Prediction of PVA fiber effect in Engineered Composite cement (ECC) by Artificial neural Network (ANN)

High tensile strength is the most vital point in Engineered Composite Cement (ECC) to reduce the dependability of steel reinforcement. In the present paper, the influence of PVA fiber for getting the compressive strength and tensile strength of ECC is predicted by Artificial Neural Networks (ANNs); specifically, the fiber properties and mixture properties are considered at the age of 28 days only. The development of the ANNs model, MATLAB R2020a software, has been used. The most influenceable parameters in the ECC mixture are considered to develop the model. A total of 79 experimental mixtures with 12 input parameters were created from the literature to develop the ANNs model, where 70 % was used for training and 30 % for testing. To ensure the model's accuracy, MSE, RMSE, and R2 were calculated for the ECC mixture. To see the influence of each parameter in the ANNs based ECC model, sensitivity analysis is also conducted. The results demonstrated that utilizing an ANNs model to estimate the compressive and tensile strength of ECC using PVA fiber is a powerful strategy. Copyright (C) 2022 Elsevier Ltd. All rights reserved. Selection and peer-review under responsibility of the scientific committee of the International Conference on Advances in Construction Materials and Structures.

Automated summary

This paper predicts the influence of PVA fiber on the compressive and tensile strength of Engineered Composite Cement (ECC) using Artificial Neural Networks (ANNs). Sensitivity analysis is conducted to study the impact of each parameter. The results demonstrate that employing an ANNs model is an effective strategy.