A fault diagnosis method and related device suitable for an onboard computer
By generating fault maps and colloquial mapping models, combined with scoring mechanisms and semantic analysis, the problem of difficult fault location under complex operating conditions of airborne computers was solved, achieving efficient and accurate fault diagnosis and maintenance.
Patent Information
- Application Number
- CN202610965037.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-06-30
- Publication Date
- 2026-08-25
AI Technical Summary
Existing fault diagnosis methods are unable to adapt to the complex operating conditions of airborne computers, resulting in difficulties in fault location, low maintenance efficiency, and an inability to meet the needs for efficient and accurate maintenance support.
By generating fault maps, a mapping model between colloquial real-world fault phenomena and fault modes is established. A fault priority management model is constructed based on a scoring mechanism. Combined with semantic analysis, non-standard fault problems are matched, and the optimal fault troubleshooting path is output.
It enables accurate diagnosis of intermittent and complex coupled faults, significantly improves the accuracy of fault location and maintenance efficiency, and meets the needs of efficient and accurate maintenance support for airborne computers.
Smart Images
Figure CN122633458A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of airborne computer reliability testing technology, and particularly relates to a fault diagnosis method and related apparatus suitable for airborne computers. Background Technology
[0002] As the complexity of equipment products continues to increase, traditional fault analysis methods that rely on manual experience are no longer able to meet the needs of large-scale, rapid on-site fault handling. This is especially true in high-reliability fields such as avionics, airborne equipment, and military products. Such equipment operates in special environments, exhibits complex fault phenomena, has diverse hierarchical relationships, and involves lengthy diagnostic chains. This results in extended fault diagnosis cycles, difficulty in establishing a closed loop of fault information, and difficulty in effectively accumulating and reusing relevant fault diagnosis knowledge and engineering experience. Consequently, these methods cannot meet the industry's development needs for efficient operation and maintenance of high-reliability equipment.
[0003] Currently, existing fault diagnosis methods in the industry are mainly divided into two categories: test-based modeling and expert reasoning. Both methods have obvious limitations: test-based modeling methods based on D-matrix reduction, while suitable for test-based and BITE system design in the development phase of airborne computers, can only achieve static analysis and cannot cope with the frequent intermittent faults that occur in actual operation; real-time diagnosis methods based on expert systems, while having the advantages of low computing power consumption and compatibility with airborne embedded platforms, and enabling rapid fault alarms, rely excessively on manually preset rules and are severely lacking in the ability to diagnose complex coupled faults and unknown faults. At the same time, current diagnostic technologies also have common application shortcomings. Airborne computers are affected by complex flight conditions, and fault manifestations are complex and diverse. Many superficial fault phenomena are difficult to match with preset fault mode descriptions, making fault location difficult; moreover, existing methods mostly focus on fault location and identification, without considering actual maintenance factors such as fault severity, troubleshooting costs, and probability of occurrence, and cannot output the most optimized fault troubleshooting path with the highest priority, resulting in low maintenance efficiency and excessively long fault troubleshooting cycles.
[0004] It is evident that existing fault diagnosis methods cannot meet the complex operating conditions required in high-voltage reliability fields such as airborne equipment, and have diagnostic limitations and application shortcomings, making it difficult to meet the actual needs of efficient and accurate maintenance and support for airborne computers. Summary of the Invention
[0005] This invention provides a fault diagnosis method and related device suitable for airborne computers. This method can adapt to the complex operating conditions of high-voltage reliability fields such as airborne equipment, and meets the actual needs of efficient and accurate maintenance and support for airborne computers.
[0006] To achieve the above objectives, the present invention adopts the following technical solution: A fault diagnosis method applicable to airborne computers, comprising: Modeling is performed based on the collected forward design fault analysis data, fault troubleshooting data, and fault repair data from the airborne computer to generate a fault map; Based on the received real-world fault phenomena, a mapping relationship between colloquial real-world fault phenomena and fault modes is established, and a comparative model of colloquial real-world fault phenomena and fault modes is generated. A fault priority management model is constructed based on a pre-defined scoring mechanism; Non-standardized fault problems are collected, and the target fault modes of non-standardized fault problems are determined through semantic analysis and matching based on fault maps and a model comparing colloquial real-world fault phenomena with fault modes. Based on the target fault mode, fault map and fault priority management model, the optimal fault investigation path is output.
[0007] Furthermore, the modeling based on the collected forward design fault analysis data, fault troubleshooting data, and fault repair data to generate a fault map includes: Collect data tables, fault zeroing reports, fault debugging records, and field maintenance records; By using Java and regular expressions to identify structured table fields, and by using a pre-built BERT+CRF named entity recognition model to extract fault modes, fault causes, fault effects, fault detection methods, fault failure rates, fault phenomenon descriptions, and fault handling measures from unstructured text, fault-related information is obtained to achieve data parsing and modeling.
[0008] Furthermore, the generation of the fault map includes: Using the Neo4j graph database as the storage medium, fault-related information is extracted as graph elements, and a hierarchical graph is constructed based on the graph elements with LRU or SRU as the root node; wherein, the hierarchical graph includes functional sub-circuit level and component level levels; Entity relationships are defined based on failure mode, failure cause, failure impact, and detection method. Failure nodes are linked together using CAUSES, AFFECTS, and LEADS_TO relationship types to form a complete failure chain and obtain a failure map.
[0009] Furthermore, the step of establishing a mapping relationship between colloquial real-world fault phenomena and fault modes based on received real-world fault phenomena in the field, and generating a comparative model of colloquial real-world fault phenomena and fault modes, includes: Based on the correlation data of real-world fault phenomena and fault modes and fault indications in product testing reports, a mapping relationship between colloquial real-world fault phenomena and fault modes is established, generating a colloquial real-world fault phenomenon and fault mode comparison model; wherein, the colloquial real-world fault phenomena include at least one of the following: the equipment cannot be turned on, the equipment occasionally crashes, it works intermittently, it freezes during operation, and it automatically shuts down.
[0010] Furthermore, the fault priority management model constructed based on the preset scoring mechanism includes: Establish a scoring mechanism that classifies failure modes into multiple levels based on their failure rates, with higher failure rates corresponding to higher levels. The failure rate, severity, probability of failure impact, and failure criteria of each failure mode are identified as characteristics for fault diagnosis. Based on the scoring mechanism, each fault diagnosis feature is scored and weighted to generate a fault priority management model.
[0011] Furthermore, the collection of non-standardized fault problems, based on fault maps and a model comparing colloquial real-world fault phenomena with fault modes, determines the target fault mode of the non-standardized fault problems through semantic analysis and matching, including: Collect non-standardized fault issues; Semantic analysis is performed on non-standardized fault problems, fault maps are used to associate fault chains, and preliminary matching is performed by combining colloquial real-world fault phenomena with fault mode comparison models. If the confidence level of the matching result is higher than the preset threshold, the target fault mode is directly determined. If the confidence level of the matching result is lower than the threshold, the fault information is supplemented through multiple rounds of follow-up questions, and then the target fault mode is determined in combination with the fault map.
[0012] Furthermore, after outputting the optimal fault diagnosis path based on the target fault mode, fault map, and fault priority management model, the method further includes: During the process of extracting fault data and modeling fault maps, performing fault semantic parsing, calling fault maps and matching fault patterns, user diagnostic feedback information is collected and the parameters of the semantic understanding model are optimized. This enables the optimization of the colloquial real-world fault phenomenon and fault pattern comparison model and the fault priority management model, which are then used in the subsequent fault diagnosis process of airborne computers.
[0013] A fault diagnosis system for airborne computers, comprising: The fault modeling unit is used to model faults based on the collected forward design fault analysis data, fault troubleshooting data, and fault maintenance data from the onboard computer, and generate fault maps. The interactive and visualization unit is used to establish a mapping relationship between spoken-language real-world fault phenomena and fault modes based on received real-world fault phenomena in the field, and to generate a comparative model of spoken-language real-world fault phenomena and fault modes. The first semantic understanding unit is used to build a fault priority management model based on a preset scoring mechanism; The second semantic understanding unit is used to collect non-standardized fault problems. Based on the fault map and the model comparing colloquial real fault phenomena with fault modes, the target fault mode of the non-standardized fault problem is determined through semantic analysis and matching. The third semantic understanding unit is used to output the optimal fault investigation path based on the target fault mode, fault map and fault priority management model.
[0014] A fault diagnosis device suitable for airborne computers, comprising: Memory, used to store computer programs; A processor is used to implement the above-described fault diagnosis method applicable to airborne computers when executing the computer program.
[0015] A computer-readable storage medium storing a computer program, which, when executed by a processor, is used to implement the above-described fault diagnosis method applicable to airborne computers.
[0016] Compared with the prior art, the present invention has the following beneficial effects: This invention provides a fault diagnosis method applicable to airborne computers. It constructs a dynamic fault map by integrating forward design fault analysis, troubleshooting, and maintenance data; establishes a mapping model between colloquial, realistic fault phenomena and fault modes; and builds a priority management model based on a scoring mechanism, incorporating factors such as fault hazard and troubleshooting costs. Furthermore, it utilizes semantic analysis to match non-standardized fault problems to determine target fault modes, and outputs the optimal troubleshooting path by combining the fault map and the priority management model. In this method, the fault map overcomes the limitations of static analysis, achieving dynamic representation of fault associations; the colloquial model solves the problem of mismatch between superficial phenomena and preset pattern descriptions through semantic adaptation; and the priority management model incorporates actual maintenance elements, ensuring the practicality and optimization orientation of the path. This method overcomes the shortcomings of existing technologies in diagnosing intermittent faults, complex coupled faults, and unknown faults, significantly improving the accuracy of fault location and maintenance efficiency, and meeting the needs of efficient and precise maintenance support for airborne computers. Attached Figure Description
[0017] Figure 1 A schematic diagram of a test report provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of an interactive question-and-answer interface provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of a fault map provided in an embodiment of the present invention; Figure 4 This is a diagram illustrating the interactive effect of a dialogue system provided in an embodiment of the present invention. Figure 5This is a structural block diagram of an intelligent fault diagnosis assistant equipped with a fault diagnosis method provided in an embodiment of the present invention; Figure 6 The core flowchart of a fault diagnosis method for airborne computers provided in an embodiment of the present invention; Figure 7 This is a schematic diagram of a fault diagnosis system for airborne computers provided in an embodiment of the present invention. Detailed Implementation
[0018] To further understand the content of this invention, the invention will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments are merely illustrative and not limiting of the invention.
[0019] The technical terms involved in this invention are explained below: FMEA: Failure Mode and Effects Analysis, which includes data sheets and failure-to-zero reports.
[0020] CAUSES: Causes, i.e., the reasons for the failure → failure mode.
[0021] AFFECTS: Effects, i.e., failure mode → failure impact.
[0022] LEADS_TO: Trigger, i.e., the impact of a fault → subsequent faults or results.
[0023] BERT: Transformer-based bidirectional encoder representation model.
[0024] CRF: Conditional Random Field.
[0025] As described in the background section, current diagnostic technologies still have significant limitations: airborne computers are affected by complex flight conditions, resulting in a wide variety of fault manifestations. Many superficial fault phenomena are difficult to match with fault patterns described in fault text, making fault location difficult. At the same time, most existing fault diagnosis methods focus on fault location and identification, failing to combine factors such as the severity of the fault, troubleshooting costs, and probability of occurrence to output the most optimized fault troubleshooting path that meets actual maintenance needs. This leads to low maintenance efficiency, long troubleshooting cycles, and difficulty in meeting the actual needs of efficient and accurate maintenance and support for airborne computers.
[0026] To address the aforementioned issues, this embodiment provides a fault diagnosis method applicable to airborne computers. This method is based on structured documents such as FMEA data tables and fault zeroing reports, and incorporates intelligent mapping technology between real fault phenomena and fault modes, as well as intelligent decision-making technology for fault priority based on multiple features. It also integrates key technologies such as Natural Language Understanding (NLU), knowledge graphs, and intelligent agent collaboration, proposing a digital and intelligent fault diagnosis method that automates the entire process from data parsing to intelligent diagnosis, assisting in the fault diagnosis and troubleshooting of airborne computers.
[0027] For example, such as Figure 6 As shown, this embodiment provides a fault diagnosis method suitable for airborne computers, including: Modeling is performed based on the collected forward design fault analysis data, fault troubleshooting data, and fault repair data from the airborne computer to generate a fault map; Based on the received real-world fault phenomena, a mapping relationship between colloquial real-world fault phenomena and fault modes is established, and a comparative model of colloquial real-world fault phenomena and fault modes is generated. A fault priority management model is constructed based on a pre-defined scoring mechanism; Non-standardized fault problems are collected, and the target fault modes of non-standardized fault problems are determined through semantic analysis and matching based on fault maps and a model comparing colloquial real-world fault phenomena with fault modes. Based on the target fault mode, fault map and fault priority management model, the optimal fault investigation path is output.
[0028] The fault diagnosis method provided in this embodiment will be further explained below: For example, this embodiment provides a fault diagnosis method suitable for airborne computers. This method is applied to fault diagnosis scenarios of airborne computers, and can also be applied to fault diagnosis scenarios of complex products such as missile-borne computer products. The specific implementation mechanism is as follows: This method comprises two parts: product management and fault diagnosis. Product management mainly involves the parsing and modeling of forward design fault analysis data, fault troubleshooting data, and fault repair data. The fault diagnosis part mainly involves the parsing, understanding, and matching of input fault data.
[0029] In product management, different areas are set up to isolate project data. Based on various types of fault data input from the product, fault maps are output and displayed interactively and visually.
[0030] Furthermore, this method also supports inputting real-world fault phenomena from the field, establishing a mapping relationship between colloquial real-world fault phenomena and fault modes in product management. Using data from product testing reports (the relationship between fault modes and fault indications), a comparative model of colloquial real-world fault phenomena and fault modes is formed. The semantic understanding unit uses the data in the comparative model to conduct multi-round interactive question-and-answer sessions, realizing the correspondence between colloquial fault phenomena and fault modes. The specific mapping relationship is shown in Table 1. Table 1 shows the mapping relationship.
[0031] In this embodiment, the economic and fault priority management decision-making technology first establishes an eight-level scoring mechanism. This mechanism is used to distinguish different levels based on the failure rate of fault modes. The higher the failure rate, the higher the level. Then, it sorts out the characteristics that affect fault diagnosis: failure mode failure rate, severity, failure impact probability, fault criteria, etc. Finally, it scores and weights the characteristics according to the scoring mechanism to form a fault priority management model.
[0032] Explainable, data such as FMECA, fault debugging records, and field maintenance records are collected, and then fault-related information is extracted from the data to complete the parsing and storage of historical fault data.
[0033] This method receives non-standardized input problems, parses them, and then uses a colloquial, realistic fault phenomenon and fault mode comparison model and a semantic understanding unit to describe and match faults, primarily leveraging the contextual understanding and domain adaptation capabilities of a large model. When the input colloquial, realistic fault phenomenon matches a similar fault mode using the comparison model, the method outputs the optimal fault investigation path using a fault priority management unit. If no match is found, a confidence threshold of 0.85 is set. If the confidence level is above the threshold, the optimal fault investigation path is output using the fault priority management unit; if it is below the threshold, multiple rounds of follow-up questions are asked through the large model to supplement information. Finally, the matched fault mode and the optimal fault investigation path are output.
[0034] It should be noted that the relevant information extracted above includes "fault mode", "fault impact", "fault cause", "fault detection method", "fault failure rate", "fault phenomenon description", and "fault handling measures". For FMEA tables in Excel / CSV format, Java + regular expressions are used to recognize structured table fields, and BERT + CRF named entity recognition model is used to extract unstructured text information such as "fault mode", "fault cause", and "fault impact".
[0035] In this embodiment, fault-related information is extracted as graph elements, and a hierarchical graph with LRU or SRU as root nodes is established using the Neo4j graph database as the storage medium. This includes functional sub-circuit level and component level. Entity relationships are defined according to "fault mode - fault cause - fault impact - detection method," and fault nodes are connected through relationship types such as CAUSES, AFFECTS, and LEADS_TO to form a complete fault chain.
[0036] For example, the fault diagnosis method for airborne computers provided in this embodiment has been specifically implemented and applied. The diagnostic method has been integrated into a specific software product to develop an intelligent fault diagnosis assistant. The specific implementation process is as follows: like Figure 5 As shown, this intelligent fault diagnosis assistant includes several core working units such as a fault modeling unit, a semantic understanding unit, an intelligent agent workflow unit, and an interactive and visualization unit.
[0037] like Figure 1 As shown, the fault modeling unit extracts key information from the FMEA data table and the zeroing report, including fault mode, fault impact, cause, detection method, and failure rate. The zeroing report includes: fault phenomenon description, handling measures, and result feedback. For Excel / CSV format FMEA tables and Word format zeroing reports, Java + regular expressions are used to identify structured table fields, and BERT + CRF named entity recognition model is used to extract unstructured text information such as "fault mode," "fault cause," and "fault impact," ensuring the accuracy of field extraction and the standardization of format.
[0038] like Figure 3 As shown, in the fault modeling unit, based on the parsed structured data, a hierarchical graph with LRU / SRU as the root node is established using the Neo4j graph database as the storage medium. According to the entity relationship definition of "fault mode - fault cause - fault impact - detection method", nodes are connected through relationship types such as CAUSES, AFFECTS, and LEADS_TO to form a complete fault chain and generate the fault graph.
[0039] Furthermore, the fault modeling unit supports both batch pre-import and single incremental import scenarios. After a new document is uploaded, an incremental modeling task is automatically triggered to update the graph nodes and relationships in real time. At the same time, data of different models is isolated by product ID to avoid confusion in cross-product diagnosis.
[0040] like Figure 2As shown, the semantic understanding unit utilizes the contextual understanding of the large model and the colloquial real-world fault phenomena and fault mode comparison model to directly map colloquial expressions (such as "the radar keeps disconnecting"), abbreviations, and vague descriptions to standard fault modes without the need for additional vector similarity calculations, thus breaking through the limitations of traditional rules on vocabulary and sentence structure.
[0041] like Figure 4 As shown, the semantic understanding unit sets a confidence threshold of 0.85. Results above the threshold are directly output; results below the threshold are supplemented through multiple rounds of follow-up questions from a larger model (e.g., "Are you referring to the main control board or the power supply board?") to reduce ambiguity. Combining product ID filtering (limiting the current product knowledge base), fault level matching (progressing by "system-component-specific fault"), and context enhancement (associating with historical dialogues) improves matching accuracy and avoids invalid recommendations. The system returns "most relevant fault mode + upstream and downstream related paths + cause list (confidence-ranked) + reasoning basis," providing interpretable decision support for subsequent reasoning and diagnosis. It should be noted here that... Figure 3 and Figure 4 This only shows the specific framework of the fault map and the dialogue-based collaborative interaction process as an example. The specific text content in the diagram will not be elaborated here and does not need to be paid attention to.
[0042] For example, the semantic understanding unit employs a bidirectional reasoning approach, combining top-down (fault phenomenon → root cause) and bottom-up (root cause → scope of impact) logic. Leveraging a path query algorithm, it supports concurrent computation across multiple paths, quickly locating root cause nodes and the complete fault chain. Using the "failure rate" from FMEA data as the core indicator, it automatically ranks and prioritizes multi-path fault causes based on a large model, prioritizing high-incidence fault modes. Custom ranking conditions such as severity and processing cost can be added later. It automatically generates a complete diagnostic chain of "symptoms → fault mode → impact path → detection methods → recommended measures," ensuring the rigor and traceability of the diagnostic logic.
[0043] For example Figure 4As shown, the interaction and visualization unit is supported by Vue + AntVG6 technology, creating an intuitive and interactive visualization interface that connects the semantic understanding layer output with user operations. It supports operations such as expanding / collapsed fault tree nodes, backtracking to parent nodes, and pruning invalid nodes; it distinguishes node states through color coding (red = current fault, orange = suspected node, green = ruled-out node), and the path highlighting function clearly displays the reasoning chain. It provides search-based interaction (quickly locating nodes by inputting the fault mode), dynamic reasoning demonstration (animated display of node highlighting and path expansion), and multi-path comparison (displaying multiple fault chains sorted by probability), adapting to different diagnostic scenarios. It adopts a "fault tree visualization on the left + natural language dialogue on the right" layout, with the semantic question-answering agent's questions and the reasoning agent's result updates synchronized to the interface in real time, realizing an interactive consultation mode.
[0044] The intelligent agent workflow unit, based on an intelligent agent building platform and component library, enables an automated diagnostic process for multi-agent collaboration. Core intelligent agent configuration: Intelligent agents are modularly divided into three types based on functional modules, with clearly defined input and output specifications to ensure module decoupling: Knowledge Auto-Modeling Agent: Inputs Excel / Word documents, outputs structured data, and supports direct generation of knowledge relationships; Fault Semantic Diagnosis Agent: An intelligent agent that takes fault description text as input and outputs matching results, etc., for fault tree diagnosis and interaction. Feedback Optimization Agent: Takes user diagnostic feedback as input; the agent collects user feedback to facilitate later optimization of model parameters. Technical Support System: Employs a self-developed intelligent agent framework, using message queues (MQ) for asynchronous communication between agents, Redis (remote dictionary service) for caching intermediate states, supporting concurrent workflow execution, and possessing the scalability to independently update modules and dynamically insert new agents.
[0045] In this embodiment, a virtuous cycle of "use-feedback-optimization" is constructed by leveraging the model capabilities of the agent and semantic understanding layer through feedback optimization. The user submits diagnostic results (correct / incorrect) feedback → the system records matching samples → incremental training of the semantic model is conducted based on the samples → diagnostic accuracy is continuously improved. This allows the system's diagnostic accuracy to gradually increase with the number of uses, reducing reliance on human experience and enhancing the level of intelligent operation and maintenance.
[0046] Therefore, this method mainly consists of two parts: product management and fault diagnosis. Product management primarily utilizes data such as FMECA (Fault Mode and Effects Analysis) results, product fault debugging records, field maintenance records, and test reports to build a product management database, supporting data isolation between projects. Product management is mainly accomplished through a fault modeling unit. The fault diagnosis part comprises a semantic understanding unit, an interaction and visualization unit, and an intelligent agent workflow unit. The semantic understanding unit receives external input, parses it, and matches it with information from the fault modeling unit, outputting the matched fault chain. The interaction and visualization unit connects the output of the semantic understanding unit with user operations. The intelligent workflow unit provides intelligent agent support for the other three units. The method employs intelligent mapping technology between real fault phenomena and fault modes, and intelligent decision-making technology for fault priority based on comprehensive multi-features, improving the accuracy and availability of fault diagnosis methods, ultimately achieving efficient fault location.
[0047] For example, such as Figure 7 As shown, this embodiment also provides a fault diagnosis system suitable for airborne computers, including: a fault modeling unit, used to model based on the collected forward design fault analysis data, fault investigation data and fault maintenance data of the airborne computer, and generate a fault map; The interactive and visualization unit is used to establish a mapping relationship between spoken-language real-world fault phenomena and fault modes based on received real-world fault phenomena in the field, and to generate a comparative model of spoken-language real-world fault phenomena and fault modes. The first semantic understanding unit is used to build a fault priority management model based on a preset scoring mechanism; The second semantic understanding unit is used to collect non-standardized fault problems. Based on the fault map and the model comparing colloquial real fault phenomena with fault modes, the target fault mode of the non-standardized fault problem is determined through semantic analysis and matching. The third semantic understanding unit is used to output the optimal fault investigation path based on the target fault mode, fault map and fault priority management model.
[0048] The present invention also provides a fault diagnosis device suitable for airborne computers, comprising: a memory for storing a computer program; and a processor for executing the computer program to implement the steps of the fault diagnosis method suitable for airborne computers.
[0049] The present invention also provides a computer program product, including a computer program / instructions that, when executed by a processor, implement the steps of the fault diagnosis method applicable to an airborne computer.
[0050] When the processor executes the computer program, it implements the above-mentioned steps for fault diagnosis applicable to airborne computers, such as: modeling based on the collected forward design fault analysis data, fault investigation data and fault repair data of the airborne computer, and generating a fault map; Based on the received real-world fault phenomena, a mapping relationship between colloquial real-world fault phenomena and fault modes is established, and a comparative model of colloquial real-world fault phenomena and fault modes is generated. A fault priority management model is constructed based on a pre-defined scoring mechanism; Non-standardized fault problems are collected, and the target fault modes of non-standardized fault problems are determined through semantic analysis and matching based on fault maps and a model comparing colloquial real-world fault phenomena with fault modes. Based on the target fault mode, fault map and fault priority management model, the optimal fault investigation path is output.
[0051] For example, the computer program may be divided into one or more modules / units, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules / units may be a series of computer program instruction segments capable of performing a preset function, the instruction segments describing the execution process of the computer program in the fault diagnosis device suitable for an airborne computer. For example, the computer program can be divided into a fault modeling unit, an interactive and visualization unit, a first semantic understanding unit, a second semantic understanding unit, and a third semantic understanding unit; the specific functions are as follows: the fault modeling unit is used to model faults based on the collected forward design fault analysis data, fault troubleshooting data, and fault maintenance data from the airborne computer, and generate a fault map; the interactive and visualization unit is used to establish a mapping relationship between spoken-language real fault phenomena and fault modes based on the received real-world fault phenomena in the field, and generate a spoken-language real fault phenomenon and fault mode comparison model; the first semantic understanding unit is used to construct a fault priority management model based on a preset scoring mechanism; the second semantic understanding unit is used to collect non-standardized fault problems, and determine the target fault mode of the non-standardized fault problems through semantic analysis and matching based on the fault map and the spoken-language real fault phenomenon and fault mode comparison model; the third semantic understanding unit is used to output the optimal fault troubleshooting path based on the target fault mode, the fault map, and the fault priority management model.
[0052] The fault diagnosis device for airborne computers can be a desktop computer, laptop, handheld computer, or cloud server, etc. The fault diagnosis device for airborne computers may include, but is not limited to, processors and memory. Those skilled in the art will understand that the above are examples of fault diagnosis devices for airborne computers and do not constitute a limitation on such devices. The device may include more components than described above, or combine certain components, or use different components. For example, the fault diagnosis device for airborne computers may also include input / output devices, network access devices, buses, etc.
[0053] The processor referred to can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor, or any conventional processor, etc. The processor is the control center for fault diagnosis of the airborne computer, connecting all parts of the fault diagnosis equipment for the airborne computer through various interfaces and lines.
[0054] The memory can be used to store the computer program and / or modules. The processor implements various functions of the fault diagnosis device suitable for airborne computers by running or executing the computer program and / or modules stored in the memory and calling the data stored in the memory.
[0055] The memory may primarily include a program storage area and a data storage area. The program storage area may store the operating system and at least one application program required for a function (such as sound playback, image playback, etc.). The data storage area may store data created based on the use of the mobile phone (such as audio data, phonebook, etc.). Furthermore, the memory may include high-speed random access memory and non-volatile memory, such as hard disks, RAM, plug-in hard disks, smart media cards (SMC), secure digital cards (SD cards), flash cards, at least one disk storage device, flash memory device, or other volatile solid-state storage devices.
[0056] The present invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the fault diagnosis method applicable to an airborne computer.
[0057] If the modules / units integrated in the fault diagnosis system for airborne computers are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium.
[0058] Based on this understanding, the present invention can implement all or part of the processes in the above-described fault diagnosis method for airborne computers, or it can be accomplished by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium. When executed by a processor, the computer program can implement the steps of the above-described fault diagnosis method for airborne computers. The computer program includes computer program code, which can be in the form of source code, object code, executable file, or a preset intermediate form, etc.
[0059] The computer-readable storage medium may include: any entity or device capable of carrying the computer program code, recording media, USB flash drive, portable hard drive, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal, and software distribution medium, etc.
[0060] It should be noted that the content contained in the computer-readable storage medium may be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the computer-readable storage medium does not include electrical carrier signals and telecommunication signals.
[0061] Compared with existing diagnostic methods, this invention provides a fault diagnosis method and related apparatus suitable for airborne computers, which has the following advantages: First, this invention can convert knowledge into assets, transforming scattered FMEA analysis reports and zeroing reports into a queryable and reasonable knowledge network, thereby achieving effective knowledge accumulation and management. Second, this invention relies on a colloquial real-fault phenomenon and fault mode comparison model to achieve rapid location of fault mode data in the knowledge network through field colloquial input, thereby improving fault diagnosis efficiency and significantly improving fault handling efficiency. Third, this invention relies on a fault priority management model to achieve the optimal fault troubleshooting path output; Fourth, it supports a continuous learning mechanism, which gradually improves diagnostic accuracy as the number of uses increases, reducing reliance on human experience; Fifth, this invention presents the reasoning path in the form of a fault tree, enhancing decision-making transparency and making it easier for users to understand and verify diagnostic results; Sixth, this invention is universal and scalable, and can be quickly extended to other product families. It can be run simply by importing the corresponding FMEA table, reducing system deployment costs and time.
[0062] The above embodiments are merely one of the implementation methods for achieving the technical solution of the present invention. The scope of protection claimed by the present invention is not limited to this embodiment, but also includes any variations, substitutions and other implementation methods that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention.
[0063] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the protection scope of the present invention.
Claims
1. A fault diagnosis method applicable to airborne computers, characterized in that, include: Modeling is performed based on the collected forward design fault analysis data, fault troubleshooting data, and fault repair data from the airborne computer to generate a fault map; Based on the received real-world fault phenomena, a mapping relationship between colloquial real-world fault phenomena and fault modes is established, and a comparative model of colloquial real-world fault phenomena and fault modes is generated. A fault priority management model is constructed based on a pre-defined scoring mechanism; Non-standardized fault problems are collected, and the target fault modes of non-standardized fault problems are determined through semantic analysis and matching based on fault maps and a model comparing colloquial real-world fault phenomena with fault modes. Based on the target fault mode, fault map and fault priority management model, the optimal fault investigation path is output.
2. The fault diagnosis method for airborne computers according to claim 1, characterized in that, The modeling based on the collected forward design fault analysis data, fault troubleshooting data, and fault repair data generates a fault map, including: Collect data tables, fault zeroing reports, fault debugging records, and field maintenance records; By using Java and regular expressions to identify structured table fields, and by using a pre-built BERT+CRF named entity recognition model to extract fault modes, fault causes, fault effects, fault detection methods, fault failure rates, fault phenomenon descriptions, and fault handling measures from unstructured text, fault-related information is obtained to achieve data parsing and modeling.
3. The fault diagnosis method for airborne computers according to claim 2, characterized in that, The generation of the fault map includes: Using the Neo4j graph database as the storage medium, fault-related information is extracted as graph elements, and a hierarchical graph is constructed based on the graph elements with LRU or SRU as the root node; wherein, the hierarchical graph includes functional sub-circuit level and component level levels; Entity relationships are defined based on failure mode, failure cause, failure impact, and detection method. Failure nodes are linked together using CAUSES, AFFECTS, and LEADS_TO relationship types to form a complete failure chain and obtain a failure map.
4. The fault diagnosis method for airborne computers according to claim 1, characterized in that, The process involves establishing a mapping relationship between spoken-language real-world fault phenomena and fault modes based on received real-world fault phenomena, and generating a comparative model of spoken-language real-world fault phenomena and fault modes, including: Based on the correlation data of real-world fault phenomena and fault modes and fault indications in product testing reports, a mapping relationship between colloquial real-world fault phenomena and fault modes is established, generating a colloquial real-world fault phenomenon and fault mode comparison model; wherein, the colloquial real-world fault phenomena include at least one of the following: the equipment cannot be turned on, the equipment occasionally crashes, it works intermittently, it freezes during operation, and it automatically shuts down.
5. A fault diagnosis method for airborne computers according to claim 1, characterized in that, The fault priority management model constructed based on the preset scoring mechanism includes: Establish a scoring mechanism that classifies failure modes into multiple levels based on their failure rates, with higher failure rates corresponding to higher levels. The failure rate, severity, probability of failure impact, and failure criteria of each failure mode are identified as characteristics for fault diagnosis. Based on the scoring mechanism, each fault diagnosis feature is scored and weighted to generate a fault priority management model.
6. The fault diagnosis method for airborne computers according to claim 1, characterized in that, The collection of non-standardized fault problems, based on fault maps and a model comparing colloquial real-world fault phenomena with fault modes, determines the target fault mode of the non-standardized fault problems through semantic analysis and matching, including: Collect non-standardized fault issues; Semantic analysis is performed on non-standardized fault problems, fault maps are used to associate fault chains, and preliminary matching is performed by combining colloquial real-world fault phenomena with fault mode comparison models. If the confidence level of the matching result is higher than the preset threshold, the target fault mode is directly determined. If the confidence level of the matching result is lower than the threshold, the fault information is supplemented through multiple rounds of follow-up questions, and then the target fault mode is determined in combination with the fault map.
7. A fault diagnosis method for airborne computers according to claim 1, characterized in that, After outputting the optimal fault diagnosis path based on the target fault mode, fault map, and fault priority management model, the following is also included: During the process of extracting fault data and modeling fault maps, performing fault semantic parsing, calling fault maps and matching fault patterns, user diagnostic feedback information is collected and the parameters of the semantic understanding model are optimized. This enables the optimization of the colloquial real-world fault phenomenon and fault pattern comparison model and the fault priority management model, which are then used in the subsequent fault diagnosis process of airborne computers.
8. A fault diagnosis system suitable for airborne computers, characterized in that, include: The fault modeling unit is used to model faults based on the collected forward design fault analysis data, fault troubleshooting data, and fault maintenance data from the onboard computer, and generate fault maps. The interactive and visualization unit is used to establish a mapping relationship between spoken-language real-world fault phenomena and fault modes based on received real-world fault phenomena in the field, and to generate a comparative model of spoken-language real-world fault phenomena and fault modes. The first semantic understanding unit is used to build a fault priority management model based on a preset scoring mechanism; The second semantic understanding unit is used to collect non-standardized fault problems. Based on the fault map and the model comparing colloquial real fault phenomena with fault modes, the target fault mode of the non-standardized fault problem is determined through semantic analysis and matching. The third semantic understanding unit is used to output the optimal fault investigation path based on the target fault mode, fault map and fault priority management model.
9. A fault diagnosis device suitable for airborne computers, characterized in that, include: Memory, used to store computer programs; A processor, configured to implement the fault diagnosis method for an airborne computer as described in any one of claims 1-6 when executing the computer program.
10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it is used to implement the fault diagnosis method applicable to an airborne computer as described in any one of claims 1-7.