Title: Statistical Comparisons of Creep and Shrinkage Prediction Models Using RILEM and NU-ITI Databases

Author(s): Akthem Al-Manaseer and Armando Prado

Publication: Materials Journal

Volume: 112

Issue: 1

Appears on pages(s): 125-136

Keywords: creep; creep prediction models; shrinkage; shrinkage prediction models

DOI: 10.14359/51686982

Date: 1/1/2015

Abstract:
Six shrinkage and creep compliance models were evaluated according to the NU-ITI and RILEM databases. The prediction models include ACI 209R-92, B3, GL 2000, CEB MC 90-99, fib MC 2010, and AASHTO 2012. Five statistical methods were used to evaluate the models. The statistical methods include the residual method, the CEB coefficient of variation VCEB, the CEB mean square error FCEB, the CEB mean deviation MCEB, and the new modified coefficient of variation method ωm. Results indicate that for shrinkage predictions, ACI 209R-92 performed best, followed by B3, CEB MC 90-99, fib MC 2010, and GL 2000 models. The AASHTO 2012 model received the lowest ranking. For creep compliance, ACI 209R-92 had the best performance, followed by the B3 and GL 2000 models. The CEB MC 90-99 model ranked third, the fib MC 2010 model ranked fourth, and the AASHTO 2012 model ranked fifth. It should be noted that the data selection criteria and the database used to assess the models can influence the final ranking conclusions. Other statistical methods might also influence the rankings.