Summary: From the massive equipment test data, the use of big data mining ideas to establish equipment big data analysis and processing and intelligent information services is one of the key technologies for the development of equipment data engineering. Analyzed the characteristics of equipment test data, combined with the big data processing analysis thoughts, put forward a kind of equipment big data model, taking the naval equipment test as an example, further analyzed the characteristics of the model, and provided a technical route for the construction of equipment data engineering. reference. 0 introduction In the process of equipment testing, a large amount of historical data resources have been accumulated, and first-hand data on equipment tests, training, and drills have been recorded. With the continuous expansion of data collection methods and the intensification of tasks of comprehensive experimental drills, tests have been conducted. The accumulation of data will increase exponentially. These data resources reflect the equipment performance and reliability, evaluate the capability level of the equipment system, and study the important basis for the development path of the equipment. However, due to the undeveloped information technology at the initial stage of construction and the constraints of network, computing, storage, and data processing technologies, there is a lack of unified system planning and design in the early days. Currently, a large number of weapon equipment test data are stored in different test terminals, and there is a lack of unified management. Analysis of the application mechanism, a large number of valuable information implicit in the data has not been effectively tapped and developed [1-2]. In the field of information technology, the big data platform ecosystem has achieved rapid development. There have been numerous open source big data distributed storage and parallel computing frameworks such as Hadoop, HBase, Storm, and Spark. Big data computing technology solves the problem of collecting, storing, calculating, and analyzing massive data. Big data technology has been applied in many fields such as medical treatment, finance, transportation, education, environmental protection, and public opinion supervision, which has driven the innovation and development of business models in related fields and achieved good economic and social benefits. In the field of military equipment, experimental data in aerospace, wind tunnels and other fields have been applied, but its essence has remained in the traditional data mining field, and no application results have been formed for big data technology. In this paper, combined with the equipment test data management requirements, a device big data model based on equipment test big data is researched and proposed. This model can guide how to extract the historical data using the equipment experiment, and use the big data processing technology to mine and establish the equipment decision analysis model knowledge. Library, and provide intelligent decision support services for equipment management based on the knowledge base, so as to improve the processing and analysis utilization of equipment data resources, and turn data resources into intelligent decision analysis services. 1 Equipment test data characteristics The equipment test data generally covers all data generated during the equipment testing process. For a long time, the test department has accumulated PB-level historical test data resources, including paper reports, electronic documents, films, electronic photographs, audio and video data, and lightning measurement data. Telemetry data, etc., are dispersed in different test terminal units, lacking a unified data organization and management specification platform. These experimental data data have the following characteristics: (1) There are many kinds of test data. With the expansion of the scope of equipment testing, the scope of test data is also becoming wider and wider. Among the major categories of tests, there are: performance test data for various types of equipment, performance evaluation data of equipment related to operational tests, and data deriving data for equipment systems; For single trials include: equipment, weather, geographical and hydrological environment, personnel protection, business processes and other data. (2) The forms of data are diversified, structured database data, unstructured texts, images, image data, and semi-structured data; in addition, there are also massive non-digitized documents. (3) High potential value and relatively stable value. Different from business data, with the passage of time, the details of historical data are no longer important. Weapons and equipment test data for equipment life cycle management cycle is very long, can be as long as ten years, data value is insensitive with time. (4) The data keeps growing. With the in-depth development of system-level equipment tests and equipment-based operational tests that are close to actual combat in the later period, the amount of experimental data resources will continue to grow rapidly at an exponential level. There will be more and more experimental data needs to be managed, analyzed and utilized. The traditional data warehouse-based data center solution will be difficult to meet the needs of future development, and a distributed data center platform based on big data needs to be established. 2 Analysis of Management Requirements for Equipment Test Data The management requirements for equipment test data mainly include the following four aspects: (1) Demand for data modeling: The value of current massive experimental data resources is in a state of slumber. Due to its unclear base and different formats, it is not yet able to apply more in-depth analysis to it, so the massive experimental data resources are modeled. The formation of effective management and governance measures is currently the most important issue to be solved. (2) Evaluation requirements for test tasks: Based on the test data of the test task, the actual results of the test task can be evaluated; based on the previous test tasks of the same type of equipment, the technical development of the equipment can be evolved and portrait analysis can be carried out to assist in the improvement of the equipment technology. Mining analysis decisions. (3) Assessment of multi-equipment systemization test needs: Through the fusion of multi-professional categories such as equipment, environment, safeguards, and task coordination, a visualized test process situation based on experimental data is developed to assist in comprehensive assessment of system effectiveness. (4) Equipment health management requirements: Based on the equipment test data, a health management model of the equipment is established. Based on the model, the remaining life of the equipment can be predicted, and an accurate on-demand maintenance and protection plan can be formulated according to the actual status of the equipment. 3 Equipment Test Big Data Model - ETBDIS The fundamental goal of equipment test big data is to provide intelligent information services and management decisions for equipment testing activities, and to provide support for the use of equipment in the later period, equipment maintenance and equipment technology development, so the equipment test big data model first needs to consider the service of the data platform. The contents include: equipment portraits, equipment modeling, equipment health management, and equipment systemization assessment. For example, as shown in Fig. 1, equipment imagery can be quantitatively and qualitatively analyzed on equipment quality, performance, applicable environment, maintenance and protection, health status and remaining life, and technology development process. This paper aims at the fundamental goal of equipment test data construction, and constructs and designs an Equipment Test Big Data Intelligent Service (ETBDIS). 3.1 ETBDIS model ETBDIS model includes experimental data layer, data preprocessing layer, data resource system layer, big data knowledge mining layer, equipment intelligent service layer, etc. (as shown in Fig. 2). The main functions of each layer are as follows. Test data layer. Test data is mainly divided into text, images, test and measurement data according to the type of data. Data preprocessing layer. According to the data type, select the appropriate data preprocessing method, such as text data, which can perform entity extraction, abstract extraction, keyword extraction, etc., and extract physical data such as equipment, quality, performance, professional technology, and events for follow-up. The entity association and event association mining are used for analysis; the test and measurement data need corresponding special data processing software for preliminary analysis and pretreatment of the data. Data resource system layer. After the original data has undergone corresponding preprocessing, the structured data that have been classified and labeled are obtained, and the equipment big data resources that can be analyzed and utilized are then constructed. Including: the data subject of equipment quality, performance, environment, technology, etc. Based on these structured data resources, the corresponding equipment knowledge base can be extracted through big data mining, machine learning and other methods. Big data knowledge mining layer. For the data subject in the equipment data resource, the corresponding machine learning methods can be used to perform dynamic model modeling, classification rule mining, pattern mining, feature extraction, and multi-entity relationship knowledge mining, and learn from the data to build equipment. Knowledge base. Equip intelligent service layer. The equipment knowledge base constructed based on data mining can provide corresponding intelligent information services based on equipment test big data. Such as: equipment portraits, equipment health management decision-making, evaluation of equipment testing tasks, and simulation analysis of equipment test simulation under different environmental systems. 3.2 Analysis of the Characteristics of Naval Equipment Test Data Management Based on ETBDIS Combined with the actual practice of naval equipment testing business, the main features of the ETBDIS model are as follows: (1) The model adopts a layered architecture and can implement data communication between different layers in the system through a standard specification interface. Within each layer, it is a modular structure, and can adopt flexible microservices for flexible intra-layer module integration. For the naval equipment test data, the first is to complete the unified management of the test data by using data such as experimental document data, video recording, and other image data and measurement tests. The test measurement data may use different acquisitions according to different test items. Analysis tools, data decoding methods are also not the same, for the formation of the distribution of experimental data acquisition framework. (2) Service-driven. It is geared to the needs of smart data services for big data equipment. Because the experimental data of naval equipment involves a large number of equipment categories, a unified test data system is formed through preprocessing on the data layer basis, and various types of thematic databases are formed, and the basic relationship between test data is established, thereby achieving the service of equipment portraits, equipment test evaluation, Equipment management, such as equipment health management and equipment evaluation under different environmental systems, not only ensures the basic construction goals of the data center, but also ensures the necessity of data collection and processing. (3) Open architecture. The equipment test data involves multiple stakeholders, including the entity departments of the equipment testing of the organizations as well as the equipment demonstration, development, and use. Therefore, each participating unit must implement its own professional field on the platform in accordance with data standards and interface standards. The storage, management, and service of data are based on an open infrastructure, and the analysis of big data fusion processing is implemented based on a professional division of labor and a unified platform. (4) Data depth fusion processing. The equipment test involves massive unstructured processing of massive pictures and documents. For example, in the navy equipment test process, various test-related conference documents, test audio and video, etc. are involved, and multi-dimensional data resources such as quality, performance, maintenance, and technical development are established so that machine learning, data mining, and other knowledge can be utilized. The library construction process improves the data processing depth. 4 in conclusion The equipment testing work has accumulated a large amount of equipment-related data. This paper aims to integrate the test data of various units, establish a system of equipment test data resources, and provide equipment information intelligence information services. A large equipment model for equipment testing is given. The model has a layered and open architecture. Each participating unit can perform test data collection and analysis according to the model on a unified platform, achieve cross-departmental multi-discipline data integration, and use the equipment data resource platform to carry out corresponding knowledge. Constructing and intelligent information services, and analyzing the main features of the model with naval equipment experimental data management as an example, further demonstrated the usability of the model. In the future research, several typical equipments will be selected for step-by-step construction of model applications. Crystal head is an important interface equipment in network connection. It is a kind of plastic joint which can be inserted along the fixed direction and can automatically prevent falling off. It is used for network communication. It is named "crystal head" because of its crystal like appearance. It is mainly used to connect network card port, hub, switch, telephone, etc. The two ends of each twisted pair are connected with network card and hub (or switch) by installing crystal head. Modular Plug ShenZhen Antenk Electronics Co,Ltd , https://www.antenksocket.com
