Artificial Neural Network Model for Predicting the Compressive Strength of Concrete using Ultrasonic Pulse Velocity

Muthanna Journal of Engineering and Technology

Volume (5), Issue (1), Year (30 April 2017), Pages (72-79)


Research Article By:

Salim T. Yousif, Omar M. Abdul-Kareem and Kaythar A. Ibrahim

Corresponding author E-mail:


This paper presents the results of study conducted with artificial neural networks (ANN) to determine the effects of the variations of concrete constituents on ultrasonic pulse velocity (UPV) and developed mathematical model to predict the compressive strength of concrete. The proposed input parameters are major factors that affect (UPV), such as cement content, water–cement ratio (W/C), the aggregate–cement ratio (A/C), maximum aggregate size, and age of concrete. The output parameter is the (UPV). The results show that (UPV) increased with the increase in concrete age. Increasing the cement content caused a rapid pulse in velocity readings, and (UPV) increased with the increase in maximum aggregate size. Aside from these factors, (W/C) negatively affected pulse velocity. Also, the ANN model was built to predict the compressive strength of the concrete using pulse velocity and the age of concrete. The results showing good rapprochement between experimental value of compressive strength with predicated value of compressive strength.

Keywords: Concrete constituents, ultrasonic pulse velocity, artificial neural networks, concrete compressive strength.

Get Full Access / pdf