Efficient production of increasingly popular high strength steel grades necessitates not just the ability to achieve a required coiling temperature, but to closely control the cooling path. The runout table plays an important role in producing steel strip which consistently meets precisely specified mechanical properties. Close inspection of the present control system used on British Steel Strip Products' Port Talbot Work's hot strip mill runout table showed that the main source of error within the system was inaccuracy in the temperature predictions used by the control system. These predictions were used as the basis for the feedfoward control system. Trials were carried out to assess the viability of using artificial neural network models instead of the present conventional mathematical model in order to increase the accuracy of temperature predictions. A methodology was developed so that data taken from the runout table could be used to build artificial neural networks capable of modelling the cooling that took place. The method used to select the training data was seen to be very important in obtaining an accurate model. Using this artificial neural network approach it was possible to develop models that were capable of predicting temperatures for groups of coils of similar grade material. The accuracy of these predictions were in general similar to the predictions provided by the present on-line model, and for one group of carbon manganese grades, they were significantly better. Finally, the overall conclusion from this work is that this approach in modelling the runout table cooling is a viable alternative in providing an on-line model, and details are given of the further work needed in order to fully optimise the benefits and reduce the risks associated with altering the present control system.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:637947 |
| Date | January 1998 |
| Creators | Loney, D. W. |
| Publisher | Swansea University |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
Page generated in 0.0024 seconds