Location:Home > Engineering science > Computer Science > Computer System Architecture > Recognition and Implementation of Multi-sintering Condition Based on Complete Binary Tree Supporting

Recognition and Implementation of Multi-sintering Condition Based on Complete Binary Tree Supporting

Downloads: []
Tutor: JiangHuiYan
School: Northeastern University
Course: Computer System Architecture
Keywords: Sintering Condition,Complete Binary Tree,Feature Extraction,Pattern Recognition,
CLC: TP391.41
Type: Master's thesis
Year:  2009
Facebook Google+ Email Gmail Evernote LinkedIn Twitter Addthis

not access Image Error Other errors

In the process of aluminum oxide production of large-scale rotary kiln, variety of condition changes and improper operations influence the internal sintering condition of rotary kiln to cause bad system performance, which results in low quality of the aluminum oxide product.In the production of aluminum oxide of large-scale rotary kiln in our country, judging sintering condition remains depends on manual fire watching experience. However, this process is restrained by workers¡¯subjective experience, which lacks objectivity and rationality. Recently, with the increasing development of computer technology, automatic control of rotary kiln has been achieved. How to make full use of computer technology to help the site operators to distinguish each kind of sintering condition accurately and quickly, make the corresponding diagnosis, and finally implement the automatic detection of sintering condition of rotary kiln is still a unsolved question now.After referring to the massive advanced research results at home and abroad, and supported by the program of "large-scale rotary kiln intelligent control system", which is a subproject of the national 863 high-tech key plan project named "low-cost intelligent control system of Process Manufacturing", this thesis proposes a recognition method for many kinds of sintering condition, which bases on the complete binary tree supporting vector machines (SVM). Firstly, sintering images are made gray, denoised and segmented, and the characteristics of the various parts of the image are extracted. Secondly, the combinatorial optimization method of the Relief Features (the ReliefF-PCA method) is used to reduce the characteristic set and evaluate the effect of the reduced characteristic set in order to get the optimal characteristic set. Finally, a classification model that is founded based on the complete binary tree supporting vector machines is used to recognize many kinds of sintering condition.The experimental results indicate that the rate of accuracy is high by using the method introduced in this thesis to distinguish different kinds of sintering condition, which has better robustness than other methods.
Related Dissertations
Last updated
Sponsored Links
Home |About Us| Contact Us| Feedback| Privacy | copyright | Back to top