There is no wealth like Knowledge
                            No Poverty like Ignorance
ARPN Journals

ARPN Journal of Engineering and Applied Sciences >> Call for Papers

ARPN Journal of Engineering and Applied Sciences

Implementation of a prediction model for the mechanical properties of low alloy steel with an artificial neural network using streamlit

Full Text Pdf Pdf
Author Hendra, Desmarita Leni, Dedi Erawadi and Hooi Peng Lim
e-ISSN 1819-6608
On Pages 89-103
Volume No. 21
Issue No. 2
Issue Date March 20, 2026
DOI https://doi.org/10.59018/012620
Keywords prediction, model, mechanical properties, low alloy steel, artificial neural network.


Abstract

Artificial neural networks (ANN) are one of the popular machine learning algorithms used for predicting material properties, particularly due to their capability in handling complex and non-linear data often found in the mechanical properties of materials. This research aims to implement a predictive model for the mechanical properties of low-alloy steel using an ANN in the Streamlit framework, enabling access through a web browser. The ANN model is trained using tensile test data of low alloy steel, which includes chemical elements, heat treatment temperature, and three major mechanical properties: YS (Tensile Strength), TS (Tensile Strength), and EL (Elongation). The model is designed to predict all three mechanical property outputs simultaneously and is evaluated using metrics such as MAE (Mean Absolute Error), RMSE (Root Mean Squared Error), and R-squared. This study also involves improving model performance through parameter tuning and feature selection using the Recursive Feature Elimination (RFE) method. The feature selection results indicate that there are 7 input variables most significant in influencing the mechanical properties of low alloy steel, including Carbon (C), Manganese (Mn), Nickel (Ni), Chromium (Cr), Molybdenum (Mo), Vanadium (V), and heat treatment temperature. The trained model results show that the combination of the Adam optimizer, learning rate of 0.001, batch size of 32, and 100 epochs provides optimal model performance, with MAE of 16.85, RMSE of 24.40, and R-squared of 0.87. Testing the model with new data also yields consistent evaluation results with the training data, with an MAE of 20.87, an RMSE of 33.070, and an R-squared of 0.858. This indicates that the developed ANN model is capable of providing accurate predictions of the mechanical properties of low alloy steel, making it applicable in manufacturing industries for designing low alloy steel mechanical properties suitable for industrial applications.

Back

GoogleCustom Search



Seperator
    arpnjournals.com Publishing Policy Review Process Code of Ethics

Copyrights
© 2026 ARPN Publishers