Dissertation > Excellent graduate degree dissertation topics show

Research of Bof End-point Optimization Control Model

Author: KongXiangRui
Tutor: YangChengZhong
School: Hangzhou University of Electronic Science and Technology
Course: Control Theory and Control Engineering
Keywords: BOF steelmaking Endpoint control RBF Neural Network BP neural network Chaos immune particle algorithm Regional optimization
CLC: TF713
Type: Master's thesis
Year: 2010
Downloads: 96
Quote: 1
Read: Download Dissertation

Abstract

The thesis topic selected from the group consisting of key scientific research projects in Zhejiang Province ─ ─ \completely rely on the artificial experience, blowing a high hit rate is not the end of the case study. Blowing process, the on-site blowing process as well as the analysis of production data, combined with neural networks based on the application of technology, the establishment of the small end of the converter optimal control model to achieve the optimization of the operation of blowing. Read a lot of literature on the basis of the paper introduces the status quo and development trend of the end of the converter control technology at home and abroad, with particular emphasis on the successful application of neural network technology in BOF endpoint control. Jiangsu Yonggang Group BOF blowing the study and analysis of the current situation and problems in the molten pool to the final temperature and carbon content in the case of a not quite successful for the end of the research object model. The object model consists of two parts: the forecasting model and control model. The paper focuses on modeling principles and processes of the forecast model and control model. Papers on modeling data sources, and pretreatment of these data, based on three methods were used to establish the prediction model for BOF endpoint temperature and carbon content: linear regression, RBF neural network, BP neural network based on chaos immune particle algorithm optimization. The first two methods are the traditional methods of modeling, but in the process RBF forecast model, this article attempts to nearest neighbor clustering method to select the center of the radial basis function, the center of the hidden layer with accuracy changes, thus avoiding a fall into the danger of local minima. The third method is the use of chaos immune particle algorithm capable of high-speed flight \During two methods of modeling the process, select Jiangsu Yonggang Group BOF shop floor 60 furnace production data for the sample to 9 influencing factors affect the late end of blowing carbon and temperature as input variables were established three RBF neural network structure and three-tier structure of the BP neural network prediction of converter steelmaking endpoint temperature and carbon content. Experiments show that BOF endpoint temperature were established based on three algorithms and the end of the carbon content prediction model based on chaos immune particle algorithm optimized BP neural network BOF endpoint temperature or the end of the carbon content prediction model faster convergence, the forecast accuracy relative higher. So this is ultimately selected as the forecasting model based control model. On the basis of the forecast model, the paper improved area optimization method based on the heat balance and oxygen balance control model, the establishment of a region-based optimization endpoint control model. Experiments show that this method has overcome the traditional control method based on the heat balance and oxygen balance control inaccurate to improve the hit rate of the end point. The thesis of the work done by the last topics of this thesis are summarized, and pointed out the inadequacies of the future work to do in the next stage of the thesis topic.

Related Dissertations

CLC: > Industrial Technology > Metallurgical Industry > Steelmaking > BOF steelmaking > Smelting process and operations
© 2012 www.DissertationTopic.Net Mobile