New energy vehicle product whole life cycle problem management method based on knowledge graph

Abstract

The application discloses a new energy automobile product whole life cycle problem management method based on a knowledge graph, which comprises the following steps: step one: inputting a problem, searching the problem in a historical problem database to determine whether the problem is a repeated problem; if yes, step seven is executed; if no, step two is executed; step two: classifying the problem based on professional knowledge and experience to determine the type and nature of the problem; step three: finding measures for solving the problem in the historical problem database and proposing a problem solution; step four: evaluating the feasibility and effectiveness of the proposed problem solution; if the evaluation is passed, step five is executed; if the evaluation is not passed, step three is executed; step five: practically verifying the problem solution; if the verification is passed, step six is executed; if the verification is not passed, step three is executed; step six: uploading problem data to a data center to update the historical problem database; and step seven: completing the problem solution.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of product information management, and specifically relates to a new energy automobile product whole life cycle problem management method based on a knowledge graph. BACKGROUND

[0002] Under the current market environment, enterprises are facing increasingly fierce competition, and discovering and solving problems and ensuring the quality of products and services have become key elements of enterprise competitiveness. The importance of problem discovery and problem solving for current enterprises has risen to the strategic level of enterprise survival and development. Enterprises should attach great importance to the ability of problem discovery and problem solving, and establish perfect mechanisms and measures to continuously improve the problem discovery and problem solving ability of enterprises, so as to achieve the strategic goal of first-class work efficiency and competitiveness and continuously approach the ultimate goal of "eliminating recovery type problems, solving high-value prevention type and pursuit type problems".

[0003] A knowledge graph is a knowledge management carrier. By associating different modalities and different cycles of knowledge in the form of triples, complex information can be efficiently organized and managed. In the whole cycle problem management of new energy vehicles, constructing a knowledge graph helps to systematically manage various types of problems and their solutions. The construction of the new energy vehicle whole cycle problem management knowledge graph enables the knowledge graph to contain a large amount of vertical field knowledge of various problems in the life cycle of new energy vehicles. These knowledge not only includes technical problems in the research and development stage, process problems in the production stage, and management problems in the supply chain, but also includes customer complaints and feedback problems in the use stage. However, the new energy vehicle life cycle involves a large amount of data, and the data comes from various sources. How to efficiently collect, organize and integrate these data is the primary challenge of constructing a knowledge graph. The accuracy, completeness and consistency of the data need to be considered; in addition, the knowledge graph needs to have strong semantic analysis and reasoning capabilities to understand and infer the relationships between different problems. This requires deep technical accumulation in natural language processing, machine learning and knowledge representation; the technology of new energy vehicles and market demand are constantly changing, and the knowledge graph needs to be dynamically updated and maintained to reflect the latest technological progress and market feedback. This involves automated data collection and updating mechanisms, as well as effective management of old knowledge.

[0004] However, most enterprises still use word processing software, email, or even human memory to promote problem discovery and solution, which has the following disadvantages:

[0005] 1. The discovery, confirmation, analysis and solution of problems completely depend on the spontaneous promotion and memory of people, which is prone to omission or forgetfulness, which reduces the efficiency of problem discovery and solution and affects the quality of problem solution.

[0006] 2. Enterprise operation and management issues are numerous and complex. Prioritizing which issues to address depends entirely on people's perceptions and preferences. It is easy to overlook important and high-priority issues, which may lead to hidden dangers or losses in enterprise operation and management.

[0007] 3. The correctness of problem analysis depends entirely on human cognition and experience, lacking a universal analytical method, which reduces the efficiency and quality of problem analysis.

[0008] 4. The discovery, confirmation, analysis, resolution, and statistical analysis (including process and results) of problems require a lot of manpower and are prone to errors.

[0009] 5. The discovery and resolution of problems are not fully integrated with the enterprise's organizational structure. On the one hand, the efficiency of the process is affected by manual intervention, and there is a risk of leakage of enterprise management information due to improper access control. On the other hand, if the work performance of enterprise employees is to be evaluated based on the discovery and resolution of problems, it will consume a huge amount of manpower (or may not be feasible).

[0010] Based on the above issues, leveraging software technology to support enterprises in efficiently identifying and resolving problems, replacing the methods people use for problem-solving such as word processing software, email, and even human memory, is an urgent problem to be solved. Summary of the Invention

[0011] In view of this, in order to address the shortcomings of existing technologies in product lifecycle problem management, the purpose of this invention is to provide a knowledge graph-based method for managing the product lifecycle problems of new energy vehicles.

[0012] To achieve the above objectives, the present invention provides the following technical solution:

[0013] A knowledge graph-based method for managing issues throughout the entire lifecycle of new energy vehicle products includes the following steps:

[0014] Step 1: Enter the question. The system will search the historical question database to see if the question is a duplicate. If yes, proceed to Step 2; otherwise, the question is a new question, and proceed to Step 3.

[0015] Step Two: Classify the problems based on professional knowledge and experience to determine their type and nature;

[0016] Step 3: For recurring problems, based on the type of problem, find solutions in the historical problem database and propose solutions; for new problems, input the new problem into the knowledge graph for new energy vehicle product lifecycle problem management built in the historical problem database, make judgments and inferences about the new problem, and propose solutions to the new problem.

[0017] Step two: classify the problem based on professional knowledge and experience, determine the type and nature of the problem;

[0018] Step three: find measures to solve the problem in the historical problem library according to the type of the problem and propose a solution to the problem;

[0019] Step four: evaluate the feasibility and effectiveness of the proposed problem solution: if the problem solution passes the evaluation, execute step five; if the problem solution fails the evaluation, execute step three;

[0020] Step five: verify the problem solution in practice: if the problem solution passes the verification, execute step six; if the problem solution fails the verification, execute step three;

[0021] Step six: upload the problem data including the problem, the problem solution and the implementation experience to the data center to update the historical problem library;

[0022] Step seven: the problem solving is completed.

[0023] Further, the historical problem library is built with a new energy vehicle product life cycle problem management knowledge graph based on problem data, and the construction method of the new energy vehicle product life cycle problem management knowledge graph is:

[0024] S1: obtain relevant data information from the whole life cycle of new energy vehicles, including product design information in the research and development stage, product production information in the manufacturing stage and user feedback information in the service stage;

[0025] S2: based on the product historical maintenance database, combine the new energy vehicle whole life cycle problem library and the case library to form a new energy vehicle problem management database;

[0026] S3: for the new energy vehicle problem management database, use a multi-modal knowledge mining algorithm to present problem data of different modalities in text form as a carrier to form a set of triples of problem data of each modality;

[0027] S4: globally encode each modality problem data triple cluster, and perform modality fusion and cross-modality alignment to realize cross-modality problem data fusion and form a new energy vehicle product life cycle problem management knowledge graph.

[0028] Further, in step S3, the method steps for converting problem data of different modalities into triples using a multi-modal knowledge mining algorithm are:

[0029] S31: explore a multi-modal industrial data natural language description method, and establish an industrial knowledge mining framework for consistency generation of modality features and natural language description;

[0030] S32: An industrial knowledge mining framework utilizing the consistency between modal features and natural language descriptions converts non-textual modal data into text or graph forms, and achieves multimodal knowledge mining through text extraction methods, including:

[0031] For text modalities: First, the PL-Marker strategy is used to encode text sentences or paragraphs. Then, Albert is used for embedding representation. Finally, a hypergraph is used to model the embedding representation and construct a hypergraph neural network to extract text modal knowledge, forming a text modal triplet group.

[0032] For image modalities: Several multi-image description models are selected, and image description tasks are performed on the same image to obtain multiple image description texts. The description texts and images are embedded and represented, and feature recognition is performed. The description texts are ranked by quality, and the top-k description texts are combined into a prompt and input into a large language model to generate a fused image description. A text modal knowledge extraction method is used to extract industrial knowledge of image modalities and form image modal triplet groups.

[0033] For graphical modalities: First, all 3D models are converted to a standard format. Then, the PCL library is used to process the 3D models into point clouds to obtain the 3D point clouds of the models. Next, a 3D object detector is used to obtain the relevant features of the point clouds, and point cloud description text is generated based on multi-head cross-attention. Finally, a text modal knowledge extraction method is used to extract industrial knowledge of graphical modalities and form graphical modal triplet groups.

[0034] For signal modes: First, the signal data is stored in a standard database. Then, the signal is processed, and typical time-frequency domain features are extracted. Guided by prior knowledge such as sensor data, signal-text association rules are constructed. Finally, a text mode knowledge extraction method is used to extract industrial knowledge of signal modes and form signal mode triplet groups.

[0035] For expression modalities: based on the different forms of expression existence, they are divided into unstructured expression data and semi-structured expression data; for semi-structured expression data, discretized information is directly extracted and a graph structure is formed; for unstructured expression data, the expression in the image is first recognized by OCR, and then discretized to construct a scene graph of discrete elements, and finally form an expression modal triplet group.

[0036] Furthermore, in step S4, the method for forming a knowledge graph for managing the product lifecycle issues of new energy vehicles includes the following steps:

[0037] Based on the correlation of the modal problem data triple clusters, modal original data features are introduced, and global encoding is performed by combining the modal triple group graph features of each modal problem data triple cluster. The features of each part are assigned weights through dynamic cross-modal weighting, and then modal fusion and cross-modal alignment operations are performed to achieve cross-modal knowledge fusion and form a knowledge graph for the management of new energy vehicle product life cycle issues.

[0038] The knowledge graph for managing the product lifecycle issues of new energy vehicles uses URI numbers as core nodes and loads text attribute descriptions, modal attribute descriptions, standardized data attributes, and feature attributes.

[0039] Furthermore, in step three, when the problem type is a production quality problem, dynamic adjustments are made to IQC sampling, failure mode and effects analysis methods, and control plan updates. Measures to solve the problem are found in the historical problem database, and a solution is proposed.

[0040] Furthermore, the failure mode and effects analysis method includes the following steps:

[0041] 31) Identify failure modes: Determine the failure modes that cause product manufacturing quality problems;

[0042] 32) Assess failure impact: Assess the impact of failure modes caused by product manufacturing quality issues on system or product performance;

[0043] 33) Determine the cause of failure: Analyze the potential causes leading to the failure mode;

[0044] 34) Risk Priority Number Calculation: Calculate the risk priority number based on the severity, frequency of occurrence, and detection difficulty of the failure mode.

[0045] RPN = Severity × Frequency of Occurrence × Difficulty of Detection

[0046] 35) Develop and implement improvement measures: When the risk priority number exceeds the set threshold, develop and implement improvement measures;

[0047] 36) Verify the effectiveness of the improvement measures and update the FMEA document regularly to reflect the latest situation.

[0048] The beneficial effects of this invention are as follows:

[0049] This invention provides a knowledge graph-based method for managing issues throughout the entire lifecycle of new energy vehicle products, which has the following technical advantages:

[0050] (1) The method of this invention can not only effectively solve many problems in traditional engineering production management systems (PMS) and quality management systems (QMS), but also improve the quality management level of the entire product lifecycle through intelligent and data-driven management. In high-tech industries such as new energy vehicles, the application of product lifecycle problem management methods based on knowledge graphs will bring significant efficiency improvements and competitive advantages, and can solve the shortcomings of existing technologies in the discovery and resolution of problems. Attached Figure Description

[0051] To make the objectives, technical solutions, and beneficial effects of this invention clearer, the following figures are provided for illustration:

[0052] Figure 1 This is a flowchart of the knowledge graph-based method for managing the entire lifecycle problems of new energy vehicle products according to the present invention;

[0053] Figure 2 This is a framework diagram of the knowledge graph-based problem management method for the entire lifecycle of new energy vehicle products in this embodiment;

[0054] Figure 3 A flowchart illustrating the method for constructing a knowledge graph for managing product lifecycle issues in new energy vehicles;

[0055] Figure 4 This is a partial diagram of a knowledge graph for managing product lifecycle issues in new energy vehicles. Detailed Implementation

[0056] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, so that those skilled in the art can better understand and implement the present invention. However, the embodiments described are not intended to limit the present invention.

[0057] like Figure 1 As shown in the figure, the knowledge graph-based method for managing issues throughout the entire lifecycle of new energy vehicle products includes the following steps.

[0058] Step 1: Enter the question. The system will search the historical question database to see if the question is a duplicate. If yes, proceed to Step 2; otherwise, the question is a new question, and proceed to Step 3.

[0059] Specifically, recurring issues may arise at each stage of the new energy vehicle product lifecycle, as well as new issues not previously included in the historical database. These issues include: R&D problems, supplier problems, manufacturing problems, and customer complaints. The system utilizes historical problem databases, product production information, product design information, and material information to assess whether the problems are recurring.

[0060] Step 2: Classify the problems based on professional knowledge and experience to determine their type and nature.

[0061] For recurring problems, engineers rely on their extensive experience and expertise to determine the type of problem and categorize it accordingly, such as material issues, vehicle quality issues, and reliability issues in the automobile manufacturing process.

[0062] Step 3: For recurring problems, based on the problem type, find solutions from the historical problem database and propose a solution. Specifically, when the problem type is a production quality problem, dynamically adjust IQC sampling, failure mode and effects analysis methods, and update control plans to find solutions from the historical problem database and propose a solution. For new problems, input the new problem into the knowledge graph for new energy vehicle product lifecycle problem management built in the historical problem database, perform judgment and reasoning on the new problem, and propose a solution to the new problem.

[0063] Specifically, dynamically adjusting IQC sampling is based on product manufacturing quality data and historical problem records. This involves dynamically adjusting the product inspection sampling plan to improve inspection efficiency and effectiveness. This includes: data collection and analysis, risk assessment, sampling plan adjustment, implementation, and monitoring.

[0064] Failure Mode and Effects Analysis (FMEA) is a systematic approach used to identify potential failure modes and assess their impact on a system or product in order to take measures to prevent or reduce failures. Specifically, the Failure Mode and Effects Analysis method includes the following steps:

[0065] 31) Identify failure modes: Determine the failure modes that cause product manufacturing quality problems;

[0066] 32) Assess failure impact: Assess the impact of failure modes caused by product manufacturing quality issues on system or product performance;

[0067] 33) Determine the cause of failure: Analyze the potential causes leading to the failure mode;

[0068] 34) Risk Priority Number Calculation: Calculate the risk priority number based on the severity, frequency of occurrence, and detection difficulty of the failure mode.

[0069] RPN = Severity × Frequency of Occurrence × Difficulty of Detection

[0070] 35) Develop and implement improvement measures: When the risk priority number exceeds the set threshold, develop and implement improvement measures;

[0071] 36) Verify the effectiveness of the improvement measures and update the FMEA document regularly to reflect the latest situation.

[0072] A control plan is a document that details the control measures and inspection methods implemented during the product manufacturing process to ensure that product quality meets requirements. It includes: identifying change requests, revising control measures, updating inspection methods, implementation, and monitoring.

[0073] Step 4: Evaluate the feasibility and effectiveness of the proposed solutions: If the solution passes the evaluation, proceed to Step 5; if the solution fails the evaluation, proceed to Step 3.

[0074] Specifically, the feasibility and effectiveness of the proposed solution can be evaluated using the cost-benefit method. By collecting and analyzing relevant data on the solution, the return on investment, cost-benefit ratio, and other indicators can be calculated. If the solution passes the evaluation indicators, proceed to step five to conduct actual verification of the solution; if the solution fails the evaluation, proceed to step three to re-propose the solution.

[0075] Step 5: Verify the solution to the problem: If the solution to the problem is verified, proceed to Step 6; if the solution to the problem is not verified, proceed to Step 3.

[0076] Step Six: Upload the problem data, including the problem, the solution, and the implementation experience, to the data center to update the historical problem database, and then update the knowledge graph of problem management for the new energy vehicle product lifecycle.

[0077] Step 7: Problem solved.

[0078] Specifically, a knowledge graph for managing the product lifecycle of new energy vehicles is constructed based on the question data within the historical question database. The method for constructing this knowledge graph is as follows:

[0079] S1: Obtain relevant data and information from the entire life cycle of new energy vehicles, including: product design information in the R&D stage, product production information in the manufacturing stage, and user feedback information in the service stage, to build a problem database and case database for the entire life cycle of new energy vehicles.

[0080] S2: Based on the product history repair database, and combined with the new energy vehicle full life cycle problem database and case database, a new energy vehicle problem management database is formed.

[0081] S3: For the new energy vehicle problem management database, multimodal knowledge mining algorithms are used to present problem data of different modalities in text form, forming triplet clusters of problem data of each modality.

[0082] In this embodiment, the method steps for converting problem data of different modalities into triples using a multimodal knowledge mining algorithm are as follows.

[0083] S31: Explore natural language description methods for multimodal industrial data and establish an industrial knowledge mining framework based on the consistency between modal features and natural language descriptions.

[0084] S32: An industrial knowledge mining framework utilizing the consistency between modal features and natural language descriptions converts non-textual modal data into text or graph forms, and achieves multimodal knowledge mining through text extraction methods, including:

[0085] For text modalities: First, the PL-Marker strategy is used to encode text sentences or paragraphs. Then, Albert is used for embedding representation. Finally, a hypergraph is used to model the embedding representation and construct a hypergraph neural network to extract text modal knowledge, forming a text modal triplet group.

[0086] For image modalities: Several multi-image description models are selected, and image description tasks are performed on the same image to obtain multiple image description texts. The description texts and images are embedded and represented, and feature recognition is performed. The description texts are ranked by quality, and the top-k description texts are combined into a prompt and input into a large language model to generate a fused image description. A text modal knowledge extraction method is used to extract industrial knowledge of image modalities and form image modal triplet groups.

[0087] For graphical modalities: First, all 3D models are converted to a standard format. Then, the PCL library is used to process the 3D models into point clouds to obtain the 3D point clouds of the models. Next, a 3D object detector is used to obtain the relevant features of the point clouds, and point cloud description text is generated based on multi-head cross-attention. Finally, a text modal knowledge extraction method is used to extract industrial knowledge of graphical modalities and form graphical modal triplet groups.

[0088] For signal modes: First, the signal data is stored in a standard database. Then, the signal is processed, and typical time-frequency domain features are extracted. Guided by prior knowledge such as sensor data, signal-text association rules are constructed. Finally, a text mode knowledge extraction method is used to extract industrial knowledge of signal modes and form signal mode triplet groups.

[0089] For expression modalities: based on the different forms of expression existence, they are divided into unstructured expression data and semi-structured expression data; for semi-structured expression data, discretized information is directly extracted and a graph structure is formed; for unstructured expression data, the expression in the image is first recognized by OCR, and then discretized to construct a scene graph of discrete elements, and finally form an expression modal triplet group.

[0090] S4: Globally encode the triplet clusters of problem data for each modality, and perform modality fusion and cross-modality alignment to achieve cross-modal problem data fusion and form a knowledge graph for problem management in the product lifecycle of new energy vehicles.

[0091] In this embodiment, the steps for forming a knowledge graph for managing the product lifecycle issues of new energy vehicles are as follows:

[0092] Based on the correlation of modal problem data triple clusters, modal original data features are introduced, and global encoding is performed by combining the modal triple group graph features of each modal problem data triple cluster. The features of each part are assigned weights through dynamic cross-modal weighting, and then modal fusion and cross-modal alignment operations are performed to achieve cross-modal knowledge fusion and form a knowledge graph for the management of new energy vehicle product life cycle issues.

[0093] The knowledge graph for managing the product lifecycle issues of new energy vehicles uses URI numbers as core nodes and loads text attribute descriptions, modal attribute descriptions, standardized data attributes, and feature attributes.

[0094] The specific implementation of this embodiment will be described below with reference to specific examples.

[0095] Case Example: Implementation of a New Energy Vehicle Manufacturing Company

[0096] When receiving input about R&D issues, such as an overheating problem in a certain engine model;

[0097] Based on information such as historical problem database, product manufacturing information, product design information, and material information, determine whether the problem is a duplicate problem.

[0098] If the problem is identified as a repetitive issue, record the problem details for subsequent analysis;

[0099] For recurring problems, experienced engineers analyze and determine that the problem is a design flaw in the engine cooling system, classifying it as a design issue. For new problems not included in the historical database, the new problems are input into the knowledge graph for new energy vehicle product lifecycle problem management, and methods such as knowledge reasoning are used to judge and reason about the new problems, proposing solutions to them.

[0100] To address this design problem, engineers used methods such as semantic parsing, graph embedding, and graph reasoning to search for solutions in existing knowledge graphs and find previous methods for solving similar problems, including optimizing heat sink design and adjusting coolant flow.

[0101] Semantic parsing: Perform deep semantic parsing on the problem description to extract key elements (such as engine location, supplier name, problem type, etc.).

[0102] Graph embedding: The parsed elements are embedded into a pre-built knowledge graph for similarity calculation and matching.

[0103] Graphical reasoning: Using historical experience and related information in a graph, we can deduce the most relevant solutions and preventative measures to the current problem.

[0104] Based on the graph, a solution to the problem was found. The cause of the problem was identified as an unreasonable heat sink design. The proposed solutions were to optimize the shape of the heat sink and increase the coolant flow rate. At the same time, preventive measures were developed, such as conducting thermal simulation analysis in future designs.

[0105] The evaluation of solutions for optimizing heat sink design and increasing coolant flow rate is crucial to determine their feasibility and effectiveness. This evaluation typically employs an expert review method, where experts in the relevant field or experienced engineers discuss and review the proposed solutions. The expert review method relies on the knowledge and experience of the experts to provide a multi-faceted evaluation of the solutions.

[0106] The optimized cooling system was tested in a real vehicle to verify the effectiveness of the solution. Test results and solution data were stored in a data center.

[0107] The relevant information, including design standards, production information, and material properties, will be updated according to the adjusted measures. For example, the design standards for heat sinks and the production information for cooling systems will be updated.

[0108] The updated relevant information and the categorized questions are used as input to update the knowledge graph. The updated knowledge graph includes new heat sink design standards and coolant flow adjustment schemes, which will serve as guidance to support future graph retrieval and problem solving.

[0109] The above-described embodiments are merely preferred embodiments provided to fully illustrate the present invention, and the scope of protection of the present invention is not limited thereto. Equivalent substitutions or modifications made by those skilled in the art based on the present invention are all within the scope of protection of the present invention. The scope of protection of the present invention is defined by the claims.

Claims

1. A knowledge graph-based new energy vehicle product full life cycle problem management method, characterized in that: Comprise the following steps: Step one: input the problem, retrieve whether the problem is a repeated problem from the historical problem library: if yes, execute step two; if no, the problem is a new problem, execute step three; Step two: classify the problem based on professional knowledge and experience to determine the type of the problem; Step three: for repeated problems, find the measures to solve the problem and propose a solution to the problem according to the type of the problem in the historical problem library; for new problems, input the new problem into the new energy vehicle product life cycle problem management knowledge graph constructed in the historical problem library, judge and reason the new problem, and propose a solution to the new problem; Step four: evaluate the feasibility and effectiveness of the proposed problem solving solution: if the problem solving solution passes the evaluation, execute step five; if the problem solving solution fails the evaluation, execute step three; Step five: verify the problem solving solution in practice: if the problem solving solution is verified, execute step six; if the problem solving solution is not verified, execute step three; Step six: upload the problem and the problem solving solution to the data center to update the historical problem library; Step seven: problem solving is completed; The new energy vehicle product life cycle problem management knowledge graph is constructed based on the multi-modal knowledge mining algorithm, which forms a cluster of triplets of problem data in each mode, and specifically comprises: For text mode: encode the text sentence or paragraph using PL-Marker strategy, embed the encoding using Albert model, model the embedded representation using hypergraph, and extract knowledge from text mode using the constructed hypergraph neural network to form a cluster of triplets of text mode; For image mode: obtain multiple image description texts, embed and identify the description texts and images, and sort the description text quality, and input the top-k description texts into a large language model to generate a fused image description; extract knowledge from the fused image description to form a cluster of triplets of image mode; For graphics mode: convert all three-dimensional models to a standard format, and then introduce PCL library to obtain three-dimensional point cloud of the three-dimensional model; generate point cloud description text based on point cloud features and multi-head cross attention; extract knowledge from the point cloud description text to form a cluster of triplets of graphics mode; For signal mode: extract typical time domain features and typical frequency domain features from signal data; extract knowledge from the typical time domain features and typical frequency domain features to form a cluster of triplets of signal mode; For expression mode: according to different forms of expression, divide into unstructured expression data and semi-structured expression data; for semi-structured expression data, directly extract discrete information and form a graph structure; for unstructured expression data, recognize the expression in the picture through OCR, and then discretize and form a graph structure; form a cluster of triplets of expression mode based on the graph structure.

2. The knowledge graph-based new energy vehicle product whole life cycle problem management method according to claim 1, characterized in that: The method for forming the new energy vehicle product life cycle problem management knowledge graph is: According to the association relationship of the modal problem data triple cluster, the modal original data feature is introduced, the global coding is carried out in combination with the modal triple group graph features of each modal problem data triple cluster, the weight distribution is allocated to each part feature through dynamic cross-modal weighting, then the modal fusion and cross-modal alignment operation are carried out, the cross-modal knowledge fusion is realized, and the new energy automobile product life cycle problem management knowledge graph is formed; The new energy automobile product life cycle problem management knowledge graph takes the URI number as the core node, and loads the text attribute description, the modal attribute description, the standardized data attribute and the characteristic attribute. 3.The knowledge graph-based new energy vehicle product whole life cycle problem management method according to claim 1, characterized in that: In step three, when the problem type is production quality problem, the dynamic adjustment of IQC sampling, failure mode and effect analysis method and control plan updating are adopted, the measures to solve the problem are found in the historical problem library, and the problem solving scheme is proposed.

4. The new energy vehicle product whole life cycle problem management method based on a knowledge graph according to claim 3, characterized in that: The failure mode and effect analysis method includes the following steps: 31) Identify failure mode: determine the failure mode of the product quality problem; 32) Evaluate failure impact: evaluate the impact of the product quality problem failure mode on system or product performance; 33) Determine failure cause: analyze the potential causes of the failure mode; 34) Risk priority calculation: calculate the risk priority according to the severity, frequency and detection difficulty of the failure mode: RPN = severity × frequency × detection difficulty 35) Develop improvement measures: when the risk priority is greater than the set threshold, develop and implement improvement measures; 36) Verify the effectiveness of the improvement measures, and update the FMEA document regularly to reflect the latest situation.

Images