,
The result of estimation by the second model is used as a direct input to the catalyst state estimation model,
A computer program for estimating the storage amount of the harmful substance at the next time as purification performance of the catalyst using the catalyst state estimation model. The catalyst state estimation device according to claim 1,
The catalyst state estimation model is
A fifth model that estimates an outflow additive amount, which is the amount of the additive that flows out of the catalyst without being consumed in the reduction control, instead of the second model,
The catalyst state estimation model is
Instead of the current reduction amount of the harmful substance estimated by the second model, the reduction amount of the harmful substance in the reduction control obtained from the current outflow additive amount estimated by the fifth model. as an input,
The estimation unit
By using the estimation result of the fourth model as an input to the fifth model to estimate the outflow additive amount, it is used as an indirect input to the catalyst state estimation model ,
A catalyst state estimating device that uses the catalyst state estimating model to estimate the storage amount of the harmful substance at the next time as the purification performance of the catalyst. The catalyst state estimation device according to claim 11,
The fifth model is
Consists of machine learning models,
In addition to the result of estimation by the fourth model, the information on the catalyst includes the temperature of the front end of the catalyst in the main flow path, the temperature of the catalyst, and the harmful substance occluded in the catalyst one hour ago. and at least one of the storage amount of nitrogen oxide as an input,
Outputting the outflow additive amount,
The catalyst state estimation device, wherein the estimation unit estimates the outflow additive amount as purification performance of the catalyst using the fifth model. The catalyst state estimation device according to claim 12, further comprising:
a second acquisition unit that acquires information about the exhaust gas flowing into the catalyst;
The estimating unit estimates the purification performance of the catalyst by applying the exhaust information acquired by the second acquiring unit to the catalyst state estimation model in addition to the information on the catalyst,
The fifth model further inputs at least one of the temperature of the exhaust gas and the flow rate of the exhaust gas as information on the exhaust gas. 14. The catalyst state estimation device according to claim 12 or 13, further comprising:
A third acquisition unit that acquires information about the additive supplied to the catalyst,
The estimating unit estimates the purification performance of the catalyst by applying the additive information acquired by the third acquiring unit to the catalyst state estimation model in addition to the information on the catalyst,
The fifth model further includes, as information on the additive, an inflow amount of the additive that flows into the catalyst and an amount of the additive that does not contribute to the reduction reaction estimated by the fourth model. A catalyst state estimating device having at least one of them as an input. The catalyst state estimation device according to any one of claims 11 to 14,
The catalyst state estimation device, wherein the fifth model is created using training data acquired from a plurality of the catalysts having different degrees of deterioration. The catalyst state estimation device according to any one of claims 1 to 8,
When a plurality of catalysts of the same type are provided in the main flow path,
The first acquisition unit acquires at least information about the catalyst positioned most upstream in the main flow path,
The estimation unit estimates purification performance of the plurality of catalysts as a whole by applying the information of the catalyst acquired by the first acquisition unit to the catalyst state estimation model,
Among the catalyst state estimation models,
The first model is a model for estimating a purification rate of the harmful substance as a whole of the plurality of catalysts,
The second model is a model for estimating a total reduction amount of the harmful substances in the reduction control for each of the plurality of catalysts,
The fourth model is a model for estimating a total amount of additives that do not contribute to a reduction reaction in the reduction control for each of the plurality of catalysts. The catalyst state estimation device according to any one of claims 11 to 15,
When a plurality of catalysts of the same type are provided in the main flow path,
The first acquisition unit acquires at least information about the catalyst positioned most upstream in the main flow path,
The estimation unit estimates purification performance of the plurality of catalysts as a whole by applying the information of the catalyst acquired by the first acquisition unit to the catalyst state estimation model,
Among the catalyst state estimation models,
The first model is a model for estimating a purification rate of the harmful substance as a whole of the plurality of catalysts,
The fifth model is a model that estimates an outflow additive amount, which is the amount of the additive that flows out from the catalyst positioned most downstream without being consumed in the reduction control,
The fourth model is a model for estimating a total amount of additives that do not contribute to a reduction reaction in the reduction control for each of the plurality of catalysts.

Images