A hybrid vehicle electrical appliance fault diagnosis positioning and repairing method and system and vehicle

Abstract

The application discloses a hybrid vehicle electric appliance fault diagnosis positioning and repairing method and system and a vehicle, relates to the technical field of hybrid vehicle electric appliance control, dynamically selects a corresponding fault determination threshold based on a current power mode, compares the working state signals of various components of the hybrid vehicle, and obtains abnormal signals to be diagnosed; the abnormal signals to be diagnosed are input into a deep learning diagnosis model to obtain fault type identification results and corresponding confidence scores, the final fault type is determined through double diagnosis logic, and the fault level of the final fault type is determined; the feature vector of the fault is obtained from the deep learning diagnosis model, the component that has the fault is determined based on the mapping relationship between the feature vector and the component position; and the component that has the fault is subjected to a hierarchical automatic repairing strategy according to the fault level. The high-voltage accessories, low-voltage electric appliances and hybrid collaborative components are subjected to real-time fault diagnosis, accurate positioning, voice prompting and automatic repairing of self-recovery faults.

Description

Technical Field

[0001] This invention relates to the field of hybrid vehicle electrical control technology, and in particular to a method, system and vehicle for diagnosing, locating and repairing electrical faults in hybrid vehicles. Background Technology

[0002] Hybrid vehicles are equipped with a 12V low-voltage system, a high-voltage battery system, a DC / DC conversion system, and a drive motor system. There are special working conditions such as high and low voltage coordination and dynamic switching of power modes among the subsystems. The number of electrical appliances and system complexity are far higher than those of traditional fuel vehicles.

[0003] In existing technologies, solutions for handling electrical faults in hybrid vehicles mostly employ single fault detection or simple alarm mechanisms, alerting the driver to abnormal information through fault codes or warning lights. Some solutions can only provide limited grading for power component faults, and the switching of power modes causes the normal operating parameters of each component to change dynamically with the operating conditions, making fixed thresholds unsuitable. Furthermore, the severity of faults in various components varies greatly, lacking a unified cross-domain grading evaluation standard, and failing to form a comprehensive fault diagnosis and handling system that simultaneously covers high-voltage accessories, low-voltage electrical appliances, and hybrid collaborative systems. Summary of the Invention

[0004] To overcome the shortcomings of the prior art, the present invention provides a method, system and vehicle for diagnosing, locating and repairing electrical faults in hybrid vehicles, applicable to hybrid electric vehicles and plug-in hybrid electric vehicles. It performs real-time fault diagnosis, precise location and voice prompts for high-voltage accessories, low-voltage electrical appliances and hybrid cooperative components, and performs automatic repair for self-healing faults.

[0005] To achieve the above objectives, one or more embodiments of the present invention provide the following technical solutions: In a first aspect, the present invention provides a method for diagnosing, locating, and repairing electrical faults in hybrid vehicles, including: Acquire the operating status signals of various components in hybrid vehicles; The current power mode of the hybrid vehicle is obtained, and the corresponding fault judgment threshold is dynamically selected based on the current power mode. The working status signal is compared to obtain the abnormal signal to be diagnosed. The abnormal signal to be diagnosed is input into the deep learning diagnostic model to obtain the fault type identification result and its corresponding confidence score. The final fault type is determined through dual diagnostic logic, and its fault level is determined based on the final fault type. The feature vector of the fault is obtained from the deep learning diagnostic model, and the faulty component is determined based on the mapping relationship between the feature vector and the component location. Based on the fault level, a graded automatic repair strategy is implemented for the faulty components.

[0006] In a further technical solution, the various components include low-voltage electrical appliances, high-voltage accessories, and hybrid collaborative systems.

[0007] A further technical solution involves selecting the corresponding fault judgment threshold from the preset pure electric mode threshold group, hybrid mode threshold group, and kinetic energy recovery mode threshold group, based on the current power mode.

[0008] A further technical solution involves setting a threshold switching stabilization time window after detecting a power mode switching signal. Within the threshold switching stabilization time window, the original threshold group remains valid. After the threshold switching stabilization time window ends, the threshold group corresponding to the new power mode is updated synchronously.

[0009] A further technical solution, the dual diagnostic logic, specifically includes: When the confidence score output by the deep learning diagnostic model is greater than or equal to the confidence threshold, the fault type identification result of the deep learning diagnostic model is used as the final fault type. When the confidence score output by the deep learning diagnostic model is less than the confidence threshold, the rule base is triggered to perform secondary verification on the same abnormal signal to be diagnosed. If the verification result of the rule base is consistent with the fault type identification result of the deep learning diagnostic model, the fault type identification result of the deep learning diagnostic model is taken as the final fault type. If they are inconsistent, the verification result of the rule base is taken as the final fault type.

[0010] A further technical solution involves determining the fault level based on the final fault type and according to a fault level determination rule table; the fault level includes transient faults, soft faults, and hard faults.

[0011] In a further technical solution, the feature vector is obtained by the deep learning diagnostic model through spatial feature extraction of the abnormal signal to be diagnosed, including amplitude features, temporal features, communication status features, and command response features.

[0012] A further technical solution is that the mapping relationship between the feature vector and the component position is pre-set during the vehicle design stage and stored in the form of a structured mapping table.

[0013] Further technical solutions will include voice announcements of the faulty component name, fault type, and current processing status.

[0014] A further technical solution involves automatically adjusting the volume of voice broadcasts based on in-vehicle noise signals.

[0015] Further technical solutions include displaying indicator lights for the faulty components according to their fault levels.

[0016] A further technical solution is to perform a reset operation on the drive module of the faulty component when the fault level is transient, clear the fault code, and restore the component to normal operation.

[0017] A further technical solution is to first detect the current vehicle speed when the fault level is a soft fault. If the vehicle speed exceeds the safe speed threshold, the restart operation is temporarily suspended and the vehicle speed is slowed down to less than or equal to the safe speed threshold before a gradual restart is performed.

[0018] A further technical solution is to first verify whether the faulty component is a non-critical high-voltage accessory when the fault level is a hard fault. The power supply circuit of the non-critical high-voltage accessory is disconnected. The power supply of the critical drive component is not disconnected, but is handled by derated operation or functional degradation.

[0019] Secondly, the present invention provides a hybrid vehicle electrical fault diagnosis, location, and repair system, comprising: The status signal acquisition module is configured to acquire the working status signals of various components of the hybrid vehicle; The anomaly detection module is configured to: obtain the current power mode of the hybrid vehicle, dynamically select the corresponding fault detection threshold based on the current power mode, compare the working status signal, and obtain the anomaly signal to be diagnosed. The fault diagnosis module is configured to: input the abnormal signal to be diagnosed into the deep learning diagnostic model, obtain the fault type identification result and its corresponding confidence score, determine the final fault type through dual diagnostic logic, and determine its fault level based on the final fault type. The fault location module is configured to: obtain the feature vector of the fault from the deep learning diagnostic model, and determine the faulty component based on the mapping relationship between the feature vector and the component location; The fault repair module is configured to execute a graded automatic repair strategy for the faulty component based on the fault level.

[0020] Thirdly, the present invention provides a vehicle, including a hybrid vehicle electrical fault diagnosis, location and repair system.

[0021] Fourthly, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the hybrid vehicle electrical fault diagnosis, location and repair method described in the first aspect.

[0022] Fifthly, the present invention provides a computer device, including 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 steps in the hybrid vehicle electrical fault diagnosis, location, and repair method described in the first aspect.

[0023] The above one or more technical solutions have the following beneficial effects: This invention addresses the technical flaw of fixed threshold diagnostic logic in hybrid vehicles, which is prone to misjudgment when dynamically switching between multiple operating conditions such as pure electric, hybrid, and kinetic energy recovery, by dynamically selecting the corresponding fault judgment threshold according to the current power mode. This ensures that the fault judgment threshold matches the actual operating conditions in real time, improving the accuracy of fault diagnosis under all operating conditions. By inputting the abnormal signal to be diagnosed into a deep learning diagnostic model and introducing confidence scoring and dual diagnostic logic, the misjudgment rate in complex and boundary fault scenarios is effectively reduced, overcoming the weakness of existing technologies in identifying multi-module linkage faults with a single rule base matching. By directly obtaining fault feature vectors from the deep learning diagnostic model and determining the faulty component based on the mapping relationship between the feature vectors and the component position, the precise location of the faulty component is achieved, avoiding the deviation introduced by secondary feature extraction. By executing a graded automatic repair strategy according to the fault level, it covers the full range of automatic processing scenarios from minor faults that can be automatically reset to hard faults that require power disconnection, significantly reducing manual intervention and improving driving safety and vehicle availability.

[0024] Advantages of additional aspects of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0025] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0026] Figure 1 This is a flowchart of a method for diagnosing, locating, and repairing electrical faults in hybrid vehicles according to an embodiment of the present invention. Detailed Implementation

[0027] It should be noted that the following detailed descriptions are exemplary and intended to provide further illustration of the invention. Unless otherwise specified, 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.

[0028] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0029] Where there is no conflict, the embodiments and features in the embodiments of the present invention can be combined with each other.

[0030] Hybrid vehicles simultaneously possess a 12V low-voltage system, a high-voltage battery system, a DC / DC conversion system, and a drive motor system. The number and complexity of electrical components far exceed those of traditional gasoline vehicles, and electrical technical problems can only be addressed with a single fault alarm. This invention addresses these pain points by proposing a self-healing, location-based, voice-prompt-enabled intelligent fault handling method and system specifically designed for hybrid vehicles, filling a gap in existing technology.

[0031] Existing technologies can only detect single electrical faults or provide simple alarms. There is no complete control method or system that can simultaneously cover all faults in the high-voltage + low-voltage + hybrid system of hybrid vehicles, perform hierarchical diagnosis, and provide voice positioning and automatic repair. Furthermore, no collaborative control logic has been designed specifically for the characteristics of hybrid vehicles.

[0032] Specifically, the existing technology has the following defects and shortcomings: Faults are only indicated by warning lights or fault codes, which are incomprehensible to ordinary users and make it difficult to quickly locate the fault. Many minor faults such as transient faults, soft faults, loose connections, and module crashes cannot be automatically recovered and must be repaired in-store, increasing the user's time and economic costs. The inability to distinguish between high-voltage dangerous faults and ordinary low-voltage faults can easily lead to user misjudgment and pose safety hazards. It lacks integrated control logic that combines fault location, voice prompts, and automatic repair, resulting in low fault handling efficiency.

[0033] Based on this, this application proposes a hybrid vehicle electrical fault diagnosis, location, and automatic repair control method and system that constructs a fully closed-loop control system of perception, diagnosis, location, repair, and feedback to solve the above-mentioned technical defects.

[0034] To address the aforementioned issues, the overall process of the technical solution provided in this application is as follows: First, the fault acquisition unit performs real-time, full-domain acquisition of the operating status signals of three types of components in the hybrid vehicle: low-voltage electrical appliances, high-voltage accessories, and the hybrid collaborative system. Second, the hybrid central control unit dynamically retrieves the corresponding fault judgment threshold group based on the current power mode, performs adaptive comparison of the acquired signals under operating conditions, and identifies the abnormal signals to be diagnosed. Third, the abnormal signals to be diagnosed are input into a deep learning diagnostic model, and a dual confirmation mechanism combining confidence scoring and rule base secondary verification is used to determine the final fault type and fault level. Subsequently, the faulty component is accurately located through the mapping relationship between fault features and component positions, and the human-machine prompting unit presents the fault information to the driver through voice broadcast and indicator light display. Finally, a graded automatic repair operation is performed according to the fault level, and the results are written into the fault feature database after repair to complete the closed-loop feedback. The core difference between this application and existing technologies lies in the introduction of an adaptive threshold dynamic adjustment mechanism, a confidence scoring and dual confirmation mechanism, and a progressive restart strategy for Level II soft faults, achieving intelligent diagnosis and graded automatic repair of hybrid vehicle electrical faults with full coverage of three domains, high precision, and adaptability.

[0035] The hybrid vehicle electrical fault diagnosis, location, and automatic repair control system implemented in this application includes a power supply unit, a fault acquisition unit, a hybrid central control unit, an automatic repair execution unit, a power supply switching and protection unit, and a human-machine prompting unit.

[0036] The power supply unit includes a 12V low-voltage battery, a high-voltage power battery system, and a DC / DC conversion module. It is electrically connected to the fault acquisition unit, the hybrid central control unit, and various electrical circuits to provide power to the system and all vehicle electrical components, demonstrating the dual high and low voltage power supply characteristics of hybrid vehicles.

[0037] The fault acquisition unit is connected to the corresponding interface signals of low-voltage electrical appliances, high-voltage accessories and various components of the hybrid collaborative system, and collects working status signals in real time and transmits them to the hybrid central control unit.

[0038] The hybrid central control unit communicates with the fault acquisition unit, automatic repair execution unit, power supply switching and protection unit, and human-machine prompting unit, and is the core processing node of the system.

[0039] The automatic repair execution unit is connected to the hybrid central control unit and each electrical module, and is responsible for receiving repair commands and executing corresponding operations.

[0040] The power supply switching and protection unit is electrically connected to the hybrid central control unit, the high-voltage power battery system and the 12V low-voltage battery, and is responsible for performing power supply path switching and fault circuit disconnection.

[0041] The human-machine prompting unit is connected to the hybrid central control unit and includes a voice prompt module and a fault location indicator panel, which are responsible for voice broadcasting and indicator display, respectively, to realize human-machine interaction.

[0042] The various units of the system are connected via wires or a communication bus. After the vehicle is powered on, the system automatically completes initialization and enters real-time monitoring mode. In this embodiment, data acquisition and processing strictly comply with relevant laws and regulations. The acquisition of vehicle electrical operating status data must be carried out within the authorized scope of the vehicle and is only used for fault diagnosis, location, and automatic repair, and not for any other unauthorized purposes.

[0043] It should be understood that Figure 1 This description is for illustrative purposes only and does not imply any limitation on the scope of protection of this application. This application can also be applied to fault management scenarios for other types of hybrid vehicles or multi-electric systems.

[0044] Example 1 like Figure 1 As shown in the figure, this embodiment discloses a method for diagnosing, locating, and repairing electrical faults in hybrid vehicles. The method includes the following steps: S1: Acquire the operating status signals of various components of the hybrid vehicle; In this embodiment, the fault acquisition unit collects signals such as current, voltage, communication status and working commands of each component in real time through sensors and communication interfaces distributed at each electrical circuit, high-voltage accessory interface and hybrid cooperative system communication node.

[0045] The fault acquisition unit collects real-time operating status signals from three core components of the hybrid vehicle: low-voltage electrical appliances, high-voltage accessories, and the hybrid collaborative system. The first category is 12V low-voltage electrical appliances, including headlights, wipers, USB ports, air conditioning, and the central control screen; the second category is high-voltage accessories, including DC / DC converter modules, OBC on-board chargers, PTC heating modules, and air conditioning compressors; the third category is the hybrid collaborative system, including the motor controller, battery management system (BMS), and engine control unit.

[0046] Signal transmission adopts a hierarchical bus architecture. Specifically, the operating status signals of high-voltage accessories and various components of the hybrid collaborative system are transmitted via the CAN bus, while the low-voltage auxiliary electrical signals are transmitted via the LIN bus. Both types of bus signals are timestamped and then converged to the bus interface node of the hybrid central control unit for unified processing to ensure the timing consistency of multiple signals. The collected operating status signals are summarized by the fault acquisition unit and transmitted to the hybrid central control unit for subsequent adaptive threshold comparison of operating conditions.

[0047] Preferably, the CAN bus communication rate is 500kbps; the LIN bus communication rate is 19.2kbps; and the timestamp synchronization accuracy is preferably no more than 1ms.

[0048] By acquiring real-time, full-coverage signals of the operating status of components across three domains and employing a hierarchical bus architecture to achieve time-synchronous aggregation of multiple signals, any electrical malfunction can be detected in a timely manner. This eliminates the blind spots in existing technologies that only perform single-component detection, providing a complete and time-consistent data foundation for subsequent accurate diagnosis.

[0049] Preferably, the fault acquisition unit acquires data at a frequency of 100Hz for high-voltage accessories and hybrid cooperative system components, and at a frequency of 10Hz for low-voltage electrical appliances. That is, it samples the operating status signals of each component 100 times and 10 times per second, respectively. High-frequency acquisition ensures that instantaneous faults on the high-voltage side are not missed, while low-frequency acquisition takes into account the resource consumption of the low-voltage side system.

[0050] S2: Obtain the current power mode of the hybrid vehicle, dynamically select the corresponding fault judgment threshold based on the current power mode, compare the working status signal, and obtain the abnormal signal to be diagnosed. In this embodiment, the hybrid central control unit obtains the current power mode of the hybrid vehicle, dynamically selects the corresponding fault judgment threshold from a set of preset threshold parameters according to the current power mode, compares the working status signal, and obtains the abnormal signal to be diagnosed.

[0051] This step enables dynamic adjustment of adaptive thresholds under specific operating conditions. The hybrid central control unit first identifies the current power mode by reading the power mode status message issued by the powertrain controller (PCU) or vehicle controller (VCU). The power modes include pure electric mode, hybrid mode, and kinetic energy recovery mode. The corresponding fault judgment thresholds are then retrieved from the preset pure electric mode threshold group, hybrid mode threshold group, and kinetic energy recovery mode threshold group, respectively.

[0052] Furthermore, the process for establishing the threshold groups for pure electric mode, hybrid mode, and kinetic energy recovery mode is as follows: During the threshold group establishment phase, for each power mode, extensive operating condition calibration data was collected for each monitored component of the hybrid vehicle (including low-voltage electrical appliances, high-voltage accessories, and components of the hybrid cooperative system) under a controlled real-vehicle test environment. Specifically, in pure electric mode, hybrid mode, and kinetic energy recovery mode, the vehicle was driven under typical load conditions such as idling, low speed, medium speed, high speed, rapid acceleration, and rapid deceleration. The fault acquisition unit continuously collected the current, voltage, communication status, and operating command signals of each component at a specified acquisition frequency, forming a normal operating state signal sample set for each component under each operating condition. The number of samples is preferably no less than 1000 sets to ensure the reliability of statistical estimation.

[0053] After obtaining the normal operating state signal sample set for each component under each power mode, statistical analysis is performed on the sample data of each monitored component under each power mode to calculate the rated operating parameters (i.e., the statistical mean of the normal operating state signal) and the fluctuation range (i.e., the standard deviation) of the normal operating state signal for that component under that mode. Using the rated operating parameters as a benchmark and a certain multiple of the standard deviation as the tolerance boundary, the fault judgment threshold for that component under that power mode is determined.

[0054] The tolerance boundary multiplier is determined based on the component type and fault severity level: for high-voltage accessories and hybrid cooperative system components, three times the standard deviation is preferred as the fault judgment tolerance boundary; for low-voltage electrical components, four times the standard deviation is preferred. A larger multiplier indicates higher tolerance for signal fluctuations and a lower false alarm rate, but a corresponding decrease in sensitivity to minor faults; a smaller multiplier indicates higher diagnostic sensitivity, but a corresponding increase in the false alarm rate. These differentiated settings effectively suppress false alarms caused by fluctuations in normal operating conditions on the low-voltage side while ensuring diagnostic sensitivity on the high-voltage side.

[0055] After the three threshold groups are established, they are stored in the non-volatile memory of the hybrid central control unit in the form of structured parameter tables, forming pure electric mode threshold groups, hybrid mode threshold groups, and kinetic energy recovery mode threshold groups, respectively. Each threshold parameter group includes the rated operating parameters of all monitored components and the corresponding fault judgment thresholds. After mass production of the vehicle, the rated operating parameters and fault judgment thresholds of each component can be iteratively corrected using the fault data accumulated from the actual vehicle through the offline batch fine-tuning mechanism of the fault feature database, further improving the matching accuracy between the threshold groups and actual operating conditions. Thus, the three threshold groups are initially calibrated and stored before the vehicle leaves the factory, and continuously optimized through a closed-loop feedback mechanism during vehicle operation, ensuring the long-term effectiveness of the dynamic adjustment mechanism for adaptive thresholds under operating conditions.

[0056] To avoid misjudgment caused by signal jitter during power mode switching, the hybrid central control unit sets a threshold switching anti-jitter time window of no less than 200ms after detecting the power mode switching signal. During the anti-jitter time window, the original threshold group remains valid. After the anti-jitter window ends, it is synchronously updated to the threshold group corresponding to the new mode, thereby eliminating the risk of diagnostic misjudgment during mode switching.

[0057] Let the current power mode be The corresponding fault determination threshold group is ,in Indicates the first Each component Fault determination threshold in mode Assign part numbers; This indicates the fault acquisition unit that acquired the first fault. Real-time operating status signal values ​​of each component; Indicates the first Each component Rated operating parameters in this mode. When acquiring signals... satisfy When this happens, the hybrid central control unit determines the signal as an abnormal signal to be diagnosed and outputs it to the subsequent dual diagnostic logic processing.

[0058] By setting independent threshold groups for different power modes, the interference of changes in the cooperative working state of high and low voltage electrical appliances during mode switching on fault judgment is eliminated, effectively avoiding misjudgment and missed judgment caused by fixed threshold diagnostic logic during operating condition switching.

[0059] S3: Input the abnormal signal to be diagnosed into the deep learning diagnostic model to obtain the fault type identification result and its corresponding confidence score. Determine the final fault type through dual diagnostic logic, and determine its fault level based on the final fault type. In this embodiment, the hybrid central control unit inputs the abnormal signal to be diagnosed into the deep learning diagnostic model to obtain the fault type identification result and the corresponding confidence score. When the confidence score is lower than the preset confidence threshold, the rule base is triggered to perform secondary verification on the abnormal signal to be diagnosed. The secondary verification result is double-confirmed with the fault type identification result to obtain the final fault type and fault level.

[0060] This application introduces dual diagnostic logic to determine the final fault type. The deep learning diagnostic model built into the hybrid central control unit adopts a CNN-LSTM hybrid architecture: the CNN convolutional layer extracts spatial features from the multidimensional features of the abnormal signal to be diagnosed (current waveform, voltage fluctuation, communication frame anomalies, etc.), the LSTM recurrent layer models the temporal anomaly pattern of the signal, and finally the fully connected classification layer outputs the fault type identification result and the corresponding confidence score. , The value range is [0,1]. The closer the value is to 1, the higher the model's confidence in the recognition result.

[0061] Let the preset confidence threshold be... ,when When the deep learning diagnostic model's identification result is used directly as the final fault type, then... At that time, the built-in rule base of the hybrid central control unit is triggered to perform secondary verification on the same abnormal signal to be diagnosed.

[0062] The rule base consists of expert rules and threshold logic rules. It defines the judgment conditions for various typical faults in the form of IF-THEN. For example, IF: The current output value of the high-voltage accessory is zero for N consecutive frames AND the corresponding voltage is abnormal THEN: It is judged as a high-voltage accessory power supply fault. After the rule base traverses all applicable rules, it outputs the matching fault type as the secondary verification result.

[0063] The validation results from the rule base and the identification results from the deep learning diagnostic model are double-verified: if they match, the result is taken as the final fault type; if they do not match, the validation result from the rule base prevails, and the confidence level of the diagnostic result is marked as pending verification. Simultaneously, the abnormal signal is included in the subsequent offline model optimization dataset. After comprehensively determining the final fault type, the fault level is classified according to the severity and scope of the fault.

[0064] Once the final fault type is determined, the hybrid central control unit classifies the fault into three levels based on its severity and scope of impact: Level I transient faults refer to faults such as transient electrical overcurrent, signal interference, and loose wiring vibrations that can be quickly recovered and do not affect normal vehicle operation; Level II soft faults refer to faults such as electrical module crashes, communication loss, and initialization abnormalities that can be recovered by restarting or reconfiguration; Level III hard faults refer to faults such as electrical component damage, open or short circuits, and high-voltage component failures that cannot be automatically recovered and require manual inspection.

[0065] Specifically, after determining the final fault type, the fault is comprehensively evaluated from two dimensions: recoverability and the scope of its impact on vehicle driving safety. Fault recoverability refers to whether the fault can be automatically restored to normal operation, while the scope of its impact on vehicle driving safety refers to whether the fault affects the vehicle's basic driving functions or driving safety. The hybrid central control unit has a built-in fault level determination rule table. This rule table is pre-calibrated by the vehicle manufacturer before the vehicle leaves the factory based on historical data of various faults, component importance, and the degree of safety impact. It is stored in the non-volatile memory of the hybrid central control unit in the form of a structured mapping table. Each record contains three fields: fault type identifier, recoverability flag, and safety impact flag.

[0066] The specific process for fault level determination is as follows: After determining the final fault type, the hybrid central control unit first queries the fault level determination rule table using the fault type identifier as an index to obtain the recoverability flag and safety impact flag corresponding to the fault type; then, it determines the fault level according to the following determination rules: if the recoverability flag indicates that it can be automatically reset and recovered and the safety impact flag indicates that it does not affect driving safety, then the fault is determined to be a Level I transient fault; if the recoverability flag indicates that it can be recovered by restarting or reconfiguration and the safety impact flag indicates that it does not affect driving safety or has a minor impact, then the fault is determined to be a Level II soft fault; if the recoverability flag indicates that it cannot be automatically recovered or the safety impact flag indicates that it affects driving safety, then the fault is determined to be a Level III hard fault.

[0067] For new fault types not included in the fault level determination rule table, the hybrid central control unit defaults to classifying them as Level III hard faults to ensure driving safety with a conservative strategy. At the same time, the new fault type is included in the fault feature database for subsequent offline analysis and rule table expansion.

[0068] In some implementations, the fault level determination rule table can be updated via OTA (Over-The-Air) upgrade to accommodate new component types or correct misjudgment rules.

[0069] By using confidence scoring and a double confirmation mechanism, the misjudgment rate of complex and transient faults is effectively reduced, overcoming the shortcomings of existing technologies that have weak ability to identify multi-module linkage faults using a single rule base matching.

[0070] Preferably, a confidence threshold is preset. Set to 0.8, which is the confidence score output by the deep learning diagnostic model. When the value is below 0.8, the rule base will automatically trigger secondary verification.

[0071] S4: Obtain the feature vector of the fault from the deep learning diagnostic model, and determine the faulty component based on the mapping relationship between the feature vector and the component location; In this embodiment, the hybrid central control unit determines the specific component that has malfunctioned based on the mapping relationship between fault characteristics and component locations, and sends the fault component identifier and fault level to the human-machine prompt unit. The voice prompt module in the human-machine prompt unit broadcasts voice prompt information including the name of the fault component, the fault type, and the processing status. The indicator lights corresponding to the fault component in the fault location indicator panel are displayed according to the color and flashing mode corresponding to the fault level.

[0072] This step enables precise fault location and intuitive human-machine interface prompts. The hybrid central control unit utilizes the pre-stored fault characteristics and component location mapping relationship to map the final fault type to the specific faulty component and line node, generate a faulty component identifier, and send it along with the fault level to the human-machine interface prompt unit.

[0073] Furthermore, when processing abnormal signals to be diagnosed, the deep learning diagnostic model uses CNN convolutional layers to extract spatial features from multi-dimensional signals. The extracted feature vectors are output along with the fault type identification results and confidence scores, and are cached and stored by the hybrid central control unit. Step S4 directly calls the feature vectors extracted and cached in step S3, eliminating the need to re-extract features from the original signal. This ensures that the features used for fault localization are completely consistent with the features used for diagnosis, avoiding feature bias that may be introduced by secondary extraction.

[0074] The feature vector includes amplitude features, timing features, communication status features, and command response features. Amplitude features represent the magnitude and direction of the deviation of the component's operating status signal (current, voltage) from its rated operating parameters. Timing features represent the duration of signal anomalies and the time distribution of anomaly occurrences. Communication status features represent the component's communication frame loss rate, number of erroneous frames, and communication interruption duration. Command response features represent the deviation between the component's actual response signal and the expected response signal after receiving a working command. These four feature dimensions together constitute the feature vector for each fault type, serving as the input for querying the fault feature-component location mapping table. This uniquely identifies the fault type's manifestation at the signal level and locates the specific faulty component and line node.

[0075] The mapping relationship between fault characteristics and component locations is pre-set during the vehicle design phase. Based on the vehicle's electrical topology diagram, OEM electrical engineers systematically organize the installation locations, electrical circuits, and corresponding signal acquisition channels of all electrical components (including low-voltage electrical appliances, high-voltage accessories, and components of the hybrid collaborative system) to form a component location information table. During bench testing and vehicle calibration, for each known fault type, fault injection is manually simulated, and the feature vectors of each signal channel at the time of the fault are recorded. These feature vectors are then associated with the corresponding fault component identifier and circuit node number. This associated and labeled data is organized into a structured fault characteristic-component location mapping table. Each record contains four fields: fault type identifier, corresponding component identifier, circuit node number, and feature vector description. This structured mapping table is stored in the non-volatile memory of the hybrid central control unit. After mass production, this mapping table can be updated synchronously with changes to the vehicle's electrical topology via OTA (Over-The-Air) updates to adapt to new component types.

[0076] After determining the final fault type, the hybrid central control unit uses the fault type identifier as an index to perform a precise matching query in the fault feature and component location mapping table to obtain the component identifier and line node number corresponding to the fault type. If the query finds a unique record, the component identifier and line node number are directly used as the fault location result, generating a fault component identifier. If the query finds multiple records (i.e., the same fault type may correspond to multiple components), the hybrid central control unit further performs a secondary screening by combining the communication channel number of the abnormal signal reported by the fault acquisition unit, matching the abnormal signal channel with the signal acquisition channel of each candidate component to determine the unique fault component and line node. If the query does not find a match (i.e., there is no corresponding record in the mapping table), the hybrid central control unit marks the fault location result as an unknown component, and simultaneously writes the fault type identifier and corresponding feature vector into the fault feature database for subsequent offline analysis and mapping table expansion. After the query is completed, the hybrid central control unit sends the fault component identifier (including component name, subsystem to which it belongs, and line node number) along with the fault level to the human-machine prompt unit, driving the voice prompt module and fault location indicator panel to execute the corresponding prompt output.

[0077] Preferably, the fault feature and component location mapping table pre-stored before leaving the factory covers no less than 90% of the total number of core electrical components in the vehicle, so as to ensure that most common fault types can be accurately located; for new faults outside the coverage, the coverage of the mapping table is gradually expanded through continuous accumulation and offline analysis of the fault feature database.

[0078] After receiving the faulty component identification and processing status information, the voice prompt module broadcasts the fault information in concise and easy-to-understand Chinese. The broadcast content clearly includes the name of the faulty component, the fault type, and the current processing status. For example: "Left front low beam headlight drive overcurrent, automatically repaired," "DC / DC low voltage output abnormal, please pay attention," "High voltage PTC module fault, safely disconnected."

[0079] The voice prompt module is also connected to the in-vehicle noise sensor, automatically adjusting the voice broadcast volume based on the collected in-vehicle noise signal to ensure the driver can clearly receive fault prompts in different driving environments. While the fault remains unresolved, the voice prompt module will continuously broadcast fault prompts at preset time intervals until the fault is repaired or a manual shutdown command is received from the user.

[0080] A fault location indicator panel is installed on the center console of the cockpit, with an independent indicator light corresponding to each core electrical component. Level I transient faults are indicated by a flashing yellow light, Level II soft faults by a flashing orange light, and Level III hard faults by a solid red light. This visually indicates the fault location, allowing users to quickly locate the faulty component without requiring specialized knowledge. This differentiated design of color and flashing patterns enables drivers to intuitively distinguish the fault level and location without any technical expertise.

[0081] By using a collaborative human-machine prompting mechanism that combines voice broadcasting and indicator light positioning, the technical shortcomings of existing technologies, which only provide fault codes or indicator lights and are difficult for ordinary users to understand, are solved, thus achieving precise positioning and intuitive presentation of faulty components.

[0082] Preferably, the loop broadcast time interval of the voice prompt module is set to 30 seconds, that is, the fault prompt information is repeated every 30 seconds while the fault is not eliminated.

[0083] S5: Execute a graded automatic repair strategy based on the fault level; In this embodiment, the hybrid central control unit sends a corresponding repair instruction to the automatic repair execution unit according to the final fault level, and the automatic repair execution unit performs the repair operation corresponding to the fault level.

[0084] Based on the fault classification, a graded automatic repair strategy is implemented to achieve rapid recovery from self-healing faults. The specific repair strategy is as follows: When a fault is classified as a Level I transient fault, the hybrid central control unit sends a reset command to the automatic repair execution unit. The automatic repair execution unit then performs a reset operation on the drive module of the faulty component, clearing the fault code and restoring the component to normal operation. The entire process takes no more than 0.5 seconds and does not affect user operation. After the reset operation is completed, the voice prompt module announces that the repair is complete, the fault location indicator light goes out, and the system returns to normal detection status.

[0085] When the fault level is a Level II soft fault, the hybrid central control unit first checks the current vehicle speed before sending a gradual restart command. ;like Exceeding the preset safe speed threshold If the vehicle slows down, the restart operation will be temporarily suspended until the vehicle decelerates to a safe speed. Then the restart process is triggered to avoid the module being temporarily offline during the restart process, which could affect high-speed driving safety.

[0086] Once the vehicle speed requirement is met, the hybrid central control unit sends a progressive restart command to the automatic repair execution unit. The automatic repair execution unit first executes a restart command on the local submodule experiencing the anomaly (i.e., a sequence of power-down → 100ms wait → power-on for the local submodule), checking if communication has been restored. If communication is restored after the local submodule restarts, the communication connection is re-established and initialization configuration is completed, thus completing the repair. If communication is not restored after the local submodule restarts, a restart command is executed on the entire faulty module (i.e., a sequence of power-down → 500ms wait → power-on for the entire module). After restarting, communication is re-established and initialization configuration is completed. If the entire module still fails to recover after restarting, a switch to a backup module is initiated, and a trigger signal prompting the user for further maintenance is sent to the human-machine interface unit. This progressive restart strategy, by prioritizing the restart of local submodules, shortens the restart time and reduces the impact on other vehicle systems.

[0087] Preferred, .

[0088] When the fault level is a Level III hard fault, the hybrid central control unit first verifies whether the faulty component is a non-critical high-voltage accessory (i.e., a component that does not affect the basic driving safety of the vehicle, including the PTC heating module, OBC on-board charger, air conditioning compressor, etc.) before sending the cut-off command. Only the power supply circuit of the non-critical high-voltage accessory is cut off. For critical drive components such as drive motor controller and battery management system (BMS), the power supply is not cut off, but is handled by derated operation or functional degradation to prevent sudden loss of power during high-speed driving.

[0089] After confirming that the target to be disconnected is compliant, the hybrid central control unit sends a disconnection command to the power supply switching and protection unit. The power supply switching and protection unit disconnects the power supply circuit of the faulty non-critical component and switches to the backup power supply path, controlling the vehicle to enter limp mode. In limp mode, the power supply to the power steering system, braking system (including ABS), and basic lighting system is unaffected. The drive motor controller operates in derating mode, and the maximum driving speed is limited to 20km / h, ensuring that the vehicle can safely travel at low speed to the maintenance point.

[0090] The voice prompt module continuously announces that the vehicle has entered limp mode and requests that the vehicle be driven at low speed to the repair shop. The indicator light corresponding to the faulty component on the fault location indicator panel is displayed in solid red to remind the user to have it repaired in time.

[0091] The three-tiered automatic repair strategy covers common electrical faults in hybrid vehicles, significantly reducing manual intervention. Specifically, the progressive restart for Level II soft faults incorporates a safety-based speed assessment, while the disconnection operation for Level III hard faults is explicitly limited to non-critical high-voltage accessories. Through redundant module switching and limp-mode safety mechanisms (uninterrupted steering, braking, and lighting, with reduced drive performance), driving safety is ensured throughout the entire tiered repair process.

[0092] Preferably, the drive module reset operation for a Level I transient fault takes no more than 0.5 seconds. Preferably, the maximum vehicle speed in limp mode is limited to 20 km / h to ensure the vehicle can safely travel at low speed to the maintenance point.

[0093] In this embodiment, after the repair operation is completed, the hybrid central control unit collects the working status signal after the repair, verifies the repair result, and writes the fault type, fault level, repair process and repair result into the fault feature database to complete the fault closed-loop feedback.

[0094] After the repair operation is completed, the fault acquisition unit re-acquires the operating status signals of the relevant components and reports them to the hybrid central control unit. The hybrid central control unit compares the repaired operating status signals with the normal operating parameters to verify the effectiveness of the repair result. After verification, the hybrid central control unit writes the type, level, repair process, and repair result of this fault into the fault feature database in the form of structured data.

[0095] The fault feature database update and model optimization employ a combined mechanism of online incremental updates and offline batch fine-tuning. Online incremental updates involve writing new fault feature data into the fault feature database in real time after each fault repair, and synchronously updating the trigger frequency and accuracy statistics of corresponding rules in the rule base to dynamically adjust rule priorities. Offline batch fine-tuning involves triggering the offline fine-tuning process of the deep learning diagnostic model when the amount of new data in the fault feature database reaches a preset batch threshold (preferably 100 records). This process uses the new fault feature data to incrementally train the CNN-LSTM model, updates the model weight parameters, and then redeploys it to the hybrid central control unit, achieving full lifecycle management of fault data and continuous iterative improvement of diagnostic capabilities. For abnormal samples where the deep learning model and the rule base make inconsistent judgments in dual diagnosis, they are prioritized for inclusion in the offline fine-tuning dataset to specifically improve the model's recognition accuracy in boundary fault scenarios.

[0096] By combining online incremental updates with offline batch fine-tuning, a fault closed-loop feedback mechanism is established to continuously accumulate fault data, dynamically optimize the rule base, and iteratively improve the recognition accuracy of the deep learning diagnostic model, thereby reducing the fault recurrence rate and providing continuous data support for the design optimization of hybrid vehicle electrical systems.

[0097] The following description, in conjunction with specific scenarios, further illustrates the above method implementations.

[0098] The sensors of the fault acquisition unit are connected to the corresponding interfaces of the vehicle's low-voltage electrical appliances, high-voltage accessories, and hybrid collaborative system, and fixed in the engine compartment and driver's cabin in unoccupied locations. The hybrid central control unit is fixed under the driver's cabin and connected to other units via wires. The voice prompt module is connected to the vehicle audio system, and the fault location indicator panel is fixed in a prominent position on the center console of the driver's cabin. The power supply switching unit is connected to the high-voltage power battery, the 12V low-voltage battery, and various electrical circuits. After the vehicle is powered on, the system automatically completes initialization and enters real-time detection mode.

[0099] In one implementation, taking a Level I transient fault scenario as an example: During vehicle operation, when the user turns on the left front low beam headlight, the fault acquisition unit detects that the current in the left front low beam headlight drive circuit momentarily exceeds the corresponding threshold and transmits this abnormal signal to the hybrid central control unit. The hybrid central control unit retrieves the corresponding threshold group based on the current power mode, determines that the signal is an abnormal signal to be diagnosed, and then inputs it into a deep learning diagnostic model. The model outputs a fault type identification result of drive overcurrent, and the confidence score reaches or exceeds the preset confidence threshold. After double confirmation, the final fault level is determined to be a Level I transient fault. The hybrid central control unit determines that the faulty component is the left front low beam headlight drive module through the mapping relationship between fault characteristics and component location, sends the faulty component identification and fault level to the human-machine prompt unit, and the voice prompt module announces "Left front low beam headlight drive overcurrent, automatically repaired," and the indicator light corresponding to the left front low beam headlight in the fault location indicator panel flashes yellow. Subsequently, the hybrid central control unit sends a reset command to the automatic repair execution unit. The automatic repair execution unit then performs a reset operation on the left front low beam headlight drive module, clearing the fault code. The left front low beam headlight resumes normal operation, and vehicle driving is unaffected. After the repair is completed, the hybrid central control unit writes the fault data to the fault characteristic database, the indicator light turns off, and the system returns to normal detection status.

[0100] In another implementation, taking a Level II soft fault scenario as an example: When the vehicle is idling, the fault acquisition unit detects an interruption in the communication signal between the DC / DC converter module and the hybrid central control unit, and transmits this abnormal signal to the hybrid central control unit. The hybrid central control unit confirms the final fault level as a Level II soft fault through dual diagnostic logic, sends the faulty component identifier to the human-machine interface unit, and the voice prompt module announces "DC / DC communication abnormal, restarting," while the indicator light corresponding to the DC / DC converter module in the fault location indicator panel flashes orange. The hybrid central control unit sends a progressive restart command to the automatic repair execution unit. The automatic repair execution unit first executes a restart command on the local sub-module of the DC / DC conversion module that experienced the anomaly, and checks the communication status. If the communication status is restored after the local sub-module restarts, the communication connection is re-established and initialization configuration is completed. The voice prompt module announces "DC / DC communication has been restored," the indicator light goes out, and the system returns to normal detection status. If the communication status is not restored after the local sub-module restarts, a restart command is executed on the entire DC / DC conversion module. After restarting, the communication connection is re-established and initialization configuration is completed. If the entire module still fails to recover after restarting, the system switches to the backup module and sends a trigger signal to the human-machine interface unit to prompt the user for subsequent maintenance. After the repair is completed, the hybrid central control unit writes the fault data into the fault characteristic database.

[0101] In another implementation, taking a Level III hard fault scenario as an example: After starting the vehicle in winter, the user turns on the air conditioning for heating. The fault acquisition unit detects that the high-voltage PTC heating module has no current output and abnormal voltage, transmitting this signal to the hybrid central control unit. The hybrid central control unit confirms the final fault level as a Level III hard fault through dual diagnostic logic, sends the faulty component identifier to the human-machine interface unit, and the voice prompt module continuously broadcasts "High-voltage PTC module fault, safely disconnected, timely repair recommended." The indicator light corresponding to the PTC heating module in the fault location indicator panel displays a solid red light. The hybrid central control unit sends a disconnection command to the power supply switching and protection unit, which disconnects the power supply circuit of the high-voltage PTC heating module, switches to the backup power supply path, and controls the vehicle to enter limp mode, ensuring that the vehicle can safely drive at low speed to the repair point. The user can quickly locate the faulty component based on the voice prompts and indicator lights, drive the vehicle to the repair point, and have it replaced and repaired by maintenance personnel. After the repair is completed, the system automatically clears the fault code and resumes normal operation. This fault data is written to the fault feature database, completing the fault closed-loop feedback.

[0102] In some implementations, the hybrid central control unit has multiple sets of fault judgment thresholds built in, including a pure electric mode threshold set, a hybrid mode threshold set, and a kinetic energy recovery mode threshold set. When the hybrid central control unit detects a power mode switch, it synchronously updates the currently used fault judgment threshold set to ensure that the fault judgment thresholds in each power mode are consistent with the actual operating conditions and to eliminate the interference of normal signal fluctuations on fault judgment during mode switching.

[0103] In some implementations, the hybrid central control unit is also connected to a fault feature database. After each fault repair is completed, the fault type, fault level, repair process and repair result are written into the fault feature database, and the updated fault feature data is used to optimize the diagnostic parameters of the deep learning diagnostic model, so as to achieve continuous iterative improvement of diagnostic capabilities.

[0104] In summary, the core technical solution of this application is to address the unique characteristics of hybrid vehicles, which feature "engine + motor" dual-power coupling and multi-module linkage of electrical systems (including high-voltage battery, motor controller, engine ECU, on-board charger, air conditioning system, steering system, etc.), by constructing a fully closed-loop intelligent control system of "perception-diagnosis-localization-repair-feedback," as detailed below: (1) Core technical solution: First, through a distributed sensor network, the operating parameters (voltage, current, temperature, signal transmission rate, actuator action feedback, etc.) of each electrical module of the hybrid vehicle are collected in real time. Combined with the dynamic working conditions of the hybrid vehicle's power switching and mode conversion (pure electric / hybrid / kinetic energy recovery), a multi-dimensional fault feature database is established. Second, the dual diagnostic logic of "deep learning algorithm + rule base matching" is adopted to analyze the collected parameters in real time and quickly identify the fault type (minor abnormality, automatically repairable fault, non-automatically repairable fault). Then, through the mapping relationship between fault features and module location, the specific component, line and fault level of the fault are accurately located. Finally, for different fault types, the corresponding automatic repair strategy (parameter calibration adjustment, redundant module switching, temporary operating mode switching, fault isolation) is executed. After the repair is completed, the fault closed-loop status is fed back in real time, and the fault feature database is updated at the same time to improve the accuracy of subsequent diagnosis.

[0105] (2) The core differences from existing technologies: 1) Targeted adaptation to the characteristics of hybrid vehicles to overcome the limitations of "generalized diagnosis" in existing technologies: Existing automotive electrical fault diagnosis technologies are mostly designed for pure gasoline vehicles or pure electric vehicles, without considering the special characteristics of hybrid vehicles such as dual power switching and the coordinated operation of high and low voltage electrical appliances (e.g., electrical signal interference between motor and engine in hybrid mode, instantaneous load fluctuations of high voltage electrical appliances during kinetic energy recovery), which can easily lead to misdiagnosis and missed diagnosis; This application is specifically adapted to the dynamic operating conditions of hybrid vehicles, incorporating factors such as power mode switching and load fluctuations into the fault diagnosis logic to avoid diagnostic deviations caused by differences in operating conditions.

[0106] 2) Diagnosis-Location-Repair Closed Loop: Overcoming the shortcomings of existing technologies that "only diagnose, not repair" or "only perform one type of repair": Existing technologies can only achieve fault diagnosis and location, requiring manual intervention for repair, or can only repair minor faults such as simple poor circuit contact, and cannot deal with complex faults such as high-voltage electrical appliances and control modules in hybrid vehicles; This application constructs a closed-loop control that can not only accurately locate faults (accurate to specific components and circuit nodes), but also automatically execute diversified repair strategies according to the fault level, and complete the repair of most common faults without manual intervention, while realizing real-time feedback of repair results and database iteration.

[0107] 3) Dual diagnostic logic improves diagnostic accuracy and solves the problem of "high false positive rate" in existing technologies: Existing technologies mostly use a single rule base for matching diagnosis, which has a weak ability to identify complex faults (such as multi-module linkage faults and transient faults); This application combines deep learning algorithms (trained through massive fault data to identify complex fault features) with a rule base (adapted to standard fault modes of hybrid vehicle electrical systems), which not only ensures rapid diagnosis of common faults, but also improves the identification accuracy of complex faults and transient faults.

[0108] 4) Redundancy design enhances the operational safety of hybrid vehicles, unlike the existing technology's "direct shutdown after failure" mode: Existing technologies often use direct shutdown when critical faults such as high-voltage electrical appliances are detected, which affects driving safety (especially at high speeds); This application sets up redundant modules for critical electrical modules (such as high-voltage batteries and motor controllers), which can automatically switch to the redundant modules when a fault occurs, maintaining the vehicle's basic driving functions and avoiding safety hazards caused by sudden shutdown.

[0109] This application, through the above-mentioned technical solution, addresses the core pain points of fault diagnosis, location, and repair in hybrid vehicle electrical systems, achieving multi-dimensional, fully closed-loop intelligent control. Specific technical effects are as follows: (1) Improve the accuracy and efficiency of fault diagnosis and location, and reduce the rate of misjudgment and missed judgment: Adopting the dual diagnostic logic of "deep learning + rule base" and combined with the hybrid vehicle working condition adaptation design, it can realize the rapid identification of various faults (minor abnormalities, instantaneous faults, complex linkage faults, high voltage faults) of hybrid vehicle electrical system. The diagnostic response time is short and the fault location accuracy is high, which effectively solves the problems of misjudgment, missed judgment and slow response of existing technology diagnosis, and avoids component damage and fault expansion caused by failure to identify faults in time.

[0110] (2) Achieve automatic fault repair, reduce manual intervention and lower maintenance costs: For more than 80% of common electrical faults (such as poor circuit contact, parameter drift, minor module abnormalities, signal interference, etc.), the repair strategy can be executed automatically without manual disassembly, inspection and repair, which greatly shortens the fault repair time (from several hours of traditional manual repair to several seconds / minutes); at the same time, it reduces the cost of manual repair, reduces the difficulty and cost of later maintenance of hybrid vehicles, and improves the convenience of vehicle use.

[0111] (3) Ensure the driving safety of hybrid vehicles and avoid sudden shutdowns caused by malfunctions: Through repair strategies such as switching redundant modules and adjusting temporary operating modes, the basic driving functions of the vehicle can be maintained when key electrical modules fail (such as switching to hybrid mode when pure electric mode fails, and switching to low-voltage emergency mode when high-voltage electrical appliances fail), thus avoiding sudden vehicle shutdowns. This is especially suitable for scenarios such as high-speed driving and complex road conditions, thereby improving the driving safety and reliability of hybrid vehicles.

[0112] (4) Extend the service life of electrical components in hybrid vehicles and reduce the recurrence rate of faults: By monitoring the operating parameters of electrical modules in real time, minor abnormalities (such as parameter drift and temperature abnormalities) can be identified in advance, and automatic parameter calibration, load adjustment and other repair operations can be performed to prevent minor abnormalities from escalating into serious faults and extend the service life of electrical components (such as high-voltage batteries, motor controllers, ECUs, etc.); at the same time, through fault closed-loop feedback, the fault feature database is continuously updated to optimize diagnosis and repair strategies and reduce the recurrence rate of faults.

[0113] (5) Adapt to the multi-condition operation requirements of hybrid vehicles and improve vehicle operation stability: Specifically designed for the characteristics of hybrid vehicles switching between pure electric, hybrid, and kinetic energy recovery modes, the diagnostic and repair logic is optimized to avoid fault misjudgment caused by signal interference during condition switching, ensuring that the vehicle can operate stably under different conditions and improving the overall operation stability and driving experience of hybrid vehicles.

[0114] (6) Realize full lifecycle management of fault data to support subsequent vehicle optimization: The system automatically records fault type, location result, repair process, repair effect and other data to form a complete fault database. It can not only be used to optimize its own diagnostic and repair algorithm, but also provide data support for the design optimization and component selection of hybrid vehicle electrical system, and help improve the product competitiveness of hybrid vehicles.

[0115] Example 2 This embodiment discloses a hybrid vehicle electrical fault diagnosis, location, and repair system, including: The status signal acquisition module is configured to acquire the working status signals of various components of the hybrid vehicle; The anomaly detection module is configured to: obtain the current power mode of the hybrid vehicle, dynamically select the corresponding fault detection threshold based on the current power mode, compare the working status signal, and obtain the anomaly signal to be diagnosed. The fault diagnosis module is configured to: input the abnormal signal to be diagnosed into the deep learning diagnostic model, obtain the fault type identification result and its corresponding confidence score, determine the final fault type through dual diagnostic logic, and determine its fault level based on the final fault type. The fault location module is configured to: obtain the feature vector of the fault from the deep learning diagnostic model, and determine the faulty component based on the mapping relationship between the feature vector and the component location; The fault repair module is configured to execute a graded automatic repair strategy for the faulty component based on the fault level.

[0116] Example 3 This embodiment discloses a vehicle, including a hybrid vehicle electrical fault diagnosis, location, and repair system.

[0117] Example 4 The purpose of this embodiment is to provide a computing device, including 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 steps of the method of Embodiment 1.

[0118] Example 5 The purpose of this embodiment is to provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the steps of the method of Embodiment 1.

[0119] The steps and methods involved in the apparatuses of Embodiments 4 and 5 above correspond to those in Embodiment 1. For specific implementation details, please refer to the relevant description section of Embodiment 1. The term "computer-readable storage medium" should be understood as a single medium or multiple media including one or more instruction sets; it should also be understood as including any medium capable of storing, encoding, or carrying an instruction set for execution by a processor and enabling the processor to perform any of the methods in this invention.

[0120] Those skilled in the art will understand that the modules or steps of the present invention described above can be implemented using general-purpose computer devices. Optionally, they can be implemented using computer-executable program code, thereby allowing them to be stored in a storage device for execution by a computer device, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. The present invention is not limited to any particular combination of hardware and software.

[0121] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

[0122] While the specific embodiments of the present invention have been described above in conjunction with the accompanying drawings, this is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art without creative effort based on the technical solutions of the present invention are still within the scope of protection of the present invention.

Claims

1. A method for diagnosing, locating, and repairing electrical faults in hybrid vehicles, characterized in that, include: Acquire the operating status signals of various components in hybrid vehicles; The current power mode of the hybrid vehicle is obtained, and the corresponding fault judgment threshold is dynamically selected based on the current power mode. The working status signal is compared to obtain the abnormal signal to be diagnosed. The abnormal signal to be diagnosed is input into the deep learning diagnostic model to obtain the fault type identification result and its corresponding confidence score. The final fault type is determined through dual diagnostic logic, and its fault level is determined based on the final fault type. The feature vector of the fault is obtained from the deep learning diagnostic model, and the faulty component is determined based on the mapping relationship between the feature vector and the component location. Based on the fault level, a graded automatic repair strategy is implemented for the faulty components.

2. The method for diagnosing, locating, and repairing electrical faults in hybrid vehicles as described in claim 1, characterized in that, The various components include low-voltage electrical appliances, high-voltage accessories, and hybrid collaborative systems.

3. The method for diagnosing, locating, and repairing electrical faults in hybrid vehicles as described in claim 1, characterized in that, Based on the current power mode, select the corresponding fault judgment threshold from the preset pure electric mode threshold group, hybrid mode threshold group, and kinetic energy recovery mode threshold group.

4. The method for diagnosing, locating, and repairing electrical faults in hybrid vehicles as described in claim 3, characterized in that, After detecting a power mode switching signal, a threshold switching stabilization time window is set. Within the threshold switching stabilization time window, the original threshold group remains valid. After the threshold switching stabilization time window ends, it is synchronously updated to the threshold group corresponding to the new power mode.

5. The method for diagnosing, locating, and repairing electrical faults in hybrid vehicles as described in claim 1, characterized in that, The dual diagnostic logic specifically includes: When the confidence score output by the deep learning diagnostic model is greater than or equal to the confidence threshold, the fault type identification result of the deep learning diagnostic model is used as the final fault type. When the confidence score output by the deep learning diagnostic model is less than the confidence threshold, the rule base is triggered to perform secondary verification on the same abnormal signal to be diagnosed. If the verification result of the rule base is consistent with the fault type identification result of the deep learning diagnostic model, the fault type identification result of the deep learning diagnostic model is taken as the final fault type. If they are inconsistent, the verification result of the rule base is taken as the final fault type.

6. The method for diagnosing, locating, and repairing electrical faults in hybrid vehicles as described in claim 1, characterized in that, Based on the final fault type, the fault level is determined according to the fault level determination rule table; the fault level includes transient faults, soft faults, and hard faults.

7. The method for diagnosing, locating, and repairing electrical faults in hybrid vehicles as described in claim 1, characterized in that, The feature vector is obtained by the deep learning diagnostic model from the spatial feature extraction of the abnormal signal to be diagnosed, including amplitude features, time series features, communication status features, and command response features.

8. The method for diagnosing, locating, and repairing electrical faults in hybrid vehicles as described in claim 1, characterized in that, The mapping relationship between the feature vectors and component positions is pre-set during the vehicle design phase and stored in the form of a structured mapping table.

9. A method for diagnosing, locating, and repairing electrical faults in hybrid vehicles as described in claim 1, characterized in that, The system will also provide voice announcements of the faulty component name, fault type, and current processing status.

10. A method for diagnosing, locating, and repairing electrical faults in a hybrid vehicle as described in claim 9, characterized in that, The voice broadcast volume is automatically adjusted based on in-vehicle noise signals.

11. The method for diagnosing, locating, and repairing electrical faults in hybrid vehicles as described in claim 1, characterized in that, Furthermore, indicator lights corresponding to the faulty components are displayed separately according to the fault level.

12. The method for diagnosing, locating, and repairing electrical faults in hybrid vehicles as described in claim 1, characterized in that, When the fault level is transient, a reset operation is performed on the drive module of the faulty component to clear the fault code and restore the component to normal operation.

13. The method for diagnosing, locating, and repairing electrical faults in hybrid vehicles as described in claim 1, characterized in that, When the fault level is a soft fault, the current vehicle speed is checked first. If the vehicle speed exceeds the safe speed threshold, the restart operation is temporarily suspended and the vehicle speed is slowed down to less than or equal to the safe speed threshold before a gradual restart is performed.

14. The method for diagnosing, locating, and repairing electrical faults in hybrid vehicles as described in claim 1, characterized in that, When the fault level is a hard fault, first verify whether the faulty component belongs to a non-critical high-voltage accessory. Perform a power supply disconnection operation on the power supply circuit of the non-critical high-voltage accessory. For critical drive components, do not perform a power supply disconnection, but handle them by derated operation or functional degradation.

15. A system for diagnosing, locating, and repairing electrical faults in hybrid vehicles, characterized in that, include: The status signal acquisition module is configured to acquire the working status signals of various components of the hybrid vehicle; The anomaly detection module is configured to: obtain the current power mode of the hybrid vehicle, dynamically select the corresponding fault detection threshold based on the current power mode, compare the working status signal, and obtain the anomaly signal to be diagnosed. The fault diagnosis module is configured to: input the abnormal signal to be diagnosed into the deep learning diagnostic model, obtain the fault type identification result and its corresponding confidence score, determine the final fault type through dual diagnostic logic, and determine its fault level based on the final fault type. The fault location module is configured to: obtain the feature vector of the fault from the deep learning diagnostic model, and determine the faulty component based on the mapping relationship between the feature vector and the component location; The fault repair module is configured to execute a graded automatic repair strategy for the faulty component based on the fault level.

16. A vehicle, characterized in that, This includes a hybrid vehicle electrical fault diagnosis, location, and repair system as described in claim 15.

17. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps in the hybrid vehicle electrical fault diagnosis, location and repair method as described in any one of claims 1-14.

18. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps in the hybrid vehicle electrical fault diagnosis, location and repair method as described in any one of claims 1-14.

Images