面向新型铜冶炼过程的电力负荷预测方法研究

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	面向新型铜冶炼过程的电力负荷预测方法研究
作者	赵力
学位类别	硕士
答辩日期	2016-05-25
授予单位	中国科学院沈阳自动化研究所
导师	王卓
关键词	富氧底吹铜熔炼机理分析电力负荷预测神经网络
其他题名	Load Forecasting Methods for the Oxygen-enriched Bottom Blowing Copper Smelting
学位专业	控制工程
中文摘要	面对当前铜冶炼行业高能耗的现状，我国自主研发了新一代铜冶炼技术，即富氧底吹熔炼技术，新技术具有节能、污染低等优点，将会在未来的铜冶炼行业中得到广泛应用。本文在面向新型铜冶炼过程的电力负荷预测方法研究中，针对目前铜冶炼行业电力负荷预测方法主要是通过经验预测或时间拟合方法获得，缺少与实际生产工艺相结合，而导致预测过程中存在精度较差的问题，提出了一种耦合双模型的电力负荷预测方法，通过核心工序的机理需氧模型与制氧工序电力负荷预测相结合的建模方法，对制氧工序的电力负荷进行预测，并通过数据的验证，证明本文提出的电力负荷预测方法预测精度较高，可以适用于新型铜冶炼生产过程电力负荷预测，为企业制定合理电力规划提供依据，提高企业的竞争力。本文针对新型铜冶炼业节能降耗目标，以充实富氧底吹熔炼新工艺的研究和完善面向铜冶炼企业的电力负荷预测方法体系。具体的研究内容如下：首先，针对新工艺发展初期基础理论研究不足的问题，根据企业能源报表和现场能源消耗过程的统计与研究，对富氧底吹铜冶炼全厂工艺流程总结，并完成企业工艺生产过程用能分析，选取制氧厂电力负荷作为单能源预测目标。其次，在深入分析熔炼反应机理的前提下，建立了基于物质流平衡的熔炼工艺机理需氧模型，并通过现场生产数据确定了模型的参数，通过数据验证了模型的准确性。机理需氧模型研究不仅是后续电力负荷预测的基础，也利于对富氧底吹熔炼新工艺的原理理解以及新技术更好的推广发展。进一步，在富氧底吹工艺需氧模型的基础上，提出了一种适合工序级别的耦合双模型预测方法，通过需氧流量预测电力负荷。将机理模型输出的氧气需量结果与制氧厂电力负荷影响因素作为神经网络输入层，通过粒子群全局优化BP神经网络初始参数，通过添加动量项，训练输入样本得到最优网络权值，优化算法提高了网络的准确性，并通过现场数据进行仿真验证，实验结果表明相比较传统的预测方法和现场经验生产，此预测方法拥有更高的精度和稳定性，尤其是对生产计划变动下的电力负荷预测的准确性有了很大提高。最后，基于B/S架构开发了电力负荷预测功能，提出了能源管理平台软件架构以及软件开发环境，实现了登陆验证模块、数据库访问、电力负荷预测模块等功能，并利用离线数据进行了验证。电力负荷预测是企业能源调度部门电力管理的前提与核心，而新型富氧底吹铜冶炼企业能源管理水平的提升，将进一步推动富氧底吹新技术的发展，本文通过对制氧工序电力负荷预测的研究，提供给富氧底吹企业一种结合工艺生产过程考虑的电力负荷预测方法。
英文摘要	Faced with the high energy consumption of the copper smelting industry, a new generation of copper smelting technology has independently developed in our country, which named oxygen-enriched bottom blowing smelting technology. The new technology has the advantage of energy saving, which will be widely used in the future copper smelting industry. Based on the study on load forecasting method for the new copper smelting process, a new power load forecasting method which coupled dual model is proved in the thesis, faced with the problem that current methods are mainly through production experience or getting fit based on time and they lack of combining the actual production process. The author combined the mechanism model and power load forecasting model in the new method, proved that the power load forecasting methods of high prediction accuracy through simulation data. We can provide the basis for enterprises to develop reasonable power plan to improve the competitiveness of enterprises by using the forecast method in this thesis. For reducing energy saving targets of copper smelting industry, specific contents are put forward as follows, in order to enrich oxygen bottom- blowing smelting new technology research and perfect for copper smelting enterprises of power load forecasting method system. Because the technology is relatively new and lacking the basic theory research, the author firstly processed the technology of rich oxygen bottom blowing copper smelting, complete enterprise energy analysis, and select the oxygen plant electrical energy as the single energy forecasting targets. Secondly, in depth analysis of the premise of the smelting reaction mechanism, the mechanism of the process material flow balance aerobic model is established. Mechanism aerobic model is not only the basis for follow-up electricity load forecasting , but also conducive to the oxygen-rich bottom blowing smelting new technology and better understanding of the principles of the promotion of the development of new technologies. Later the author determined the parameters of the model through the field by producing data and verified the accuracy of the data model. Further, on the basis of the mechanism of oxygen-rich bottom blowing process model, the author proposed coupling double model for power load forecasting. The author demanded oxygen plant power load factors and mechanism of the model output oxygen as a result of the neural network input layer, and optimized BP neural network parameters through the initial particle swarm globally by adding momentum to obtain the optimal network weights optimization algorithms to improve the accuracy of the network. Through field data simulation, experimental results showed that compared with traditional forecasting methods and experience in the field of production, the coupling double model had higher accuracy and stability, particularly for electricity production plans under load fluctuation prediction accuracy has been greatly improved. Finally, the authors developed the energy forecasting capability to enable a login authentication modules, database access, power load forecasting module and other functions, and use the offline data was verified. Power load forecasting is a prerequisite and core power management, and the development of enterprise energy management technology, which will further promote low energy consumption and promote new technologies. Research of the thesis is to provide an oxygen-rich bottom blowing technology enterprise energy forecasting method under conditions suitable for complex production.
语种	中文
产权排序	1
页码	60页
内容类型	学位论文
源URL	[http://ir.sia.cn/handle/173321/19657]
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	赵力. 面向新型铜冶炼过程的电力负荷预测方法研究[D]. 中国科学院沈阳自动化研究所. 2016.