Prediction Model of Concrete Compressive Strength Using Artificial Neural Networks with Backpropagation Algorithm

Unnahar, Zhia and Setiawan, Agustinus Agus and Akbar, Rufman Iman (2022) Prediction Model of Concrete Compressive Strength Using Artificial Neural Networks with Backpropagation Algorithm. Prediction Model of Concrete Compressive Strength Using Artificial. pp. 54-67.

Text Agus_Setiawan_-_Prediction_Model_of_Concrete_Compressive.pdf Download (1MB)

Abstract

This research was conducted to determine the accuracy of the prediction of the artificial neural network on the compressive strength of normal concrete. This study employed the backpropagation algorithm as a form of network architecture to make a predictive model of the compressive strength of concrete. This study used 2 types of data: secondary data which is the composition of the normal concrete mixture as well as the results of the compressive strength of 28 days old concrete and the primary data which is the concrete mix design based on SNI 7656:2012 that had been tested in the laboratory to obtain the concrete compressive strength as validation data from secondary data. The results of this study indicate that the concrete compressive strength prediction model had a percentage of 86.10% for training data; while for data 1 to 3, it had a percentage of 85.77%, 86.77% and 80.76% respectively. Meanwhile, the MSE value generated by the prediction model was 0.0023, 0.0051 and 0.0106 and had a difference between the predicted data and the target data of 0.045, 0.064 and 0.096. In conclusion, the ANN model is quite accurate in predicting the compressive strength of concrete.

Item Type:	Artikel
Subjects:	A General Works > AC Collections. Series. Collected works
Divisions:	Universitas Pembangunan Jaya
Depositing User:	Alexandro Andika
Date Deposited:	25 May 2023 03:44
Last Modified:	25 May 2023 03:44
URI:	http://eprints.upj.ac.id/id/eprint/5285

Actions (login required)

View Item