Pratama, Rachmat Ariestyo Putra (2016) Prediksi Nilai Level Dan Pressure Pada Steam Drum Boiler Dengan Pendekatan Neural Network Di PLTU Paiton Unit 5 Dan 6. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.
Preview |
Text
2412100005-Undergraduate_Thesis.pdf - Accepted Version Download (2MB) | Preview |
Abstract
Steam Drum Boiler merupakan bagian penting dari suatu Power Plant dimana pada Steam Drum ini terdapat dua fase yang berbeda dalam satu tempat yaitu air dan uap. Steam Drum juga memiliki karakter yang kompleks untuk dimodelkan. Level dan Pressure merupakan dua variabel utama yang harus dipantau kondisinya pada sebuah Steam Drum. Oleh karena itu, pada Tugas Akhir ini dilakukan prediksi nilai level dan pressure dalam pemodelannya menggunakan Jaringan Syaraf Tiruan dengan metode training Levenberg Marquadt. Kelebihan dari sistem Jaringan Syaraf Tiruan adalah mampu menghitung secara pararel dengan cara belajar dari pola-pola yang diajarkan. Rancangan sistem Jaringan Syaraf Tiruan mempunyai struktur Multi Layer Perceptron. Arsitektur untuk prediksi level menghasilkan RMSE pada saat training sebesar 0,196 , dan error rata-rata saat testing sebesar 0,623%. Sedangkan untuk prediksi pressure menghasilkan RMSE pada saat training sebesar 0,05 , dan error rata-rata saat testing sebesar 0,973%
========================================================================================================
Steam Boiler Drum is an important part of a Power Plant where the steam drum, there are two different phases in one place, namely water and steam. Steam Drum also has a complex character to be modeled. Level and Pressure are the two main variables that should be monitored condition in a Steam Drum. Therefore, in this final project is done predictive value and the level of pressure in the modeling using artificial neural network with Levenberg Marquadt training methods. Advantages of Neural Network system is capable of calculating in parallel by learning from the patterns that are taught. The design of the system has a structure Neural Networks Multi Layer Perceptron. Architecture to generate the prediction RMSE level during training at 0.196, and the average error when testing amounted to 0.623%. As for the prediction of pressure generating RMSE during training at 0.05, and the average error when testing amounted to 0.973%.
| Item Type: | Thesis (Undergraduate) |
|---|---|
| Additional Information: | RSF 629.836 Pra p |
| Uncontrolled Keywords: | Steam Drum Boiler , Jaringan Syaraf Tiruan , Levenberg-Marquardt , Mean Squared Error |
| Subjects: | T Technology > TD Environmental technology. Sanitary engineering > TD899.S68 Steam power plants |
| Divisions: | Faculty of Industrial Technology > Physics Engineering > 30201-(S1) Undergraduate Thesis |
| Depositing User: | Yeni Anita Gonti |
| Date Deposited: | 12 Jun 2020 02:43 |
| Last Modified: | 12 Jun 2020 02:43 |
| URI: | http://repository.its.ac.id/id/eprint/76154 |
Actions (login required)
 |
View Item |