基于多模态异构算法组态工具的流程行业智能化解决方案  

Intelligent Solution for Process Industry Based on Multi-Modal Heterogeneous Algorithm Configuration Tool

作　　者：李戎 孙继超 

机构地区：[1]杭州和利时自动化有限公司,陕西西安710075

出　　处：《自动化博览》2024年第3期64-69,共6页Automation Panorama1

摘　　要：本文首先分析了流程行业应用场景从自动化向智能化转型升级所需要的条件,并从数据模态、模型特征、异构实现、工程化以及人员能力方面总结了工业智能应用面临的问题和挑战;其次,针对性地提出了多模态数据融合、机理模型数据驱动及专家异构整合、算法技能,符合用户习惯的工程化组态方式、不同人员协作方式等方案,并进一步给出了针对复杂工业场景的工具方案;最后,对智能优化控制、设备异常检测、多模态数据闭环控制等场景进行了分析,并总结展望了多模态异构算法组态工具在流程行业向智能化转型升级中的作用。This paper initially analyzes the transformation and upgrading conditions of process industry application scenarios from automation to intelligence.It then summarizes the problems and challenges faced by industrial inteligence applications from the aspects of data modality,model characteristics,heterogeneous implementation,engineering practices,and personnel capabilities.Furthermore,we propose multi-modal data fusion,mechanism model data drive,expert heterogeneous integration and algorithm skills.In line with the user's habit of engineering configuration,different personnel cooperation and other schemes,and further give the tool scheme for complex industrial scenes.Finally,analyzing intelligent optimization control,equipment anomaly detection,multi-modal data closed-loop control and other scenarios,this paper outlines the pivotal role of multi-modal heterogeneous algorithm configuration tools in the transformation and upgrading of process industryto intelligence.

关 键 词：数据模态 多模态数据融合 智能优化控制 多模态数据闭环控制 

分 类 号：TP399[自动化与计算机技术—计算机应用技术] TH16[自动化与计算机技术—计算机科学与技术]