Abstract: Deep learning technology has many application scenarios in modern discrete manufacturing industry,and it is one of the key technologies to support the transformation of discrete manufacturing from digital and network stage to intelligent stage.For practitioners of discrete manufacturing industry,the learning cost of using traditional development methods and applying deep learning technology to solve problems in production scenarios is very high.Practitioners are required to have both deep development ability and deep learning knowledge background.To reduce the cost of deep learning function development for practitioners in discrete manufacturing as much as possible,a deep learning codeless development platform for discrete manufacturing was proposed and implemented based on the characteristics of discrete manufacturing industry,deep learning in discrete manufacturing and the application of artificial intelligence development mode to fully understand.The application effect of enterprise-level projects was compared,which proved the excellent performance of the development platform.

Key words: discrete manufacturing, deep learning, codeless development, artificial intelligence, ductile iron pipe

摘要： 深度学习技术在现代离散制造业中有很多应用场景,是支撑离散制造业由数字化、网络化阶段向智能化转型的关键技术之一。对离散制造业的从业者而言,使用传统开发方式,应用深度学习技术解决生产场景中问题的学习成本很高,需要从业者同时具备较深的开发能力和深度学习知识背景。为了尽可能地降低离散制造业从业人员进行深度学习功能开发的成本,在对离散制造业的特性、深度学习在离散制造业的应用情况和人工智能开发模式进行充分了解之后,提出了适用于离散制造业的深度学习无代码开发平台的构建方案,并进行了实现,最后应用企业级的项目进行了应用效果对比,证明了该开发平台的优异性能。

关键词: 离散制造业, 深度学习, 无代码开发, 人工智能, 球墨铸管