Vehicle cycle fault code standardization method and system for electric vehicle

Abstract

The invention provides a method and a system for standardizing a cycle fault code of an electric vehicle. The method comprises the following steps: defining a standardized fault code; collecting user fault record data and enterprise after-sales data, then carrying out data labeling, and generating an entity identification data set and an intention classification data set; generating a knowledge graph database; loading a pre-trained open source BERT basic model by using a natural language model development tool Transformers library, and performing model training on the entity recognition data set and the intention classification data set to generate a trained fault entity recognition model and a trained intention classification recognition model; acquiring voice information input by a user; constructing knowledge graph query conditions; querying from a knowledge graph database according to the knowledge graph query condition to obtain fault detail node information; and according to a preset template, obtaining corresponding attribute field information from the fault detail node information, and generating a fault knowledge question and answer result.

Description

Technical Field

[0001] This invention relates to the field of electric vehicle technology, and in particular to a method and system for standardizing fault codes during the vehicle's service life. Background Technology

[0002] Currently, the vehicle-side system software / hardware fault codes for two-wheeled and three-wheeled vehicles are relatively well-developed and displayed on the vehicle-side screen. However, most fault codes are technical in nature and have limited readability, making them difficult for users, maintenance personnel, and after-sales staff to understand. Furthermore, there is no unified online system that describes the causes of the fault codes and recommended actions.

[0003] Currently, the definition and management of fault codes in the software / hardware systems of electric two-wheelers and three-wheelers throughout the entire lifecycle of vehicle networking services are fragmented. Each business system focuses on its own business scope, and there is no unified construction of fault codes for the entire vehicle networking business software / hardware system based on the perspective of the user's entire vehicle use cycle.

[0004] Currently, fault codes are generally only found in the user manual. When a fault occurs in a vehicle usage scenario, it is difficult for users and after-sales personnel to quickly assess the damage and impact of the fault, and they are also unable to take targeted measures and maintenance operations. Furthermore, the detailed fault records throughout the user's entire vehicle network usage cycle are not collected and analyzed, making it impossible to provide early warnings and notifications to other users with similar potential faults. Summary of the Invention

[0005] In view of this, the purpose of this invention is to provide a standardized method and system for fault codes during the vehicle's service life. It employs a three-stage design to standardize fault codes, enabling rapid identification of the specific business system to which the fault belongs and the specific fault itself. A knowledge graph model is used for standardized management of the fault code knowledge base, providing users and enterprise after-sales teams with a more efficient knowledge base management system. Furthermore, it provides a natural language interaction-based question-and-answer knowledge base service, combining technologies such as speech recognition, knowledge graph text generalization, question-and-answer knowledge base querying, and speech synthesis to offer users a convenient interaction method, thereby improving user experience.

[0006] In a first aspect, embodiments of the present invention provide a method for standardizing fault codes during the vehicle usage cycle of electric vehicles, the method comprising: Define standardized fault codes, and update the standardized fault code data to the database when the standardized fault codes meet the requirements; After collecting user fault record data and enterprise after-sales data, the data is labeled to generate entity recognition dataset and intent classification dataset. Using the Neo4j graph database tool, a knowledge graph database is generated according to the data attributes, fault entities, and relation types of the standardized fault codes. The Transformers library, a natural language model development tool, is used to load a pre-trained open-source BERT base model to train the model on the entity recognition dataset and the intent classification dataset, generating a trained fault entity recognition model and a trained intent classification recognition model. Obtain user-inputted voice information; Based on the voice information, the trained fault entity recognition model, and the trained intent classification recognition model, construct knowledge graph query conditions; Based on the knowledge graph query conditions, a query is performed from the knowledge graph database to obtain fault detail node information; According to the preset template, the corresponding attribute field information is obtained from the fault details node information, and the fault knowledge Q&A results are generated.

[0007] Furthermore, standardized fault codes are defined, including: Obtain fault codes from the software / hardware systems of the Internet of Vehicles (IoV) for electric two-wheelers and three-wheelers throughout their entire vehicle lifecycle. The fault codes are defined according to a three-segment standard, with each segment marked using English punctuation. The first segment uses a two-letter uppercase English letter string to identify the specific source of the fault code; the second segment uses the same two-letter uppercase English letter string to identify the system to which the fault code belongs; and the third segment uses a four-digit number string to identify the specific fault corresponding to the fault code.

[0008] Furthermore, after collecting user fault record data and enterprise after-sales data, data annotation is performed to generate entity recognition datasets and intent classification datasets, including: The fault entity features are extracted from the user fault record data and labeled to obtain the entity recognition dataset; The intent classification of customer service and user question dialogue texts is extracted from the enterprise's after-sales data and labeled to obtain the intent classification dataset.

[0009] Furthermore, based on the voice information, the trained fault entity recognition model, and the trained intent classification recognition model, knowledge graph query conditions are constructed, including: The speech information is converted into text corpus using a speech recognition model; The trained fault entity recognition model and the trained intent classification recognition model are invoked to recognize the text corpus and obtain the recognition result; The knowledge graph query conditions are constructed based on the recognition results.

[0010] Furthermore, the trained fault entity recognition model is used to provide search targets for the user question and answer, and the trained intent classification recognition model is used to provide search directions for the user question and answer.

[0011] Furthermore, the knowledge graph database is a database for online fault entity querying based on detailed fault code data structure attribute information.

[0012] Furthermore, the method also includes: The results of the fault knowledge Q&A will be fed back to the user.

[0013] Secondly, embodiments of the present invention provide a standardized fault code system for the vehicle usage cycle of electric vehicles, the system comprising: The definition module is used to define standardized fault codes. When the standardized fault codes meet the requirements, the data of the standardized fault codes is updated to the database. The data acquisition module is used to collect user fault record data and enterprise after-sales data, and then perform data annotation to generate entity recognition datasets and intent classification datasets. The first generation module is used to generate a knowledge graph database using the Neo4j graph database tool, according to the data attributes, fault entities, and relation types of the standardized fault codes. The second generation module is used to load a pre-trained open-source BERT base model using the Transformers library, a natural language model development tool, to train the model on the entity recognition dataset and the intent classification dataset, and generate a trained fault entity recognition model and a trained intent classification recognition model. The first acquisition module is used to acquire the voice information input by the user; The construction module is used to construct knowledge graph query conditions based on the voice information, the trained fault entity recognition model, and the trained intent classification recognition model. The query module is used to query the knowledge graph database according to the knowledge graph query conditions to obtain fault detail node information; The second acquisition module is used to obtain the corresponding attribute field information from the fault details node information according to the preset template, and generate fault knowledge Q&A results.

[0014] Thirdly, embodiments of the present invention provide an electronic device, including a memory and a processor, wherein the memory stores a computer program that can run on the processor, and the processor executes the computer program to implement the method described above.

[0015] Fourthly, embodiments of the present invention provide a computer-readable medium having processor-executable non-volatile program code that causes the processor to perform the method described above.

[0016] This invention provides a method and system for standardizing fault codes throughout the usage cycle of electric vehicles, including: defining standardized fault codes; updating the data of standardized fault codes to a database when they meet the requirements; collecting user fault record data and enterprise after-sales data and then annotating the data to generate an entity recognition dataset and an intent classification dataset; generating a knowledge graph database using the Neo4j graph database tool according to the data attributes, fault entities, and relation types of the standardized fault codes; loading a pre-trained open-source BERT basic model using the Transformers library, a natural language model development tool, to train the model on the entity recognition dataset and the intent classification dataset, generating a trained fault entity recognition model and a trained intent classification recognition model; acquiring user-input voice information; and updating the data based on the voice information... The trained fault entity recognition model and the trained intent classification recognition model are used to construct knowledge graph query conditions. Based on these conditions, queries are performed from the knowledge graph database to obtain fault detail node information. Following a preset template, corresponding attribute fields are retrieved from the fault detail node information, and fault knowledge question-and-answer results are generated. A three-stage design is adopted to standardize fault codes, enabling rapid location of the specific business system to which the fault belongs and the specific fault itself. A knowledge graph model is used for standardized management of the fault code knowledge base, providing users and enterprise after-sales teams with a more efficient knowledge base management system. Externally, a natural language interaction-based question-and-answer knowledge base service is provided, combining technologies such as speech recognition, knowledge graph text generalization, question-and-answer knowledge base query, and speech synthesis to provide users with a convenient interaction method, thereby improving user experience.

[0017] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention are realized and obtained in accordance with the structures particularly pointed out in the description, claims and drawings.

[0018] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0019] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0020] Figure 1 This is a flowchart of the method for standardizing fault codes during the vehicle usage cycle of an electric vehicle, provided in Embodiment 1 of the present invention. Figure 2 This is a schematic diagram illustrating the three-segment fault code composition format and example provided in Embodiment 1 of the present invention; Figure 3 This is a schematic diagram of the basic process for standardizing fault codes provided in Embodiment 1 of the present invention; Figure 4 This is a schematic diagram of the interaction logic of the standard fault code knowledge base system provided in Embodiment 1 of the present invention; Figure 5 This is a schematic diagram of the overall system module architecture of the unified and standardized fault code platform provided in Embodiment 1 of the present invention; Figure 6 This is a schematic diagram of the electric vehicle usage cycle fault code standardization system provided in Embodiment 2 of the present invention. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] The following explains the technical terms used in this application: Neo4j Graph Database: A high-performance non-relational graph database. Neo4j is an open-source database system based on a native graph storage architecture. Its core data model consists of a three-layer structure of nodes, relationships, and properties. It is designed to handle large-scale "entity-relationship" data and can express and store complex relationships between different entities in a near-natural way. It supports efficient graph traversal and relationship mining operations. Transformers Library: The Transformers Library is an open-source core tool library for natural language processing (NLP) and multimodal tasks. Its core value lies in the standardized encapsulation of mainstream Transformer architecture pre-trained models, providing a unified and efficient development interface. It aims to lower the development and application threshold of Transformer-like models, support the entire process of technical implementation from model loading, data preprocessing, inference computation to fine-tuning and optimization, and is widely compatible with various deep learning frameworks. It is the mainstream tool for the engineering implementation of Transformer models in the current field of artificial intelligence. The BERT (Bidirectional Encoder Representations from Transformers) model is a bidirectional pre-trained language model based on the Transformer encoder architecture. Its core feature is the bidirectional self-attention mechanism, which achieves deep semantic understanding of natural language text through a two-stage paradigm of "pre-training-fine-tuning".

[0023] With the implementation of the new national standard GB 17761-2024 for electric bicycles, the standard stipulates that vehicles should be equipped with 4G or 5G public network communication modules, and electric bicycles should be able to send dynamic safety monitoring information to information management platforms built by enterprises. On this basis, vehicles can report fault information during use to the enterprise's cloud platform through the network channel, so that the enterprise's cloud data platform can instantly perceive the real-time fault status of the vehicle. With the establishment of enterprise cloud-based vehicle networking data platforms, electric two-wheeler companies can standardize fault codes for the entire vehicle lifecycle by combining fault sources, software / hardware system categories, and specific faults in the process of using mobile apps, third-party wearable devices, vehicle systems, and cloud-based vehicle networking business systems. They can also uniformly define, collect, analyze, and manage fault codes. With the development of AI technology, electric two-wheeler companies can uniformly manage and analyze data collected from multiple terminals, throughout the entire vehicle lifecycle, and standardized software / hardware fault codes. They can also use standardized knowledge graph models to generalize and associate various user query text corpora with knowledge base query conditions, providing users with accurate after-sales support services. With the development of voice technology, electric two-wheeler manufacturers can combine voice recognition and synthesis technologies to provide users with intelligent voice knowledge Q&A services through natural language interaction on interactive channels such as mobile apps, vehicle systems, and mini-programs. This will provide users with convenient fault analysis and basic maintenance guidance services, thereby improving the user experience. With the establishment of a unified, standardized fault code knowledge base system for the entire lifecycle of electric vehicles, the customer service and maintenance systems of electric two-wheeler companies, based on this unified and standardized fault knowledge base management system, can quickly support user feedback and fault location, improve enterprise work efficiency, and provide users with a better vehicle experience.

[0024] This application is based on a three-segment unified standard fault code design, which abstracts the fault codes in the entire life cycle of vehicle networking services. Each segment is separated by the punctuation mark ".", resulting in a clear structure and distinct segments. Each segment belongs to the previous segment, and the relationship between segments is progressive. By combining the segments, the specific business system to which the fault belongs and the specific fault can be quickly located, which facilitates the rapid diversion and fault location by the enterprise's after-sales, maintenance, R&D, customer service and other departments. The three-segment unified standard fault code adopted in this application specifies that the first two segments are identified by uppercase letters, and the third segment is identified by a four-digit string. The uppercase letters in the first two segments are composed of common English word abbreviations, which are easy to understand. After the first two segments are determined, the four-digit string in the third segment is sufficient to identify 9,999 specific faults, which is easy to expand in the future. This application uses a three-segment identification system to cover almost all software and hardware system business scenarios of electric two-wheeled and three-wheeled vehicles throughout the entire life cycle of vehicle networking interaction, including mobile phone terminals, vehicle terminals, enterprise cloud platforms, and third-party access terminals. It collects fault codes generated at each stage through network channels and manages them uniformly to the enterprise data platform, greatly reducing the cost of problem analysis and localization. The fault code acquisition solution in this application includes a unified data format for users, devices, system time, and fault codes, which can completely restore the on-site environment when the fault occurred, better provide data support for users and enterprise after-sales teams, and facilitate fault location and analysis. This application standardizes the management of the cause, solution, recommended handling method, and corresponding early warning measures for each fault code. It uses a knowledge graph model to standardize the management of the fault code knowledge base, providing users and enterprise after-sales teams with a more efficient knowledge base management system. Based on the construction of a unified and standardized fault code knowledge base data system, this system provides external users with a question-and-answer knowledge base service based on natural language interaction, combining technologies such as speech recognition, knowledge graph text generalization, question-and-answer knowledge base query, and speech synthesis to provide users with a convenient interaction method and improve user experience.

[0025] To facilitate understanding of this embodiment, the embodiments of the present invention will be described in detail below.

[0026] Example 1: Figure 1The flowchart is a method for standardizing fault codes during the vehicle usage cycle of an electric vehicle, as provided in Embodiment 1 of the present invention.

[0027] Reference Figure 1 The method includes the following steps: Step S101: Define standardized fault codes. When the standardized fault codes meet the requirements, update the data of the standardized fault codes to the database. Step S102: After collecting user fault record data and enterprise after-sales data, perform data annotation to generate entity recognition dataset and intent classification dataset; Step S103: Using the Neo4j graph database tool, generate a knowledge graph database according to the data attributes, fault entities, and relation types of standardized fault codes; Step S104: Load the pre-trained open-source BERT base model using the Transformers library, a natural language model development tool, and train the model on the entity recognition dataset and the intent classification dataset to generate the trained fault entity recognition model and the trained intent classification recognition model. Step S105: Obtain the voice information input by the user; Step S106: Construct knowledge graph query conditions based on voice information, the trained fault entity recognition model, and the trained intent classification recognition model; Step S107: Query the knowledge graph database according to the knowledge graph query conditions to obtain fault detail node information; Step S108: According to the preset template, obtain the corresponding attribute field information from the fault details node information and generate fault knowledge Q&A results.

[0028] Furthermore, step S101 includes the following steps: Step S201: Obtain the fault codes of the vehicle network software / hardware system throughout the entire life cycle of electric two-wheeled and three-wheeled vehicles; Step S202: Define the fault code according to the three-segment standard, and mark each segment with English punctuation. The first segment uses a two-letter uppercase English letter string to identify the specific source of the fault code; the second segment uses a two-letter uppercase English letter string to identify the system to which the fault code belongs; and the third segment uses a four-digit numeric string to identify the specific fault corresponding to the fault code.

[0029] This application applies to the usage scenarios of electric two-wheelers and electric tricycles. This application focuses on defining a standardized method for full-service fault codes in the Internet of Vehicles (IoV) system; and the construction scheme for a standardized system for collecting, storing, and querying fault codes generated during vehicle use.

[0030] This application defines fault codes for the entire lifecycle of electric two-wheeled and three-wheeled vehicle networking software / hardware systems according to a three-segment standard. Each segment is marked with an English punctuation mark ".". The first segment is identified by a two-letter uppercase English letter string, indicating the specific source of the fault code (divided according to physical units, such as AP for mobile APP, CL for enterprise cloud platform, VC for vehicle software system, IL for vehicle instrument panel, etc.). The second segment is identified by a two-letter uppercase English letter string, indicating the system to which the fault code belongs (based on the specific system under the source of the first segment; the identified system belongs to the source of the first segment, for example, AP.NW indicates the network system under the mobile APP, and LC.OT indicates the OTA remote upgrade system under the cloud platform). The third segment is identified by a four-digit string of numbers 0-9, indicating the specific fault corresponding to the fault code (the specific fault belongs to the second segment of business system under the source of the first segment; the systems are isolated from each other).

[0031] For example, both AP.NW and LC.OT can have fault code 0001. For instance, AP.NW.0001 fault code indicates fault number 0001 in the network connection system under the mobile APP, while LC.OT.0001 fault code indicates fault number 0001 in the OTA remote upgrade system under the cloud platform. The format and examples of the three-part fault code are as follows: Figure 2 As shown.

[0032] Standard fault codes are uniformly applied for and managed on the standardization platform of electric two-wheeled and three-wheeled vehicle enterprises. The enterprise standardization management team is responsible for maintenance and management to ensure global consensus and avoid cross-referencing or duplication of codes. The first and second parts of the fault code are applied for by the business team to the standardization team. After the application, the standardization team conducts unified scheduling and approval. The third part is formulated by the business system internally based on the analysis and discussion of the fault details, and then submitted to the standardization system for filing and storage by the system manager.

[0033] The fault code standardization system, as a static management system, is responsible for the unified management of adding, modifying, and discarding standard fault codes, and maintains the detailed information of standard fault codes in the database table. One fault code corresponds to one row of data in the database table. The specific fault code data structure is shown in Table 1. Table 1

[0034] After the underlying data of the fault code standardization system is built, a text-based fault knowledge base is established for data such as fault codes, root causes of faults, operation suggestions, and maintenance guidance. This base is then used to enrich the accuracy and response efficiency of the knowledge graph query by generalizing the text corpus. The basic process of fault code standardization is as follows: Figure 3 As shown.

[0035] The question-and-answer knowledge base is applied to after-sales scenarios for electric two-wheelers and electric tricycles. 2. Construction scheme for a standardized system for collecting, storing, and querying fault codes generated during vehicle use; The question-and-answer knowledge base system of this application relies on the standardized definition of fault codes. After the standardization of fault codes is completed, the data records of the standardized fault codes are stored in a database.

[0036] Furthermore, step S102 includes the following steps: Step S301: Extract fault entity features from user fault record data and label them to obtain entity recognition dataset; Step S302: Extract the intent classification of customer service and user question dialogue text from the enterprise's after-sales data, and label it to obtain the intent classification dataset.

[0037] Specifically, fault record data is collected throughout the user's vehicle usage process according to the standardized definition of fault codes, and the fault data is uploaded to the cloud data platform. The cloud data platform records the fault details of all users, and the detailed fault data format is shown in Table 2: Table 2

[0038] Collect the dialogue text data from the 400 customer service system, and collect the text content of fault descriptions and repair operations from maintenance / repair staff in the maintenance management system. Store and input the data into the system according to the formatted data. Text annotation tools were used to annotate the text content of user fault record data and enterprise after-sales data collection. The annotations were divided into two categories: the first category was fault entity feature labeling (including fault codes, fault symptoms, causative components, recommended operations, fault levels, root causes, etc.); the second category was customer service and user question dialogue text intent classification labeling (including intent classifications such as query definition, asking for reasons, asking for solutions, asking for maintenance, etc.). Entity recognition datasets and intent classification datasets were generated. Based on the standardized fault code definition data, Neo4j graph database tools were used to generate a knowledge graph database that allows online fault entity queries based on detailed fault code data structure attribute information, according to the standardized fault code data attributes, fault entities, and their relationship types.

[0039] Furthermore, step S106 includes the following steps: Step S401: Convert speech information into text corpus using a speech recognition model; Step S402: Call the trained fault entity recognition model and the trained intent classification recognition model to recognize the text corpus and obtain the recognition result; Step S403: Construct knowledge graph query conditions based on the recognition results.

[0040] Furthermore, the trained fault entity recognition model is used to provide search targets for user question answering, and the trained intent classification recognition model is used to provide search directions for user question answering.

[0041] Furthermore, the knowledge graph database is a database for online fault entity querying based on the detailed data structure attribute information of fault codes.

[0042] Furthermore, the method also includes the following steps: Step S501: Feedback the results of the fault knowledge Q&A to the user.

[0043] Specifically, the Transformers library, a natural language model development tool, is used to load a pre-trained open-source BERT base model. Cleaned and labeled text data is then used to train the model, generating a trained intent classification and recognition model (providing search direction for user questions) and a fault entity recognition model (providing search targets for user questions). For example, if a user inputs "What should I do if the car's infotainment system screen turns yellow while I'm riding?", the intent classification model determines whether to search for "cause" or "measures," while the fault entity recognition model locates the specific fault within the intent range.

[0044] The system integrates speech recognition model capabilities. When a user inputs voice in the voice interaction mode of the electric vehicle company's APP system or vehicle-side system, the system calls the speech recognition model to convert the speech into text. Using the text corpus after speech recognition, it calls the intent recognition model and the fault entity recognition model for model recognition. Combining the recognition results of both, it constructs knowledge graph query conditions (including query intent and query entity conditions), retrieves fault detail node information from the knowledge graph database, extracts corresponding attribute field information from the nodes according to a preset template to generate an answer, and provides feedback to the user. The interaction logic of the standard fault code knowledge base system is as follows: Figure 4 and Figure 5 As shown.

[0045] Example 2: Figure 6 This is a schematic diagram of the electric vehicle usage cycle fault code standardization system provided in Embodiment 2 of the present invention.

[0046] Reference Figure 6 The system includes: The definition module is used to define standardized fault codes. When the standardized fault codes meet the requirements, the data of the standardized fault codes is updated to the database. The data acquisition module is used to collect user fault record data and enterprise after-sales data, and then perform data annotation to generate entity recognition datasets and intent classification datasets. The first generation module is used to generate a knowledge graph database using the Neo4j graph database tool, based on the data attributes, fault entities, and relation types of standardized fault codes. The second generation module is used to load the pre-trained open-source BERT base model using the Transformers library, a natural language model development tool, to train the model on the entity recognition dataset and the intent classification dataset, and generate the trained fault entity recognition model and the trained intent classification recognition model. The first acquisition module is used to acquire the voice information input by the user; The module is used to construct knowledge graph query conditions based on voice information, the trained fault entity recognition model, and the trained intent classification recognition model. The query module is used to query the knowledge graph database based on the query conditions of the knowledge graph to obtain fault detail node information; The second acquisition module is used to retrieve the corresponding attribute field information from the fault details node information according to the preset template, and generate fault knowledge Q&A results.

[0047] This invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the electric vehicle usage cycle fault code standardization method provided in the above embodiments.

[0048] This invention also provides a computer-readable medium having processor-executable non-volatile program code, on which a computer program is stored. When the computer program is run by a processor, it executes the steps of the electric vehicle usage cycle fault code standardization method of the above embodiments.

[0049] The computer program product provided in this embodiment of the invention includes a computer-readable storage medium storing program code. The instructions included in the program code can be used to execute the methods described in the preceding method embodiments. For specific implementation details, please refer to the method embodiments, which will not be repeated here.

[0050] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the system and apparatus described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0051] Furthermore, in the description of the embodiments of the present invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in the present invention based on the specific circumstances.

[0052] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, essentially, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0053] In the description of this invention, it should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing the invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0054] Finally, it should be noted that the above-described embodiments are merely specific implementations of the present invention, used to illustrate the technical solutions of the present invention, and not to limit it. The scope of protection of the present invention is not limited thereto. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can still modify or easily conceive of changes to the technical solutions described in the foregoing embodiments within the technical scope disclosed in the present invention, or make equivalent substitutions for some of the technical features; and these modifications, changes, or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for standardizing fault codes during the usage cycle of an electric vehicle, characterized in that, The method includes: Define standardized fault codes, and update the standardized fault code data to the database when the standardized fault codes meet the requirements; After collecting user fault record data and enterprise after-sales data, the data is labeled to generate entity recognition dataset and intent classification dataset. Using the Neo4j graph database tool, a knowledge graph database is generated according to the data attributes, fault entities, and relation types of the standardized fault codes. The Transformers library, a natural language model development tool, is used to load a pre-trained open-source BERT base model to train the model on the entity recognition dataset and the intent classification dataset, generating a trained fault entity recognition model and a trained intent classification recognition model. Obtain user-inputted voice information; Based on the voice information, the trained fault entity recognition model, and the trained intent classification recognition model, construct knowledge graph query conditions; Based on the knowledge graph query conditions, a query is performed from the knowledge graph database to obtain fault detail node information; According to the preset template, the corresponding attribute field information is obtained from the fault details node information, and the fault knowledge Q&A results are generated.

2. The method for standardizing fault codes during the vehicle's service life as described in claim 1, characterized in that, Define standardized fault codes, including: Obtain fault codes from the software / hardware systems of the Internet of Vehicles (IoV) for electric two-wheelers and three-wheelers throughout their entire vehicle lifecycle. The fault codes are defined according to a three-segment standard, with each segment marked using English punctuation. The first segment uses a two-letter uppercase English letter string to identify the specific source of the fault code; the second segment uses the same two-letter uppercase English letter string to identify the system to which the fault code belongs; and the third segment uses a four-digit number string to identify the specific fault corresponding to the fault code.

3. The method for standardizing fault codes during the vehicle's service life according to claim 1, characterized in that, After collecting user fault record data and enterprise after-sales data, the data is labeled to generate entity recognition datasets and intent classification datasets, including: The fault entity features are extracted from the user fault record data and labeled to obtain the entity recognition dataset; The intent classification of customer service and user question dialogue texts is extracted from the enterprise's after-sales data and labeled to obtain the intent classification dataset.

4. The method for standardizing fault codes during the vehicle's service life according to claim 1, characterized in that, Based on the voice information, the trained fault entity recognition model, and the trained intent classification recognition model, knowledge graph query conditions are constructed, including: The speech information is converted into text corpus using a speech recognition model; The trained fault entity recognition model and the trained intent classification recognition model are invoked to recognize the text corpus and obtain the recognition result; The knowledge graph query conditions are constructed based on the recognition results.

5. The method for standardizing fault codes during the vehicle's service life according to claim 1, characterized in that, The trained fault entity recognition model is used to provide search targets for the user question and answer, and the trained intent classification recognition model is used to provide search directions for the user question and answer.

6. The method for standardizing fault codes during the vehicle's service life according to claim 1, characterized in that, The knowledge graph database is a database for online fault entity querying based on detailed fault code data structure attribute information.

7. The method for standardizing fault codes during the vehicle's service life according to claim 1, characterized in that, The method further includes: The results of the fault knowledge Q&A will be fed back to the user.

8. A standardized fault code system for the vehicle's service life, characterized in that, The system includes: The definition module is used to define standardized fault codes. When the standardized fault codes meet the requirements, the data of the standardized fault codes is updated to the database. The data acquisition module is used to collect user fault record data and enterprise after-sales data, and then perform data annotation to generate entity recognition datasets and intent classification datasets. The first generation module is used to generate a knowledge graph database using the Neo4j graph database tool, according to the data attributes, fault entities, and relation types of the standardized fault codes. The second generation module is used to load a pre-trained open-source BERT base model using the Transformers library, a natural language model development tool, to train the model on the entity recognition dataset and the intent classification dataset, and generate a trained fault entity recognition model and a trained intent classification recognition model. The first acquisition module is used to acquire the voice information input by the user; The construction module is used to construct knowledge graph query conditions based on the voice information, the trained fault entity recognition model, and the trained intent classification recognition model. The query module is used to query the knowledge graph database according to the knowledge graph query conditions to obtain fault detail node information; The second acquisition module is used to obtain the corresponding attribute field information from the fault details node information according to the preset template, and generate fault knowledge Q&A results.

9. An electronic device comprising a memory and a processor, wherein the memory stores a computer program executable on the processor, characterized in that, When the processor executes the computer program, it implements the method described in any one of claims 1 to 7.

10. A computer-readable medium having processor-executable non-volatile program code, characterized in that, The program code causes the processor to execute the method described in any one of claims 1 to 7.

Images