Darmaatmaja, Abdussalam (2019) Pemodelan API Separator pada PT SAKA Indonesia Pangkah Limited Berbasis Jaringan Syaraf Tiruan. Other thesis, Institut Teknologi Sepuluh Nopember.
Preview |
Text
02311540000049-Undergraduate Thesis.pdf Download (4MB) | Preview |
Abstract
Pemodelan API Separator yang merupakan bagian dari sistem Produced Water Treatment pada PT Saka Indonesia Pangkah Limited dilakukan dengan menggunakan Jaringan Syaraf Tiruan (JST). JST yang digunakan berstruktur feedforward backpropagation, dengan beralgoritma Levenberg-Marquardt. JST ini memiliki 3 noda input (pH, Temperature, dan TSS) pada input layer, dan 1 noda output (TSS) pada output layer. Karena metode untuk menentukan hidden node masih belum ada, sehingga dilakukan 20 variasi hidden node pada hidden layer, untuk mencari RMSE validasi terbaik. Hasil simulasi menunjukkan bahwa nilai RMSE terbaik berada pada JST dengan struktur 3-9-1 (input layer-hidden layer-output layer). Nilai RMSE yang dihasilkan bernilai 32.65, yang mana sudah lebih baik dari model JST referensi dengan parameter input yang hampir sama bernilai 37.38. Model JST yang telah dibuat mampu menunjukkan hubungan antara 3 parameter input (pH, Temperature, dan TSS) dengan 1 parameter output (TSS)
=================================================================================================================================
In this final project, a neural network model has been designed to predict the output of API Separator which is a part of produced water treatment at PT Saka Indonesia Pangkah Limited. This neural network has a feedforward structure, and use a Levenberg-Marquardt algorithm. The neural network will have 3 input node in the input layer, and 1 output node in the output layer. Because the method to determine hidden node is still not available, the hidden node is varied until 20 hidden node. The simulation result shows that the best value of RMSE lies in neural network with 3-9-1 structure (Input layer-Hidden layer-Output layer). The RMSE has value of 32.65, which is better than other neural network model found in the references which has a value of 37.38. The neural network that has been designed able to show a relationship between the 3 input parameter (pH, Temperature, TSS) and 1 output parameter (TSS)
| Item Type: | Thesis (Other) |
|---|---|
| Additional Information: | RSF 006.32 Dar p-1 2019 |
| Uncontrolled Keywords: | Produced Water, API Separator, JST |
| Subjects: | T Technology > TD Environmental technology. Sanitary engineering > TD433 Water treatment plants |
| Divisions: | Faculty of Industrial Technology > Physics Engineering > 30201-(S1) Undergraduate Thesis |
| Depositing User: | Abdussalam darmaatmaja |
| Date Deposited: | 17 May 2024 02:31 |
| Last Modified: | 17 May 2024 02:31 |
| URI: | http://repository.its.ac.id/id/eprint/67867 |
Actions (login required)
 |
View Item |