PERANCANGAN MODEL PERAMALAN DAYA PLTS ON-GRID MENGGUNAKAN METODE LONG SHORT TERM MEMORY (Studi Kasus: PT Idec Abadi Wood Industries)

Qarini, Ayuk (2026) PERANCANGAN MODEL PERAMALAN DAYA PLTS ON-GRID MENGGUNAKAN METODE LONG SHORT TERM MEMORY (Studi Kasus: PT Idec Abadi Wood Industries). Bachelor thesis, Institut Teknologi Kalimantan.

	Text 04221069_cover.pdf Restricted to Repository staff only until 4 October 2028. Download (486kB) | Request a copy
	Text 04221069_statement_of_authenticity.pdf Restricted to Repository staff only until 4 October 2028. Download (229kB) | Request a copy
	Text 04221069_publishing_agreement.pdf Restricted to Repository staff only until 4 October 2028. Download (529kB) | Request a copy
	Text 04221069_approval_sheet.pdf Restricted to Repository staff only until 4 October 2028. Download (229kB) | Request a copy
	Text 04221069_preface.pdf Restricted to Repository staff only until 4 October 2028. Download (743kB) | Request a copy
	Text 04221069_abstract_id.pdf Restricted to Repository staff only until 4 October 2028. Download (515kB) | Request a copy
	Text 04221069_abstract_en.pdf Restricted to Repository staff only until 4 October 2028. Download (505kB) | Request a copy
	Text 04221069_table_of_content.pdf Restricted to Repository staff only until 4 October 2028. Download (689kB) | Request a copy
	Text 04221069_illustrations.pdf Restricted to Repository staff only until 4 October 2028. Download (955kB) | Request a copy
	Text 04221069_tables.pdf Restricted to Repository staff only until 4 October 2028. Download (488kB) | Request a copy
	Text 04221069_notations.pdf Restricted to Repository staff only until 4 October 2028. Download (248kB) | Request a copy
	Text 04221069_chapter_1.pdf Restricted to Repository staff only until 4 October 2028. Download (2MB) | Request a copy
	Text 04221069_chapter_2.pdf Restricted to Repository staff only until 4 October 2028. Download (6MB) | Request a copy
	Text 04221069_chapter_3.pdf Restricted to Repository staff only until 4 October 2028. Download (5MB) | Request a copy
	Text 04221069_conclusions.pdf Restricted to Repository staff only until 4 October 2028. Download (913kB) | Request a copy
	Text 04221069_bibliography.pdf Restricted to Repository staff only until 4 October 2028. Download (963kB) | Request a copy
	Text 04221069_paper.pdf Restricted to Repository staff only until 4 October 2028. Download (601kB) | Request a copy
	Text 04221069_presentation.pdf Restricted to Repository staff only until 4 October 2028. Download (2MB) | Request a copy
	Text 04221069_Form. TA-020.pdf Restricted to Repository staff only until 4 October 2028. Download (82kB) | Request a copy
	Text 04221069_enclosure.pdf Restricted to Repository staff only until 4 October 2028. Download (9MB) | Request a copy
	Text 04221069_chapter_4.pdf Restricted to Repository staff only until 4 October 2028. Download (4MB) | Request a copy

Abstract

Meningkatnya kebutuhan energi listrik mendorong pemanfaatan energi terbarukan melalui sistem Pembangkit Listrik Tenaga Surya (PLTS) on-grid. Namun, daya keluaran PLTS bersifat fluktuatif karena dipengaruhi oleh iradiasi matahari, suhu, dan kondisi cuaca, sehingga diperlukan metode peramalan yang akurat. Penelitian ini bertujuan mengembangkan model peramalan daya keluaran PLTS on-grid menggunakan metode Long Short-Term Memory (LSTM), menganalisis pengaruh optimasi hyperparameter, serta membandingkan kinerjanya dengan metode Gated Recurrent Unit (GRU). Data yang digunakan merupakan data historis PLTS berkapasitas 974,32 kWp di PT Idec Abadi Wood Industries selama periode Juli–Desember 2025 dengan interval satu jam sebanyak 1.840 data. Variabel masukan meliputi waktu, iradiasi matahari, suhu udara, dan data historis daya. Tahapan penelitian meliputi pre-processing data, normalisasi Min-Max Scaling, pembentukan fitur sliding window, optimasi hyperparameter melalui Random Search, serta evaluasi model menggunakan metrik MSE, RMSE, MAE, dan MAPE. Hasil penelitian menunjukkan bahwa model LSTM terbaik diperoleh dengan konfigurasi arsitektur 256–128–64 neuron, learning rate 0,001, dropout rate 0,2, dan batch size 16, menghasilkan MAPE data latih 28,84%, MAPE data uji 30,52%, dan MSE 8.427,43 kW², lebih baik dibandingkan GRU dengan MAPE data uji 30,57% dan MSE 8.847,47 kW². Pengujian berdasarkan kondisi cuaca menunjukkan akurasi model meningkat pada kondisi data cerah dengan MAPE 15,85%, sedangkan kondisi data mendung menghasilkan MAPE 36,69%. Model LSTM direkomendasikan untuk peramalan daya keluaran PLTS on-grid jangka pendek dan bekerja lebih efektif pada kondisi cuaca cerah, sedangkan pada kondisi mendung diperlukan penambahan variabel seperti kelembapan udara untuk meningkatkan akurasi prediksi dan mendukung pengelolaan energi yang lebih efektif dan adaptif.

Item Type:	Thesis (Bachelor)
Subjects:	A General Works > AI Indexes (General)
Divisions:	Jurusan Teknologi Industri dan Proses > Teknik Elektro
Depositing User:	Ayuk Qarini
Date Deposited:	16 Jul 2026 03:22
Last Modified:	16 Jul 2026 03:22
URI:	http://repository.itk.ac.id/id/eprint/26525

Actions (login required)

View Item