Pemodelan Profil Temperatur Bubuk Basah Teh Hitam Grade 1 Pada Proses Oksidasi Enzimatis Menggunakan Jaringan Syaraf Tiruan

Alatas, Fathiah (2008) Pemodelan Profil Temperatur Bubuk Basah Teh Hitam Grade 1 Pada Proses Oksidasi Enzimatis Menggunakan Jaringan Syaraf Tiruan. Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 1106201722-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (37MB) | Request a copy

Abstract

Oksidasi enzimatis dalam pengolahan teh hitam berpengaruh terhadap kualitas dari teh hitam.Pemodelan profil temperaturSalah satu faktor yang merrtpengaruhi oksidasi enzimatis teh hitam adalah ternperatur ruangan. Temperatur ruangan selama proses oksidasi enzimatis berpengaruh terhadap temperatur b d1Juk basah (hasil pelayuan dan penggilingan), Penelitian ini bertujuan untuk meng~tahui pr6fil temperatur bubuk basah selama proses oksidasi enzimatis dengan melakukan pemodelan. Pemodelan dengan teknik identifikasi yang didasarkan hasil pengukuran masukan keluaran sistem (black box identification). Proses oksidasi enzitnatis pada pengolahan teh hitam merupakan sistem non-linier, maka identifikasi dilakukan dengan menggunakan jaringan syaraf tiruan . Sebagai masukan dan keluaran sistem berturut-turut adalah temperatur ruangan dan temperatur bubuk basah. Agar pengamatan proses oksidasi enzimatis mendekati kondisi sebenarnya maka pengambilan data dilakukan secara on-line. Pengamatan dilak-ukan dengan pos.isi sensor 2 cm dan 5 cm dari dasar paparan pada ketebalan paparan 9 cm. Model profil temperatur bubuk basah grade I pada proses oksidasi ehzimatis menggunakan JST diperoleh. Struktur model JST diperoleh satu lapisan masukan dengan 12 unit, satu Iapisan tersembunyi dengan I2 unit dan satu lapisan keluaran dengan satu unit. Dari seluruh proses training didapatkan nilai error terkecil sebesar 9, 99844 X I 0"5 dengan target error I 0"5. Hasil validasi menunjukkan bahwa nilai rerata dari residu mendekati nol (white noise terpenuhi) dan memiliki nilai variansi kecil. Nilai statistik autokorelasi dari residu dan korelasi silang antara residu dan masukan telahmemenuhi kriteria dengan tingkat kepercayaan 95 % kecuali pada pengujian pertama
===================================================================================================================================
Enzymatic oxidation in black tea processing affects the quality of black tea. Temperature profile modeling One of the factors that affect the enzymatic oxidation of black tea is the room temperature. The room temperature during the enzymatic oxidation process affects the temperature of the wet powder (the result of withering and grinding). This study aims to determine the temperature profile of the wet powder during the enzymatic oxidation process by modeling. Modeling with identification techniques based on the results of system input and output measurements (black box identification). The enzymatic oxidation process in black tea processing is a non-linear system, so identification is carried out using an artificial neural network. The input and output of the system are room temperature and wet powder temperature, respectively. In order for the observation of the enzymatic oxidation process to approach the actual conditions, data collection is carried out on-line. Observations are made with sensor positions of 2 cm and 5 cm from the base of the exposure at an exposure thickness of 9 cm. The temperature profile model of grade I wet powder in the enzymatic oxidation process using ANN is obtained. The structure of the ANN model obtained one input layer with 12 units, one hidden layer with I2 units and one output layer with one unit. From the entire training process, the smallest error value was obtained at 9.99844 X I 0"5 with a target error of I 0"5. The validation results show that the average value of the residual is close to zero (white noise is met) and has a small variance value. The statistical value of autocorrelation of the residuals and cross-correlation between the residuals and inputs has met the criteria with a confidence level of 95% except in the first test

Item Type:	Thesis (Masters)
Additional Information:	RTFi 006.32 Ala p-1 2008
Uncontrolled Keywords:	Bubuk basah, ldentifikllsi Sistem, Jaringan Syaraf Tiruan, Oksidasi, Enzimatis, Pemodelan; Wet Powder, Identification System, Artificial Neural Network, Oxidation, Enzymatic, Modeling
Subjects:	Q Science > QA Mathematics > QA336 Artificial Intelligence Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science)
Divisions:	Faculty of Mathematics and Science > Physics > 45101-(S2) Master Thesis
Depositing User:	EKO BUDI RAHARJO
Date Deposited:	16 Oct 2025 08:21
Last Modified:	16 Oct 2025 08:21
URI:	http://repository.its.ac.id/id/eprint/128601

Actions (login required)

View Item