基于智能预测算法的烧结工艺三维能耗监控系统研究与设计

发布时间：2018-03-07 21:19

本文选题：烧结工艺　切入点：神经网络　出处：《东华大学》2017年硕士论文　论文类型：学位论文

【摘要】：钢铁企业的能源消耗问题以及工业流程中的动态监控问题始终是钢铁行业关注的问题。烧结是铁前三工艺之一,其生产过程是典型的物质和能量的连续、非连续变化的复杂过程,且一些流程参数存在滞后性、测量不准确等特点,因此加强烧结过程的监督与控制能够有效地控制能源消耗以及烧结矿的质量。目前工业中运用的三维监控还比较少,大多运用二维平面监控,为了更真实的展现烧结工艺,本文利用三维技术进行监控系统的设计。整体思路为:针对烧结工艺的动态流程,利用神经网络建立模型,完成工艺流程的动态预测;同时利用三维建模技术对流程进行三维模型的建立完成监控部分的设计。具体得到研究与设计结果如下:1.参数选取:针对烧结工艺流程进行详细分析与了解,依据流程特点以及涉及到的各类工业参数,选取出对最终能耗影响较大的9个参数,进行数据收集。2.系统总体设计与数据库设计:基于整个系统需求,进行系统的总体设计,并且对数据库进行了设计。通过进行数据库的需求分析,完成基础表功能的确定,再进行概念结构的设计,转化为E-R图,最终根据E-R图进行用户表的设计。3.能耗预测:针对烧结工艺流程特点以及参数特点,运用小波神经网络进行能耗的预测。针对参数的数据特点,进行神经网络的学习与训练,完成能耗的静态预测。此处的静态预测是预测一天当中平均的能耗(kgce/t),由于烧结工艺的连续性与时滞性,在静态预测的基础上进行动态的预测,即预测某采样期的能耗值。动态预测中,神经网络的输入为要预测的当前采样点的影响参数以及前几采样点的影响参数,输出为当前采样期的能耗值。这样能够充分考虑到工艺的时滞性,使能耗预测更为及时准确。4.针对监控环节进行了三维监控系统的研究与设计。功能包括流程显示、数据的显示、场景的切换、报警提示等。它涉及的主要模块有:1)3D模型设计模块:实现工业流程中各器件的三维再现,依据流程进行三维模型的建立,包括烧结机、环冷机等模型的设计与创建;2)动画及界面显示模块:将创建好的3D模型导入Unity3D中,运用自带的粒子系统以及GUI系统进行三维动画特效和显示界面的添加与制作,最终完成初步的显示界面;3)报警模块:依据显示框中的能耗预测值与真实值进行比较,最终实现当两者差值较大时,报警灯闪烁,提示操作者进行参数调整;4)数据显示模块:将Unity3D和数据库进行连接,实现两者之间的数据的交互。本论文最终实现了设计的三维监控系统,为烧结工业流程的监控提供较好的依据。
[Abstract]:The problem of energy consumption in iron and steel enterprises and the problem of dynamic monitoring in industrial process are always concerned by the iron and steel industry. Sintering is one of the first three iron processes, and its production process is a typical continuous process of material and energy. The complex process of discontinuous change, and some process parameters have the characteristics of lag, inaccurate measurement and so on. Therefore, strengthening the supervision and control of sintering process can effectively control the energy consumption and the quality of sinter. The whole idea is: according to the dynamic process of sintering process, the neural network is used to establish the model to complete the dynamic prediction of the process flow. At the same time, the 3D modeling technology is used to build the 3D model of the process to complete the design of the monitoring part. The specific research and design results are as follows: 1. Parameter selection: detailed analysis and understanding of the sintering process, According to the process characteristics and all kinds of industrial parameters involved, 9 parameters which have a great impact on the final energy consumption are selected for data collection .2.The overall design and database design of the system: based on the requirements of the whole system, the overall design of the system is carried out. And the database is designed. By analyzing the requirements of the database, the function of the basic table is determined, and then the conceptual structure is designed, which is transformed into the E-R diagram. Finally, according to E-R diagram, the user table is designed .3.Energy consumption prediction: according to the characteristics of sintering process and parameters, wavelet neural network is used to predict energy consumption. According to the data characteristics of parameters, the study and training of neural network are carried out. The static prediction is to predict the average energy consumption during the day. Because of the continuity and time delay of the sintering process, the static prediction is based on the static prediction. In dynamic prediction, the input of the neural network is the influence parameters of the current sampling points to be predicted and the influence parameters of the previous sampling points. The output is the energy consumption value of the current sampling period. In this way, the time-delay of the process can be fully taken into account, and the energy consumption prediction can be more timely and accurate .4. the research and design of the three-dimensional monitoring system for the monitoring link is carried out. Data display, scene switching, alarm warning and so on. The main modules involved in it are: 1: 1 / 1 3D model design module: realize 3D reproduction of each device in the industrial process, build 3D model according to the process, including sintering machine, Design and creation of models such as ring cooling machine) Animation and Interface display Module: the created 3D model is imported into Unity3D, and the special effects of 3D animation and the adding and making of display interface are carried out by using the particle system and GUI system. Finally, the initial display interface (3) alarm module is completed: according to the energy consumption prediction value in the display box and the real value, the alarm lamp flashes when the difference between the two values is large. The data display module: connecting Unity3D and database to realize the data interaction between the two. Finally, this paper realizes the design of a three-dimensional monitoring system. It provides a good basis for the monitoring of sintering process.
【学位授予单位】：东华大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP277

【参考文献】