Penentuan Pembagian Beban Operasiona Turbin Gas Menggunakan Artificial Neural Network - Multiple Layer Perceptron

Hadi, Safwanul (2026) Penentuan Pembagian Beban Operasiona Turbin Gas Menggunakan Artificial Neural Network - Multiple Layer Perceptron. Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6010231024-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (2MB) | Request a copy

Abstract

Turbin gas merupakan peralatan utama dalam sistem jaringan kelistrikan. Untuk memenuhi kebutuhan beban pada jaringan, pembangkit listrik mempunyai beberapa unit turbin gas, dengan kapasitas yang sama ataupun beragam . Turbin gas suatu pembangkit di lapangan minyak di Riau memiliki 5 unit turbin gas yang dihubungkan dengan heat recovery steam generator. Semua unit beroperasi secara simultan sesuai dengan permintaan beban dari sistem kelistrikan dan permintaan uap untuk keperluan lapangan minyak dengan metode steam flood. Pengaturan beban untuk tiap-tiap unit turbin gas dilakukan dengan perhitungan numerik berdasarkan kebutuhan beban, parameter operasi, serta konsumsi bahan bakar. Pada aktualnya, nilai rekomendasi dari perhitungan numerik selalu berada diatas operasi gas turbin, mengakibatkan pemakaian bahan bakar menjadi tidak efisien. Perhitungan efisensi turbin gas menggunakan metode numerik menjadi tidak sesuai dengan kondisi unit terkini, ketika nilai tersebut tidak memperhitungkan perubahan dinamis dan berbagai multi-variabel lainnya. Penelitian Chen dan Huang menunjukkan optimasi turbin gas bisa dihitung menggunakan machine learning. Pengoperasian yang optimum bisa didapat dengan memasukkan data parameter operasi ke metode Machine Learning dengan menggunakan Artificial neural network yang di kombinasikan dengan multi layer perceptron untuk mendapat output daya yang maksimal berdasarkan historical data. Machine learning akan memprediksi kemampuan maksimum dari tiap unit. Untuk pembagian beban, ditentukan dengan proportional scalling dengan mempertimbangkan kemampuan dari tiap unit. Penelitian ini menghasilkan pemerataan dalam pembebanan sehingga bisa menghasilkan penghematan konsumsi bahan bakar sebesar 9,5%. Pemerataan beban pada unit turbin gas menghasilkan produksi uap yang lebih merata.
===================================================================================================================================
Gas turbines are the main equipment in the electrical grid system. To meet the load needs on the grid, power plants have several gas turbine units, with the same or different capacities. Gas turbine a power plant in an oil field in Riau has 5 units of gas turbines connected to heat recovery steam generators. All units operate simultaneously according to the load demand from the electrical system and the steam demand for oil field purposes by the steam flood method. The load adjustment for each gas turbine unit is carried out by numerical calculation based on load needs, operating parameters, and fuel consumption. In actuality, the recommended value of the numerical calculation is always above the turbine gas operation, resulting in inefficient fuel consumption. The calculation of gas turbine efficiency using the numerical method becomes inappropriate for the current unit conditions, when the value does not take into account dynamic changes and various other multi-variables. Chen and Huang's research shows that gas turbine optimization can be calculated using machine learning. Optimal operation can be obtained by entering operation parameter data into the Machine Learning method using an Artificial neural network combined with multi-layer perceptron to get maximum power output based on historical data. Machine learning will predict the maximum capability of each unit. For load sharing, it is determined by proportional scaling by considering the capabilities of each unit. This research results in an equal distribution in charging so that it can result in fuel consumption savings of 9.5%. Equal distribution of the load on the gas turbine unit results in more even steam production.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Turbin Gas, Jaringan Kelistrikan, Artificial neural Network, Multi-layer perceptron, fuel saving, Operasi, Machine Learning ============================================================ Gas Turbine, Electrical Grid, Artificial neural Network, Multi-layer perceptron, fuel saving, Operation, Machine Learning
Subjects:	Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines.
Divisions:	Faculty of Industrial Technology and Systems Engineering (INDSYS) > Industrial Engineering > 26101-(S2) Master Thesis
Depositing User:	Safwanul Hadi
Date Deposited:	02 Feb 2026 04:23
Last Modified:	02 Feb 2026 04:23
URI:	http://repository.its.ac.id/id/eprint/131564

Actions (login required)

View Item