Click for new scientific resources and news about Corona[COVID-19]

Paper Information

Title: 

USING ANNS AND REGRESSION TREES APPROACHES TO ESTIMATE SCOUR DEPTH AROUND A CIRCULLAR PILE DUE TO WAVE IN MEDIUM DENSE SILT AND SAND BED

Type: PAPER
Author(s): AYOUBLOO MOHAMMAD KARIM,ETEMAD SHAHIDI AMIR,MAHJOOBI JAVAD
 
 
 
Name of Seminar: INTERNATIONAL CONFERENCE ON COASTS, PORTS AND MARINE STRUCTURES (ICOPMAS) PORTS & MARITIME ORGANIZATION
Type of Seminar:  CONFERENCE
Sponsor:  PORTS AND MARITIME ORGANIZATION
Date:  2008Volume 8
 
 
Abstract: 

PREDICTION OF SCOUR AROUND A PILE DUE TO OSCILLATORY WAVE ACTION IS VERY IMPORTANT IN MANY OFFSHORE AND MARINE ENGINEERING PROBLEMS. BECAUSE OF COMPLEXITY OF SCOUR PROCESS, MOST OF THE EMPIRICAL FORMULAS ARE UNABLE TO ESTIMATE SCOUR HOLE DEPTH ACCURATELY. ARTIFICIAL NEURAL NETWORKS (ANNS) AND REGRESSION TREES ARE EFFICIENT PROCEDURES TO UNDERSTAND AND MODEL COMPLEX SYSTEMS WITH AMBIGUOUS RELATIONS. A MULTI LAYER PERCEPTRON (MLP) IS ONE OF THE MOST COMMON KINDS OF ANNS AND HAS BEEN USED TO MAP INPUT-OUTPUT SYSTEMS WHILE CART ALGORITHM WAS EMPLOYED FOR BUILDING AND EVALUATING REGRESSION TREES. IN THE PRESENT STUDY, TWO INPUT SETS WERE EMPLOYED TO ESTIMATE SCOUR DEPTH: (A) DIMENSIONAL PARAMETERS SUCH AS BED GRAIN SIZE, PILE DIAMETER, WAVE PERIOD, WAVE HEIGHT, MAXIMUM FLOW VELOCITY AND MAXIMUM SHEAR VELOCITY (B) NONDIMENSIONAL PARAMETERS SUCH AS PILE REYNOLDS NUMBER, SHIELDS PARAMETER, KEULEGAN- CARPENTER NUMBER, GRAIN REYNOLDS NUMBER, SEDIMENT NUMBER AND RELATIVE DENSITY. OUTPUT PARAMETER WAS NONDIMENSIONAL EQUILIBRIUM SCOUR DEPTH. THE TESTS RESULTS REVEAL THAT A MLP WITH BACK PROPAGATION LEARNING RULE AND CART MODEL BASED ON NONDIMENSIONAL PARAMETERS CAN PREDICT SCOUR HOLE DEPTH BETTER THAN THE EXISTING EMPIRICAL FORMULA. ALSO, A SENSITIVITY ANALYSIS WAS CARRIED OUT AND IT SHOWED THAT KEULEGAN-CARPENTER NUMBER AND WAVE HEIGHT ARE THE MOST IMPORTANT PARAMETERS IN SCOUR PROCESS.

 
Keyword(s): ARTIFICIAL NEURAL NETWORK, PILE, REGRESSION TREE, SCOUR, WAVE
 
 
Yearly Visit 33   pdf-file tarjomyar
 
Latest on Blog
Enter SID Blog