Satellite Fault Diagnosis Method

Abstract

The present invention provides a satellite fault diagnosis method, comprising: performing threshold preprocessing and envelope boundary preprocessing to diagnose the satellite's explicit faults according to the classification of stationary data, periodic data and gradual data; performing machine learning prediction and data-driven diagnosis according to the satellite's explicit faults to diagnose the satellite's implicit faults; performing machine learning prediction includes: using historical data to train a function model and self-learning to continuously approach a real model, and finally using the trained function model to predict future trends; performing data-driven diagnosis includes: constructing a data training library, using a data-driven classification method to train a fault classifier, generating a fault diagnosis rule based on current facts after feature extraction and pattern analysis, and using historical data to test and optimize the rule to further enrich the rule diagnostic device.

Description

Technical Field

[0001] The present invention relates to the technical field of satellite autonomous health management, and in particular to a satellite fault diagnosis method. Background Art

[0002] As a complex, technically extensive and expensive integrated system, satellites will inevitably experience some anomalies or failures during their in-orbit operation due to the uncertainty of the space environment and the limitations of pre-launch testing. For satellites that operate in orbit for a long time, in order to avoid the failure of some key single components causing the entire system to fail, they must be handled quickly and effectively at the beginning of the failure or before it occurs, so as to maintain the basic normal functions of the system and ensure the safe and reliable completion of the mission.

[0003] Limited by the scope of measurement and control of my country's tracking and control stations, medium and low-orbit satellites have their own unmeasurable or uncontrollable arcs. When satellites have abnormalities during this period, the ground cannot detect and handle them in time, which may cause the fault to spread and shorten the service life of the satellite. For spacecraft such as meteorological satellites and navigation satellites, there is also a requirement for service continuity. If the interruption time is too long, it will cause huge military or economic losses. Summary of the invention

[0004] The purpose of the present invention is to provide a satellite fault diagnosis method to solve the problem that the interruption time of an on-orbit satellite due to a fault is too long, thereby causing huge military or economic losses.

[0005] In order to solve the above technical problems, the present invention provides a satellite fault diagnosis method, comprising:

[0006] According to the classification of stationary data, periodic data and gradual data, threshold preprocessing and envelope boundary preprocessing are performed to diagnose the satellite's obvious faults;

[0007] Based on the satellite’s visible faults, perform machine learning prediction and data-driven diagnosis to diagnose the satellite’s hidden faults;

[0008] Executing machine learning predictions includes: using historical data to train the function model and self-learning to continuously approach the real model, and finally using the trained function model to predict future trends;

[0009] Executing data-driven diagnosis includes: building a data training library, using data-driven classification methods to train fault classifiers, generating fault diagnosis rules based on current facts after feature extraction and pattern analysis, and using historical data to test and optimize the rules to further enrich the rule diagnostics.

[0010] Optionally, the satellite fault diagnosis method further includes:

[0011] The data receiving module integrates the data of all satellite subsystems to obtain steady-state data, periodic data and gradual-change data;

[0012] The data receiving module receives key telemetry data and business data of all subsystems on the satellite through a high-speed bus, including business information of the payload subsystem, business information of the satellite service subsystem, measurement and control information of the measurement and control subsystem, control information of the attitude and orbit control subsystem, telemetry information of the thermal control subsystem and telemetry information of the energy subsystem.

[0013] Optionally, the satellite fault diagnosis method further includes:

[0014] The data preprocessing module directly compares the steady data with the threshold to determine the health status of the system;

[0015] The data preprocessing module analyzes the periodic data through the envelope threshold of the periodic data and extracts the health characteristics. If the measured value of the satellite parameter is within the empirical interval, the health condition passes. If the measured value accumulated for a period of time is outside the empirical interval, the health condition fails.

[0016] The data preprocessing module sends the gradual change data to the fault diagnosis module for analysis and prediction.

[0017] Optionally, the satellite fault diagnosis method further includes:

[0018] If the variance of a data is close to 0, it is classified as stationary data, which maintains long-term stability.

[0019] If a certain data has a strong dependence on time and shows periodic changes, it is classified as periodic data;

[0020] If the development trend of a certain data gradually weakens or strengthens with the life of a single machine, it is classified as gradual data.

[0021] Optionally, the satellite fault diagnosis method further includes:

[0022] The fault classification management module classifies the fault diagnosis results;

[0023] The fault recovery module performs various recovery operations according to the fault classification results.

[0024] Optionally, in the satellite fault diagnosis method, the fault classification management module classifies the explicit faults and implicit faults into four levels according to the output result of the fault diagnosis module, wherein:

[0025] The first level is the failure that affects the safety of the entire satellite and cannot be recovered autonomously;

[0026] The second level is a fault that affects the payload mission and cannot be recovered autonomously;

[0027] The third level is the failure that can be recovered by reconstruction;

[0028] The fourth level is a fault whose impact can be recovered by resetting.

[0029] Optionally, in the satellite fault diagnosis method, when the satellite is unable to perform autonomous recovery, autonomous onboard protection measures are implemented for first-level and second-level faults to ensure the safety of the entire satellite; autonomous recovery measures by the satellite are implemented for third-level and fourth-level faults, including resetting, power-off and reconstruction measures, while ensuring the normality of the payload mission.

[0030] Optionally, in the satellite fault diagnosis method, the first level fault includes energy failure, attitude control failure and satellite service failure, and the satellite will start to enter the safety mode to implement self-protection;

[0031] The second level of failure includes time-frequency failure and amplifier failure, and the satellite will adjust the transmission power to implement self-protection;

[0032] The third level of failure involves all subsystems of the entire satellite, and initiates a fault recovery command to the platform computer to switch the faulty single machine to the backup work;

[0033] The fourth level of faults can be recovered through autonomous software compensation, resetting or powering on and off. A fault recovery command is sent to the platform computer or load task processor to autonomously reset or power on the faulty unit.

[0034] In the satellite fault diagnosis method provided by the present invention, by classifying steady data, periodic data and gradual data, and separately diagnosing explicit faults and latent faults, different fault types can be quickly separated according to different data types, thereby avoiding huge military or economic losses caused by long interruptions.

[0035] At present, the health management of domestic satellites is still mainly based on traditional ground monitoring stations, and ground staff make status judgments and handle faults based on the telemetry data transmitted by satellites. However, with the complexity of satellite functions and the long-term operation on orbit, the possibility of on-orbit faults is increasing, and people are paying more and more attention to the autonomous health management of satellite systems.

[0036] At present, the satellite autonomous health management function mainly uses integrated electronics as a platform to carry out fault diagnosis, fault classification and fault recovery research and engineering implementation for the power subsystem and attitude control subsystem, which can ensure the most basic survivability of the spacecraft. Its shortcomings are:

[0037] 1) With the development of aerospace applications, the demand for autonomous health management functions of payloads for spacecraft with mission continuity requirements, such as communication satellites, navigation satellites, and meteorological satellites, is urgent. Existing engineering applications and academic research lack autonomous diagnosis and recovery of payload failures.

[0038] 2) Autonomous health management of key subsystems is realized based on an integrated electronic platform. The main task function overlaps with the fault diagnosis function, which occupies a lot of FPGA / CPU resources, has a complex execution strategy and low reliability.

[0039] In the satellite fault diagnosis method provided by the present invention, a satellite centralized autonomous health management architecture composed of a data receiving module, a data preprocessing module, a fault diagnosis module, a fault classification management module and a fault recovery module is formed to form a dedicated autonomous health management unit, which integrates the data of all subsystems of the satellite, integrates data collection, data preprocessing, health prediction, fault diagnosis, fault classification and fault recovery functions, predicts and diagnoses autonomous on-orbit faults of satellite-borne products, and improves professional processing capabilities and efficiency.

[0040] The present invention analyzes the characteristics of the satellite subsystem services and telemetry data, which have certain commonalities, designs a dedicated autonomous health management unit, and adopts a fault detection algorithm to achieve autonomous fault diagnosis and recovery of all subsystem tasks. At the same time, it does not affect the main tasks of the integrated electronics or payload core processors, reduces the operational complexity of ground control personnel, and ensures that the system switches stably after autonomous recovery from an abnormality, can work normally, and complete specific tasks.

[0041] Satellite autonomous health management includes health status detection, diagnosis, prediction and recovery. First, by adding autonomous health management technology, it can make full use of the redundant structure and analytical redundancy characteristics between subsystems. After a satellite fails, it can promptly detect, isolate and restore the faulty equipment, extend the satellite's service life, and ensure that the satellite provides normal services and functions; secondly, the autonomous health management system can also provide on-orbit health status monitoring for satellites, predict and avoid failures as much as possible; finally, the autonomous health management system can provide detection data for the design of high-reliability and long-life satellites, detect and analyze the causes of failures, and provide a basis for the design of new satellites in the future. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 is a schematic diagram of a satellite fault diagnosis method according to an embodiment of the present invention;

[0043] Figure 2 is a schematic diagram of a data-driven fault diagnosis process according to an embodiment of the present invention;

[0044] As shown in the figure: 10 - data receiving module; 20 - data pre-processing module; 30 - fault diagnosis module; 40 - fault classification management module; 50 - fault recovery module. DETAILED DESCRIPTION

[0045] The satellite fault diagnosis method proposed by the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. The advantages and features of the present invention will become more apparent from the following description and claims. It should be noted that the accompanying drawings are all in a very simplified form and are not in exact proportions, and are only used to conveniently and clearly assist in explaining the purpose of the embodiments of the present invention.

[0046] In addition, unless otherwise specified, the features in different embodiments of the present invention may be combined with each other. For example, a feature in the second embodiment may be used to replace a corresponding feature in the first embodiment or a feature with the same or similar function, and the resulting embodiment also falls within the disclosure or description scope of the present application.

[0047] The core idea of ​​the present invention is to provide a satellite fault diagnosis method to solve the problem that the on-orbit satellite is interrupted for too long due to a fault, thereby causing huge military or economic losses.

[0048] To realize the above idea, the present invention provides a satellite fault diagnosis method, comprising: a data receiving module, configured to collect subsystem data; a data preprocessing module, configured to preprocess the subsystem data; a fault diagnosis module, configured to perform health prediction and fault diagnosis according to the preprocessing result of the subsystem data to form a fault diagnosis result; a fault classification management module, configured to classify the fault diagnosis result; and a fault recovery module, configured to perform a variety of recovery operations according to the fault classification result.

[0049] The present invention overcomes the shortcomings of the prior art and proposes a centralized autonomous health management framework for satellites. Through a dedicated autonomous health management unit, the telemetry data and business data of each subsystem are integrated to form a fault knowledge base. Through threshold monitoring, data-driven, machine learning and other algorithms, satellite explicit and implicit faults are identified. According to the fault level classification, different levels of fault recovery strategies are implemented to achieve autonomous fault diagnosis and recovery of platforms, payloads, intersatellite missions and working units. After recovery, the normal execution of the mission can still be guaranteed, reducing the operational complexity of ground control personnel and improving the autonomous survivability of spacecraft.

[0050] The satellite centralized autonomous health management architecture includes a dedicated autonomous health management unit, which integrates data from all satellite subsystems and integrates functions such as data collection, data preprocessing, health prediction, fault diagnosis, fault classification and fault recovery. It predicts and diagnoses autonomous on-orbit faults of satellite-borne products and improves professional processing capabilities and efficiency.

[0051] The system includes a data receiving module 10 , a data preprocessing module 20 , a fault diagnosis module 30 , a fault classification management module 40 and a fault recovery module 50 .

[0052] The data receiving module 10 receives key telemetry data and business data of all subsystems on the satellite through a high-speed bus, including business information of the payload subsystem 60, business information of the satellite service subsystem 61, measurement and control information of the measurement and control subsystem 65, control information of the attitude and orbit control subsystem 62, and telemetry information of the thermal control subsystem 63 and the energy subsystem 64. The data are divided into three categories: steady type, periodic type and gradual type to ensure the validity and completeness of the information after collection.

[0053] The data preprocessing module 20 sequentially reads data from the data receiving module 10, and diagnoses the satellite's dominant faults by using preprocessing techniques such as thresholds, envelope boundaries, etc. according to the three classified data.

[0054] The fault diagnosis module 30 is equipped with a high-speed processing unit and a transmission bus. Based on the output results of the data preprocessing module 20, it integrates and applies machine learning and data-driven diagnosis methods, gradually adopts highly adaptable diagnosis methods for different types of telemetry data, and deeply analyzes and predicts deep-seated faults that are difficult to judge by intuitive means.

[0055] The fault classification management module 40 and the fault recovery module 50 first classify the faults according to the results output by the fault diagnosis module 30, and classify the faults into four levels according to the degree of impact, namely, affecting the safety of the entire satellite and being unable to recover autonomously, affecting the payload mission and being unable to recover autonomously, being able to recover through reconstruction, and being able to recover through resetting. The fault levels are: A-1 (first level fault), A-2 (second level fault), B (third level fault), and C (fourth level fault). The A-1 fault is the most serious, followed by the A-2 fault. When autonomous recovery is not possible, autonomous protection measures are implemented on board for these two levels of faults to ensure the safety of the entire satellite. The B and C faults are recovered autonomously by the satellite, including taking measures such as resetting, power off, and reconstruction, while ensuring the normality of the payload mission.

[0056] The satellite centralized autonomous health management architecture is implemented by a dedicated autonomous health management unit, which includes a data receiving module, a data preprocessing module, a fault diagnosis module, and a fault classification and recovery module.

[0057] The data receiving module receives key telemetry data and business data of all subsystems on the satellite through a high-speed bus, including business information of the payload subsystem, measurement and control information of the measurement and control subsystem, control information of the attitude and orbit control subsystem, and telemetry information of the thermal control and energy subsystems. These data mainly present the following three characteristics:

[0058] (1) Stable type: data that maintains long-term stability, such as analog voltage telemetry information of a single device. If the variance of the data is close to 0, it is classified as stable data;

[0059] (2) Periodic: the data is strongly dependent on time and shows periodic changes, such as the temperature information of a single device;

[0060] (3) Gradual change type: the data development trend gradually weakens or strengthens with the life of a single machine, such as the light intensity information of a satellite-borne atomic clock.

[0061] The data receiving module first divides these large amounts of data into three categories to ensure the validity and completeness of the collected information.

[0062] The data preprocessing module reads data from the data receiving unit. For steady data, the health status of the system is directly judged through the threshold, such as power supply voltage, current and temperature. For periodic data, the health characteristics are extracted after analyzing the data through the envelope threshold. The envelope threshold is an upper and lower boundary range, which is an empirical interval obtained after the satellite successfully completes the ground test. If the measured value of the satellite parameter is within this interval, the health status passes and there is a low failure rate. On the contrary, if the measured value accumulated for a period of time is not within this interval, it can be predicted that the health status is poor and the failure rate is high. For gradual data, it cannot be directly used to judge the health status of the system and needs to be sent to the fault diagnosis unit for in-depth analysis and prediction.

[0063] Based on the output results of the preprocessing module, the fault diagnosis module integrates and applies machine learning and data-driven diagnostic methods, gradually adopts highly adaptable diagnostic methods for different types of telemetry data, and deeply analyzes and predicts deep-seated faults that are difficult to judge by intuitive means.

[0064] The working process of prediction technology based on machine learning is: use historical data to train the function model and self-learn, continuously approach the real model, and finally use the trained function model to predict future trends.

[0065] The data-driven diagnosis method is to build a data training library, use the data-driven classification method to train the fault classifier, generate fault diagnosis rules based on current facts after feature extraction and pattern analysis, and use historical data to test and optimize the rules to further enrich the rule diagnostics. The diagnosis process is as follows: Figure 2 shown.

[0066] The fault classification and recovery module classifies the faults into four levels: A-1, A-2, B, and C according to the degree of impact, based on the results output by the fault diagnosis module, with A-1 being the most serious.

[0067] Level A-1 refers to failures that affect the safety of the entire satellite and cannot be recovered autonomously, including energy failures, attitude control failures, and satellite service failures. The satellite will start to enter safety mode and implement self-protection.

[0068] Level A-2 refers to failures that affect the payload mission and cannot be recovered autonomously, including timing and frequency failures, amplifier failures, etc. The satellite will adjust the transmission power and implement self-protection.

[0069] Class B refers to failures that can only be restored by autonomously switching to cold / hot standby (isolation), involving all subsystems of the entire satellite. The autonomous health management unit initiates a fault recovery command to the platform computer to switch the faulty single machine to backup work.

[0070] Class C faults are those that can be recovered through software self-compensation, reset or power on / off, involving all subsystems of the entire satellite. The autonomous health management unit sends a fault recovery command to the platform computer or payload task processor to self-reset or power on the faulty single machine.

[0071] In summary, the above embodiments describe in detail different configurations of the satellite fault diagnosis method. Of course, the present invention includes but is not limited to the configurations listed in the above embodiments. Any content that is transformed based on the configurations provided in the above embodiments belongs to the scope of protection of the present invention. Those skilled in the art can draw inferences based on the contents of the above embodiments.

[0072] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the system disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0073] The above description is only a description of the preferred embodiments of the present invention, and is not intended to limit the scope of the present invention. Any changes or modifications made by a person skilled in the art in the field of the present invention based on the above disclosure shall fall within the scope of protection of the claims.

Claims

1. A satellite fault diagnosis method, characterized in that: include: The data preprocessing module performs threshold preprocessing and envelope boundary preprocessing according to the classification of stationary data, periodic data and gradual data to diagnose the dominant fault of the satellite; The high-speed computing unit of the fault diagnosis module performs machine learning prediction and data-driven diagnosis based on the satellite's explicit faults to diagnose the satellite's implicit faults; Executing machine learning predictions includes: using historical data to train the function model and self-learning to continuously approach the real model, and finally using the trained function model to predict future trends; Executing data-driven diagnosis includes: building a data training library, using data-driven classification methods to train fault classifiers, generating fault diagnosis rules based on current facts after feature extraction and pattern analysis, and using historical data to test and optimize the rules to further enrich the rule diagnostics.

2. The satellite fault diagnosis method according to claim 1, characterized in that: Also includes: The data receiving module integrates the data of all satellite subsystems to obtain steady data, periodic data and gradual data; The data receiving module receives key telemetry data and business data of all subsystems on the satellite through a high-speed bus, including business information of the payload subsystem, business information of the satellite service subsystem, measurement and control information of the measurement and control subsystem, control information of the attitude and orbit control subsystem, telemetry information of the thermal control subsystem and telemetry information of the energy subsystem.

3. The satellite fault diagnosis method according to claim 1, characterized in that: Also includes: The data preprocessing module directly compares the steady data with the threshold to determine the health status of the system; The data preprocessing module analyzes the periodic data through the envelope threshold of the periodic data to extract health features. If the measured value of the satellite parameter is within the empirical interval, the health condition passes. If the measured value accumulated over a period of time is outside the empirical interval, the health condition fails. The data preprocessing module sends the gradual change data to the fault diagnosis module for analysis and prediction.

4. The satellite fault diagnosis method according to claim 1, characterized in that: Also includes: If the variance of a data is close to 0, it is classified as stationary data, which maintains long-term stability. If a certain data has a strong dependence on time and shows periodic changes, it is classified as periodic data; If the development trend of a certain data gradually weakens or strengthens with the life of a single machine, it is classified as gradual data.

5. The satellite fault diagnosis method according to claim 1, characterized in that: Also includes: The fault diagnosis result is graded by a fault classification management module; The fault recovery module performs various recovery operations according to the fault classification results.

6. The satellite fault diagnosis method according to claim 5, characterized in that: The fault classification management module classifies the explicit faults and implicit faults into four levels according to the output result of the fault diagnosis module, wherein: The first level is the failure that affects the safety of the entire satellite and cannot be recovered autonomously; The second level is a fault that affects the payload mission and cannot be recovered autonomously; The third level is the failure that can be recovered by reconstruction; The fourth level is a fault whose impact can be recovered by resetting.

7. The satellite fault diagnosis method according to claim 6, characterized in that: In the event that the satellite is unable to recover autonomously, autonomous protection measures will be implemented onboard for level 1 and level 2 faults to ensure the safety of the entire satellite. Autonomous recovery measures will be implemented for level 3 and level 4 faults, including resetting, power-off and reconstruction measures, while ensuring the normality of the payload mission.

8. The satellite fault diagnosis method according to claim 7, characterized in that: The first level of failures include energy failures, attitude control failures, and satellite service failures. The satellite will start to enter the safety mode and implement self-protection. The second level of failure includes time-frequency failure and amplifier failure, and the satellite will adjust the transmission power to implement self-protection; The third level of failure involves all subsystems of the entire satellite, and initiates a fault recovery command to the platform computer to switch the faulty single machine to the backup work; The fourth level of faults can be recovered through autonomous software compensation, resetting or powering on and off. A fault recovery command is sent to the platform computer or load task processor to autonomously reset or power on the faulty unit.

Images