航空气象多源异构数据存储管理平台的设计及应用

Design and Application of an Aviation Meteorological Multi-Source Heterogeneous Data Storage and Management Platform

摘要: 针对航空气象业务中多源异构探测数据类型复杂、增长迅速，以及传统存储架构在可靠性、安全性和扩展性方面存在不足的问题，设计并实现了一套航空气象多源异构数据存储管理平台。平台由数据接入、数据治理、分布式存储、安全防护和业务应用支撑等部分组成，可实现多源气象探测设备和相关业务数据的统一接入、标准化处理、集中存储和安全管理。在数据接入方面，平台支持多种通信协议和异构数据格式解析；在数据治理方面，构建了质量控制、时空对齐和多源融合处理流程；在数据存储方面，基于MinIO分布式对象存储架构，结合纠删码和校验和机制提高数据可靠性；在安全防护方面，采用SM2算法和SSL/TLS机制实现数据传输与访问控制保护。以机场低能见度监测预警为应用场景，利用葵花8号卫星资料、数值预报资料、气象局加密自动站资料以及双流机场和天府机场自动观测资料开展应用验证。结果表明：该平台能够较好支撑多源气象数据的协同处理与应用调用，在典型个例中可提前24 h给出起雾概率预测，预测结果与实况基本一致。研究结果说明，该平台可为航空气象智能化业务提供稳定的数据基础支撑。

Abstract: To address the problems of complex data types, rapid growth of multi-source heterogeneous observation data, and the limitations of traditional storage architectures in reliability, security and scalability in aviation meteorological services, an aviation meteorological multi-source heterogeneous data storage and management platform was designed and implemented. The platform consists of data access, data governance, distributed storage, security protection and service application support modules, and realizes unified access, standardized processing, centralized storage and secure management of multi-source meteorological observation equipment and related business data. In terms of data access, the platform supports multiple communication protocols and heterogeneous data format parsing. In terms of data governance, a processing workflow including quality control, spatiotemporal alignment and multi-source data fusion was established. In terms of data storage, the platform is built on the MinIO distributed object storage architecture, and data reliability is improved by erasure coding and checksum mechanisms. In terms of security protection, the SM2 algorithm and SSL/TLS mechanism are adopted to protect data transmission and access control. Airport low-visibility monitoring and warning was selected as an application scenario. Himawari-8 satellite data, numerical weather prediction data, dense automatic weather station data from the meteorological bureau, and automatic observation data from Shuangliu Airport and Tianfu Airport were used for validation. The results show that the platform can effectively support collaborative processing and application use of multi-source meteorological data. In a typical case, the platform can provide fog probability prediction 24 h in advance, and the prediction is generally consistent with observations. The results indicate that the proposed platform can provide a stable data foundation for intelligent aviation meteorological services.