【作者】 钱新华；

【导师】 姚平经； 王克峰；

【作者基本信息】 大连理工大学 ， 化学工程， 2010， 博士

【摘要】 乙烯生产分离装置是典型的大型连续性化工生产装置,工艺程序复杂,设备多,易发生故障,如何实现其过程的严格在线动态模拟,获得最接近生产实际的动态模型,进而建立其数字化工厂,为节能优化、安全生产和优化生产做出有效指导,是非常具有现实意义的问题。本文的研究课题就是在这一背景下提出的,着重研究了以下几个问题：一、针对乙烯装置建立以数据为中心的数字工厂,运用生产运行管理系统(MES);同时利用自主知识产权软件SIMACS,建立以严格稳态和动态模型为基础的数字工厂,实现安全和优化生产。二、建立以数据为中心的数字工厂的关键问题是如何实现数据从各种型号的分布式控制系统(DCS)或工厂实验室数据系统(LIMS)传输到实时数据库。传统的乙烯装置DCS和LIMS的数据采集方式存在着滞后、包容性差等特点。针对以上问题,本文提出和改进了传统的OPC (OLE for Process Control)技术进行实时数据的采集,采用了MATRICON OPC技术。它自身所拥有的传输数据量大、速度快、高稳定、数据安全等特点,完全可以满足建立以数据为中心的数字工厂的各种需求。三、建立以严格稳态和动态模型为基础的数字工厂的关键问题是如何将稳态模型和动态模型相结合,利用实时上传的实际工厂的实际数据和动态模拟的数据值进行对比,实时对模型参数进行自适应校正,实时调整模型参数,保证实时动态模型的准确度,从而使严格稳态和动态模型更能反映工厂的实际情况。针对上述问题,本文在软件SIMACS的基础上提出了动态模拟与稳态模拟技术相结合的具有模型参数自适应校正功能的化工过程在线动态模拟技术。四、以严格稳态和动态模型为基础的数字工厂的基础上,对建立智能数字工厂进行了初步研究。利用数据挖掘的方法,定义了智能故障诊断的决策树知识表达和C4.5决策树获取算法,并在乙烯装置中进行了验证,取得了良好的经济效益。本文通过对以上实时数据传输技术和动态模拟技术这两个关键技术的研究,建立了乙烯装置的数字工厂。此数字工厂已成功在某石化企业得到应用,为企业的生产管理带来了很好的收益。更多还原

【Abstract】 Ethylene producing and separation unit is the typical large-scale continuous chemical production plant; its process program is complex and frequent malfunction. How to realize the strict on-line dynamic simulation of the process, obtain the simulation data close to the produce reality, then guide for production effectively, it has great realistic meaning. Research subject of this paper is proposed based on the background above, focusing on the following questions:1 Digital factory of data-centric for ethylene plant can realize Manufacturing Execution System (MES). Safe and optimized production can be achieved by Digital factory based on strict steady-state and dynamic models using Proprietary software SIMACS.2 The key issue for establishing a data-centric digital factory is how to transfer data from various types of distributed control system (DCS) or Laboratory Information Management System (LIMS) to the real-time database. The DCS and LIMS data collection method for conventional ethylene plant is lagging, unstable, poor-compatible. To solve the problems, MATRICON OPC(OLE for Process Control) technology is presented here to improve the conventional OPC technology for real-time data acquisition. It has so many excellent characteristics such as large quantity of data transmission, high speed, high stability, data safety that will meet the demands of upper application program completely.3 The key issue for establishing a digital factory based on strict Steady-state and dynamic model is how to combine the two models. Comparing the real-time data from actual plants and data from dynamic models, adapting and correcting the model parameters in real-time, ensure the accuracy of dynamic model, so that the strict Steady-state and dynamic models can reflect the actual situation of the factory better. Chemical process on-line dynamic simulation technology is presented here. It is combined by Steady-state and dynamic models based on Proprietary software SIMACS.4 We studied the establishment of intelligent digital factories based on digital factories which is based on strict steady state and dynamic models. Using data mining method, we can define the intelligent fault diagnosis knowledge representation and C4.5 decision tree. The method has been used in ethylene plant. Good economic results were achieved.更多还原