Please use this identifier to cite or link to this item:
|Title:||Prediction of Wave Transmission and Damage Level of Tandem Breakwater Using Soft Computing Techniques|
|Keywords:||Department of Applied Mechanics and Hydraulics;Tandem breakwater;Wave transmission|Damage level;Artificial Neural Network (ANN);Support Vector Machine (SVM);Adaptive Neuro-Fuzzy Inference System (ANFIS);Particle Swarm Optimization (PSO)|
|Publisher:||National Institute of Technology Karnataka, Surathkal|
|Abstract:||Tranquility condition inside the port and harbor has to be maintained for loading and unloading cargo and embarking and de-embarking passengers. To maintain calm condition inside the port and harbor, breakwater has to be constructed to dissipate energy of incoming waves. The alignment of the breakwater must be carefully considered after examining the predominant direction of approach of waves and winds, degree of protection required, magnitude and direction of littoral drift and the possible effect of these breakwaters on the shoreline. In general these studies are invariably conducted in a physical model test where various alternatives are studied and the final selection will be based on performance consistent with cost. Mathematical modeling of these complex interactions is difficult while physical modeling will be laborious and uneconomical. Hence soft computing techniques are employed in the present study. Soft computing techniques, such as, Artificial Neural Network (ANN), Support Vector Machine (SVM), Adaptive Neuro-Fuzzy Inference System (ANFIS) and Particle Swarm Optimization (PSO) have been efficiently used for modeling coastal engineering problems. For developing soft computing models for predicting wave transmission and damage level of tandem breakwater, experimental data set by Rao et al. (2004) is collected from M. S. Lab of Applied Mechanics and Hydraulics Department, NITK Surathkal, India. They have conducted the study on the tandem breakwater subjected to regular waves in wave mechanics lab. These data sets are divided into two groups, one for training and other for testing. The input parameters, that influence the wave transmission (H t /H tmax ) over a submerged reef and damage level (S) of conventional rubble mound breakwater of tandem breakwater are relative wave steepness (H i /gT 2 ), the relative spacing (X/d), stability number (H/D n50 ), relative crest widths (B/d), (B/L o ), relative crest height (h/d), relative submergence (F/H i ), relative water depth (d/gT 2 ), which are considered in the present study. The performance of the all models in the prediction of wave transmission and damage level is compared with the measured values using statistical measures, such as, Correlation Coefficient (CC), Root Mean Square Error (RMSE), Nash-Sutcliffe Efficiency (NSE) and Scatter Index (SI). The ANN model is developed for prediction of wave transmission and damage level of tandem breakwater. Predictions of ANN(8-3-1) for wave transmission and ANN(8-5-1) for damage level showed good performance compared with other ANN models with different hidden neurons for 100 epochs. It is observed that the CC test of 0.99 and CC test of 0.956 is obtained between predicted and observed wave transmission and damage levels respectively. Further, to predict wave transmission and damage level of tandem breakwater, SVM model is developed. This technique works on structural risk minimization principle that has greater generalization ability and is superior to the empirical risk minimization principle as adopted in conventional neural network models. Support Vector Machines (SVM) are based on statistical learning theory. The basic idea of support vector machines is to map the original data set into a feature space with three dimensional through a non-linear mapping function and construct an optimal hyper-plane in new space. Four SVM models are constructed using four different kernel functions. In order to know the effectiveness of each kernel function in predicting wave transmission and damage level of tandem breakwater, SVM is trained by applying these kernel functions. Performance of SVM is based on the best setting between of SVM and kernel parameters. SVM model with radial basis kernel function gives CC test of 0.965 for wave transmission and CC test of 0.935 for damage level is considerably better than SVM models with other kernel functions. Adaptive Neuro-Fuzzy Inference System (ANFIS) uses hybrid learning algorithm, which is more effective than the LMA approach used in ANN. ANFIS models with different membership functions namely Triangular-shaped built-in membership function (TRIMF), Trapezoidal-shaped built-in membership function (TRAPMF), Generalized bell-shaped built-in membership function (GBELLMF), and Gaussian curve built-in membership function (GAUSSMF) are developed to predict wave transmission and damage level of tandem breakwater. ANFIS model with GAUSSMF gave realistic prediction compared with the observed values with CC test of 0.935 for wave transmission and CC test of 0.875 for damage level. It is observed that the ANN models yield higher CC compared to ANFIS models. However, it is noticed that ANN model in isolation cannot capture all data patterns easily. ANN models are developed with particle swarm optimization (PSO). PSO tuned ANN (PSO-ANN) model is developed to predict wave transmission and damage level of tandem breakwater. The performance of the PSO-ANN(8-3-1) model in the prediction of wave transmission gives CC test of 0.879 and PSO-ANN(8-2-1) model gives CC test of 0.589 for damage level compared with the measured values. PSO-ANN(8-3-1) and PSO-ANN(8-2- 1) models prediction are not matched well with observed values and performed poor compared to the all other hybrid models. Further, PSO is applied to avoid over-fitting or under-fitting of the SVM model due to the improper selection of SVM and kernel parameters. SVM and kernel parameters are optimized using PSO optimization technique. PSO-SVM model is developed to predict wave transmission and damage level of tandem breakwater. The performance of the PSO- SVM model with polynomial kernel function gives CC test of 0.984 for wave transmission prediction and PSO-SVM model with radial basis kernel function give CC test of 0.941 for damage level compared with the measured values. The results are found to be reliable. The different soft computing models namely ANN, SVM, ANFIS, PSO-ANN and PSO- SVM and results are compared in terms of CC, RMSE, SI and NSE. ANN performed well and showed good results compared to SVM models in prediction of wave transmission and damage level of tandem breakwater. PSO-SVM model performed better in both cases of wave transmission and damage level prediction compared to other hybrid models with higher CC, NSE and lower RMSE, and SI. PSO-SVM with polynomial kernel and PSO- SVM with radial basis kernel function give higher CC, NSE and lower RMSE, SI compared to all the Individual and hybrid soft computing models. Therefore, PSO-SVM model can be utilized as an alternate soft computing technique to provide accurate and reliable solution in prediction of the wave transmission and damage level of tandem breakwater.|
|Appears in Collections:||1. Ph.D Theses|
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.