CN113961612A - Satellite health data comprehensive analysis system and method based on deep learning - Google Patents
Satellite health data comprehensive analysis system and method based on deep learning Download PDFInfo
- Publication number
- CN113961612A CN113961612A CN202110824648.0A CN202110824648A CN113961612A CN 113961612 A CN113961612 A CN 113961612A CN 202110824648 A CN202110824648 A CN 202110824648A CN 113961612 A CN113961612 A CN 113961612A
- Authority
- CN
- China
- Prior art keywords
- data
- fault
- load
- unit
- analysis
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2458—Special types of queries, e.g. statistical queries, fuzzy queries or distributed queries
- G06F16/2465—Query processing support for facilitating data mining operations in structured databases
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/28—Databases characterised by their database models, e.g. relational or object models
- G06F16/283—Multi-dimensional databases or data warehouses, e.g. MOLAP or ROLAP
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N5/00—Computing arrangements using knowledge-based models
- G06N5/02—Knowledge representation; Symbolic representation
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Databases & Information Systems (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- Fuzzy Systems (AREA)
- Probability & Statistics with Applications (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention relates to the technical field of satellite health data management, in particular to a satellite health data comprehensive analysis system and method based on deep learning; load data of the satellite is obtained through a data mining module, the data are managed and normalized, a generated data warehouse is uploaded to a load data fault knowledge base, and real-time load data are uploaded to an effective load fault detection module; the load data fault knowledge base analyzes the change data of the load parameters of the data warehouse through a big data mining technology, and establishes a fault sample set, a fault model and a fault type; the effective load fault detection module obtains real-time load data, carries out fault analysis on the real-time load data, finally imports analysis parameters into a fault model and a fault type for matching, and finally outputs a result for workers to obtain satellite health information in real time, so that the labor intensity of workers is reduced.
Description
Technical Field
The invention relates to the technical field of satellite health data management, in particular to a satellite health data comprehensive analysis system and method based on deep learning.
Background
At present, the aerospace industry of China is rapidly developed, the tasks of satellites are more and more complicated, the telemetry data of the satellites is increased at a very high speed, and then the acquired satellite parameters are numerous, but the current satellite mission system cannot efficiently meet daily requirements.
The load parameter correlation analysis technology is the basis of fault detection, health analysis and prediction technology, and in the aspect of effective load health management of a satellite, the existing system processing has the problems of low efficiency, low data interpretation accuracy, unreasonable model construction and the like, so that the accuracy of numerical values is judged manually according to experience in many cases at present, and manpower is greatly wasted.
Disclosure of Invention
The invention aims to provide a satellite health data comprehensive analysis system and method based on deep learning, and aims to solve the technical problems that in the aspect of effective load health management of satellites in the prior art, system processing has many problems, numerical accuracy is judged manually according to experience in many cases, and manpower is greatly wasted.
In order to achieve the purpose, the invention adopts a satellite health data comprehensive analysis system based on deep learning, which comprises a data mining module, a load data fault knowledge base and an effective load fault detection module;
the data mining module is connected with the payload fault detection module, and the load data fault knowledge base is respectively connected with the data mining module and the payload fault detection module;
the data mining module is used for acquiring load data, performing data management and normalization on the load data to acquire recombined data, uploading the recombined data to the effective load fault detection module, establishing a data warehouse according to the recombined data, and uploading the data warehouse to the load fault knowledge base;
the load fault data knowledge base is used for acquiring the data warehouse, analyzing the change trend of load parameters in the data warehouse according to a certain specific state or a specific time interval or a non-artificial event, and establishing a fault sample set, a fault model and a fault type;
and the payload fault detection module is used for acquiring the recombined data, performing fault analysis on the recombined data, acquiring set parameters, matching the fault model and the fault type according to the set parameters and outputting a result.
The data mining module is used for acquiring load data of a satellite, wherein the load data comprises a satellite real-time load database, a satellite full-life-cycle load database and a satellite data comprehensive basic database, extracting, cleaning, converting and summarizing the load data to obtain recombined data, establishing the recombined data into a data warehouse, uploading the data to the load data fault knowledge base for storage, and uploading the real-time data in the recombined data to the effective fault detection module; the load data fault knowledge base establishes a fault sample set, a fault model and a fault type based on a data warehouse; the effective load fault detection module is used for carrying out fault analysis on the real-time data, extracting set parameters according to an analysis result, inputting the set parameters into a fault model and a fault type for matching, and finally outputting a matching result for workers to find satellite abnormity in real time so as to reduce manpower observation.
The data mining module comprises a data analysis unit and a standardization unit, and the standardization unit is connected with the data analysis unit;
the data analysis unit is used for acquiring the load data, translating the binary load data into load analysis service data in a standard format, and uploading the load analysis service data to the standardization unit;
and the standardization unit is used for carrying out format standardization processing on the existing load analysis service data to obtain standard load analysis data.
The data analysis unit translates the load data, and the standardization unit standardizes the translated load data, so that the next operation is facilitated.
The data mining module further comprises a data cleaning unit and a data preprocessing unit, the data cleaning unit is connected with the standardization unit, the data preprocessing unit is connected with the data cleaning unit, and the preprocessing unit is further connected with the load data fault knowledge base and the payload fault detection module;
the data cleaning unit is used for removing repeated rows and columns of the standard load analysis data, filling default values and time mark alignment, and obtaining cleaning load analysis data;
the data preprocessing unit is used for discretizing, dualizing, normalizing and carrying out variable transformation on the cleaning load analysis data to obtain the recombined data, uploading the recombined data to the payload fault detection module, establishing the recombined data into the data warehouse, and uploading the data warehouse to the load data fault knowledge base.
The data cleaning unit is used for cleaning the standardized load data, and the data preprocessing unit is used for converting and summarizing the cleaned load data, so that the next operation is facilitated.
The data mining module further comprises a feature extraction unit and an index tag unit, wherein the feature extraction unit is connected with the data preprocessing unit, and the index tag unit is connected with the feature extraction unit;
the characteristic extraction unit is used for extracting the characteristics of the recombined data and storing the acquired characteristics into a characteristic library;
and the index tag unit is used for establishing an index tag for the recombined data.
The characteristic unit extracts the characteristics of the ethnic data, solves the problem of dimension disaster, reduces the difficulty of learning tasks and provides data support for subsequent tasks, and the index tag unit establishes an index tag for the collected recombined data and provides a basis for subsequent loading data storage, retrieval and query.
The load data fault knowledge base comprises a fault sample unit and a fault model unit, the fault sample unit is connected with the data preprocessing unit, and the fault model unit is respectively connected with the payload fault detection module and the fault sample unit;
the fault sample unit is used for acquiring the data warehouse, analyzing the change trend of the load parameters in the data warehouse according to a certain specific state or a specific time interval or a non-artificial event, and establishing a fault sample set;
and the fault model unit is used for establishing a fault model and a judgment rule according to the fault sample set.
The fault model unit is based on a data warehouse, analyzes the variation trend of the load parameters in the data warehouse, establishes a fault sample set, and establishes a fault model and a judgment rule according to the fault sample set.
A satellite health data comprehensive analysis method based on deep learning comprises the following steps:
acquiring load data, performing data management and normalization on the load data to obtain recombined data, uploading the recombined data to an effective load fault detection module, establishing a data warehouse according to the recombined data, and uploading the data warehouse to a load fault knowledge base;
acquiring the data warehouse, analyzing the change trend of the load parameters in the data warehouse according to a certain specific state or a specific time interval or a non-artificial event, and establishing a fault sample set, a fault model and a fault type;
and acquiring the recombined data, performing fault analysis on the recombined data, acquiring set parameters, matching the fault model and the fault type according to the set parameters, and outputting a result.
Firstly, load data of a satellite are obtained, wherein the load data comprise a real-time load database, a full-life-cycle load database and a satellite data comprehensive basic database, the data are managed and normalized, a generated data warehouse is uploaded to a load fault knowledge base, and the real-time load data are uploaded to an effective load fault detection module for analysis; then, the load fault knowledge base analyzes data in a massive data warehouse through a big data mining technology, models, creates a fault sample set, a fault model, a judgment rule and a fault type, and provides data support for algorithm analysis; the effective load fault detection module receives and analyzes real-time load data, extracts set parameters, changes fault models and fault types which import the set parameters into the load data fault knowledge base to be matched, and finally outputs matching results to enable workers to obtain satellite fault information in real time and check fault phenomena and reason information, so that the labor intensity of workers is reduced.
The invention relates to a satellite health data comprehensive analysis system and a method based on deep learning, wherein a real-time load database, a full-life-cycle load database and a micro data comprehensive basic database of a satellite are obtained through a data mining module, the databases are managed and normalized, a generated data warehouse is uploaded to a load data fault knowledge base, and real-time load data is uploaded to an effective load fault detection module; the load data fault knowledge base analyzes mass data of a data warehouse through a big data mining technology, models, analyzes change data of load parameters, and establishes a fault sample set, a fault model and a fault type; the effective load fault detection module acquires real-time load data, performs fault analysis on the real-time load data, extracts set parameters, finally imports the set parameters into the load fault data knowledge base, performs matching in a fault model and a fault type, and finally outputs a result to enable workers to acquire satellite health information in real time, so that the labor intensity of workers is reduced.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
Fig. 1 is a schematic structural diagram of a deep learning-based satellite health data comprehensive analysis system of the present invention.
FIG. 2 is a block diagram of the data mining module of the present invention.
Fig. 3 is a schematic diagram of the structure of the load data failure knowledge base of the present invention.
Fig. 4 is a schematic structural diagram of a payload failure detection module of the present invention.
Fig. 5 is a step diagram of the deep learning-based satellite health data comprehensive analysis method of the present invention.
The method comprises the steps of 1-a data mining module, 2-a load data fault knowledge base, 3-a payload fault detection module, 11-a data analysis unit, 12-a standardization unit, 13-a data cleaning unit, 14-a data preprocessing unit, 15-a feature extraction unit, 16-an index tag unit, 21-a fault sample unit, 22-a fault model unit, 23-a fault type unit, 31-a fault analysis unit, 32-a parameter setting unit and 100-a deep learning-based satellite health data comprehensive analysis system.
Detailed Description
Reference will now be made in detail to embodiments of the present invention, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the same or similar elements or elements having the same or similar function throughout. The embodiments described below with reference to the drawings are illustrative and intended to be illustrative of the invention and are not to be construed as limiting the invention.
Referring to fig. 1 to 4, the invention provides a deep learning-based satellite health data comprehensive analysis system 100, which includes a data mining module 1, a load data fault knowledge base 2 and an effective load fault detection module 3;
the data mining module 1 is connected with the payload fault detection module 3, and the load data fault knowledge base 2 is respectively connected with the data mining module 1 and the payload fault detection module 3;
the data mining module 1 is configured to acquire load data, perform data management and normalization on the load data, acquire recombined data, upload the recombined data to the payload fault detection module 3, establish a data warehouse according to the recombined data, and upload the data warehouse to the load fault knowledge base;
the load fault data knowledge base is used for acquiring the data warehouse, analyzing the change trend of load parameters in the data warehouse according to a certain specific state or a specific time interval or a non-artificial event, and establishing a fault sample set, a fault model and a fault type;
and the payload fault detection module 3 is configured to obtain the regrouped data, perform fault analysis on the regrouped data, obtain a set parameter, match the fault model and the fault type according to the set parameter, and output a result.
In the embodiment, the data mining module 1 is used for acquiring load data of a satellite, wherein the load data comprises a real-time load database, a full-life-cycle load database and a satellite data comprehensive basic database of the satellite, data management and normalization of extraction, cleaning, conversion and induction are performed on the databases to generate recombined data, the recombined data is established into a data warehouse and then uploaded to the load data fault knowledge base 2, and the real-time load data of the satellite in the recombined data is uploaded to the effective load fault detection module 3; the load data failure knowledge base 2 acquires a data base, analyzes and models the data base through a big data mining technology, establishes a health state knowledge base, provides data support for algorithm analysis, analyzes the variation trend of parameters according to a certain state or a specific time interval or a non-artificial event, calculates the prediction trend of each parameter by establishing a prediction model of the load, inputting the prediction trend of each parameter into the model to calculate to acquire the trend prediction of load health, obtains a newly trained data set through a training data set and a Bagging algorithm, calculates through an initialization weight, calculates according to a sub-model fusion function and the initialization weight, establishes a failure sample set, a failure model and a failure type, wherein the failure model has rules of threshold judgment and envelope judgment, the fault type comprises a fault phenomenon and a fault reason; the effective load fault detection module 3 acquires real-time load data of a satellite, performs fault analysis on the real-time load data, extracts variation parameters, generates setting parameters, introduces the setting parameters into the load data fault knowledge base 2, and matches the setting parameters with a fault model and a fault type to obtain the fault type, the fault phenomenon and the fault reason, so that maintenance personnel can quickly position and eliminate the fault; further, a deep self-learning algorithm is provided, mass data are effectively utilized, and a model is accurately constructed; the method supports flexible configuration, realizes expandability, supports integration of external algorithms, provides multi-party support for diagnosis means, and can realize integration of external algorithm libraries by reserving a system interface, introducing and integrating external algorithm program packages on line, configuring information such as paths, accessing and participating and the like, and calling and configuring system services; the correlation mining technology can analyze data with different dimensions, can greatly improve the analysis efficiency and accuracy, and can guarantee a high-timeliness result during fault diagnosis; through the integrated information fusion method, massive heterogeneous data is analyzed, analyzed and judged, effective data analysis and interpretation are effectively achieved, the fault analysis efficiency is improved, and the amount of manual labor is greatly reduced.
Further, referring to fig. 2, the data mining module 1 includes a data parsing unit 11 and a normalizing unit 12, and the normalizing unit 12 is connected to the data parsing unit 11;
the data analysis unit 11 is configured to obtain the load data, translate the binary load data into load analysis service data in a standard format, and upload the load analysis service data to the normalization unit 12;
the standardizing unit 12 is configured to perform format standardization processing on the existing load analysis service data to obtain standard load analysis data.
Further, referring to fig. 2, the data mining module 1 further includes a data cleaning unit 13 and a data preprocessing unit 14, the data cleaning unit 13 is connected to the normalization unit 12, the data preprocessing unit 14 is connected to the data cleaning unit 13, and the preprocessing unit is further connected to the payload data failure knowledge base 2 and the payload failure detection module 3;
the data cleaning unit 13 is configured to remove repeated rows and columns of the standard load analysis data, fill default values and align time marks, and obtain cleaning load analysis data;
the data preprocessing unit 14 is configured to discretize, binarize, normalize, and perform variable transformation on the cleaning load analysis data to obtain the restructured data, upload the restructured data to the payload fault detection module 3, establish the restructured data as the data warehouse, and upload the data warehouse to the load data fault knowledge base 2.
Further, referring to fig. 2, the data mining module 1 further includes a feature extraction unit 15 and an index tag unit 16, the feature extraction unit 15 is connected to the data preprocessing unit 14, and the index tag unit 16 is connected to the feature extraction unit 15;
the feature extraction unit 15 is configured to extract features of the restructured data, and store the acquired features in a feature library;
the index tag unit 16 is configured to create an index tag for the reassembled data.
In this embodiment, the data parsing unit 11 translates the obtained binary load data into load analysis service data in a standard format, where the data format of the binary load data may be described by an XML file, and the software parses data in different formats based on the XML file and transmits the load analysis service data to the normalizing unit 12; the standardization unit 12 is used for carrying out format standardization processing on the existing load analysis service data to realize uniform format, so that the next operation is facilitated; the data cleaning unit 13 is used for removing the heavy rows and the heavy columns in the standard load analysis data, filling the missing values and performing time scale alignment to obtain cleaning load analysis data; the data preprocessing module is used for discretizing, dualizing, normalizing and carrying out variable transformation on the cleaning load analysis data to obtain recombined data, providing data support for algorithm analysis, establishing the recombined data into a data warehouse, uploading the data warehouse to the load data fault knowledge base 2 for storage, and uploading real-time load data in the recombined data to the effective load fault detection module 3 for detection; the feature extraction unit 15 is used for extracting and selecting features of the reconstructed data so as to solve the problem of dimension disaster and reduce the difficulty of learning tasks, and storing the obtained features into a feature library in a database so as to provide data support for subsequent tasks; the index tag unit 16 establishes an index tag for the grouped data, provides a basis for warehousing, retrieving and inquiring subsequently acquired load data, and analyzes and processes mass data through a big data mining technology, so that effective data analysis and interpretation are effectively realized, the analysis efficiency of faults is improved, and the amount of manual labor is reduced.
Further, referring to fig. 3, the load data failure knowledge base 2 includes a failure sample unit 21 and a failure model unit 22, the failure sample unit 21 is connected to the data preprocessing unit 14, and the failure model unit 22 is connected to the payload failure detection module 3 and the failure sample unit 21, respectively;
the fault sample unit 21 is configured to obtain the data warehouse, analyze a variation trend of the load parameter in the data warehouse according to a certain specific state or a specific time interval or in a non-artificial event, and establish a fault sample set;
and the fault model unit 22 is configured to establish a fault model and a judgment rule according to the fault sample set.
Further, referring to fig. 3, the load data failure knowledge base 2 further includes a failure type unit 23, and the failure type unit 23 is connected to the payload failure detection module 3 and the failure model unit 22 respectively;
the fault type unit 23 is configured to classify the fault sample set based on the fault model to obtain a fault type.
In this embodiment, the fault sample unit 21 obtains a data warehouse, and analyzes the variation trend of the parameters according to a certain specific state or a specific time interval or in a non-artificial event by using spatial data mining (tree structure) and multidimensional time series correlation analysis (graph structure) through a structured data correlation mining technology, so as to establish a fault sample set, and update the data warehouse in real time to enrich the fault sample set; the fault model unit 22 suggests a fault model and a judgment rule for threshold judgment and envelope judgment based on a fault sample set; the fault type module classifies the fault sample set based on the fault model, associates the fault type, the fault phenomenon and the fault reason correspondingly, and provides data support for the payload fault detection module 3.
Further, referring to fig. 4, the payload failure detection module 3 includes a failure analysis unit 31 and a parameter setting unit 32, where the failure analysis unit 31 is connected to the data preprocessing unit 14, and the parameter setting unit 32 is connected to the failure analysis unit 31, the failure model unit 22, and the failure type unit 23, respectively;
the fault analysis unit 31 is configured to obtain the regrouped data, perform fault analysis within a load and between loads on a parameter abnormal condition in the regrouped data, and upload an analysis result to the parameter setting unit 32;
the parameter setting unit 32 is configured to obtain the analysis result, generate corresponding parameters, upload the parameters to the fault model unit 22 and the fault type unit 23 for matching, and finally output a matching result.
In this embodiment, the fault analysis unit 31 obtains real-time load data, performs intra-load and inter-load fault analysis on the real-time load data, and uploads an analysis result to the parameter setting unit 32, the parameter setting unit 32 extracts abnormal conditions of each parameter according to the analysis result, generates corresponding abnormal parameter data, detects abnormal parameters by calling a fault model in the fault model unit 22, and matches the output fault type with the fault type unit 23 to give out a fault type, a fault phenomenon, and a fault reason, so that maintenance personnel can quickly locate and eliminate faults, and reduce the amount of manual labor.
Referring to fig. 5, a method for comprehensively analyzing satellite health data based on deep learning includes the following steps:
s101: acquiring load data, performing data management and normalization on the load data to obtain recombined data, uploading the recombined data to an effective load fault detection module 3, establishing a data warehouse according to the recombined data, and uploading the data warehouse to a load fault knowledge base.
S102: and acquiring the data warehouse, analyzing the change trend of the load parameters in the data warehouse according to a certain specific state or a specific time interval or a non-artificial event, and establishing a fault sample set, a fault model and a fault type.
S103: and acquiring the recombined data, performing fault analysis on the recombined data, acquiring set parameters, matching the fault model and the fault type according to the set parameters, and outputting a result.
In this embodiment, first, load data of a satellite is obtained, where the load data includes a real-time load database, a full-life-cycle load database, and a satellite data comprehensive basic database, and the data is managed and normalized, and a generated data warehouse is uploaded to the load fault knowledge base, and the real-time load data is also uploaded to the payload fault detection module 3 for analysis; then, the load fault knowledge base analyzes data in a massive data warehouse through a big data mining technology, models, creates a fault sample set, a fault model, a judgment rule and a fault type, and provides data support for next algorithm analysis; then, the payload fault detection module 3 receives the real-time load data, analyzes the real-time load data, extracts the set parameters, changes the fault model and the fault type which lead the set parameters into the load data fault knowledge base 2 to match, and finally outputs the matching result to give out the fault type, the fault phenomenon and the fault reason, so that maintenance personnel can quickly position and eliminate the fault, and the labor intensity of workers is reduced.
While the invention has been described with reference to a preferred embodiment, it will be understood by those skilled in the art that various changes in form and detail may be made therein without departing from the spirit and scope of the invention as defined by the appended claims.
Claims (6)
1. A satellite health data comprehensive analysis system based on deep learning is characterized by comprising a data mining module, a load data fault knowledge base and an effective load fault detection module;
the data mining module is connected with the payload fault detection module, and the load data fault knowledge base is respectively connected with the data mining module and the payload fault detection module;
the data mining module is used for acquiring load data, performing data management and normalization on the load data to acquire recombined data, uploading the recombined data to the effective load fault detection module, establishing a data warehouse according to the recombined data, and uploading the data warehouse to the load fault knowledge base;
the load fault data knowledge base is used for acquiring the data warehouse, analyzing the change trend of load parameters in the data warehouse according to a certain specific state or a specific time interval or a non-artificial event, and establishing a fault sample set, a fault model and a fault type;
and the payload fault detection module is used for acquiring the recombined data, performing fault analysis on the recombined data, acquiring set parameters, matching the fault model and the fault type according to the set parameters and outputting a result.
2. The deep learning based satellite health data integrated analysis system of claim 1,
the data mining module comprises a data analysis unit and a standardization unit, and the standardization unit is connected with the data analysis unit;
the data analysis unit is used for acquiring the load data, translating the binary load data into load analysis service data in a standard format, and uploading the load analysis service data to the standardization unit;
and the standardization unit is used for carrying out format standardization processing on the existing load analysis service data to obtain standard load analysis data.
3. The deep learning based satellite health data integrated analysis system of claim 2,
the data mining module further comprises a data cleaning unit and a data preprocessing unit, the data cleaning unit is connected with the standardization unit, the data preprocessing unit is connected with the data cleaning unit, and the preprocessing unit is further connected with the load data fault knowledge base and the payload fault detection module;
the data cleaning unit is used for removing repeated rows and columns of the standard load analysis data, filling default values and time mark alignment, and obtaining cleaning load analysis data;
the data preprocessing unit is used for discretizing, dualizing, normalizing and carrying out variable transformation on the cleaning load analysis data to obtain the recombined data, uploading the recombined data to the payload fault detection module, establishing the recombined data into the data warehouse, and uploading the data warehouse to the load data fault knowledge base.
4. The deep learning based satellite health data integrated analysis system of claim 3,
the data mining module also comprises a feature extraction unit and an index tag unit, wherein the feature extraction unit is connected with the data preprocessing unit, and the index tag unit is connected with the feature extraction unit;
the characteristic extraction unit is used for extracting the characteristics of the recombined data and storing the acquired characteristics into a characteristic library;
and the index tag unit is used for establishing an index tag for the recombined data.
5. The deep learning based satellite health data integrated analysis system of claim 3,
the load data fault knowledge base comprises a fault sample unit and a fault model unit, the fault sample unit is connected with the data preprocessing unit, and the fault model unit is respectively connected with the effective load fault detection module and the fault sample unit;
the fault sample unit is used for acquiring the data warehouse, analyzing the change trend of the load parameters in the data warehouse according to a certain specific state or a specific time interval or a non-artificial event, and establishing a fault sample set;
and the fault model unit is used for establishing a fault model and a judgment rule according to the fault sample set.
6. A deep learning based satellite health data integrated analysis method, the deep learning based satellite health data integrated analysis system according to claim 1, comprising the steps of:
acquiring load data, performing data management and normalization on the load data to obtain recombined data, uploading the recombined data to an effective load fault detection module, establishing a data warehouse according to the recombined data, and uploading the data warehouse to a load fault knowledge base;
acquiring the data warehouse, analyzing the change trend of the load parameters in the data warehouse according to a certain specific state or a specific time interval or a non-artificial event, and establishing a fault sample set, a fault model and a fault type;
and acquiring the recombined data, performing fault analysis on the recombined data, acquiring set parameters, matching the fault model and the fault type according to the set parameters, and outputting a result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110824648.0A CN113961612A (en) | 2021-07-21 | 2021-07-21 | Satellite health data comprehensive analysis system and method based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110824648.0A CN113961612A (en) | 2021-07-21 | 2021-07-21 | Satellite health data comprehensive analysis system and method based on deep learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113961612A true CN113961612A (en) | 2022-01-21 |
Family
ID=79460402
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110824648.0A Pending CN113961612A (en) | 2021-07-21 | 2021-07-21 | Satellite health data comprehensive analysis system and method based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113961612A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114442057A (en) * | 2022-01-25 | 2022-05-06 | 国家卫星海洋应用中心 | Microwave scatterometer load fault positioning method and system based on HY2 satellite |
| CN115293459A (en) * | 2022-09-26 | 2022-11-04 | 北京开运联合信息技术集团股份有限公司 | Digital twin satellite payload health management system |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109934130A (en) * | 2019-02-28 | 2019-06-25 | 中国空间技术研究院 | The in-orbit real-time fault diagnosis method of satellite failure and system based on deep learning |
| CN111874268A (en) * | 2020-07-22 | 2020-11-03 | 中国科学院微小卫星创新研究院 | Satellite centralized type autonomous health management system |
| CN112152698A (en) * | 2020-09-24 | 2020-12-29 | 中国科学院空天信息创新研究院 | Health monitoring system and method for remote sensing satellite ground station |
| CN112487058A (en) * | 2020-11-30 | 2021-03-12 | 北京航空航天大学 | Numerical control machine tool fault monitoring and diagnosing system based on data mining |
-
2021
- 2021-07-21 CN CN202110824648.0A patent/CN113961612A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109934130A (en) * | 2019-02-28 | 2019-06-25 | 中国空间技术研究院 | The in-orbit real-time fault diagnosis method of satellite failure and system based on deep learning |
| CN111874268A (en) * | 2020-07-22 | 2020-11-03 | 中国科学院微小卫星创新研究院 | Satellite centralized type autonomous health management system |
| CN112152698A (en) * | 2020-09-24 | 2020-12-29 | 中国科学院空天信息创新研究院 | Health monitoring system and method for remote sensing satellite ground station |
| CN112487058A (en) * | 2020-11-30 | 2021-03-12 | 北京航空航天大学 | Numerical control machine tool fault monitoring and diagnosing system based on data mining |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114442057A (en) * | 2022-01-25 | 2022-05-06 | 国家卫星海洋应用中心 | Microwave scatterometer load fault positioning method and system based on HY2 satellite |
| CN115293459A (en) * | 2022-09-26 | 2022-11-04 | 北京开运联合信息技术集团股份有限公司 | Digital twin satellite payload health management system |
| CN115293459B (en) * | 2022-09-26 | 2022-12-06 | 北京开运联合信息技术集团股份有限公司 | Digital twin satellite payload health management system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110659173B (en) | Operation and maintenance system and method | |
| CN112395170A (en) | Intelligent fault analysis method, device, equipment and storage medium | |
| CN113961612A (en) | Satellite health data comprehensive analysis system and method based on deep learning | |
| CN111047082A (en) | Early warning method and device for equipment, storage medium and electronic device | |
| CN114118224A (en) | Neural network-based system-wide remote measurement parameter anomaly detection system | |
| CN116599857B (en) | Digital twin application system suitable for multiple scenes of Internet of things | |
| CN110674189A (en) | Method for monitoring secondary state and positioning fault of intelligent substation | |
| CN111224840A (en) | Gateway system with fault diagnosis performance and method thereof | |
| CN112685215A (en) | Cloud platform abnormal log analysis method | |
| CN115576293B (en) | Pressure-sensitive adhesive on-line production analysis method and system based on data monitoring | |
| CN110895506A (en) | Construction method and construction system of test data | |
| CN116361147A (en) | Method for positioning root cause of test case, device, equipment, medium and product thereof | |
| EP4052446A1 (en) | Information acquiring method, apparatus, and system | |
| WO2024104406A1 (en) | Anomaly detection method and cloud network platform | |
| Dutta et al. | Big data architecture for environmental analytics | |
| CN115809795B (en) | Method and device for evaluating bearing capacity of production team based on digitalization | |
| CN111708756A (en) | Method, device and equipment for automatically processing data warehouse and storage medium | |
| US20230376795A1 (en) | Device, computing platform and method of analyzing log files of an industrial plant | |
| CN114245895A (en) | Method for generating consistent representation for at least two log files | |
| CN113807704A (en) | Intelligent algorithm platform construction method for urban rail transit data | |
| CN113434295B (en) | Farmland monitoring method, device, equipment and storage medium based on edge calculation | |
| CN115080669B (en) | Method, device, equipment and storage medium for generating full error-prevention information point table | |
| CN112329937B (en) | GIS fault diagnosis method based on case and fault reasoning | |
| Lei et al. | Application of distributed machine learning model in fault diagnosis of air preheater | |
| CN113779312B (en) | Unstructured power grid data processing method and device based on knowledge graph |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20220121 |
|
| RJ01 | Rejection of invention patent application after publication |