Rancang Bangun Hidung Elektronik (E-Nose) untuk Klasifikasi Mutu Kopi Arabika

Cahyafazri, Muhammad Rafi (2026) Rancang Bangun Hidung Elektronik (E-Nose) untuk Klasifikasi Mutu Kopi Arabika. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 2042221060-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (9MB) | Request a copy

Abstract

Sistem E-NOSE yang dikembangkan terdiri atas delapan sensor gas berbasis metal-oxide semiconductor (MOS), dua pompa, dua katup (valve), dan sebuah chamber pengukuran yang dirancang untuk menghasilkan pola sinyal volatile yang stabil dan berulang. Pengukuran dilakukan selama 300 detik dalam tiga fase, yaitu baseline, exposure, dan purging. Sampel kopi Arabika seberat enam gram dipanaskan pada suhu 100 °C untuk memaksimalkan pelepasan senyawa volatile sebelum dialirkan ke dalam chamber. Hasil pengujian menunjukkan bahwa respons keseluruhan sensor pada sampel kopi Arabika mutu tinggi (High Grade) berada pada rentang 700–4500 mV, sedangkan pada sampel mutu rendah (Low Grade) berada pada rentang 500–3500 mV, yang mengindikasikan perbedaan karakteristik volatile yang signifikan antar kelas mutu. Data sinyal sensor diproses dalam bentuk time-series dan disusun menjadi dataset yang terdiri atas 48 sampel data pelatihan dan 12 sampel data validasi untuk dua kelas mutu. Model CNN dilatih untuk mengekstraksi fitur temporal, termasuk puncak respons, fase stabilisasi, dan pola penurunan konsentrasi senyawa volatile. Evaluasi model menunjukkan nilai precision, recall, dan F1-score sebesar 1,0 pada kedua kelas, baik pada training set maupun validation set, dengan akurasi mencapai 100%. Hasil tersebut menunjukkan kemampuan generalisasi model yang sangat baik dalam membedakan mutu kopi Arabika. Secara keseluruhan, integrasi sistem E-NOSE dan CNN terbukti efektif sebagai metode evaluasi mutu kopi yang cepat, non-destruktif, dan akurat berbasis karakteristik volatile aroma.
====================================================================================================================================
The developed E-NOSE system consists of eight metal-oxide semiconductor (MOS) gas sensors, two pumps, two valves, and a measurement chamber designed to generate stable and repeatable volatilee signal patterns. Measurements were conducted for 300 seconds in three phases, namely baseline, exposure, and purging. Six grams of Arabica coffee samples were heated at 100 °C to maximize the release of volatilee compounds before being introduced into the chamber. Experimental results show that the overall sensor responses for high-grade Arabica coffee samples ranged from 700 to 4500 mV, while those for low-grade samples ranged from 500 to 3500 mV, indicating significant differences in volatilee characteristics between grade classes. Sensor signals were processed as time-series data and organized into a dataset consisting of 48 training samples and 12 validation samples for two grade classes. The CNN model was trained to extract temporal features, including response peaks, stabilization phases, and decay patterns of volatilee compound concentrations. Model evaluation demonstrated precision, recall, and F1-score values of 1.0 for both classes on the training and validation sets, with overall accuracy reaching 100%. These results indicate excellent generalization capability in distinguishing Arabica coffee grade. Overall, the integration of the E-NOSE system and CNN proves to be an effective, fast, non-destructive, and accurate method for coffee grade evaluation based on aroma volatilee characteristics.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Arabica coffee, E-NOSE, volatile compounds, CNN, grade classification, gas sensor
Subjects:	T Technology > T Technology (General) T Technology > T Technology (General) > T57.5 Data Processing T Technology > T Technology (General) > T57.74 Linear programming T Technology > T Technology (General) > T58.62 Decision support systems T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK7882.P3 Pattern recognition systems T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK7895.G36 Field programmable gate arrays--Design and construction.
Divisions:	Faculty of Vocational > Instrumentation Engineering
Depositing User:	Muhammad Rafi Cahyafazri
Date Deposited:	04 Aug 2026 01:17
Last Modified:	04 Aug 2026 01:17
URI:	http://repository.its.ac.id/id/eprint/142728

Actions (login required)

View Item