Regression model for back-calculation of rigid-pavement properties

Abstract

Computer algorithms and graphical solutions for deflection-based closed-form back-calculation of rigid pavement properties are available. However, a solution in the form of regression equations would have useful practical applications because of the speed and convenience in computation. This work presents a regression model for this purpose based on the closed-form back-calculation algorithm NUS-BACK. The input parameters remain the same as the original closed-form algorithm, with no requirements for seed values or initial guess of pavement properties. The database for the development of the regression model was established by forward computation of deflection for practical ranges of pavement properties. The verification of the regression model is presented by checking against the solutions of the closed-form algorithm NUS-BACK. The reliability of the regression equations was assessed against the closed-form algorithm for two conditions of input deflections: one for exact deflection measurements with no errors, another for deflection measurements with random errors. The analysis showed that the regression model could provide comparable reliability and accuracy to the analytical closed-form solution by NUS-BACK.