铝电解过程能效优化方法研究

CORC > 沈阳自动化研究所 > 中国科学院沈阳自动化研究所 > 信息服务与智能控制技术研究室

题名	铝电解过程能效优化方法研究
作者	孔磊
学位类别	硕士
答辩日期	2015-05-26
授予单位	中国科学院沈阳自动化研究所
授予地点	中国科学院沈阳自动化研究所
导师	王卓
关键词	铝电解能效优化电流效率价值虚拟对标软测量集成建模
其他题名	Research on Method of Energy Efficiency Optimization of Aluminum Electrolysis Process
学位专业	控制理论与控制工程
中文摘要	铝电解是传统的高耗能过程，面对当前严峻的能源形势，节能降耗已成为其首要目标。目前铝电解能效优化途径主要是从电解机理分析或实际生产经验中获得，忽略了过程数据对标对能效优化所起的作用。针对以上情况，本文提出了一种基于电流效率价值虚拟对标的铝电解过程能效优化方法，以充实和完善能效优化方法体系。具体研究的内容如下：首先，针对电流效率的测量周期较长问题，建立了铝电解过程电流效率软测量智能集成模型。由于模糊C均值聚类和非监督聚类方法存在不足，提出一种模糊C均值监督聚类改进算法，并在此基础上建立了监督分布式支持向量机电流效率软测量智能模型；同时，在深入分析铝电解反应机理的前提下，基于部分合理的假设条件以及电荷守恒定律，建立了电流效率软测量机理模型。将以上两种模型以训练样本均方误差最小化为准则进行加权集成，即得到电流效率软测量智能集成模型，并应用现场数据验证了模型的精度和泛化性能。其次，在电流效率软测量智能集成模型的基础上，提出了一种铝电解过程电流效率价值虚拟对标能效优化方法。电流效率价值将电流效率与其影响因素组合地放在同一平台，利用支持度权重分配数据融合改进算法得到了电流效率影响因素的最佳匹配，然后基于电流效率软测量智能集成模型得到其对应的电流效率，二者组合即得到铝电解过程电流效率价值标杆。利用现场数据进行对标仿真实验，结果表明电流效率价值虚拟标杆更实用合理，能指出当前电解槽提高电流效率的有效途径，以降低吨铝直流电耗，优化能效。最后，开发了铝电解过程能效优化系统。提出了铝电解过程能效优化系统的软件架构，实现了登录验证模块、电流效率价值对标主模块和电流效率改进建议模块等，同时利用实际数据进行了验证。
索取号	O224/K48/2015
英文摘要	Aluminum electrolysis is a traditional energy intensive process. Considering the serious energy situation, the energy saving has be promoted to be the primary target of this process. Currently, the energy efficiency optimization approachs are mainly from the mechanism analysis and the practical experience, ignoring the role of the process data benchmarking. In view of the above situation, an energy efficiency optimization method of current efficiency value virtual benchmarking is proposed in this thesis. The concrete research contents are as follows: First, the intelligent integrated soft sensor model of the current efficiency is established focusing on the long cycle of the measurement of the current efficiency. Due to the deficiency of the fuzzy C means algorithm and unsupervised clustering method, a fuzzy C means supervised clustering algorithm is proposed. Moreover,a supervised distributed support vector machine is constructed to establish the current efficiency intelligent soft sensor model. Based on the in-depth analysis of aluminum electrolysis mechanism and some reasonable assumptions and the law of conservation of the charge, the mechanism model for current efficiency is established. Finally, the intelligent integrated model of current efficiency is achieved via weighted integration of the above two ones. The precision of the integrated model is demonstrated by the result of the simulation with the field data. It means the model can be used to forecast the actual aluminum electrolysis current efficiency. Secondly, the energy efficiency optimization method of current efficiency value virtual benchmarking is proposed on the basis of the intelligent integrated soft sensor model of the current efficiency. This paper combines the current efficiency and its influence factors to present the current efficiency value. With the improved support-degree based weight distribution data fusion method, the best match between the factors of the current efficiency is determined, and its corresponding current efficiency can be attained with the established intelligent integrated soft sensor model, so the virtual benchmark of current efficiency value is achieved by the combination of the best match and its current efficiency. The simulation experiment of the current efficiency value benchmarking is conducted with the field data, and the result indicates that compared with traditional methods, the virtual benchmark is more practical and reasonable, and can more effectively guide the operation of the cell to improve its current efficiency. Finally, the system software for the current efficiency value benchmarking of the aluminum electrolysis process is developed. The software architecture is first put up with, and then the Login-Authentication module, the current efficiency value benchmarking module and the current efficiency improvement module are introduced in detail. The software is verified with the field data and shows good performance.
语种	中文
产权排序	1
页码	55页
内容类型	学位论文
源URL	[http://ir.sia.ac.cn/handle/173321/16755]
专题	沈阳自动化研究所_信息服务与智能控制技术研究室
推荐引用方式 GB/T 7714	孔磊. 铝电解过程能效优化方法研究[D]. 中国科学院沈阳自动化研究所. 中国科学院沈阳自动化研究所. 2015.