New strategies for estimation of cut and fill areas in road design with different ground cross-section offsets |
| |
Authors: | Razieh Babapour Ismael Ghajar Reza Ghodsi |
| |
Institution: | 1. Department of Forestry, Faculty of Natural Resources, University of Guilan, Sowmeh, Iran;2. Engineering Department, Central Connecticut State University, New Britain, CT, USA |
| |
Abstract: | Earthwork operations accounted for great amount of forest road construction costs. Any Optimization in planning road profile needs a reliable objective function that calculates earthwork volume in an appropriate time and accuracy. This study aimed at incorporating the shape of the ground cross sections in calculating cut/fill areas using three alternative methods including: Multiple Linear Regression (MLR), Artificial Neural Network (ANN) and heuristic programming in MATLAB environment. The accuracy of results was validated via AutoCAD as a manual method for area calculation of road’s cross-section. Trapezoidal method applied to calculate the area of cut/fill in the heuristic method. Paired sample T-Test and statistical analysis (ANOVA) with confidence level of 0.95 indicated that the heuristic method not only reduced computing time but also provided accurate estimations as same as the manual AutoCAD based method. It can be concluded that although MLR and ANN estimations were as well as AutoCAD results within a reasonable time, they need to be retrained by adding, removing or changing entrance parameters. |
| |
Keywords: | Forest road earthwork volume artificial neural network MATLAB multiple linear regression heuristic |
|
|