Huda, Ahmad Miftahul (2023) Prediksi Konsumsi Energi Listrik Berbasis Citra Satelit Pada Bangunan Di Kec. Klojen Kota Malang Dengan Metode Machine Learning. Other thesis, Institut Teknologi Sepuluh Nopember.
Text
02311940000058-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2025. Download (4MB) | Request a copy |
Abstract
Prediksi konsumsi energi listrik dibutuhkan untuk pengembangan infrastruktur kelistrikan. Penelitian tugas akhir ini menggunakan citra bangunan dari satelit untuk mendapatkan luas, keliling, dan jumlah bangunan yang merupakan fitur prediksi, fitur ini dideteksi menggunakan U-Net dengan backbone ResNet-34. Fitur-fitur yang didapatkan dari bangunan terdeteksi kemudian diagregasi per kelurahan bersama dengan data PLN yang meliputi jumlah konsumsi daya listrik dan jumlah pelanggan. Dalam penelitian ini dibuat tiga model machine learning yaitu jaringan syaraf tiruan (JST), regresi linier, dan random forest regression, dengan tiap model memiliki dua variasi yaitu prediksi dengan data pelanggan PLN dan prediksi dengan data citra satelit. Hasil dari setiap model dievaluasi menggunakan nilai R2 dan RMSE. Hasil penelitian menunjukkan bahwa model random forest regression memberikan prediksi paling akurat dengan R2 sebesar 0,9866 dan RMSE sebesar 0,0295. Selain itu, fitur luas bangunan memiliki pengaruh yang signifikan terhadap hasil prediksi model random forest regression. Secara keseluruhan, penelitian ini menunjukkan bahwa menggunakan citra satelit dan metode machine learning dapat memberikan prediksi yang akurat untuk konsumsi energi listrik di Kecamatan Klojen.
=================================================================================================================================
The prediction of electricity consumption is essential for the development of power infrastructure. This final project research utilizes satellite imagery of buildings to obtain features for prediction, such as area, perimeter, and the number of buildings. These features are detected using U-Net with ResNet-34 as the backbone. The extracted features from the detected buildings are then aggregated per neighborhood along with PLN data, including the number of power consumption and customers. Three machine learning models are developed: artificial neural network, linear regression, and random forest regression. Each model has two variations: one predicts using PLN customer data and the other using satellite imagery data. The evaluation of each model is based on R2 and RMSE values. The research results show that the random forest regression model provides the most accurate prediction with an R2 of 0.9866 and an RMSE of 0.0295. Furthermore, the feature of building area significantly influences the prediction results of the random forest regression model. Overall, this research demonstrates that using satellite imagery and machine learning methods can yield accurate predictions for electricity consumption in Klojen District.
Item Type: | Thesis (Other) |
---|---|
Uncontrolled Keywords: | energi listrik, Kota Malang, machine learning, prediksi, electrical energy, Malang City, prediction. |
Subjects: | Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines. |
Divisions: | Faculty of Industrial Technology and Systems Engineering (INDSYS) > Physics Engineering > 30201-(S1) Undergraduate Thesis |
Depositing User: | Ahmad Miftahul Huda |
Date Deposited: | 10 Nov 2023 07:15 |
Last Modified: | 10 Nov 2023 07:15 |
URI: | http://repository.its.ac.id/id/eprint/102018 |
Actions (login required)
View Item |