Parametric Regression Model and ANN (Artificial Neural Network) Approach in Predicting Concrete Compressive Strength by SonReb Method Lucio Nobile and Mario Bonagura Department DICAM, University of Bologna-Campus of Cesena ,Via Cavalcavia 61, 47521 Cesena, Italy Email: lucio.nobile@unibo.it Abstract The commonly used NDT methods to predict concrete compressive strength include the rebound hammer test and the Ultrasonic Pulse Velocity (UPV) test. The poor reliability of rebound hammer and ultrasonic pulse velocity due to different aspects could be partially contrasted by using both methods together, as proposed.in the SonReb method, developed by RILEM Technical Committees 7 NDT and TC-43 CND. There are three techniques that are commonly used to predict fc based on the SonReb measurements: computational modeling, artificial intelligence, and parametric multi-variable regression models. The aim of this study is to verify the accuracy of some reliable parametric multi-variable regression models and ANN approach comparing the estimated compressive strength based on NDT measured parameters with the effective compressive strength based on DT results on core drilled in adjacent locations. The comparisons show the best performance of ANN approach. Index Terms—concrete strength, SonReb method, parametric regression model, ANN approach I. INTRODUCTION Recent seismic codes give relevance to procedure and methods to establish the performance levels of existing structures. To this end detailed inspections and tests on materials are required. Different sets of material and structural safety factors are therefore required, as well as different analysis procedures, depending on the completeness and reliability of the information available. To this purpose, codes require that a Knowledge Level (KL) is defined in order to choose the admissible type of analysis and the appropriate Confidence Factor (CF) values in the evaluation. The commonly used Non-Destructive Testing (NDT) methods to predict concrete compressive strength fc include the Rebound Hammer test and the Ultrasonic Pulse Velocity (UPV) test. The poor reliability of rebound hammer and ultrasonic pulse velocity methods due to different aspects could be partially contrasted by using both methods together. One of the most employed Manuscript received November 30, 2015; revised January 27, 2016. NDT combined methods in practice is the SonReb method, developed by RILEM Technical Committees 7 NDT and TC-43 CND [1]. There are three techniques that are commonly used to predict fc based on the SonReb measurements: computational modeling, artificial intelligence, and parametric multi-variable regression models. Computational modeling is based on the modeling of complex physical phenomena and thus is often not practical. Parametric multi-variable regression models, on the other hand, can be more easily implemented and used in practice for future applications (such as the reliability assessment of RC structures incorporating field data). Artificial intelligence including the Artificial Neural Network (ANN) is a nonparametric statistical tool without knowing the theoretical relationships between the input and the output. The aim of this study is to verify the accuracy of some reliable parametric multi-variable regression models and ANN approach comparing the estimated compressive strength based on NDT measured parameters with the effective compressive strength based on DT results on core drilled in adjacent locations. To this end a relevant number of DT tests and NDT tests have been performed on many reinforced concrete structures. II. PARAMETRIC REGRESSION MODEL A number of parametric regression models using the SonReb measurements (UPV and RN) to predict the concrete compression strength have been developed. The combined method SonReb can evaluate the concrete compression strength by combining the experimentally obtained non-destructive parameters with correlations as follow: f c =f 0 e a V b (RI) c (1) where: f c is the concrete compression strength, [MPa]; f 0 is the units conversion factor, [usually f 0 = 1MPa s/m]; V is the ultrasonic pulse velocity [m/s]; RI is the rebound index; a,b,c are dimensionless correlation parameters to be determined by regression analysis. International Journal of Structural and Civil Engineering Research Vol. 5, No. 3, August 2016 © 2016 Int. J. Struct. Civ. Eng. Res. 183 doi: 10.18178/ijscer.5.3.183-186 —