面向钢铁制造行业的大数据平台架构研究

The Research on the Architecture of Big Data Platform for the Steel Manufacturing Industry

由于人工智能高投入、高技术门槛的特点,钢铁制造业中的大型企业需要一个可以共享的大数据平台支撑,实现人工智能技术共性能力的复用。提出一种基于容器集群技术的大数据平台架构,详细阐述大数据平台架构组成及数据计算引擎、数据模型工厂、数据模型市场等3个模块的功能实现,从而构建企业统一人工智能算力调度、数据标注、模型训练、在线推理、云边协同等核心能力,支撑现代大型制造企业各类人工智能应用场景。

Due to the characteristics of high investment and high technical threshold of AI,large enterprises in the steel manufacturing industry need a shared big data platform support to realize the reuse of AI technology common capabilities and reduce the cost of building AI applications.A big data platform architecture based on container cluster technology is proposed to elaborate the data platform architecture composition and the functions of three modules,such as data model engine,data model factory and data model market.It realizes the core capabilities of arithmetic scheduling,data annotation,model training,online inference and cloud-side collaboration to support various AI scenarios in modern large manufacturing enterprises.