Permodelan Estimasi Top-Oil Temperatur Transformator Distribusi Akibat Fluktuasi Beban Ekstrim untuk Mendukung Aplikasi Smart Grid pada Sistem Distribusi

Wijaksono, Yohanes Andri (2025) Permodelan Estimasi Top-Oil Temperatur Transformator Distribusi Akibat Fluktuasi Beban Ekstrim untuk Mendukung Aplikasi Smart Grid pada Sistem Distribusi. Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6022231090_Master_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (7MB) | Request a copy

Abstract

Semakin majunya teknologi saat ini, berbanding lurus terhadap kebutuhan daya listrik pada pelanggan. Transformator menajdi bagian penting dalam melayani kebutuhan listrik pelanggan. Beban listrik yang memerlukan beban tinggi saat ini adalah pengisian kendaraan listrik, yang memiliki perubahan beban ekstrim dan fluktuatif. Perubahan beban yang ekstrim yang terjadi secara terus menerus akan menghasilkan panas yang berlebih di dalam transformator, yang secara jangka panjang dapat mempercepat kerusakan isolasinya, yang mempengaruhi usia transformator. Pada penelitian ini akan dilakukan percobaan efek perubahan yang fluktuatif dan ekstrem dari permodelan beban resistif yang diuji di laboratorium arus tinggi PT. PLN Pusertif. Penelitian dilakukan pada kondisi lingkungan indoor dan outdoor, menggunakan 6 titik lokasi sensor thermal yang dipasang pada tangki transformator, yaitu kantong thermometer, top cover, radiator atas, radiator bawah, ambient 2 titik. Kemudian datanya akan diproses menggunakan Machine Learning metode LSTM (Long-Short Term Memory) dan hybrid PSO-LSTM (Particle Swarm Optimization - Long-Short Term Memory) dengan data training sebanyak 548 sample dan data testing sebanyak 200 sample. Evaluasi akurasi rata-rata menggunakan LSTM, MSE = 2.56 (indoor) dan MSE = 1.35 (outdoor). Evaluasi akurasi rata-rata menggunakan PSO-LSTM, MSE = 1.58 (indoor) dan MSE = 0.77 (outdoor). Perhitungan HST tertinggi diperoleh 70.220C, sehingga diperoleh perhitungan FAA terbesar 30.67 jam/tahun dan LoL sebesar 0.35%. Dengan mengetahui kondisi temperatur pada transformator, diharapkan dapat melakukan langkah preventif untuk menjaga kehandalan transformator, sehingga dapat menunjang pengaplikasian smart grid pada sistem distribusi.
================================================================================================================================
The development of modern technology, customer need electricity more. Transformers are a crucial component in serving customer electricity needs. The charging of electric vehicle, which the character is extreme and fluctuating load changes, is a major electrical load requirement. In the future, extreme load changes generate excessive heat within the transformer, which in the long term can accelerate insulation deterioration and impact its lifespan. This study will research the effects of fluctuating and extreme changes in resistive load modeling, tested in the high-current laboratory of PT. PLN Pusertif. The study was conducted under indoor and outdoor conditions, using six thermal sensor locations installed on the transformer tank: the thermometer pocket, top cover, top radiator, bottom radiator, and two ambient locations. The data will then be processed using machine learning methods using LSTM (Long-Short Term Memory) and a hybrid PSO-LSTM (Particle Swarm Optimization - Long-Short Term Memory) with 548 training samples and 200 testing samples. Average accuracy evaluation using LSTM, MSE = 2.56 (indoor) and MSE = 1.35 (outdoor). Average accuracy evaluation using PSO-LSTM, MSE = 1.58 (indoor) and MSE = 0.77 (outdoor). The highest HST calculation was obtained at 70.220C, resulting in the largest FAA calculation of 30.67 hours/year and LoL of 0.35%. By knowing the temperature conditions of the transformer, it is hoped that preventive measures can be taken to maintain the reliability of the transformer, so that it can support the application of smart grids in the distribution system.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Transformator, Isolasi, Fluktuatif, Ekstrim, Temperatur, Machine Learning, Top-oil, Smart Grid Transformer, Isolation, Fluctuation, Extreme, Temperature, Machine Learning, Top-oil, Smart Grid
Subjects:	T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK2551 Electric transformers.
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Electrical Engineering > 20101-(S2) Master Thesis
Depositing User:	Yohanes Andri Wijaksono
Date Deposited:	25 Jul 2025 07:17
Last Modified:	25 Jul 2025 07:18
URI:	http://repository.its.ac.id/id/eprint/121550

Actions (login required)

View Item