JP2024004934A - Deterioration diagnostic device of exhaust emission control device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a deterioration diagnostic device for an exhaust emission control device capable of calculating an index value contributing to diagnosis of deterioration in the exhaust emission control device by thermal stress.

SOLUTION: A data center 200 being a deterioration diagnostic device of an exhaust emission control device includes a processing circuit 210 and a storage device 220. The storage device 220 stores a machine learned model for calculating maximum temperature difference between two points in a catalyst carrier on the basis of input data composed of time series data for a period of existing length of an explanatory variable including information of an operation state of an engine. The pressing circuit 210 executes estimation processing for inputting the input data to the machine learned model stored in the storage device 220 to calculate the maximum temperature difference, and calculation processing for calculating an index value of a damage of the catalyst carrier so as to be larger as maximum temperature difference becomes large on the basis of the calculated maximum temperature difference.

SELECTED DRAWING: Figure 1

COPYRIGHT: (C)2024,JPO&INPIT

Description

The present invention relates to a deterioration diagnosing device for an exhaust purification device that diagnoses deterioration of an exhaust purification device installed in an exhaust passage of an engine.

Patent Document 1 discloses a diagnostic device for diagnosing an abnormality in an exhaust purification device provided in an exhaust passage of an engine. This diagnostic device calculates the oxygen storage capacity of the catalyst of the exhaust gas purification device as an index value of the performance of the exhaust gas purification device.

Japanese Patent Application Publication No. 2012-31761

By the way, the exhaust purification device is exposed to high temperature exhaust discharged from the engine. Therefore, thermal stress due to temperature changes acts on the catalyst carrier supporting the catalyst. Then, due to repeated temperature changes, damage due to distortion due to thermal stress accumulates on the catalyst carrier. In order to replace the exhaust gas purification device before it breaks down, it is necessary to estimate the damage caused by such thermal stress.

However, the temperature and flow rate of the exhaust gas flowing through the exhaust purification device change depending on the operating state of the engine. Therefore, the temperature distribution in the catalyst carrier also changes depending on the operating state of the engine. Therefore, it has been difficult to diagnose deterioration of exhaust gas purification devices due to such thermal stress.

Below, means for solving the above problems and their effects will be described.
A deterioration diagnosis device for an exhaust gas purification device for solving the above problem is based on input data consisting of time series data for a predetermined length period of an explanatory variable including information on the operating state of the engine. It includes a storage device storing a machine-learned model for calculating the maximum temperature difference between two points separated by a predetermined distance on a catalyst carrier installed in a passageway, and a processing circuit. In this deterioration diagnosis device, the processing circuit performs an estimation process of inputting the input data into the machine learned model stored in the storage device to calculate the maximum temperature difference, and a process of calculating the maximum temperature difference through the estimation process. Based on the maximum temperature difference, a calculation process of calculating a damage index value of the catalyst carrier such that the larger the maximum temperature difference is, the larger the index value of damage to the catalyst carrier is executed.

The greater the temperature difference between two points on the catalyst carrier, the greater the damage caused by thermal strain at that location. Therefore, in the deterioration diagnosis device described above, the maximum temperature difference is calculated, and an index value of damage to the catalyst carrier is calculated based on the maximum temperature difference. Thereby, the deterioration diagnosis device can express the damage accumulated on the catalyst carrier using an index value.

Further, in the deterioration diagnosis device described above, the maximum temperature difference is calculated by a machine learned model using time series data of explanatory variables including information on the operating state of the engine. Therefore, this deterioration diagnosing device can calculate the maximum temperature difference by reflecting the transition of the operating state of the engine. As a result, the deterioration diagnosis device calculates the maximum temperature difference that also reflects the effects of changes in the temperature and flow rate of the exhaust gas flowing into the catalyst carrier. Therefore, according to the deterioration diagnosis device described above, it is possible to calculate the index value of damage to the catalyst carrier by reflecting changes in the temperature distribution of the catalyst carrier depending on the operating state of the engine. That is, this deterioration diagnosis device can calculate an index value that contributes to diagnosis of deterioration of the exhaust gas purification device due to thermal stress.

FIG. 1 is a schematic diagram showing the configuration of a data center, which is an embodiment of a deterioration diagnosis device for an exhaust gas purification device, and a vehicle equipped with the exhaust gas purification device. FIG. 2 is a schematic diagram showing the relationship between an engine mounted on a vehicle and an exhaust purification device. FIG. 3 is a schematic diagram showing an engine control device and devices and sensors connected to the engine control device. FIG. 4 is a schematic diagram of the catalyst carrier for explaining the parts where the temperature difference is calculated. FIG. 5 is a schematic diagram of a long-term short-term memory neural network. FIG. 6 is a flowchart showing the flow of training processing. FIG. 7 is a flowchart showing the flow of a series of processes related to estimation processing. FIG. 8 is a flowchart showing the flow of a series of processes related to calculation processing and notification processing. FIG. 9 is a flowchart showing the flow of a series of processes related to calculation processing and notification processing in a modified example of the deterioration diagnosis device. FIG. 10 is a schematic diagram of a catalyst carrier for explaining a portion where the deterioration diagnosis device calculates a temperature difference in another modification.

A data center 200, which is an embodiment of a deterioration diagnosis device for an exhaust gas purification device, will be described below with reference to FIGS. 1 to 8.
<About the deterioration diagnosis system>
FIG. 1 shows the configuration of a deterioration diagnosis system including a data center 200 according to an embodiment. As shown in FIG. 1, the deterioration diagnosis system includes a data center 200 and a vehicle communication device 140 mounted on a vehicle 500. Data center 200 is communicably connected to vehicle communication device 140 mounted on vehicle 500 via communication network 300 .

<Configuration of data center 200>
As shown in FIG. 1, the data center 200 includes a storage device 220 in which programs are stored, and a processing circuit 210 that executes the programs stored in the storage device 220 to perform various processes. . The data center 200 also includes a communication device 230. The communication device 230 is executed as hardware such as a network adapter, various communication software, or a combination thereof. The communication device 230 is configured to implement wired or wireless communication via the communication network 300.

Note that the data center 200 may be configured using a plurality of computers. For example, data center 200 may be configured with multiple server devices.
<About vehicle 500>
As shown in FIG. 1, a vehicle 500 is equipped with an engine control device 100 and a vehicle communication device 140.

FIG. 2 shows the engine 10 controlled by the engine control device 100. An exhaust pipe 30 is connected to the exhaust manifold 11 of the engine 10. The exhaust manifold 11 and the exhaust pipe 30 constitute an exhaust passage. An exhaust purification device 35 is installed in the middle of the exhaust passage. The exhaust purification device 35 includes a catalyst device 31 and a filter 32.

As shown in FIG. 3, the engine control device 100 includes a storage device 120 in which programs are stored, and a processing circuit 110 that executes the programs stored in the storage device 120 to perform various controls. ing.

Various sensors that detect the state of the engine 10 and the state of the vehicle 500 are connected to the engine control device 100. For example, the engine control device 100 is connected to a vehicle speed sensor 101 that detects a vehicle speed SPD that is the speed of the vehicle. Engine control device 100 acquires information on vehicle speed SPD detected by vehicle speed sensor 101.

A crank position sensor 102 is connected to the engine control device 100 . The crank position sensor 102 outputs a crank angle signal according to a change in the rotational phase of the crankshaft, which is the output shaft of the engine 10. The engine control device 100 calculates the engine rotation speed NE, which is the rotation speed of the crankshaft, based on a detection signal of the rotation angle of the crankshaft input from the crank position sensor 102.

An air flow meter 103 is connected to the engine control device 100 . The air flow meter 103 detects an intake air temperature THA, which is the temperature of air taken into the cylinder through the intake passage of the engine 10, and an intake air amount Ga, which is the mass of the air taken in. The engine control device 100 acquires information on the intake air temperature THA and the intake air amount Ga detected by the air flow meter 103.

Engine control device 100 calculates engine load factor KL based on engine speed NE and intake air amount Ga. The engine load factor KL is an index value of the air filling rate in the combustion chamber of the engine 10. Specifically, the engine load factor KL is the ratio of the amount of inflowing air per one combustion cycle of one cylinder to the reference amount of inflowing air. Note that the reference inflow air amount is variably set according to the engine rotation speed NE.

A water temperature sensor 104 is connected to the engine control device 100 . Water temperature sensor 104 detects water temperature THW, which is the temperature of the cooling water of engine 10 . The engine control device 100 acquires the water temperature THW detected by the water temperature sensor 104.

An air-fuel ratio sensor 105 is connected to the engine control device 100. The air-fuel ratio sensor 105 is installed in the exhaust passage upstream of the exhaust purification device 35. The air-fuel ratio sensor 105 detects the air-fuel ratio of exhaust gas. Engine control device 100 acquires the air-fuel ratio detected by air-fuel ratio sensor 105.

An exhaust temperature sensor 106 is connected to the engine control device 100 . The exhaust gas temperature sensor 106 is installed in a portion of the exhaust passage upstream of the exhaust purification device 35. Exhaust temperature sensor 106 detects exhaust temperature TEX. The engine control device 100 acquires information on the exhaust temperature TEX detected by the exhaust temperature sensor 106.

Engine control device 100 is connected to intake side VVT 12 of engine 10 . The engine control device 100 controls the intake side VVT 12 to adjust the opening/closing timing of the intake valve of the engine 10. The engine control device 100 acquires information on the opening/closing timing of the intake valve from the intake side VVT 12. Engine control device 100 is connected to exhaust side VVT 13. The engine control device 100 controls the exhaust side VVT 13 to adjust the opening/closing timing of the exhaust valve of the engine 10. The engine control device 100 acquires information on the opening/closing timing of the exhaust valve from the exhaust side VVT 13.

The engine 10 includes an EGR mechanism that recirculates a portion of exhaust gas from the exhaust manifold 11 to the intake passage. The engine control device 100 is connected to an EGR valve 14 that adjusts the amount of exhaust gas recirculated to the intake passage. Engine control device 100 controls the EGR valve 14 to adjust the amount of exhaust gas recirculated to the intake passage. The engine control device 100 calculates an EGR rate that indicates the opening amount of the EGR valve 14 as a percentage based on the control amount of the EGR valve 14.

Further, the engine control device 100 is connected to an injector 15 and a spark plug 16 of the engine 10. Engine control device 100 operates injector 15 to control the fuel injection amount. Engine control device 100 operates spark plug 16 to control ignition timing.

As shown in FIG. 3, a vehicle communication device 140 is connected to the engine control device 100. The vehicle communication device 140 is implemented as hardware such as a network adapter, various communication software, or a combination thereof. The vehicle communication device 140 is configured to realize wired or wireless communication via the communication network 300.

<About the catalyst carrier 40>
The catalyst device 31 includes a catalyst carrier 40 supporting a three-way catalyst. As shown in FIG. 4, the catalyst carrier 40 is a ceramic carrier formed into a cylindrical shape, and has a honeycomb structure through which exhaust gas passes.

The exhaust purification device 35 is exposed to high-temperature exhaust gas discharged from the engine 10. Therefore, thermal stress due to temperature changes acts on the catalyst carrier 40. Then, due to repeated temperature changes, damage due to distortion due to thermal stress accumulates on the catalyst carrier 40. In order to replace the catalyst carrier 40 before the exhaust gas purification device 35 breaks down, it is necessary to estimate the damage caused by such thermal stress.

Therefore, in this deterioration diagnosis system, data on the operating state of the engine 10 is transmitted to the data center 200 through the vehicle communication device 140. Then, the data center 200 that has received the data calculates an index value of damage to the catalyst carrier 40.

The storage device 120 of the engine control device 100 stores vehicle speed SPD, engine rotation speed NE, intake air amount Ga, EGR rate, engine load rate KL, ignition timing, water temperature THW, and intake temperature THA while the vehicle 500 is operating. and air-fuel ratio information is accumulated. The storage device 120 also stores information on the operation amount of the intake side VVT 12, the operation amount of the exhaust side VVT 13, and the exhaust temperature TEX. Vehicle communication device 140 transmits these data together with information identifying vehicle 500 to data center 200 each time the data of these information accumulated in storage device 120 reaches a certain amount.

Then, the data center 200 calculates an index value of damage to the catalyst carrier 40 based on input data do. The storage device 220 of the data center 200 stores data of a machine-learned model that calculates the maximum temperature difference ΔTmax between two points on the catalyst carrier 40 separated by a predetermined distance based on the input data X. The data center 200 calculates an index value of damage to the catalyst carrier 40 by integrating the calculated maximum temperature difference ΔTmax.

<About machine learned models>
The data center 200 uses a long-term short-term memory neural network that can handle time-series data while retaining information about changes along the time axis as a machine-learned model. The long-term short-term memory neural network is a so-called LSTM (Long Short-Term Memory) neural network. LSTM neural network is a type of recurrent neural network.

The temperature and flow rate of the exhaust gas flowing into the catalyst carrier 40 change depending on the operating state of the engine 10. Therefore, the data center 200 uses data including information on the operating state of the engine 10 described above as an explanatory variable. Then, the time series data of the explanatory variables is input into the machine learned model. As described above, the explanatory variables include the vehicle speed SPD, the engine speed NE, the intake air amount Ga, the engine load factor KL, the EGR rate, the operation amount of the intake side VVT 12, and the operation amount of the exhaust side VVT 13. The explanatory variables also include the air-fuel ratio, intake temperature THA, water temperature THW, and exhaust temperature TEX. That is, the explanatory variables include these 11 types of information.

In this way, in the data center 200, the above-mentioned 11 types of information are included in the explanatory variables for calculating the output y of the LSTM neural network. The output y is the maximum temperature difference ΔTmax. Note that the maximum temperature difference ΔTmax is the maximum value of the temperature difference ΔT between two points on the catalyst carrier 40 separated by a predetermined distance. In this deterioration diagnosis system, as shown in FIG. 4, a plurality of temperature measurement points are provided on the catalyst carrier 40, and a plurality of data on the temperature difference ΔT between two adjacent temperature measurement points are collected. Then, among all the collected temperature differences ΔT, the maximum value is set as the maximum temperature difference ΔTmax.

In FIG. 4, the direction in which the exhaust gas flows is indicated by a white arrow. That is, in the case of the example shown in FIG. 4, the left side of the catalyst carrier 40 in FIG. 4 is the exhaust upstream side. The right side of the catalyst carrier 40 in FIG. 4 is the downstream side of the exhaust gas. In the example shown in FIG. 4, eight temperature measurement points A to H are set at equal intervals in the circumferential direction along the outer peripheral surface in a portion of the catalyst carrier 40 upstream of the exhaust gas from the center. Furthermore, eight temperature measurement points I to P are set at equal intervals in the circumferential direction along the outer circumferential surface in a portion of the catalyst carrier 40 on the downstream side of the exhaust gas from the center.

In this deterioration diagnosis system, the temperature difference ΔT between the temperature T at the temperature measurement point A and the temperature T at the temperature measurement point B is defined as the temperature difference ΔT_AB between the two points. In this deterioration diagnosis system, the temperature difference ΔT between the temperature T at the temperature measurement point B and the temperature T at the temperature measurement point C is defined as the temperature difference ΔT_BC between the two points. In this deterioration diagnosis system, the temperature difference ΔT between the temperature T at the temperature measurement point C and the temperature T at the temperature measurement point D is defined as the temperature difference ΔT_CD between the two points. In this deterioration diagnosis system, the temperature difference ΔT between the temperature T at the temperature measurement point D and the temperature T at the temperature measurement point E is defined as the temperature difference ΔT_DE between the two points. In this deterioration diagnosis system, the temperature difference ΔT between the temperature T at the temperature measurement point E and the temperature T at the temperature measurement point F is defined as the temperature difference ΔT_EF between the two points. In this deterioration diagnosis system, the temperature difference ΔT between the temperature T at the temperature measurement point F and the temperature T at the temperature measurement point G is defined as the temperature difference ΔT_FG between the two points. In this deterioration diagnosis system, the temperature difference ΔT between the temperature T at the temperature measurement point G and the temperature T at the temperature measurement point H is defined as the temperature difference ΔT_GH between the two points. In this deterioration diagnosis system, the temperature difference ΔT between the temperature T at the temperature measurement point H and the temperature T at the temperature measurement point A is defined as the temperature difference ΔT_HA between the two points.

Further, in this deterioration diagnosis system, the temperature difference ΔT between the temperature T at the temperature measurement point I and the temperature T at the temperature measurement point J is defined as the temperature difference ΔT_IJ between the two points. In this deterioration diagnosis system, the temperature difference ΔT between the temperature T at the temperature measurement point J and the temperature T at the temperature measurement point K is defined as the temperature difference ΔT_JK between the two points. In this deterioration diagnosis system, the temperature difference ΔT between the temperature T at the temperature measurement point K and the temperature T at the temperature measurement point L is defined as the temperature difference ΔT_KL between the two points. In this deterioration diagnosis system, the temperature difference ΔT between the temperature T at the temperature measurement point L and the temperature T at the temperature measurement point M is defined as the temperature difference ΔT_LM between the two points. In this deterioration diagnosis system, the temperature difference ΔT between the temperature T at the temperature measurement point M and the temperature T at the temperature measurement point N is defined as the temperature difference ΔT_MN between the two points. In this deterioration diagnosis system, the temperature difference ΔT between the temperature T at the temperature measurement point N and the temperature T at the temperature measurement point O is defined as the temperature difference ΔT_NO between the two points. In this deterioration diagnosis system, the temperature difference ΔT between the temperature T at the temperature measurement point O and the temperature T at the temperature measurement point P is defined as the temperature difference ΔT_OP between the two points. In this deterioration diagnosis system, the temperature difference ΔT between the temperature T at the temperature measurement point P and the temperature T at the temperature measurement point I is defined as the temperature difference ΔT_PI between the two points.

In this deterioration diagnosis system, the maximum value among the temperature differences ΔT measured at these 16 locations during a period of a predetermined length is set as the maximum temperature difference ΔTmax. That is, in this deterioration diagnosis system, the largest temperature difference ΔT among the temperature differences ΔT at each of a plurality of pairs of a first point and a second point separated by a predetermined distance set on the catalyst carrier 40 is defined as the maximum temperature difference ΔTmax. It is said that Then, the data center 200, which is a deterioration diagnosis device, calculates an index value of damage to the catalyst carrier 40 using the value of this maximum temperature difference ΔTmax.

Note that in this deterioration diagnosis system, a period of 10 minutes after the engine 10 is started is set as a predetermined length period. That is, each time the engine 10 is started, the maximum temperature difference ΔTmax in the catalyst carrier 40 during a period of 10 minutes after the engine 10 is started is calculated.

As described above, vehicle communication device 140 transmits data accumulated in storage device 120 of engine control device 100 to data center 200. Data center 200 receives this data via communication network 300 by communication device 230 . The data center 200 organizes and stores the received data for each vehicle 500 in the storage device 220.

The data center 200 uses this accumulated data as an explanatory variable for calculating the maximum temperature difference ΔTmax used for calculating the index value of damage to the catalyst carrier 40. The explanatory variables include 11 types of information as described above.

In order to calculate the maximum temperature difference ΔTmax, the processing circuit 210 of the data center 200 creates input data X from time series data of explanatory variables acquired during 10 minutes after the engine 10 is started. That is, here, the predetermined length of time is 10 minutes after the engine 10 is started. Note that in vehicle 500, explanatory variables are acquired at predetermined intervals. For example, in the case where explanatory variables are acquired every second in vehicle 500, input data X for 10 minutes is a set of 600 explanatory variables that are continuously collected in chronological order. Specifically, the input data It is a gathering.

Note that each explanatory variable includes 11 types of information as described above. Therefore, each piece of collected data is an 11-dimensional vector made up of these 11 values. Therefore, in this case, the input layer of the neural network that constitutes the machine learned model has 11 nodes.

FIG. 5 schematically shows the configuration of a recurrent neural network. Note that the arrows extending vertically in FIG. 5 indicate the forward propagation direction of the neural network into which explanatory variables are input. Note that "n" in FIG. 5 indicates the chronological order of the explanatory variables in the input data X.

The neural network shown at the right end in FIG. 5 to which the collected data X(n) is input receives the 11-dimensional vector that is the collected data X(n) and outputs the maximum temperature difference ΔTmax as the output y. In this case, collected data X(n) is a neural network into which collected data X(600) is input. This neural network is a fully connected neural network that propagates forward to the output layer that outputs the maximum temperature difference ΔTmax.

As shown in Figure 5, the collected data X (599) collected at the previous timing is input to the hidden layer of the neural network where the collected data X (600) that was last collected in 10 minutes is input. The output of the hidden layer of the neural network is reflected.

The hidden layer of the neural network to which collected data X (599) is input reflects the output of the hidden layer of the neural network to which collected data X (598) collected at the previous timing is input. It has become. In this way, each neural network to which each collected data is input reflects the output of the hidden layer in the neural network to which the collected data collected at the previous timing is input.

An LSTM neural network is a recurrent neural network in which a mechanism called an LSTM block is provided in each hidden layer of the recurrent neural network to adjust the propagation of time-series information. Note that the LSTM block has a cell for keeping errors inside to prevent the gradient from disappearing, an input gate for controlling the input to the cell, an output gate for controlling the output from the cell, and a cell for preventing the gradient from disappearing by keeping the error inside. It consists of an oblivion gate that prevents it from becoming stagnant.

The machine learned model stored in the storage device 220 has undergone supervised learning in advance using training data X_tr including information on actual measurement data of the maximum temperature difference ΔTmax of the catalyst carrier 40. Note that an experimental vehicle equipped with a plurality of sensors that measure the temperature T of each part of the catalyst carrier 40 is used to collect data for creating the training data X_tr. A large amount of data will then be collected while repeating driving experiments using this experimental vehicle in various driving conditions.

<About training processing>
Next, a training process for obtaining a machine learned model used for calculating the maximum temperature difference ΔTmax will be described with reference to FIG. 6. The training data X_tr is generated based on a large amount of data collected from the results of experiments conducted in advance. A computer that executes a neural network training process generates training data X_tr based on the collected data, and trains the neural network based on the training data X_tr.

As shown in FIG. 6, in the training process, the computer first reads time-series data of explanatory variables in step S100. Then, in the next step S110, the computer generates a plurality of training data X_tr from the read time series data. Note that the training data X_tr here is data including information on the actual measurement data of the maximum temperature difference ΔTmax in the above-mentioned set of 600 pieces of collected data. Specifically, collected data, which is an explanatory variable, collected for 10 minutes after the engine 10 is started is substituted into X(1) to X(600) in chronological order. That is, the data substituted into X(1) is the oldest collected data among the 10 minutes' worth of data. The data substituted into X (600) is the latest collected data among the 10 minutes' worth of data. In this way, time series data in the same format as input data X is generated. Then, one training data X_tr is generated by combining the time series data of the explanatory variables for 10 minutes after the engine 10 is started and the actually measured maximum temperature difference ΔTmax as the correct answer label.

Next, the computer advances the process to step S120. Then, among the training data X_tr, data from X(1) to X(600) corresponding to the input data X is input to the LSTM neural network to calculate the maximum temperature difference ΔTmax.

The computer performs learning in the next step S130. Specifically, the weights in the neural network are set so that the error between the maximum temperature difference ΔTmax calculated through the process of step S120 and the actually measured maximum temperature difference ΔTmax, which is the correct label in the training data X_tr used for estimation, is small. Adjust.

The computer repeats the calculation of the maximum temperature difference ΔTmax in the process of step S120 and the adjustment of the weight in the process of step S130. The computer determines that learning is complete when the error in the maximum temperature difference ΔTmax calculated using the LSTM neural network becomes sufficiently small. Then, the trained LSTM neural network data is stored in the storage device 220, and the training process is ended.

The storage device 220 stores trained LSTM neural network data whose weights have been adjusted through the training process. That is, the storage device 220 stores a machine learned model.

<About estimation processing>
Next, a process for estimating the maximum temperature difference ΔTmax executed by the processing circuit 210 will be described. The processing circuit 210 calculates the maximum temperature difference ΔTmax by repeatedly executing the routine shown in FIG. Then, the calculated maximum temperature difference ΔTmax is recorded in the storage device 220.

When the routine shown in FIG. 7 is started, the processing circuit 210 first reads time-series data of explanatory variables stored in the storage device 220 in step S200. Then, in the process of step S210, the processing circuit 210 formats the input data X from the read data. Specifically, the processing circuit 210 assigns the read 600 pieces of collected data for 10 minutes to X(1) to X(600) in chronological order. The processing circuit 210 formats the input data X in which 600 pieces of collected data are arranged in chronological order.

Next, in the process of step S220, the processing circuit 210 inputs the input data X to the machine learned model stored in the storage device 220, and calculates the maximum temperature difference ΔTmax. The process in step S220 is an estimation process that inputs the input data X into the machine learned model stored in the storage device 220 and calculates the maximum temperature difference ΔTmax. After calculating the maximum temperature difference ΔTmax in this way, the processing circuit 210 stores the calculated value of the maximum temperature difference ΔTmax in the storage device 220 in the process of step S230.

The data center 200 repeatedly executes this routine, shapes the input data X at predetermined time intervals, and calculates the maximum temperature difference ΔTmax. Then, the calculated maximum temperature difference ΔTmax is stored in the storage device 220. For example, the data center 200 repeatedly executes this routine every time the engine 10 of the vehicle 500 finishes operating.

<About calculation processing and notification processing>
Next, calculation processing and notification processing executed by the processing circuit 210 will be described with reference to FIG. 8. The processing circuit 210 of the data center 200 executes the routine shown in FIG. 8 every time the maximum temperature difference ΔTmax is stored in the storage device 220 through the process of step S230 in the routine described with reference to FIG.

When the routine shown in FIG. 8 is started, the processing circuit 210 first updates the integrated value ΣΔTmax of the maximum temperature difference ΔTmax in the process of step S300. The initial value of the integrated value ΣΔTmax is "0". In the process of step S300, the processing circuit 210 adds the newly calculated maximum temperature difference ΔTmax to the integrated value ΣΔTmax stored in the storage device 220. Then, the sum is stored in the storage device 220 as a new integrated value ΣΔTmax. Thereby, the processing circuit 210 updates the integrated value ΣΔTmax through the process of step S300. That is, each time the maximum temperature difference ΔTmax is newly calculated, the processing circuit 210 calculates the integrated value ΣΔTmax by integrating the maximum temperature difference ΔTmax. In this deterioration diagnosis system, the data center 200 calculates an integrated value ΣΔTmax as an index value of damage to the catalyst carrier 40. A large maximum temperature difference ΔTmax indicates that large distortion has occurred in the catalyst carrier 40 due to thermal stress. That is, the larger the maximum temperature difference ΔTmax, the more damage will accumulate on the catalyst carrier 40. Then, the maximum temperature difference ΔTmax becomes larger in the integrated value ΣΔTmax. Therefore, the integrated value ΣΔTmax is an index value indicating the magnitude of damage accumulated on the catalyst carrier 40.

In this way, the process of step S300 is a calculation process that calculates the damage index value of the catalyst carrier 40 such that it increases as the maximum temperature difference ΔTmax increases, based on the maximum temperature difference ΔTmax calculated through the estimation process. ing.

After calculating the integrated value ΣΔTmax in this way, the processing circuit 210 advances the process to step S310. In the process of step S310, the processing circuit 210 determines whether the integrated value ΣΔTmax is greater than or equal to the threshold value Xth. The threshold value Xth is a threshold value for determining that damage has accumulated to the catalyst carrier 40 to the extent that the catalyst carrier 40 needs to be replaced based on the fact that the integrated value ΣΔTmax is greater than or equal to the threshold value Xth. The magnitude of the threshold value Xth is set based on the results of experiments conducted in advance.

If the processing circuit 210 determines in the process of step S310 that the integrated value ΣΔTmax is equal to or greater than the threshold value Xth (step S310: YES), the process proceeds to step S320. Then, the processing circuit 210 executes the notification process in the process of step S320. Specifically, in this notification process, the processing circuit 210 performs a process of prompting the user of the vehicle 500 to replace the catalyst carrier 40 for the vehicle 500 for which the integrated value ΣΔTmax is determined to be equal to or greater than the threshold value Xth. Outputs the command to be executed. This command is sent by communication device 230. Then, in the vehicle 500 that receives this command via the communication network 300, a process such as displaying a message prompting the replacement of the catalyst carrier 40 on the display is executed, thereby prompting the user of the vehicle 500 to perform maintenance.

After executing the notification process in this way, the processing circuit 210 temporarily ends this routine.
Further, in the process of step S310, if it is determined that the integrated value ΣΔTmax is less than the threshold value Xth (step S310: NO), the processing circuit 210 ends this routine without executing the process of step S320. . That is, the processing circuit 210 executes the notification process when the integrated value ΣΔTmax becomes equal to or greater than the threshold value Xth. On the other hand, when the integrated value ΣΔTmax is not equal to or greater than the threshold value Xth, the notification process is not executed.

<Action of this embodiment>
In the data center 200, based on the driving state data transmitted from the vehicle 500, the maximum temperature difference ΔTmax is calculated based on input data X, which is time series data of explanatory variables for 10 minutes after the engine 10 is started. Ru. That is, every time the engine 10 finishes operating, the data center 200 calculates the maximum temperature difference ΔTmax in the catalyst carrier 40 for 10 minutes from startup. At this time, the data center 200 inputs the input data X into the machine learned model and calculates the maximum temperature difference ΔTmax.

After calculating the maximum temperature difference ΔTmax, the data center 200 calculates an integrated value ΣΔTmax of the maximum temperature difference ΔTmax as an index value of damage to the catalyst carrier 40. The maximum temperature difference ΔTmax is the maximum value among the temperature differences ΔT at two points separated by a predetermined distance on the catalyst carrier 40. Then, the data center 200 integrates the maximum temperature difference ΔTmax and calculates an integrated value ΣΔTmax as an index value of damage to the catalyst carrier 40.

Thereby, according to the data center 200, damage to the catalyst carrier 40 can be estimated.
<Effects of this embodiment>
(1) The greater the temperature difference ΔT between two points on the catalyst carrier 40, the greater the damage caused by thermal strain at that location. Therefore, in the data center 200, the maximum temperature difference ΔTmax is calculated, and based on the maximum temperature difference ΔTmax, an integrated value ΣΔTmax, which is an index value of damage to the catalyst carrier 40, is calculated. Thereby, the data center 200, which is a deterioration diagnosis device, can express the damage accumulated in the catalyst carrier 40 using an index value.

(2) The data center 200 calculates the maximum temperature difference ΔTmax using time series data of explanatory variables including information on the operating state of the engine 10 and the machine learned model. Therefore, the data center 200 can calculate the maximum temperature difference ΔTmax by reflecting the change in the operating state of the engine 10. As a result, the data center 200 calculates the maximum temperature difference ΔTmax that also reflects the effects of changes in the temperature and flow rate of the exhaust gas flowing into the catalyst carrier 40.

Therefore, according to the data center 200, it is possible to calculate the index value of damage to the catalyst carrier 40 by reflecting changes in the temperature distribution of the catalyst carrier 40 depending on the operating state of the engine 10. That is, the data center 200 can calculate an index value that contributes to diagnosis of deterioration of the exhaust gas purification device 35 due to thermal stress.

(3) The machine learned model calculates the maximum temperature difference ΔT among the temperature differences ΔT at each of the plurality of pairs of the first point and the second point set on the catalyst carrier 40 separated by a predetermined distance. This is a model that calculates ΔTmax. The data center 200 calculates the index value using the largest temperature difference ΔT among the temperature differences ΔT between the plurality of locations where the pair of the first point and the second point is set. Therefore, even if the position where the maximum temperature difference ΔTmax occurs changes depending on the operating state, the index value can be calculated appropriately.

(4) The processing circuit 210 calculates the maximum temperature difference ΔTmax based on the input data X for 10 minutes from the start of the engine 10. Then, by integrating the maximum temperature difference ΔTmax, an integrated value ΣΔTmax, which is an index value of damage, is calculated. After the engine 10 is started, the catalyst carrier 40, which has been cold, is warmed by the exhaust gas. Therefore, a temperature gradient occurs within the catalyst carrier 40. Therefore, the maximum temperature difference ΔTmax in the catalyst carrier 40 tends to become large. That is, according to the data center 200, the index value can be calculated based on information when the maximum temperature difference ΔTmax becomes large and damage due to thermal stress is likely to occur.

(5) When the integrated value ΣΔTmax, which is a damage index value, exceeds the threshold value Xth, the processing circuit 210 executes a notification process to prompt maintenance. Therefore, maintenance can be prompted based on information on the degree of damage accumulation on the catalyst carrier 40.

<Example of change>
This embodiment can be implemented with the following modifications. This embodiment and the following modified examples can be implemented in combination with each other within a technically consistent range.

- In the above embodiment, 11 types of data are used as explanatory variables. On the other hand, the explanatory variables are not limited to these 11 types of data. For example, some of the 11 types of information illustrated in the above embodiment may be omitted. Further, data other than the 11 types of data illustrated may be added to the explanatory variables. For example, the value of the rate of change of the illustrated data may be calculated and the calculated value may be added to the explanatory variables.

- In the above embodiment, an example was shown in which one maximum temperature difference ΔTmax is calculated using one machine learning model. The machine learned model is not limited to the one exemplified in the above embodiment. For example, a machine learned model prepared for each of a plurality of pairs is stored to calculate the maximum temperature difference ΔTmax between a plurality of pairs of a first point and a second point separated by a predetermined distance set on the catalyst carrier 40. It is stored in the device 220. That is, a machine-learned model is prepared for each part for which the temperature difference ΔT is to be calculated. The maximum temperature difference ΔTmax at each site is calculated using each machine learned model. Then, the maximum temperature difference ΔTmax used for calculating ΣΔTmax may be calculated by selecting the largest value among the plurality of calculated maximum temperature differences ΔTmax.

In this case, referring to FIG. 7, the routine shown in FIG. 9 may be executed instead of the routine. In this case, as shown in FIG. 9, the processing circuit 210 executes the process of step S400 prior to the process of step S300. In the process of step S400, the processing circuit 210 selects the largest maximum temperature difference ΔTmax from among the plurality of maximum temperature differences ΔTmax calculated by the plurality of machine learning models. The processing circuit 210 then executes the same processing as steps S300 to S320 in the routine of FIG.

With this configuration, as in the above embodiment, even if the position where the maximum temperature difference ΔTmax occurs changes depending on the operating state, the index value can be appropriately calculated.
- In the above embodiment, an example was shown in which the maximum temperature difference ΔTmax is calculated from the temperature differences ΔT at 16 locations. On the other hand, in order to calculate the maximum temperature difference ΔTmax, the number and positions of the parts for calculating the temperature difference ΔT can be changed as appropriate. For example, the location where the maximum temperature difference ΔTmax occurs varies depending on the shape of the exhaust pipe 30, the layout of the exhaust system, the displacement of the engine 10, the number of cylinders, etc. For example, if the location where the maximum temperature difference ΔTmax occurs can be narrowed down due to the influence of the layout of the exhaust system, the configuration of the engine 10, etc., then the number of locations for calculating the temperature difference ΔT is narrowed down to that location. may be decreased. Furthermore, if the part most susceptible to damage is known in advance, only the maximum temperature difference ΔTmax of that part may be calculated, for example. Further, the number of locations where the temperature difference ΔT is calculated may be increased.

Further, in the above embodiment, an example was shown in which the temperature difference ΔT between a portion of the catalyst carrier 40 on the upstream side of the exhaust gas and a portion on the downstream side of the exhaust gas from the center of the catalyst carrier 40 is calculated. Since exhaust gas flows into the catalyst carrier 40 from the upstream side of the exhaust gas, temperature changes in the catalyst carrier 40 are more likely to occur in the upstream part of the exhaust gas than in the center of the catalyst carrier 40. Therefore, the temperature difference ΔT that can reach the maximum temperature difference ΔTmax may not occur on the downstream side of the exhaust gas from the center of the catalyst carrier 40. In such a case, only the temperature difference ΔT at a portion upstream of the exhaust gas from the center of the catalyst carrier 40 may be calculated.

- An example is shown in which the maximum temperature difference ΔTmax is calculated based on time series data of explanatory variables for 10 minutes after the engine 10 is started, assuming that this is a period in which the maximum temperature difference ΔTmax is likely to become large. The period for which the maximum temperature difference ΔTmax is calculated is not limited to the illustrated one. For example, when a condition in which the maximum temperature difference ΔTmax is likely to become large is satisfied, an estimation process for calculating the maximum temperature difference ΔTmax based on an explanatory variable of a predetermined length may be executed. For example, an example of a condition where the maximum temperature difference ΔTmax is likely to be large is that acceleration is started near a junction with a main road on an expressway. The maximum temperature difference ΔTmax for a predetermined period of time may be calculated based on time-series data of explanatory variables for a predetermined period from when such conditions are satisfied. Further, the predetermined length is not limited to 10 minutes as exemplified in the above embodiment. It may be a longer period or a shorter period.

- In the above embodiment, the difference between the temperature T at the first point and the temperature T at the second point adjacent to each other in the circumferential direction of the catalyst carrier 40 is used as the temperature difference ΔT. On the other hand, the temperature difference ΔT may be any temperature difference between two points separated by a predetermined distance. For example, the temperature difference ΔT at two points separated by a predetermined distance in the radial direction may be used instead of in the circumferential direction.

For example, as shown in FIG. 10, a temperature measurement point a is set a predetermined distance away from the temperature measurement point A on the central axis side of the catalyst carrier 40, and the temperature T at the temperature measurement point A and the temperature T at the temperature measurement point a are set. Let the temperature difference ΔT between the two points be the temperature difference ΔT_Aa between the two points. Similarly, regarding the temperature measurement points B to P, the temperature measurement points are set at a predetermined distance apart from each other on the central axis side of the catalyst carrier 40, and the temperature difference between the two points is determined. In addition, in FIG. 10, dots are attached within a predetermined distance range from the first temperature measurement point to the central axis side to indicate the depth of the location where the second point is located. In this manner, the temperature differences between the two points can be ΔT_Aa, ΔT_Bb, ΔT_Cc, ΔT_Dd, ΔT_Ee, ΔT_Ff, ΔT_Gg, and ΔT_Hh. Similarly, the temperature differences between two points can be ΔT_Ii, ΔT_Jj, ΔT_Kk, ΔT_Ll, ΔT_Mm, ΔT_Nn, ΔT_Oo, and ΔT_Pp.

Even when the temperature difference ΔT between two points is set in this way, the machine learned model can be trained in the same way as in the above embodiment to obtain the machine learned model for calculating the maximum temperature difference ΔTmax. can. The data center 200 can then use the machine learned model to calculate the integrated value ΣΔTmax, which is an index value of damage to the catalyst carrier 40.

- In the above embodiment, an example is shown in which the integrated value ΣΔTmax obtained by integrating the maximum temperature difference ΔTmax is used as an index value of damage to the catalyst carrier 40. On the other hand, the damage index value is not limited to the integrated value ΣΔTmax. The index value may be a value calculated such that it increases as the maximum temperature difference ΔTmax increases.

- Instead of the notification process, the data center 200 may execute a restriction process that outputs a command to the engine control device 100 to limit the rate of change of the intake air amount Ga per unit time. By executing such restriction processing, changes in the flow rate and temperature of exhaust gas can be suppressed. As a result, the temperature difference ΔT in the catalyst carrier 40 can be suppressed and damage accumulation can be suppressed. Note that if vehicle 500 is a hybrid vehicle, when the rate of change in intake air amount Ga is limited by such a limiting process, the output of the motor may be changed to suppress changes in acceleration. Further, the data center 200 may execute such a restriction process in addition to the notification process.

- In the above embodiment, the data center 200 is illustrated as the deterioration diagnosis device. The deterioration diagnosis device may be mounted on vehicle 500. In this case, for example, the data of the machine learned model is stored in the storage device 120 of the engine control device 100 mounted on the vehicle 500. Then, the processing circuit 110 of the engine control device 100 executes estimation processing, calculation processing, and notification processing.

- In the above embodiment, the data center 200, which is a deterioration diagnosis device, includes a processing circuit 210 and a storage device 220, and executes software processing. However, this is only an example. For example, the deterioration diagnosis device may include a dedicated hardware circuit (for example, an ASIC) that processes at least a portion of the software processing executed in the above embodiments. That is, the deterioration diagnosis device may have any of the following configurations (a) to (c). (a) The deterioration diagnosis device includes an execution device that executes all processes according to a program, and a storage device that stores the program. That is, the deterioration diagnosis device includes a software execution device. (b) The deterioration diagnosis device includes an execution device that executes a part of processing according to a program, and a storage device. Furthermore, the deterioration diagnosis device includes a dedicated hardware circuit that performs the remaining processing. (c) The deterioration diagnosis device includes a dedicated hardware circuit that executes all processing. Here, there may be a plurality of software execution devices and/or dedicated hardware circuits. That is, the processing may be performed by processing circuitry comprising one or more software execution devices and/or one or more dedicated hardware circuitry. The storage device or computer readable medium for storing the program includes any available medium that can be accessed by a general purpose or special purpose computer.

The above-described embodiments and modified examples include configurations described in the following supplementary notes.
[Additional note 1] Based on input data consisting of time-series data for a period of a predetermined length of an explanatory variable that includes information on the operating state of the engine, The apparatus includes a storage device storing a machine learned model for calculating a maximum temperature difference between two points, and a processing circuit, the processing circuit converting the input data into the machine learning model stored in the storage device. an estimation process that calculates the maximum temperature difference by inputting the data into a completed model, and an index of damage to the catalyst carrier that increases as the maximum temperature difference increases. A deterioration diagnosis device for an exhaust purification device that performs a calculation process for calculating a value.

[Additional Note 2] The machine-learned model determines the largest temperature difference among the temperature differences at each of the plurality of pairs of the first point and the second point set on the catalyst carrier and separated by the predetermined distance, as the maximum temperature difference. The deterioration diagnosis device for an exhaust gas purification device according to [Appendix 1], which is a model that calculates as a temperature difference.

[Additional note 3] The storage device stores a plurality of pairs of points set on the catalyst carrier to calculate the maximum temperature difference between the plurality of pairs of the first point and the second point separated by the predetermined distance. The machine learned model for each pair is stored, and the calculation process calculates the largest value among the maximum temperature differences for each pair calculated by the machine learned models through the estimation process. The deterioration diagnosis device for an exhaust gas purification device according to [Appendix 1] which is selected and used for calculating the index value.

[Additional note 4] The processing circuit performs the estimation process of calculating the maximum temperature difference based on the input data for the predetermined length period from the start of the engine, and integrates the maximum temperature difference. The deterioration diagnosis device for an exhaust gas purification device according to any one of [Appendix 1] to [Appendix 3], which executes the calculation process of calculating the index value.

[Additional Note 5] When the index value becomes equal to or greater than a threshold value, the processing circuit executes a notification process to prompt maintenance. Diagnostic equipment.

10...Engine 11...Exhaust manifold 12...Intake side VVT
13...Exhaust side VVT
14... EGR valve 15... Injector 16... Spark plug 30... Exhaust pipe 31... Catalyst device 32... Filter 35... Exhaust purification device 40... Catalyst carrier 100... Engine control device 101... Vehicle speed sensor 102... Crank position sensor 103... Air flow meter 104 ...Water temperature sensor 105...Air-fuel ratio sensor 106...Exhaust temperature sensor 110...Processing circuit 120...Storage device 140...Vehicle communication device 200...Data center 210...Processing circuit 220...Storage device 230...Communication device 300...Communication network 500...Vehicle

Claims (5)

Based on input data consisting of time series data for a period of a predetermined length of an explanatory variable that includes information on the operating state of the engine, the maximum of two points separated by a predetermined distance on the catalyst carrier installed in the exhaust passage of the engine is determined. It is equipped with a storage device that stores a machine-learned model for calculating temperature differences, and a processing circuit.
The processing circuit calculates the maximum temperature difference by inputting the input data into the machine learned model stored in the storage device, and the processing circuit calculates the maximum temperature difference based on the maximum temperature difference calculated through the estimation process. A deterioration diagnosing device for an exhaust gas purification device, wherein the larger the maximum temperature difference, the larger the index value of damage to the catalyst carrier. The machine learned model calculates, as the maximum temperature difference, the largest temperature difference among the temperature differences at each of the plurality of pairs of the first point and the second point set on the catalyst carrier and separated by the predetermined distance. The deterioration diagnosis device for an exhaust gas purification device according to claim 1, which is a model that performs the following. The storage device stores the machine for each of a plurality of pairs of points set on the catalyst carrier to calculate the maximum temperature difference at each of the plurality of pairs of the first point and the second point separated by the predetermined distance. The trained model is memorized,
The calculation process selects the largest value among the maximum temperature differences for each of the plurality of pairs calculated by the plurality of machine-learned models through the estimation process and uses the largest value to calculate the index value. A deterioration diagnosis device for an exhaust purification device as described in . The processing circuit,
the estimation process of calculating the maximum temperature difference based on the input data for a period of the predetermined length from the start of the engine;
The deterioration diagnosis device for an exhaust gas purification device according to claim 1, wherein the calculation process calculates the index value by integrating the maximum temperature difference. The deterioration diagnosis device for an exhaust gas purification device according to claim 1, wherein the processing circuit executes a notification process for prompting maintenance when the index value becomes equal to or greater than a threshold value.

Images