| Popis: |
This research was conducted to investigate the suitability of using two models; namely the artificial neural network (ANN) and adaptive system network model Neuro-Fuzzy (ANFIS) methods to predict the flexible pavement layer moduli. The falling weight deflectometer (FWD) data obtained from the Butterworth–Kulim Expressway (BKE) in year 2011 were used to test the models. The ANN and ANFIS models have been trained and tested continuously using different parameters until the optimum output is obtained. Thereafter, a validation process has been carried out to test the ability of whole models by calculating the coefficient of determination (R2), the mean absolute error (MAE), the mean squared error (MSE) and root mean squared error (RMSE). A total of 270 data sets have been used to develop the models. Based on the analysis conducted, it was found that both models are able to predict well the flexible pavement layer moduli(R2 > 0.90). However, a further comparison suggested that the ANFIS model yield a higher precision compared to the ANN model, with lower MAE, MSE and RMSE values. Therefore, it can be inferred that the neural network models have an excellent potential to replace the existing analytical and empirical models in predicting the flexible pavement layer moduli. |
