Rancang Bangun Web Application Prediksi Cuaca Buruk Untuk Operasional Penerbangan Di Bandara Juanda Berdasarkan Unsur Cuaca Menggunakan Long Short-Term Memory

Azizah, Salsabila Aulia (2026) Rancang Bangun Web Application Prediksi Cuaca Buruk Untuk Operasional Penerbangan Di Bandara Juanda Berdasarkan Unsur Cuaca Menggunakan Long Short-Term Memory. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 2043221067-Undergraduate_Thesis.pdf Restricted to Repository staff only Download (9MB) | Request a copy

Abstract

Pasca pandemi COVID-19 di Indonesia, industri penerbangan terus mengalami peningkatan permintaan transportasi udara, termasuk Bandara Udara Juanda Surabaya. Peningkatan yang sangat tinggi tersebut harus diimbangi dengan peningkatan kualitas layanan, khususnya ketepatan waktu penerbangan. Namun, keadaan nyatanya masih sering terjadi delay penerbangan. Salah satu faktor penyebab delay adalah faktor cuaca yang ditandai dengan adanya aerodrome dan wind shear warning. Oleh karena itu, penelitian ini bertujuan untuk membangun rancangan web application untuk memprediksikan cuaca buruk untuk operasional penerbangan di Bandara Udara Juanda berdasarkan unsur cuaca. Unsur cuaca yang digunakan meliputi arah dan kecepatan angin, suhu, tekanan udara, visibility, titik embun, dan kondisi cuaca. Periode data peringatan cuaca buruk dan data unsur cuaca yang digunakan pada rentang Januari 2021 hingga September 2025. Pembangunan model prediksi menggunakan algoritma Long Short-Term Memory (LSTM) yang mampu mengenali pola data sekuensial dan ketergantungan jangka panjang. Pada pengembangan model tersebut juga dilakukan optimasi dengan menerapkan Bayesian Optimization dan optimizer Adam. Pengembangan model juga dilakukan dalam 3 skenario pembagian data, yaitu 70:30, 80:20, dan 90:10. Hasil penelitian menunjukkan bahwa durasi cuaca buruk cenderung muncul pada musim penghujan. Hal ini dipengaruhi oleh hubungan nonlinear antara kejadian cuaca buruk dengan interaksi unsur cuaca secara simultan. Berdasarkan hasil evaluasi model, Skenario C (90:10) dipilih sebagai model terbaik karena menunjukkan performa umum yang lebih baik serta memiliki kemampuan deteksi kejadian cuaca buruk yang lebih tinggi dibandingkan skenario lainnya. Model terbaik kemudian diimplementasikan ke dalam web application yang mampu menghasilkan prediksi cuaca buruk untuk operasional penerbangan berdasarkan data cuaca.
=================================================================================================================================
In the wake of the COVID-19 pandemic in Indonesia, the aviation industry has continued to experience an increase in demand for air travel, including at Surabaya’s Juanda International Airport. This significant increase must be matched by improvements in service quality, particularly flight punctuality. However, in reality, flight delays still occur frequently. One of the factors causing delays is weather, characterized by aerodrome and wind shear warnings. Therefore, this study aims to develop a web application to predict adverse weather conditions affecting flight operations at Juanda International Airport based on weather parameters. The weather parameters used include wind direction and speed, temperature, air pressure, visibility, dew point, and weather conditions. The data period for severe weather warnings and weather element data used spans from January 2021 to September 2025. The prediction model was developed using the Long Short-Term Memory (LSTM) algorithm, which is capable of recognizing patterns in sequential data and long-term dependencies. The model was further optimized by applying Bayesian Optimization and the Adam optimizer. Model development was also conducted using three data split scenarios: 70:30, 80:20, and 90:10. The research results show that periods of inclement weather tend to occur during the rainy season. This is influenced by the nonlinear relationship between inclement weather events and the simultaneous interaction of weather elements. Based on the results of the model evaluation, Scenario C (90:10) was selected as the best model because it demonstrated better overall performance and had a higher ability to detect severe weather events compared to the other scenarios. The best model was then implemented into a web application capable of generating adverse weather forecasts for flight operations based on weather data.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Bandara Juanda, Cuaca Buruk, Long Short-Term Memory (LSTM), Operasional Penerbangan, Prediksi, Aviation Operations, Juanda Airport, Long Short-Term Memory (LSTM), Predicting, Servere Weather
Subjects:	Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science)
Divisions:	Faculty of Vocational > 49501-Business Statistics
Depositing User:	Salsabila Aulia Azizah
Date Deposited:	29 Jul 2026 02:32
Last Modified:	29 Jul 2026 02:32
URI:	http://repository.its.ac.id/id/eprint/139002

Actions (login required)

View Item