Aulia Ilham, Willy (2023) STUDI SWELLING CAPACITY DARI SUPERABSORBENT POLYMER PADA POLIMERISASI RADIKAL SKALA INDUSTRI DENGAN BASIS MACHINE LEARNING. Master thesis, UNIVERSITAS SULTAN AGENG TIRTAYASA.

	Text (Tesis) 01. Willy Aulia Ilham _ 7780210002 _ Fulltext.pdf Restricted to Registered users only Download (6MB) | Request a copy
	Text (Tesis) 02. Willy Aulia Ilham _ 7780210002 _ CP.pdf Restricted to Registered users only Download (337kB) | Request a copy
	Text (Tesis) 03. Willy Aulia Ilham _ 7780210002 _ 01.pdf Restricted to Registered users only Download (2MB) | Request a copy
	Text (Tesis) 04. Willy Aulia Ilham _ 7780210002 _ 02.pdf Restricted to Registered users only Download (623kB) | Request a copy
	Text (Tesis) 05. Willy Aulia Ilham _ 7780210002 _ 03.pdf Restricted to Registered users only Download (183kB) | Request a copy
	Text (Tesis) 06. Willy Aulia Ilham _ 7780210002 _ 04.pdf Restricted to Registered users only Download (988kB) | Request a copy
	Text (Tesis) 07. Willy Aulia Ilham _ 7780210002 _ 05.pdf Restricted to Registered users only Download (68kB) | Request a copy
	Text (Tesis) 08. Willy Aulia Ilham _ 7780210002 _ REFF.pdf Restricted to Registered users only Download (109kB) | Request a copy
	Text (Tesis) 09. Willy Aulia Ilham _ 7780210002 _ LAMP.pdf Restricted to Registered users only Download (170kB) | Request a copy

Abstract

Superabsorbent polymer (SAP) is a polymer that has many functions in modern human’s life, one of which is as an absorbent material in baby diapers. The growing demand for SAP with high quality SAP makes it necessary to study the effects of production variables on the quality of SAP on an industrial scale. The quality of SAP is determined by the number of Swelling Capacity (SC) where the ideal number is greater than 50 g/g. Many previous studies regarding the effect of synthesis parameters on the number of SC of SAP were limited to the laboratory scale where the polymerization conditions were different from the industrial scale. Also the number of variables and SC data obtained from laboratory scale synthesis were limited. The radical polymerization stage in the reactor with its accompanying operating variables such as reaction temperature, monomer concentration, retention time, crosslinker concentration, addition of nitrogen gas and initiator concentration affects the value of SC in the industrial-scale SAP production process. This study aims to analyse the correlation between the reactor operating conditions and the number of SAP SC obtained. Correlation data analysis was done using microsoft excel with Data Analysis toolpak/python by utilizing production data with a total of 1,562 entries where the correlation between each variables was calculated using the Pearson’s correlation coefficient. Based on the analysis, it was obtained that the variables that had the strongest correlation with the SC were retention time and reaction temperature with correlation coefficient values of 0.31 and -0.26. The weakest correlation was obtained from the addition of nitrogen gas and an initiator with a value of -0.07 and -0.02. Positive values indicate a relationship that is directly proportional while negative values indicate an inverse relationship. Based on the correlation analyses and trends, the ideal polymerization conditions have been identified. Artificial intelligence offers a solution whereby a large data generated from the daily SAP production conditions is correlated to the swelling capacity using machine learning (artificial neural network). The generated mathematical correlation can be used for predictive quality improvement of SAP at an industrial scale. ANN machine learning model, multi-layer feed forward neural network type, backpropagation algorithm with an optimization of architecture for the number of neurons 44 in 1 hidden layer and 4 independent variables for reactor operating conditions, levenberg-marquardt optimation learning function could predict SC values with performance of MAPE 1.4207%, MSE 0.9303, RMSE 0.9645 & R 0.76092. The rate constant of 1st order reaction for SAP radical polymerization with a temperature range of 71-77oC is k = 25.8*106*exp(-45.75/RT) minute-1.

Item Type:	Thesis (Master)
Contributors:	Contribution Contributors NIP/NIM Thesis advisor Kurniawan, Teguh 198305062006041002 Thesis advisor Saepurahman, Saepurahman 198208132020121001
Additional Information:	Superabsorbent polymer (SAP) merupakan polimer yang memiliki banyak fungsi dalam kehidupan manusia modern, salah satunya sebagai bahan penyerap pada popok bayi. Peningkatan permintaan SAP dengan kualitas yang tinggi membuat perlu untuk mempelajari pengaruh variabel produksi terhadap kualitas SAP pada skala industri. Kualitas SAP ditentukan oleh nilai Swelling Capacity (SC) dimana angka idealnya lebih besar dari 50 g/g. Banyak penelitian sebelumnya mengenai pengaruh parameter sintesis terhadap nilai SC SAP akan tetapi masih terbatas pada skala laboratorium dimana kondisi polimerisasinya berbeda dengan skala industri. Selain itu jumlah variabel dan data SC yang diperoleh dari sintesis skala laboratorium terbatas. Tahap polimerisasi radikal pada reaktor dengan variabel kondisi operasi yang menyertainya seperti suhu reaksi, konsentrasi monomer, waktu tinggal, konsentrasi pengikat silang, penambahan gas nitrogen dan konsentrasi inisiator mempengaruhi nilai SC dalam proses produksi SAP skala industri. Penelitian ini bertujuan untuk menganalisis korelasi antara kondisi operasi reaktor dengan nilai SC SAP yang diperoleh. Analisis korelasi data dilakukan menggunakan microsoft excel dengan toolpak/python. Analisis data dengan memanfaatkan data produksi dengan total 1,562 data dimana korelasi antar variabel dihitung menggunakan koefisien korelasi Pearson. Berdasarkan analisis diperoleh variabel yang memiliki korelasi paling kuat dengan SC adalah waktu tinggal dan suhu reaksi dengan nilai koefisien korelasi 0.31 dan -0.26. Korelasi terlemah diperoleh dari penambahan gas nitrogen dan inisiator dengan nilai -0.07 dan -0.02. Nilai positif menunjukkan hubungan yang berbanding lurus sedangkan nilai negatif menunjukkan hubungan yang berbanding terbalik. Berdasarkan analisis korelasi dan trend, kondisi polimerisasi yang ideal telah diidentifikasi. Kecerdasan buatan menawarkan solusi dimana data yang banyak dari kondisi harian produksi SAP dikorelasikan dengan swelling capacity menggunakan pembelajaran mesin (jaringan syaraf tiruan). Korelasi matematis yang dihasilkan dapat digunakan untuk prediksi peningkatan kualitas SAP pada skala industri. Model pembelajaran mesin jaringan syaraf tiruan jenis multi layer feed forward neural network algoritma backpropagation dengan optimasi arsitektur jumlah neuron 44 pada 1 hidden layer dan 4 variabel bebas kondisi operasi reaktor, fungsi pembelajaran optimasi levenberg-marquardt bisa melakukan prediksi nilai SC dengan performance MAPE 1.4207%, MSE 0.9303, RMSE 0.9645 & R 0.76092. Nilai konstanta laju reaksi orde 1 untuk polimerisasi radikal SAP dengan rentang suhu 71-77oC adalah k = 25.8*106*exp(-45.75/RT) menit-1.
Uncontrolled Keywords:	Keywords : superabsorbent polymer, swelling capacity, machine learning, artificial neural network, radical polymerization, correlation data analysis Kata kunci : superabsorbent polymer, swelling capacity, pembelajaran mesin, jaringan syaraf tiruan, polimerisasi radikal, analisis korelasi data
Subjects:	Q Science > QD Chemistry T Technology > TP Chemical technology
Divisions:	08-Pascasarjana 08-Pascasarjana > 84104-Magister Teknik Kimia
Depositing User:	Willy Aulia Ilham
Date Deposited:	09 Jun 2023 16:29
Last Modified:	09 Jun 2023 16:29
URI:	http://eprints.untirta.ac.id/id/eprint/25187

Actions (login required)

View Item