Automobile exhaust system temperature prediction and protection control method based on modular modeling

Abstract

The invention relates to the technical field of automobile electronic control, and discloses an automobile exhaust system temperature prediction and protection control method based on modular modeling, which comprises the following steps: configuring an exhaust system modular model, collecting sensor data of an engine exhaust system in real time, and calculating the temperature of the engine exhaust system; calculating a temperature prediction value of each node of the exhaust system based on the modular model; a configurable exhaust system model is constructed by adopting a modular modeling method, exhaust systems of different structures are flexibly adapted, including turbocharging, natural air suction and various types with and without particle traps, and the universality and deployment efficiency of the model are improved; model calculation is driven based on real-time sensor data, dynamic prediction of the temperature of each node of the exhaust system is realized, the problem of insufficient prediction precision of a traditional single model is avoided, and a reliable basis is provided for subsequent protection control.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of automobile electronic control, in particular to a kind of automobile exhaust system temperature prediction and protection control method based on modular modeling. BACKGROUND

[0002] The automobile exhaust system is mainly to discharge the exhaust gas discharged by engine operation, while reducing the exhaust gas pollution and noise reduction; The automobile exhaust system is mainly used for light vehicle, micro vehicle and passenger car, motorcycle and other motor vehicles; The automobile exhaust system refers to the system for collecting and discharging exhaust gas, which is generally composed of exhaust manifold, exhaust pipe, catalytic converter, exhaust temperature sensor, automobile muffler and exhaust tail pipe.

[0003] At present, due to the diversification of the architecture of the automobile exhaust system and the dynamic nature of the working condition, when realizing the exhaust temperature prediction and protection control, the traditional method adopts a fixed structure model, which cannot adapt to different exhaust system configurations such as turbocharging, naturally aspirated, with particle trap and without particle trap, resulting in insufficient model generalization ability.

[0004] Therefore, the present application provides an automobile exhaust system temperature prediction and protection control method based on modular modeling to solve the above problems. SUMMARY

[0005] In view of the deficiencies of the prior art, the present application provides an automobile exhaust system temperature prediction and protection control method based on modular modeling, which solves the problem that the model cannot adapt to different exhaust system configurations, resulting in insufficient model generalization ability.

[0006] To achieve the above purpose, the present application provides the following technical solutions: an automobile exhaust system temperature prediction and protection control method based on modular modeling, comprising the following steps: Configure the exhaust system modular model, define multiple module types through configurable parameters, including pipe module, catalytic converter module, turbine module and particle trap module, adapt to different exhaust system architectures; On the basis of configuring the model, real-time acquisition of engine exhaust system sensor data, including temperature, pressure and flow data, calculation of exhaust system node temperature prediction value based on modular model; After calculating the temperature prediction value, implement dynamic compensation algorithm, temperature prediction compensation for fuel cut condition and scavenging condition, handle the lag and overshoot phenomenon in engine dynamic response; When the predicted temperature value exceeds the preset protection threshold, trigger the protection control logic, adjust the engine operating parameters, including fuel injection amount, boost pressure and air-fuel ratio; After the engine is stopped, predict the residual temperature based on the current model state and execute the cooling control strategy; Meanwhile, through a closed-loop optimization mechanism, the actual temperature data is compared with the prediction results, and the modular model parameters are dynamically updated. The modular model supports a plurality of exhaust assembly combinations and transmits and fuses temperature data through vectorized data routing.

[0007] Preferably, the configuration of the exhaust system modular model comprises the following steps: Define a module type vector to identify and classify key components in the exhaust system, including a pipe module, a catalyst module, a turbine module, a particulate filter module, a muffler module, a sensor interface module, and a connector module; Set up a mapping vector to route and transmit temperature data between modules, including an input vector, an output vector, a state vector, and an error vector; Adjust module properties and connection relationships through configurable parameters to adapt to different exhaust architectures, such as turbocharging, naturally aspirated, with a particulate filter, and without a particulate filter; Establish a modular database to store static parameters and dynamic response characteristics of each module, suitable for initializing the model and supporting real-time calculation.

[0008] Preferably, the real-time collection of sensor data of the engine exhaust system comprises the following steps: Obtain real-time data of key nodes of the exhaust system through a distributed sensor array, including the front end of the turbine, the inlet of the catalyst, the outlet of the particulate filter, and the exhaust tail pipe; The sensor data includes temperature sensor data, pressure sensor data, flow sensor data, and environmental parameter data, wherein the environmental parameter data includes humidity, altitude, and pollutant concentration; Preprocess the collected data, including data cleaning, outlier removal, and normalization processing, to generate a standardized data set; Align the processed data by timestamp and input it into the modular model for calculation.

[0009] Preferably, the implementation of the dynamic compensation algorithm comprises the following steps: For the fuel cut condition, detect the engine fuel cut event, and apply a compensation factor to correct the temperature prediction value, the compensation factor is dynamically adjusted based on the learning results of historical fuel cut event data; For the scavenging condition, analyze the temperature fluctuation characteristics during the scavenging process, and adopt a scavenging compensation strategy, including increasing the response weight of the prediction model and shortening the sampling interval; Quantify the influence coefficient of the working condition change on the temperature prediction, construct a dynamic influence matrix, and identify the priority of high sensitivity conditions; Evaluate the comprehensive effect of multiple conditions through a decision tree algorithm, output the compensation value, and inject it into the modular model.

[0010] Preferably, the trigger protection control logic comprises the following steps: Setting multi-level protection thresholds, including early warning thresholds, critical thresholds and failure thresholds, corresponding to different severity of overheating risks; When the predicted temperature value exceeds the early warning threshold, generating primary control instructions, including adjusting fuel injection amount and optimizing boost pressure; When the predicted temperature value exceeds the critical threshold, activating high-level protection mode, performing fuel cut control and power reduction operation; When the predicted temperature value exceeds the failure threshold, triggering failure diagnosis logic, outputting maintenance warning and recording failure code; Visualizing the protection control instructions and historical effects for comparison, assisting user decision-making and intervention.

[0011] Preferably, the predicted residual temperature and the execution of cooling control strategy comprise the following steps: Based on the modular model state at shutdown, predicting the residual temperature decay curve and outputting cooling control instructions; Applying interpolation algorithm to generate cooling curve to guide the operation parameters of fan and cooling system; Monitoring the residual temperature until it is within the safe range; Comparing the predicted residual temperature with the actual cooling data to optimize the cooling strategy parameters.

[0012] Preferably, the dynamic updating of modular model parameters comprises the following steps: Establishing an error analysis unit to calculate the deviation value between actual temperature data and prediction results; Based on the deviation value, generating optimization instructions to adjust the weight coefficients and connection parameters of the modular model; Using adaptive learning algorithm to update the mapping vector to respond to sudden environmental changes and exhaust system aging: ; Here, represents the learning rate at time t, represents the initial learning rate, represents the decay coefficient, represents the error change, represents the current error value; Building a case library management system to store historical temperature data and protection cases, and calling reference cases to optimize control strategy through similarity matching.

[0013] Preferably, the definition of module type vector comprises: The module type vector adopts a structured array format, supporting at least 7 types of module configuration; Each module type is associated with static property data, including heat capacity coefficient, thermal conductivity and material characteristics; The module connection relationship is achieved through topology mapping, and data flow is seamlessly transmitted. The configurable parameters include a module activation flag, a priority parameter and a fault tolerance parameter.

[0014] Preferably, the influence coefficient of the quantitative working condition change on the temperature prediction includes: An environmental sensitivity matrix is constructed to identify key working condition factors, including throttle opening rate of change, engine speed mutation and environmental temperature fluctuation. The influence coefficient is fitted through regression analysis, and the matrix weight is dynamically updated. In the scavenging condition compensation, a wind speed correction strategy is applied to adjust the model output to adapt to the dust diffusion effect. The decision tree algorithm takes working condition data as input features and temperature prediction error as output target to generate a compensation rule library.

[0015] Preferably, the case library management system includes the following steps: Store historical exhaust temperature case data, including normal working condition cases, fuel cut working condition cases, scavenging working condition cases and fault cases. The feature extraction unit analyzes the case data to generate a feature vector set. Use a similarity matching algorithm to compare the current working condition with historical cases and output a reference control strategy. The decision support unit feeds back the reference strategy to the protection control logic.

[0016] Compared with the prior art, the present application provides a vehicle exhaust system temperature prediction and protection control method based on modular modeling, which has the following beneficial effects: 1. In the present application, when predicting and protecting the temperature of the vehicle exhaust system, a configurable exhaust system model is constructed by using a modular modeling method, which can flexibly adapt to different architectures of the exhaust system, including turbocharging, naturally aspirated, and various types with and without particle traps, improving the universality and deployment efficiency of the model; based on real-time sensor data driving model calculation, the dynamic prediction of the temperature of each node of the exhaust system is realized, avoiding the problem of insufficient prediction accuracy of traditional single models, and providing a reliable basis for subsequent protection control.

[0017] 2. In the present application, when the temperature prediction and protection control of the automobile exhaust system is carried out, the dynamic compensation algorithm is introduced, the temperature prediction value is compensated for the special working conditions such as engine fuel cut and scavenging, the prediction deviation problem caused by engine dynamic response lag and overshoot is solved, and the system can maintain prediction stability under complex working conditions; combined with the multi-level protection threshold mechanism, when the predicted temperature exceeds the threshold, the protection logic is automatically triggered, the fuel injection amount and the supercharging pressure parameter are adjusted in real time, the risk of overheating damage of exhaust system parts is reduced, and the reliability and safety of engine operation are improved.

[0018] 3. In the present application, when the temperature prediction and protection control of the automobile exhaust system is carried out, the actual temperature and the prediction result are continuously compared through the closed-loop optimization mechanism, and the model parameters are dynamically updated, so that the system has self-learning and adaptive ability, can cope with the aging of the exhaust system, environmental mutation and uncertain factors, and can maintain the prediction accuracy for a long time; further, the residual temperature is predicted after the engine is stopped and cooling control is performed, the temperature management of the exhaust system throughout its life cycle is realized, and the robustness and comprehensive protection effect of the system are enhanced. BRIEF DESCRIPTION OF DRAWINGS

[0019] Figure 1 The flowchart of the automobile exhaust system temperature prediction and protection control method based on modular modeling of the present application. DETAILED DESCRIPTION

[0020] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0021] Specific embodiment: an automobile exhaust system temperature prediction and protection control method based on modular modeling, comprising the following steps: Configure the exhaust system modular model, define multiple module types through configurable parameters, including pipe module, catalyst module, turbine module and particulate filter module, and adapt to different exhaust system architectures; On the basis of configuring the model, real-time acquisition of engine exhaust system sensor data is carried out, including temperature, pressure and flow data, and temperature prediction values of each node of the exhaust system are calculated based on the modular model; After calculating the temperature prediction value, a dynamic compensation algorithm is implemented to compensate the temperature prediction for fuel cut and scavenging conditions, and to handle the lag and overshoot phenomenon in the dynamic response of the engine; When the predicted temperature value exceeds the preset protection threshold, the protection control logic is triggered, and the engine operating parameters, including fuel injection amount, supercharging pressure and air-fuel ratio, are adjusted; After engine shutdown, the residual temperature is predicted based on the current model state and a cooling control strategy is executed; At the same time, through a closed-loop optimization mechanism, the actual temperature data is compared with the prediction results, and the modular model parameters are dynamically updated; Among them, the modular model supports multiple exhaust component combinations, and transmits and fuses temperature data through vectorized data routing.

[0022] Configuring the modular model of the exhaust system includes the following steps: Define a module type vector to identify and classify key components in the exhaust system, including pipe modules, catalyst modules, turbine modules, particulate filter modules, muffler modules, sensor interface modules, and connector modules; Set up a mapping vector to route and transfer temperature data between modules, including input vectors, output vectors, state vectors, and error vectors; Adjust module attributes and connection relationships through configurable parameters to adapt to different exhaust architectures such as turbocharging, naturally aspirated, with particulate filter, and without particulate filter; Establish a modular database to store static parameters and dynamic response characteristics of each module, suitable for initializing the model and supporting real-time calculation.

[0023] Real-time acquisition of sensor data of the engine exhaust system includes the following steps: Obtain real-time data of key nodes of the exhaust system through a distributed sensor array, including the front end of the turbine, the inlet of the catalyst, the outlet of the particulate filter, and the exhaust tail pipe; Sensor data includes temperature sensor data, pressure sensor data, flow sensor data, and environmental parameter data, including humidity, altitude, and pollutant concentration; Preprocess the collected data, including data cleaning, outlier removal, and normalization processing, to generate a standardized data set; Align the processed data by timestamp and input it into the modular model for calculation.

[0024] Implementing a dynamic compensation algorithm includes the following steps: For the fuel cut condition, detect the engine fuel cut event and apply a compensation factor to correct the temperature prediction value, and the compensation factor is dynamically adjusted based on the learning results of historical fuel cut event data; Compensation factor calculation: ; Where is the compensation factor, is the maximum compensation reference value, is the decay coefficient, is the fuel cut duration; For scavenging conditions, analyze the temperature fluctuation characteristics during the scavenging process, and adopt scavenging compensation strategies, including increasing the response weight of the prediction model and shortening the sampling interval; Quantify the influence coefficient of condition changes on temperature prediction, construct a dynamic influence matrix, and identify the priority of high sensitivity conditions; The matrix elements are defined as: ; Where is the sensitivity of the temperature node to the condition , , is the partial derivative of temperature with respect to the condition parameter, is the duration weight factor of condition j; Evaluate the combined effect of multiple conditions through a decision tree algorithm, output the compensation value and inject it into the modular model; The implementation steps of the decision tree algorithm are as follows: First, build a binary tree structure, with the throttle opening rate and engine speed mutation as input features, and the temperature prediction error as output target; Select the split node by information gain rate to generate a decision rule library; When the scavenging condition is detected, call the rule library to output the compensation value and inject it into the model.

[0025] The trigger protection control logic includes the following steps: Set multiple protection thresholds, including warning thresholds, critical thresholds and fault thresholds, corresponding to different severity levels of overheating risks; When the predicted temperature value exceeds the warning threshold, generate primary control instructions, including adjusting the fuel injection amount and optimizing the boost pressure; When the predicted temperature value exceeds the critical threshold, activate the advanced protection mode and perform fuel cut control and power reduction operations; When the predicted temperature value exceeds the fault threshold, trigger the fault diagnosis logic, output the maintenance alert and record the fault code; Visually display the protection control instructions and historical effects for comparison to assist user decision-making and intervention.

[0026] Predicting the residual temperature and executing cooling control strategies includes the following steps: Based on the state of the modular model at shutdown, predict the residual temperature decay curve and output the cooling control instructions; Apply interpolation algorithms to generate cooling curves to guide the operating parameters of fans and cooling systems: ; Where is the residual temperature at time , is the shutdown time, , 、 are spline coefficients; Monitoring residual temperature until the safety range; Compare the predicted residual temperature with the actual cooling data, optimize the cooling strategy parameters.

[0027] The dynamic updating of the modular model parameters includes the following steps: An error analysis unit is established to calculate the deviation value of the actual temperature data and the predicted results: ; Wherein is the comprehensive error value, is the weight of the node and the turbine node weight = 2.0, and others = 1.0, is the actual temperature, is the predicted temperature; Based on the deviation value, generate optimization instructions to adjust the weight coefficients and connection parameters of the modular model; Adopting adaptive learning algorithm to update the mapping vector, responding to sudden environmental changes and exhaust system aging: ; Here, denotes the learning rate at time t, denotes the initial learning rate, denotes the decay coefficient, denotes the error change amount, denotes the current error value; Construct a case library management system to store historical temperature data and protection cases, and call reference cases to optimize the control strategy through similarity matching.

[0028] The definition of the module type vector includes: The module type vector adopts a structured array format, supporting at least 7 module type configurations; Each module type is associated with static attribute data, including heat capacity coefficient, thermal conductivity and material properties; The module connection relationship is realized through topological mapping, seamlessly transferring data flow; Configurable parameters include module enable flag, priority parameter and fault tolerance parameter.

[0029] Quantifying the influence coefficient of working condition changes on temperature prediction includes: Build an environmental sensitivity matrix to identify key working condition factors, including throttle opening rate, engine speed mutation and environmental temperature fluctuation; Fit the influence coefficient through regression analysis and dynamically update the matrix weight: ; wherein is a wind speed influence coefficient, is a throttle opening rate, , , is a regression coefficient; In the scavenging condition compensation, a wind speed correction strategy is applied to adjust the model output to adapt to the dust diffusion effect: ; wherein is the corrected dust removal pulse frequency, is a reference frequency, is a wind speed; The decision tree algorithm takes the working condition data as the input feature and the temperature prediction error as the output target to generate a compensation rule library. The specific steps of the decision tree algorithm are as follows: A CART classification tree is constructed, taking the particulate matter type and environmental humidity as features and the dust removal efficiency loss rate as output. The Gini index minimization principle is used to split the nodes to generate a feature importance ranking table, which quantifies the contribution threshold of each pollutant.

[0030] The case library management system includes the following steps: Store historical exhaust temperature case data, including normal working condition cases, oil-off working condition cases, scavenging working condition cases, and fault cases; The feature extraction unit analyzes the case data to generate a feature vector set; Use a similarity matching algorithm to compare the current working condition with historical cases and output a reference control strategy; The steps of the similarity matching algorithm are as follows: Use the dynamic time warping algorithm: align the time series of the current working condition data with the historical cases, calculate the minimum path distance, and set a similarity threshold to output the Top-3 matching cases; The decision support unit feeds back the reference strategy to the protection control logic to realize model self-optimization and knowledge transfer.

[0031] The operation steps of the method are as follows: Step 1, modular modeling system configuration First, a configurable modular modeling system is constructed, which decomposes the exhaust system into independent functional modules through predefined parameters. Specifically, it defines the types of core components of pipe modules, catalyst modules, turbine modules, and particulate traps. Each module is associated with static parameters such as thermal capacity characteristics and material properties. Through topology mapping, data routing logic between modules is established to support flexible adaptation of different architectures such as turbocharging and naturally aspirated. This step forms the basic framework of the model, ensuring the generality and extensibility of subsequent calculations.

[0032] Step two, multi-source data collaborative collection and dynamic preprocessing Real-time capture of key node data in the exhaust system based on distributed sensor networks: Data sources: turbine front section temperature, catalytic converter inlet pressure, particulate filter flow rate, and environmental parameters; Preprocessing mechanism: detect outliers, complete missing data through interpolation, and normalize heterogeneous data to generate a standardized data set aligned in time and space. This step provides high-integrity input for temperature prediction and eliminates noise interference.

[0033] Step three, temperature prediction and dynamic working condition compensation Input standardized data into modular models and perform multi-level calculations: Core prediction: calculate node temperature values based on real-time data streams and identify engine working condition states in synchronization; Dynamic compensation: increase model response weight to offset airflow disturbance for scavenging conditions and inject compensation factors to correct hysteresis effects for fuel cut events. This step improves prediction stability in complex dynamic environments by integrating working condition characteristics and model output.

[0034] Step four, intelligent protection control strategy execution Establish a multi-level threshold protection mechanism: Early warning response: when the predicted temperature exceeds the preset threshold, dynamically adjust parameters such as fuel injection quantity and boost pressure to suppress overheating risks; Emergency protection: if the temperature approaches the critical value, trigger power reduction, fuel cut operation, and force cooling system; Shutdown management: predict the residual temperature decay curve after engine stoppage and control the cooling fan operation time until it reaches a safe range. This step forms a full life cycle protection closed loop to ensure component safety.

[0035] Step five, closed-loop optimization and knowledge transfer Continuous optimization of the system based on actual operation data: Error tracing: compare predicted temperature with actual monitoring values to locate the source of deviation from model parameter lag and environmental mutations; Case library linkage: call historical similar working condition cases to generate reference control strategies through feature matching; Model self-evolution: update module weights according to optimization instructions to enhance the system's ability to cope with aging, pollution, and other uncertain factors. This step enables the system to have continuous learning ability, maintaining high-precision prediction in the long term.

[0036] It is to be understood that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting; it is not intended to exclude myriad other embodiments of the present application that other inventors can develop based on the same general inventive concepts embodied by the described embodiments. That is, although the present application is described in terms of particular embodiments and implementations, it is to be understood that the terminology used is for the purpose of descriptive clarity and that it is intended to be limited only by the words recited in the appended claims. The scope of the present application shall be limited only by the claims.

[0037] While the embodiments of the application have been shown and described herein, it is to be understood that the application is not limited to these embodiments. Rather, numerous modifications are possible without departing from the spirit and scope of the present application as delineated by the claims and their equivalents.

Claims

1. A method for temperature prediction and protection control of an automotive exhaust system based on modular modeling, characterized in that: Includes the following steps: Configure a modular exhaust system model, defining multiple module types through configurable parameters, including pipe modules, catalyst modules, turbine modules, and particulate filter modules, to adapt to different exhaust system architectures; Based on the configuration model, sensor data from the engine exhaust system, including temperature, pressure, and flow data, are collected in real time, and the predicted temperature values ​​of each node in the exhaust system are calculated based on the modular model. After calculating the predicted temperature value, a dynamic compensation algorithm is implemented to perform temperature prediction compensation for fuel cut-off and scavenging conditions, and to handle the lag and overshoot phenomena in the engine dynamic response. When the predicted temperature value exceeds the preset protection threshold, the protection control logic is triggered to adjust the engine operating parameters, including fuel injection quantity, boost pressure and air-fuel ratio. After the engine stops, the residual temperature is predicted based on the current model state and a cooling control strategy is executed. Meanwhile, through a closed-loop optimization mechanism, the parameters of the modular model are dynamically updated by comparing actual temperature data with prediction results. The modular model supports multiple exhaust component combinations and transmits and fuses temperature data through vectorized data routing.

2. The method of claim 1, wherein: The modular model for configuring the exhaust system includes the following steps: Define a module type vector to identify and classify key components in the exhaust system. The module types include pipe modules, catalyst modules, turbine modules, particulate filter modules, muffler modules, sensor interface modules, and connector modules. Set up mapping vectors to route and transmit temperature data between modules, including input vectors, output vectors, state vectors, and error vectors; The module attributes and connection relationships can be adjusted by configurable parameters to adapt to different exhaust architectures, including turbocharged, naturally aspirated, with particulate filter, and without particulate filter. Establish a modular database to store the static parameters and dynamic response characteristics of each module, which is suitable for initializing the model and supports real-time calculation.

3. The method of claim 1, wherein: The real-time acquisition of sensor data from the engine exhaust system includes the following steps: Real-time data of key nodes in the exhaust system are acquired through a distributed sensor array, including the turbine front end, catalytic converter inlet, particulate filter outlet, and exhaust tailpipe. The sensor data includes temperature sensor data, pressure sensor data, flow sensor data, and environmental parameter data, wherein the environmental parameter data includes humidity, altitude, and pollutant concentration; The collected data is preprocessed, including data cleaning, outlier removal, and normalization, to generate a standardized dataset. The processed data is aligned by timestamps and then input into a modular model for calculation.

4. The method of claim 1, wherein: The implementation of the dynamic compensation algorithm includes the following steps: For fuel cut-off conditions, engine fuel cut-off events are detected, and a compensation factor is applied to correct the temperature prediction value. The compensation factor is dynamically adjusted based on the learning results of historical fuel cut-off event data. For scavenging conditions, the temperature fluctuation characteristics during the scavenging process are analyzed, and a scavenging compensation strategy is adopted, including increasing the response weight of the prediction model and shortening the sampling interval. The influence coefficient of changes in working conditions on temperature prediction is quantified, a dynamic influence matrix is ​​constructed, and the priority of highly sensitive working conditions is identified. The combined effect of multiple operating conditions is evaluated using a decision tree algorithm, and the compensation value is output and injected into the modular model.

5. The method for predicting and protecting the temperature of an automotive exhaust system based on modular modeling as described in claim 1, characterized in that: The trigger protection control logic includes the following steps: Set up multi-level protection thresholds, including warning thresholds, critical thresholds and fault thresholds, which correspond to different levels of overheating risk; When the predicted temperature exceeds the warning threshold, a primary control command is generated, including adjusting the fuel injection quantity and optimizing the boost pressure. When the predicted temperature exceeds the critical threshold, the advanced protection mode is activated, and fuel cut-off control and power reduction operations are performed. When the predicted temperature value exceeds the fault threshold, the fault diagnosis logic is triggered, a maintenance alarm is output and a fault code is recorded. Visualize the comparison between protection and control commands and historical effects to assist users in decision-making and intervention.

6. The method for predicting and protecting the temperature of an automotive exhaust system based on modular modeling as described in claim 1, characterized in that: The process of predicting residual temperature and implementing cooling control strategies includes the following steps: Based on the modular model state at the time of shutdown, the residual temperature decay curve is predicted, and cooling control commands are output. Interpolation algorithms are used to generate cooling curves to guide the operating parameters of fans and cooling systems. Monitor residual temperature until it reaches a safe range; By comparing the predicted residual temperature with the actual cooling data, the cooling strategy parameters can be optimized.

7. The method for predicting and protecting the temperature of an automotive exhaust system based on modular modeling as described in claim 1, characterized in that: The dynamic updating of modular model parameters includes the following steps: Establish an error analysis unit to calculate the deviation between the actual temperature data and the predicted results; Optimization instructions are generated based on the deviation values ​​to adjust the weight coefficients and connection parameters of the modular model; An adaptive learning algorithm is used to update the mapping vector in response to sudden environmental changes and exhaust system aging. ； Here, The learning rate represents the time interval t. This represents the initial learning rate. Indicates the attenuation coefficient. Indicates the amount of error change. Indicates the current error value; Build a case library management system to store historical temperature data and protection cases, and use similarity matching to call reference cases to optimize control strategies.

8. The method for predicting and protecting the temperature of an automotive exhaust system based on modular modeling as described in claim 2, characterized in that: The defined module type vector includes: The module type vector adopts a structured array format and supports at least 7 module type configurations; Each module type is associated with static attribute data, including thermal capacity coefficient, thermal conductivity, and material properties; Module connections are achieved through topology mapping, enabling seamless data flow. The configurable parameters include module enable flag, priority parameters, and fault tolerance parameters.

9. The method for predicting and protecting the temperature of an automotive exhaust system based on modular modeling as described in claim 4, characterized in that: The influence coefficient of the quantitative chemical condition change on temperature prediction includes: Construct an environmental sensitivity matrix to identify key operating condition factors, including throttle opening change rate, sudden engine speed change, and ambient temperature fluctuation; The influence coefficients are fitted using regression analysis, and the matrix weights are dynamically updated. In the scavenging operation compensation, a wind speed correction strategy is applied to adjust the model output to adapt to the dust diffusion effect. The decision tree algorithm uses operating condition data as input features and temperature prediction error as the output target to generate a compensation rule base.

10. The method for predicting and protecting the temperature of an automotive exhaust system based on modular modeling as described in claim 7, characterized in that: The construction of the case library management system includes the following steps: Store historical exhaust temperature case data, including normal operating condition cases, oil cut-off operating condition cases, scavenging operating condition cases, and fault cases; The case data is analyzed by the feature extraction unit to generate a feature vector set; By using a similarity matching algorithm, the current operating conditions are compared with historical cases, and a reference control strategy is output. The decision support unit feeds back the reference strategy to the protection control logic.

Images