CN110543489A

CN110543489A - Solid rocket engine reliability data analysis mining and application software tool

Info

Publication number: CN110543489A
Application number: CN201910821059.XA
Authority: CN
Inventors: 魏龙; 宋晓茜; 马群; 刘伟; 刘鑫生; 张富强; 沈杰
Original assignee: Inner Mongolia Power Machinery Research Institute
Current assignee: Inner Mongolia Power Machinery Research Institute
Priority date: 2019-09-02
Filing date: 2019-09-02
Publication date: 2019-12-06
Anticipated expiration: 2039-09-02
Also published as: CN110543489B

Abstract

The invention relates to a solid rocket engine reliability data analysis mining and application software tool, which is divided into three layers of a basic platform layer, a data management layer and a data application layer, wherein the basic platform is a server layer and comprises an application server, a database server, a file server and an integrated interface; the data management layer is a database layer and has a data management function on basic information of the engine, reliability data of charging and internal insulation, reliability data of the spray pipe, reliability data of the shell and external insulation, and reliability data of the whole engine and direct parts; the data application layer is a user layer and is used for data query, reliability data analysis, reliability data two-dimensional graph display and reliability data three-dimensional graph display. The method provides data support for the prejudgment of the quality of the engine product, further improves the stability and consistency of the product quality, and lays a foundation for improving the capabilities of the solid rocket engine in various aspects such as fine management, accurate manufacturing, quality guarantee, informatization management and the like.

Description

Solid rocket engine reliability data analysis mining and application software tool

Technical Field

The invention relates to the technical field of solid rocket engines, in particular to a solid rocket engine reliability data analysis mining and application software tool.

Background

The reliability data is used as the source of product quality information, reflects the real-time quality condition of the product, is the premise and the basis of the reliability management of modern enterprises, and the work of analyzing and mining the reliability data becomes the basis and the key for effectively carrying out various reliability management activities of the enterprises. The reliability data statistical analysis can scientifically and accurately analyze and predict the current and future quality conditions of a product or a process, the quality condition of the product is controlled to be kept in a stable and controlled statistical equilibrium state, the aim of controlling and improving the product quality is finally achieved, and data analysis and mining are the basis and important means for reflecting quality control through data.

The collection of traditional enterprise reliability data generally adopts manual record mode, and this kind of mode collection speed is slow, and the real-time is poor, and the precision is low moreover, is unfavorable for product quality's real-time dynamic management and control. To solve these problems, foreign researchers have conducted a great deal of research on the theory of reliable data collection and its application. With the continuous development and deepening of computer technology, the computer technology is popularized to the aspect of reliability data acquisition, and the reliability data acquisition theory and technology are greatly developed. A Krenzel investigated a computer-aided acquisition system to assist in the analysis of crisis conditions. Billo RE et al studied an automatic data acquisition system based on code technology. The German BEISSBARTH four-wheel aligner, the full-automatic headlamp detector and the like have the advantages of high detection speed, high detection precision, function integration and the like, and effectively ensure the quality of products of modern enterprises. At present, many researchers in foreign countries have conducted intensive research on integration, intelligence and networking of the reliability data acquisition technology.

During world war II, the statistical method rate is widely applied in the quality management of American, English and other national enterprises. At that time, in order to ensure the quality of a large number of military supplies, the U.S. military requires that all military product manufacturers must popularize statistical quality control methods, and establishes national standards such as "ASAZ 1.1 quality control guidelines", "ASAZ 1.2 control graph method for data analysis", and the like. At the same time, the uk has also worked out "statistical methods in BS600 industrial standardization and quality management" national standards. In order to meet the challenge of high quality and cost of products from japanese enterprises in the end of the 20 th century and the 80 th century, the U.S. motorola company has created a famous sigma reliability management concept and a corresponding reliability management system, and by adopting a modern advanced statistical quality analysis control method, the quality level of the products of the company is ensured to reach the six sigma standard, so that the competitiveness of the enterprises in the international market is greatly improved. In addition, the data analysis mining bears a great deal of valuable knowledge information such as product structure composition, variety, key reliability data indexes and the like, so the method is also a basis for supporting scientific research and is an important component of core competitiveness. For the above reasons, data analysis mining is increasingly being emphasized.

Compared with developed countries, the technological level and the reliability data analysis mining and application level of the spaceflight in China are still in gap, and the gap is large especially in the aspect of basic conditions. China proposed an idea of constructing a national science and technology basic condition platform in 2002, and is supported by related departments and widely praised by the science and technology field. Through serious investigation and research, advanced experiences at home and abroad are summarized, and a framework and a development target of national science and technology basic condition platform construction are creatively provided. In order to effectively manage and analyze product reliability data and fully mine potential values in the data in the face of massive product reliability data, domestic colleges and universities, enterprises and scientific research institutions research the management, analysis and application technologies of the product reliability data in different industries, and a series of research results are obtained.

Yang and the like of Beijing Shenzhou space software technology Limited company promotes the collection and management of reliability data of space model products, provides a model product reliability data packet construction and management target, establishes a space model data packet, can realize the recording, transmission, processing and utilization of the reliability data of the model products, is beneficial to the development of product engineering, improves the analysis, evaluation and improvement capacity of the product data, and realizes the tracing and application of the reliability data of the whole life cycle of the product model. Various reliability data generated in the whole life cycle process of the aero-engine are researched by Hover and Liqiang of northwest industrial university, data information is classified and analyzed based on the constituent elements of the engine, quality information is divided into product types, component types and part types, an aero-engine reliability data information system structure is established, and integrated management of the aero-engine reliability data from collection, classification and transmission to sharing is achieved. Aiming at the problems of lack of unified data sources, difficult data tracking, poor design quality evaluation means and the like in the traditional reliability data management of radar structures, the thirty-eight institute Chen Jiang and the like of China electronic science and technology group company, a radar equipment reliability data classification and system construction method and a reliability data tracking and transmission method based on multi-level and multivariate analysis are analyzed and researched, a radar structure design quality comprehensive evaluation method is provided, a prototype system supporting radar equipment design quality evaluation and improved design and an integrated reliability data model of key parts of radar equipment are established, and the improved design of the radar equipment structure is guided, so that the aims of improving the design quality, reducing the maintenance cost and improving the quality control capability are fulfilled. The method for integrating the reliability data facing to the whole life cycle of the product is researched by Beijing aerospace university Zhengyao and WangMeiqing, a product whole life cycle reliability data mapping semantic library and an integration mapping model are constructed based on the ideas of data extraction, conversion and transfer (ETL), on the basis, an integration subsystem QQ-DI facing to the product whole life cycle reliability data is designed and developed, and for heterogeneous reliability data from the whole life cycle of the product, the system can realize statistical analysis and tracing facing to the product structure and the organization structure and automatic generation of a reliability data packet. Through analysis of research on product reliability data analysis mining and application of domestic colleges and universities, enterprises and scientific research institutions, the research on reliability data analysis and application technology is developed in the research and development process of various industrial products in China, corresponding reliability data models, reliability data analysis and evaluation systems, system architectures and the like are preliminarily constructed, and effective support is provided for quality characteristics and reliability of corresponding products. However, in the field of the aerospace power system, because the special data of the engine is more and the attention degree on the related reliability data is not enough, the analysis and application technical research on the reliability data of the corresponding product is not basically carried out at present, so that the research on the reliability data analysis and mining technology of the solid rocket engine is urgently needed to be carried out.

During the development process of the solid rocket engine, a large amount of reliability data is accumulated. However, due to the fact that the special data in the engine reliability data are many, the pertinence is strong, and the particularity of the solid engine is added, the existing data general data analysis mining method is not suitable for being used for the solid rocket engine, so that the existing engine reliability data are various, complex in structure, scattered in data, heterogeneous in multisource, idle in a large amount, and mixed in high-value data and low-value data, and difficult to distinguish; and the data cannot be deeply analyzed and processed. Meanwhile, data analysis and mining are the basis and prerequisite conditions of data engineering application, and the existing engine data application level is low due to the fact that no good data analysis and mining software tool exists, the actual problem of the engine cannot be solved by data guidance, and the due value of the data is not brought into play. On one hand, the problems seriously hinder the rapid development of the informatization technology and the digitization technology of the solid rocket engine, and on the other hand, the problems result in low reliability data utilization rate of the solid rocket engine and are not beneficial to the quality and reliability evaluation of the solid rocket engine.

Disclosure of Invention

The invention aims to solve the technical problem of providing a solid rocket engine reliability data analysis mining and application software tool to solve the problem of realizing engine reliability data management and analysis mining on the basis of a solid rocket engine reliability data system.

In order to solve the existing technical problems, the technical scheme adopted by the invention is as follows: a solid rocket engine reliability data analysis mining and application software tool is divided into three layers of a basic platform layer, a data management layer and a data application layer, wherein the basic platform is a server layer and comprises an application server, a database server, a file server and an integrated interface, and platform functions of tree definition user management, authority setting, log management and parameter setting are provided; the data management layer is a database layer and provides data management functions of importing, exporting and editing basic information of the engine, reliability data of charging and internal insulation, reliability data of the spray pipe, reliability data of the shell and external insulation and reliability data of the complete engine and direct parts; the data application layer is a user layer, is a visual tool and an operation interface which are interacted with a user, and is used for the reliability data application functions of data query, reliability data analysis, reliability data two-dimensional graph display and reliability data three-dimensional graph display.

particularly, the reliability data of the charging and the internal insulation, the reliability data of the spray pipe, the reliability data of the shell and the external insulation and the reliability data of the complete engine and the direct parts can be further divided into product data, process data and fault data according to the life cycle and the application,

the product data is divided into engine design research and development data, production manufacturing data and test data, the engine design research and development data system is divided into engine basic data, design data, temporary technical data and calculation data, the basic data is divided into engine structure index data, engine performance index data, main size data of the whole engine and parts of the engine, engine performance parameter data, engine working environment data, engine matching data, part assembly replacement data, history data of the whole engine and parts of the engine, data labels and special release data of the over-tolerance substitute materials, the design data is divided into engine pattern data, engine design task drawing data, technical condition data, development summary data, development transition evaluation data and characteristic analysis data, the temporary technical data is divided into technical problem processing data, temporary task working data, development transition evaluation data and characteristic analysis data, Changing parameter data and changing production data, wherein the calculation data is divided into conventional calculation data, simulation calculation data and other calculation data; the production and manufacturing data are divided into engine recorded data and process data, the engine recorded data are divided into component measurement data, procedure process recorded data and final assembly test recorded data, and the process data are divided into process file data, process report data, process summary data, process analysis data and key process data; the test data comprises test data of an engine development process, outfield test data and other types of test data, the test data of the engine development process comprises part and part assembly test scheme data, part and part assembly test record data, part and part assembly test result analysis data, complete machine test scheme data, complete machine test record data, complete machine test result analysis data and the like, the outfield test data comprises outfield flight test data and outfield refurbishment test data, and the other types of test data are improvement measure verification test data;

The process data is divided into engine comprehensive guarantee data and health monitoring detection data, the engine comprehensive guarantee data is divided into guarantee data, review and reexamination data, outsourcing data, material use data and scientific and technological achievement data, and the health monitoring detection data is divided into engine performance parameter extreme value data, engine structure parameter extreme value data, engine health monitoring data and engine nondestructive detection data;

The fault data comprises typical fault and analysis data of the engine and data for improving the quality of products, wherein the typical fault and analysis data comprises typical fault data, fault analysis data and product quality improvement data of the engine, and the data for improving the quality of the engine comprises data for improving the whole engine and data for improving the quality of the engine, data for improving the.

particularly, the data general functions of the data management function and the data application function of the invention can also comprise data import, data export, data display, data query retrieval, data analysis, addition, editing and deletion, wherein the data import comprises a data import template and a data import verification function; the data display comprises BOM tree display and list display functions; the data analysis is the analysis of the packaged reliability data and comprises mean value method packaging, minimum value method packaging, maximum value method packaging, power method packaging, evolution method packaging, fitting method packaging and interpolation method packaging.

In particular, the reliability data analysis of the present invention may further include mathematical operation analysis, data statistical analysis, envelope analysis, regression analysis, and neural network algorithms.

particularly, the data application layer of the invention can also comprise a reliability index alarm function, and the reliability index alarm function is divided into reliability index definition and risk early warning tree.

Particularly, the solid rocket engine reliability data analysis mining and application software tool can further comprise algorithm plug-and-play support, and the algorithm plug-and-play support has an editing management function on a Python algorithm, parameters of the Python algorithm and return values of the Python algorithm.

Particularly, the solid rocket engine reliability data analysis, mining and application software tool can further comprise data security management, wherein the data security management comprises login management, access security control, data domain security setting and system administrator permission adoption of 'three-person management'.

The invention aims at the problems of various reliable data storage types, dispersed storage positions, large data volume, irregular management, low utilization rate, incapability of providing effective support for engine research and development and reliability improvement and the like in the whole life cycle of a solid engine, and develops engine reliability data analysis, mining and application technical researches.

the invention is based on a solid engine reliability data system, aims at realizing the management and analysis and mining of the engine reliability data, and can perform the functions of data storage and management of the engine reliability data, data visualization display, reliability index early warning, algorithm plug and play, data dynamic tabulation, data analysis and mining and the like.

The method takes the reliability data of the solid rocket engine as a research object, takes the problem of solving the actual engineering of the engine as a main line, constructs a set of highly-targeted engine reliability data analysis mining method and software tool, establishes an engine reliability database, realizes visualization of engine data, deeply mines the internal relation and the potential law of various reliability data of the engine, provides data support for prejudging the quality of engine products, further improves the stability and consistency of the product quality, and lays a foundation for improving the capabilities of the solid rocket engine in various aspects such as fine management, accurate manufacturing, quality guarantee, information management and the like.

Advantageous effects

The invention aims at the problems of various storage types, dispersed storage positions, large data volume, non-standard management, low utilization rate, incapability of providing effective support for engine research and development and reliability improvement and the like of the reliability data of the solid engine, and carries out the research on the engine reliability data analysis mining technology and the development of a software tool thereof. The patent provides a set of software tools including engine reliability data management, analysis, mining and application, and compared with the conventional general data analysis and mining software tools, the software tools have strong pertinence; the method covers all the contents from research to engineering application, such as classification construction of an engine reliability data system, a data analysis and mining method, data application, software function module development and the like, and is friendly to workers in a progressive mode; meanwhile, the software of the invention also comprises the functions of data universal function, data dynamic modeling, data analysis and mining method, data visual display, reliability index alarm, algorithm plug and play, data safety design and the like, and is in accordance with the current mainstream digital research and development mode, thereby laying a very favorable foundation for the subsequent data application work.

Drawings

FIG. 1 is a schematic diagram of a general scheme for analyzing, mining and applying research on reliability data of a solid engine;

FIG. 2 is a schematic diagram of data type and usage classification;

FIG. 3 is a schematic diagram of a reliability data analysis and mining technology study;

FIG. 4 is a schematic diagram of a reliability data visualization display technology study;

FIG. 5 is a diagram illustrating mathematical operations;

FIG. 6Python algorithm support diagram;

FIG. 7 is a schematic diagram of a reliability data analysis mining and application software system architecture;

FIG. 8 is a schematic diagram of a three-level physical architecture of a reliability data analysis mining software tool;

FIG. 9 is a diagram of a software data management page;

FIG. 10 illustrates an interface diagram of a new master table of the software;

FIG. 11 is an edit and delete interface diagram of the main table;

FIG. 12 is a two-dimensional diagram showing an interface diagram;

FIG. 13 is a three-dimensional diagram showing interface view one;

FIG. 14 is a three-dimensional view of interface map two;

FIG. 15 is a reliability early warning indicator definition interface diagram;

FIG. 16 is a schematic view of a risk early warning tree;

FIG. 17 is an algorithm customization and import interface diagram;

FIG. 18 is an audit log viewing interface diagram;

FIG. 19 is a schematic diagram of approval of data import and export business processes.

Detailed Description

The invention is further described below with reference to the accompanying drawings and examples.

The main content of the invention is an engine reliability data analysis mining software tool and the visualization thereof, and the main construction idea of the overall scheme of the solid engine reliability data analysis mining and application research is shown in figure 1. The main development process comprises the following steps: dividing the life cycle of the engine into different stages of design, research and development, production, manufacture, test, after-sale use and the like according to the actual state of the engine, determining the type of reliability data in each stage, and inducing and carding the reliability data; for different types of reliability data, reliability data management method researches in different modes based on files, database data transmission, distributed storage and the like are carried out, distributed storage and management of the reliability data are realized, quality control is carried out on the transmitted reliability data, error reliability data are detected, repeated reliability data are removed, and the like; by utilizing the stored reliability data, the research on the analysis and mining technology of the reliability data is developed, and the analysis and processing capacity of the reliability data of different types is improved; a method for safely managing and controlling the access of all data in the architecture by adopting a network or system physical isolation mode is researched, and a top-level design method from top to bottom is adopted to carry out top-level design on reliability data analysis and application architecture in the whole life cycle of the solid engine, so that an engine reliability data analysis, mining and application software tool is formed.

1 reliability data Classification and architecture

In recent years, with the improvement of quality management level and the deepening of understanding of the quality of missile weaponry, new requirements on the construction of reliability data resources are put forward under new situations. The classification of the reliability data of the solid rocket engine and the construction of a system are carried out, on one hand, the urgent requirements of quality situation analysis and equipment quality comprehensive evaluation in the missile weapon equipment industry are met, the consistency of an analysis and evaluation result and the actual performance of equipment is improved, and the scientificity of quality management decision is ensured; on the other hand, reliability data are fed back to research and development processes such as design, production, test and the like, and the reliability data are used for improving the product quality and improving the research and development level.

The project adopts a 'fractal structure and type' method to carry out data system construction on the reliability data of the solid engine. The 'structure division' is to classify the engine reliability data according to the product structure (generally, the structure can be divided into the engine body, the combustion chamber, the nozzle, the safety ignition device, the thrust termination mechanism, the self-destruction device, the ground equipment and accessories and the like). The classification is to divide the reliability data of the engine into product data, process data and fault data according to the life cycle and the application (the product data is directly related to the product quality and comprises design, research, development, production, manufacturing, test data and the like; the process data mainly comprises comprehensive guarantee data and product health monitoring data; and the fault data is related to the quality improvement of the product and mainly comprises typical fault and analysis data of the engine, one against three data and the like).

The engine can be divided into an engine body, a combustion chamber, a spray pipe, a safe ignition device, a thrust termination mechanism, a self-destruction device, ground equipment, accessories and the like according to the product structure.

the engine reliability data is divided into product data, process data and fault data according to life cycle and use. The engine reliability data system is classified by data type and use as shown in fig. 2.

(1) Product data

The product data includes engine design development data, manufacturing data, and test data.

1) solid rocket engine design research and development data

The engine design research and development data system comprises engine basic data, design data, temporary technical data and calculation data.

The basic data is the basis of the work of engine design, production, renovation and modification, field test and the like, whether basic data is comprehensive or not, whether query is convenient or not and whether data is accurate or not are related to the aspect of daily work of the engine. The basic data of the engine comprises data of structural index of the engine, data of performance index of the engine, data of main sizes of the whole engine and parts of the engine, data of performance parameters of the engine, data of working environment of the engine, data of matching of the engine, data of replacement of parts of the parts, data of history of the whole engine and parts of the engine, data labels, data of special conditions such as the existence of out-of-tolerance substitute special release and the like.

The design data is the main basis of the work of engine design, use, maintenance and the like, and is also one of the important contents of the engine research and development data system. The design data comprises the aspects of engine drawing pattern data, engine design task data, technical condition data, development summary data, development transition evaluation data, characteristic analysis data and the like.

the temporary technical data is additionally added technical data for ensuring the smooth operation of tasks in the production and use processes of the engine, and is an important supplement to design class data. The temporary technical data comprises technical problem processing data (unqualified product processing sheets, questioning sheets and other problem processing files), temporary task working data (technical notice sheet data, scheduling notice sheet data, process notice sheets and the like), parameter changing data, production changing data and the like.

The calculated data is also one of the important contents of an engine design research and development data system, and the data plays an important role in the processes of engine scheme design, batch use, simulation modeling and the like. The calculation data comprises conventional calculation data (calculation data of parameters of the engine and the part components, calculation results, calculation programs and calculation reports), simulation calculation data and other calculation data.

2) Manufacturing data for solid rocket engine

the production and manufacturing data system of the solid rocket engine comprises engine record data and process data.

The recorded data refers to data extracted from original records in the process of engine production and manufacturing. The record data comprises the aspects of component measurement data, procedure process record data, final assembly test record data and the like.

the process data is one of important contents of an engine production and manufacturing data system, and plays an important role in the aspects of engine manufacturing, mass production and the like. The process data includes process file data, process report data, process summary data, process analysis data, key process data and the like.

3) Solid rocket engine test data

The solid rocket engine test data system comprises engine development process test data, outfield test data and other types of test data.

the test data of the machine development process comprises part (assembly) test scheme data, part (assembly) test record data, part (assembly) test result analysis data, complete machine test scheme data, complete machine test record data, complete machine test result analysis data and the like.

The outfield test data comprises outfield flight test data (review before entry, detection before flight, data processing after flight), outfield refurbishment test data (decomposition work, detection work and reloading work) and the like.

Other types of test data would include: and improving measures to verify the contents of the test data and the like.

(2) Process data architecture

The process data comprises engine comprehensive guarantee data and health monitoring and detecting data

the engine will use different life cycle data at different stages of the life cycle to evaluate the actual state of the engine. The engine quality data comprises the aspects of guarantee data, review and review data, outsourcing and outsourcing data, material use data, scientific and technical result data and the like.

the diagnosis data comprises the extreme value data of the engine performance parameters, the extreme value data of the engine structure parameters, the engine health monitoring data, the nondestructive testing data of the engine (an ultrasonic flaw detection film, an accelerator flaw detection film) and the like.

(3) Fault data architecture

The fault data system comprises typical fault and analysis data of the engine and three data.

typical fault and analysis data includes: typical engine fault data, fault analysis data, product quality improvement data, and the like.

the three-way data comprises: the engine complete machine has three data, the engine combustion chamber has three data, the engine spray pipe has three data, the engine shell has three data, the engine safety ignition device has three data, etc.

2 reliability data analysis and mining technique

The analysis and the mining of the reliability data of the solid rocket engine are the key points of reliability data processing, a large amount of reliability data have no practical significance, and the data can only play a role by analyzing the data aiming at a specific application state and converting the data into a useful result. Due to the fact that reliability data of the solid rocket engine are various and multi-source heterogeneous, the traditional data analysis method is difficult to adapt to the requirements of engine reliability data analysis and application. On the basis of the research on reliability data acquisition, storage and management technologies, the method effectively improves the capability of mining the internal connection and potential rules of the reliability data of different types of the engine by developing the research on data analysis and mining methods such as classification clustering analysis, regression analysis, association rule analysis and the like, and provides effective data support for data application problems such as product performance trend analysis, quality early warning, fault difference troubleshooting and the like. The reliability data analysis and mining technique study is shown in fig. 3.

(1) research on conventional mathematical analysis and mining technology of reliability data

The conventional mathematical analysis method for the engine reliability data is to analyze the engine reliability data by utilizing a relatively mature mathematical method, the analysis method is a type of analysis and mining method which is most used, most in demand and most widely applied at present, and the conventional mathematical analysis method comprises contents of mathematical operation analysis, data statistical analysis, two-dimensional graph analysis, envelope analysis, deviation analysis and the like. Through the conventional mathematical analysis mining technology, a large amount of simple and visual data can be provided for the aspects of engine design, production, manufacture, test, report compiling and the like, and data support is provided for the smooth daily development work of the engine.

(2) research of special reliability data analysis and mining technology

In the general concept of the reliability data, the special reliability data and the general reliability data are respectively provided, and for the solid rocket engine, the key data influencing the quality of the solid rocket engine are basically engine special data (such as heat insulation layer bonding data, grain structure integrity data, propellant mechanical property data and the like). Therefore, research on the engine-specific reliability data analysis mining technology is an important part of the present subject. The method for analyzing and mining the reliability data special for the engine combines the characteristics of the engine, carries out the analysis and mining of the specific forming process and the specific professional data of the engine, and summarizes and continues the traditional reliability data analysis algorithm special for the engine.

(3) Research on classification, clustering and analysis and mining technology of reliability data

The classification and clustering analysis method of the engine reliability data is to find out common features and different points of one group or a plurality of groups of data in a database, and to classify the data into a plurality of categories according to similarity and difference, and aims to ensure that the similarity between the data in the same category is as large as possible and the similarity between the data in different categories is as small as possible. The analysis and mining method can be applied to the aspects of engine batch management, general problem analysis, difference search of faults or problems and the like.

(4) Reliability data association rule analysis and mining technology research

The engine reliability data association rule analysis method is that certain items in one data set can cause other items to appear in the same data set according to the appearance of the certain items, and the purpose is to find out the association and the interrelation hidden in the data. By the analysis and mining technology, key factors influencing different working processes of the engine can be found out from a large amount of data, and data support is provided for engine product positioning, test risk analysis and production flight decision.

3 reliable data application

(1) reliability data early warning and evaluation technology research

and (3) obtaining the enveloping ranges of the material characteristics, the structure size, the main performance characterization parameters of the key components and the like of the engine through a reliability data enveloping analysis function, thereby giving out the engine reliability judgment criterion. On the basis, by using the functions of mathematical operation analysis, data statistical analysis, two-dimensional graph analysis, envelope analysis, deviation analysis, data classification clustering analysis, regression analysis, association rule analysis and the like, the variation trend and fluctuation rule of the reliability data are obtained, the quality state of the engine is predicted according to the quality judgment criterion, and a quality fluctuation early warning signal is given, so that powerful technical support is provided for improving the quality reliability, stability and consistency of the solid engine product, corresponding reference and basis are provided for improving the capacity of designing and manufacturing the solid engine, and the problems of integration and applicability of reliability data service, integration and analysis such as product performance trend analysis, quality early warning, fault difference investigation and the like of the solid engine are solved.

(2) Reliability data visualization technical study

The reliability data of the whole life cycle of the solid rocket engine are mined and analyzed, and the data are displayed in a specific mode, so that a user can conveniently, intuitively and quickly acquire the internal relation, the potential law and the like of the engine data. For traditional structured data, data can be represented in the forms of direct data value display, data table display, various statistical graph display and the like, while non-structured data is difficult to represent due to the fact that the traditional display method is various in types, heterogeneous in multiple sources and complex in relation, a large number of data tables and a messy relation graph can make users feel lost, and even users can be misled. By developing data visual display technical researches such as computer graphics and image processing, displaying a conventional data two-dimensional graph, and displaying the change trend of the internal unit of the engine model through color change, the capabilities of product performance trend analysis, quality early warning, fault difference troubleshooting and the like can be effectively improved, and data support is provided for solving the technical problems encountered in the design, production, test, storage and flight of the solid rocket engine. The reliability data visualization presents a technical study as shown in fig. 1.

4 engine reliability data management, analysis and application tool design

By integrating several key technologies such as integrating and condensing reliability data classification, reliability data storage and management, reliability data analysis and mining, and deep application of reliability data, researches such as solid engine reliability data management, analysis and application platform system architecture, functional module design and customization are systematically developed, a software platform capable of effectively supporting solid engine reliability data application is constructed, reliability data application research is developed on the basis, and scientificity, effectiveness and rationality of the subject result are comprehensively verified.

(1) Operating environment

TABLE 1 System hardware

TABLE 2 System software

(2) System architecture

The system architecture of the engine reliability data analysis and application is divided into: the basic platform layer, the data management layer and the data application layer are specifically explained as follows:

A basic platform: a management software development platform (origin. sun) which provides platform functions of tree definition user management, authority setting, log management, parameter setting and the like;

The data management layer is used for providing data management functions of importing, exporting, editing and the like for basic information of the engine, reliability data of charging and internal insulation, reliability data of the spray pipe, reliability data of the shell and external insulation, reliability data of the complete engine and direct parts and the like;

A data application layer: the reliability data application functions specifically comprise functions of data query, reliability data analysis, reliability data two-dimensional graph display, reliability data three-dimensional graph display and the like.

The architecture of the system is a mainstream three-layer architecture based on B/S, the three-layer architecture can well realize an MCV (Model/Controller/View) design mode advocated in software engineering, namely, a database layer is used for realizing data storage (Model), a server layer is used for realizing service logic and service flow (Controller), and a user layer is used for realizing a data display function (View).

The three-layer structure can be respectively operated on different computers.

The database layer adopts an international popular Oracle relational database. The Oracle database has strong data storage and query capabilities, is suitable for mass data management, and has good stability and expandability;

The server layer mainly comprises an application server, a database server, a file server, an integrated interface and the like;

The user layer is mainly a visualization tool and an operation interface for interaction with a user. See in particular fig. 7-8.

(3) Data general function design

Data import

Importing data into a template;

And performing customized development according to the provided data template, analyzing the data template, and importing data.

Data import check;

And data verification is carried out during data import, if the problem exists in the template imported at this time, if a certain line of missing data exists, the import is rolled back, and the number of the problem lines is printed and reported in the page.

Data export

The system supports exporting the reliability test data stored in the system in the forms of EXCEL, WORD and the like, and stores the reliability test data to the local, and the function only has the authority of a data administrator and has no export authority of other people.

data presentation

displaying a BOM tree;

The system is used for displaying data navigation in a multidimensional way according to the situation of the BOM tree, and is convenient for a user to check, including product codes, pattern codes and the like.

displaying the list;

The reliability test data imported into the system will be shown in the tabulated situation.

data query retrieval

Ordinary inquiry;

The system provides a query area in the reliability data management page, displays the query conditions defined in the background in the area, inputs query contents, and can query by clicking a [ query ] button.

advanced querying;

The switching can be performed in the query area by clicking [ advanced query ], and in the advanced query, a user can select a column to be queried and configure a complex query expression. After configuration is finished, clicking the (inquiry) button to inquire. The configured query condition information system can be reserved, and the next query is facilitated.

Data analysis

The system provides a packaged reliability data analysis method, which comprises the following steps:

Packaging by a mean value method;

Selecting a certain data column in a data management page, clicking a mean value packaging method, and automatically calculating the mean value of the data column by the system;

Packaging by a minimum value method;

Selecting a certain data column in a data management page, clicking a minimum value packaging method, and automatically calculating the minimum value of the data column by the system;

Packaging by a maximum value method;

selecting a certain data column in a data management page, clicking a maximum packaging method, and automatically calculating the minimum value of the data column by the system;

Packaging by a power method;

And selecting certain data, clicking a packaging method (power), and automatically calculating the power result of the data by the system.

packaging by using a squaring method;

Selecting certain data, clicking the packaging method, and automatically calculating the opening result of the data by the system.

Packaging by a fitting method;

And selecting certain row of data, firstly generating a line graph by the system, packaging by a click (fitting) method, and displaying the fitted line graph by the system.

Packaging by an interpolation method;

selecting a certain row of data, firstly generating a line graph by the system, clicking a certain point on the line to perform interpolation operation, packaging by a click (interpolation) method, inputting a value to be interpolated, and displaying the interpolated line graph by the system.

Adding new

The information list interface can check the applied filing reservation information; clicking (detailed) to check application detailed information and audit conditions; inputting query conditions, and clicking and screening;

Editing

The user can edit and modify the imported reliability data, select a certain piece of reliability data in a page, and click the [ edit ] button to modify the piece of data. The function is controlled by the authority, only a data manager has an editing function, other people do not have editing and modifying authority, the operation records can be kept in the log, and system management personnel can check the log.

deleting

the user can delete the imported reliability data, check a certain piece of reliability data in the page, and click the [ delete ] button to delete the piece of data. The function is controlled by the authority, only a data administrator has a deleting function, other people do not have deleting authority, the operation records can be kept in the log, and system management personnel can check the log.

(4) Dynamic modeling of data

And a data dynamic modeling (dynamic data table building) function is supported.

Dynamic table building

under the query condition, a series of input boxes appear, and the existing tables can be queried by clicking the query after corresponding data is input. Clicking an 'add' button to enter an interface of an added table on a self-defined table list page:

adding fields

Clicking an 'add column' button to add fields to the data table, and saving after finishing the operation.

editing and deleting field information

And selecting the fields needing to be edited, carrying out double-click editing operation, selecting and clicking a delete button if the fields need to be deleted, and clicking a save button to save after the operation is finished.

(5) data analysis and mining method

analysis of mathematical operations

Mathematical operational analysis includes data analysis from the simplest and most basic functions (e.g., mean, maximum, minimum, power, square, sine, cosine, etc.) to complex functions (e.g., interpolation, differentiation, integration, convolution). The mathematical operation is shown in figure 5.

Statistical analysis of data

the data statistical analysis comprises descriptive statistical analysis and hypothesis testing analysis.

Descriptive statistical analysis includes maximum, minimum, median, sum, and common variance and standard deviation. Hypothesis testing includes single sample t testing, multiple sample t testing, variance chi-squared testing.

Envelope analysis

For a large amount of complex data, besides the methods of extreme value comparison, algorithm formula and the like, the method can also collect historical data for envelope analysis, especially for military products with higher requirements on reliability and service life, grasp the development trend of the electrical characteristics of the products, and can be used as evidence for analyzing the overall performance of the products and preventing hidden dangers of the products.

The envelope analysis process comprises the steps of defining parameters participating in envelope analysis, defining parameter operation rules, boundary accuracy and the like, selecting according to the model and the product code number concerned by a user, generating an envelope analysis result table of a product, and automatically extracting the upper limit and the lower limit of the envelope according to historical success data by the user.

regression analysis

regression analysis is a method for determining the interdependent quantitative relationship between two or more variables and has wide application. According to the type of relationship between independent variables and dependent variables, linear regression and nonlinear regression can be classified. The regression analysis, which includes only one independent variable and one dependent variable according to the number of independent variables and dependent variables, is called a univariate regression analysis, and if two or more independent variables are included in the regression analysis, it is called a multivariate regression analysis. The partial least square regression method is a novel multivariate statistical data analysis method, mainly researches the regression modeling of multiple independent variables by multiple dependent variables, and is more effective by using the partial least square regression method particularly when the internal height of each variable is highly linearly related, and the system adopts the partial least square method to carry out regression analysis.

Neural network algorithm

Neural networks provide a relatively efficient method for solving complexity problems, and can easily solve problems with a large number of parameters, which are commonly used in classification and regression.

A neural network may be structurally divided into an input layer, an output layer, and a hidden layer. Each node of the input layer corresponds to an independent variable, and the number of the nodes of the output layer corresponds to a target variable. An implied layer is arranged between the input layer and the output layer, and the number of layers of the implied layer and the number of nodes of each layer determine the complexity of the neural network. Each neuron of the hidden layer is an independent unit with a similar structure, receives data transmitted from the previous layer, inputs the weighting of the data into the nonlinear function, and finally transmits the output result of the nonlinear function to the output layer.

(6) Data visualization display

The visualized display of the engine reliability data is divided into a two-dimensional graph data display function and a three-dimensional graph data display function.

Two-dimensional graph data presentation

the system realizes the display of the reliability data two-dimensional graph through an integrated report design tool FineReport. The support data field is associated with the graphic form, and when the format and the content of the two-dimensional graph need to be changed, the two-dimensional graph can be quickly realized by modifying the template.

the setting is carried out through a FineReport of a design tool, and the display of various types of two-dimensional graphs such as column graphs, pie graphs and broken line graphs can be supported for the same data.

(7) Reliability index alarm function

The reliability index alarm function comprises two parts: reliability index definition and risk early warning tree.

reliability index definition

The system provides a reliability index definition function, supports the definition of names, data sources (calculation methods), alarm threshold values (target values), index categories and the like of reliability indexes (such as reliability, failure degree, failure rate, service life, average service life and the like) by users, and provides a function of inquiring the indexes according to the categories and the index names.

risk early warning tree

And the early warning information of the reliability index of the engine can be checked through the risk early warning tree. The risk early warning tree is constructed by the system according to levels, the system carries out graphical early warning on the standard condition according to indexes in the system, the standard condition can be displayed in different colors according to the early warning level, and the standard condition can be unfolded/folded in a hierarchical mode according to needs.

(8) algorithm plug and play support

In addition to the basic mathematical analysis algorithm and reliability algorithm that are already supported in the current system, the system needs to provide plug-and-play support of the algorithm, i.e., the user can conveniently use the custom algorithm in the system. At present, Python is the most common programming language for data analysis and mining, and can construct various reliability algorithms to meet business requirements, so that the system realizes algorithm plug and play through the support of the Python language, transmits the algorithm name and parameters to the Python, and returns the result after performing reliability calculation. The Python algorithm support is shown in figure 6.

In order to realize convenient calling of the system to the Python algorithm, the system also needs to support the editing management function of the Python algorithm and the parameters and return values thereof.

9) data security design

In order to guarantee data safety, a safety design is carried out on the reliability data analysis application tool:

the system provides login management and carries out log recording on the IP address, the login time and a login person;

The system carries out safety control on access, provides a user interface related to the authority of the user, and only presents menus and buttons conforming to the authority of the user;

The system sets the security of the data domain, namely, which data records can be accessed by the user;

the system administrator authority adopts 'three-person management', and accords with the national confidentiality regulation.

In order to guarantee data security, when data is imported, exported and modified, corresponding approval processes need to be carried out, audit logs need to be matched with related operations, and the operations of importing, exporting, modifying and the like without approval are forbidden.

And the related users initiate the task application for importing, exporting and modifying, submit the task application to related personnel for auditing, transfer the task application to the related users for operation if the auditing of the related personnel is passed, and refute the task application to the application node if the auditing of the related personnel is not passed, and resubmit the application by the related users according to the refute opinions.

Specific design and function shows are shown in figures 7-19.

Claims

1. A solid rocket engine reliability data analysis mining and application software tool is divided into three layers of a basic platform layer, a data management layer and a data application layer, and is characterized in that:

the basic platform is a server layer and comprises an application server, a database server, a file server and an integrated interface, and provides platform functions of tree definition user management, authority setting, log management and parameter setting;

The data management layer is a database layer and provides data management functions of importing, exporting and editing basic information of the engine, reliability data of charging and internal insulation, reliability data of the spray pipe, reliability data of the shell and external insulation and reliability data of the complete engine and direct parts;

The data application layer is a user layer, is a visual tool and an operation interface which are interacted with a user, and is used for the reliability data application functions of data query, reliability data analysis, reliability data two-dimensional graph display and reliability data three-dimensional graph display.

2. The solid-rocket engine reliability data analysis, mining and application software tool of claim 1, wherein: the reliability data of the charging and the internal insulation, the reliability data of the spray pipe, the reliability data of the shell and the external insulation and the reliability data of the complete engine and the direct parts are divided into product data, process data and fault data according to the life cycle and the application,

3. The solid-rocket engine reliability data analysis, mining and application software tool of claim 1, wherein: the data management function and the data application function have data general functions including data import, data export, data display, data query retrieval, data analysis, addition, editing and deletion, wherein the data import comprises a data import template and a data import verification function; the data display comprises BOM tree display and list display functions; the data analysis is the analysis of the packaged reliability data and comprises mean value method packaging, minimum value method packaging, maximum value method packaging, power method packaging, evolution method packaging, fitting method packaging and interpolation method packaging.

4. the solid-rocket engine reliability data analysis, mining and application software tool of claim 1, wherein: the reliability data analysis comprises mathematical operation analysis, data statistics analysis, envelope analysis, regression analysis and neural network algorithm.

5. The solid-rocket engine reliability data analysis, mining and application software tool of claim 1, wherein: the data application layer further comprises a reliability index alarming function, and the reliability index alarming function is divided into reliability index definition and a risk early warning tree.

6. the solid-rocket engine reliability data analysis, mining and application software tool of claim 1, wherein: the solid rocket engine reliability data analysis, mining and application software tool further comprises algorithm plug-and-play support, and the algorithm plug-and-play support has an editing management function on a Python algorithm, parameters of the Python algorithm and return values of the Python algorithm.

7. The solid-rocket engine reliability data analysis, mining and application software tool of claim 1, wherein: the solid rocket engine reliability data analysis, mining and application software tool further comprises data security management, wherein the data security management comprises login management, access security control, data domain security setting and three-person management of system administrator permission.