A data-driven radar health management system modeling implementation method

Abstract

The present invention proposes a data-driven radar health management system modeling implementation method. Aiming at the needs of rapid customization and continuous upgrading of health management software for different models of radars, the present invention uses configuration files or visual logic modeling tools to create health management models based on radar system analysis data, sets radar fault diagnosis basis, trend prediction algorithm, functional status evaluation and maintenance guarantee measures, and generates model files; and uses database management tools to set collection points, configure data sources and associate BIT data, so as to achieve the association between logical data in the model file and actual collected physical data, and finally uses Web or desktop health management information display software to monitor the operating status of the radar system in real time, so as to achieve integrated management and control of the health status of the radar equipment system. The present invention significantly improves the development efficiency of the radar health management software system, and is of great significance to improving the comprehensive test and diagnosis capabilities of the radar system.

Description

Technical Field

[0001] The invention belongs to the technical field of radar management. Background Art

[0002] The performance of modern radar equipment systems is constantly improving, and the composition is becoming increasingly complex. The problems exposed in fault diagnosis and maintenance assurance are becoming increasingly prominent. How to reduce safety and reliability issues caused by faults, shorten downtime, reduce maintenance costs, and improve work efficiency is one of the key issues in the design process of new radars. The concept of fault prognostics and health management (PHM) has also emerged and has been valued in fields such as aerospace. Health management technology has been deeply studied and promoted in Western countries such as the United States and the United Kingdom because it has the ability to conduct real-time assessment of current health conditions and predict the time of future faults. It is becoming an important part of the new generation of radar equipment systems.

[0003] China started late in the research of basic theories and methods for design and verification of health management systems. The "Research and Implementation of Fault Prediction and Radar Health Management System Based on Attention" proposed the design idea of ​​radar health management system based on fault prediction algorithm. The "Research on Radar Health Information Perception and Management Technology Based on Data Mining" uses data mining methods to perceive radar health status information, explains the perception technology of health information generated during the radar life cycle, and the management technology of radar health information data and maintenance activities, and designs and develops a radar health management system based on the Django framework.

[0004] During the use of radar, a huge amount of equipment operation information is generated, and a large amount of equipment failure data is accumulated. However, the utilization of these data and information is not sufficient at present. Therefore, it is necessary to use data-driven methods such as statistical models or machine learning to carry out research on radar equipment health status assessment and life prediction methods, dig out the value behind the data, and help radar equipment health management and operation and maintenance guarantee work. For different types of radars, a complete and applicable health management system modeling implementation method is given, which can realize the rapid construction and continuous upgrading of the radar health management platform, which is of great significance to improving the comprehensive test and diagnosis capabilities of radar equipment, the autonomous guarantee capabilities, and improving the system operation safety and availability. Summary of the invention

[0005] Modern radar equipment is highly complex and integrated, with a large number of components and modules, a wide variety of types, and high fault coupling. The health status of the equipment directly affects the system's continuous duty combat capability. Health management software can enhance the comprehensive test and diagnosis capabilities and autonomous support capabilities of radar equipment, improve system operation safety and availability, and reduce support costs. It is necessary to design a complete and applicable health management system modeling implementation method to achieve rapid customization and continuous upgrading of health management software for different types of radars.

[0006] In order to solve the problems existing in the prior art, such as unclear functional module division and boundaries, non-universal implementation methods of different radar models, cumbersome and inefficient upgrade iteration, etc., the present invention proposes a data-driven radar health management system modeling implementation method, which includes:

[0007] Step 1: Analyze the radar system. According to the composition and structure of the radar system, sort out the radar failure modes at different levels, the factors affecting the radar health status under different mission modes, and the corresponding maintenance and support information. At the same time, use big data feature extraction and artificial intelligence algorithms to complete the algorithm training of fault diagnosis and health prediction, and output the algorithm optimization results.

[0008] Step 2: Create a radar health management model in the form of a configuration file or a visual logic modeling tool, set the fault diagnosis basis, trend prediction algorithm, functional status evaluation and maintenance guarantee measures for the corresponding radar equipment system, subsystem and each component, and package the built system model to generate a callable model file;

[0009] Step 3: Based on the generated system health management model file, use the database management tool Navicat to set the collection points, configure the data source, and associate the BIT data, so as to associate the logical data in the constructed model file with the physical data actually collected and monitored, and complete the management of the model;

[0010] Step 4: Based on the constructed model, configuration information and algorithm, with real-time monitoring data as input, the health management information display software in Web or desktop form is used to monitor the operating status of the radar system in real time, and corresponding maintenance and guarantee decisions are made according to the system operating status to achieve integrated management and control of the health status of the radar equipment system.

[0011] The present invention has clear internal logic, can quickly customize different types of equipment, and is convenient for updating the software system, which is conducive to building a radar equipment health management system with strong stability and versatility. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Attached Figure 1 The main technical processes for radar health management modeling implementation.

[0013] Attached Figure 2 Schematic diagram of hierarchical modeling of radar health management system. DETAILED DESCRIPTION

[0014] The present invention proposes a data-driven radar health management system modeling implementation method, the preferred implementation process is as follows Figure 1 ,include:

[0015] S1: Radar system analysis: Starting from the structural composition of the radar equipment, the three levels of radar system-subsystem-each group (LRU) are determined. The analytic hierarchy process (AHP) is used to analyze the FMEA of different systems and components at each level to determine the failure mode, failure cause, failure impact, failure rate, failure level and other information of the objects at each level. At the same time, the main functions and mission information of the radar, as well as maintenance and support information are obtained. The system's diagnostic analysis results are used as the input for model design, monitoring, diagnosis, evaluation and predictive analysis of radar system health management.

[0016] S2: Model creation: Based on the analysis of the diagnostic object, the system structure tree is constructed, the original data set definition and data conversion, fault diagnosis criteria setting, trend prediction, function and task evaluation, maintenance support resource setting and model inspection and verification are carried out in sequence in the form of configuration files or visual logic modeling tools, so as to build a health management model for the equipment and generate a model file that supports calling. The preferred process includes the following 5 steps:

[0017] 1. Based on the radar equipment system analysis, create a system composition structure tree, define the levels including system, subsystem, components and bottom-level fault units, realize the hierarchical system structure creation, and describe the system composition structure and fault units accordingly according to the product design information.

[0018] 2. The original data is converted through embedded feature functions to generate model data for fault diagnosis, evaluation, and prediction, and to build the node association relationship between the fault unit and the model data.

[0019] 3. Set the fault mode and fault diagnosis basis, and select the trend fitting algorithm to fit and predict the changing trend of the model data value, and seek the best estimate for fault diagnosis and fault prediction.

[0020] 4. Create radar system functions and system tasks, and build the relationship between functions and tasks, including the addition and deletion of functions during the specific time period of task operation; create component and function evaluation models, and complete the status evaluation of system functions or components through evaluation algorithms; at the same time, define the cause of the fault, the impact of the fault, the operating measures and other information when the fault occurs, to provide information support for subsequent troubleshooting.

[0021] 5. Define maintenance units, maintenance behaviors, work cards, spare parts and other information, associate maintenance behavior information when a fault occurs, set corresponding maintenance methods and maintenance time, and provide maintenance support for users.

[0022] S3: Model management: Based on the constructed equipment system health management model, the database management tool is applied to build the association between physical data and logical data through collection point setting, data source configuration and BIT data association. At the same time, the mapping function of the constructed model file to the database is completed, and the model file is stored in the database in the form of a data table structure, which is convenient for later viewing, management and reading and calling of the health management information display software, so as to realize the background management of the system.

[0023] S4: Operation monitoring: Based on the constructed system health management model files and related data configuration information stored in the database, the data information of the database and the real-time data of the system are directly read with the help of Web or desktop health management information display software to complete real-time monitoring and fault diagnosis. At the same time, the health status of the system is evaluated, the life prediction is completed, and a maintenance plan is given in combination with maintenance resources to achieve integrated comprehensive integrated management and control.

Claims

1. A data-driven radar health management system modeling implementation method, characterized by: Step 1: Analyze the radar system. According to the composition and structure of the radar system, sort out the radar failure modes at different levels, the factors affecting the radar health status under different mission modes, and the corresponding maintenance and support information. At the same time, use big data feature extraction and artificial intelligence algorithms to complete the algorithm training of fault diagnosis and health prediction, and output the algorithm optimization results. Step 2: Create a radar health management model in the form of a configuration file or a visual logic modeling tool, set the fault diagnosis basis, trend prediction algorithm, functional status evaluation and maintenance guarantee measures for the corresponding radar equipment system, subsystem and each component, and package the built system model to generate a callable model file; Step 3: Based on the generated system health management model file, use the database management tool Navicat to set the collection points, configure the data source, and associate the BIT data, so as to associate the logical data in the constructed model file with the physical data actually collected and monitored, and complete the management of the model; Step 4: Based on the constructed model, configuration information and algorithm, with real-time monitoring data as input, the health management information display software in Web or desktop form is used to monitor the operating status of the radar system in real time, and corresponding maintenance and guarantee decisions are made according to the system operating status to achieve integrated management and control of the health status of the radar equipment system.

2. A data-driven radar health management system modeling implementation method according to claim 1, characterized in that: The radar system analysis process in step 1 includes: starting from the structural composition of the radar equipment, determining the three levels of radar system-subsystem-groups, using the hierarchical analysis method to analyze the FMEA of different systems and components at each level, determining the failure mode, failure cause, failure impact, failure rate and failure level information of objects at each level, and obtaining its main function and task information, maintenance and support information at the same time, using the diagnostic analysis results of the radar system as the input for the design of the system health management model.

3. The data-driven radar health management system modeling implementation method according to claim 1 is characterized by: The process of creating the radar health management model in step 2 includes: Step 2-1: Based on the radar equipment system analysis, create a system composition structure tree, define the hierarchy including system, subsystem, component and bottom fault unit, realize the hierarchical system structure creation, and describe the system composition structure and fault unit accordingly according to the product design information; Step 2-2: convert the original data through the embedded characteristic function, generate model data for fault diagnosis, evaluation and prediction, and build the node association relationship between the fault unit and the model data; Step 2-3: Set the fault mode and fault diagnosis basis, and build the fault diagnosis and trend fitting algorithm components for fault diagnosis and fault prediction; Step 2-4: Create an evaluation model of component status and system function based on the relationship between system functions and tasks, and define the cause and impact of a fault when it occurs, to provide information support for subsequent troubleshooting; Step 2-5: Define maintenance units, maintenance behaviors, work cards, and spare parts, associate maintenance behavior information when a fault occurs, set corresponding maintenance methods and maintenance duration, and provide maintenance support for users.

4. The data-driven radar health management system modeling implementation method according to claim 1 is characterized by: The management of the radar health management model includes: based on the constructed equipment system radar health management model, applying database management tools, building the association between physical data and logical data through collection point setting, data source configuration and BIT data association, and completing the mapping function of the constructed model file to the database, storing the model file in the database in the form of a data table structure, and later viewing, management and health management information display software reading and calling.

5. The data-driven radar health management system modeling implementation method according to claim 1 is characterized by: The real-time monitoring and display in step 4 includes: based on the constructed radar system health management model file and related data configuration information stored in the database, directly reading the data information in the database and the real-time data of the system with the help of health management information display software in the form of Web or desktop, completing real-time monitoring and fault diagnosis, and at the same time, evaluating the health status of the system according to the optimal status evaluation algorithm model, completing life prediction, and providing a maintenance plan in combination with maintenance resources to achieve integrated comprehensive integrated management and control.

Images