dc.contributor.advisor | Amir | |
dc.contributor.author | Farzin, Amir | |
dc.date.accessioned | 2021-10-16T16:41:13Z | |
dc.date.issued | 2021 | |
dc.identifier | no.usn:wiseflow:6341482:45759367 | |
dc.identifier.uri | https://hdl.handle.net/11250/2823474 | |
dc.description | Full text not available | |
dc.description.abstract | Data-driven modeling of the decarbonization section in an ammonia plant was the main purpose of this project. The available data from plant historians was used for the prediction of CO2 slip from the decarbonization section. Both linear and non-linear modeling approaches was used. FIR modeling with different orders and different regression techniques were used for the linear part. For non-linear modeling, long short-term memory (LSTM) networks and convolutional neural networks (CNNs) were used. All the models illustrated the data is not sufficient for the modeling purpose. None of the models could predict the target values in an acceptable manner. The problem with the data is discussed and explained why the modeling was impossible with available data. | |
dc.description.abstract | | |
dc.language | eng | |
dc.publisher | University of South-Eastern Norway | |
dc.title | Comparison between finite impulse response (FIR) and deep neural network (DNN) models for prediction of the CO2 slip in the CO2 removal system of ammonia plant | |
dc.type | Master thesis | |