Prediksi Gelombang Laut Berbasis Metode ANN-LSTM Studi Lokasi : Selat Lombok Dan Selat Makassar

Aqilah Dewi, Tsabitah Sahda (2026) Prediksi Gelombang Laut Berbasis Metode ANN-LSTM Studi Lokasi : Selat Lombok Dan Selat Makassar. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5020221066-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (3MB) | Request a copy

Abstract

Selat Makassar dan Selat Lombok merupakan jalur pelayaran strategis Alur Laut Kepulauan Indonesia II (ALKI II) yang memiliki karakteristik oseanografi bertolak belakang. Selat Makassar bersifat semi-tertutup dengan dominasi gelombang angin lokal (wind-sea) berperiode pendek, sedangkan Selat Lombok merupakan perairan terbuka yang dipengaruhi rambatan gelombang alun (swell) berperiode panjang dari Samudra Hindia. Perbedaan dinamika gelombang ini menuntut adanya prediksi cuaca laut yang akurat demi keselamatan pelayaran. Penelitian ini bertujuan untuk memprediksi tinggi dan periode gelombang laut menggunakan algoritma Artificial Neural Network-Long Short-Term Memory (ANN-LSTM). Data yang digunakan merupakan data time-series oseanografi per jam dari ERA5 ECMWF periode 2020–2025, mencakup parameter Significant Wave Height (Hs), Maximum Wave Height (Hmax), Mean Wave Period (Tz), dan Peak Wave Period (Tp). Dataset dibagi dengan rasio 80% data training, 10% validation, dan 10% testing. Hasil penelitian menunjukkan bahwa arsitektur ANN-LSTM memiliki akurasi peramalan yang sangat tinggi dan tangguh (robust). Evaluasi Correlation Coefficient (CC) aktual dan prediksi pada parameter Hs menunjukkan korelasi nyaris sempurna sebesar 0,998 (Lombok) dan 0,996 (Makassar). Tingkat kesalahan model juga sangat minim, dibuktikan dengan nilai Mean Absolute Percentage Error (MAPE) untuk Hs sebesar 0,67% di Selat Lombok dan 1,94% di Selat Makassar, jauh di bawah toleransi galat 10%. Model ini terbukti peka dalam merespons anomali lonjakan ombak ekstrem seketika maupun pergerakan alun yang stabil. Secara keseluruhan, algoritma ANN-LSTM sangat layak diimplementasikan sebagai algoritma inti pada Early Warning System (EWS) jangka pendek untuk mendukung mitigasi keselamatan dan penentuan navigasi maritim taktis di jalur ALKI II.
================================================================================================================================
The Makassar Strait and Lombok Strait are strategic shipping routes of the Indonesian Archipelagic Sea Lane II (ALKI II) exhibiting fundamentally distinct oceanographic characteristics. The Makassar Strait is a semi-enclosed waterbody dominated by short-period local wind seas, whereas the Lombok Strait is an open ocean region influenced by long-period swells propagating from the Indian Ocean. These contrasting wave dynamics necessitate accurate marine weather forecasting for navigational safety. This study aims to predict wave height and period using the Artificial Neural Network-Long Short-Term Memory (ANN-LSTM) algorithm. The model utilized hourly oceanographic time-series data from ERA5 ECMWF spanning 2020–2025, comprising Significant Wave Height (Hs), Maximum Wave Height (Hmax), Mean Wave Period (Tz), and Peak Wave Period (Tp). The dataset was systematically split into 80% training, 10% validation, and 10% testing. The results demonstrated that the ANN-LSTM architecture possesses exceptionally high forecasting accuracy and robustness. The Correlation Coefficient (CC) evaluation between actual and predicted Hs revealed near-perfect correlations of 0.998 (Lombok) and 0.996 (Makassar). The model's error rate was remarkably low, evidenced by Mean Absolute Percentage Error (MAPE) values for Hs of 0.67% in the Lombok Strait and 1.94% in the Makassar Strait, well below the 10% error tolerance. The model proved highly sensitive in instantaneously detecting both sudden extreme wave spikes and stable swell movements. Overall, the ANN-LSTM algorithm is proven to be highly viable for implementation as a core algorithm in a short-term marine Early Warning System (EWS) to support safety mitigation and tactical maritime navigation decisions along the ALKI II route.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Gelombang Signifikan, ANN-LSTM, Selat Makassar, Selat Lombok, Prediksi Gelombang. Significant Wave Height, ANN-LSTM, Makassar Strait, Lombok Strait, Wave Prediction
Subjects:	Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines. T Technology > T Technology (General) > T174 Technological forecasting T Technology > T Technology (General) > T58.5 Information technology. IT--Auditing
Divisions:	Faculty of Marine Technology (MARTECH) > Ocean Engineering > 38201-(S1) Undergraduate Thesis
Depositing User:	Tsabitah Sahda Aqilahdewi
Date Deposited:	24 Jul 2026 06:31
Last Modified:	24 Jul 2026 06:31
URI:	http://repository.its.ac.id/id/eprint/137187

Actions (login required)

View Item