A neural network approach to the modeling of blast furnace
Year: 1999
Type: dissertation
Abstract: The inside of a blast furnace process is very complicated and impossible to model mathematically. A new approach in this field relies on using artificial neural networks which have proven successful in dealing with complex systems that involve many variables. One of the most important parameters to model is the hot metal temperature - the temperature the output of the blast furnace, pig iron. This thesis presents a multi-layer perceptron neural network which predicts hot metal temperature based on 11 input variables and the actual output value from the previous time period. The best result achieved was a mean squared error of 0.0086 between the predicted hot metal temperature and the actual hot metal temperature. 2 Acknowledgment I would like to thank the following people: Dr. Amar Gupta for providing direction and guidance on this project. Brad Banks and Ravi Sarma for collaborating with me on this work. Ashish Agarwal, Sanjeev Vadhavkar for their help and input on blast furnace and neural networks. Most importantly, I want to thank my parents for their love and support for the past twenty two years. 3 Table of Contents 1.0
Author: Angela X. Ge
Institution: Massachusetts Institute of Technology
Language: English
Subfield: Metals and Alloys
Field: Materials Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W2246183835
