Modelling the temperature of a steel strip after roughing mill with Bayesian networks and neural networks. P., L. Ph.D. Thesis, 2000.
abstract   bibtex   
This work deals with the walking beam furnace and roughing mill of the hot strip mill at Rautaruukki Steel mill in Raahe. The goal of the work was to create a model predicting temperature of a steel strip after roughing mill based on measurements done in the furnace. The amount of variables measured in walking beam furnace is large and there is lots of data available. The modelling work was done with Bayesian networks and neural networks. Relationships between variables, groupings of variables and conditional distributions were modelled with Bayesian networks. An adaptive model capable of predicting the temperature after the roughing mill was implemented with neural networks. The Bayesian network formed from the data clarifies the complex process of heating the steel slabs. With the knowledge presented within the network even a person who is not an expert in the area, e.g. a researcher modelling the data, can achieve a basic comprehension of the phenomena. For an expert the network functions as a model compressing knowledge and bringing interesting information of the conditional distributions between variables. An adaptive model based on neural networks was implemented. The goals set for the prediction were met and the results achieved are considered to be very good. The input vectors of neural network models were formed based on the groupings of variables in the Bayesian network. Key words: Bayesian networks, neural networks, adaptive modelling, on-line learning, steel industry, rolling, heating furnace.
@phdthesis{
 title = {Modelling the temperature of a steel strip after roughing mill with Bayesian networks and neural networks},
 type = {phdthesis},
 year = {2000},
 id = {01fd3d0c-bb6c-3305-a0e1-9ed2db6cba63},
 created = {2019-11-19T13:01:22.684Z},
 file_attached = {false},
 profile_id = {bddcf02d-403b-3b06-9def-6d15cc293e20},
 group_id = {17585b85-df99-3a34-98c2-c73e593397d7},
 last_modified = {2019-11-19T13:46:21.495Z},
 read = {false},
 starred = {false},
 authored = {false},
 confirmed = {true},
 hidden = {false},
 citation_key = {isg:575},
 source_type = {mastersthesis},
 notes = {M.Sc. thesis, Department of Mathematics, Statistics, University of Oulu, Finland (in Finnish)},
 private_publication = {false},
 abstract = {This work deals with the walking beam furnace and roughing mill of the hot strip mill at Rautaruukki Steel mill in Raahe. The goal of the work was to create a model predicting temperature of a steel strip after roughing mill based on measurements done in the furnace. The amount of variables measured in walking beam furnace is large and there is lots of data available.

The modelling work was done with Bayesian networks and neural networks. Relationships between variables, groupings of variables and conditional distributions were modelled with Bayesian networks. An adaptive model capable of predicting the temperature after the roughing mill was implemented with neural networks.

The Bayesian network formed from the data clarifies the complex process of heating the steel slabs. With the knowledge presented within the network even a person who is not an expert in the area, e.g. a researcher modelling the data, can achieve a basic comprehension of the phenomena. For an expert the network functions as a model compressing knowledge and bringing interesting information of the conditional distributions between variables.

An adaptive model based on neural networks was implemented. The goals set for the prediction were met and the results achieved are considered to be very good. The input vectors of neural network models were formed based on the groupings of variables in the Bayesian network.

Key words: Bayesian networks, neural networks, adaptive modelling, on-line learning, steel industry, rolling, heating furnace.},
 bibtype = {phdthesis},
 author = {P., Laurinen}
}
Downloads: 0