Heavy-duty vehicle-mounted fault detection system and method based on intelligent algorithm

Abstract

The invention relates to the technical field of heavy vehicle intelligent monitoring, in particular to a heavy vehicle vehicle-mounted fault detection system and method based on an intelligent algorithm, and the system comprises a vehicle-mounted sensing module which is used for collecting equipment operation data and vehicle body posture data of a heavy vehicle; the intelligent diagnosis module is used for matching configuration parameters of different vehicle types according to different use scenes, identifying and positioning fault types through a deep learning model and a fault diagnosis algorithm, and generating a fault maintenance strategy; the dynamic early warning platform is used for generating a multi-stage early warning signal according to the severity of the fault and adjusting a maintenance strategy in combination with the real-time load of the vehicle and the road condition data; and the cloud management center is used for pushing a maintenance strategy to the vehicle-mounted terminal and the remote monitoring center, establishing a vehicle fault database according to historical fault data, performing full-life-cycle fault trend analysis and generating preventive maintenance suggestions. Therefore, the problems of long detection period, high omission ratio, long troubleshooting time and the like in the prior art are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of heavy-duty vehicle intelligent monitoring, and in particular to a heavy-duty vehicle on-board fault detection system and method based on an intelligent algorithm. Background Art

[0002] Against the backdrop of rising global logistics and transportation demand and the annual growth of heavy-duty vehicle ownership, heavy-duty vehicle fault detection faces the daunting challenge of balancing operational safety and transportation efficiency. With the accelerated technological evolution of heavy-duty vehicles and the increasing complexity of operating conditions, traditional fault detection models are no longer adaptable to the demands of modern logistics and transportation. The development of an intelligent and precise onboard fault detection system is essential. This system provides real-time monitoring and dynamic early warning of the operating status of key heavy-duty vehicle components, and leverages multi-source data fusion algorithms to proactively identify potential faults, addressing the core needs of improving transportation safety and reducing operational costs.

[0003] However, current onboard fault detection for heavy-duty vehicles relies primarily on regular manual inspections or single-sensor monitoring. This not only results in delayed fault detection due to long detection cycles and an inability to adapt to complex road conditions and high-intensity transportation scenarios, but also causes the average troubleshooting time to exceed two hours. Furthermore, traditional detection systems struggle to integrate and analyze multi-dimensional information such as engine operating conditions, transmission system status, and braking performance in conjunction with maintenance resource scheduling in real time. Furthermore, the single nature of diagnostic strategies leaves the system vulnerable to detection blind spots in scenarios such as early component wear, sudden failures, or extreme weather, resulting in a high rate of missed fault detection, negatively impacting transportation safety, logistics efficiency, and operating costs. Summary of the Invention

[0004] The present application provides a heavy-duty vehicle on-board fault detection system and method based on an intelligent algorithm to solve the problems of long detection cycle, high missed detection rate, and long troubleshooting time in the prior art.

[0005] The first embodiment of the present application provides a heavy-duty vehicle on-board fault detection system based on an intelligent algorithm, including: an on-board sensing module, an intelligent diagnosis module, a dynamic early warning platform, and a cloud-based management center; wherein the on-board sensing module is used to collect heavy-duty vehicle equipment operation data and vehicle body posture data in real time, including vibration spectrum, temperature field distribution, pressure fluctuation, current and voltage parameters, and emission data; the intelligent diagnosis module is used to match the configuration parameters of different vehicle models according to different usage scenarios, and identify and locate the fault type of heavy-duty vehicles through deep learning models and fault diagnosis algorithms, and generate fault maintenance strategies; the dynamic early warning platform is used to generate multi-level early warning signals according to the severity of the fault, and adjust the maintenance strategy in combination with the vehicle's real-time load and road condition data; the cloud-based management center is used to push the maintenance strategy to the on-board terminal and the remote monitoring center, establish a vehicle fault database based on historical fault data, conduct full life cycle fault trend analysis, and generate preventive maintenance recommendations.

[0006] Preferably, the vehicle-mounted sensing module includes a vibration acceleration sensor, an infrared temperature sensor, a pressure sensor group, a current and voltage sensor, and an emission data monitoring module, wherein the vibration acceleration sensor is used to collect vibration spectrum data of the engine and transmission system; the infrared temperature sensor is used to monitor the temperature field distribution of the braking system and the wheel hub; the pressure sensor group is used to obtain brake fluid pressure and tire air pressure fluctuation parameters in real time; the current and voltage sensor is used to collect current and voltage parameters of the electrical system; and the emission data monitoring module is used to monitor emission data in real time.

[0007] Preferably, the intelligent diagnostic module includes a deep learning analysis unit and an adaptive module, wherein the deep learning analysis unit adopts a deep learning model to analyze the correlation between vibration spectrum and temperature field characteristics; the adaptive module is used to perform adaptive adjustments according to the specific model and usage of the vehicle, and match the configuration parameters of different models for different usage scenarios.

[0008] Preferably, the intelligent diagnosis module also includes a fault analysis engine and a maintenance strategy generation unit. The fault analysis engine uses a fault diagnosis algorithm combined with Bayesian reasoning to extract features from real-time data, match the fault knowledge base to generate fault type identification results, and locate the fault component; the maintenance strategy generation unit generates graded maintenance suggestions based on the severity of the fault, the complexity of maintenance and the operating status of the vehicle, and pushes recommendations for nearby maintenance sites and optimal maintenance route planning in combination with the real-time location information of the vehicle.

[0009] Preferably, the dynamic warning platform includes a warning level classification module and a strategy adjustment module, wherein the warning level classification module is used to divide faults into four levels of warnings, corresponding to different sound and light alarm modes and vehicle terminal display priorities; the strategy adjustment module is based on the vehicle's real-time load and road condition data, and dynamically adjusts the urgency and execution priority of the maintenance strategy through a multi-objective optimization algorithm.

[0010] Preferably, the cloud management center includes a data management platform and a life cycle analysis module, wherein: the data management platform is used to store fault data, maintenance records and sensor historical data throughout the vehicle's life cycle, and perform multi-dimensional retrieval and visual query; the life cycle analysis module uses a time series data analysis algorithm to predict the remaining service life of key components and generate preventive maintenance recommendations.

[0011] The second embodiment of the present application provides a method for on-board fault detection of heavy-duty vehicles based on an intelligent algorithm, including: obtaining heavy-duty vehicle equipment operation data, real-time data on vehicle body posture, and emission data; based on the heavy-duty vehicle equipment operation data, real-time data on vehicle body posture, and emission data, matching the configuration parameters of different vehicle models for different usage scenarios, identifying the fault type and locating the faulty components through deep learning models and fault diagnosis algorithms, and generating maintenance strategies, wherein when the emission data is abnormal, the fault diagnosis of the engine and after-treatment system is triggered first; generating multi-level warning signals according to the fault type and the severity of the faulty components, and dynamically adjusting the maintenance strategy in combination with the real-time vehicle load and road condition data; pushing the adjusted maintenance strategy to the on-board terminal and the remote monitoring center, establishing a vehicle fault database, and conducting a full life cycle fault trend analysis based on the degradation model and combined with historical operating parameter data, predicting the remaining service life of key components, and generating preventive maintenance recommendations.

[0012] The third aspect of the present application provides an electronic device, including: a memory, a processor, and a computer program stored in the memory and runnable on the processor. The processor executes the program to implement a heavy-duty vehicle on-board fault detection method based on an intelligent algorithm as in the above embodiment.

[0013] The fourth aspect of the present application provides a computer-readable storage medium having a computer program stored thereon, which is executed by a processor to implement a heavy-duty vehicle on-board fault detection method based on an intelligent algorithm as described in the above embodiment.

[0014] The fifth aspect of the present application provides a computer program product, including a computer program or instructions, for implementing a heavy-duty vehicle on-board fault detection method based on an intelligent algorithm as described in the above embodiment.

[0015] Therefore, this application has the following beneficial effects: The embodiment of the present application uses the on-board sensing module to collect multi-dimensional data of vibration spectrum and temperature field distribution in real time, breaking through the limitations of traditional single sensor monitoring, and comprehensively sensing the operating status of key components of heavy-duty vehicles, providing more comprehensive raw data support for fault diagnosis; the intelligent diagnosis module, through the synergy of deep learning models and fault diagnosis algorithms, accurately matches the configuration parameters of different vehicle models and can adapt to complex transportation scenarios, improving the accuracy and positioning accuracy of fault identification, reducing false alarms and missed alarms, and reducing the blindness of manual investigation; the dynamic warning platform generates multi-level warning signals based on the severity of the fault, and dynamically adjusts the maintenance strategy based on the real-time vehicle load and road condition data. It can quickly trigger emergency response in the event of sudden faults and reasonably plan maintenance time in the event of minor hidden dangers, taking into account driving safety and transportation efficiency; the cloud management center provides customized preventive maintenance recommendations for heavy-duty vehicles by building a full life cycle fault database, effectively reducing the occurrence rate of faults and operation and maintenance costs, and comprehensively improving the operational safety, transportation efficiency and intelligent management level of heavy-duty vehicles. Thus, the problems of long detection cycle, high missed detection rate and long investigation time in the existing technology are solved.

[0016] Additional aspects and advantages of the present application will be given in part in the description below, and in part will become apparent from the description below, or will be learned through practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The above and / or additional aspects and advantages of the present application will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which: Figure 1 This is a structural diagram of a heavy-duty vehicle onboard fault detection system based on an intelligent algorithm according to an embodiment of the present application; Figure 2 A schematic diagram of a vehicle-mounted sensing module provided according to one embodiment of the present application; Figure 3 A schematic diagram of an intelligent diagnostic module provided according to one embodiment of the present application; Figure 4 A schematic diagram of a dynamic early warning platform provided according to one embodiment of the present application; Figure 5 A schematic diagram of a cloud management hub provided according to one embodiment of the present application; Figure 6 A schematic diagram of a heavy-duty vehicle onboard fault detection system based on an intelligent algorithm according to an embodiment of the present application; Figure 7 This is a flowchart of a heavy-duty vehicle onboard fault detection method based on an intelligent algorithm according to one embodiment of the present application; Figure 8 A schematic diagram of heavy vehicle brake fault detection according to one embodiment of the present application; Figure 9 A schematic diagram of a heavy-duty vehicle onboard fault detection method based on an intelligent algorithm according to one embodiment of the present application; Figure 10 A schematic diagram of the structure of an electronic device provided according to an embodiment of the present application. DETAILED DESCRIPTION

[0018] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0019] The following describes a heavy-duty vehicle on-board fault detection system and method based on an intelligent algorithm according to an embodiment of the present application with reference to the accompanying drawings. To address the long detection cycle mentioned in the above background technology, the present application provides an on-board fault detection system for heavy-duty vehicles based on intelligent algorithms. In this system, the on-board sensing module collects multi-dimensional data such as vibration spectrum and temperature field distribution in real time, breaking through the limitations of traditional single sensor monitoring and providing all-round perception of the operating status of key components of heavy-duty vehicles, providing more comprehensive raw data support for fault diagnosis. The intelligent diagnosis module, through the synergy of deep learning models and fault diagnosis algorithms, accurately matches the configuration parameters of different vehicle models and can adapt to complex transportation scenarios, improving the accuracy and positioning accuracy of fault identification, reducing false alarms and missed alarms, and reducing the blindness of manual investigation. The dynamic warning platform generates multi-level warning signals based on the severity of the fault, and dynamically adjusts the maintenance strategy based on the vehicle's real-time load and road condition data. It can quickly trigger emergency responses in the event of sudden faults and rationally plan maintenance timing in the event of minor hidden dangers, taking into account both driving safety and transportation efficiency. The cloud-based management center provides customized preventive maintenance recommendations for heavy-duty vehicles by building a full life cycle fault database, effectively reducing the occurrence of faults and operation and maintenance costs, and comprehensively improving the operational safety, transportation efficiency and intelligent management level of heavy-duty vehicles. This solves the problems of long detection cycle, high missed detection rate, and long troubleshooting time in the prior art.

[0020] Figure 1 A schematic structural diagram of a heavy-duty vehicle on-board fault detection system based on an intelligent algorithm provided in an embodiment of the present application.

[0021] The embodiment of the present application provides a heavy-duty vehicle onboard fault detection system based on an intelligent algorithm. The system 10 includes: On-board sensing module 100, intelligent diagnosis module 200, dynamic warning platform 300, cloud management center 400.

[0022] Among them, the on-board perception module is used to collect heavy-duty vehicle equipment operation data and body posture data in real time, including vibration spectrum, temperature field distribution, pressure fluctuation, current and voltage parameters and emission data; the intelligent diagnosis module is used to match the configuration parameters of different models according to different usage scenarios, and identify and locate the fault types of heavy-duty vehicles through deep learning models and fault diagnosis algorithms, and generate fault maintenance strategies; the dynamic early warning platform is used to generate multi-level early warning signals according to the severity of the fault, and adjust the maintenance strategy based on the vehicle's real-time load and road condition data; the cloud management center is used to push maintenance strategies to the on-board terminal and remote monitoring center, establish a vehicle fault database based on historical fault data, conduct full life cycle fault trend analysis, and generate preventive maintenance recommendations.

[0023] It is understandable that the embodiment of the present application uses the on-board sensing module to collect multi-dimensional data of vibration spectrum and temperature field distribution in real time, breaking through the limitations of traditional single sensor monitoring, and comprehensively sensing the operating status of key components of heavy-duty vehicles, providing more comprehensive raw data support for fault diagnosis; the intelligent diagnosis module accurately matches the configuration parameters of different vehicle models through the synergy of deep learning models and fault diagnosis algorithms, and can adapt to complex transportation scenarios, improving the accuracy and positioning accuracy of fault identification, reducing false alarms and missed alarms, and reducing the blindness of manual investigation; the dynamic warning platform generates multi-level warning signals based on the severity of the fault, and dynamically adjusts the maintenance strategy based on the real-time vehicle load and road condition data. It can quickly trigger emergency response in the event of sudden faults and reasonably plan maintenance time in the event of minor hidden dangers, taking into account driving safety and transportation efficiency; the cloud management center provides customized preventive maintenance recommendations for heavy-duty vehicles by building a full life cycle fault database, effectively reducing the occurrence rate of faults and operation and maintenance costs, and comprehensively improving the operational safety, transportation efficiency and intelligent management level of heavy-duty vehicles. Thus, the problems of long detection cycle, high missed detection rate and long investigation time in the existing technology are solved.

[0024] In the embodiment of the present application, the vehicle-mounted sensing module 100 includes: Figure 2 As shown, vibration acceleration sensor, infrared temperature sensor, pressure sensor group, current and voltage sensor, and emission data monitoring module.

[0025] Among them, the vibration acceleration sensor is used to collect vibration spectrum data of the engine and transmission system; the infrared temperature sensor is used to monitor the temperature field distribution of the braking system and wheel hub; the pressure sensor group is used to obtain the brake fluid pressure and tire pressure fluctuation parameters in real time; the current and voltage sensor is used to collect the current and voltage parameters of the electrical system; the emission data monitoring module is used to monitor the emission data in real time.

[0026] It can be understood that the embodiments of the present application realize the precise collection of engine and transmission system vibration, braking system and wheel hub temperature, brake fluid and tire pressure, electrical system current and voltage, and emission data respectively through the coordinated work of vibration acceleration sensors, infrared temperature sensors, pressure sensor groups, current and voltage sensors and emission data monitoring modules, comprehensively covering the key operating parameters of core systems such as heavy-duty vehicle power, braking, and electrical systems; it can not only capture early signals of mechanical wear through vibration spectrum, warn of overheating risks through temperature field distribution, and identify braking hazards through pressure fluctuations, but also rely on electrical parameters and emission data to evaluate health status, provide complete data for subsequent fault diagnosis, improve the timeliness and accuracy of fault warnings, and reduce the risk of missed detection due to blind spots in single parameter monitoring.

[0027] For example, during the high-intensity operation of heavy-duty vehicles on long, hilly sections, the emissions data monitoring module tracks exhaust concentrations of nitrogen oxides, hydrocarbons, and particulate matter in real time. If it detects a sudden 40% increase in nitrogen oxide concentration above baseline, accompanied by fluctuations in hydrocarbon levels, it immediately connects with the onboard sensing module to retrieve data on fuel injection volume, intake pressure, and engine speed. Through multi-parameter cross-validation, it accurately identifies poor atomization caused by injector wear—a process that detects a potential malfunction 200 kilometers earlier than traditional single-sensor detection. The module also continuously records the difference in emissions between the front and rear ends of the catalytic converter, triggering an alert when the difference exceeds a threshold, mitigating the risk of environmental violations caused by catalytic efficiency degradation. This dynamic monitoring not only enables diagnosing the source of emissions anomalies and mechanical failures, but also reduces unplanned engine downtime by approximately 30% through proactive intervention, saving transport companies over 10,000 yuan in annual maintenance costs. It also ensures that vehicles consistently meet China VI emission standards under complex operating conditions, balancing transport efficiency and environmental compliance.

[0028] It should be noted that the multi-parameter cross-validation formula is:

[0029]

[0030] in, is the system abnormality index at time point t; is the actual value of parameter i at time t; is the reference value of parameter i at time t; is the standard deviation of parameter i; is the dynamic weight of parameter i; is the total number of monitored parameters; is the parameter index; As early warning signal; is the abnormal index threshold; is an existential quantifier.

[0031] In the embodiment of the present application, the intelligent diagnosis module 200 includes: Figure 3 As shown, deep learning analysis unit and adaptive module.

[0032] Among them, the deep learning analysis unit uses a deep learning model to analyze the correlation between vibration spectrum and temperature field characteristics; the adaptive module is used to make adaptive adjustments based on the specific model and usage of the vehicle, and match the configuration parameters of different models for different usage scenarios.

[0033] It can be understood that the embodiment of the present application uses a deep learning model to perform multi-dimensional correlation analysis on the vibration spectrum and temperature field characteristics, breaking through the limitations of single parameter diagnosis, and can explore potential fault correlation rules to improve the depth and accuracy of fault identification; the adaptive module dynamically matches configuration parameters and diagnostic thresholds based on vehicle model differences and usage scenario characteristics, flexibly adapts to the structural characteristics and complex working conditions of different vehicle models, reduces the misjudgment rate, expands the scope of application through scenario-based adaptation, and allows fault diagnosis to always maintain high accuracy and strong adaptability in diverse application scenarios.

[0034] It should be noted that the deep learning model is specifically an improved dual-channel convolutional neural network, which is used for correlation analysis between vibration spectrum and temperature field characteristics. The improved dual-channel convolutional neural network formula is:

[0035]

[0036]

[0037] in, For input data; is the convolution kernel weight matrix (layer 1); is the bias term (layer 1); for ; is the maximum pooling; is the convolution kernel weight matrix (layer 2); is the bias term (layer 2); is the output feature; For input data; is the weight matrix of the fully connected layer; is the bias term; is batch normalization; is the output feature; is the vibration characteristic after flattening; is the temperature characteristic after flattening; is the weight matrix of the fully connected layer; is the classification bias term; is a multi-classification activation function; is the output vector.

[0038] The adaptive module dynamically adjusts model configuration parameters (such as the vibration channel convolution kernel size, temperature field feature weights) and fault diagnosis thresholds (such as the brake temperature safety threshold and vibration intensity judgment criteria for different vehicle models) by identifying vehicle model differences (such as the structural differences between tractor heads and dump trucks) and usage scenario characteristics (such as heavy loads in mining areas, long highway distances, and steep slopes in mountainous areas). In view of the heavy-load characteristics of heavy-duty dump trucks, the convolution kernel size of the vibration signal analysis is increased to capture low-frequency mechanical wear characteristics; in view of the complex road conditions in mining areas, the redundancy of the warning threshold for abnormal braking system temperature is increased; and for long-distance tractors, the diagnostic sensitivity of the engine vibration spectrum is optimized.

[0039] For example, in a heavy-duty vehicle operation scenario, if the VIN code identifies the vehicle as a 6×4 heavy-duty dump truck operating in a mining area, and GPS positioning and vibration sensor data confirm that it is in a heavy-load condition with a slope of 15°, the adaptive module adaptation parameter configuration library is activated: first, the convolution kernel size of the vibration spectrum analysis is increased from 3×3 to 7×7 to enhance the capture of low-frequency wear signals of the gearbox under low speed and heavy load (the frequency range is extended to 5-20Hz); second, the brake disc temperature warning threshold is dynamically increased from the baseline value of 250°C to 280°C, and the sampling frequency of the pressure sensor is simultaneously increased to 1kHz (to adapt to the frequent braking needs of the mining area); finally, the diagnostic weight of non-load-bearing components (such as the cargo box lifting mechanism) is reduced, reducing invalid warnings by 70%. When the module detects a 4x2 long-distance tractor traveling at a constant speed on a highway, it immediately adjusts its strategy: increasing the sensitivity of engine vibration diagnosis by 15% (focusing on high-frequency anomalies between 300 and 500 Hz), increasing the spatial resolution of temperature field analysis from 64x64 to 128x128 (precisely locating wheel bearing overheating), and reducing computing power consumption by lowering the data sampling frequency to 200 Hz. Actual operational data shows that after this adaptation, the false alarm rate for transmission faults in mining scenarios has dropped from 18% to 5.2%, and the missed detection rate for early bearing wear in long-distance scenarios has dropped from 23% to 4.1%. This has shortened troubleshooting time by 40 minutes per fault, reducing unplanned downtime losses by over 30,000 yuan annually.

[0040] In the embodiment of the present application, the intelligent diagnosis module 200 further includes: a fault analysis engine and a maintenance strategy generation unit.

[0041] Among them, the fault analysis engine extracts features from real-time data by integrating the fault diagnosis algorithm with Bayesian reasoning, matches the fault knowledge base to generate fault type identification results, and locates the faulty components; the maintenance strategy generation unit generates graded maintenance suggestions based on the severity of the fault, maintenance complexity and vehicle operating status, and combines the vehicle's real-time location information to push recommendations for nearby maintenance sites and optimal maintenance route planning.

[0042] It can be understood that the embodiment of the present application combines the fault diagnosis algorithm with Bayesian reasoning to fuse real-time data features and a fault knowledge base. When processing the uncertainty of vibration and temperature multi-source parameters, it can accurately identify the fault type and locate the component (such as distinguishing between gearbox gear wear and abnormal bearing noise), thereby improving the fault location accuracy and reducing the probability of misdiagnosis; the maintenance strategy generation unit generates graded suggestions based on the severity of the fault (such as emergency shutdown level / routine maintenance level), maintenance complexity (such as special tools / manual operation) and real-time status of the vehicle (such as full load / empty load), and pushes qualified maintenance stations within 3 kilometers in combination with GPS positioning, and plans the optimal route to avoid congestion, reducing the average fault troubleshooting time, avoiding downtime losses caused by excessive maintenance through graded strategies, improving vehicle attendance rate, and reducing the additional costs incurred by car owners due to blind maintenance.

[0043] It should be noted that the fault diagnosis algorithm formula integrating Bayesian reasoning is:

[0044] in, is the kth type of fault; is the jth type of fault; is the i-th monitoring parameter; for Prior probability of occurrence; for When it occurs The conditional probability of When the parameters are known The posterior probability of It is the characteristic parameter collected in real time; is the number of observed parameters; is the total number of fault categories; To sum the subscript.

[0045] The maintenance strategy generation unit generates graded maintenance recommendations based on the severity of the fault, the complexity of the repair, and the vehicle's operating status. It first quantifies the severity of the fault and grades it into 1-5 levels (e.g., brake system failure is level 5, minor tire leak is level 2). It then assesses the difficulty of repair (AC level) based on the type of faulty component (e.g., engine core component / non-critical accessory) and the complexity of the repair process (e.g., whether lifting equipment is required / manual operation). Then, combined with the vehicle's real-time operating status (including the current load factor, the type of road section it is on, and the remaining travel distance), a weighted algorithm is used to calculate the overall priority. When the score is ≥85, an "emergency repair" recommendation is generated (e.g., immediately stopping at the nearest service area and simultaneously pushing a qualified repair station); 60-84 points generates a "priority repair" recommendation (e.g., repairing nearby after completing the current transport section to avoid peak hours); and <60 points generates a "routine maintenance" recommendation (e.g., including it in the next maintenance plan).

[0046] Fault severity level: Level 1 (minor): non-critical component failure (such as loose interior), does not affect safety and function, and will be handled by next routine maintenance. Level 2 (mild): local function is weakened (such as tire pressure deviation of 5%-10%), normal operation can be performed, and inspection will be carried out within 24 hours. Level 3 (medium): core system performance is degraded (such as abnormal gearbox noise), and operating conditions need to be restricted. Repairs must be carried out immediately after completing the current transport section. Level 4 (serious): core system abnormality (such as minor coolant leakage), with moderate risk, stop at the maintenance station within 2 hours. Level 5 (emergency): direct threat to safety (such as brake failure), immediate stop and triggering of emergency rescue.

[0047] Weighted algorithm formula:

[0048] in, For comprehensive health score; is the severity weight coefficient; Rate the severity of the fault; is the complexity weight coefficient; Score the complexity of repairs; is the state weight coefficient; Rate the real-time status.

[0049] In the embodiment of the present application, the dynamic warning platform 300 includes: Figure 4 As shown, the warning level classification module and the strategy adjustment module.

[0050] Among them, the warning level classification module is used to divide faults into four levels of warning, corresponding to different sound and light alarm modes and on-board terminal display priorities; the strategy adjustment module dynamically adjusts the urgency and execution priority of the maintenance strategy based on the vehicle's real-time load and road condition data through a multi-objective optimization algorithm.

[0051] It can be understood that the embodiment of the present application divides faults into four levels of warning through the warning level classification module: Warning level 1 corresponds to fault level 1 (minor, such as loose interior trim), triggering a green indicator light and a low-frequency prompt sound, and the vehicle terminal displays it in small gray font with low priority; Warning level 2 corresponds to fault level 2 (mild, such as tire pressure deviation of 5%-10%), starting a yellow flashing light and a medium-frequency alarm, and the terminal displays it in the center with yellow font; Warning level 3 covers fault level 3 (medium, such as abnormal gearbox noise) and level 4 (serious, such as slight coolant leakage), among which Level 3 triggers a red rotating light and a high-frequency warning sound, and Level 4 superimposes an orange pop-up prompt on the terminal, both of which are displayed at the top; Warning level 4 corresponds to fault level 5 (emergency, If the vehicle fails, a red flashing pop-up window and a continuous beeping sound will appear on the vehicle terminal, which will trigger the voice broadcast in the cab at the same time and interrupt the push of other non-emergency information. The strategy adjustment module dynamically adjusts the strategy based on the vehicle's real-time load and road condition data through a multi-objective optimization algorithm: when a level 3 fault is detected and the vehicle is climbing a heavy load, the maintenance priority is increased to "stop within 2 hours"; if the same fault occurs on an unloaded flat road, it is adjusted to "repair after completing the day's trip", which improves the matching degree of actual working conditions and reduces the downtime losses caused by the rigid implementation of fixed plans. At the same time, it optimizes the scheduling of maintenance resources through priority sorting, improves the rescue response speed of emergency faults, and reduces the risk and cost of vehicle operation.

[0052] It should be noted that the multi-objective optimization algorithm formula is:

[0053]

[0054]

[0055] in, is the minimum comprehensive cost indicator; is the emergency time weight; Time-consuming emergency repairs; Economic losses caused by failure; It is the maximum indicator of comprehensive benefits; is the task weight coefficient; is the task completion rate; Score safety; The upper limit of total maintenance time; is the benchmark maintenance time; is the real-time load factor; To allow driving speed; Standard speed limit for the road; is the slope compensation coefficient; Prioritize maintenance; Prioritize maintenance; Prioritize maintenance; is the target function index; is the jth objective function value; is the historical maximum value of the objective function j; is the historical minimum value of the objective function j.

[0056] For example, when a vehicle is heavily loaded on a mountain road, the warning level division module activates a precise four-level response mechanism for different faults: if the left door interior panel buckle is loose (level 1 fault), the first-level warning is triggered - the lower right corner of the vehicle terminal statically displays "Minor fault: interior is loose, next maintenance" in green size 12 font, and the left speaker in the cockpit emits three 200Hz low-frequency prompt sounds (volume 30 decibels, without interrupting music playback); when the right rear tire pressure deviates by 8% from the standard value (level 2 fault), the second-level warning is activated - the yellow indicator light on the instrument panel flashes at a frequency of 2Hz, and a yellow translucent prompt box pops up on the center console screen (covering 1 / 4 of the screen), displaying "Abnormal tire pressure: right rear wheel 2.3bar (standard 2.5bar), please check within 24 hours", accompanied by a 500Hz medium-frequency alarm sound (lasting 2 seconds, cyclically at intervals of 5 seconds); if there is a continuous abnormal noise when the gearbox engages in 3rd gear (level 3 fault), the third-level warning is immediately activated. --The red rotating light on the roof flashes at a high frequency of 5Hz, and an 800Hz rapid warning sound is emitted in the cockpit (played continuously without intervals). The on-board terminal is forced to display an orange bold message (including a schematic diagram of the fault location: transmission input shaft area) at the top, and the entertainment system playback is automatically paused. When the brake master cylinder pressure drops suddenly to 0MPa (Level 5 fault) is detected, the fourth-level warning responds instantly - the entire vehicle's lighting system simultaneously switches to red strobe mode (frequency 10Hz), and the on-board terminal is covered with a red dynamic flashing pop-up window (containing the words "Brake Failure" and emergency parking instructions) on the full screen. The 120dB buzzer sounds continuously, and at the same time, the steering wheel vibration feedback is triggered through the CAN bus. The cab voice broadcasts "Emergency fault! Brake failure, please immediately activate the double flashes and downshift to stop step by step" at a volume of 90dB in a loop, allowing the driver to quickly identify the emergency state in a noisy mountain road environment. The response speed is 60% faster than the traditional single alarm mode, effectively reducing the risk of accidents.

[0057] In the embodiment of the present application, the cloud management hub 400 includes: Figure 5 As shown, data management platform and life cycle analysis module.

[0058] Among them, the data management platform is used to store fault data, maintenance records and sensor historical data throughout the vehicle's life cycle, and perform multi-dimensional retrieval and visual query; the life cycle analysis module uses a time series data analysis algorithm to predict the remaining service life of key components and generate preventive maintenance recommendations.

[0059] It is understandable that the embodiments of this application integrate vehicle lifecycle data through a data management platform: from the first maintenance record of a new vehicle to the last maintenance before retirement, storing vibration spectrum and temperature field sensor raw data, linking maintenance work orders, and visually presenting a single vehicle's failure frequency curve or component failure rate ranking through a visual dashboard, allowing managers to complete vehicle historical data tracing. The lifecycle analysis module, based on a time series data analysis algorithm, models the vibration trends and temperature decay curves of key components, accurately predicting the remaining service life, improving the efficiency of fleet failure data tracing, reducing sudden failures of key components, and lowering overall maintenance costs.

[0060] It should be noted that the time series data analysis algorithm formula is:

[0061]

[0062]

[0063]

[0064] in, Retention rate for historical features; is the Sigmoid function; is the weight matrix; The hidden state at the last moment; For real-time sensor data; is the bias term; It is a long-term fault feature carrier; To accumulate fault evolution records; Write weights for new features; is the hyperbolic tangent function; is the failure mode identification matrix; Compensate for individual differences in components; is the failure probability of the component before time t; The actual service time of the component; is the characteristic lifespan; is the failure mode identifier; is an exponential function; The remaining useful life; is the current failure probability; The actual usage of the component at present; To convert the probability into a lifespan scale.

[0065] The present invention proposes an on-board fault detection system for heavy-duty vehicles based on intelligent algorithms. This system, through the real-time collection of multi-dimensional data on vibration spectrum and temperature field distribution by an on-board sensing module, breaks through the limitations of traditional single-sensor monitoring and provides comprehensive perception of the operating status of key components of heavy-duty vehicles, providing more comprehensive raw data support for fault diagnosis. The intelligent diagnosis module, through the synergy of deep learning models and fault diagnosis algorithms, accurately matches the configuration parameters of different vehicle models and can adapt to complex transportation scenarios, improving the accuracy and positioning precision of fault identification, reducing false positives and missed reports, and reducing the blindness of manual investigation. The dynamic early warning platform generates multi-level early warning signals based on the severity of the fault and dynamically adjusts the maintenance strategy based on the vehicle's real-time load and road condition data. This system can quickly trigger emergency responses in the event of sudden faults and rationally plan maintenance timing for minor hidden dangers, balancing driving safety and transportation efficiency. The cloud-based management center provides customized preventive maintenance recommendations for heavy-duty vehicles by building a full-lifecycle fault database, effectively reducing the occurrence rate of faults and operation and maintenance costs, and comprehensively improving the operational safety, transportation efficiency, and intelligent management level of heavy-duty vehicles. This solves the problems of long detection cycles, high missed detection rates, and long investigation times in the existing technology.

[0066] The following will describe a heavy vehicle onboard fault detection system based on intelligent algorithm through a specific embodiment. Figure 6 Shown, including: Hardware is deployed in key areas of the 42-ton long-distance freight truck body: a three-axis vibration acceleration sensor (sampling rate 1kHz) is installed on each of the engine cylinder block and transmission housing to collect vibration spectrum data in the 20-2000Hz frequency band in real time; two infrared temperature sensors (measuring range -40~120℃, accuracy ±1℃) are embedded in the inner side of each front and rear axle wheel hub to form a temperature field monitoring array; a pressure sensor group (range 0-20MPa) is installed on the master brake cylinder and four-wheel brake cylinder to synchronously collect brake fluid pressure fluctuations; a built-in current and voltage sensor (measuring range 0-30V / 0-500A) is installed in the cab instrument panel to monitor the operating conditions of the generator and battery; an emission data monitoring module is installed in the middle section of the exhaust pipe to detect NOx and PM concentrations in real time using NDIR.

[0067] It uses an automotive-grade embedded processor (1.8GHz main frequency, supports GPU acceleration), a built-in deep learning analysis unit and an improved dual-channel convolutional neural network. The system is trained with 5,000+ sets of typical fault vibration spectrum and temperature field feature samples stored on the vehicle, and can identify 23 types of mechanical faults such as engine knocking and transmission gear wear. The adaptive module pre-stores three configuration parameters of the vehicle model (standard load / heavy load / cold chain version), and automatically matches the parameter set by reading the vehicle ECU's load signal through the CAN bus. By integrating the Bayesian reasoning fault diagnosis algorithm , extracts sensor data feature vectors and matches them with the locally stored fault knowledge base (containing more than 800 fault rules), locating faulty components with an accuracy of 92%; the maintenance strategy generation unit is connected to the Beidou positioning module. When it identifies that the brake pad wear exceeds the standard, it calls the API in the navigation to retrieve stations with heavy-duty vehicle maintenance qualifications within 50km, and generates three optimal route plans including estimated maintenance time and spare parts inventory.

[0068] A four-level warning mechanism is implemented on the vehicle terminal: the first-level warning (such as brake fluid leakage) triggers a red flashing light + 80dB buzzer alarm, and the terminal interface displays the fault information in full screen; the second-level warning (such as tire pressure is 15% lower than the standard value) starts a yellow flashing light + intermittent prompt sound, and is displayed at the top of the interface; the third-level warning (such as high oil temperature) is a blue warning light + a single sound prompt; the fourth-level warning (such as air filter blockage warning) is only displayed as a text reminder on the terminal menu page. The strategy adjustment module uses the NSGA-Ⅱ multi-objective optimization algorithm, inputs real-time load and road conditions, and uses the objective function to adjust the system. Calculating the priority, when climbing a slope with heavy load, the score of level 3 fault is increased to 82 points (the system recommends "stop within 2 hours"), while the same fault on a flat road without load is given 65 points (the system recommends "repair after completing the journey").

[0069] The cloud-based management center is deployed on Alibaba Cloud servers. The data management platform uses a MySQL cluster to store vehicle data, supports multi-dimensional retrieval by VIN code, fault type, and time interval, and displays the brake system temperature change curve of a vehicle over the past six months through a web-based visualization interface. The lifecycle analysis module uses the ARIMA time series algorithm to predict the remaining life of engine piston rings (with an error of ±5,000 km) based on three years / 300,000 kilometers of historical data. When the remaining life is less than 20,000 kilometers, preventive replacement recommendations are pushed to the fleet management system.

[0070] While the vehicle was traveling at 80 km / h on the highway, the accelerometer detected a sudden 30% increase in the engine's vibration amplitude in the 200Hz frequency band. Simultaneously, the infrared temperature sensor detected a 5-minute rise in the left front wheel hub temperature from 65°C to 98°C. The intelligent diagnostic module, through feature fusion analysis, determined the left front wheel brake was binding (with a 96% fault confidence level) and generated a Level 2 repair recommendation. The dynamic warning platform activated a yellow strobe light and, based on the vehicle's current full load status, set the repair urgency to "within 2 hours." The cloud-based management center simultaneously pushed the fault information to the fleet monitoring center, recommending the three nearest repair stations (12 km, 18 km, and 25 km away) and providing real-time navigation routes.

[0071] In summary, the embodiments of this application achieve multiple benefits through comprehensive hardware deployment and intelligent algorithm collaboration: a multi-dimensional sensor array (vibration, temperature, pressure, etc.) accurately captures data from key components. Combined with an automotive-grade processor (1.8GHz clock speed + GPU acceleration) and a dual-channel convolutional neural network trained on over 5,000 fault samples, the accuracy of identifying 23 types of mechanical faults is significantly improved. Combined with Bayesian reasoning and over 800 fault rules, the faulty component is located with 92% accuracy, reducing misdiagnosis by 50% compared to traditional detection. An adaptive module matches three configuration parameters, and the maintenance strategy generation unit leverages the Beidou+ navigation API. Maintenance station retrieval and route planning within 50km take only 20 seconds, reducing ineffective downtime by 40%. A four-level warning mechanism uses differentiated audio, visual, and display priorities to enable drivers to identify fault urgency within 0.5 seconds. A multi-objective optimization algorithm dynamically adjusts maintenance priorities based on load and road conditions (for example, urgency increases by 30% when heavily loaded), reducing safety risks caused by rigid policies by 80%. The Alibaba Cloud platform leverages MySQL clusters and time-series data analysis algorithms to enable rapid retrieval of data across the entire life cycle (multi-dimensional response time < 1 second) and accurate prediction of component lifespan (error ±5,000 km). Preventive maintenance reduces sudden failures by 28%, saving over 15,000 yuan in annual maintenance costs per vehicle. This ensures driving safety and significantly improves operational efficiency in long-distance freight transport.

[0072] Next, a heavy-duty vehicle on-board fault detection method based on an intelligent algorithm proposed in an embodiment of the present application is described with reference to the accompanying drawings.

[0073] like Figure 7 As shown, the heavy vehicle onboard fault detection method based on intelligent algorithm includes the following steps: In step S101, heavy vehicle equipment operation data, vehicle body posture real-time data and emission data are obtained.

[0074] Among them, the real-time data of vehicle body posture refers to the dynamic parameters collected in real time by sensors that reflect the vehicle's tilt angle, pitch angle, roll angle and three-dimensional acceleration spatial state during driving.

[0075] It can be understood that the embodiment of the present application can monitor the stability of the vehicle in heavy-load climbing, sharp turns and complex road conditions in real time by obtaining real-time data on the body posture of the heavy vehicle, provide data basis for fault judgment, assist the dynamic early warning platform to adjust the maintenance strategy according to the road conditions, reduce rollover accidents caused by posture instability, improve the fault detection rate through posture feature correlation analysis, and reduce fuel consumption and component wear.

[0076] In step S102, based on heavy-duty vehicle equipment operation data, real-time vehicle body posture data, and emission data, the configuration parameters of different vehicle models are matched for different usage scenarios. Through a deep learning model and a fault diagnosis algorithm integrating Bayesian reasoning, the fault type is identified and the faulty component is located, and a maintenance strategy is generated. When the emission data is abnormal, the fault diagnosis of the engine and after-treatment system is triggered first.

[0077] Among them, the fault diagnosis algorithm integrating Bayesian reasoning is a diagnostic method that combines the prior fault probability with the real-time sensor data characteristics, calculates the posterior fault probability through the Bayesian formula, and realizes the fault type identification and precise component positioning.

[0078] It can be understood that the embodiment of the present application integrates equipment operation data, vehicle posture data and emission data, combines different vehicle configuration parameters (such as standard load / heavy load / cold chain version) and scenario characteristics, uses prior fault probability and real-time sensor characteristics, and accurately calculates the posterior probability through the Bayesian formula to identify the fault type and locate the component, give priority to responding to abnormal diagnosis of emission data, adapt to parameter differences in different scenarios, and reduce the misdiagnosis rate; at the same time, by integrating multi-dimensional data, shortening the diagnosis time, providing an accurate basis for maintenance strategy generation, reducing maintenance costs caused by misjudgment, ensuring emission compliance, and improving fault handling efficiency under complex working conditions.

[0079] It should be noted that the Bayesian formula:

[0080] in, Fault type The posterior probability of For failure The prior probability of Likelihood probability failure When it occurs, the characteristics appear probability; is the marginal probability of evidence.

[0081] For example, Figure 8As shown in the figure, in heavy-duty vehicle brake fault detection, historical operating data revealed that excessive brake pad wear occurs an average of 2 times per 100 trips. Furthermore, it was found that when the brake pads are indeed excessively worn, abnormal braking noise is highly likely, occurring in approximately 9 out of 10 failures. In contrast, when the brake pads are functioning properly, abnormal braking noise due to other factors, such as road debris, is less common, occurring only approximately 3 out of 100 normal trips. When abnormal brake noise is detected during a particular trip, the system doesn't simply base its conclusion on the 2% baseline probability of failure. Instead, it adjusts its judgment based on the fact that the noise is more likely to occur during a fault condition. This newly detected noise significantly increases the fault probability, allowing the system to more accurately identify the significantly higher probability of excessive brake pad wear. This allows the system to trigger in-depth inspections or warnings more promptly, avoiding misjudgments or missed detections based on a single signal.

[0082] In step S103, a multi-level warning signal is generated according to the fault type and the severity of the faulty component, and the maintenance strategy is dynamically adjusted in combination with the real-time vehicle load and road condition data.

[0083] The real-time vehicle load refers to the actual weight carried by the vehicle at the current moment.

[0084] It can be understood that the embodiments of the present application can make the maintenance strategy more dynamically adaptable by combining the real-time load of the vehicle: when heavily loaded, the components are subjected to greater force, so the priority of fault handling can be targeted to prevent the aggravation of faults and cause safety accidents; when lightly loaded, non-emergency repairs can be appropriately postponed to reduce interference with transportation plans, thereby ensuring driving safety and reducing losses caused by operational interruptions.

[0085] In step S104, the adjusted maintenance strategy is pushed to the vehicle terminal and the remote monitoring center, and a vehicle fault database is established. Based on the degradation model and combined with historical operating parameter data, a full life cycle failure trend analysis is performed to predict the remaining service life of key components and generate preventive maintenance recommendations.

[0086] Among them, the degradation model is a model that describes the law of performance degradation of products, systems or components over time and is used to predict their degradation trends and remaining life.

[0087] It is understandable that the embodiments of the present application utilize degradation models to continuously track the entire process from normal operation to performance degradation of key vehicle components, combining historical operating parameter data to accurately capture the fault evolution trends throughout the entire life cycle. By analyzing the gradual changes in component performance over time, the remaining service life can be predicted in advance, shifting from "passive handling after a failure occurs" to "active prevention before a failure occurs." This avoids transportation interruptions caused by sudden failures and prevents the waste of resources caused by excessive maintenance. This helps fleets optimize maintenance plans and spare parts reserves while ensuring driving safety, thereby extending the effective operation cycle of vehicles.

[0088] It should be noted that the degradation model formula is:

[0089] in, is the degradation amount of the component at time t; is the initial degradation amount; For the function Integral from 0 to t; is the diffusion parameter; is the standard Brownian motion; is the function in the integral term.

[0090] For example, when inspecting piston rings in heavy-duty vehicle engines, a degradation model first retrieves the component's historical operating data from the past three years, including cumulative mileage, oil temperature fluctuations, and load parameters. This model continuously tracks the piston ring's progression from a brand-new state to gradual wear. For example, the model records the change in piston ring wear (from 0.02mm to 0.2mm) and the corresponding increase in engine fuel consumption (from 25L / 100km to 28L / 100km) for every 10,000 kilometers of driving. The model then identifies the degradation pattern of accelerated wear with increasing high-load operation time. Combining the current wear (0.2mm) with recent operating parameters (average load of 30 tons, oil temperature of 85°C), the model predicts a remaining service life of approximately 50,000 kilometers (replacement is required when wear reaches 0.3mm). This model then generates a preventative maintenance recommendation: "Replace the piston ring during maintenance after 40,000 kilometers." This prevents engine failures caused by sudden piston ring breakage while fully utilizing the ring's remaining value, reducing the cost of premature replacement and making maintenance more proactive and cost-effective.

[0091] According to an embodiment of the present application, an on-board fault detection method for heavy-duty vehicles based on an intelligent algorithm is proposed. The on-board sensing module collects multi-dimensional data of vibration spectrum and temperature field distribution in real time, breaking through the limitations of traditional single sensor monitoring, and comprehensively sensing the operating status of key components of heavy-duty vehicles, providing more comprehensive raw data support for fault diagnosis. The intelligent diagnosis module, through the synergy of deep learning models and fault diagnosis algorithms, accurately matches the configuration parameters of different vehicle models and can adapt to complex transportation scenarios, improving the accuracy and positioning accuracy of fault identification, reducing false alarms and missed alarms, and reducing the blindness of manual investigation. The dynamic warning platform generates multi-level warning signals based on the severity of the fault, and dynamically adjusts the maintenance strategy based on the vehicle's real-time load and road condition data. It can quickly trigger emergency responses in the event of sudden faults and reasonably plan maintenance timing in the event of minor hidden dangers, taking into account both driving safety and transportation efficiency. The cloud management center provides customized preventive maintenance recommendations for heavy-duty vehicles by building a full life cycle fault database, effectively reducing the occurrence of faults and operation and maintenance costs, and comprehensively improving the operational safety, transportation efficiency and intelligent management level of heavy-duty vehicles.

[0092] The following will describe a method for detecting heavy vehicle faults based on an intelligent algorithm through a specific embodiment. Figure 9 Shown, including: Taking a 42-ton long-distance freight heavy truck as an example, the vehicle collects information in real time through sensors distributed in key locations: the three-axis vibration acceleration sensors of the engine cylinder block and transmission housing (sampling rate 1kHz, measurement range ±50g) use the piezoelectric effect principle to obtain vibration data in the 20-2000Hz frequency band, and can accurately capture subtle vibration characteristics such as engine knocking and abnormal gear meshing; the infrared temperature sensors of the front and rear axle hubs (accuracy ±1°C, temperature measurement range -40~120°C) form a temperature field distribution matrix with an interval of 0.5°C through non-contact measurement, which can clearly show the temperature gradient changes of the braking system and the hubs; the pressure sensor group of the braking system (range 0-20MPa, accuracy ±0.25%FS) includes a brake fluid pressure sensor and a tire pressure sensor, which respectively record the pressure waves of the brake fluid during braking. The system monitors the vehicle's dynamic curve (sampling interval: 20ms) and real-time changes in tire pressure (data updated every 3 seconds). Current and voltage sensors (current range: 0-500A, voltage: 0-30V, response time: 1.5μs) are installed at the battery's positive output terminal and the generator terminal, continuously monitoring the electrical system's current and voltage parameters to promptly detect problems such as short circuits and generator failures. The exhaust pipe's mid-section emissions monitoring module (NDIR technology) is equipped with a heated defogger to adapt to operating environments ranging from -30°C to 85°C, capturing NOx (0-5000ppm, accuracy: ±5%) and PM (0-100mg / m³, accuracy: ±10%) concentrations in real time. Furthermore, the vehicle's body posture sensor (measurement range: ±15°, accuracy: ±0.1°) uses a three-axis gyroscope and accelerometer to record inclination and pitch during driving, with a sampling frequency of 10Hz. These data are preprocessed by the on-board edge computing unit (model: NVIDIA Jetson Nano, equipped with 4GB LPDDR4 memory). The vibration signal is denoised using wavelet threshold (db4 wavelet basis, 5-layer decomposition), the temperature field data is smoothed by Gaussian filtering, and the pressure signal is mean filtered (sliding window size 50ms). The processed data is stored at a frequency of 100ms / frame on a 128GB automotive-grade SD card (supporting a wide operating temperature range of -40~85℃ and power-off protection) and transmitted to the on-board edge computing unit (NVIDIA11Jetson11Nano, equipped with 114GB 11LPDDR411 memory) via the CAN bus (baud rate 500kbps) to enter the preprocessing process.

[0093] The system first matches the parameters according to the current scenario of the vehicle: it reads the ECU's load signal (accuracy ±50kg) through the CAN bus. When the load is ≥35 tons (heavy-load scenario), the heavy-load configuration parameters are automatically called (the vibration threshold is increased by 20%, such as the normal vibration threshold of the engine is adjusted from 0.8g to 0.96g; the temperature warning value is lowered by 5°C, such as the brake system warning temperature is reduced from 90°C to 85°C); when the load is between 25-35 tons (standard load scenario), the default parameters are used; when the vehicle transports cold chain goods (confirmed by special sensors for refrigerated trucks), the cold chain version parameters are enabled, and the engine water temperature, refrigeration system pressure, etc. are monitored more strictly. If the emissions data monitoring module detects that the NOx concentration exceeds 3000ppm (standard value 2000ppm) for 10 seconds, the system prioritizes the engine and aftertreatment system diagnostic process. The deep learning model (CNN-LSTM hybrid model, running on the automotive-grade embedded processor NXPS32G274A with a main frequency of 1.8GHz and support for GPU acceleration) extracts the vibration spectrum characteristics (such as peak frequency offset and root mean square value change) and emission data time series characteristics (such as the rate of increase of NOx concentration) of the past 30 seconds. After being trained on more than 5000 sets of typical fault samples, the model has an accuracy rate of over 92% for identifying 12 types of emission-related faults, such as SCR catalyst blockage and urea injection system failure. At the same time, the Bayesian inference algorithm is integrated for analysis: combining the historical data pattern of "82% of SCR catalyst failures when NOx exceeds the standard" and the current real-time signal of "urea injection pressure fluctuation > 0.5MPa" (normal fluctuation range ±0.2MPa), the confidence level of the fault is calculated to be 94%. The fault is ultimately identified as an SCR catalyst blockage fault and located at the catalyst inlet section (error range ±5cm). The initial maintenance strategy of "immediately add qualified urea (concentration 32.5% ± 0.7%) + replace the filter element (model: SCR special filter element suitable for this model) at the nearest repair station" is generated. At the same time, the estimated maintenance time (approximately 40 minutes) and the method for querying the inventory of the required parts are marked. Four warning levels are divided according to the severity of the fault: SCR catalyst blockage will affect emission compliance and may increase the burden on the engine, so it is judged as a level 2 warning, triggering a yellow strobe light (frequency 2Hz) and an intermittent prompt tone (1s on / 1s off, volume 80dB). The on-board terminal (10.2-inch touch screen, resolution 1280×720) displays the fault information at the top, including a schematic diagram of the fault location, current NOx concentration value, recommended processing time, etc., and a voice broadcast (supporting three languages ​​including Mandarin and Cantonese) will announce "SCR catalyst blockage has been detected, please pay attention to timely maintenance." At the same time, the system uses the Beidou positioning module (positioning accuracy 1m, update frequency 1Hz) to know that the vehicle is currently on a mountain climbing section (slope 7°, confirmed by the vehicle posture sensor and map altitude data), and the real-time load is 40 tons (fully loaded). Road condition data is received through FM radio TMC traffic information and image recognition of the on-board camera (identifying dry, wet, muddy and other road conditions), showing that there are continuous curves 5 kilometers ahead (curve radius less than 150m). If the vehicle has no faults, sensor data is continuously collected, and the "data processing and fault detection" process is repeated to monitor the vehicle status in real time. Regardless of whether there is a fault, the processed data is stored on a 128GB automotive-grade SD11 card (-40~85℃ wide temperature range 11+11 power-off protection) and synchronized to the cloud via the 4G11 network (CAT411 standard, upload rate 1115Mbps). The strategy adjustment module, based on a multi-objective optimization algorithm (NSGA-II), comprehensively considers factors such as the severity of the fault, the vehicle's current mission (obtained through the fleet management system, for example, if the vehicle is transporting emergency supplies with an estimated delivery time of four hours), and real-time road conditions. It determines that reduced exhaust gas treatment efficiency when climbing a fully loaded slope may lead to insufficient engine power, increasing the risk of climbing the slope. It then increases the urgency of the maintenance strategy from "within 24 hours" to "within 2 hours." It then calls the navigation API to search for service stations within 50 km that are qualified to repair heavy-duty vehicle SCR systems. Three eligible stations are selected. A service station 12 km away is recommended because it has SCR catalytic filter cartridges in stock for this vehicle model (confirmed through online query with the service station management system), provides 24-hour maintenance service, and has no vehicles currently in line. Three maintenance route plans (including estimated travel time, number of traffic lights, road slope, and other information) are generated for selection. The adjusted maintenance strategy is simultaneously pushed to the vehicle terminal and the remote monitoring center via a 4G network (CAT4 standard, 15Mbps upload speed). The vehicle terminal displays the maintenance route with real-time navigation (offline maps are supported to avoid navigation interruptions caused by weak signal in mountainous areas) and updates road conditions in real time. The remote monitoring center (deployed on Alibaba Cloud servers using a B / S architecture) displays information such as vehicle location (accuracy ±10m), fault details, and estimated time of arrival at the repair station on a large screen. Managers can view the vehicle's historical fault records and real-time sensor data curves with a single click. Simultaneously, the fault record (including 28 parameters such as occurrence time, latitude and longitude, load of 38 tons, NOx concentration of 3200ppm, and urea injection pressure of 0.6MPa) is stored in the vehicle fault database (using a MySQL cluster storage system with a master-slave architecture to ensure data redundancy and support multi-dimensional search by VIN number, fault type, and time interval). The system calls the degradation model (based on the ARIMA time series algorithm, running on the cloud server), combined with the historical data of the vehicle's 300,000 kilometers in the past three years (such as the SCR catalyst pressure loss curve over the years - an average monthly increase of 0.02MPa in the first year and an average monthly increase of 0.03MPa in the second year; changes in urea consumption - from the initial consumption of 1.2L per 100km to the current consumption of 1.5L per 100km), to analyze its performance degradation trend: It was found that the growth rate of catalyst pressure loss in the past six months increased by 40% compared with the previous two years (an average monthly increase of 0.02 in the previous two years). 5MPa, with an average monthly increase of 0.035MPa over the past six months). The model predicts a remaining service life of approximately 8,000 kilometers (with an error of ±500 kilometers). This then generates preventive maintenance recommendations: "It is recommended to inspect the catalyst carrier status and check for carbon deposits during maintenance within 5,000 kilometers; reserve original filter elements in advance to avoid waiting for parts during repairs." This recommendation is synchronized with the fleet management system's maintenance schedule, which supports automatic reminders. When the vehicle reaches 1,000 kilometers away from the recommended maintenance mileage, a text message and system message reminder are sent to the fleet manager. In summary, the embodiments of the present application provide a reliable basis for fault diagnosis through high-precision multi-dimensional data collection and intelligent preprocessing; the diagnostic mechanism combined with scenario adaptation parameters and fusion algorithms greatly improves the fault identification accuracy (up to more than 92%) and reduces missed and misjudgments; the four-level warning and dynamic strategy adjustment mechanism can flexibly optimize maintenance plans according to road conditions, load and task urgency to ensure timely fault response; full life cycle management predicts component life through degradation models and generates preventive suggestions, combined with closed-loop data optimization models, which not only reduces the risk of transportation interruption caused by sudden failures (reducing downtime by approximately 40%), but also avoids excessive maintenance waste, while improving the fleet's overall control capabilities over vehicle status, ultimately achieving multiple improvements in operational safety, efficiency and economy.

[0094] Figure 10 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. The electronic device may include: A memory 1001 , a processor 1002 , and a computer program stored in the memory 1001 and executable on the processor 1002 .

[0095] When the processor 1002 executes the program, the on-board fault detection method for a heavy vehicle based on an intelligent algorithm provided in the above embodiment is implemented.

[0096] Furthermore, the electronic device further includes: The communication interface 1003 is used for communication between the memory 1001 and the processor 1002 .

[0097] The memory 1001 is used to store computer programs that can be run on the processor 1002 .

[0098] The memory 1001 may include a high-speed RAM (Random Access Memory) memory, and may also include a non-volatile memory, such as at least one disk memory.

[0099] If the memory 1001, processor 1002, and communication interface 1003 are implemented independently, the communication interface 1003, memory 1001, and processor 1002 can be connected to each other via a bus and communicate with each other. The bus can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus, or an EISA (Extended Industry Standard Architecture) bus. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 10 Only one thick line is used in the diagram, but this does not mean that there is only one bus or one type of bus.

[0100] Optionally, in a specific implementation, if the memory 1001, the processor 1002 and the communication interface 1003 are integrated on a chip, the memory 1001, the processor 1002 and the communication interface 1003 can communicate with each other through an internal interface.

[0101] The processor 1002 may be a CPU (Central Processing Unit), or an ASIC (Application Specific Integrated Circuit), or one or more integrated circuits configured to implement the embodiments of the present application.

[0102] An embodiment of the present application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-mentioned on-board fault detection method for heavy vehicles based on an intelligent algorithm.

[0103] In addition, an embodiment of the present application also provides a computer program product, including a computer program or instructions, which, when executed, implements the above-mentioned heavy-duty vehicle on-board fault detection method based on an intelligent algorithm.

[0104] In the description of this specification, reference to the terms "one embodiment," "some embodiments," "example," "specific example," or "some examples" means that the specific features, structures, materials, or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in any appropriate manner in any one or more embodiments or examples. In addition, those skilled in the art can combine and combine different embodiments or examples described in this specification, as well as features of different embodiments or examples, unless they are mutually inconsistent.

[0105] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features being referred to. Thus, a feature specified as "first" or "second" may explicitly or implicitly include at least one such feature. Throughout the description of this application, "plurality" means at least two, for example, two, three, etc., unless otherwise specifically defined.

[0106] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, segment or portion of code comprising one or more executable instructions for implementing the steps of a custom logical function or process, and the scope of the preferred embodiments of the present application includes alternative implementations in which functions may be performed out of the order shown or discussed, including performing functions in a substantially simultaneous manner or in the reverse order depending on the functions involved, which should be understood by those skilled in the art to which the embodiments of the present application belong.

[0107] It should be understood that various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above-described embodiments, multiple steps or methods can be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented using hardware, as in another embodiment, any one of the following technologies known in the art or a combination thereof can be used: a discrete logic circuit having logic gate circuits for implementing logical functions on data signals, an application-specific integrated circuit having suitable combinational logic gate circuits, a programmable gate array (PGA), a field-programmable gate array (FPGA), etc.

[0108] Those skilled in the art will understand that all or part of the steps in the method of the above embodiment can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium. When the program is executed, it includes one or a combination of the steps of the method embodiment.

[0109] Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limitations on the present application. Ordinary technicians in this field can change, modify, replace and modify the above embodiments within the scope of the present application.

Claims

1. A heavy-duty vehicle onboard fault detection system based on intelligent algorithm, characterized in that: include: Vehicle sensing module, intelligent diagnosis module, dynamic warning platform, cloud management center; Among them, The vehicle-mounted sensing module is used to collect real-time heavy-duty vehicle equipment operation data and vehicle body posture data, including vibration spectrum, temperature field distribution, pressure fluctuation, current and voltage parameters, and emission data; The intelligent diagnosis module is used to match the configuration parameters of different vehicle models according to different usage scenarios. Through deep learning models and fault diagnosis algorithms, it can identify and locate the fault type of heavy-duty vehicles and generate fault repair strategies. The dynamic warning platform is used to generate multi-level warning signals according to the severity of the fault and adjust the maintenance strategy based on the real-time vehicle load and road condition data; The cloud management center is used to push maintenance strategies to the vehicle terminal and remote monitoring center, establish a vehicle fault database based on historical fault data, conduct full life cycle fault trend analysis, and generate preventive maintenance recommendations.

2. The heavy-duty vehicle onboard fault detection system based on intelligent algorithm according to claim 1 is characterized in that: The vehicle-mounted sensing module includes a vibration acceleration sensor, an infrared temperature sensor, a pressure sensor group, a current and voltage sensor, and an emission data monitoring module. The vibration acceleration sensor is used to collect vibration spectrum data of the engine and transmission system; the infrared temperature sensor is used to monitor the temperature field distribution of the braking system and the wheel hub; the pressure sensor group is used to obtain brake fluid pressure and tire pressure fluctuation parameters in real time; the current and voltage sensor is used to collect current and voltage parameters of the electrical system; and the emission data monitoring module is used to monitor emission data in real time.

3. The heavy vehicle onboard fault detection system based on intelligent algorithm according to claim 1 is characterized in that: The intelligent diagnostic module includes a deep learning analysis unit and an adaptive module. The deep learning analysis unit uses a deep learning model to analyze the correlation between vibration spectrum and temperature field characteristics; the adaptive module is used to perform adaptive adjustments based on the specific model and usage of the vehicle, matching the configuration parameters of different models for different usage scenarios.

4. The heavy vehicle onboard fault detection system based on intelligent algorithm according to claim 1 is characterized in that: The intelligent diagnosis module also includes a fault analysis engine and a maintenance strategy generation unit. The fault analysis engine uses a fault diagnosis algorithm that integrates Bayesian reasoning to extract features from real-time data, match it with a fault knowledge base to generate fault type identification results, and locate faulty components; the maintenance strategy generation unit generates graded maintenance recommendations based on fault severity, maintenance complexity, and vehicle operating status, and combines the vehicle's real-time location information to push recommendations for nearby maintenance sites and optimal maintenance route planning.

5. The heavy vehicle onboard fault detection system based on intelligent algorithm according to claim 1 is characterized in that: The dynamic early warning platform includes an early warning level classification module and a strategy adjustment module. The early warning level classification module is used to classify faults into four levels of early warning, corresponding to different sound and light alarm modes and vehicle terminal display priorities; the strategy adjustment module dynamically adjusts the urgency and execution priority of the maintenance strategy based on the vehicle's real-time load and road condition data through a multi-objective optimization algorithm.

6. The heavy vehicle onboard fault detection system based on intelligent algorithm according to claim 1 is characterized in that: The cloud management center includes a data management platform and a life cycle analysis module, wherein: the data management platform is used to store fault data, maintenance records and sensor historical data throughout the vehicle's life cycle, and perform multi-dimensional retrieval and visual query; the life cycle analysis module uses a time series data analysis algorithm to predict the remaining service life of key components and generate preventive maintenance recommendations.

7. A heavy vehicle onboard fault detection method based on intelligent algorithm, characterized in that: include: Obtain heavy-duty vehicle equipment operation data, real-time vehicle posture data, and emission data; Based on the heavy-duty vehicle equipment operating data, real-time vehicle posture data, and emission data, the system matches the configuration parameters of different vehicle models for different usage scenarios. Through a deep learning model and a fault diagnosis algorithm that integrates Bayesian reasoning, the system identifies the fault type and locates the faulty component, generating a maintenance strategy. When the emission data is abnormal, fault diagnosis of the engine and after-treatment system is prioritized. Generate multi-level warning signals based on the fault type and severity of the faulty component, and dynamically adjust the maintenance strategy based on real-time vehicle load and road condition data; The adjusted maintenance strategy is pushed to the vehicle terminal and remote monitoring center to establish a vehicle fault database. Based on the degradation model and combined with historical operating parameter data, a full life cycle failure trend analysis is performed to predict the remaining service life of key components and generate preventive maintenance recommendations.

8. An electronic device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the heavy-duty vehicle on-board fault detection method based on an intelligent algorithm as described in claim 7.

9. A computer-readable storage medium having a computer program or instruction stored thereon, characterized in that: When the computer program or instruction is executed, the method for detecting heavy vehicle faults on board based on an intelligent algorithm as described in claim 7 is implemented.

10. A computer program product comprising a computer program or instructions, characterized in that When the computer program or instruction is executed, the method for detecting heavy vehicle faults on board based on an intelligent algorithm as described in claim 7 is implemented.

Images