PREDICTION OF COMPRESSIVE STRENGTHOF METAKAOLIN BLENDED WITH CONCRETE USING ANN
Pasupuleti Revathi1, B. Ajitha2
1PG-Scholar, Department of CIVIL Engineering, JNTUA College of Engineering (Autonomous) Ananthapuramu, India.
2Assistant Professor, Department of CIVIL Engineering, JNTUA College of Engineering (Autonomous) Ananthapuramu, India.
ABSTRACT: Compressive strength mainly depends on the ingredients of concrete mix design. Concrete is generally used as construction material. Due to the vast construction in urban areas, there is high demand of concrete. The work is experimentally carried out by partial replacement of Ordinary Portland Cement (OPC) with Metakaolin (MK) additive and total replacement of fine aggregate that is river sand with steel slag sand. The cement content will be replaced by 0%,5%,10%,15% and 20% of Metakaolin in the grade of concrete M60 at 3 days,7 days and 28 days. Due to the replacement of the Pozzolanic material and fine aggregate the strength properties will be achieved. Artificial Neural Networks (ANN) is used to predict the strength properties. ANN has three layers which include output, input and hidden layer. The input layer consists of the quantity of cement, coarse aggregate, water content, percentage of Metakaolin and steel slag sand. The output consists of compressive strength of concrete. While developing ANN model 45 samples will be used as training testing data sets. Two assessments will be carried out one is to determine the effective number of neurons in the hidden layer for predicting the network system and second is to evaluate the accuracy of predicted network will be done under different load conditions. Generally Artificial neural network learns from training and gives extremely good results. ANN can be used to escalate the experimental data to determine the compressive strength of concrete. High accuracy outcomes might be observed when compared with the experimental results and results obtained after training of neural network.
KEYWORDS: Compressive strength, Metakaolin (MK), Steel slag sand, Artificial Neural Networks (ANN).