Overload protection control method and system for commercial vehicle engine

Abstract

The invention belongs to the technical field of vehicle control, and particularly relates to a commercial vehicle engine overload protection control method and system, and the method comprises the steps: S1, deploying a distributed sensor network, combining the preprocessing and feature extraction of S2, and a weighted feature fusion algorithm of S3, and integrating multi-source data into a unified engine health state vector. The deep fusion mechanism can capture fine state changes under high-load and multi-scene switching, misjudgment caused by local data distortion is avoided, and the comprehensiveness and reliability of state sensing are remarkably improved. According to the method, the dynamic overload risk index is calculated through the two-factor model in the step S4, and the future temperature trend is predicted based on the road condition matching algorithm in the step S5, so that the protection threshold is dynamically adjusted along with the working condition, the hysteresis of the fixed threshold is avoided, and the performance loss under the extreme working condition is reduced.

Description

Technical Field

[0001] This invention belongs to the field of vehicle control technology, specifically relating to an overload protection control method and system for a commercial vehicle engine. Background Technology

[0002] As the core of a commercial vehicle's power system, the long-term reliable operation of the engine directly determines the safety and economy of transportation operations, and overload protection is a key technology to ensure the engine's service life. Traditional commercial vehicle engine overload protection systems mostly rely on fixed thresholds for a single physical quantity for judgment, typically monitoring only isolated parameters such as coolant temperature and oil pressure. When a parameter exceeds a preset threshold, it triggers a simple power reduction or alarm action.

[0003] This single-threshold protection mode has significant limitations: on the one hand, it cannot integrate multi-dimensional information such as temperature change rate, load fluctuation, and environmental conditions. Under complex working conditions with high load and multiple scene switching, it is easy to cause protection delays or false triggers due to local data distortion. For example, when climbing continuously in mountainous areas, it is difficult to predict the temperature rise trend by only monitoring the coolant temperature, which can easily lead to failures such as piston cylinder scoring. Summary of the Invention

[0004] To address the aforementioned shortcomings of the prior art, this invention provides an overload protection control method, system, terminal, and storage medium for commercial vehicle engines.

[0005] In a first aspect, the present invention provides an overload protection control method for a commercial vehicle engine, comprising: S1. By using sensors deployed in the engine compartment and body, the engine block temperature, oil pan area temperature, coolant flow / temperature, oil pressure, ambient temperature, vehicle speed, real-time load, throttle opening and driving operation signals are collected. S2. Preprocess the multimodal data collected in step S1, and extract the time-domain features of temperature change rate and load fluctuation amplitude from the preprocessed multimodal data to output a standardized dataset. S3. The standardized dataset output in step S2 is weighted and fused using a weighted feature association algorithm to obtain the engine health status vector. S4. Input the engine health state vector obtained in step S3 into the pre-trained two-factor model to calculate the overload risk index. S5. Combining the overload risk index output in step S4 and the vehicle speed and throttle opening data collected in step S1, the current driving scenario of the commercial vehicle is identified through the pre-stored road condition matching algorithm, and the temperature rise trend is predicted within a preset time in the future. S6. Combining the overload risk index output in step S4 and the temperature rise trend output in step S5, generate graded protection instructions. S7. Based on the graded protection instructions, the engine cooling system is linked to adjust the commercial vehicle's fan or water pump, fuel supply module, and driver's cabin alarm unit to execute corresponding protection actions.

[0006] Further improvements to this technical solution include step S1, which includes: S11. Install a first thermocouple in the main bearing housing area of ​​the engine block to obtain the cylinder block temperature. A second thermocouple is installed at the midpoint of the oil level in the oil pan to obtain the oil temperature. A turbine flow meter is connected in series in the coolant line to obtain the flow rate. With temperature A piezoresistive sensor is installed in the main oil passage to obtain the oil pressure. The ambient temperature is obtained by installing a thermal resistor in the area of ​​the vehicle body longitudinal beam where there is no heat source interference. A Hall effect sensor is installed on the drive shaft to obtain the vehicle speed v; a strain gauge load cell is installed between the leaf spring and the frame to obtain the real-time load m; an angular displacement sensor is installed on the accelerator pedal shaft to obtain the throttle opening θ; and microswitches are installed on the steering column and brake pedal to obtain driving operation signals. Furthermore, a channel index i is established for the above nine types of parameters, and an independent analog-to-digital conversion gain coefficient is configured for each channel. Where i = 1 to 9 correspond to the parameters mentioned above, and are driving operation signals. Includes coded values ​​for steering, braking, and gear shifting actions; S12. Obtain the current engine speed n via the CAN bus, and determine the sampling frequency according to the physical change rate characteristics of each parameter. Temperature parameters are sampled at low frequency, pressure and flow parameters at medium frequency, and vehicle speed and throttle opening parameters at high frequency. The sampling frequency is calculated using the following formula: ; in, This represents the data acquisition frequency of the i-th channel; n is the real-time engine speed. These are low-frequency sampling coefficients; These are the intermediate frequency sampling coefficients; These are high-frequency sampling coefficients; S13, based on sampling frequency The collected raw data undergoes a triple verification process: physical threshold judgment, gradient mutation detection, and sensor fault code verification. Data that passes verification is stored in a specified length. Circular cache queue Cache queue length The calculation formula is: ; in, This represents the total number of data points that the i-th channel buffer queue can hold; η is the cache time window; η is the safety factor.

[0007] Further improvements to this technical solution include step S2, which includes: S21. From the established nine-channel circular buffer queue In the process, continuous valid data points are selected based on the data quality label q=0, and the extraction length is [length missing]. The data segment, and based on the sampling frequency of each channel. With the smallest time base unit Perform linear interpolation alignment to construct a 9×M dimensional synchronization data matrix. ;in, The highest sampling frequency among the nine channels; extraction length The calculation formula is: ; in, This represents the number of data points extracted from the i-th channel; For feature extraction time window; synchronized data matrix The number of columns M is equal to the minimum extraction length min among all channels. ); S22. Synchronize the data matrix in step S21. The three rows of data corresponding to cylinder block temperature, oil temperature, and coolant temperature are used to calculate the first-order backward difference point by point according to the time series, so as to obtain the instantaneous rate of change of each temperature parameter within the feature extraction time window, and then synthesize a three-dimensional temperature gradient vector. The formula for calculating the cylinder block temperature change rate is: ; in, This represents the rate of change of cylinder block temperature at time t; The cylinder block temperature at time t; For a unified sampling interval after synchronization, the temperature gradient vector The j-th component corresponds to the average rate of change of the three temperatures, i.e. ; S23, Synchronization data matrix The corresponding line of data for the real-time load is in the format of a line with a length of [length missing]. The standard deviation of the load data within the sliding window is calculated segment by segment as the load fluctuation amplitude characteristic Am, and the load dynamic stability index is calculated. The formula for calculating the magnitude of load fluctuations to quantify their severity is as follows: ; in, This represents the load fluctuation range at time t; This represents the load value at a historical moment. The width of the sliding window; The average load within this window is used to calculate the load dynamic stability index. ; in, It is a dimensionless load dynamic stability index; This represents the load fluctuation range within the current time window. γ represents the baseline fluctuation range under the vehicle's calibration unloaded state; γ is the load correction factor. S24. The original data matrix obtained in step S21 is... With the extracted temperature gradient vector Load fluctuation range Am and stability index The vectors are concatenated to form a 12-dimensional feature vector, and deviation standardization based on historical extreme values ​​is performed on each feature dimension to finally output a standardized dataset. The deviation standardization mapping formula is: ; in, These are the standardized eigenvalues; These are the original feature values ​​or the extracted time-domain feature values. and These are the maximum and minimum values ​​of this type of feature, respectively.

[0008] Further improvements to this technical solution include step S3, which includes: S31. Based on the data cleaned in S24, core parameters are selected using a feature importance ranking algorithm. For temperature-related features, cylinder block temperature, coolant temperature, and exhaust temperature change rate are retained; for mechanical status features, oil pressure, crankshaft speed, and piston vibration amplitude are retained; for environmental features, ambient temperature, atmospheric pressure, and humidity are retained. Principal component analysis is then performed on similar features. Two principal components are extracted for the temperature category, and sub-weights are assigned based on the variance contribution rate after calculating the scores of each principal component. The principal component score calculation is based on a linear transformation formula: ; in, This represents the score of the k-th principal component; This represents the loading coefficient of the k-th principal component on the i-th original feature; This represents the i-th original feature value; n represents the number of features in this class; the weight allocation formula is based on the variance contribution rate: ; in, This represents the weight of the k-th principal component; The eigenvalues ​​of the k-th principal component are represented by m; m represents the total number of principal components. S32. Based on the principal component scores and weights from step S31, a weighted feature association algorithm is used to multiply the comprehensive score of the temperature category by a weight of 0.4, the mechanical state category by a weight of 0.3, and the environmental category by a weight of 0.3, and then concatenate them in dimensional order to form an engine health state vector.

[0009] Further improvements to this technical solution include step S4, which includes: The engine health state vector output from step S3 is input into a pre-trained two-factor model. The overload risk index is calculated through weight calculation and dynamic threshold comparison. ; in, Indicates the overload risk index; This represents the threshold deviation weighting coefficient; D represents the normalized threshold deviation. , The dynamic protection threshold for the q-th core parameter in the engine health state vector. This represents the mapping value of the q-th core parameter in the engine health state vector. , This represents the original value of the q-th core parameter in the engine health state vector. This is the maximum value of the core parameter; This represents the weighting coefficient for the acceleration due to temperature change; A represents the acceleration due to temperature change. , This represents the mapping value of the core temperature parameters at the current moment. These are the mapping values ​​of the core temperature parameters at the first and second sampling times, respectively. This is the scaling factor; This indicates the normalized oil pressure.

[0010] Further improvements to this technical solution include step S5, which includes: S51. Based on the vehicle speed sequence and throttle opening change data collected in real time in step S1, driving characteristic parameters are extracted through a sliding window with a length of 60 seconds. The driving characteristic parameters include the average vehicle speed fluctuation coefficient, the standard deviation of throttle opening, and the shift frequency. S52. Perform Euclidean distance matching between the extracted driving feature parameters and the pre-stored road condition feature library, and select the scene with the smallest distance and similarity ≥ 0.75 as the current driving scene; the road condition feature library contains standard feature vectors of three typical scenes: highway, national highway and mountain road; S53. Based on the current driving scenario identified in step S52 and the overload risk index output in step S4, the temperature rise coefficient matrix of the corresponding scenario is called. The temperature increase per unit time is calculated through piecewise linear interpolation, and then an ambient temperature correction term is added to generate a temperature trend curve for the next 5 minutes. The temperature rise trend prediction formula is: ; in, This represents the predicted temperature at a future time Δt. Indicates the current measured temperature; Indicates the road condition coefficient; Indicates the load factor; Indicates the weight of throttle opening; Indicates the prediction time step; Indicates ambient temperature.

[0011] Further improvements to this technical solution include step S6, which includes: S61. Combining the overload risk index RI output in step S4 and the temperature rise trend output in step S5. The protection level is determined through a tiered triggering mechanism: when the RI is in the 60-75 range, a warning level protection is activated; when the RI is in the 75-90 range, an intervention level protection is activated; when the RI exceeds 90 or... Forced protection will be activated when the temperature is predicted to exceed 130℃ within 5 minutes. ; in, Indicates the protection level; RI indicates the overload risk index; t represents the predicted temperature for the next 5 minutes; t represents the current time. S62. Based on the protection level determined in step S61, generate corresponding control command parameters: For the warning level, the red indicator light on the control panel flashes and the buzzer emits an intermittent warning sound; for the intervention level, send a command to the cooling system via the CAN bus to force the cooling fan to enter full-speed operation mode and increase the flow rate of the cooling pump to 120% of the rated value; for the forced level, the fuel injection system reduces the fuel injection quantity by 15%-20% by adjusting the injection pulse width, while limiting the maximum speed to no more than 2200 rpm.

[0012] Secondly, the present invention provides an overload protection control system for a commercial vehicle engine, comprising: The sensor array, deployed in the engine compartment and body, is used to collect engine block temperature, oil pan area temperature, coolant flow / temperature, oil pressure, ambient temperature, vehicle speed, real-time load, throttle opening and driving operation signals. The preprocessing module, which communicates with the sensor array, is used to preprocess multimodal data, extract the time-domain features of temperature change rate and load fluctuation amplitude, and output a standardized dataset. The weighted fusion module, which communicates with the preprocessing module, is used to perform weighted fusion on the standardized dataset using a weighted feature association algorithm to obtain the engine health status vector. The risk calculation module, which communicates with the weighted fusion module, is used to input the engine health state vector into the pre-trained two-factor model to calculate the overload risk index. The scene recognition and prediction module communicates with the risk calculation module and sensor array. It is used to combine the overload risk index and vehicle speed and throttle opening data, and identify the current driving scene of the commercial vehicle through a pre-stored road condition matching algorithm to predict the temperature rise trend within a preset time in the future. The instruction generation module communicates with the risk calculation module and the scene recognition and prediction module to generate graded protection instructions by combining the overload risk index and the temperature rise trend. The execution control module is connected to the command generation module and is used to link the engine cooling system with the graded protection command to adjust the fan or water pump, control the fuel supply module and the cockpit alarm unit, and execute the corresponding protection actions.

[0013] Thirdly, the present invention provides a terminal, comprising: Processor, memory, among which, This memory is used to store computer programs. The processor is used to retrieve and run the computer program from memory, causing the terminal to perform the terminal method described above.

[0014] Fourthly, the present invention provides a computer storage medium storing instructions that, when executed on a computer, cause the computer to perform the methods described in the above aspects.

[0015] The beneficial effects of this invention are as follows: Traditional protection systems monitor only isolated parameters such as coolant temperature or oil pressure, making it difficult to comprehensively reflect the engine's status under complex operating conditions. This invention deploys a distributed sensor network (e.g., cylinder block temperature, oil pan temperature, ambient temperature, load, and other multi-dimensional parameters) in step S1, combined with preprocessing and feature extraction (e.g., temperature change rate, load fluctuation amplitude) in step S2, and a weighted feature fusion algorithm (e.g., principal component analysis and cross-modal weight allocation) in step S3, integrating multi-source data into a unified engine health status vector. This deep fusion mechanism can capture subtle state changes under high load and multi-scenario switching, avoiding misjudgments caused by local data distortion, and significantly improving the comprehensiveness and reliability of state perception.

[0016] Traditional fixed thresholds cannot adapt to environmental changes or vehicle aging. This invention calculates a dynamic overload risk index using a two-factor model in step S4 (e.g., a formula combining threshold deviation, temperature change acceleration, and oil pressure). Step S5 predicts future temperature trends based on road condition matching algorithms (e.g., Euclidean distance scene recognition) using a piecewise linear interpolation model, allowing the protection threshold to dynamically adjust with operating conditions. For example, when climbing hills in mountainous areas, the system can predict the temperature rise trend 5 minutes in advance, triggering an early warning instead of reactive intervention. This avoids the lag of fixed thresholds and reduces performance degradation under extreme conditions.

[0017] Through the tiered triggering mechanism in step S6 (early warning, intervention, and mandatory levels) and the coordinated execution in step S7 (e.g., adjusting the fan, fuel supply, and alarm unit), the system intervenes precisely according to the risk level. For example, when the risk index is 60-75, only an alarm is triggered; when it is 75-90, the cooling system is forced to cool down and the radiator pump is accelerated; and when it exceeds 90, the fuel injection quantity and engine speed are limited. This tiered strategy avoids the power interruption or economic loss caused by traditional "one-size-fits-all" protection, maximizing the operating efficiency of commercial vehicles while ensuring engine safety. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a schematic flowchart illustrating a method according to an embodiment of the present invention.

[0020] Figure 2 This is a schematic block diagram of a system according to an embodiment of the present invention.

[0021] Figure 3 This is a schematic diagram of the structure of a terminal provided in an embodiment of the present invention. Detailed Implementation

[0022] To make the objectives, features, and advantages of this invention more apparent and understandable, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings of the specific embodiments. Obviously, the embodiments described below are only some embodiments of this invention, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0023] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention.

[0024] Figure 1 This is a schematic flowchart illustrating an overload protection control method for a commercial vehicle engine provided by the present invention. Figure 1 The implementing entity can be an overload protection control system for a commercial vehicle engine. Depending on different requirements, the order of steps in this flowchart can be changed, and some steps can be omitted.

[0025] like Figure 1 As shown, the method includes: S1. By using sensors deployed in the engine compartment and body, the engine block temperature, oil pan area temperature, coolant flow / temperature, oil pressure, ambient temperature, vehicle speed, real-time load, throttle opening and driving operation signals are collected. S2. Preprocess the multimodal data collected in step S1, and extract the time-domain features of temperature change rate and load fluctuation amplitude from the preprocessed multimodal data to output a standardized dataset. S3. The standardized dataset output in step S2 is weighted and fused using a weighted feature association algorithm to obtain the engine health status vector. S4. Input the engine health state vector obtained in step S3 into the pre-trained two-factor model to calculate the overload risk index. S5. Combining the overload risk index output in step S4 and the vehicle speed and throttle opening data collected in step S1, the current driving scenario of the commercial vehicle is identified through the pre-stored road condition matching algorithm, and the temperature rise trend is predicted within a preset time in the future. S6. Combining the overload risk index output in step S4 and the temperature rise trend output in step S5, generate graded protection instructions. S7. Based on the graded protection instructions, the engine cooling system is linked to adjust the commercial vehicle's fan or water pump, fuel supply module, and driver's cabin alarm unit to execute corresponding protection actions.

[0026] To facilitate understanding of the present invention, the overload protection control method for commercial vehicle engines provided by the present invention will be further described below, based on the principle of the overload protection control method for commercial vehicle engines of the present invention and in conjunction with the process of performing overload protection control on commercial vehicle engines in the embodiments.

[0027] First, step S1 includes: S11. Install a first thermocouple in the main bearing housing area of ​​the engine block to obtain the cylinder block temperature. A second thermocouple is installed at the midpoint of the oil level in the oil pan to obtain the oil temperature. A turbine flow meter is connected in series in the coolant line to obtain the flow rate. With temperature A piezoresistive sensor is installed in the main oil passage to obtain the oil pressure. The ambient temperature is obtained by installing a thermal resistor in the area of ​​the vehicle body longitudinal beam where there is no heat source interference. A Hall effect sensor is installed on the drive shaft to obtain the vehicle speed v; a strain gauge load cell is installed between the leaf spring and the frame to obtain the real-time load m; an angular displacement sensor is installed on the accelerator pedal shaft to obtain the throttle opening θ; and microswitches are installed on the steering column and brake pedal to obtain driving operation signals. Furthermore, a channel index i is established for the above nine types of parameters, and an independent analog-to-digital conversion gain coefficient is configured for each channel. Where i = 1 to 9 correspond to the parameters mentioned above, and are driving operation signals. Includes coded values ​​for steering, braking, and gear shifting actions; S12. Obtain the current engine speed n via the CAN bus, and determine the sampling frequency according to the physical change rate characteristics of each parameter. Temperature parameters are sampled at low frequency, pressure and flow parameters at medium frequency, and vehicle speed and throttle opening parameters at high frequency. The sampling frequency is calculated using the following formula: ; in, This represents the data acquisition frequency of the i-th channel; n is the real-time engine speed. These are the low-frequency sampling coefficients (values ​​ranging from 1.0 to 2.0). The intermediate frequency sampling coefficient (values ​​range from 2.5 to 3.5); These are the high-frequency sampling coefficients (values ​​ranging from 4.0 to 5.0). S13, based on sampling frequency The collected raw data undergoes a triple verification process: physical threshold judgment, gradient mutation detection, and sensor fault code verification. Data that passes verification is stored in a specified length. Circular cache queue Cache queue length The calculation formula is: ; in, This represents the total number of data points that the i-th channel buffer queue can hold; η is the cache time window; η is the safety factor.

[0028] Specifically, a K-type armored thermocouple is embedded in the main bearing housing area of ​​the engine block (the area with the highest heat load and most prone to cylinder scoring) as the first temperature sensor. Its temperature sensing end is tightly fitted to the metal substrate and connected to the ECU analog input interface via a shielded twisted pair cable to obtain the cylinder block temperature. A thermocouple of the same type is installed at the median level of the oil level inside the oil pan as a second temperature sensor to obtain the oil temperature. To avoid measurement distortion caused by oil level fluctuations, an integrated turbine flow meter is connected in series in the coolant pipeline, with a built-in PT1000 thermal resistor synchronously outputting the coolant temperature. With traffic A piezoresistive oil pressure sensor (range 0–1.5 MPa) is screwed into the threaded hole of the main oil passage, outputting a 4–20 mA current signal which is then converted by I / V and fed into the ADC. An ambient temperature sensor, using a Pt100 RTD, is installed on the longitudinal beam of the chassis away from heat sources such as the exhaust pipe and turbocharger to reduce radiation interference. Vehicle speed is obtained by detecting the number of gear pulses using a Hall effect speed sensor mounted at the end of the driveshaft; real-time load is measured by a strain gauge load cell array installed at the connection between the front / rear leaf springs and the chassis, outputting a differential voltage via a Wheatstone bridge; throttle opening is acquired by a high-precision angular displacement sensor (resolution 0.1°) mounted on the accelerator pedal shaft; driving operation signals are collected by a steering encoder switch below the steering column, a microswitch at the bottom of the brake pedal, and a gear shift lever position sensor, generating binary codes for steering, braking, and shifting respectively, which are combined into an 8-bit operation signal. The above nine types of physical quantities correspond to channel indices i=1,2,…,9, and each channel is configured with an independent analog-to-digital conversion gain coefficient during the ECU initialization phase. (e.g., temperature channel) Pressure Channel Speed ​​lane This is to match the output amplitude range of each sensor and make full use of the ADC's 12-bit resolution.

[0029] The ECU periodically reads the real-time engine speed n (unit: rpm) broadcast by the engine control module via the CAN bus. Based on the physical dynamic characteristics of each parameter, the nine channels are divided into three sampling levels: Low-frequency parameters (i=1,2,4,5: cylinder temperature, oil temperature, coolant temperature, ambient temperature) change slowly, and the sampling frequency is set to [value missing]. ,in The default value is 1.5; Medium-frequency components (i=3,4: coolant flow rate, oil pressure) have a faster response; the sampling frequency is set to [value missing]. ,in The default value is 3.0; For high-frequency data (i=6,7,8,9: vehicle speed, load, throttle opening, driving operation), transient behavior needs to be captured, and the sampling frequency is set to [value missing]. ,in The default value is 4.5.

[0030] For example, when n=1800rpm, the sampling rate for the low-frequency channel is 45 Hz, for the mid-frequency channel it is 90 Hz, and for the high-frequency channel it is 135 Hz. This mechanism ensures increased sampling density of critical dynamic signals at high speeds, while reducing system load at idle speeds.

[0031] For the raw data collected from each channel at its sampling frequency, execute the following sequentially: Physical threshold determination: Check whether it is within the sensor's measurement range (e.g.: ); Gradient mutation detection: Calculate the difference between the current value and the previous valid value. If it exceeds the preset jump threshold (e.g., temperature > 10℃ / s), it is marked as suspicious. Sensor fault code verification: Check if the corresponding sensor DTC (Diagnostic Trouble Code) exists in the CAN bus or local diagnostic register.

[0032] Only when all three verifications pass is the data assigned a quality label q=0 and written to the channel's dedicated circular buffer queue. .

[0033] Secondly, step S2 includes: S21. From the established nine-channel circular buffer queue In the process, continuous valid data points are selected based on the data quality label q=0, and the extraction length is [length missing]. The data segment, and based on the sampling frequency of each channel. With the smallest time base unit Perform linear interpolation alignment to construct a 9×M dimensional synchronization data matrix. ;in, The highest sampling frequency among the nine channels; extraction length The calculation formula is: ; in, This represents the number of data points extracted from the i-th channel; For feature extraction time window; synchronized data matrix The number of columns M is equal to the minimum extraction length min among all channels. This alignment operation ensures that the nine dimensions of data, including cylinder temperature, load, and throttle opening, are strictly correlated at the same time reference. S22. Synchronize the data matrix in step S21. The three rows of data corresponding to cylinder block temperature (channel index i=1), oil temperature (i=2), and coolant temperature (i=4) are used to calculate the first-order backward difference point by point according to the time series, so as to obtain the instantaneous rate of change of each temperature parameter within the feature extraction time window, and then synthesize a three-dimensional temperature gradient vector. The formula for calculating the cylinder block temperature change rate is: ; in, This represents the rate of change of cylinder temperature at time t (the calculation principle for the other two rate of change of temperature is the same as that for cylinder temperature). The cylinder block temperature at time t; For a unified sampling interval after synchronization, the temperature gradient vector The j-th component corresponds to the average rate of change of the three temperatures, i.e. ; S23, Synchronization data matrix The corresponding row of data for the real-time load (channel index i=7) uses a length of [missing information]. The standard deviation of the load data within the sliding window is calculated segment by segment as the load fluctuation amplitude characteristic Am, and the load dynamic stability index is calculated. The formula for calculating the magnitude of load fluctuations to quantify their severity is as follows: ; in, This represents the load fluctuation range at time t; This represents the load value at a historical moment. The width of the sliding window; The average load within this window is used to calculate the load dynamic stability index. ; in, It is a dimensionless load dynamic stability index; This represents the load fluctuation range within the current time window. The baseline fluctuation range for vehicle calibration under no-load conditions (preset value 30~50kg); γ is the load correction factor ( , For the current load, (Rated load capacity). The larger the value, the more severe the shaking of the cargo or the impact of the road surface, which causes additional stress to the engine support system. S24. The original data matrix obtained in step S21 is... With the extracted temperature gradient vector Load fluctuation range Am and stability index The vectors are concatenated to form a 12-dimensional feature vector, and deviation standardization based on historical extreme values ​​is performed on each feature dimension to finally output a standardized dataset. The deviation standardization mapping formula is: ; in, These are the standardized eigenvalues; These are the original feature values ​​or the extracted time-domain feature values. and These are the maximum and minimum values ​​of this type of feature, respectively.

[0034] The preprocessing module first iterates through each channel. Filter out continuous valid data segments with quality label q=0. Starting from the current time point, backtrack one feature extraction time window. Calculate the number of data points to be extracted for each channel: ; in, The sampling frequency determined for this channel in S12 (e.g., cylinder block temperature channel) ,but Take the smallest value among all channels. The value is used as the column number of the synchronization matrix. Subsequently, using the smallest time base unit (For example ,but To unify the time reference, linear interpolation is performed on the data from each channel, aligning the originally asynchronously acquired nine-dimensional signals (cylinder temperature, oil temperature, coolant temperature / flow rate, engine oil pressure, ambient temperature, vehicle speed, load, throttle opening, and driving operation) at the same time, thus constructing a unified time reference. 3D Synchronous Data Matrix This operation ensures that subsequent feature extraction is based on strictly time-synchronized multimodal data, avoiding spurious correlations caused by asynchronous sampling.

[0035] Synchronization data matrix Extract the first row (cylinder block temperature) ), Line 2 (oil temperature) ) and line 4 (coolant temperature) For each row, calculate the first-order backward difference point by point according to the time series: ; in, This results in three lengths of M. The instantaneous rate of change sequence of 1. Further synthesis of a three-dimensional temperature gradient vector. Each component represents the average rate of change of the corresponding temperature within a 10-second window: ; This vector represents the overall trend of engine thermal load increase or decrease and is a key input for S3 health status fusion.

[0036] Next, step S3 includes: S31. Based on the data cleaned in S24, core parameters are selected using a feature importance ranking algorithm. For temperature-related features, cylinder block temperature, coolant temperature, and exhaust temperature change rate are retained; for mechanical status features, oil pressure, crankshaft speed, and piston vibration amplitude are retained; for environmental features, ambient temperature, atmospheric pressure, and humidity are retained. Principal component analysis is then performed on similar features. Two principal components are extracted for the temperature category, and sub-weights are assigned based on the variance contribution rate after calculating the scores of each principal component. The principal component score calculation is based on a linear transformation formula: ; in, This represents the score of the k-th principal component; This represents the loading coefficient of the k-th principal component on the i-th original feature; This represents the i-th original feature value; n represents the number of features in this class; the weight allocation formula is based on the variance contribution rate: ; in, This represents the weight of the k-th principal component; The eigenvalues ​​of the k-th principal component are represented by m; m represents the total number of principal components. S32. Based on the principal component scores and weights from step S31, a weighted feature association algorithm is used to multiply the comprehensive score for temperature by a weight of 0.4, the score for mechanical status by a weight of 0.3, and the score for environmental status by a weight of 0.3. These weights are then concatenated in dimensional order to form an engine health status vector (specifically a 12-dimensional engine health status vector). Simultaneously, the Euclidean modulus of the vector is calculated as the comprehensive health status index (a modulus ≤ 8 indicates basic health). ; in, The comprehensive score representing the engine health status vector is used to determine the engine status based on H and adjust the sampling frequency accordingly. , The principal component weights are for the temperature mode. , ; , These are the scores of the first and second principal components for the temperature category, respectively. , , For mechanical state feature sub-weights, ; These are the normalized value of oil pressure, the speed fluctuation coefficient, and the logarithmic transformation value of vibration amplitude, respectively. , For environmental feature sub-weights, ; These represent the standardized value of ambient temperature and the atmospheric pressure correction factor, respectively. The 12-dimensional engine health state vector is... ,in, This refers to the characteristics of load fluctuation amplitude; The load dynamic stability index; This represents the average rate of change of cylinder block temperature.

[0037] Set the threshold range for H, for example: When H≥7.5, the engine is judged to be in good condition, triggering the basic monitoring mode. The system only performs low-frequency data acquisition (e.g., reducing the sampling frequency of S1) to reduce the computational load.

[0038] When 5≤H<7.5, the status is determined to be normal, triggering the standard monitoring mode, and the system runs step S4 according to the default process.

[0039] When H < 5, the status is determined to be abnormal, triggering the enhanced intervention mode. The system prioritizes the two-factor model calculation in step S4 and increases the frequency of risk assessment (e.g., shortening the call interval of S4).

[0040] The structure constructed in step S3 of this invention is not an end-to-end deep neural network, but a feature fusion engine composed of two statistical learning modules that are highly interpretable, have low computational overhead, and are suitable for automotive embedded deployment. Its structure is as follows: Input: The standardized multidimensional feature vector output by S24; Module A: Feature Importance Filter (Offline Training, Vehicle-Mounted Installation); Output: 9 core features (grouped into three categories: temperature, mechanical, and environmental); Module B: Grouped Principal Component Analysis and Weighted Fusion Tool (Execute Online); Temperature submodule: PCA dimensionality reduction → Principal component score → Sub-weighting; Mechanical submodule: Direct weighting (no dimensionality reduction); Environment submodule: Direct weighting (without dimensionality reduction); Output: A 12-dimensional engine health state vector and its Euclidean modulus H; The XGBoost (Extreme Gradient Boosting) classifier was used as a feature importance evaluation tool. Its purpose was not for real-time prediction, but rather for offline analysis of historical fault data to quantify the ability of each sensor feature to discriminate engine overload conditions, thereby selecting the most representative core parameters.

[0041] Training data source: Positive samples: 5000 historical operation segments labeled as "overload failure occurred" (e.g., piston cylinder scoring, bearing erosion), each segment containing the complete raw S1 signal 10 minutes before the failure; Negative samples: 45,000 sets of "normal operation" segments, covering typical working conditions such as highways, national roads, mountain roads, empty loads, full loads, high temperatures, and high altitudes; All samples are processed through the S1–S2 process to generate 13-dimensional standardized feature vectors as model input.

[0042] Training objectives and tags: Tag y: Binary category label (1 = overload, 0 = normal); Optimization objective: Maximize AUC (Area Under ROC Curve) to balance minority class (fault) recognition rate and false alarm rate.

[0043] After training, the "Gain" importance metric is extracted (representing the average information gain of this feature across all trees). The top 9 features are selected in descending order of importance and then categorized into three classes based on their physical attributes: Temperature-related parameters: Cylinder block temperature, coolant temperature, and exhaust temperature change rate (ranked 1-3 in importance). Mechanical: Oil pressure, crankshaft speed fluctuation coefficient, piston vibration amplitude (ranked 4-6); Environmental factors: ambient temperature, atmospheric pressure, humidity (ranked 7–9).

[0044] In a preferred embodiment of the present invention, step S3 is performed by a weighted fusion module in the vehicle ECU, and its input is entirely derived from the standardized dataset output in step S24. This dataset contains the original synchronization signals and their derived time-domain features. The core objective of S3 is to compress high-dimensional, redundant multimodal data into a structured, low-redundancy engine health state vector, specifically implemented as follows: The weighted fusion module first evaluates the importance of the 13-dimensional features output by S24 based on a feature importance ranking model trained on a historical operating database (containing 5 million labeled operating condition samples), using the XGBoost algorithm and engine fault labels as supervision signals. According to the ranking results, the following nine core parameters are retained and divided into three categories based on their physical meaning: Temperature category (3D): Cylinder block temperature (From S11 thermocouple, synchronized by S21, standardized by S24); Coolant temperature (From S11 turbine flow meter integrated PT1000, processed as above); Exhaust temperature change rate (Although not explicitly listed in S1, commercial vehicles are equipped with exhaust temperature sensors as standard, and their raw signals are added to the output of S24 as derived features after differential processing in S22.) Mechanical condition category (3D): Normalized value of engine oil pressure ( The crankshaft speed fluctuation coefficient is derived from the S11 piezoresistive sensor, normalized by S24 and mapped to [0.2, 0.8]. (n is from the CAN bus, and the ratio of the standard deviation to the mean of the 10-second window is calculated after synchronization with S21); Logarithmic transformation value of piston vibration amplitude ( The original signal from the engine cylinder head vibration sensor is synchronized by S21, standardized by S24, and then logarithmically compressed to reduce the dynamic range. Environmental category (3D): Standardized value of ambient temperature (From S11 RTD, normalized by S24 deviation); Atmospheric pressure correction factor ( (Source: vehicle barometer, standardized according to S24); relative humidity (From an integrated temperature and humidity sensor, standardized by S24).

[0045] Perform principal component analysis (PCA) independently for each type of feature: Temperature category: Construct a 3×N matrix (N is the number of samples), calculate the covariance matrix, and solve for the eigenvalues. Take the first two principal components (cumulative variance contribution rate ≥ 85%), and calculate their scores as follows: ; in, Let be the loading coefficient of the k-th principal component on the i-th original feature. These are the standardized feature values.

[0046] Mechanical state and environment classes: Since they are only 3-dimensional and have independent physical meanings, they are not reduced in dimensionality and are directly assigned weights.

[0047] The three types of features are fused according to preset modal weights: Overall score for temperature category: ; Overall mechanical condition score: ; Overall environmental score: ; The final result is a 12-dimensional health state vector.

[0048] In addition, step S4 includes: The engine health state vector output from step S3 (the aforementioned 12-dimensional engine health state vector) is input into the pre-trained two-factor model. The overload risk index is calculated through weight calculation and dynamic threshold comparison. ; in; Indicates the overload risk index; This represents the threshold deviation weighting coefficient; D represents the normalized threshold deviation. , The dynamic protection threshold for the q-th core parameter in the engine health state vector. This represents the mapping value of the q-th core parameter in the engine health state vector. , This represents the original value of the q-th core parameter in the engine health state vector. This is the maximum value of the core parameter; This represents the weighting coefficient for the acceleration due to temperature change; A represents the acceleration due to temperature change. , This represents the mapping value of the core temperature parameters at the current moment. These are the mapping values ​​of the core temperature parameters at the first and second sampling times, respectively. This is the scaling factor; It represents the normalized oil pressure; the two factors include: threshold deviation D and temperature change acceleration A.

[0049] Dynamic weight adjustment trigger: The real-time value of H is used as the weight of the S4 two-factor model. The adjustment is based on the following: For example, when H is low (e.g., H < 6), the threshold deviation weight is automatically increased. This makes the model pay more attention to the risk of parameter deviation; when H is high (e.g., H>8), the weight of temperature change acceleration is increased. Strengthen trend prediction.

[0050] The weighting coefficients in the threshold comparison are dynamically adjusted based on real-time operating conditions. The weighting adjustment is based on the current engine load and environmental parameters. The formula for calculating the weighting coefficients is as follows: ; in, This represents the threshold deviation weighting coefficient; The ambient temperature is used; when the ambient temperature deviates from the standard temperature by 25°C, the weighting is adjusted. Automatic adjustment enables the system to optimize the sensitivity of threshold comparisons based on actual environmental conditions. For example, in high-temperature environments (… At temperatures above 25℃, Increase the weight of threshold deviation in the system to more sensitively detect overheating risks; low temperature environment ( When <25℃, Appropriately reducing the weighting coefficients avoids over-protection under low-temperature conditions; this temperature-adaptive characteristic enables the system to overcome the limitations of fixed weighting coefficients under different climatic conditions, significantly improving its adaptability to different geographical environments and seasonal changes.

[0051] The value is calculated in segments based on the real-time value of H: .

[0052] In extreme environments, moderate amplification Adjustments to increase adaptability: .

[0053] Furthermore, based on the three-dimensional operating condition vector composed of engine speed n, real-time load m, and vehicle speed v, the dynamic protection threshold for each core parameter is calculated using a pre-stored operating condition-threshold mapping surface. The dynamic threshold calculation formula is as follows: : ; This represents the dynamic protection threshold (dimensionless) for the q-th core parameter. The static reference threshold for this parameter is 0.75 for cylinder block temperature, 0.65 for oil pressure, and 0.70 for coolant temperature, all determined by thermal balance tests on a calibration bench. n is the real-time engine speed (unit: rpm) obtained in step S12. The engine's rated speed is 1900~2100 rpm, and m is the real-time load value in the buffer queue in step S13. The rated load of the vehicle is in tons, and v is the vehicle speed collected. This is the vehicle's maximum speed. , , The operating condition coupling coefficient ( The value ranges from 0.1 to 0.15. The value ranges from 0.2 to 0.25. (Values ​​range from -0.05 to 0.05). This formula explicitly establishes a positive correlation between the dynamic threshold and the engine load intensity, avoiding premature triggering of the fixed threshold under heavy load and low speed conditions or delayed response under light load and high speed conditions.

[0054] Then, step S5 includes: S51. Based on the vehicle speed sequence and throttle opening change data collected in real time in step S1, driving characteristic parameters are extracted through a sliding window with a length of 60 seconds. The driving characteristic parameters include the average vehicle speed fluctuation coefficient, the standard deviation of throttle opening, and the shift frequency. S52. Perform Euclidean distance matching between the extracted driving feature parameters and the pre-stored road condition feature library, and select the scene with the smallest distance and similarity ≥ 0.75 as the current driving scene; the road condition feature library contains standard feature vectors of three typical scenes: highway, national highway and mountain road; S53. Based on the current driving scenario identified in step S52 and the overload risk index output in step S4, the temperature rise coefficient matrix of the corresponding scenario is called. The temperature increase per unit time is calculated through piecewise linear interpolation, and then an ambient temperature correction term is added to generate a temperature trend curve for the next 5 minutes. The temperature rise trend prediction formula is: ; in, This represents the predicted temperature at a future time Δt. Indicates the current measured temperature; Indicates the road condition coefficient; Indicates the load factor; Indicates the weight of throttle opening; Indicates the prediction time step; Indicates ambient temperature.

[0055] The Execution Control Module (ECU) continuously monitors and caches vehicle speed v and throttle opening θ data in a separate task, Task_SceneExtract (priority OS_PRIO_MEDIUM, period 100 ms), within the real-time operating system. Two circular buffers, v_buffer

[6000] and θ_buffer

[6000] , are allocated in the ECU's SRAM. Each unit is a 16-bit signed integer, storing data from the most recent 6000 sampling points (corresponding to a data length of 60 seconds at a maximum sampling frequency of 100 Hz). When the buffers are full, new data overwrites the oldest data, maintaining a constant "sliding window" length of W_{ext} = 6000.

[0056] When the task cycle arrives, extract the data segment L=60 seconds prior to the current time from the buffer, with an extraction length of... Dynamically calculated based on the current vehicle speed and the actual sampling frequencies f_v and f_θ of the throttle channel: =min(f_v, f_θ)×60. Taking an engine speed of n=1500 rpm as an example, the vehicle speed sampling frequency f_v=75 Hz and the throttle sampling frequency fθ=225 Hz, the minimum value is taken. =4500 points.

[0057] Average vehicle speed fluctuation coefficient The extraction method is as follows: First, calculate the mean value of the vehicle speed data within 60 seconds, μ_v = (1 / First, calculate Σv(t), then calculate the difference sequence Δv(t) = |v(t) - v(t-1)| between adjacent sampling points, and finally calculate the mean of the difference sequence and normalize it. ; in, This is a dimensionless fluctuation coefficient, ranging from 0 to 1. During high-speed cruising... ≈0.05, but can reach over 0.3 during frequent acceleration and deceleration. Throttle opening standard deviation. Calculated directly as follows: ; in, The average throttle opening over 60 seconds. The unit is degrees (°). In mountainous climbing conditions, the angle can reach more than 15°, while in constant speed driving conditions, it is usually less than 5°.

[0058] Shift frequency This is achieved by detecting step changes in the gear position sensor signal (analog signal or CAN message). The ECU internally maintains the gear position status variable `gear_status`, and counts when the number of changes in this variable exceeds a threshold within 60 seconds. The shift frequency calculation formula is as follows: ; in, The unit is times per minute, where N{shift} is the number of gear changes within 60 seconds. (Under mountain road conditions) It can reach 8-12 times per minute, while it approaches 0 during high-speed cruising.

[0059] The standard feature vectors were obtained through statistical analysis of more than 50 sets of real vehicle test data. For example, the standard vector for highway scenes is [0.05, 3.0, 0.2], and the variance vector is [0.01, 0.5, 0.1]; the standard vector for national highway scenes is [0.15, 8.0, 2.0], and the variance vector is [0.03, 1.2, 0.5]; the standard vector for mountain road scenes is [0.35, 18.0, 10.0], and the variance vector is [0.08, 2.5, 1.0].

[0060] Euclidean distance matching is performed in the Task_SceneMatch task (priority OS_PRIO_HIGH, period 1 second). The currently extracted driving feature parameter vector is calculated. =[ , , Weighted Euclidean distance to the standard vectors of the three scene types: ; in, This represents the distance between the current state and the k-th scenario. Let j be the j-th component of the standard feature vector of the k-th scene. This corresponds to the variance components. The variance weighting mechanism gives higher weight to dimensions with smaller fluctuations (such as vehicle speed fluctuation coefficient) in distance calculation, thus improving matching accuracy.

[0061] The matching logic is as follows: Select the minimum distance d{min} = min(d1, d2, d3), and calculate the similarity S = exp(-d{min}). If S ≥ 0.75, then scene_id is assigned the scene code corresponding to the one with the smallest distance; otherwise, it is judged as "unknown scene," and the system defaults to using national road scene parameters to ensure conservatism. This threshold of 0.75 was determined through ROC curve analysis, which can ensure a correct recognition rate of over 95% while controlling the false recognition rate to within 5%.

[0062] The temperature rise coefficient matrix is ​​stored in Flash addresses 0x080E0400-0x080E0BFF, using a three-dimensional lookup table structure TempCoeff[scene_id][risk_level][temp_index]. Scene_id indexes three scene categories (1-3), risk_level represents the four discretized levels of the overload risk index R_overload (0: <0.3, 1: 0.3-0.5, 2: 0.5-0.7, 3: >0.7), and temp_index is the cylinder block temperature zone index (0: 60-80°C, 1: 80-100°C, 2: 100-120°C, 3: >120°C). Matrix elements are dimensionless temperature rise coefficients α_rise, representing the rate of temperature rise per unit throttle opening weight under the current scene and risk level.

[0063] Piecewise linear interpolation is performed in the Task_TempPredict task (priority OS_PRIO_HIGH, period 1 second). Assume the current cylinder block temperature T_{cyl} is in the interval [T_low, T_high] (e.g., 95°C is in the interval [80,100]), and the corresponding index temp_index=1.

[0064] Additionally, step S6 includes: S61. Combining the overload risk index RI output in step S4 and the temperature rise trend output in step S5. The protection level is determined through a tiered triggering mechanism: when the RI is in the 60-75 range, a warning level protection is activated; when the RI is in the 75-90 range, an intervention level protection is activated; when the RI exceeds 90 or... Forced protection will be activated when the temperature is predicted to exceed 130℃ within 5 minutes. ; in, Indicates the protection level; RI indicates the overload risk index; t represents the predicted temperature for the next 5 minutes; t represents the current time. S62. Based on the protection level determined in step S61, generate corresponding control command parameters: For the warning level, the red indicator light on the control panel flashes and the buzzer emits an intermittent warning sound; for the intervention level, send a command to the cooling system via the CAN bus to force the cooling fan to enter full-speed operation mode and increase the flow rate of the cooling pump to 120% of the rated value; for the forced level, the fuel injection system reduces the fuel injection quantity by 15%-20% by adjusting the injection pulse width, while limiting the maximum speed to no more than 2200 rpm.

[0065] Finally, step S7 includes: S71. The execution control module first receives the protection instruction data packet from the instruction generation module and parses the following content: Protection level ; List of target execution units (e.g., cooling fan, electric water pump, fuel injector, instrument panel, buzzer); The target control parameters for each unit (e.g., fan speed percentage, fuel injection pulse width correction, alarm frequency, etc.).

[0066] The instruction structure is strictly consistent with the S62 definition, for example: Early warning level: Includes only alarm unit control parameters; Intervention level: Includes cooling system + alarm unit parameters; Forced level: Includes parameters for cooling system, fuel system, and alarm unit.

[0067] S72: Perform cooling system adjustment: Based on the protection level, send control commands to the engine cooling subsystem: like =Intervention level or mandatory level: The PWM duty cycle command is sent to the electronic cooling fan controller via the CAN bus to force the fan into full-speed operation mode (duty cycle 100%, corresponding speed ≥2800 rpm). Send a flow control signal to the electronic water pump drive module to increase the coolant circulation flow rate to 120% of the rated value (achieved by increasing the water pump motor speed). Start cooling system status feedback monitoring: continuously read the actual fan speed and water pump current. If the target value is not reached within 5 seconds, record the DTC (diagnostic fault code) and upgrade the alarm level.

[0068] S73, Perform fuel supply regulation: Only when =Activated at the mandatory level: Send a fuel injection correction request to the fuel injection submodule of the engine control unit (ECU); Adjust the injection pulse width to reduce the amount of fuel injected in a single injection by 15%–20% (the specific value is dynamically interpolated based on the current speed and load). A speed limit command is sent synchronously to clamp the engine's maximum permissible speed to 2200 rpm (achieved by modifying the ignition timing and throttle opening in coordination). The torque smooth transition algorithm is activated to gradually reduce power within 2 seconds, avoiding sudden power changes that could lead to loss of vehicle control.

[0069] Safety mechanism: If the vehicle speed is >60 km / h or the brake pedal is not depressed, the fuel limit will be delayed to prioritize driving safety.

[0070] S74, Execute cockpit alarm: Activate the human-computer interaction unit according to the protection level: like =Warning Level: The red overload indicator light on the control panel flashes at a frequency of 1Hz. The buzzer emits an intermittent beep (sounds for 0.5 seconds, pauses for 1.5 seconds, and remains on for 30 seconds before turning constantly lit). A text message pops up on the LCD screen: "Engine overload, please reduce throttle."

[0071] like =Intervention level or mandatory level: The red indicator light remains constantly on; The buzzer sounds continuously for 5 seconds, then switches to an intermittent 1 Hz tone. The display screen showed: "Engine severely overloaded! Protection measures have been activated. Please pull over and check."

[0072] S75. Execution Status Feedback and Exit Mechanism: Continuously monitor the overload risk index RI output in step S4 and the predicted temperature in step S5. The current protection level will automatically exit when any of the following conditions are met: RI < 55 and predicted temperature < 120℃ for 60 seconds; The driver manually presses the "Protection Reset" button (located on the instrument panel); When exiting, gradually remove the restrictions using a gradual recovery strategy (e.g., reduce fan speed by 10% every 10 seconds until normal control is restored). Record this protection event (time, highest RI, duration, operation response) to non-volatile memory for after-sales diagnostics.

[0073] In some embodiments, the overload protection control system 200 for the commercial vehicle engine may include multiple functional modules composed of computer program segments. The computer programs for each program segment in the overload protection control system 200 for the commercial vehicle engine may be stored in the memory of a computer device and executed by at least one processor to perform (see details). Figure 1 (Description) The overload protection control function of commercial vehicle engines.

[0074] In this embodiment, the overload protection control system 200 for the commercial vehicle engine can be divided into multiple functional modules according to its functions, such as... Figure 2 As shown. The functional modules may include: a sensor array 210, a preprocessing module 220, a weighted fusion module 230, a risk calculation module 240, a scene recognition and prediction module 250, an instruction generation module 260, and an execution control module 270. The module referred to in this invention is a series of computer program segments that can be executed by at least one processor and perform a fixed function, stored in memory. In this embodiment, the functions of each module will be described in detail in subsequent embodiments.

[0075] Specifically, sensor array 210, deployed in the engine compartment and vehicle body, is used to collect engine block temperature, oil pan temperature, coolant flow / temperature, oil pressure, ambient temperature, vehicle speed, real-time load, throttle opening, and driving operation signals. Preprocessing module 220, communicatively connected to the sensor array, is used to preprocess the multimodal data, extracting time-domain features of temperature change rate and load fluctuation amplitude, and outputting a standardized dataset. Weighted fusion module 230, communicatively connected to the preprocessing module, is used to perform weighted fusion of the standardized dataset using a weighted feature association algorithm to obtain an engine health state vector. Risk calculation module 240, communicatively connected to the weighted fusion module, is used to input the engine health state vector into pre-training. The system employs a two-factor model to calculate the overload risk index; a scene recognition and prediction module 250, which communicates with the risk calculation module and sensor array, combines the overload risk index with vehicle speed and throttle opening data, and uses a pre-stored road condition matching algorithm to identify the current driving scenario of the commercial vehicle and predict the temperature rise trend within a preset time period; an instruction generation module 260, which communicates with the risk calculation module and scene recognition and prediction module, combines the overload risk index and temperature rise trend to generate graded protection instructions; and an execution control module 270, which communicates with the instruction generation module, uses the graded protection instructions to link the engine cooling system to adjust the fan or water pump, control the fuel supply module and the cockpit alarm unit, and execute corresponding protection actions.

[0076] Figure 3This is a schematic diagram of the structure of a terminal 300 provided in an embodiment of the present invention. The terminal 300 can be used to execute the overload protection control method for commercial vehicle engines provided in the embodiment of the present invention.

[0077] The terminal 300 may include a processor 310, a memory 320, and a communication module 330. These components communicate via one or more buses. Those skilled in the art will understand that the server structure shown in the figure does not constitute a limitation of the present invention. It may be a bus topology or a star topology, and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0078] The memory 320 can be used to store the execution instructions of the processor 310. The memory 320 can be implemented by any type of volatile or non-volatile memory terminal or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. When the execution instructions in the memory 320 are executed by the processor 310, the terminal 300 is able to perform some or all of the steps in the above method embodiments.

[0079] The processor 310 serves as the control center of the storage terminal, connecting various parts of the electronic terminal via various interfaces and lines. It executes software programs and / or modules stored in the memory 320, and calls data stored in the memory to perform various functions of the electronic terminal and / or process data. The processor can be composed of integrated circuits (ICs), such as a single packaged IC or multiple packaged ICs with the same or different functions connected together. For example, the processor 310 may consist only of a central processing unit (CPU). In this embodiment of the invention, the CPU may have a single processing core or include multiple processing cores.

[0080] The communication module 330 is used to establish a communication channel, enabling the storage terminal to communicate with other terminals. It receives user data sent by other terminals or sends user data to other terminals.

[0081] The present invention also provides a computer storage medium, wherein the computer storage medium may store a program, which, when executed, may include some or all of the steps provided in the embodiments of the present invention. The storage medium may be a magnetic disk, an optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0082] Those skilled in the art will clearly understand that the techniques in the embodiments of the present invention can be implemented using software plus necessary general-purpose hardware platforms. Based on this understanding, the technical solutions in the embodiments of the present invention, or the parts that contribute to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium such as a USB flash drive, mobile hard drive, read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk, or other media capable of storing program code. It includes several instructions to cause a computer terminal (which may be a personal computer, server, or a second terminal, network terminal, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention.

[0083] The same or similar parts between the various embodiments in this specification can be referred to mutually. In particular, the terminal embodiments are basically similar to the method embodiments, so the description is relatively simple, and the relevant parts can be referred to the description in the method embodiments.

[0084] In the embodiments provided by this invention, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the system embodiments described above are merely illustrative. For instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between systems or modules may be electrical, mechanical, or other forms.

[0085] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules; that is, they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0086] In addition, the functional modules in the various embodiments of the present invention can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0087] Although the present invention has been described in detail with reference to the accompanying drawings and preferred embodiments, the present invention is not limited thereto. Various equivalent modifications or substitutions can be made to the embodiments of the present invention by those skilled in the art without departing from the spirit and essence of the invention, and such modifications or substitutions should all be within the scope of the present invention. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should also be covered within the protection scope of the present invention.

Claims

1. An overload protection control method for a commercial vehicle engine, characterized in that, include: S1. By using sensors deployed in the engine compartment and body, the engine block temperature, oil pan area temperature, coolant flow / temperature, oil pressure, ambient temperature, vehicle speed, real-time load, throttle opening and driving operation signals are collected. S2. Preprocess the multimodal data collected in step S1, and extract the time-domain features of temperature change rate and load fluctuation amplitude from the preprocessed multimodal data to output a standardized dataset. S3. The standardized dataset output in step S2 is weighted and fused using a weighted feature association algorithm to obtain the engine health status vector. S4. Input the engine health state vector obtained in step S3 into the pre-trained two-factor model to calculate the overload risk index. S5. Combining the overload risk index output in step S4 and the vehicle speed and throttle opening data collected in step S1, the current driving scenario of the commercial vehicle is identified through the pre-stored road condition matching algorithm, and the temperature rise trend is predicted within a preset time in the future. S6. Combining the overload risk index output in step S4 and the temperature rise trend output in step S5, generate graded protection instructions. S7. Based on the graded protection instructions, the engine cooling system is linked to adjust the commercial vehicle's fan or water pump, fuel supply module, and driver's cabin alarm unit to execute corresponding protection actions.

2. The overload protection control method for a commercial vehicle engine according to claim 1, characterized in that, Step S1 includes: S11. Install a first thermocouple in the main bearing housing area of ​​the engine block to obtain the cylinder block temperature. A second thermocouple is installed at the midpoint of the oil level in the oil pan to obtain the oil temperature. A turbine flow meter is connected in series in the coolant line to obtain the flow rate. With temperature A piezoresistive sensor is installed in the main oil passage to obtain the oil pressure. The ambient temperature is obtained by installing a thermal resistor in the area of ​​the vehicle body longitudinal beam where there is no heat source interference. A Hall effect sensor is installed on the drive shaft to obtain the vehicle speed v; a strain gauge load cell is installed between the leaf spring and the frame to obtain the real-time load m; an angular displacement sensor is installed on the accelerator pedal shaft to obtain the throttle opening θ; and microswitches are installed on the steering column and brake pedal to obtain driving operation signals. Furthermore, a channel index i is established for the above nine types of parameters, and an independent analog-to-digital conversion gain coefficient is configured for each channel. Where i = 1 to 9 correspond to the parameters mentioned above, and are driving operation signals. Includes coded values ​​for steering, braking, and gear shifting actions; S12. Obtain the current engine speed n via the CAN bus, and determine the sampling frequency according to the physical change rate characteristics of each parameter. Temperature parameters are sampled at low frequency, pressure and flow parameters at medium frequency, and vehicle speed and throttle opening parameters at high frequency. The sampling frequency is calculated using the following formula: ； in, This represents the data acquisition frequency of the i-th channel; n is the real-time engine speed. These are low-frequency sampling coefficients; These are the intermediate frequency sampling coefficients; These are high-frequency sampling coefficients; S13, based on sampling frequency The collected raw data undergoes a triple verification process: physical threshold judgment, gradient mutation detection, and sensor fault code verification. Data that passes verification is stored in a specified length. Circular cache queue Cache queue length The calculation formula is: ； in, This represents the total number of data points that the i-th channel buffer queue can hold; η is the cache time window; η is the safety factor.

3. The overload protection control method for a commercial vehicle engine according to claim 2, characterized in that, Step S2 includes: S21. From the established nine-channel circular buffer queue In the process, continuous valid data points are selected based on the data quality label q=0, and the extraction length is [length missing]. The data segment, and based on the sampling frequency of each channel. With the smallest time base unit Perform linear interpolation alignment to construct a 9×M dimensional synchronization data matrix. ;in, The highest sampling frequency among the nine channels; extraction length The calculation formula is: ； in, This represents the number of data points extracted from the i-th channel; For feature extraction time window; synchronized data matrix The number of columns M is equal to the minimum extraction length min among all channels. ); S22. Synchronize the data matrix in step S21. The three rows of data corresponding to cylinder block temperature, oil temperature, and coolant temperature are used to calculate the first-order backward difference point by point according to the time series, so as to obtain the instantaneous rate of change of each temperature parameter within the feature extraction time window, and then synthesize a three-dimensional temperature gradient vector. The formula for calculating the cylinder block temperature change rate is: ； in, This represents the rate of change of cylinder block temperature at time t; The cylinder block temperature is given at time t. For a unified sampling interval after synchronization, the temperature gradient vector The j-th component corresponds to the average rate of change of the three temperatures, i.e. ; S23, Synchronization data matrix The corresponding line of data for the real-time load is in the format of a line with a length of [length missing]. The standard deviation of the load data within the sliding window is calculated segment by segment as the load fluctuation amplitude characteristic Am, and the load dynamic stability index is calculated. The formula for calculating the magnitude of load fluctuations to quantify their severity is as follows: ； in, This represents the load fluctuation range at time t; This represents the load value at a historical moment. The width of the sliding window; The average load within this window is used to calculate the load dynamic stability index. ； in, It is a dimensionless load dynamic stability index; This represents the load fluctuation range within the current time window. γ represents the baseline fluctuation range under the vehicle's calibration unloaded state; γ is the load correction factor. S24. The original data matrix obtained in step S21 is... With the extracted temperature gradient vector Load fluctuation range Am and stability index The vectors are concatenated to form a 12-dimensional feature vector, and deviation standardization based on historical extreme values ​​is performed on each feature dimension to finally output a standardized dataset. The deviation standardization mapping formula is: ； in, These are the standardized eigenvalues; These are the original feature values ​​or the extracted time-domain feature values. and These are the maximum and minimum values ​​of this type of feature, respectively.

4. The overload protection control method for a commercial vehicle engine according to claim 3, characterized in that, Step S3 includes: S31. Based on the data cleaned in S24, core parameters are selected using a feature importance ranking algorithm. For temperature-related features, cylinder block temperature, coolant temperature, and exhaust temperature change rate are retained; for mechanical status features, oil pressure, crankshaft speed, and piston vibration amplitude are retained; for environmental features, ambient temperature, atmospheric pressure, and humidity are retained. Principal component analysis is then performed on similar features. Two principal components are extracted for the temperature category, and sub-weights are assigned based on the variance contribution rate after calculating the scores of each principal component. The principal component score calculation is based on a linear transformation formula: ； in, This represents the score of the k-th principal component; This represents the loading coefficient of the k-th principal component on the i-th original feature; This represents the i-th original feature value; n represents the number of features in this class; the weight allocation formula is based on the variance contribution rate: ； in, This represents the weight of the k-th principal component; The eigenvalues ​​of the k-th principal component are represented by m; m represents the total number of principal components. S32. Based on the principal component scores and weights from step S31, a weighted feature association algorithm is used to multiply the comprehensive score of the temperature category by a weight of 0.4, the mechanical state category by a weight of 0.3, and the environmental category by a weight of 0.3, and then concatenate them in dimensional order to form an engine health state vector.

5. The overload protection control method for a commercial vehicle engine according to claim 1, characterized in that, Step S4 includes: The engine health state vector output from step S3 is input into a pre-trained two-factor model. The overload risk index is calculated through weight calculation and dynamic threshold comparison. ； in, Indicates the overload risk index; This represents the threshold deviation weighting coefficient; D represents the normalized threshold deviation. , The dynamic protection threshold for the q-th core parameter in the engine health state vector. This represents the mapping value of the q-th core parameter in the engine health state vector. , This represents the original value of the q-th core parameter in the engine health state vector. This is the maximum value of the core parameter; This represents the weighting coefficient for the acceleration due to temperature change; A represents the acceleration due to temperature change. , This represents the mapping value of the core temperature parameters at the current moment. These are the mapping values ​​of the core temperature parameters at the first and second sampling times, respectively. This is the scaling factor; This indicates the normalized oil pressure.

6. The overload protection control method for a commercial vehicle engine according to claim 5, characterized in that, Step S5 includes: S51. Based on the vehicle speed sequence and throttle opening change data collected in real time in step S1, driving characteristic parameters are extracted through a sliding window with a length of 60 seconds. The driving characteristic parameters include the average vehicle speed fluctuation coefficient, the standard deviation of throttle opening, and the shift frequency. S52. Perform Euclidean distance matching between the extracted driving feature parameters and the pre-stored road condition feature library, and select the scene with the smallest distance and similarity ≥ 0.75 as the current driving scene; the road condition feature library contains standard feature vectors of three typical scenes: highway, national highway and mountain road; S53. Based on the current driving scenario identified in step S52 and the overload risk index output in step S4, the temperature rise coefficient matrix of the corresponding scenario is called. The temperature increase per unit time is calculated through piecewise linear interpolation, and then an ambient temperature correction term is added to generate a temperature trend curve for the next 5 minutes. The temperature rise trend prediction formula is: ； in, This represents the predicted temperature at a future time Δt. This indicates the current measured temperature; Indicates the road condition coefficient; Indicates the load factor; Indicates the weight of throttle opening; Indicates the prediction time step; Indicates ambient temperature.

7. The overload protection control method for a commercial vehicle engine according to claim 6, characterized in that, Step S6 includes: S61. Combining the overload risk index RI output in step S4 and the temperature rise trend output in step S5. The protection level is determined through a tiered triggering mechanism: when the RI is in the 60-75 range, a warning level protection is activated; when the RI is in the 75-90 range, an intervention level protection is activated; when the RI exceeds 90 or... Forced protection will be activated when the temperature is predicted to exceed 130℃ within 5 minutes. ； in, Indicates the protection level; RI indicates the overload risk index; t represents the predicted temperature for the next 5 minutes; t represents the current time. S62. Based on the protection level determined in step S61, generate corresponding control command parameters: For the warning level, the red indicator light on the control panel flashes and the buzzer emits an intermittent warning sound; for the intervention level, send a command to the cooling system via the CAN bus to force the cooling fan to enter full-speed operation mode and increase the flow rate of the cooling pump to 120% of the rated value; for the forced level, the fuel injection system reduces the fuel injection quantity by 15%-20% by adjusting the injection pulse width, while limiting the maximum speed to no more than 2200 rpm.

8. An overload protection control system for a commercial vehicle engine, characterized in that, include: The sensor array, deployed in the engine compartment and body, is used to collect engine block temperature, oil pan area temperature, coolant flow / temperature, oil pressure, ambient temperature, vehicle speed, real-time load, throttle opening and driving operation signals. The preprocessing module, which communicates with the sensor array, is used to preprocess multimodal data, extract the time-domain features of temperature change rate and load fluctuation amplitude, and output a standardized dataset. The weighted fusion module, which communicates with the preprocessing module, is used to perform weighted fusion on the standardized dataset using a weighted feature association algorithm to obtain the engine health status vector. The risk calculation module, which communicates with the weighted fusion module, is used to input the engine health state vector into the pre-trained two-factor model to calculate the overload risk index. The scene recognition and prediction module communicates with the risk calculation module and sensor array. It is used to combine the overload risk index and vehicle speed and throttle opening data, and identify the current driving scene of the commercial vehicle through a pre-stored road condition matching algorithm to predict the temperature rise trend within a preset time in the future. The instruction generation module communicates with the risk calculation module and the scene recognition and prediction module to generate graded protection instructions by combining the overload risk index and the temperature rise trend. The execution control module is connected to the command generation module and is used to link the engine cooling system with the graded protection command to adjust the fan or water pump, control the fuel supply module and the cockpit alarm unit, and execute the corresponding protection actions.

9. A terminal, characterized in that, include: processor; Memory used to store the processor's execution instructions; The processor is configured to perform the method of any one of claims 1-7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the program is executed by the processor, it implements the method as described in any one of claims 1-7.

Images