Dicky, Ghazy (2020) Analisa Pengaruh Komposisi Kimia Terhadap Sifat-Sifat Penyimpanan Hidrogen Logam Paduan AB2 Menggunakan Metode Machine Learning. Other thesis, Institut Teknologi Sepuluh Nopember.
Preview |
Text
02111640000013_Undergraduate_Thesis.pdf Download (2MB) | Preview |
Abstract
Aplikasi hydrogen sebagai bahan bakar kendaraan bermotor membutuhkan hydrogen untuk disimpan pada tangki penyimpanan eksternal, dimana logam hidrida adalah material yang memiliki potensi tinggi untuk menyimpan hydrogen. Dataset logam hidrida AB2 yang terdiri dari komposisi kimia dari logam paduan beserta sifat-sifat penyimpanan hydrogen yang didapatkan dari 50 jurnal penelitian terdahulu dianalisa menggunakan metode machine learning. Prediksi pengaruh dari unsur paduan terhadap energi panas pembentukan (ΔH dan ΔS), phase abundance, dan kapasitas hydrogen (wt%), serta klasifikasi pengaruh dari setiap unsur serta nilai B/A terhadap sifat penyimpanan hidrogen didapatkan dari analisa yang dilakukan menggunakan tiga algoritma machine learning, yaitu multivariate linear regression, decision tree, dan random forest. Skor R2 dan metric error diperoleh dari evaluasi model yang dilakukan. Model random forest menghasilkan skor R2 tertinggi dan error terendah untuk memprediksi seluruh sifat penyimpanan hydrogen pada dataset dengan unsur-unsur independen. Skor R2tertinggi dan error terendah diraih model decision tree untuk prediksi delta entropi, phase abundance, dan persen massa hydrogen, serta model random forest untuk prediksi delta entalphi pada dataset dengan unsur-unsur yang dikelompokkan menjadi unsur A dan B.
========================================================================================================================
Hydrogen is one of the most potential clean and sustainable energy sources to be utilized. The application of hydrogen as fuels requires hydrogen to be stored in several forms, with metal hydrides being the form of storage with high potency. The alloying elements of metal hydrides are frequently studied to discover alloy with the most beneficial and desired properties. However, until recently, standard methods to ascertain the alloying composition is yet present, leaving researchers with the logical design based alloying element determination. In an effort to assist researchers, machine learning approaches are developed to analyze metal hydrides dataset consisting of chemical compositions of alloys as well as hydrogen storage properties obtained from 50 previous research journals, aiming to predict the effect of the alloying element on the heat of formation (ΔH and ΔS), phase abundance, and the hydrogen capacity (wt% H) of the alloy as well as to reveal the importance of the features. Three models are being employed namely multivariate linear regression, decision tree, and random forest which were evaluated by comparing predicted values with the actual values resulting in R2 scores and error metrics. The random forest model generates the best score out of all machine learning models.
| Item Type: | Thesis (Other) |
|---|---|
| Uncontrolled Keywords: | Logam Paduan AB2, Sifat-sifat Hidrogenasi, Machine Learning, Metal Hydrides, Hydrogen Storage, Hydrogen Energy, AB2 Alloys. |
| Subjects: | T Technology > TJ Mechanical engineering and machinery > TJ165 Energy storage. |
| Divisions: | Faculty of Industrial Technology > Mechanical Engineering > 21201-(S1) Undergraduate Thesis |
| Depositing User: | Ghazy Dicky |
| Date Deposited: | 27 Aug 2020 02:37 |
| Last Modified: | 05 Jul 2023 00:44 |
| URI: | http://repository.its.ac.id/id/eprint/80201 |
Actions (login required)
 |
View Item |