Eine Plattform für die Wissenschaft: Bauingenieurwesen, Architektur und Urbanistik
Eine Plattform für die Wissenschaft: Bauingenieurwesen, Architektur und Urbanistik
Predicting Non-deposition Sediment Transport in Clean Pipes: Assessing Heuristic Models
https://orcid.org/http://orcid.org/0000-0002-1240-9036
Sediment transport without deposition at sewer system is mandatory to find a reliable model to prevent sedimentation. A comprehensive examination of different input combinations, as well as the training model, is done to overcome the limitation of existing models by developing a simple yet accurate model for forecasting sediment transport in clean pipes. In this study, the Froude number of three-phase flow in circular sewage channels is estimated using the extreme learning machine (ELM) model. This method is one of the powerful and rapid artificial intelligence methods in predicting complex and nonlinear phenomena. The extreme learning machine acts very rapid in the learning process compared to other learning algorithms and has an acceptable performance in processing the generation function. Also, using the parameters affecting on the Froude number, 127 different ELM models are created. Furthermore, four modes of training and test including 50% training 50% test, 60% training 40% test, 70% training 30% test and 80% training and 20% test are used for modeling 127 ELM models. Subsequently, by analyzing the ELM models, the superior model is introduced. For example, the values of R2, MARE and RMSE for the superior model in the test mode are calculated as 0.856, 0.117 and 0.738, respectively. Furthermore, the results of the superior model are compared with the artificial neural network models, the support vector machine and ANFIS model. Studying of the modeling results indicates the high accuracy of the ANFIS model.
Predicting Non-deposition Sediment Transport in Clean Pipes: Assessing Heuristic Models
https://orcid.org/http://orcid.org/0000-0002-1240-9036
Sediment transport without deposition at sewer system is mandatory to find a reliable model to prevent sedimentation. A comprehensive examination of different input combinations, as well as the training model, is done to overcome the limitation of existing models by developing a simple yet accurate model for forecasting sediment transport in clean pipes. In this study, the Froude number of three-phase flow in circular sewage channels is estimated using the extreme learning machine (ELM) model. This method is one of the powerful and rapid artificial intelligence methods in predicting complex and nonlinear phenomena. The extreme learning machine acts very rapid in the learning process compared to other learning algorithms and has an acceptable performance in processing the generation function. Also, using the parameters affecting on the Froude number, 127 different ELM models are created. Furthermore, four modes of training and test including 50% training 50% test, 60% training 40% test, 70% training 30% test and 80% training and 20% test are used for modeling 127 ELM models. Subsequently, by analyzing the ELM models, the superior model is introduced. For example, the values of R2, MARE and RMSE for the superior model in the test mode are calculated as 0.856, 0.117 and 0.738, respectively. Furthermore, the results of the superior model are compared with the artificial neural network models, the support vector machine and ANFIS model. Studying of the modeling results indicates the high accuracy of the ANFIS model.
Predicting Non-deposition Sediment Transport in Clean Pipes: Assessing Heuristic Models
https://orcid.org/http://orcid.org/0000-0002-1240-9036
Sediment transport without deposition at sewer system is mandatory to find a reliable model to prevent sedimentation. A comprehensive examination of different input combinations, as well as the training model, is done to overcome the limitation of existing models by developing a simple yet accurate model for forecasting sediment transport in clean pipes. In this study, the Froude number of three-phase flow in circular sewage channels is estimated using the extreme learning machine (ELM) model. This method is one of the powerful and rapid artificial intelligence methods in predicting complex and nonlinear phenomena. The extreme learning machine acts very rapid in the learning process compared to other learning algorithms and has an acceptable performance in processing the generation function. Also, using the parameters affecting on the Froude number, 127 different ELM models are created. Furthermore, four modes of training and test including 50% training 50% test, 60% training 40% test, 70% training 30% test and 80% training and 20% test are used for modeling 127 ELM models. Subsequently, by analyzing the ELM models, the superior model is introduced. For example, the values of R2, MARE and RMSE for the superior model in the test mode are calculated as 0.856, 0.117 and 0.738, respectively. Furthermore, the results of the superior model are compared with the artificial neural network models, the support vector machine and ANFIS model. Studying of the modeling results indicates the high accuracy of the ANFIS model.
Predicting Non-deposition Sediment Transport in Clean Pipes: Assessing Heuristic Models
Iran J Sci Technol Trans Civ Eng
Yosefvand, Fariborz (Autor:in) / Rajabi, Ahmad (Autor:in) / Shabanlou, Saeid (Autor:in)
01.02.2022
15 pages
Aufsatz (Zeitschrift)
Elektronische Ressource
Englisch
Sediment transport in pressure pipes
| UB Braunschweig | 2021
Graded Sediment Transport at Limit Deposition in Clean Pipe Channel
| British Library Conference Proceedings | 1999
Sediment transport in corrugated pipes
| British Library Conference Proceedings | 2002
Optimal design of sediment transport in pipes
| British Library Conference Proceedings | 1995