Prediction of Concrete Strength Based on Design Parameters, Hammer Test and Test UPV by Using Artificial Neural Network (ANN).

Abstract

This study aims to predict the compressive strength of existing concrete without using destructive tests which can damage the surface of the concrete. Destructive testing has the disadvantage of damaging the surface of the concrete, requires a long time and need expensive cost, while the Non Destructive Test (NDT) has the advantage of not damaging the surface of the concrete and faster when combined with the Artificial Neural Network (ANN) method.

In this research, the Non Destructive Test (NDT) result such as hammer test and UPV were combined with concrete mix design properties and used to predict the compressive strength of concrete at three and 28 days. The Artificial Neural Network (ANN) method is used to make correlation of mix design properties data and NDT. In this study experimental tests were performed using variation of design parameters such as water per cement ratio and weight ratio of fly ash. The water per cement ratio used in this research was in range 0.45 until 0.55. Furthermore, the weight ratio of fly ash was in range 0% until 25%. Based on the modeling result using ANN method, it found that that the neural network method successfully predicts the compressive strength of concrete at three and 28 days with the mean square error (MSE) value and regression of concrete at three days are5.83 and 0.89 respectively. While at 28 days the MSE and regression value are 4.7 and 0.87 respectively.