Development of Fuzzy Model to Backcalculate Surface Modulus from Fwd Data

Vol. 1 No.1

Year: 2011

Issue: Dec-Feb

Title: Development of Fuzzy Model to Backcalculate Surface Modulus from Fwd Data 

Author Name: Mesbah Ahmed, Rafiqul A. Tarefder 

Synopsis: 

Backcalculation of modulus values from Falling Weight Deflectometer (FWD) data is one of the most popular practices all over the world for pavement evaluation. Pavement rehabilitation method and timing depend on the existing stiffness or modulus of pavement layer materials. This study develops a fuzzy model to backcalculate modulus of pavement layers. Specifically, Modified Learning From Example (MLFE) based fuzzy rule is employed to determine modulus from the magnitude of FWD test loads and maximum deflection. To generate training dataset, an axi-symmetric Finite Element Model (FEM) is developed to simulate surface deflections in response to given load and trial layer modulus. These loads, deflections, and modulus of elasticity are then trained by MLFE rule to develop a fuzzy model. Recursive Least Square (RLS) error is used to improve the accuracy of the model by updating the MLFE parameters. Surface moduli from this fuzzy model are compared to those from  BAKFAA, a backcalculation software developed by Federal Aviation Administration (FAA). Results from fuzzy model and BAKFAA are reasonably comparable.