Rancang Bangun Alat Deteksi Bahan Baku Gelatin Menggunakan Multisensor Gas Berbasis Machine learning

Febby Trishe Ananda (Departemen Kimia, Fakultas Matematika dan Ilmu Pengetahuan Alam, Universitas Andalas)
Harmadi Harmadi (Departemen Kimia, Fakultas Matematika dan Ilmu Pengetahuan Alam, Universitas Andalas)
Wildan Panji Tresna (Badan Riset dan Inovasi Nasional)
Aditya Eka Mulyono (Badan Riset Inovasi Nasional (BRIN))

Jurnal Fisika Unand
Vol 15 No 4 (2026)

In The demand for gelatin for halal food ingredients in Indonesia is very high. Pork-based gelatin, which is difficult to identify after undergoing the production process, is still widely used in the food and pharmaceutical industries. Analytical methods such as PCR and FTIR offer high accuracy, but they require high operational costs and specialized laboratory facilities. This study develops a machine learning-based electronic nose (e-nose) system to detect and classify gelatin types based on their raw material sources. This system integrates eight types of gas sensors to detect volatile organic compounds (VOCs) produced by each type of gelatin, which are then classified by PCA and predicted by LDA. The study will use gelatin with raw materials sourced from cows, pigs, and fish with varying gelatin concentrations of 1%, 3%, 5%, and 7%. Testing will be conducted for 30 minutes for each sample. The results of the testing prove that the device is capable of detecting and distinguishing gelatin based on its raw materials with a detection accuracy of 95%.