PERANCANGAN MODEL PERAMALAN DAYA PLTS OFF-GRID DENGAN EKSTERNAL VARIABEL MENGGUNAKAN METODE LONG SHORT-TERM MEMORY - Submit Jurnal

Rahmatulah, Muhammad Rubian (2026) PERANCANGAN MODEL PERAMALAN DAYA PLTS OFF-GRID DENGAN EKSTERNAL VARIABEL MENGGUNAKAN METODE LONG SHORT-TERM MEMORY - Submit Jurnal. Bachelor thesis, Institut Teknologi Kalimantan.

[img] Text
04221044_cover.pdf
Restricted to Repository staff only until 4 October 2028.

Download (527kB) | Request a copy
[img] Text
04221044_statement_of_authenticity.pdf
Restricted to Repository staff only until 4 October 2028.

Download (90kB) | Request a copy
[img] Text
04221044_publishing_agreement.pdf
Restricted to Repository staff only until 4 October 2028.

Download (405kB) | Request a copy
[img] Text
04221044_approval_sheet.pdf
Restricted to Repository staff only until 4 October 2028.

Download (71kB) | Request a copy
[img] Text
04221044_preface.pdf
Restricted to Repository staff only until 4 October 2028.

Download (389kB) | Request a copy
[img] Text
04221044_abstract_id.pdf
Restricted to Repository staff only until 4 October 2028.

Download (391kB) | Request a copy
[img] Text
04221044_abstract_en.pdf
Restricted to Repository staff only until 4 October 2028.

Download (379kB) | Request a copy
[img] Text
04221044_table_of_content.pdf
Restricted to Repository staff only until 4 October 2028.

Download (490kB) | Request a copy
[img] Text
04221044_illustrations.pdf
Restricted to Repository staff only until 4 October 2028.

Download (412kB) | Request a copy
[img] Text
04221044_tables.pdf
Restricted to Repository staff only until 4 October 2028.

Download (416kB) | Request a copy
[img] Text
04221044_notations.pdf
Restricted to Repository staff only until 4 October 2028.

Download (409kB) | Request a copy
[img] Text
04221044_chapter_1.pdf
Restricted to Repository staff only until 4 October 2028.

Download (506kB) | Request a copy
[img] Text
04221044_chapter_2.pdf
Restricted to Repository staff only until 4 October 2028.

Download (1MB) | Request a copy
[img] Text
04221044_chapter_3.pdf
Restricted to Repository staff only until 4 October 2028.

Download (1MB) | Request a copy
[img] Text
04221044_chapter_4.pdf
Restricted to Repository staff only until 4 October 2028.

Download (5MB) | Request a copy
[img] Text
04221044_conclusions.pdf
Restricted to Repository staff only until 4 October 2028.

Download (459kB) | Request a copy
[img] Text
04221044_bibliography.pdf
Restricted to Repository staff only until 4 October 2028.

Download (376kB) | Request a copy
[img] Text
04221044_enclosure.pdf
Restricted to Repository staff only until 4 October 2028.

Download (1MB) | Request a copy
[img] Text
04221044_paper.pdf
Restricted to Repository staff only until 4 October 2028.

Download (525kB) | Request a copy
[img] Text
04221044_presentation.pdf
Restricted to Repository staff only until 4 October 2028.

Download (1MB) | Request a copy
[img] Text
04221044_Form TA. 020.pdf
Restricted to Repository staff only until 4 October 2028.

Download (145kB) | Request a copy

Abstract

Pembangkit Listrik Tenaga Surya (PLTS) tipe Off-grid memiliki tantangan besar berupa ketidakpastian daya keluaran akibat fluktuasi cuaca yang dinamis, yang berpotensi mengganggu stabilitas sistem dan manajemen beban. Penelitian ini bertujuan untuk merancang, mengimplementasikan, dan memvalidasi model peramalan daya berbasis deret waktu (time-series) menggunakan metode Long Short-Term Memory (LSTM) dengan mengintegrasikan variabel eksternal lokal. Penelitian dilakukan pada sistem PLTS Off-grid di Kelompok Tani Sumber Harapan, Kecamatan Samboja. Model dirancang menggunakan 6 fitur Input utama, yaitu daya keluaran historis, iradiasi matahari, suhu panel, kelembaban lingkungan, arus, dan tegangan, ditambah fitur lag temporal, dengan arsitektur dua hidden layer LSTM (128 dan 64 memory blocks). Untuk menguji keandalan, dilakukan validasi silang melalui tiga skenario rasio pembagian data (60:40, 70:30, dan 80:20) serta pengujian komparatif terhadap metode Transformer Forecasting dan ARIMA. Hasil penelitian menunjukkan bahwa integrasi variabel eksternal secara signifikan meningkatkan kemampuan model dalam menangkap dinamika cuaca harian, dengan iradiasi matahari sebagai fitur paling dominan. Pada data pelatihan, metode LSTM mencatatkan nilai kesalahan terendah pada metrik MSE dan MAPE dibandingkan ARIMA dan Transformer Forecasting, sedangkan pada data pengujian ARIMA menunjukkan akurasi yang sedikit lebih unggul namun dengan waktu komputasi jauh lebih lama (88 menit) dibandingkan LSTM (19 menit). Dengan mempertimbangkan keseimbangan antara akurasi dan efisiensi waktu, LSTM dipilih sebagai metode utama dalam penelitian ini. Hasil prediksi daya keluaran dan prakiraan beban puncak ini dapat dimanfaatkan secara optimal untuk mendukung efisiensi operasional dan manajemen baterai pada PLTS Off-grid.

Item Type: Thesis (Bachelor)
Subjects: T Technology > TK Electrical engineering. Electronics Nuclear engineering
Divisions: Jurusan Teknologi Industri dan Proses > Teknik Elektro
Depositing User: Muhammad Rubian Rahmatulah
Date Deposited: 20 Jul 2026 08:08
Last Modified: 20 Jul 2026 08:08
URI: http://repository.itk.ac.id/id/eprint/27901

Actions (login required)

View Item View Item