Studi Sistem Monitoring Prediksi Keandalan Real-Time Pada Sistem Pengendalian Kecepatan Generator Turbin Angin Dengan Kesalahan Sensor

Ayurani, Lilik (2019) Studi Sistem Monitoring Prediksi Keandalan Real-Time Pada Sistem Pengendalian Kecepatan Generator Turbin Angin Dengan Kesalahan Sensor. Masters thesis, Institut Teknologi Sepuluh Nopember.

[img]
Preview
Text
02311650012002-Master_Thesis.pdf

Download (2MB) | Preview

Abstract

Peningkatan jumlah penduduk seiring dengan peningkatan kebutuhan energi listrik yang ada di Indonesia. Salah satu cara yang sedang digalakkan untuk pemenuhan kebutuhan listrik melalui energi alternatif adalah dengan menggunakan turbin angin. Sementara sistem turbin angin sedang digunakan, keandalannya akan menurun secara bertahap. Penelitian ini bertujuan untuk menentukan parameter desain yang mempengaruhi prediksi keandalan real-time pada sistem pengendalian kecepatan generator turbin angin. Tiga langkah yang perlu dilakukan yaitu perancangan observer, perancangan algoritma prediksi kesalahan sensor dan perancangan algoritma prediksi keandalan. Perancangan observer digunakan untuk mengestimasi kesalahan sensor dari variabel yang terukur, kemudian hasil estimasi digunakan untuk menghitung prediksi kesalahan sensor melalui algoritma exponential smoothing. Hasil dari prediksi kesalahan sensor ini secara langsung digunakan untuk prediksi keandalan real-time. Waktu kegagalan riil ketika kecepatan generator lebih besar dari 1.2 pu dan kecepatan generator dibawah 0.5 pu menandakan keandalan mulai turun. Variasi diberikan terhadap time interval sebesar 1 detik, 5 detik, 10 detik dan 50 detik, sedangkan setiap time interval diberikan variasi jumlah prediksi sebesar 1, 5 dan 10. Hasil simulasi menunjukkan bahwa prediksi keandalan dengan time interval sebesar 1 detik lebih tepat jika dibandingkan variasi time interval sebesar 5 detik, 10 detik, dan 50 detik. Jumlah prediksi mempengaruhi ketelitian prediksi keandalan real-time yang dihasilkan, semakin besar jumlah prediksi maka semakin teliti. Dengan time interval 1 detik, jumlah prediksi sebesar 10 menghasilkan prediksi keandalan real-time lebih tepat dibandingkan jumlah prediksi sebesar 1 dan 5. =============================================================================================== Indonesia is known as the largest archipelagic country in the world. It has a significant growth in its population every year. As the result, the energy demand in Indonesia continues to develop. One of the method of helping meet energy needs is by introducing wind turbine as alternative energy source. However, the reliability on the turbine system will be decreasing steadly if it is being continually exploited. This study aims to determine the design parameters that is affecting the real-time reliability prediction for wind turbine generator speed control systems. There are three steps that need to be done, namely observer design, sensor fault prediction algorithm design, and reliability prediction algorithm design. Observer design was used to estimate sensor fault from the measured variable and the estimated result was used to calculate sensor fault prediction using exponential smoothing algorithm. The calculated result will then be directly exploited to predict real-time reliability. We obtained that the failure time correspond with a decrease in the reliability when the generator speed is greater than 1.2 pu and below 0.5 pu. Time interval was varied by 1s, 5s, 10s, and 50s. In each of the time interval, the variation number of prediction given was of 1, 5, and 10. Simulation result has shown that reliability prediction with time interval 1s was more precise compared to time interval variation of 5s, 10s, and 50s. The number of prediction affects real-time reliability prediction fidelity resulted. With the time interval of 1s, the number of prediction of 10 results in accurate real-time reliability prediction compared to number of prediction of 1 and 5.

Item Type: Thesis (Masters)
Additional Information: RTF 620.004 52 Ayu s-1 2019
Uncontrolled Keywords: Wind turbine generator, observer, sensor fault prediction, real-time reliability prediction.
Subjects: T Technology > TJ Mechanical engineering and machinery > TJ828 Wind turbines
T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK7870.23 Reliability. Failures
Divisions: Faculty of Industrial Technology > Physics Engineering > 30101-(S2) Master Thesis
Depositing User: Lilik Ayurani
Date Deposited: 07 Jun 2021 07:05
Last Modified: 07 Jun 2021 07:05
URI: https://repository.its.ac.id/id/eprint/60378

Actions (login required)

View Item View Item