Prediksi Tinggi Gelombang Laut Menggunakan Model Hybrid VAR-LSTM

Pradana, Muhammad Anugrah Agustian (2026) Prediksi Tinggi Gelombang Laut Menggunakan Model Hybrid VAR-LSTM. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5020221089-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (5MB) | Request a copy

Abstract

Selat Makassar sebagai bagian dari Alur Laut Kepulauan Indonesia II (ALKI II) merupakan jalur pelayaran internasional yang sangat strategis namun memiliki kondisi gelombang yang dinamis akibat pengaruh Arlindo dan pola angin monsun. Gelombang tinggi di perairan ini telah menyebabkan beberapa kecelakaan maritim dalam tahun-tahun terakhir, sehingga prediksi parameter gelombang yang akurat menjadi kebutuhan penting untuk mendukung keselamatan pelayaran. Penelitian ini mengembangkan model hybrid Vector Autoregressive Long Short-Term Memory (VAR-LSTM) untuk memprediksi tinggi gelombang signifikan (Hs), tinggi gelombang maksimum (Hmax), rata-rata periode lintas nol (Tz), dan periode puncak gelombang (Tp) di perairan Selat Makassar secara simultan. Pendekatan hybrid ini menggabungkan kemampuan VAR dalam menangkap hubungan linear dan dinamika timbal balik antar keempat parameter gelombang dengan kemampuan LSTM dalam mempelajari pola non-linear dan dependensi jangka panjang dari data historis gelombang. Data yang digunakan merupakan data reanalysis ERA5 dari European Centre for Medium-Range Weather Forecasts (ECMWF). Model dievaluasi menggunakan metrik Root Mean Square Error (RMSE) dan Mean Absolute Percentage Error (MAPE) untuk masing-masing parameter gelombang secara individual. Hasil pengujian out-sample menunjukkan bahwa model hybrid VAR-LSTM mampu memprediksi keempat parameter gelombang secara simultan dengan nilai MAPE berturut-turut sebesar 11,35% (Hs), 11,57% (Hmax), 4,94% (Tz), dan 12,46% (Tp), meskipun menggunakan data yang terbatas. Model ini diharapkan dapat memberikan informasi prediksi parameter gelombang yang lebih lengkap dan komprehensif untuk mendukung keselamatan pelayaran di jalur ALKI II Selat Makassar.
=====================================================================================================================================
The Makassar Strait, as part of the Indonesian Archipelago Sea Lane II (ALKI II), is a highly strategic international shipping lane but has dynamic wave conditions due to the influence of Arlindo and monsoon wind patterns. High waves in these waters have caused several maritime accidents in recent years, making accurate wave parameter prediction essential for supporting shipping safety. This study develops a hybrid Vector Autoregressive Long Short-Term Memory (VAR-LSTM) model to simultaneously predict significant wave height (Hs), maximum wave height (Hmax), average zero-crossing period (Tz), and peak wave period (Tp) in the Makassar Strait. This hybrid approach combines VAR's ability to capture linear relationships and reciprocal dynamics among the four wave parameters with LSTM's ability to learn non-linear patterns and long-term dependencies from historical wave data. The data used is ERA5 reanalysis data from the European Centre for Medium-Range Weather Forecasts (ECMWF). The model was evaluated using Root Mean Square Error (RMSE) and Mean Absolute Percentage Error (MAPE) metrics for each wave parameter individually. Out-sample testing results show that the hybrid VAR-LSTM model is able to simultaneously predict all four wave parameters with MAPE values of 11.35% (Hs), 11.57% (Hmax), 4.94% (Tz), and 12.46% (Tp), despite the use of limited data. This model is expected to provide more complete and comprehensive wave parameter prediction information to support shipping safety along the ALKI II Makassar Strait shipping lane.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	VAR-LSTM, Machine Learning, Artificial Intelligence, Prediksi Gelombang,Tinggi Gelombang Signifikan, Selat Makassar, ALKI II, VAR-LSTM, Machine Learning, Artificial Intelligence, Wave Prediction,Significant Wave Height, Makassar Strait, ALKI II
Subjects:	Q Science > QA Mathematics > QA336 Artificial Intelligence T Technology > TC Hydraulic engineering. Ocean engineering T Technology > TC Hydraulic engineering. Ocean engineering > TC147 Ocean wave power. T Technology > TC Hydraulic engineering. Ocean engineering > TC424 Water levels
Divisions:	Faculty of Marine Technology (MARTECH) > Ocean Engineering > 38201-(S1) Undergraduate Thesis
Depositing User:	Muhammad Anugrah Agustian Pradana
Date Deposited:	21 Jul 2026 01:34
Last Modified:	21 Jul 2026 01:34
URI:	http://repository.its.ac.id/id/eprint/135843

Actions (login required)

View Item