Principle Component Analysis in Conjuction with Data Driven Methods for Sediment Load Prediction
- PDF / 383,507 Bytes
- 14 Pages / 439.37 x 666.142 pts Page_size
- 58 Downloads / 192 Views
Principle Component Analysis in Conjuction with Data Driven Methods for Sediment Load Prediction Gokmen Tayfur & Yashar Karimi & Vijay P. Singh
Received: 7 November 2012 / Accepted: 29 January 2013 / Published online: 9 February 2013 # Springer Science+Business Media Dordrecht 2013
Abstract This study investigates sediment load prediction and generalization from laboratory scale to field scale using principle component analysis (PCA) in conjunction with data driven methods of artificial neural networks (ANNs) and genetic algorithms (GAs). Five main dimensionless parameters for total load are identified by using PCA. These parameters are used in the input vector of ANN for predicting total sediment loads. In addition, nonlinear equations are constructed, based upon the same identified dimensionless parameters. The optimal values of exponents and constants of the equations are obtained by the GA method. The performance of the so-developed ANN and GA based methods is evaluated using laboratory and field data. Results show that the expert methods (ANN and GA), calibrated with laboratory data, are capable of predicting total sediment load in field, thus showing their transferability. In addition, this study shows that the expert methods are not transferable for suspended load, perhaps due to insufficient laboratory data. Yet, these methods are able to predict suspended load in field, when trained with respective field data. Keywords Principle component analysis . Sediment load . Artificial neural network . Genetic algorithm . Transferability
1 Introduction Since estimates of sediment loads are required in many fields of water resource engineering, considerable effort has been devoted to sedimentation engineering (Jain 2001; Tayfur 2002; G. Tayfur (*) Department of Civil Engineering, Izmir Institute of Technology, Izmir, Turkey e-mail: [email protected] Y. Karimi Department of Civil Engineering, Ege University, Izmir, Turkey e-mail: [email protected] V. P. Singh Department of Civil Engineering, Texas A and M University, College Station, USA e-mail: [email protected]
2542
G. Tayfur et al.
Bhattacharya et al. 2007; Dogan et al. 2009, among others). Most of the existing models are based on combinations of several characteristics of flow and channel geometry, and sediment dynamics parameters. Zhu et al. (2006) summarizes the parameters used in several commonly employed models. Using artificial neural networks (ANNs), Bhattacharya et al. (2007) estimated sediment loads employing dimensionless parameters based mainly on studies of Yalin (1977) and van Rijn (1984). Bhattacharya et al. (2007) considered two scenarios by employing different sets of input variables to predict dimensionless total sediment transport rate. In the first scenario, they employed dimensional parameters of u (flow velocity), h (flow depth), D (particle diameter), and I (slope); and in the second scenario, they used D* (particle parameter), T (transport stage parameter), and h/D to predict ϕt (dimensionless total sediment transport rate).
Data Loading...
