Article Per Year (5 Year)
p-Index From 2021 - 2026
0.444
P-Index
This Author published in this journals
All Journal Journal of Physics: Theories and Applications
Eko Sulistya
Physics Department FMIPA UGM Yogyakarta
Education Materials Science & Nanotechnology Physics

Published : 2 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 2 Documents
Search

 1 
Predicting Newtonian cooling with machine learning: a comparative analysis of gradient boosting and random forest models Eko Sulistya
Journal of Physics: Theories and Applications Vol 9, No 2 (2025): Journal of Physics: Theories and Applications
Publisher : Universitas Sebelas Maret

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.20961/jphystheor-appl.v9i2.105932

Abstract

This study investigates the use of artificial intelligence, specifically machine learning models, to predict temperature reduction in Newtonian cooling experiments involving varying volumes of water. Two regression models, Gradient Boosting Regression and Random Forest Regressor, were utilized to learn from empirical data. The findings indicate that both models are capable of accurately predicting cooling behavior, with the Random Forest model demonstrating superior accuracy for the dataset used. The machine learning models effectively represent the theoretical model of Newton’s Law of Cooling, which is characterized by an exponential decay curve. Furthermore, the cooling constant for each volume was estimated using curve fitting techniques. This research underscores the potential of AI in modeling complex physical processes, particularly in real-world scenarios where the relationships between physical variables are intricate and challenging to express analytically. With sufficient data, AI can adeptly predict variable changes based on fluctuations in others. As technology continues to advance, AI is poised to assume an increasingly critical role in experimental and industrial applications involving complex physical systems. The novelty of this study lies in its comparative analysis to identify the optimal machine learning model—Gradient Boosting Regression or Random Forest Regressor—for accurately predicting Newtonian cooling behavior. Additionally, this research introduces an automated data acquisition approach using a datalogger, significantly enhancing precision and practicality compared to traditional manual methods involving a stopwatch and thermometer.
A study on the characteristics of 1H NMR spectra and evaluation of the sensitivity of an electromagnetic induction system to differences in the research octane number of petroleum fuels Rohmah Insyirah Handayani; Bambang Murdaka Eka Jati; Eko Sulistya
Journal of Physics: Theories and Applications Vol 10, No 1 (2026): Journal of Physics: Theories and Applications
Publisher : Universitas Sebelas Maret

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.20961/jphystheor-appl.v10i1.114659

Abstract

The Research Octane Number (RON) is a key quality parameter of gasoline that reflects the molecular structural characteristics of hydrocarbons; however, its conventional determination relies on standardized engine testing, which is impractical for rapid laboratory analysis. This study aims to analyze the characteristics of 1H NMR spectra of gasoline with different RON values and to evaluate the potential of alternative approaches based on electromagnetic induction and capacitive methods as discriminative parameters for RON. The methodology includes 1H NMR spectral analysis using region-based integration of chemical shifts without individual compound identification, as well as evaluation of the response of mutual induction systems and RC and RLC capacitive circuits to various gasoline samples. The results show that 1H NMR spectra exhibit clear differences in the distribution of aliphatic and aromatic signals among gasoline samples with different RON values, whereas the electromagnetic induction system does not demonstrate sufficient sensitivity due to the non-magnetic nature of gasoline. The capacitive approach is capable of detecting media with large permittivity contrast but is not yet sufficiently sensitive to discriminate subtle variations among gasoline samples. This study provides a methodological basis for the application of 1H NMR in the molecular structural characterization of gasoline and a critical evaluation of the limitations of sensor-based approaches relying on electromagnetic responses.
Co-Authors Bambang Murdaka Eka Jati Rohmah Insyirah Handayani