Automatic rule generated data transmission system and method for judging rocket fault

The rocket fault judgment data transmission system generated by automated rules integrates the data of each subsystem of the rocket, and uses intelligent models and expert systems to achieve automated fault detection, solving the problems of data dispersion and human work dependence in pre-launch testing, and improving faults Check efficiency and system reliability.

CN118981652AInactive Publication Date: 2024-11-19SHANGHAI AEROSPACE EQUIPMENTS MANUFACTURER CO LTD
View PDF 3 Cites 0 Cited by

Patent Information

Application Number
CN202411471942.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-10-22
Publication Date
2024-11-19
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

During the rocket pre-launch test, due to the dispersed test data of each subsystem, it is difficult for on-site testers to fully understand the system status, making it difficult to quickly judge and process the fault conditions. The test system relies on human work, reducing the degree of automation.

Method used

Design a data transmission system that automatically generates rules to determine rocket failures, including data acquisition, preprocessing, intelligent model training, rule generation and optimization modules, and generate interpretable Automate fault judgment rules and use expert systems to perform fault detection.

Benefits of technology

It realizes the integrated transmission of data of various subsystems of the rocket, improves the efficiency of troubleshooting and system reliability, reduces labor costs, solves the problem of human dependence, and improves the degree of automation of rocket fault detection.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118981652A_ABST
    Figure CN118981652A_ABST
Patent Text Reader

Abstract

The invention provides a data transmission system and method for judging a rocket fault based on automatic rule generation, and the system comprises a data collection module which is used for collecting rocket monitoring data; the data transmission module is used for transmitting the rocket monitoring data to the data preprocessing module; the data preprocessing module is used for receiving and preprocessing the rocket monitoring data; the preprocessing comprises data cleaning and trend prediction-oriented high-dimensional data dimension reduction; and the intelligent model training module is used for training an intelligent model based on the preprocessed rocket monitoring data and capturing the relevance between the data. According to the invention, the collected test data of each subsystem of the rocket can be integrated and transmitted, so that an operator can master the overall condition of the system, the efficiency of troubleshooting and the reliability of system operation are improved, and the manpower cost is reduced.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of fault diagnosis, and in particular to a data transmission system and method for judging rocket faults using automated rule generation. Background Art

[0002] The pre-launch test of a rocket covers the test of each subsystem of the rocket and the overall test of the system. With the refinement of professional division of labor and the influence of traditional patterns, the test of each subsystem of the rocket is usually self-contained. The test of each subsystem needs to be carried out by professionals who are very familiar with the system. In the overall pre-launch test, due to the scattered test data of each subsystem, it is difficult for on-site testers to fully understand the system test status, and it is also difficult to quickly judge and deal with fault conditions. According to statistics, human factors account for as much as 40% of the abnormal causes of all industrial systems. Therefore, it is urgent to develop an unmanned monitoring system for each subsystem of the pre-launch test to improve the reliability of system operation.

[0003] However, the monitoring of the test system operation process still relies on manual work, which reduces the automation level of the entire system. Summary of the invention

[0004] In view of the defects in the prior art, the purpose of the present invention is to provide a data transmission system and method for judging rocket failures with automated rule generation.

[0005] According to the present invention, a data transmission system for determining rocket failures by automatic rule generation is provided, comprising: Data acquisition module, used to collect rocket monitoring data; A data transmission module, used to transmit rocket monitoring data to a data preprocessing module; The data preprocessing module receives and preprocesses the rocket monitoring data, and sends the preprocessed rocket monitoring data to the intelligent model training module; The preprocessing includes data cleaning and high-dimensional data dimensionality reduction for trend prediction; The data cleaning includes: Processing missing values ​​of data, replacing missing data with preset values ​​through a missing value replacement method; the missing value replacement method includes mean filling, median filling, maximum value filling, minimum value filling and deletion of missing values; Perform outlier processing on the data, and replace the outliers with preset values ​​by an outlier replacement method; the outlier replacement method includes a Z-score method, an IQR method, and a data distribution method; Processing duplicate values ​​on the data, deleting or merging duplicate data by using a duplicate value processing method; the duplicate value processing method includes a deduplication method and a clustering method; Performing type conversion on the data, and performing data conversion by a data conversion method; the data conversion method includes type conversion and format conversion; the data conversion includes integer to floating point number conversion, string to integer conversion, date format conversion and time format conversion; Normalizing the data, scaling the range of the data by a data normalization method; the normalization method includes maximum and minimum normalization and Z-score normalization; Intelligent model training module, which trains the intelligent model based on the pre-processed rocket monitoring data to capture the correlation between the data; The rule generation and optimization module uses the trained intelligent model to generate rules and perform optimization; The rule application module constructs an expert system based on the optimized rules; the expert system can be converted into a script language to run and output fault judgment results.

[0006] Preferably, the high-dimensional data dimensionality reduction includes: Based on the PCA method of entropy weight method, the standardized matrix P is obtained: ; Where m is the number of data items that need to be reduced in dimension, n is the data feature dimension, is the value of the jth feature of the i-th data after standardization, calculated as:

[0007] in is the jth feature of the i-th data in the original data. Then the information entropy value is calculated for each feature j in the data set : ; Using information entropy Calculate feature weights : n.

[0008] Preferably, the rule generation and optimization module includes: Load the trained intelligent model, input the preprocessed data into the model and obtain the prediction results; Based on the processed data and prediction structure, the decision tree model is trained to simulate the decision-making process of the intelligent model and generate explainable rules; The rules are optimized by genetic algorithm and simulated annealing technology.

[0009] Preferably, the rule application module is based on a rule-based expert system, which converts the rules into Python language and runs them in an infinite loop. The diagnostic function consists of all rules extracted by automatic rule generation, and the result of each rule is saved as an array; the diagnostic function can detect multiple faults simultaneously.

[0010] A data transmission method for determining rocket failures by automatic rule generation provided by the present invention includes: Step S1: Collecting rocket monitoring data and transmitting it; Step S2: receiving rocket monitoring data and preprocessing; The preprocessing includes data cleaning and high-dimensional data dimensionality reduction for trend prediction; The data cleaning includes: Processing missing values ​​of data, replacing missing data with preset values ​​through a missing value replacement method; the missing value replacement method includes mean filling, median filling, maximum value filling, minimum value filling and deletion of missing values; Perform outlier processing on the data, and replace the outliers with preset values ​​by an outlier replacement method; the outlier replacement method includes a Z-score method, an IQR method, and a data distribution method; Processing duplicate values ​​on the data, deleting or merging duplicate data by using a duplicate value processing method; the duplicate value processing method includes a deduplication method and a clustering method; Performing type conversion on the data, and performing data conversion by a data conversion method; the data conversion method includes type conversion and format conversion; the data conversion includes integer to floating point number conversion, string to integer conversion, date format conversion and time format conversion; The data is normalized, and the range of the data is scaled by a data normalization method; the normalization method includes maximum and minimum normalization and Z-score normalization.

[0011] Step S3: training the intelligent model based on the preprocessed rocket monitoring data to capture the correlation between the data; Use the trained intelligent model to generate rules and optimize them; An expert system is constructed based on the optimized rules; the expert system can be converted into a script language to run and output a fault judgment result.

[0012] Preferably, the high-dimensional data dimensionality reduction includes: The PCA method based on the entropy weight method calculates the value of the standardized data features:

[0013] Get the normalized matrix: ; Calculate the information entropy value for each feature j in the data set : ; Using information entropy Calculate feature weights : n.

[0014] Compared with the prior art, the present invention has the following beneficial effects: 1. The present invention can integrate and transmit the collected test data of various subsystems of the rocket, so that the operator can grasp the overall status of the system, improve the efficiency of troubleshooting and the reliability of system operation, and reduce the expenditure of labor costs.

[0015] 2. The present invention uses an intelligent model to perform model training on preprocessed data to capture complex patterns and relationships in the data, and then generate automated rules; by optimizing the generated rules, the problem of manual operation dependence of rocket fault detection is solved.

[0016] 3. The present invention can utilize recorded unlabeled behavior sequence data for training, and at the same time combines supervised learning to improve training efficiency, thereby ensuring that normal use by users is not affected and having good practicality.

[0017] Other beneficial effects of the present invention will be explained in the specific implementation manner through the introduction of specific technical features and technical solutions. Through the introduction of these technical features and technical solutions, those skilled in the art should be able to understand the beneficial technical effects brought about by the technical features and technical solutions. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Other features, objects and advantages of the present invention will become more apparent from the detailed description of non-limiting embodiments made with reference to the following drawings: Figure 1 A schematic diagram of a flow chart of the rule-based automatic generation of results of the present invention.

[0019] Figure 2 The figure is an operation flow chart of an embodiment of the present invention. DETAILED DESCRIPTION

[0020] The present invention is described in detail below in conjunction with specific embodiments. The following embodiments will help those skilled in the art to further understand the present invention, but are not intended to limit the present invention in any form. It should be noted that, for those of ordinary skill in the art, several changes and improvements can also be made without departing from the concept of the present invention. These all belong to the protection scope of the present invention.

[0021] Example 1 Reference Figure 1 As shown, a data transmission system for determining rocket failures by automatic rule generation includes: Data preprocessing module, intelligent model training module, rule generation and optimization module, rule application module; The data preprocessing module is used to preprocess rocket-related detection data through data cleaning and data dimensionality reduction, including data cleaning and high-dimensional data dimensionality reduction for trend prediction, making it suitable for model training; The steps for data cleaning are as follows: S1. Process missing values ​​of data by replacing missing data with appropriate values ​​through missing value replacement methods to improve data integrity. The missing value replacement methods include mean filling, median filling, maximum value filling, minimum value filling, and deletion of missing values; S2. Processing outliers on the data by replacing outliers with appropriate values ​​through an outlier replacement method to improve the consistency of the data. The outlier replacement method includes a Z-score method, an IQR method, and a data distribution method; S3, performing duplicate value processing on the data, deleting or merging duplicate data by a duplicate value processing method to improve the simplicity of the data, wherein the duplicate value processing method includes a deduplication method and a clustering method; S4, performing type conversion on the data, and realizing data conversion through a data conversion method; the data conversion method includes type conversion and format conversion; the data conversion includes integer to floating point number conversion, string to integer conversion, date format conversion, and time format conversion; S5. Normalize the data and scale the data range by a data normalization method, wherein the normalization method includes maximum and minimum normalization and Z-score normalization.

[0022] The specific method of high-dimensional data dimensionality reduction is: PCA method combined with entropy weight method, and its steps include: S1. Calculate the value of the standardized data feature ; S2. Obtaining the standardized matrix ; S3. Calculate the information entropy value for each feature j in the data set ; S4. Using information entropy Calculate feature weights n.

[0023] The intelligent model training module uses intelligent models to perform model training on preprocessed data to capture complex patterns and relationships in the data; intelligent models include natural language understanding models such as BERT.

[0024] The rule generation and optimization module uses the trained intelligent model to automatically generate rules from new data and optimizes them using genetic algorithms and simulated annealing techniques to improve their accuracy and efficiency. Specifically: S1. Use the trained model to make predictions, including loading the trained intelligent model, inputting the pre-processed new data into the model and obtaining the prediction results; S2. Use new data and prediction structures to train a decision tree model to simulate the decision-making process of the intelligent model and generate explainable rules; S3. Optimize the rules through genetic algorithms and simulated annealing techniques.

[0025] The rule application module builds a rule-based expert system based on rules, converts the system into a scripting language for operation, and obtains the fault judgment result; the rule application module is based on the rule-based expert system (RBES), which converts the rules into Python language and runs in an infinite loop. The diagnostic function consists of all rules extracted by automatic rule generation. The result of each rule is saved as an array. The diagnostic function can detect multiple faults at the same time.

[0026] The present invention can integrate and transmit the collected test data of various subsystems of the rocket, so that the operating personnel can grasp the overall status of the system, improve the efficiency of troubleshooting and the reliability of system operation, and reduce the expenditure of labor costs.

[0027] Example 2 Reference Figure 1 and Figure 2 As shown, a data transmission system for determining rocket failures by automatic rule generation includes the following modules: data preprocessing module, intelligent model training module, rule generation and optimization module, and rule application module. The implementation of each module and the judgment process in this embodiment are described in detail below: 1. Data preprocessing module In this embodiment, multiple steps are performed to achieve data preprocessing and data cleaning: S1. Process missing data and replace the missing data with appropriate values ​​to improve data integrity, including methods for processing missing data such as mean filling, median filling, maximum filling, minimum filling and deleting missing values.

[0028] S2. Process abnormal data and replace abnormal values ​​with appropriate values ​​to improve data consistency, including methods for processing abnormal data such as Z-score method, IQR method and data distribution.

[0029] S3. Process duplicate data by deleting or merging duplicate data to improve the simplicity of data, including methods for processing duplicate data such as deduplication and clustering.

[0030] S4, converting data types, and implementing data conversion through a data conversion method. The data conversion method includes type conversion and format conversion; the data conversion includes integer to floating point conversion, string to integer conversion, date format conversion, and time format conversion.

[0031] S5. Normalize the data by scaling the data range through a variety of methods to achieve data normalization. The normalization methods include maximum and minimum normalization and Z-score normalization.

[0032] 2. Intelligent model training module In this embodiment, the PCA method combined with the entropy weight method is used to complete the intelligent model training for automatic rule generation, and the steps include: S1. Calculate the value of the standardized data feature ; S2. Obtaining the standardized matrix ; S3. Calculate the information entropy value for each feature j in the data set ; S4. Using information entropy Calculate feature weights n, and pass it to the model for training the intelligent model.

[0033] By using natural language understanding intelligent models such as BERT, model training is performed on preprocessed data to facilitate subsequent rule generation and optimization.

[0034] 3. Rule generation and optimization module The operation steps of the rule generation and optimization module are as follows: S1. Use the trained intelligent model to make predictions, including loading the trained intelligent model, inputting the pre-processed new data into the model and obtaining the prediction results; S2. Use the obtained data and prediction results to train a decision tree model to simulate the decision-making process of the intelligent model, thereby generating explainable rules; S3. Further optimize the rules generated by the decision tree model through genetic algorithms, simulated annealing technology and other methods.

[0035] (IV) Rule application module The rules generated and optimized are applied using a rule-based expert system (RBES). The generated and optimized rules are converted into Python statements and passed into the expert system to run in an infinite loop. The diagnostic function consists of all the rules extracted by automatic rule generation. The results of each rule are saved as an array, and the array can be used to determine whether there is a fault. With the help of the loop operation and multi-result storage of the expert system, the diagnostic function can detect multiple faults at the same time. Specifically, the expert system judges the current data in sequence according to the passed-in rules, analyzes the current situation, and executes jumps according to the specific situation to obtain the fault result or make the next judgment, and finally obtains the correct fault result.

[0036] In this embodiment, for the abnormal state of the device ID M0, the current device data information is first obtained and processed, and passed to a pre-trained intelligent model such as BERT. The intelligent model analyzes the current device status data, determines the potential cause of the fault, and generates and optimizes the fault judgment rules in combination with the decision tree. After obtaining the judgment rules, they are converted into programming languages ​​such as Python statements and applied to the expert system for looping, and finally the fault judgment result is obtained in combination with the current device status.

[0037] Example 3 Reference Figure 1 and Figure 2 As shown, a data transmission method for determining rocket failure by automatic rule generation includes: Step S1: Collecting rocket monitoring data and transmitting it; Step S2: receiving rocket monitoring data and preprocessing; The preprocessing includes data cleaning and high-dimensional data dimensionality reduction for trend prediction; Step S3: Train the intelligent model based on the preprocessed rocket monitoring data to capture the correlation between the data.

[0038] Use the trained intelligent model to generate rules and optimize them; An expert system is constructed based on the optimized rules; the expert system can be converted into a script language to run and output a fault judgment result.

[0039] Data cleaning includes: Processing missing values ​​of data, replacing missing data with preset values ​​through a missing value replacement method; the missing value replacement method includes mean filling, median filling, maximum value filling, minimum value filling and deletion of missing values; Perform outlier processing on the data, and replace the outliers with preset values ​​by an outlier replacement method; the outlier replacement method includes a Z-score method, an IQR method, and a data distribution method; Processing duplicate values ​​on the data, deleting or merging duplicate data by using a duplicate value processing method; the duplicate value processing method includes a deduplication method and a clustering method; Performing type conversion on the data, and performing data conversion by a data conversion method; the data conversion method includes type conversion and format conversion; the data conversion includes integer to floating point number conversion, string to integer conversion, date format conversion and time format conversion; The data is normalized, and the range of the data is scaled by a data normalization method; the normalization method includes maximum and minimum normalization and Z-score normalization.

[0040] Dimensionality reduction of high-dimensional data includes: The PCA method based on the entropy weight method calculates the value of the standardized data features:

[0041] Get the normalized matrix: ; Calculate the information entropy value for each feature j in the data set : ; Using information entropy Calculate feature weights : n.

[0042] Those skilled in the art know that, in addition to realizing the system and its various devices, modules, and units provided by the present invention in a purely computer-readable program code, it is entirely possible to realize the same functions in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers by logically programming the method steps. Therefore, the system and its various devices, modules, and units provided by the present invention can be considered as a hardware component, and the devices, modules, and units included therein for realizing various functions can also be regarded as structures within the hardware component; the devices, modules, and units for realizing various functions can also be regarded as both software modules for realizing the method and structures within the hardware component.

[0043] In the description of the present application, it should be understood that the terms "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc., indicating orientations or positional relationships, are based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the referred device or element must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as a limitation on the present application.

[0044] The above describes the specific embodiments of the present invention. It should be understood that the present invention is not limited to the above specific embodiments, and those skilled in the art can make various changes or modifications within the scope of the claims, which does not affect the essence of the present invention. In the absence of conflict, the embodiments of the present application and the features in the embodiments can be combined with each other arbitrarily.

Claims

1. A data transmission system for determining rocket failures using automated rule generation, characterized in that: include: Data acquisition module, used to collect rocket monitoring data; A data transmission module, used to transmit rocket monitoring data to a data preprocessing module; The data preprocessing module receives and preprocesses the rocket monitoring data, and sends the preprocessed rocket monitoring data to the intelligent model training module; The preprocessing includes data cleaning and high-dimensional data dimensionality reduction for trend prediction; The data cleaning includes: Processing missing values ​​of data, replacing missing data with preset values ​​through a missing value replacement method; the missing value replacement method includes mean filling, median filling, maximum value filling, minimum value filling and deletion of missing values; Perform outlier processing on the data, and replace the outliers with preset values ​​by an outlier replacement method; the outlier replacement method includes a Z-score method, an IQR method, and a data distribution method; Processing duplicate values ​​on the data, deleting or merging duplicate data by using a duplicate value processing method; the duplicate value processing method includes a deduplication method and a clustering method; Performing type conversion on the data, and performing data conversion by a data conversion method; the data conversion method includes type conversion and format conversion; the data conversion includes integer to floating point number conversion, string to integer conversion, date format conversion and time format conversion; Normalizing the data, scaling the range of the data by a data normalization method; the normalization method includes maximum and minimum normalization and Z-score normalization; Intelligent model training module, which trains the intelligent model based on the pre-processed rocket monitoring data to capture the correlation between the data; The rule generation and optimization module uses the trained intelligent model to generate rules and perform optimization; The rule application module constructs an expert system based on the optimized rules; the expert system can be converted into a script language to run and output fault judgment results.

2. The data transmission system for determining rocket failures generated by automated rules according to claim 1 is characterized in that: The high-dimensional data dimensionality reduction includes: The PCA method based on the entropy weight method calculates the value of the standardized data features: Get the normalized matrix: ; Calculate the information entropy value for each feature j in the data set : ; Using information entropy Calculate feature weights : n。 3. The data transmission system for determining rocket failures according to the automated rule generation according to claim 1 is characterized in that: The rule generation and optimization module includes: Load the trained intelligent model, input the preprocessed data into the model and obtain the prediction results; Based on the processed data and prediction structure, the decision tree model is trained to simulate the decision-making process of the intelligent model and generate explainable rules; The rules are optimized by genetic algorithm and simulated annealing technology.

4. The data transmission system for determining rocket failures generated by automated rules according to claim 1 is characterized in that: The rule application module is based on a rule-based expert system, which converts the rules into Python language and runs them in an infinite loop. The diagnostic function consists of all rules extracted by automatic rule generation, and the result of each rule is saved as an array; the diagnostic function can detect multiple faults at the same time.

5. A data transmission method for determining rocket failure by automated rule generation, characterized in that: include: Step S1: Collecting rocket monitoring data and transmitting it; Step S2: receiving rocket monitoring data and preprocessing; The preprocessing includes data cleaning and high-dimensional data dimensionality reduction for trend prediction; The data cleaning includes: Processing missing values ​​of data, replacing missing data with preset values ​​through a missing value replacement method; the missing value replacement method includes mean filling, median filling, maximum value filling, minimum value filling and deletion of missing values; Perform outlier processing on the data, and replace the outliers with preset values ​​by an outlier replacement method; the outlier replacement method includes a Z-score method, an IQR method, and a data distribution method; Processing duplicate values ​​on the data, deleting or merging duplicate data by using a duplicate value processing method; the duplicate value processing method includes a deduplication method and a clustering method; Performing type conversion on the data, and performing data conversion by a data conversion method; the data conversion method includes type conversion and format conversion; the data conversion includes integer to floating point number conversion, string to integer conversion, date format conversion and time format conversion; Normalizing the data, scaling the range of the data by a data normalization method; the normalization method includes maximum and minimum normalization and Z-score normalization; Step S3: training the intelligent model based on the preprocessed rocket monitoring data to capture the correlation between the data; Use the trained intelligent model to generate rules and optimize them; An expert system is constructed based on the optimized rules; the expert system can be converted into a script language to run and output a fault judgment result.

6. The data transmission method for determining rocket failures generated by automated rules according to claim 5 is characterized in that: The high-dimensional data dimensionality reduction includes: The PCA method based on the entropy weight method calculates the value of the standardized data features: Get the normalized matrix: ; Calculate the information entropy value for each feature j in the data set : ; Using information entropy Calculate feature weights : n。

Citation Information

Patent Citations

  • Fault detection method used in test-launching stage of launch vehicle

    CN102722170A

  • Entropy weight method feature selection-based residential electricity load classification method

    CN106096805A

  • Electric charge income prediction method based on dynamic weight optimization

    CN116452238A