Evaluation of the Strength of Recycled Aggregate Concrete using Artificial Neural Network

Recycled Aggregate Concrete (RAC), obtained by replacing Natural Aggregates (NA) with Recycled Aggregates (RA) obtained from different sources is highly complex in nature making it hard to predict the strength and study the performance of RAC. Many researchers have reported alternate statistical or mathematical methods in studying the concrete properties and Artificial Neural Network (ANN) is one such tool. Inspired by the biological neural network, ANN has the ability to learn from past examples, capture highly non-linear relationships and adapt itself to a similar situation effectively. Hence the study aims to determine the 7 day and 28 day compressive strength of RAC with suitable architecture under MATLAB program. In the present investigation 83 sets of data from 10 different published literature sources is used to train, test and validate the network. Here, two separate ANN models were developed for 7 and 28 day strength determination. The major conclusion was that the compressive strength of RAC can be determined with fairly high accuracy using ANN tool avoiding large amount of human effort and time into the experimental work.