Pengembangan Algoritma Random Forest untuk Klasifikasi Potensi Recharge Area Lapangan Panas Bumi Lumut Balai Sumatera Selatan

Pratama, Dandi Syahtia (2024) Pengembangan Algoritma Random Forest untuk Klasifikasi Potensi Recharge Area Lapangan Panas Bumi Lumut Balai Sumatera Selatan. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5017201044-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2026. Download (2MB) | Request a copy

Abstract

Panas bumi merupakan salah satu renewable energi yang saat ini banyak digunakan. Namun, energi panas bumi di suatu lapangan dapat dikatakan renewable apabila dilakukan pengelolaan, pemeliharaan, dan monitoring secara memadai. Salah satu aspek yang harus diperhatikan adalah recharge area yang menjadi faktor penting produksi panas bumi. Penelitian ini bertujuan untuk memetakan potensi recharge area di lapangan panas bumi Lumut Balai. Potensi rechare area ditentukan berdasarkan tiga parameter geologi seperti Complete Bouger Anomaly (CBA) infiltrasi jenis batuan, dan kepadatan linemanet. Dengan adanya beberapa parameter berbeda, digunakan metode Machine Learning algoritma Random Forest (RF) dalam penelitian ini untuk klasifikasi memodelkan potensi recharge area berdasarkan multiparameter. Model dibagun melalui beberapa tahap, dataset dari tiga parameter dinormalisasi untuk mempermudah proses training algoritma. Dataset dibagi menjadi data training dan testing dengan rasio 45:55. Model dibangun berdasakan dua hyperparameter model n_estimators atau banyaknya pohon keputusan yang dibangun dan max_depth atau kedalaman maksimal dari pohon keputusan yang nilainya didapat melalui motode grid search cross-validation. Berdasarkan metode evaluasi confusion matrix, hasil paling optimal didapatkan dengan akurasi mencapai 0.930, precision 0.935, dan F1-Score 0.929. Model menujukan performa yang baik. Namun, korelasi antar parameter training rendah dan tingkat pengaruh masing-masing parameter dalam membangun model tidak merata. Kondisi parameter yang tidak terlalu kompleks ini juga menjadi indikasi mengapa performa algoritma sangat tinggi dalam membangun klasifikasi potensi recharge area.
=================================================================================================================================
Geothermal is one of the renewable energies that is currently widely used. However, geothermal energy in a field can be said to be renewable if adequate management, maintenance, and monitoring are carried out. One aspect that must be considered is the recharge area which is an important factor in geothermal production. This research aims to map the potential recharge area in the Lumut Balai geothermal field. The recharge area potential is determined based on three geological parameters such as Complete Bouguer Anomaly (CBA), rock type infiltration, and lineament density. With several different parameters, the Random Forest (RF) algorithm Machine Learning method is used in this research for classification to model the potential recharge area based on multiparameter. The model is built through several stages, the dataset of three parameters is normalized to facilitate the algorithm training process. The dataset is divided into training and testing data with a ratio of 45:55. The model is built based on two hyperparameter models n_estimators or the number of decision trees built and max_dept or the maximum depth of the decision tree whose values are obtained through the grid search cross-validation method. Based on the confusion matrix evaluation method, the most optimal results were obtained with accuracy reaching 0.930, precision 0.935, and F1-Score 0.929. The model performed well. However, the correlation between training parameters is low and the degree of influence of each parameter in building the model is uneven. The less complex condition of the parameters is also an indication of why the performance of the algorithm is very high in building the classification of potential recharge areas.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Confusion Matrix, Hyperparameter, Machine Learning, Random Forest, Recharge Area
Subjects:	Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines. Q Science > QE Geology > QE601 Geology, Structural
Divisions:	Faculty of Civil, Planning, and Geo Engineering (CIVPLAN) > Geophysics Engineering > 33201-(S1) Undergraduate Thesis
Depositing User:	Dandi Syahtia Pratama
Date Deposited:	15 Aug 2024 03:18
Last Modified:	15 Aug 2024 03:18
URI:	http://repository.its.ac.id/id/eprint/114452

Actions (login required)

View Item