Electronic nose usage on pattern recognition of volatile organic compounds for prediction contaminated and non-contaminated bacterial cellulose fermentation

Volatile organic compounds (VOC) are released during bacterial cellulose (BC) fermentation.Fermentation time influences the VOC produced.The gas profile in BC fermentation, consisting of various VOC, is the main criterion used to assess BC fermentation quality.The occurrence of contamination will cause the VOC profile in the fermenter to change.The ability to differentiate VOC due to contamination is another approach to detect contamination more quickly.This research was conducted to detect and predict the detection accuracy for the success or failure of the BC fermentation process through a gas profile consisting of VOC formed using an electronic nose.Fermentation was treated with contamination and non-contamination treatments.The sensor response data will then be processed using signal pre-processing, feature extraction, and the principal component analysis (PCA) method.The experimental results prove that the enose can detect the success or failure of the BC fermentation process on the third day for PCA analysis, which uses the best feature extraction parameter, namely the mean value.The accuracy value for prediction with the neural model was 87.5% using MindsDB machine learning.