Desain Dan Simulasi Mikro Innverter Berbasis Recurrent Neural Network (RNN) Pada Mikrogrid

Shalsabila, Atsila (2025) Desain Dan Simulasi Mikro Innverter Berbasis Recurrent Neural Network (RNN) Pada Mikrogrid. Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6022231073-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (6MB) | Request a copy

Abstract

Meningkatnya populasi global dan tantangan ekonomi yang signifikan telahmenyebabkan peningkatan permintaan energi. Sumber energi utama, bahan bakar fosil, berkontribusi terhadap dampak negatif terhadap lingkungan seperti emisi gas rumah kaca dan perubahan iklim. Transisi ke sumber energi terbarukan sangat penting untuk mengurangi dampak tersebut. Energi surya, sebagai sumber energi terbarukan, telah muncul sebagai solusi untuk memenuhi kebutuhan energi global. Panel fotovoltaik (PV) memiliki potensi untuk menghasilkan energi bersih, tetapi efisiensinya sering kali dipengaruhi oleh kondisi lingkungan seperti fluktuasi suhu dan curah hujan. Pelacakan Titik Daya Maksimum (MPPT) digunakan untuk memastikan bahwa panel surya dapat beroperasi pada tingkat daya maksimum dalam berbagai kondisi lingkungan. Jaringan mikro, sebagai sistem energi terbarukan terdistribusi, dianggap sebagai solusi yang efektif untuk meningkatkan efisiensi energi. Inverter mikro memainkan peran penting dalam menjaga stabilitas dan mengurangi kegagalan jaringan. Teknologi AI dan Machine Learning, seperti Recurrent Neural Network (RNN), memainkan peran penting dalam mengoptimalkan sistem jaringan mikro berbasis energi terbarukan.Mengintegrasikan RNN ke dalam algoritma MPPT dapat meningkatkan efisiensi dan responsif terhadap perubahan lingkungan. Penelitian ini akan berfokus pada perancangan dan simulasi sistem mikro inverter yang menggunakan RNN untuk mengoptimalkan MPPT dan meningkatkan performa sistem. Penelitian inimenggunakan algoritma MPPT berbasis RNN untuk mengoptimalkan output panel fotovoltaik, yang meliputi input tegangan dan arus. Output RNN adalah duty cycle untuk mengontrol MOSFET dalam konverter Boost, memungkinkan kontrol panel surya yang lebih efisien dan responsif. Algoritma MPPT memiliki daya pelacakan relatif ±0,14 detik, yang lebih cepat daripada algoritma P&O dengan Waktu pelacakan ±0,38 detik.
============================================================================================================================================
The increasing global population and significant economic challenges have led to an increase in energy demand. The main source of energy, fossil fuels, contributes to negative environmental impacts such as greenhouse gas emissions and climate change. Transitioning to renewable energy sources is essential to mitigate such impacts. Solar energy, as a renewable energy source, has emerged as a solution to meet global energy needs. Photovoltaic (PV) panels have the potential to produce clean energy, but their efficiency is often affected by environmental conditions such as temperature fluctuations and rainfall. Maximum Power Point Tracking (MPPT) is used to ensure that solar panels can operate at maximum power levels under various environmental conditions. Micro-grids, as a distributed renewable energy system, are considered an effective solution to improve energy efficiency. Microinverters play an important role in maintaining stability and reducing grid failures. AI and Machine Learning technologies, such as Recurrent Neural Network (RNN), play an important role in optimizing renewable energy-based micro-grid systems. Integrating RNN into MPPT algorithms can improve efficiency and responsiveness to environmental changes. This research will focus on designing and simulating a micro inverter system that uses RNN to optimize MPPT and improve system performance. This research uses an RNN-based MPPT algorithm to optimize the output of photovoltaic panels, which includes voltage and current inputs. The RNN output is the duty cycle to control the MOSFETs in the Bppst konverter, enabling more efficient and responsive control of the solar panel. The MPPT algorithm has a relative tracking power of ±0.14 seconds, which is faster than the P&O algorithm with a tracking time of ±0.38 seconds.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Mikro Inverter, Mikro Grid, Recurrent Neural Network, Energi Terbarukan, Micro Inverter, Micro Grid, Recurrent Neural Network, Renewable Energy
Subjects:	T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK1007 Electric power systems control T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK1087 Photovoltaic power generation T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK2692 Inverters T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK7872 Electric current converters, Electric inverters.
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Electrical Engineering > 20101-(S2) Master Thesis
Depositing User:	Atsila Shalsabila
Date Deposited:	23 Jul 2025 04:12
Last Modified:	23 Jul 2025 04:12
URI:	http://repository.its.ac.id/id/eprint/120755

Actions (login required)

View Item