Perancangan Soft Sensor Untuk Memprediksi Komposisi Produk Pada Kolom Distilasi Melalui Distributed Control Systems dan Ole For Process Control

Pamungkas, Rahadian Agnies Septanto (2017) Perancangan Soft Sensor Untuk Memprediksi Komposisi Produk Pada Kolom Distilasi Melalui Distributed Control Systems dan Ole For Process Control. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.

[img] Text
2414106025-Undergraduate-Theses.pdf
Restricted to Repository staff only

Download (1MB) | Request a copy

Abstract

Kualitas komposisi produk merupakan salah satu parameter keberhasilan dari suatu industri proses. Untuk memperoleh kualitas komposisi produk yang sesuai yaitu dengan menerapkan instrumen proses kontrol. Namun pengukuran yang dilakukan oleh instrumen analiser memiliki kinerja respon yang lambat, kurangnya kehandalan, dan mahal. Oleh karena itu dibuatlah estimasi berupa soft sensor yaitu sensor yang berbasis model dari masukan berupa temperatur dan tekanan yang diperoleh dari proses kolom distilasi debutanizer. Pemodelan soft sensor dilakukan menggunakan metode jaringan syaraf tiruan (JST) dan didapatkan nilai RMSE dari keluaran berupa komposisi distilat (Xd) sebesar 0.00000709 kgmole/h dan komposisi bawah (Xb) sebesar 0.00002617 kgmole/h. Dengan memanfaatkan Distributed Control systems dan OLE for Process Control maka soft sensor dapat dirancang dengan menanamkan bobot yang telah diperoleh sehingga hasil prediksi komposisi dapat dipantau dan dievaluasi. Nilai prediksi komposisi yang dihasilkan dari soft sensor divalidasi dengan data komposisi proses kolom debutanizer dan menghasilkan penyimpangan maksimal sebesar 0.61%. ==================================================================================== The quality of product composition is one of the parameters to be achieved of an industrial process. To obtain the appropriate quality product composition that is able to implement process control instruments. But measurements made by the analyzer instruments have slow response performance, lack of reliability, and expensive. Therefore made estimates in the form of soft sensor is a sensor based on models of the input of temperature and pressure were obtained from the distillation column debutanizer. Soft sensor modeling is using by method of artificial neural network (ANN) and RMSE values obtained from the output of the distillate composition (Xd) of 0.00000709 kgmole / h and bottom composition (Xb) of 0.00002617 kgmole / h. By using Distributed Control systems and OLE for Process Control, then soft sensors can be designed to implant a weight that has been obtained so that the predicted composition can be monitored and evaluated. Predictive value of the resulting composition of the soft sensor is validated with composition data distillation column debutanizer process and produces a maximum error of 0.61%.

Item Type: Thesis (Undergraduate)
Uncontrolled Keywords: Artificial Neural Network; Distillation Column; Distributed Control Systems; Soft Sensor; Jaringan Syaraf Tiruan; Kolom Distilasi
Subjects: T Technology > TJ Mechanical engineering and machinery > TJ217.6 Predictive Control
T Technology > TJ Mechanical engineering and machinery > TJ223_Programmable controllers
Divisions: Faculty of Industrial Technology > Physics Engineering > (S1) Undergraduate Theses
Depositing User: - RAHADIAN AGNIES SEPTANTO PAMUNGKAS
Date Deposited: 17 Apr 2017 03:00
Last Modified: 22 Dec 2017 06:50
URI: http://repository.its.ac.id/id/eprint/3578

Actions (login required)

View Item View Item