Optimisasi Konsumsi Energi Pada Proses Kolom Depropanizer Berbasis Neural Network - Particle Swarm Optimization (NN-PSO)

Amir Akbar, Ian Haikal (2020) Optimisasi Konsumsi Energi Pada Proses Kolom Depropanizer Berbasis Neural Network - Particle Swarm Optimization (NN-PSO). Undergraduate thesis, Institut Teknologi Sepuluh Nopember.

[img] Text
02311640000021-Undergraduate_Thesis.pdf - Accepted Version
Restricted to Repository staff only

Download (4MB) | Request a copy

Abstract

Kolom Distilasi merupakan plant yang digunakan untuk memisahkan komponen senyawa hydrocarbon pada crude oil dan natural gas berdasarkan nilai volatilitasnya. Penggunaan energi yang tinggi pada equipment Condenser dan Reboiler menuntut adanya optimisasi konsumsi energi kolom distilasi untuk menekan biaya utilitas cooling water dan steam. Metode optimisasi dilakukan dengan cara mencari nilai variabel operasional pada kolom distilasi tersebut yang menghasilkan konsumsi energi minimum. Proses pencarian nilai variabel operasional kolom distilasi dilakukan dengan menggunakan metode Stochastic Algorithm yaitu Particle Swarm Optimization (PSO). Kolom distilasi yang bersifat kompleks dan non linier dimodelkan dengan Neural Network (NN), dan keluaran dari model NN digunakan untuk penentuan fungsi obyektif pada algoritma PSO. Variabel operasional dari kolom depropanizer yang akan dilakukan optimisasi yaitu Feed – Molar Flow (F), Feed – Temperature (Tf), Condenser – Pressure (Pc), Reboiler – Pressure (Pr). Performansi arsitektur Neural Network Feed Forward - Backpropagation dengan algoritma training Lavenberg Marquart pada jumlah neuron 19 menghasilkan nilai RMSE Qc, RSME Qr, dan RMSE total berturut – turut sebesar 5,541 x 10-4 ; 6,310 x 10-4 ; 8,398 x 10-4. Metode optimisasi konsumsi energi menggunakan algoritma PSO menghasilkan penghematan konsumsi energi pada Reboiler dengan nilai efisiensi Er sebesar 29,27 %, sedangkan pada equipment Condenser terjadi peningkatan konsumsi energi dibandingkan dengan kondisi eksisting sehingga nilai efisiensi Ec sebesar -33,75 %, sehingga secara keseluruhan dapat memberikan penghematan biaya utilitas sebesar 16,92 %. ======================================================================================================================== Distillation column is a plant that is used to separate components of hydrocarbon compounds in crude oil and natural gas based on their volatility values. The high energy usage of Condenser and Reboiler equipment demands the optimization of the distillation column energy consumption to reduce the utility cost of cooling water and steam. Optimization methods performed by searching operational condition variable values distillation column to produce minimum energy consumption, the process of finding the value of distillation column operational variables using Stochastic Algorithm, that is Particle Swarm Optimization (PSO). Distillation columns that are complex and non-linear can be modeled using Neural Network (NN), and the output from the NN model is used to determine the objective function in the PSO algorithm. The operational variables of the Depropanizer Column to be optimized are Feed - Molar Flow (F), Feed - Temperature (Tf), Condenser - Pressure (Pc), Reboiler - Pressure (Pr). Architecture performance of Neural Network Feed Forward - Backpropagation with Lavenberg Marquart's training algorithm on the number of neurons 19 produces RMSEQc, RSMEQr, and RMSEtotal values respectively 5,541 x 10-4; 6,310 x 10-4; 8,398 x 10-4. Energy consumption optimization method using the PSO algorithm result in savings of energy consumption in the reboiler with Er efficiency value of 29.27%, while in Condenser equipment there is an increase in energy consumption compared to existing conditions, so the Ec efficiency value of -33.75%, so that as a whole can provide utility cost savings amounting 16.92%.

Item Type: Thesis (Undergraduate)
Uncontrolled Keywords: Depropanizer Column, Distillation Column, Neural Network, Optimization, Particle Swarm Optimization, Kolom Depropanizer, Kolom Distilasi, Neural Network, Optimisasi, Particle Swarm Optimization.
Subjects: Q Science > QC Physics > QC320 Heat transfer
Divisions: Faculty of Industrial Technology and Systems Engineering (INDSYS) > Physics Engineering > 30201-(S1) Undergraduate Thesis
Depositing User: Ian Haikal Amir Akbar
Date Deposited: 06 Aug 2020 05:33
Last Modified: 06 Aug 2020 05:35
URI: https://repository.its.ac.id/id/eprint/77077

Actions (login required)

View Item View Item