基于工况的操作模式优化及在石化工业中的应用  

Working condition-based optimization framework for operational patterns and its application in petrochemical industry

作　　者：蒋白桦[1] 刘伟[2] 戴志军[2] 王宏安[2] 

机构地区：[1]石化盈科信息技术有限责任公司,北京100007 [2]中国科学院软件研究所,北京100190

出　　处：《化工学报》2012年第12期3978-3984,共7页CIESC Journal

基　　金：国家科技支撑计划项目(2012BAE05B03);国家重点基础研究发展计划项目(2009CB320804)~~

摘　　要：提出了一种基于工况的操作模式优化框架,框架包括三部分:数据预处理、优化操作模式库的形成和基于工况的实时优化。在操作模式优化控制的框架和相关概念的基础上重点研究了催化过程中干点温度指标在线优化的应用。针对催化的工业特点,提出SVM与AdaBoost相结合的两步结合的操作模式的发现方法,试验证实该方法具有较高的预测性能。A new working condition-based optimization framework for operational patterns was proposed.The framework is composed of three components which are data preprocessing,the creation of library of optimized operational patterns and real-time optimization based on working conditions.On the basis of optimization and control framework,the application of online operational optimization for the dry-point temperature of FCCU was focused.Base on the FCCU properties,a two-step optimization approach for operational patterns which combines SVM and AdaBoost was proposed.Experimental results show that the new approach has higher prediction performance.

关 键 词：操作模式 软测量 支持向量机 ADABOOST 模式发现 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]