Fault Identification Method for Information Processing System

Abstract

The present invention provides a method for identifying faults in an information processing system, which is applicable to a dual-redundancy electromechanical management system. The method detects various faults in the electromechanical information processing system through power-on self-test and real-time monitoring during operation. The method adopts a fault reset accumulation method. When a module reports faults three or more times in a row, the method uses a machine learning method to determine the fault type, output a switching signal, and switch to a standby electromechanical information processing unit to take over its work and complete the interactive work. At the same time, the fault information is sent to a recording device to facilitate subsequent analysis and troubleshooting during ground maintenance. The present invention greatly ensures the safety, reliability, and stability of the aircraft during flight. The method uses a machine learning method to predict faults and improves identification accuracy. At the same time, the recording of fault classification facilitates subsequent reproduction, troubleshooting, and other work by on-site personnel.

Description

Technical Field

[0001] The present invention belongs to the technical field of avionics, and in particular relates to a method for identifying faults in an information processing system. Background Art

[0002] During flight, each onboard device generates a large amount of data and parameters. This data must be integrated, processed, and output to the integrated display, flight parameters, and other devices for interaction through the electromechanical information processing system. The electromechanical information processing system plays a crucial role in the aircraft's integrated management, automatic control, fault detection, and data exchange, and is essential for the safe operation of the entire aircraft.

[0003] Currently, many new aircraft are equipped with two electromechanical information processing units (EMPUs), known as dual-redundant systems, to ensure safe and stable operation. However, due to the numerous uncertainties in the airborne environment, the EMPs can experience various failures, leading to abnormalities in data processing and interaction systems, and potentially unsafe incidents. Therefore, a method is urgently needed to accurately identify system failures, facilitate timely switching to a backup EMP, and facilitate subsequent maintenance.

[0004] In view of this, the present invention is proposed. Summary of the Invention

[0005] The present invention aims to provide a method for identifying faults in an information processing system, thereby improving the functionality of existing dual-redundancy systems and enabling the identification of specific faults in the system. The technical solution of this case has numerous beneficial effects, as described below:

[0006] Provided is an information processing system fault identification method applicable to a dual-redundant electromechanical management system, wherein the dual-redundant electromechanical management system includes a primary first electromechanical information processor and a backup second electromechanical information processor, wherein the first electromechanical information processor includes a first self-test module, a first identification and detection module, a first signal processing module, and a first automatic switching module;

[0007] The second electromechanical information processor includes a second self-test module, a second identification detection module, a second signal processing module and a second automatic switching module;

[0008] The first automatic switching module and the second automatic switching module are in communication connection, and when the first electromechanical information processor fails, the first electromechanical information processor and the second electromechanical information processor perform a primary-backup switch;

[0009] The first self-test module and the second self-test module are used to test the internal functions and data interfaces of the corresponding electromechanical information processor after power is turned on, and output self-test signals to the first identification and detection module and the second identification and detection module respectively;

[0010] The first signal processing module and the second signal processing module are used to send and receive signal data, interact with the onboard auxiliary system and the power system after pre-processing, monitor the fault signal in real time, output feedback signals, and send the fault signal to the corresponding first identification and detection module and the first identification and detection module for identification;

[0011] The first identification and detection module and the second identification and detection module communicate with the machine learning model respectively. The machine learning model obtains standard airborne fault parameters and performs training. When the fault signals from the first self-test module and the first signal processing module are received, the specific fault category is determined and output.

[0012] Compared with the prior art, the technical solution provided by the present invention has the following beneficial effects:

[0013] This invention uses machine learning to train a fault model for the information processor model based on offline processing of large amounts of fault data to determine the system's fault state. This not only saves prediction time but also improves the system's recognition accuracy by leveraging the characteristic states trained from a large amount of data.

[0014] The present invention writes the determined fault type into the recording device, and facilitates the ground analysis and processing work by reading the data in the recorder, and is also beneficial for on-site personnel to carry out subsequent analysis, reproduction, and troubleshooting.

[0015] At the same time, if the fault count accumulates for more than 3 times, it indicates that the electromechanical information processor is faulty, and the system will automatically switch to another electromechanical information processor to ensure the safe and efficient operation of the aircraft. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0017] Figure 1 This is a structural diagram of the information processing system fault identification method based on machine learning;

[0018] Figure 2 This is a flowchart of the information processing system fault identification method based on machine learning in a specific implementation method. DETAILED DESCRIPTION

[0019] The following describes the embodiments of the present invention through specific examples. Those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that, in the absence of conflict, the following embodiments and features in the embodiments can be combined with each other. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0020] It should be noted that various aspects of the embodiments within the scope of the appended claims are described below. It should be apparent that the aspects described herein can be embodied in a wide variety of forms, and any specific structure and / or function described herein is merely illustrative. Based on the present invention, those skilled in the art will appreciate that an aspect described herein can be implemented independently of any other aspect, and two or more of these aspects can be combined in various ways. For example, any number of aspects described herein can be used to implement an apparatus and / or practice a method. In addition, other structures and / or functionalities other than one or more of the aspects described herein can be used to implement this apparatus and / or practice this method.

[0021] It should also be noted that the illustrations provided in the following embodiments are merely schematic illustrations of the basic concept of the present invention. The illustrations only show components related to the present invention and are not drawn according to the number, shape, and size of components in actual implementation. In actual implementation, the type, quantity, and proportion of each component may be changed arbitrarily, and the component layout may also be more complex.

[0022] In addition, in the following description, specific details are provided to facilitate a thorough understanding of the examples. However, those skilled in the art will understand that aspects can be practiced without these specific details. In order to enable those skilled in the art to better understand the present invention, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments. The terms "first" and "second" are used for descriptive purposes only and are not to be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, unless otherwise stated, "multiple" means two or more.

[0023] like Figure 1The machine learning-based information processing system fault identification method shown is applicable to data exchange in an airborne dual-redundant electromechanical management system. The dual-redundant electromechanical management system includes a primary first electromechanical information processor and a backup second electromechanical information processor. The dual-redundant system is used in a 1553 network, which can connect multiple dual-redundant systems, including:

[0024] The first electromechanical information processor includes a first self-test module, a first identification and detection module, a first signal processing module and a first automatic switching module; the second electromechanical information processor includes a second self-test module, a second identification and detection module, a second signal processing module and a second automatic switching module;

[0025] The first automatic switching module is communicatively connected to the second automatic switching module, and when the first electromechanical information processor fails, the first electromechanical information processor and the second electromechanical information processor perform a primary-backup switch;

[0026] The specific operation process is as follows Figure 2 As shown, the first self-test module and the second self-test module are used to test the internal functions and data interfaces of the corresponding electromechanical information processor after power is turned on, and output self-test signals to the first identification detection module and the second identification detection module respectively for fault type determination. The faults of the self-test signal and feedback signal include one or more of power-on self-test faults, data anomalies, data reception faults, data transmission faults, system communication faults, electromechanical information processor power failures, and electromechanical information processor periodic self-test faults.

[0027] The first and second signal processing modules are used to send and receive signal data. After pre-processing, they interact with the onboard auxiliary systems and power system, monitor fault signals in real time, and output feedback signals. These fault signals are then sent to the corresponding first identification and detection modules for identification. Specifically, the signal processing modules are used to receive signal data from the fuel system, environmental control system, hydraulic system, power supply system, nose wheel steering system, landing gear system, braking system, de-icing system, etc. Due to the wide variety of data types in various aircraft systems, these signals can be categorized as analog, AFDX, discrete, 429, and 422 data.

[0028] The first identification and detection module and the second identification and detection module communicate with the machine learning model respectively. The machine learning model obtains standard airborne fault parameters and performs training. When receiving the fault signals from the first self-test module and the first signal processing module, it determines the specific fault category and outputs it.

[0029] The machine learning model processes large amounts of fault data offline, training a fault model tailored to the processor model to determine the system's fault status. This not only reduces prediction time but also improves the system's recognition accuracy by leveraging the characteristic states trained from this massive amount of data. When switching between the primary and backup processors, the specific fault type or category can be determined, facilitating timely troubleshooting.

[0030] Typically, after powering on, the first and second mechatronic information processors perform power-on self-tests using the first and second self-test modules, respectively. To prevent system misjudgments, the system restarts when a fault occurs. If the fault persists after three restarts, it indicates a genuine fault in the information processor system. The fault signal is then fed into the machine learning model for analysis. The self-test module utilizes conventional self-test and latching circuits. The self-test module first tests the information processor's hardware and software to enhance operational safety. If the accumulated fault count exceeds three, indicating a fault in the mechatronic information processor, the system automatically switches to the other mechatronic information processor, ensuring safe and efficient aircraft operation. The fault cause is transmitted to a recorder, and a switch signal is output. If everything is normal, the mechatronic information processing system begins data processing and interaction.

[0031] Furthermore, the machine learning model includes a basic model and a temporary model of the communication connection, wherein:

[0032] The basic model acquires and trains standard airborne fault parameters, extracting features from each fault to build an appropriate model. Specifically, for power-on self-test faults, features extracted include system initialization anomalies, multi-process task initialization anomalies, subcard initialization anomalies, and Ethernet communication establishment anomalies. During signal processing, features extracted include data anomalies, data reception failures, data transmission failures, system communication failures, electromechanical information processor power outages, and electromechanical information processor periodic self-test failures. If the fault signal transmitted by the first self-test module or the first signal processing module is recognized by the basic model, the specific fault type is output. If not, the signal is transmitted to a temporary model for training using machine learning with supervised classification. After a preset period of time, the training results are fed back to the basic model, updating the model database with new fault type features. Simultaneously, the determined fault type and cause are recorded in a recording device. Reading the data from the recorder facilitates ground-based analysis and processing, and also facilitates subsequent analysis, reproduction, and troubleshooting by field personnel.

[0033] By adding machine learning model functionality, when the base model fails to identify the current fault, the system deems it a new fault type, feeds back the signal, and transmits the fault information characteristics to a temporary model for online learning. This paper primarily utilizes the perceptron algorithm from the standard online machine learning algorithm library for supervised learning. This algorithm uses a hyperplane for classification. Each time a new data instance is used, predictions are made, comparisons are made, and updates are made to adjust the position of the hyperplane, ensuring that new faults can be accurately and efficiently identified. Generally, online learning takes 3-5 hours. The trained structure is then fed back to the base model for data updates for subsequent fault identification.

[0034] The basic model uses linear discriminant analysis (LDA) to train a prediction model suitable for electromechanical information processing systems using various fault information features collected offline. This allows real-time online prediction of the type of fault in the electromechanical information processor and allows switching to another backup electromechanical information processor based on the fault content. Linear discriminant analysis is an effective method for prediction based on sample category analysis. The basic principle of LDA is to find the most appropriate projection axis so that the distance between the projections of different samples on this axis is as far as possible, while the projections of samples within each category are as compact as possible, thereby achieving the best classification effect. That is, it maximizes the distance between classes while minimizing the distance within classes.

[0035] Assume that the total number of features of the sample training set and the test sample is In order to find the best projection direction, it is necessary to calculate the mean of various samples , the sample class dispersion matrix and the total inter-class dispersion matrix , the sample inter-class dispersion matrix According to the criteria, make the samples of each type as far away as possible, and try to concentrate the samples of each type to find the best projection vector Finally, we need to find the boundary points of the Y space. After finding the boundary points, we can project the sample to be tested into the Y space, determine the relationship between its projection point and the boundary point, classify it, and determine which type of fault it belongs to.

[0036] As a specific implementation method provided in this case, when the first self-test module or the first signal processing module transmits a fault signal, the first automatic switching module communicates with the second automatic switching module to perform active-standby switching between the first electromechanical information processor and the second electromechanical information processor.

[0037] As a specific implementation method provided in this case, when the first electromechanical information processor does not feedback a fault signal, the second electromechanical information processor does not go online and is used as the RT of the 1553 bus network; when the first electromechanical information processor feedbacks a fault signal, the second electromechanical information processor goes online and enters the working mode, and the first electromechanical information processor enters the standby state.

[0038] The identification and detection module is connected to the automatic switching module. When one mechatronic processor receives a switching signal from another mechatronic processor, it immediately switches between the two processors. After the switching task is executed, the faulty mechatronic processor is switched from the master system to the slave system, and the healthy mechatronic processor is switched from the slave system to the master system for data exchange.

[0039] The above is a detailed introduction to the product provided by the present invention. Specific examples are used herein to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only intended to help understand the core ideas of the present invention. It should be pointed out that, for those skilled in the art, without departing from the principles of the invention, several improvements and modifications can be made to the invention, and these improvements and modifications also fall within the scope of protection of the invention claims.

Claims

1. A method for identifying faults in an information processing system, applicable to a dual-redundant electromechanical management system comprising a primary first electromechanical information processor and a backup second electromechanical information processor, characterized in that: The first electromechanical information processor includes a first self-test module, a first identification and detection module, a first signal processing module and a first automatic switching module; The second electromechanical information processor includes a second self-test module, a second identification detection module, a second signal processing module and a second automatic switching module; The first automatic switching module and the second automatic switching module are in communication connection, and when the first electromechanical information processor fails, the first electromechanical information processor and the second electromechanical information processor perform a primary-backup switch; The first self-test module and the second self-test module are used to test the internal functions and data interfaces of the corresponding electromechanical information processor after power is turned on, and output self-test signals to the first identification and detection module and the second identification and detection module respectively; The first signal processing module and the second signal processing module are used to send and receive signal data, interact with the onboard auxiliary system and the power system after pre-processing, monitor the fault signal in real time, output feedback signals, and send the fault signal to the corresponding first identification and detection module and the first identification and detection module for identification; The first identification and detection module and the second identification and detection module communicate with the machine learning model respectively. The machine learning model obtains standard airborne fault parameters and performs training. When the fault signals from the first self-test module and the first signal processing module are received, the specific fault category is determined and output. The identification and detection module is connected to the automatic switching module. The basic model of the machine learning model adopts linear discriminant analysis and uses various fault information features collected offline to train a prediction model suitable for the electromechanical information processing system. It predicts the faults of the electromechanical information processor in real time online and switches to another spare electromechanical information processor according to the fault content.

2. The information processing system fault identification method according to claim 1, characterized in that: After the first electromechanical information processor and the second electromechanical information processor are powered on, they perform power-on self-tests through the first self-test module and the second self-test module respectively. When a fault occurs, the system is restarted. If the fault is still reported after three restarts, the fault signal is sent to the identification and detection module.

3. The information processing system fault identification method according to claim 1, characterized in that: The machine learning model includes a basic model and a temporary model of the communication connection, wherein: The basic model obtains standard airborne fault parameters and performs training; if the fault signal transmitted by the first self-test module or the first signal processing module is recognized by the basic model, the specific fault type is output; if it is not recognized by the basic model, it is transmitted to the temporary model for machine learning with supervised classification function and training, and the training results are transmitted to the basic model after a preset time period, and the fault type is stored and copied to the ground maintenance equipment in the form of a mobile hard disk.

4. The information processing system fault identification method according to claim 3, characterized in that: The first automatic switching module communicates with the second automatic switching module to perform active / standby switching between the first electromechanical information processor and the second electromechanical information processor.

5. The information processing system fault identification method according to claim 4, characterized in that: The basic model is trained by a linear learning method to identify the fault model.

6. The information processing system fault identification method according to claim 5, characterized in that: When the first electromechanical information processor does not feedback a fault signal, the second electromechanical information processor does not go online and is used as the RT of the 1553 bus network; when the first electromechanical information processor feedbacks a fault signal, the second electromechanical information processor goes online and enters the working mode, and the first electromechanical information processor enters the standby state.

Images