A method and system for diagnosing the critical state of a three-way catalytic converter

Abstract

This invention discloses a method and system for diagnosing the critical state of a three-way catalytic converter, belonging to the field of vehicle control technology. This invention makes judgments based on the driver's power demand signal. In steady-state control mode, the temperature difference vector between the front and rear exhaust temperature signals is compared with the difference between the front and rear exhaust temperature signals obtained from a calibration table to obtain a first comparison result. In transient control mode, the front, rear, and exhaust temperature signals, as well as the exhaust temperature difference signal, are obtained after processing the operating condition data through a deep learning model. The exhaust temperature difference signal is compared with the value from the calibration table to obtain a second comparison result. The first or second comparison result is set as Terror. If the value of Terror meets a threshold range, the three-way catalytic converter is determined to be in a critical state. This invention can intelligently identify the temperature under dynamic operating conditions without relying on feedback signals from the rear oxygen sensor, thus improving identification efficiency.

Description

Technical Field

[0001] This invention relates to the field of three-way catalytic converter diagnostic and control technology, and in particular to a three-way catalytic converter critical state diagnostic system and method. Background Technology

[0002] Increasingly stringent emission regulations place higher demands on on-board diagnostics (OBD) systems. To ensure the pollutant conversion efficiency of the emission control system, it is crucial to promptly illuminate the malfunction indicator lamp (OBD) before vehicle emissions exceed the OBD threshold, prompting the driver to conduct inspections and repairs. Because the three-way catalytic converter (TWC) of a gasoline engine is affected by temperature aging and mileage, its catalytic conversion efficiency for pollutants continuously declines. In fact, the fundamental reason for this decline in efficiency is the decrease in the oxygen storage capacity (OSC) of the catalytic converter. OSC reflects the ability of the catalytic converter's carrier coating to store oxygen and depends on the activity of the coating material. The electronic control system cannot directly measure the OSC value; it can only estimate the OSC model through calibration during the development phase. Simultaneously, it monitors the OSC to monitor whether emissions exceed the OBD threshold when it drops to a certain value; this point is set as the OBD alarm point.

[0003] The engine of an EREV (Extended Range Hybrid Electric Vehicle) is not always running while driving. This unique characteristic makes it difficult to diagnose the critical state of the catalyst, necessitating the development of a new method and system for diagnosing the critical state of the catalyst.

[0004] The principle of diagnosing the critical state of a catalyst is to estimate the OSC value of the catalyst. If it is lower than a certain value, it is considered that the catalyst no longer has the ability to catalyze the conversion of gaseous pollutants. The OSC depends on factors such as the coating material, formula, coating process, precious metal content and ratio of the TWC. Moreover, as the temperature of the TWC increases, the degree of aging increases, and external pollution continues to deteriorate, there is no accurate model to describe the OSC and its deterioration process. The current technical solution is: the amount of OSC can be calculated by formula (1), where 0.23 represents the mass ratio of oxygen in the air, ml represents the mass flow rate of the engine intake air, λ represents the exhaust air-fuel ratio upstream of the catalytic converter, t1 represents the moment when the air-fuel ratio changes from rich to lean, and t2 represents the moment when the voltage of the downstream oxygen sensor drops.

[0005]

[0006] The basic principle of formula (1) is that when the catalyst has a strong catalytic ability to convert gaseous pollutants, it has a high oxygen storage capacity, that is, the OSC value will be relatively large. The time length t2 minus t1 is the time difference that is measured at the downstream oxygen sensor after active dilution upstream of the catalyst, due to the catalyst's certain oxygen storage capacity. However, as the catalyst becomes contaminated and thermally aged, its oxygen storage capacity will gradually decrease, and eventually it will basically lose its oxygen storage capacity. At this point, the catalyst should be replaced.

[0007] The existing solutions have the following drawbacks:

[0008] 1. According to formula (1), the enrichment and leaning of the air-fuel ratio depends on the accurate response of the rear oxygen sensor. Relying solely on the signal from the rear oxygen sensor may result in deviations, thus affecting the calculation results of the OSC model. In addition, the engine intake air mass flow rate ml used to calculate the OSC uses the model value of the controller, which itself has a large error, and will also affect the accuracy of the calculation results of the OSC model.

[0009] 2. The diagnosis is based on daily driving conditions. The OSC calculation needs to be within a relatively stable window (engine speed and load). If the operating conditions change frequently, the diagnosis will be interrupted often, affecting the timeliness of the diagnosis.

[0010] 3. When the catalyst deteriorates to a critical state, it means that the coating on its carrier has almost lost its catalytic conversion function and the oxygen storage capacity has dropped to a low level. At this time, the exhaust gas entering the catalyst hardly undergoes chemical reaction on the carrier. Therefore, the exhaust temperature at the catalyst outlet will be significantly lower than when the catalyst is in normal condition. The current technical solution does not take this characteristic into account.

[0011] 4. Due to the above facts, there may be abnormal alarms from the OBD system, which may cause customers to complain about economic losses due to premature replacement of the after-processor, or the after-processor may be damaged and the electronic control system may fail to identify and prompt it in time, thus causing air pollution. Summary of the Invention

[0012] In view of the above, the present invention aims to provide a critical state diagnosis method and system for three-way catalytic converters to solve problems such as inaccurate determination of the critical state of EREV catalytic converters, failure to consider the exhaust temperature of the catalytic converter outlet, and abnormal alarms of the OBD system.

[0013] The technical solution adopted in this invention is as follows:

[0014] This invention provides a method for diagnosing the critical state of a three-way catalytic converter, comprising the following steps:

[0015] Based on key input signals, obtain the driver's power demand signal;

[0016] Based on the driver's power demand signal, a steady-state control mode is adopted when the demanded power is less than or equal to the power threshold; and a transient control mode is adopted when the demanded power is greater than the power threshold.

[0017] In steady-state control mode, the temperature difference vector between the front and rear exhaust temperature signals is compared with the difference between the front and rear exhaust temperature signals obtained from the calibration table to obtain the first comparison result. In transient control mode, the front exhaust temperature signal, rear exhaust temperature signal, and exhaust temperature difference signal are obtained after processing the operating point data through a deep learning model. The exhaust temperature difference signal is compared with the difference between the front and rear exhaust temperature signals obtained from the calibration table to obtain the second comparison result.

[0018] The first or second comparison result is set as Terror. If the value of Terror meets a threshold range, the three-way catalytic converter is determined to be in a critical state.

[0019] Optionally, the key input signals include accelerator pedal opening, vehicle speed, battery state of charge, and engine coolant temperature.

[0020] Optionally, the threshold range is set to -25. <Terror<-5。

[0021] Optionally, the front and rear temperature signals in steady-state or transient operating conditions are measured by thermocouples, and the front and rear temperature signals measured by thermocouples are input into a deep learning model for verification.

[0022] Optionally, in steady-state control mode, the front exhaust temperature signal and the rear exhaust temperature signal are scanned at intervals with air-fuel ratios of 1.0, 0.9, 0.8, 1.05 and 1.1, respectively, to obtain the temperature difference vector.

[0023] Optionally, the deep learning model is constructed through the following steps:

[0024] Set the load parameters of the hub to simulate the load of the vehicle, assuming the standard hub parameter is L;

[0025] Under standard hub parameters L, different driving cycles were used to simulate road patterns in vehicle use to obtain the first set of training data.

[0026] The hub parameter L was configured to 1.05*L, 1.1*L, 1.15*L, and 1.2*L respectively. The driving cycle simulation steps were repeated to obtain the second set of training data, the third set of training data, the fourth set of training data, the second set of training data, and the fifth set of training data respectively.

[0027] A deep learning model is trained based on the first set of training data, the second set of training data, the third set of training data, the fourth set of training data, the second set of training data, and the fifth set of training data.

[0028] Optionally, in transient control mode, after processing the operating point data through a deep learning model, the front exhaust temperature signal and the rear exhaust temperature signal are obtained. After interpolating the front exhaust temperature signal and the rear exhaust temperature signal, the exhaust temperature difference signal is compared with the difference between the front exhaust temperature signal and the rear exhaust temperature signal obtained from the calibration table.

[0029] Optionally, in determining whether the three-way catalytic converter is in a critical state, a driving cycle time range is set to N seconds, and the calculation is performed once per second to obtain N operating points. If the count of the three-way catalytic converter being in a critical state is greater than or equal to 0.95*N, then the fault diagnosis result is output.

[0030] The present invention also provides a critical state diagnostic system for a three-way catalytic converter, comprising:

[0031] The acquisition module acquires the driver's power demand signal based on key input signals;

[0032] The judgment module, connected to the acquisition module, is used to make a judgment based on the driver's power demand signal. When the demand power is less than or equal to the power threshold, a steady-state control mode is adopted; when the demand power is greater than the power threshold, a transient control mode is adopted.

[0033] The control module, connected to the judgment module, is configured to: in steady-state control mode, compare the temperature difference vector between the front and rear exhaust temperature signals with the difference between the front and rear exhaust temperature signals obtained from the calibration table to obtain a first comparison result; in transient control mode, after processing the operating point data through a deep learning model, obtain the front exhaust temperature signal, the rear exhaust temperature signal, and the exhaust temperature difference signal, and compare the exhaust temperature difference signal with the difference between the front and rear exhaust temperature signals obtained from the calibration table to obtain a second comparison result;

[0034] The diagnostic module is connected to the control module and sets either the first comparison result or the second comparison result as Terror. If the value of Terror meets a threshold range, the three-way catalytic converter is determined to be in a critical state.

[0035] Optionally, the three-way catalytic converter critical state diagnostic system also includes:

[0036] The early warning module, connected to the diagnostic module, is used to store fault codes indicating that the three-way catalytic converter is in a critical state and to illuminate the engine emission fault light based on the fault diagnosis results output by the diagnostic module.

[0037] The critical state diagnosis method and system for three-way catalytic converters provided by the present invention have the following technical advantages:

[0038] (1) A deep learning model based on multiple sets of training data was established, which can intelligently identify the temperature under dynamic working conditions and simulate the load of the vehicle by using the load parameters of the rotating hub, so that the established deep learning model has better predictive performance and can accurately obtain the values ​​of the front and rear exhaust temperatures.

[0039] (2) Based on the different power demands of drivers, steady-state control mode or transient control mode is adopted, taking into account the different effects of different power demands of drivers and different working modes of the range extender on the critical catalyst diagnosis.

[0040] (3) Deep learning models do not need to rely on the feedback signal of the post-oxygen sensor, which simplifies the model and improves the efficiency of recognition.

[0041] (4) In the process of determining whether the three-way catalytic converter is in a critical state, the length of the driving cycle time range and the calculation frequency are set to realize the fault determination based on probability statistics. A determination method based on 95% confidence is designed, which improves the stability of identification and avoids misdiagnosis. Attached Figure Description

[0042] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described below with reference to the accompanying drawings, wherein:

[0043] Figure 1 This is a flowchart illustrating the critical state diagnosis method for a three-way catalytic converter provided in an embodiment of the present invention.

[0044] Figure 2 This is a schematic diagram of the data acquisition process for the critical state diagnosis method of a three-way catalytic converter provided in an embodiment of the present invention;

[0045] Figure 3 A schematic diagram of a deep learning model for a three-way catalytic converter critical state diagnosis method provided in an embodiment of the present invention. Detailed Implementation

[0046] Embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0047] Current technologies for diagnosing the critical state of three-way catalytic converters rely on estimating the OSC (Oxygen Consumption Capacity) value. A value below a certain threshold is considered insufficient for catalytic conversion of gaseous pollutants. OSC depends on factors such as the coating material, formulation, coating process, and precious metal content and proportion of the TWC (Total Wastewater Coating). Furthermore, as the TWC deteriorates with temperature, aging, and external pollution, there is no precise model to describe OSC and its degradation process. Existing solutions use OSC calculation formulas, but these formulas rely solely on signals from the rear oxygen sensor, which can introduce biases and affect the OSC model's results. Frequent changes in operating conditions can disrupt the diagnosis, impacting timeliness and leading to false alarms in the vehicle's fault diagnosis system. This invention addresses this issue by designing exhaust temperature models for different operating modes of the range extender, based on varying driver power demands. A deep learning network model intelligently identifies temperatures under dynamic conditions, enabling OBD diagnosis of the critical catalytic converter based on a combination of steady-state and transient data.

[0048] This invention proposes an embodiment of a critical state diagnosis method for three-way catalytic converters, specifically, as follows: Figure 1 As shown, it includes the following steps:

[0049] Based on key input signals, obtain the driver's power demand signal, which includes accelerator pedal opening, vehicle speed, battery state of charge and engine coolant temperature.

[0050] Based on the driver's power demand signal, the system determines whether steady-state control is activated when the demanded power is less than or equal to a power threshold, and whether transient control is activated when the demanded power exceeds the power threshold. Specifically, for example... Figure 2 As shown, the power threshold is set to 30KW;

[0051] In steady-state control mode, the temperature difference vector between the front exhaust temperature signal and the rear exhaust temperature signal is compared with the difference between the front exhaust temperature signal and the rear exhaust temperature signal obtained from the calibration table to obtain the first comparison result. In transient control mode, after processing the operating point data through a deep learning model, the front exhaust temperature signal, the rear exhaust temperature signal, and the exhaust temperature difference signal are obtained. The exhaust temperature difference signal is compared with the difference between the front exhaust temperature signal and the rear exhaust temperature signal obtained from the calibration table to obtain the second comparison result. Specifically, the front exhaust temperature vector is obtained based on the front exhaust temperature signal, and the rear exhaust temperature vector is obtained based on the rear exhaust temperature signal.

[0052] Front exhaust temperature vector: Tfront = [Tf1, Tf2, ..., TfN];

[0053] Rear exhaust temperature vector: Tbehind = [Tb1, Tb2, ..., TbN];

[0054] Temperature difference vector: Terror = [Te1,Te2,…,TeN]; defined as Terror = Tbehind – Tfront.

[0055] Based on the operating points in Tables 1 and 2 below, parameter data acquisition of the front and rear exhaust temperature signals of the actual critical catalyst was carried out on the test bench.

[0056] Table 1. Steady-state operating condition front exhaust temperature MAP (°C) @ air-fuel ratio = 1

[0057]

[0058]

[0059] Table 2. Steady-state exhaust temperature MAP (°C) @ air-fuel ratio = 1

[0060]

[0061] Tables 1 and 2 show the front and rear exhaust temperatures when the air-fuel ratio is 1.0, as shown in the coordinates of Table 1. The horizontal axis represents engine speed (rpm), and the vertical axis represents engine torque (Nm). The calibration value is the front exhaust temperature signal; Table 2 shows the rear exhaust temperature signal. The blank spaces in the tables have different values ​​obtained based on the calibration at different engine speeds and load points. During calibration, the air-fuel ratio is also adjusted to 0.9 (10% richer than the stoichiometric air-fuel ratio, resulting in lower front and rear exhaust temperatures), 0.8 (20% richer than the stoichiometric air-fuel ratio, resulting in lower front and rear exhaust temperatures), 1.05 (5% leaner than the stoichiometric air-fuel ratio, resulting in higher front and rear exhaust temperatures), and 1.1 (10% leaner than the stoichiometric air-fuel ratio, resulting in higher front and rear exhaust temperatures), thus obtaining four more sets of calibration tables. During vehicle operation, under the condition of APU startup, the engine speed and torque nodes corresponding to the operating points in Table 1 are selected according to the actual power demand. Then, the temperature difference vector is obtained by scanning at 20-second intervals with air-fuel ratios of 1.0, 0.9, 0.8, 1.05, and 1.1 respectively.

[0062] The first or second comparison result is set as Terror. If the value of Terror meets a threshold range, the three-way catalytic converter is determined to be in a critical state.

[0063] In one embodiment of the present invention, the threshold range is set to -25. <Terror<-5。

[0064] In one embodiment of the present invention, thermocouples are used to measure the front and rear exhaust temperature signals under steady-state or transient operating conditions, and these signals are then input into a deep learning model for verification. Although the front and rear exhaust temperature sensors can measure exhaust temperature signals in real time, thus providing input for the deep learning model's diagnosis, the reliability of exhaust temperature sensors still has some issues. Therefore, this invention uses high-precision thermocouples to measure the front and rear exhaust temperature signals under various operating conditions. The actual measured signals are used for the first step of verifying the exhaust temperature of the deep learning model before being used for the critical catalytic converter diagnosis. Otherwise, the system prompts the user to first check the rationality of the exhaust temperature sensor signal and replace it with a new one before proceeding with the catalytic converter diagnosis. This method of simultaneously acquiring and calculating actual measurement signals and model signals ensures system reliability and avoids misdiagnosis.

[0065] In one embodiment of the present invention, the deep learning model is constructed through the following steps:

[0066] Set the load parameters of the hub to simulate the load of the vehicle, assuming the standard hub parameter is L;

[0067] Under standard hub parameters L, different driving cycles are used to simulate road patterns in vehicle use to obtain the first set of training data. Specifically, driving cycles such as NEDC, WLTC, US06, and HWFET are used to simulate road patterns in vehicle use. NEDC, WLTC, US06, and HWFET are several different test cycles for evaluating vehicle fuel consumption and emission performance. These cycles represent different driving conditions and road conditions in order to more comprehensively evaluate the vehicle's performance in various real-world usage scenarios.

[0068] The hub parameter L was configured to 1.05*L, 1.1*L, 1.15*L, and 1.2*L respectively. The driving cycle simulation steps were repeated to obtain the second set of training data, the third set of training data, the fourth set of training data, the second set of training data, and the fifth set of training data respectively.

[0069] The deep learning model is trained using the first set of training data, the second set of training data, the third set of training data, the fourth set of training data, the second set of training data, and the fifth set of training data.

[0070] Specifically, the deep learning model in this invention is configured as an LSTM (Long Short-Term Memory) deep learning model, specifically, as follows: Figure 3The diagram shows a deep learning model, which includes an input layer, a forward layer, a backward layer, an activation layer, and an output layer. The predicted variables X of the deep learning model are: engine speed, vehicle speed, air volume, coolant temperature, air-fuel ratio, state of charge (SOC), accelerator pedal opening, front oxygen sensor signal, and ignition angle. These predicted variables X are set as input variables X0. t-1 To X T The response variable Y is the front and rear temperature signals. The response variable Y is set as the output layer. t-1 To Y T The activation layer processes the linear output of the previous layer through a non-linear activation function, thereby simulating arbitrary functions and enhancing the network's representational ability. Forward propagation is the process by which the neural network calculates the output from the input layer to the output layer during training. Backward propagation is the process by which the neural network updates its weights based on the error in the output layer during training. The hyperparameters of the deep learning model are shown in Table 3.

[0071] Table 3. Hyperparameter list of the deep learning model for exhaust temperature

[0072] name number of floors Training times Number of neurons Minimum unit Learning rate LSTM 5 50 50 50 0.005

[0073] In deep learning models, hyperparameters are parameters set before the learning process begins, rather than being obtained from training data. These hyperparameters define higher-level concepts of the model, such as complexity or learning capacity, and cannot be learned directly from the data used in the training of standard models; they need to be predefined.

[0074] In one embodiment of the present invention, in transient control mode, after processing the operating condition point data through a deep learning model, the front exhaust temperature signal and the rear exhaust temperature signal are obtained. After interpolating the front exhaust temperature signal and the rear exhaust temperature signal, the exhaust temperature difference signal is compared with the difference between the front exhaust temperature signal and the rear exhaust temperature signal obtained from the calibration table.

[0075] In one embodiment of the present invention, in the process of determining whether the three-way catalytic converter is in a critical state, a driving cycle time range is set to N seconds, and the calculation is performed once per second to obtain N operating points. If the count of the three-way catalytic converter being in a critical state is greater than or equal to 0.95*N, then the fault diagnosis result is output.

[0076] In one embodiment of the present invention, during vehicle operation, the signal waveform of the exhaust temperature downstream of the catalytic converter is compared with that of the deep learning model. The method involves calculating the Euclidean distance between the measured exhaust temperature signal and the model signal. If the distance is less than a calibrated threshold, the catalytic converter is determined to be in a critical state and needs to be replaced promptly. Specifically, the Euclidean distance refers to the actual distance between two points in m-dimensional space. In general engineering practice, the exhaust temperature value is obtained by interpolating the model's operating point (speed / torque) based on the actual operating conditions (speed / torque). The technical solution of this invention calculates the Euclidean distance between the model exhaust temperature and the measured exhaust temperature in three-dimensional space, providing greater accuracy. The calculation formula is:

[0077]

[0078] Where d, n, T, and temp represent the Euclidean distance, rotational speed, torque, and exhaust temperature in three-dimensional space, respectively; the subscripts m and t represent the model value and the test value, respectively.

[0079] Table 4 Catalyst Critical State Determination Threshold (MAP)

[0080]

[0081] As shown in Table 4, at a constant speed, the calibration threshold decreases as the torque increases, because a larger torque means a larger intake volume and a more stable exhaust temperature.

[0082] This invention also provides a critical state diagnostic system for a three-way catalytic converter, comprising:

[0083] The acquisition module acquires the driver's power demand signal based on key input signals;

[0084] The judgment module, connected to the acquisition module, is used to make judgments based on the driver's power demand signal. When the demanded power is less than or equal to the power threshold, a steady-state control mode is adopted; when the demanded power is greater than the power threshold, a transient control mode is adopted.

[0085] The control module, connected to the judgment module, is configured to: in steady-state control mode, compare the temperature difference vector between the front and rear exhaust temperature signals with the difference between the front and rear exhaust temperature signals obtained from the calibration table to obtain a first comparison result; in transient control mode, after processing the operating point data through a deep learning model, obtain the front exhaust temperature signal, the rear exhaust temperature signal, and the exhaust temperature difference signal, and compare the exhaust temperature difference signal with the difference between the front and rear exhaust temperature signals obtained from the calibration table to obtain a second comparison result;

[0086] The diagnostic module is connected to the control module. If the first comparison result or the second comparison result is set to Terror, and the value of Terror satisfies: -25 < Terror < -5, it is determined that the three-way catalytic converter is in a critical state. Within the time range of a driving cycle with a length of N seconds, it is calculated once per second, obtaining N operating points. If the count of the three-way catalytic converter being in a critical state is greater than or equal to 0.95 * N, the fault diagnosis result is output.

[0087] Specifically, the relevant technical features of the three-way catalytic converter critical state diagnosis system of the present invention are described in detail in the corresponding description of the three-way catalytic converter critical state diagnosis method, and will not be elaborated further here. In addition, the three-way catalytic converter critical state diagnosis system of the present invention further includes: a warning module, connected to the diagnostic module, for storing the fault code of the three-way catalytic converter being in a critical state and lighting the engine emission fault lamp based on the fault diagnosis result output by the diagnostic module to prompt the user to perform maintenance in a timely manner.

[0088] The three-way catalytic converter critical state diagnosis method and system provided by the embodiments of the present invention establish a deep learning model based on multiple sets of training data, which can intelligently identify the temperature of dynamic working conditions, simulate the load of the vehicle through the load parameters of the chassis dynamometer, so that the established deep learning model has better prediction performance to accurately obtain the values of the front and rear exhaust gas temperatures; based on the different power requirements of the driver, a steady-state control mode or a transient control mode is adopted, and the different working modes of the range extender bring differences to the diagnosis of the critical catalytic converter, making the diagnosis result more accurate; the deep learning model does not need to rely on the feedback signal of the post-oxygen sensor, simplifies the model, and improves the recognition efficiency; in the process of determining whether the three-way catalytic converter is in a critical state, the length and calculation frequency of the driving cycle time range are set to achieve fault determination based on probability statistics, improve the recognition stability, and avoid misdiagnosis.

[0089] This technical solution is mainly used in the field of OBD diagnosis of the critical catalytic converter of range-extended hybrid vehicles equipped with gasoline engines, and can also be extended to the field of OBD diagnosis of the critical catalytic converter of hybrid vehicles with other configurations equipped with gasoline engines.

[0090] The structure, features and effects of the present invention have been described in detail based on the embodiments shown in the drawings above. However, the above are only the preferred embodiments of the present invention. It should be noted that for the technical features involved in the above embodiments and their preferred modes, those skilled in the art can reasonably combine and match them into various equivalent solutions without departing from and changing the design concept and technical effects of the present invention; Therefore, the present invention is not limited by the scope shown in the drawings. Any changes made according to the concept of the present invention or modified into equivalent embodiments with equivalent changes still fall within the spirit covered by the description and the drawings, and should be within the protection scope of the present invention.

Claims

1. A method for diagnosing the critical state of a three-way catalytic converter, characterized in that, Includes the following steps: Based on key input signals, obtain the driver's power demand signal; Based on the driver's power demand signal, a steady-state control mode is adopted when the demanded power is less than or equal to the power threshold. When the power demand exceeds the power threshold, a transient control mode is adopted. In steady-state control mode, the temperature difference vector between the front and rear exhaust temperature signals is compared with the difference between the front and rear exhaust temperature signals obtained from the calibration table to obtain the first comparison result. In transient control mode, the front exhaust temperature signal, rear exhaust temperature signal, and exhaust temperature difference signal are obtained after processing the operating point data through a deep learning model. The exhaust temperature difference signal is compared with the difference between the front and rear exhaust temperature signals obtained from the calibration table to obtain the second comparison result. The first or second comparison result is set as Terror. If the value of Terror meets a threshold range, the three-way catalytic converter is determined to be in a critical state.

2. The critical state diagnosis method for a three-way catalytic converter according to claim 1, characterized in that, The key input signals include accelerator pedal opening, vehicle speed, battery state of charge, and engine coolant temperature.

3. The critical state diagnosis method for a three-way catalytic converter according to claim 1, characterized in that, The threshold range is set to -25. <Terror<-5。 4. The critical state diagnosis method for a three-way catalytic converter according to claim 1, characterized in that, The front and rear temperature signals were measured by thermocouples under steady-state and transient operating conditions, and the measured front and rear temperature signals were input into a deep learning model for verification.

5. The critical state diagnosis method for a three-way catalytic converter according to claim 1, characterized in that, In steady-state control mode, the front exhaust temperature signal and the rear exhaust temperature signal are scanned at intervals with air-fuel ratios of 1.0, 0.9, 0.8, 1.05 and 1.1 respectively to obtain the temperature difference vector.

6. The critical state diagnosis method for a three-way catalytic converter according to claim 1, characterized in that, The deep learning model is constructed through the following steps: Set the load parameters of the hub to simulate the load of the vehicle, assuming the standard hub parameter is L; Under standard hub parameters L, different driving cycles were used to simulate road patterns in vehicle use, and the first set of training data was obtained. The hub parameter L was configured to 1.05*L, 1.1*L, 1.15*L, and 1.2*L respectively. The driving cycle simulation training steps were repeated to obtain the second set of training data, the third set of training data, the fourth set of training data, the second set of training data, and the fifth set of training data respectively. A deep learning model is trained based on the first set of training data, the second set of training data, the third set of training data, the fourth set of training data, the second set of training data, and the fifth set of training data.

7. The critical state diagnosis method for a three-way catalytic converter according to claim 1, characterized in that, In transient control mode, the front exhaust temperature signal and the rear exhaust temperature signal are obtained after processing the operating condition point data through a deep learning model. After interpolating the front exhaust temperature signal and the rear exhaust temperature signal, the exhaust temperature difference signal is compared with the difference between the front exhaust temperature signal and the rear exhaust temperature signal obtained from the calibration table.

8. The critical state diagnosis method for a three-way catalytic converter according to claim 1, characterized in that, In determining whether the three-way catalytic converter is in a critical state, a driving cycle time range is set to N seconds. The calculation is performed once per second to obtain N operating points. If the count of the three-way catalytic converter being in a critical state is greater than or equal to 0.95*N, the fault diagnosis result is output.

9. A critical state diagnostic system for a three-way catalytic converter, characterized in that, include: The acquisition module acquires the driver's power demand signal based on key input signals; The judgment module, connected to the acquisition module, is used to make a judgment based on the driver's power demand signal. When the demand power is less than or equal to the power threshold, a steady-state control mode is adopted. When the power demand exceeds the power threshold, a transient control mode is adopted. The control module, connected to the judgment module, is configured to: in steady-state control mode, compare the temperature difference vector between the front and rear exhaust temperature signals with the difference between the front and rear exhaust temperature signals obtained from the calibration table to obtain a first comparison result; in transient control mode, after processing the operating point data through a deep learning model, obtain the front exhaust temperature signal, the rear exhaust temperature signal, and the exhaust temperature difference signal, and compare the exhaust temperature difference signal with the difference between the front and rear exhaust temperature signals obtained from the calibration table to obtain a second comparison result; The diagnostic module is connected to the control module and sets either the first comparison result or the second comparison result as Terror. If the value of Terror meets a threshold range, the three-way catalytic converter is determined to be in a critical state.

10. The critical state diagnostic system for a three-way catalytic converter according to claim 9, characterized in that, Also includes: The early warning module, connected to the diagnostic module, is used to store fault codes indicating that the three-way catalytic converter is in a critical state and to illuminate the engine emission fault light based on the fault diagnosis results output by the diagnostic module.

Images