Putra, I Gusti Agung Krisna Manggala (2020) Pemodelan Gas Turbine Generator Berbasis Jaringan Saraf Tiruan. Other thesis, Institut Teknologi Sepuluh Nopember.
Preview |
Text
02311540000107-Undergraduate_Thesis.pdf Download (2MB) | Preview |
Abstract
Gas turbine generator adalah suatu turbin yang dapat merubah energi kalor menjadi energi mekanis. GTG memiliki 3 bagian penting yakni kompresor, ruang bakar, dan turbin. Ketiga bagian ini memiliki variabel input dan output dimana dinamika proses ini sulit dimodelkan dengan menggunakan hukum fisika. Maka dari itu pasangan data yang diambil dari track record GTG ini dimodelkan dengan menggunakan jaringan saraf tiruan (JST). Masing-masing bagian GTG memiliki struktur JST tersendiri dimana JST dalam kasus ini dibangin dengan membagi data input dan output ke dalam data training dan validasi. Training dilakukan pada JST untuk mengenali pola input dan output, kemudian validasi dilakukan untuk membuktikan bahwa JST dapat memodelkan dinamika yang terjadi pada output setiap bagian GTG jika terjadi perubahan input. kinerja dari JST ini dievaluasi berdasarkan nilai RMSE. Pada penelitian ini variabel untuk kompresor yaitu inlet temperature dan pressure, inlet guide vane, dan PCD, sementara pada ruang bakar yaitu fuel flow rate, fuel pressure, cc outlet temperature, kemudian pada turbin yakni temperature inlet turbine, lube pressure, bearing temperature, dan power output (beban). Hasil simulasi menunjukkan bahwa untuk kompresor berada pada struktur JST 3-14-1 (input-hidden node-output) dengan nilai RMSE 0,01854, pada ruang bakar didapatkan struktur JST 3-3-1 dengan nilai RMSE 0,03126, dan pada bagian turbin dengan struktur 3-2-1 dengan nilai RMSE 0,0583. Kata kunci: gas turbine generator, jaringan saraf tiruan (JST), root mean square error
================================================================================================================================
Gas turbine generator is a turbine that can convert heat energy into mechanical energy. GTG has 3 important parts namely compressor, combustion chamber, and turbine. These three parts have input and output variables where the dynamics of this process are difficult to model using the laws of physics.Therefore the pair of data taken from the GTG track record is modeled using an artificial neural network (ANN). Each of these sections has its own ANN structure, where ANN, in this case, analyzes the pattern of input and output pairs. This structure is built by dividing input and output data into training and validation data. Training is conducted at ANN to recognize input and output patterns. Validation is carried out to prove that ANN can model the dynamics that occur in the output of each GTG section if input changes occur. This ANN performance is evaluated based on the RMSE value. In this study, the variables found for the compressor are inlet temperature and pressure, inlet guide vane, PCD, while in the combustion chamber are fuel flow rate, fuel pressure, cc outlet temperature, then in the turbine namely turbine inlet temperature, lube pressure, bearing temperature, and power output. The simulation results show that the compressor is in the ANN structure 3-14-1 (input-hidden node-output) with an RMSE value of 0.01854, the combustion chamber obtained ANN structure 3-3-1 with an RMSE value of 0.03126 and in the turbine section with a structure 3-2-1 with an RMSE value of 0.0583. keywords: gas turbine generator, artificial neural networks, root mean square error
| Item Type: | Thesis (Other) |
|---|---|
| Additional Information: | RSF 621.433 Put p-1 2020 |
| Uncontrolled Keywords: | gas turbine generator, jaringan saraf tiruan (JST), root mean square error |
| Subjects: | Q Science > Q Science (General) > Q337.5 Pattern recognition systems Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science) T Technology > TJ Mechanical engineering and machinery > TJ213 Automatic control. |
| Divisions: | Faculty of Industrial Technology > Physics Engineering > 30201-(S1) Undergraduate Thesis |
| Depositing User: | Mr. Krisna Manggala |
| Date Deposited: | 06 Mar 2025 07:28 |
| Last Modified: | 06 Mar 2025 07:28 |
| URI: | http://repository.its.ac.id/id/eprint/74550 |
Actions (login required)
 |
View Item |