Comprehensive intelligent fault diagnosis method for airplane

Abstract

The invention belongs to the field of aircraft comprehensive guarantee work, and particularly relates to an aircraft comprehensive intelligent fault diagnosis method which comprises a man-machine interface, a knowledge base, a database and an inference engine. The input end of the man-machine interface and the input end of the knowledge base are both connected with a diagnosed object, the output end of the man-machine interface is connected with the database, the database and the knowledge base are both interconnected with the inference engine, and knowledge of various fault types is stored in the knowledge base; aircraft working data and fault type data are stored in the database, and the inference engine can obtain data in the knowledge base and the database to diagnose the current state of a diagnosed object and give a diagnosis result; according to the method, the whole system of the aircraft can be subjected to quick fault positioning, detection, troubleshooting and the like, so that the aircraft can be quickly repaired after landing, the diagnosis, maintenance and logistics guarantee of the aircraft are quickly and efficiently connected to form intelligent, comprehensive and integrated comprehensive guarantee, and finally, the fault positioning analysis time efficiency of the aircraft is reduced.

Description

Technical Field

[0001] The present application belongs to the field of aircraft integrated support work, and in particular relates to an aircraft integrated intelligent fault diagnosis method. Background Art

[0002] Aircraft maintainability and supportability are two important indicators for considering aircraft dispatch capabilities. Modern aircraft airborne systems generally use airborne digital computers to achieve real-time identification and isolation of aircraft system failures. Typical aircraft systems include flight control systems, power supply systems, lighting systems, warning systems, atmospheric systems, etc. When a failure occurs in any aircraft system, the airborne digital computer will perform fault detection and record the failure, providing aircraft system fault diagnosis and positioning for aircraft ground maintenance personnel.

[0003] An existing method for diagnosing faults of aircraft fuel system accessories can quickly handle faults of aircraft fuel system accessories; however, the disadvantage is that it does not propose a universal diagnostic idea and process for aircraft faults and does not have the breadth of aircraft fault handling.

[0004] Therefore, how to achieve universal aircraft fault diagnosis is a problem that needs to be solved. Summary of the invention

[0005] The purpose of this application is to provide an integrated intelligent aircraft fault diagnosis method to solve the problem that existing aircraft fault diagnosis methods are difficult to generalize.

[0006] The technical solution of the present application is: a comprehensive intelligent fault diagnosis method for aircraft, including a human-machine interface, a knowledge base, a database and an inference engine; the input ends of the human-machine interface and the knowledge base are connected to the diagnosed object, the output end of the human-machine interface is connected to the database, the database and the knowledge base are interconnected with the inference engine, and the knowledge base stores knowledge of various fault types; the database stores aircraft working data and fault type data, and the inference engine can obtain the data in the knowledge base and the database to diagnose the current state of the diagnosed object and give a diagnosis result; a knowledge acquisition unit is also provided between the human-machine interface and the knowledge base, and the knowledge acquisition unit can capture various fault data transmitted by the human-machine interface, obtain different categories of knowledge after sorting, and transmit them to the knowledge base.

[0007] Preferably, the knowledge stored in the knowledge base includes: fault case knowledge, FMECA knowledge, FTA knowledge and IETM knowledge.

[0008] Preferably, the knowledge acquisition unit includes an IETM data acquisition module, a fault tree data acquisition module, an FMECA data acquisition module and a fault case knowledge acquisition module; the IETM data acquisition module can acquire knowledge data from the IETM data; the fault tree data acquisition module can extract knowledge data from the fault tree model; the FMECA data acquisition module can extract knowledge data from the FMECA table; and the fault case knowledge acquisition module can acquire all fault cases in the aircraft system equipment.

[0009] Preferably, an IETM data storage module is provided in front of the human-machine interface. The IETM data storage module is loaded with all IETM data that conforms to the standards and is capable of parsing and running the IETM data. The specific method of parsing and running the IETM data is to embed the S1000D standard IETM operating platform. When the IETM data storage module is running, records are made through the IETM operating platform and feedback is given to the IETM maintainer according to the situation.

[0010] Preferably, after receiving the fault diagnosis request sent by the inference engine, the IETM data acquisition module retrieves the corresponding data from the IETM data storage module through the human-machine interface; when the IETM data is updated, the IETM data storage module can synchronously send the updated data to the IETM data acquisition module.

[0011] Preferably, the fault tree data acquisition module imports the fault tree analysis data through a human-machine interface, and converts the fault tree analysis data into structured data to form and store fault tree knowledge.

[0012] Preferably, the fault tree analysis data includes the analysis object, the fault tree model, the minimum cut set of each node of the fault tree, the failure probability of each minimum cut set, the failure probability of each node of the fault tree, the importance, the critical importance, the diagnostic importance, and the fault phenomenon description, the fault symptom set, and the maintenance measures of each level of the fault tree node.

[0013] Preferably, the FMECA data acquisition module obtains the FMECA table from the human-machine interface and extracts fault diagnosis knowledge data from the FMECA table, including generating structural data of the hierarchical relationship of each system, subsystem or sub-subsystem, product component or part of the aircraft; diagnostic object knowledge of the expert system; the failure rate, fault judgment criteria, and whether each diagnostic object belongs to the minimum equipment list; the failure mode, mean interval time between failure modes, failure cause, impact of failure, fault detection method, failure mode characteristic value, failure mode severity, failure mode frequency ratio, maintenance measures, and mission stage compensation measures of each diagnostic object.

[0014] Preferably, the fault case knowledge acquisition module obtains fault case knowledge data from the human-machine interface, and analyzes and classifies and stores it to form a fault case knowledge base; the fault case knowledge base includes the name of the fault, fault phenomenon description, fault code, system to which the fault belongs, time of fault discovery, cause of the fault, fault characteristic value, fault component, fault mode, fault handling measures and troubleshooting effect.

[0015] Preferably, when the knowledge base performs knowledge diagnosis, it first obtains the fault case knowledge in the fault case knowledge acquisition module, and applies the fault case knowledge to case matching. If the best matching result cannot be obtained, the IETM data in the IETM data acquisition module is obtained; the IETM data is used to retrieve and isolate the fault. If the fault is successfully eliminated, the fault diagnosis process and results are organized into a fault case form and added to the fault case knowledge base again.

[0016] Preferably, if the diagnostic result still cannot be obtained, the fault diagnosis knowledge is obtained from the FMECA data acquisition module, and the fault tree knowledge is obtained from the fault tree data acquisition module; the fault diagnosis knowledge and fault tree knowledge are used to perform intelligent fault diagnosis. If the fault is successfully eliminated through the diagnostic result, the diagnostic process and results are organized into a fault case form and added to the fault case knowledge base.

[0017] Preferably, the method for the inference engine to perform fault diagnosis is: first receiving the fault symptom provided by the user, then obtaining case-based expert knowledge stored in the fault case knowledge base, selecting the required reasoning strategy to perform fault diagnosis based on the case matching method, judging whether the fault mode and recommended maintenance measures can be given, and if so, outputting the diagnosis result;

[0018] If the fault mode and recommended maintenance measures cannot be given, the fault diagnosis is performed based on IETM, and it is determined whether fault isolation measures or maintenance measures are obtained. If so, the diagnosis results are output;

[0019] If the fault isolation measures or maintenance measures cannot be obtained through IETM, the intelligent fault diagnosis based on the computer system is used to determine whether the fault cause is isolated and maintenance measures are given. If so, the diagnosis result is given; if not, the knowledge data is supplemented and adjusted from the knowledge base and the judgment is repeated;

[0020] After the diagnosis results are given, the diagnosis results and maintenance measures suggestions are sent to the user through the database and human-computer interface.

[0021] Preferably, before each diagnosis is completed and the diagnosis result is given, a judgment is made as to whether the maintenance measures are effective. If effective, the diagnosis result is output; if not effective, the next fault diagnosis step is still performed.

[0022] Preferably, the reasoning strategies selected by the inference engine include: rule reasoning, forward reasoning, backward reasoning and forward and backward mixed reasoning.

[0023] Preferably, the method for fault diagnosis based on case matching is: obtain the fault phenomenon, symptom data, and the system to which the fault belongs, match the acquired data information with the fault case knowledge according to the rules, and determine whether the match is successful. If so, output the diagnosis result; if not, first prompt the user to perform detection operations or start on-board self-detection based on the matched approximate case, supplement the data symptom data, and perform rule matching again; if the match is still unsuccessful, obtain the IETM data in the IETM data acquisition module and switch to IETM diagnosis.

[0024] Preferably, before performing fault diagnosis based on the case matching method, all fault case knowledge in the fault case knowledge base is first obtained, and then the feature vector set of all fault case knowledge is determined. When matching, the fault phenomenon needs to be decomposed and entered according to the category of the feature vector and matched; then the fault case knowledge is classified and stored according to the feature vector set; at the same time, each fault case knowledge is described with quantitative data or enumerated qualitative data.

[0025] Preferably, after each fault diagnosis is completed, the fault diagnosis result data is saved in a fault case knowledge base.

[0026] Preferably, the specific method of IETM-based fault diagnosis is: obtaining the fault phenomenon, symptom data, and the system to which the fault belongs, and determining the type of fault; after obtaining the fault type, combining the reasoning engine of the IETM reader to guide maintenance personnel to complete the fault diagnosis; if the fault type cannot be obtained, switching to intelligent fault diagnosis.

[0027] Preferably, the fault types include isolated faults, detection faults, observation faults and related faults.

[0028] Preferably, the isolated faults are input into the IETM operating platform for fault retrieval by entering the fault code or fault phenomenon description information transmitted through the data link on the air-to-ground data link receiver, and the corresponding troubleshooting and maintenance measures are obtained; the diagnosis results are then presented to the user through the human-machine interface. After the user confirms that the maintenance measures given for the fault type have been successfully eliminated, the fault information is organized into a fault case form and added to the case class knowledge base.

[0029] Preferably, a fault isolation procedure is given for detection-type faults, and questions are asked according to the fault isolation steps. After the user answers the questions, the system gives the next step of action or continues to ask questions based on the answers, and through such interaction, the cause of the fault is found and maintenance measures are given. If IETM gives multiple LRUs in sequence, the user is prompted to perform corresponding tests or detection operations, and judge whether the LRU is faulty based on the test results fed back by the user. If it is faulty, corresponding maintenance measures are given for the LRU. After the user confirms that the fault has been successfully eliminated according to the given maintenance measures, the fault code, fault phenomenon description, and test results and maintenance measures generated during the IETM isolation diagnosis process are added to the case-type knowledge base in the form of fault cases.

[0030] Preferably, the observed faults are interactively guided and isolated through IETM, and if the final maintenance measures can successfully eliminate the fault, the fault phenomenon description of the fault, various test results generated during the IETM-guided isolation process, and maintenance measures are added to the case class knowledge base in the form of fault cases.

[0031] Preferably, relevant faults are guided and isolated using IETM. If the maintenance measures finally given can successfully prevent these warning messages and BIT detection information from appearing again, then these warning messages, BIT detection information and their corresponding fault codes, and various test results and maintenance measures generated during the IETM guided isolation process are organized into the form of fault cases and added to the case class knowledge base.

[0032] Preferably, before the intelligent fault diagnosis based on the computer system, a knowledge base of fault diagnosis is first constructed according to the relevant data of FMECA and FTA of the aircraft design department and the data accumulated by the troops; based on the status data collected by MIM, a set of all possible fault modes that cause the abnormality can be given, and the level of each fault mode can be calculated; finally, the maintenance aggregation model continuously obtains different types of fault case data to obtain the efficiency of each maintenance method in repairing a fault mode.

[0033] Preferably, the steps for calculating the level of the fault mode are: obtaining abnormal data, obtaining a set of fault modes from the MIM according to the input abnormal data, and then calculating the level of each fault mode according to the PHM status data; then obtaining the corresponding maintenance task from the MIM according to the level of the fault mode, and then using supplementary learning technology to calculate the level of the maintenance task; finally, entering the autonomous support information system to determine the best maintenance plan.

[0034] The aircraft comprehensive intelligent fault diagnosis method of the present application can quickly locate, detect, and troubleshoot faults in the entire aircraft system, so that the aircraft can be quickly repaired after landing, thereby quickly and efficiently connecting the aircraft's diagnosis, maintenance, and logistics support to form an intelligent, comprehensive, and integrated comprehensive support, ultimately achieving the effect of reducing the time required for aircraft fault location analysis, shortening the aircraft troubleshooting cycle, and reducing labor costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the technical solution provided by the present application, the following is a brief introduction to the accompanying drawings. Obviously, the accompanying drawings described below are only some embodiments of the present application.

[0036] Figure 1 This is the overall system structure diagram of this application;

[0037] Figure 2 This is a comprehensive intelligent diagnosis flow chart for this application;

[0038] Figure 3 This is a schematic diagram of the working principle of the aircraft comprehensive intelligent fault diagnosis method of this application;

[0039] Figure 4 This is a diagnostic flowchart based on the case matching method for this application;

[0040] Figure 5 Schematic diagram of the interactive electronic technical manual for this application. DETAILED DESCRIPTION

[0041] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0042] An integrated intelligent fault diagnosis system for aircraft, such as Figure 1-Figure 3 , including a human-machine interface, a knowledge base, a database and an inference engine. The input ends of the human-machine interface and the knowledge base are connected to the object to be diagnosed, and the output end of the human-machine interface is connected to the database. The database and the knowledge base are interconnected with the inference engine. The knowledge base stores knowledge of various fault types; the database stores aircraft working data and fault type data. The inference engine can obtain the data in the knowledge base and the database to diagnose the current state of the object to be diagnosed and give the diagnosis result. A knowledge acquisition unit is also set between the human-machine interface and the knowledge base. The knowledge acquisition unit can capture various fault data transmitted by the human-machine interface, obtain different categories of knowledge after sorting, and transmit them to the knowledge base.

[0043] This design is not only suitable for diagnosing faults discovered by air and ground crews and faults monitored on the ground, but also for diagnosing faults by receiving various sensor parameters, status information and warning prompt information recorded by the portable support equipment (hereinafter referred to as PMA) on the air-ground data link in real time.

[0044] Preferably, the knowledge stored in the knowledge base includes: fault case knowledge (fault phenomenon, fault feature set, fault cause, troubleshooting steps, etc.), FMECA knowledge (diagnostic object, fault mode, fault phenomenon description, fault mode feature set, fault mode frequency ratio, maintenance measures, etc.), FTA knowledge (diagnostic object, i.e., top event, fault tree hierarchical relationship, feature set of faults at each node of the fault tree, failure probability of nodes at each level, maintenance measures for faults at each node, etc.), and IETM knowledge.

[0045] Preferably, the knowledge acquisition unit includes an IETM data acquisition module, a fault tree data acquisition module, an FMECA data acquisition module and a fault case knowledge acquisition module; the IETM data acquisition module can acquire knowledge data from the IETM data; the fault tree data acquisition module can extract knowledge data from the fault tree model; the FMECA data acquisition module can extract knowledge data from the FMECA table; and the fault case knowledge acquisition module can acquire all fault cases in the aircraft system equipment.

[0046] Preferably, an IETM data storage module is provided in front of the human-machine interface. The IETM data storage module is loaded with all IETM data that conform to the S1000D standard and can parse and run the IETM data. The specific method of parsing and running the IETM data is to embed the S1000D standard IETM operating platform. When the IETM data storage module is running, the IETM operating platform records the data and feeds back to the IETM maintainer according to the situation.

[0047] After receiving the fault diagnosis request from the inference engine, the IETM data acquisition module retrieves the corresponding data from the IETM data storage module through the human-machine interface; when the IETM data is updated, the IETM data storage module can synchronously send the updated data to the IETM data acquisition module.

[0048] Preferably, the fault tree analysis method (hereinafter referred to as FTA) is an important method for analyzing the reliability and safety of aircraft systems. The fault tree model is a behavioral model based on the structure and functional characteristics of the research object. It is a qualitative causal model. It takes the most undesirable event of the aircraft system as the top event, and other events that may cause the top event as the bottom event. It uses logic gates to represent the relationship between events. It is an inverted tree-structured logic diagram. The purpose of qualitative fault tree analysis is to find the cause or combination of causes that lead to the occurrence of the top event, and to identify all fault modes that lead to the occurrence of the top event, that is, all minimum cut sets. Combined with these characteristics, FTA data can be used as fault diagnosis knowledge in the diagnosis process.

[0049] Traditional FTA exists in graphic form, which is not convenient for computer systems to share and use FTA data. The fault tree data acquisition module imports fault tree analysis data through the human-computer interface, and converts the fault tree analysis data into structured data to form and store fault tree knowledge.

[0050] Fault tree analysis data includes analysis object (top event, i.e., fault description), fault tree model (node ​​description at each level, logical relationship with upper level node), minimum cut set of each node of fault tree, failure probability of each minimum cut set, failure probability of each node of fault tree, importance (Importance Factor is a measure of the impact of bottom event on top event), critical importance (MarginalImportanceFactor, MIF refers to the difference in system reliability between failure and normal conditions of a certain module), diagnosis importance (DiagnosisImportanceFactor, DIF is a measure of the impact of a specific module failure on system reliability, expressed as a probability value), and then supplemented with fault phenomenon description, fault symptom set, maintenance measures and other information of fault tree nodes at each level. Extended fault tree knowledge is formed and stored in a multi-type shared knowledge base in a structured form.

[0051] Preferably, during the development and design phase of an aircraft, not all faults of the aircraft are analyzed by fault tree analysis. Due to resource and time constraints, generally, only catastrophic faults and fatal faults are analyzed as top events. Therefore, the FTA data available to users is limited, and if more fault tree data is to be obtained, it can only be supplemented by the user during the use phase. To solve this problem, the FMECA data acquisition module can obtain fault diagnosis knowledge from FMECA data.

[0052] The FMECA data acquisition module obtains the FMECA table from the human-machine interface and extracts fault diagnosis knowledge data from the FMECA table, including the structure data of the hierarchical relationship between the aircraft's systems, subsystems or sub-systems, product components or parts; the expert system's diagnostic object knowledge; the failure rate of each diagnostic object (system or subsystem, component), the failure criteria (performance parameters and parameter limit range), whether it belongs to the minimum equipment list; the failure mode of each diagnostic object, the average interval time between failure modes, the cause of failure, the impact of failure (mode), the failure (mode) detection method, the failure mode characteristic value (detection parameter or failure symptom set), the severity of the failure mode, the failure mode frequency ratio, the maintenance measures, the mission stage compensation measures, etc. These data are stored in the knowledge base in a structured form.

[0053] Extract knowledge data from the FMECA table, generally extract FMECA data at the LRU level and above.

[0054] Preferably, the fault cases in the fault case knowledge acquisition module are faults that have occurred in the use history of aircraft system equipment. The fault case knowledge acquisition module acquires fault case knowledge data from the human-machine interface, analyzes and classifies and stores them to form a fault case knowledge base; the fault case knowledge base includes the name of the fault, fault phenomenon description, fault code, system to which the fault belongs, fault discovery time, fault cause, fault characteristic value (symptom set), fault component, fault mode, fault handling measures and troubleshooting effect, etc.

[0055] Preferably, when the knowledge base performs knowledge diagnosis, it first obtains the fault case knowledge in the fault case knowledge acquisition module, and applies the fault case knowledge to case matching. If the best matching result cannot be obtained, the IETM data in the IETM data acquisition module is obtained; the IETM data is used to retrieve and isolate the fault. If the fault is successfully eliminated, the fault diagnosis process and results are organized into a fault case form and added to the fault case knowledge base again.

[0056] If the diagnosis result still cannot be obtained, the fault diagnosis knowledge is obtained from the FMECA data acquisition module, and the fault tree knowledge is obtained from the fault tree data acquisition module; the fault diagnosis knowledge and fault tree knowledge are used to perform intelligent fault diagnosis (the structure of fault mode knowledge and fault tree knowledge shows that the expression of these two kinds of knowledge is very similar. The fault mode data of an object can be regarded as a simplified fault tree with only one layer of sub-nodes, or the fault mode data of these objects can be connected into a large fault tree through the hierarchical relationship of systems, subsystems or sub-subsystems, components, and parts). If the fault is successfully eliminated through the diagnosis result, the diagnosis process and results of this time are organized into a fault case form and added to the fault case knowledge base.

[0057] Therefore, the knowledge base of the aircraft comprehensive intelligent fault diagnosis method is a multi-type shared knowledge base.

[0058] With the application of integrated intelligent fault diagnosis in aircraft, the knowledge of fault case types will become more and more comprehensive, the matching success rate of fault cases will become higher and higher, and the speed of fault diagnosis will become faster and faster.

[0059] The inference engine is the brain of the aircraft's comprehensive intelligent fault diagnosis. The method by which the inference engine performs fault diagnosis is as follows: first, it receives the fault symptoms provided by the user, then obtains the case-based expert knowledge stored in the fault case knowledge base, selects the required inference strategy to perform fault diagnosis based on the case matching method, and determines whether it can give a fault mode and recommend maintenance measures. If so, it outputs the diagnosis results.

[0060] If the fault mode and recommended maintenance measures cannot be given, the fault diagnosis is performed based on IETM, and it is determined whether fault isolation measures or maintenance measures are obtained. If so, the diagnosis results are output;

[0061] If the fault isolation measures or maintenance measures cannot be obtained through IETM, the intelligent fault diagnosis based on the computer system is used to determine whether the fault cause is isolated and maintenance measures are given. If so, the diagnosis result is given; if not, the knowledge data is supplemented and adjusted from the knowledge base and the judgment is repeated;

[0062] After the diagnosis results are given, the diagnosis results and maintenance measures suggestions are sent to the user through the database and human-computer interface.

[0063] Preferably, before each diagnosis is completed and the diagnosis result is given, a judgment is made as to whether the maintenance measures are effective. If effective, the diagnosis result is output; if not effective, the next fault diagnosis step is still performed.

[0064] Preferably, the reasoning strategies selected by the inference engine include: rule reasoning, forward reasoning, reverse reasoning and forward and reverse mixed reasoning. Rule reasoning is to use the knowledge in the knowledge base as rules for reasoning. Forward reasoning is to match the premise to the conclusion, while reverse reasoning first assumes that a conclusion is established to see if its conditions are met. The reasoning engine is like the way experts think about solving problems, and the knowledge base realizes its value through the reasoning engine. The comprehensive ladder reasoning engine of the aircraft comprehensive intelligent fault diagnosis method mainly adopts rule reasoning and forward reasoning.

[0065] The input of the fault diagnosis method is the abnormal phenomenon of aircraft failure, that is, the fault symptom, and the output is a set of fault modes (with fault mode levels) and a set of maintenance tasks (with maintenance task levels).

[0066] Preferably, when fault diagnosis is performed based on the case matching method, only rule reasoning is applied for reasoning. If the input features are successfully matched with the rules (ie, fault case knowledge), a unique solution (fault diagnosis result: fault cause and repair measures) is obtained. Figure 4 It is the working process of fault diagnosis based on case matching method.

[0067] Specifically: obtain data such as fault phenomenon, symptom data, and the system to which the fault belongs, match the acquired data information with the fault case knowledge according to the rules, and determine whether the match is successful. If so, output the diagnosis result; if not, first prompt the user to perform detection operations or start on-board self-detection based on the matched approximate case, supplement the data symptom data, and perform rule matching again; if the match is still unsuccessful, obtain the IETM data in the IETM data acquisition module and switch to IETM diagnosis.

[0068] Before fault diagnosis based on case matching method, all fault case knowledge in the fault case knowledge base is obtained first, and then the feature vector set of all fault case knowledge is determined. When matching, the fault phenomenon needs to be decomposed and entered according to the category of feature vectors and matched; then the fault case knowledge is classified and stored according to the feature vector set; at the same time, each fault case knowledge is described with quantitative data or enumerated qualitative data. This can achieve fast and accurate fault case knowledge matching.

[0069] After each fault diagnosis is completed, the fault diagnosis result data is saved to the fault case knowledge base, and the fault case knowledge base is continuously updated and accumulated. The success rate and credibility of fault diagnosis must gradually increase with the accumulation of knowledge. Therefore, case matching can be used as the first step of fault diagnosis for maintenance personnel to perform rapid diagnosis and switch to other diagnostic methods when matching is unsuccessful.

[0070] Preferably, the interactive electronic technical manual (IETM) occupies a very important position in the aircraft autonomous support information system. It is an important data basis to support the effective operation of the autonomous support information system. At the same time, it provides data information such as troubleshooting logic, maintenance guidance logic and programming language logic for the aircraft integrated intelligent fault diagnosis system, and provides knowledge for fault analysis and support decision-making.

[0071] IETM is divided into five levels, such as Figure 5 According to the positioning requirements of IETM in the autonomous support information system, the IETM delivered to the user must be an integrated IETM database, which must be able to be loaded into the user's support database (MLSDB) and managed by the database management system, and can be integrated with the support decision-making system.

[0072] The specific method of fault diagnosis based on IETM is to obtain data such as fault phenomenon, symptom data, and the system to which the fault belongs, and determine the type of fault; after obtaining the fault type, the reasoning engine of the IETM reader is combined to guide maintenance personnel to complete the fault diagnosis. If the fault type cannot be obtained, it will be transferred to intelligent fault diagnosis.

[0073] The technical data that the aircraft will be delivered to the user in the future is completely digital, interactive electronic data, which meets the requirements of Level 4 IETM and has some functions of Level 5 IETM. According to the S1000D standard, IETM mainly provides diagnostic isolation and troubleshooting guidance for four types of faults, namely: isolated faults, detection faults, observation faults and related faults.

[0074] For isolated faults, there are fault codes, fault phenomenon descriptions, and the fault has been isolated and located to a specific LRU (Least Recently Used), and maintenance measures for the LRU can be given. This type of fault is mainly the fault that can be detected and isolated by the onboard PMA. For this type of fault, the user enters the fault code or fault phenomenon description information transmitted by the onboard data link or receives it through the air-to-ground data link and enters it into the IETM operating platform for fault retrieval to obtain the corresponding troubleshooting and maintenance measures. The diagnostic results are then presented to the user through the human-machine interface. After the user confirms that the maintenance measures given for this fault type have been successfully eliminated, the fault information is organized into a fault case form and added to the case class knowledge base.

[0075] Detection-type faults have fault codes and fault phenomenon descriptions, but the faults are not isolated to the LRU. This type of fault is generally detectable by the onboard PMA or other ground detection equipment. In the IETM data, the fault isolation procedure is generally given for detection-type faults, or multiple LRUs that may have faults are given in sequence. If the IETM gives a fault isolation procedure, questions are asked according to the fault isolation steps. After the user answers the questions, the system gives the next step or continues to ask questions based on the answers. Through such interaction, the cause of the fault is found and the maintenance measures are given. If the IETM gives multiple LRUs in sequence, the user is prompted to perform corresponding tests or detection operations, and the LRU is judged to be faulty based on the test results fed back by the user. If it is faulty, the corresponding maintenance measures for the LRU are given. After the user confirms that the fault has been successfully eliminated according to the given maintenance measures, the fault code, fault phenomenon description, and the test results (fault characteristic values) and maintenance measures generated during the IETM isolation diagnosis process are organized into a fault case and added to the case knowledge base.

[0076] Observational faults have fault phenomenon description, fault context information and isolation information. Observational faults are generally given by the ground crew IETM. The fault isolation information may be a fault diagnosis procedure, multiple LRUs to be tested, the cause of the fault diagnosis and repair measures, or a specific fault LRU and its repair measures. For observational faults, if interactive guided isolation is performed through IETM, and the final maintenance measures can successfully eliminate the fault, the fault phenomenon description, various test results (fault feature values) generated during the IETM guided isolation process, and maintenance measures are organized into a fault case and added to the case knowledge base.

[0077] Correlated faults refer to multiple warning messages or BIT detection messages reported by the onboard BIT. These warning messages and BIT messages have their own fault codes, but these fault codes all correspond to the same fault isolation procedure or all correspond to the same multiple LRUs to be detected. For such faults, IETM is used for guided isolation. If the final maintenance measures can successfully prevent these warning messages and BIT detection messages from reappearing, these warning messages, BIT detection information and their corresponding fault codes, and various test results (fault feature values) generated during the IETM guided isolation process, maintenance measures and other information are organized into fault case forms and added to the case class knowledge base.

[0078] Preferably, before the intelligent fault diagnosis based on the computer system, a knowledge base of fault diagnosis is first constructed, namely, a maintenance integrated model (MIM). The knowledge base is a fault tree type database, and its data source is mainly the relevant data of FMECA (failure mode, effects and criticality analysis) and FTA (fault tree analysis) of the aircraft design department. It can also be gradually supplemented and improved according to the data accumulated by the troops. According to this knowledge base, auxiliary fault location (determine the cause of the fault) and auxiliary troubleshooting (determine the troubleshooting method) are carried out; secondly, according to the status data collected by MIM, a set of all possible fault modes causing abnormalities can be given, and the level of each fault mode (the probability of the fault mode causing the abnormality) can be calculated; finally, the maintenance integrated model continuously obtains different types of fault case data to obtain the efficiency (success rate) of each maintenance method in repairing a fault mode.

[0079] The specific steps for calculating the level of the fault mode are as follows: obtain abnormal data such as fault phenomenon, symptom data, and the system to which the fault belongs; obtain the fault mode set from the MIM based on the input abnormal data; and then calculate the level of each fault mode based on the PHM status data; then obtain the corresponding maintenance task from the MIM based on the level of the fault mode, and then use supplementary learning technology to calculate the level of the maintenance task; finally, enter the autonomous support information system to determine the best maintenance plan.

[0080] For the cases where the fault mode cannot be accurately inferred or the fault tree fault cause can only provide a set of fault modes, if the knowledge base has the false alarm rate (FAR) and true failure rate (RFP) of the corresponding symptom parameters, the system will calculate the level according to the fault mode and adjust the result according to the failure rate.

[0081] In summary, the present application has the following advantages: it can quickly locate, detect, and troubleshoot the entire aircraft system, so that the aircraft can be quickly repaired after landing, thereby quickly and efficiently connecting the aircraft's diagnosis, maintenance, and logistics support to form an intelligent, comprehensive, and integrated comprehensive support, ultimately achieving the effect of reducing the time required for aircraft fault location analysis, shortening the aircraft troubleshooting cycle, and reducing labor costs.

[0082] Finally, it should be noted that: the drawings of the embodiments disclosed in the present invention only involve structures related to the embodiments disclosed in the present invention, and other structures can refer to the general design. In the absence of conflict, the same embodiment and different embodiments of the present invention can be combined with each other;

[0083] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. An aircraft comprehensive intelligent fault diagnosis method, characterized in that: It includes a human-machine interface, a knowledge base, a database and an inference engine; the input ends of the human-machine interface and the knowledge base are connected to the diagnosed object, the output end of the human-machine interface is connected to the database, the database and the knowledge base are interconnected with the inference engine, and the knowledge base stores knowledge of various fault types; The database stores aircraft operating data and fault type data. The inference engine can obtain data from the knowledge base and database to diagnose the current state of the diagnosed object and give a diagnostic result. A knowledge acquisition unit is also provided between the human-machine interface and the knowledge base. The knowledge acquisition unit can capture various fault data transmitted by the human-machine interface, organize them, and obtain different categories of knowledge, which are then transmitted to the knowledge base.

2. The aircraft comprehensive intelligent fault diagnosis method according to claim 1, characterized in that: The knowledge stored in the knowledge base includes: fault case knowledge, FMECA knowledge, FTA knowledge and IETM knowledge.

3. The aircraft comprehensive intelligent fault diagnosis method as claimed in claim 2, characterized in that: The knowledge acquisition unit includes IETM data acquisition module, fault tree data acquisition module, FMECA data acquisition module and fault case knowledge acquisition module; the IETM data acquisition module can acquire knowledge data from IETM data; the fault tree data acquisition module can extract knowledge data from the fault tree model; the FMECA data acquisition module can extract knowledge data from the FMECA table; the fault case knowledge acquisition module can acquire all fault cases in the aircraft system equipment.

4. The aircraft comprehensive intelligent fault diagnosis method as claimed in claim 3, characterized in that: An IETM data storage module is provided in front of the human-machine interface. The IETM data storage module is loaded with all IETM data that conform to the standards and is capable of parsing and operating the IETM data. The specific method of parsing and operating the IETM data is to embed the S1000D standard IETM operating platform. When the IETM data storage module is operated, records are made through the IETM operating platform and feedback is given to the IETM maintainer according to the situation.

5. The aircraft comprehensive intelligent fault diagnosis method as claimed in claim 4, characterized in that: After receiving the fault diagnosis request from the inference engine, the IETM data acquisition module retrieves the corresponding data from the IETM data storage module through the human-machine interface; when the IETM data is updated, the IETM data storage module can synchronously send the updated data to the IETM data acquisition module.

6. The aircraft comprehensive intelligent fault diagnosis method as claimed in claim 3, characterized in that: The fault tree data acquisition module imports the fault tree analysis data through the human-machine interface, and converts the fault tree analysis data into structured data to form and store fault tree knowledge.

7. The aircraft comprehensive intelligent fault diagnosis method according to claim 6, characterized in that: Fault tree analysis data includes analysis object, fault tree model, minimum cut set of each node of the fault tree, failure probability of each minimum cut set, failure probability of each node of the fault tree, importance, critical importance, diagnostic importance, as well as fault phenomenon description, fault symptom set and maintenance measures of each level of fault tree nodes.

8. The aircraft comprehensive intelligent fault diagnosis method as claimed in claim 3, characterized in that: The FMECA data acquisition module obtains the FMECA table from the human-machine interface and extracts the fault diagnosis knowledge data from the FMECA table, including generating the structural data of the hierarchical relationship of each system, subsystem or sub-system, product component or part of the aircraft; Expert system's knowledge of diagnostic objects; failure rate of each diagnostic object, fault criteria, and whether it belongs to the minimum equipment list; Failure mode of each diagnosis object, mean time between failure modes, cause of failure, impact of failure, fault detection method, characteristic value of failure mode, severity of failure mode, frequency ratio of failure mode, maintenance measures, and compensation measures in the mission stage.

9. The aircraft comprehensive intelligent fault diagnosis method as claimed in claim 3, characterized in that: The fault case knowledge acquisition module obtains fault case knowledge data from the human-machine interface, analyzes and classifies and stores it to form a fault case knowledge base; the fault case knowledge base includes the name of the fault, fault phenomenon description, fault code, system to which the fault belongs, the time of fault discovery, the cause of the fault, fault characteristic value, fault component, fault mode, fault handling measures and troubleshooting effect.

10. The aircraft comprehensive intelligent fault diagnosis method according to claim 3, characterized in that: When the knowledge base performs knowledge diagnosis, it first obtains the fault case knowledge in the fault case knowledge acquisition module, and applies the fault case knowledge to case matching. If the best matching result cannot be obtained, the IETM data in the IETM data acquisition module is obtained; the IETM data is used to retrieve and isolate the fault. If the fault is successfully eliminated, the fault diagnosis process and results are organized into a fault case form and added to the fault case knowledge base again.

11. The aircraft comprehensive intelligent fault diagnosis method according to claim 10, characterized in that: If the diagnosis result still cannot be obtained, the fault diagnosis knowledge is obtained from the FMECA data acquisition module and the fault tree knowledge is obtained from the fault tree data acquisition module; Intelligent fault diagnosis is performed using fault diagnosis knowledge and fault tree knowledge. If the fault is successfully eliminated through the diagnosis results, the diagnosis process and results are organized into a fault case form and added to the fault case knowledge base.

12. The aircraft comprehensive intelligent fault diagnosis method according to claim 3, characterized in that: The method of fault diagnosis by the inference engine is: first receiving the fault symptoms provided by the user, then obtaining the case-based expert knowledge stored in the fault case knowledge base, selecting the required reasoning strategy to perform fault diagnosis based on the case matching method, judging whether the fault mode and recommended maintenance measures can be given, and if so, outputting the diagnosis results; If the fault mode and recommended maintenance measures cannot be given, the fault diagnosis is performed based on IETM, and it is determined whether fault isolation measures or maintenance measures are obtained. If so, the diagnosis results are output; If the fault isolation measures or maintenance measures cannot be obtained through IETM, the intelligent fault diagnosis based on the computer system is used to determine whether the fault cause is isolated and maintenance measures are given. If so, the diagnosis result is given; if not, the knowledge data is supplemented and adjusted from the knowledge base and the judgment is repeated; After the diagnosis results are given, the diagnosis results and maintenance measures suggestions are sent to the user through the database and human-computer interface.

13. The aircraft comprehensive intelligent fault diagnosis method according to claim 12, characterized in that: Before each diagnosis is completed and the diagnostic results are given, a judgment is made as to whether the maintenance measures are effective. If they are effective, the diagnostic results are output; if not, the next fault diagnosis step is still performed.

14. The aircraft comprehensive intelligent fault diagnosis method according to claim 12, characterized in that: The reasoning strategies selected by the inference engine include: rule reasoning, forward reasoning, backward reasoning and forward and backward mixed reasoning.

15. The aircraft comprehensive intelligent fault diagnosis method according to claim 12, characterized in that: The method for fault diagnosis based on case matching is: obtain the fault phenomenon, symptom data, and the system to which the fault belongs, match the obtained data information with the fault case knowledge according to the rules, and determine whether the match is successful. If so, output the diagnosis result; if not, first prompt the user to perform detection operations or start on-board self-detection based on the matched approximate case, supplement the data symptom data, and perform rule matching again; if the match is still unsuccessful, obtain the IETM data in the IETM data acquisition module and switch to IETM diagnosis.

16. The aircraft comprehensive intelligent fault diagnosis method according to claim 15, characterized in that: Before performing fault diagnosis based on the case matching method, all fault case knowledge in the fault case knowledge base is first obtained, and then the feature vector set of all fault case knowledge is determined. When matching, the fault phenomenon needs to be decomposed and entered according to the category of the feature vector and matched; then the fault case knowledge is classified and stored according to the feature vector set; at the same time, each fault case knowledge is described with quantitative data or enumerated qualitative data.

17. The aircraft comprehensive intelligent fault diagnosis method according to claim 16, characterized in that: After each fault diagnosis is completed, the fault diagnosis result data is saved in the fault case knowledge base.

18. The aircraft comprehensive intelligent fault diagnosis method according to claim 12, characterized in that: The specific method of IETM-based fault diagnosis is: obtain the fault phenomenon, symptom data, the system to which the fault belongs, and determine the fault type; after obtaining the fault type, combine the reasoning engine of the IETM reader to guide maintenance personnel to complete the fault diagnosis; if the fault type cannot be obtained, switch to intelligent fault diagnosis.

19. The aircraft comprehensive intelligent fault diagnosis method according to claim 18, characterized in that: The fault types include isolated faults, detection faults, observation faults and related faults.

20. The aircraft comprehensive intelligent fault diagnosis method according to claim 19, characterized in that: Isolated faults are retrieved by entering the fault code or fault phenomenon description information transmitted through the data link on the air-ground data link receiver into the IETM operating platform to obtain corresponding troubleshooting and maintenance measures; the diagnosis results are then presented to the user through the human-machine interface. After the user confirms that the maintenance measures given for the fault type have been successfully eliminated, the fault information is organized into a fault case and added to the case-based knowledge base.

21. The aircraft comprehensive intelligent fault diagnosis method according to claim 19, characterized in that: For detection-type faults, a fault isolation procedure is given and questions are asked according to the fault isolation steps. After the user answers the questions, the system gives the next step of action or continues to ask questions based on the answer. Through such interaction, the cause of the fault is found and maintenance measures are given. If IETM gives multiple LRUs in sequence, the user is prompted to perform corresponding tests or detection operations, and judge whether the LRU is faulty based on the test results fed back by the user. If it is faulty, the corresponding maintenance measures are given for the LRU. After the user confirms that the fault has been successfully eliminated according to the given maintenance measures, the fault code, fault phenomenon description, and test results and maintenance measures generated during the IETM isolation diagnosis process are added to the case-type knowledge base in the form of fault cases.

22. The aircraft comprehensive intelligent fault diagnosis method according to claim 19, characterized in that: Observation-type faults are interactively guided and isolated through IETM. If the final maintenance measures can successfully eliminate the fault, the fault phenomenon description, various test results generated during the IETM-guided isolation process, and maintenance measures are added to the case-type knowledge base in the form of fault cases.

23. The aircraft comprehensive intelligent fault diagnosis method according to claim 19, characterized in that: Related faults are guided and isolated using IETM. If the final maintenance measures can successfully prevent these warning messages and BIT detection information from appearing again, these warning messages, BIT detection information and their corresponding fault codes, as well as various test results and maintenance measures generated during the IETM-guided isolation process, are organized into fault case forms and added to the case-based knowledge base.

24. The aircraft comprehensive intelligent fault diagnosis method according to claim 19, characterized in that: Before intelligent fault diagnosis based on computer systems, a knowledge base for fault diagnosis is first built based on the relevant data of FMECA and FTA from the aircraft design department and the data accumulated by the troops; based on the status data collected by MIM, a set of all possible fault modes that cause abnormalities can be given, and the level of each fault mode can be calculated; finally, the maintenance aggregation model continuously obtains different types of fault case data to obtain the efficiency of each maintenance method in repairing a fault mode.

25. The aircraft comprehensive intelligent fault diagnosis method according to claim 24, characterized in that: The steps for calculating the level of the fault mode are as follows: obtain the abnormal data, obtain the fault mode set from the MIM according to the input abnormal data, and then calculate the level of each fault mode according to the PHM status data; then obtain the corresponding maintenance task from the MIM according to the level of the fault mode, and then use the supplementary learning technology to calculate the level of the maintenance task; finally, enter the autonomous support information system to determine the best maintenance plan.

Images