Evaluation of fermentation parameters in black beer industrial production : a case study at My Loi An Co., Ltd (Vietnam)
Tran, Ngoc My Linh (2026)
Diplomityö
2026
School of Engineering Science, Kemiantekniikka
Kaikki oikeudet pidätetään.
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi-fe2026052958180
https://urn.fi/URN:NBN:fi-fe2026052958180
Tiivistelmä
This thesis studies how to optimize fermentation parameters for black beer production through a data-driven case study approach. The research uses industrial fermentation data of 12 batches from My Loi An Co., Ltd. (Vietnam), of which some are black beer and others golden beer.
The impact of temperature, pitching rate, and oxygenation during fermentation process was first studied by exploratory data analysis, then combined with multiple linear regression modeling to provide analytical results. Fermentation time, apparent attenuation, and alcohol by volume (ABV) were the main performance parameters used.
The regression model was a quite good predictor of the fermentation time (R² = 0.779, p = 0.0053) and it was indication that the fermentation time was shorter when the yeast was exposed to higher levels of oxygen. Oxygenation was proven as the only significant factor in the model (p = 0.037). The attenuation model showed the best result (R² = 0.854, p = 0.0010) where pitching rate was displayed as a significant the attenuation was negatively impacted by pitching rate (p = 0.0163). On the other hand, the ABV model illustrated poor explanatory power due to the limited variability of ABV values produced batches (R² = 0.545, p = 0.084). Strong multicollinearity was also observed between pitching rate and oxygenation leading to the conclusion that these two parameters should be considered together (r = -0.902). Type of beer had a significant influence on the fermentation behavior in industrial conditions which was explained by the clear separation of black and golden beer in the clustering.
Due to the small size of the dataset and the study being observational, optimization was done by interpreting regression results rather than using formal optimization techniques. Some practical operating ranges are suggested to help inform process improvement under industrial conditions. Findings here are initial ones, and it is advisable to carry out further research with bigger datasets.
The impact of temperature, pitching rate, and oxygenation during fermentation process was first studied by exploratory data analysis, then combined with multiple linear regression modeling to provide analytical results. Fermentation time, apparent attenuation, and alcohol by volume (ABV) were the main performance parameters used.
The regression model was a quite good predictor of the fermentation time (R² = 0.779, p = 0.0053) and it was indication that the fermentation time was shorter when the yeast was exposed to higher levels of oxygen. Oxygenation was proven as the only significant factor in the model (p = 0.037). The attenuation model showed the best result (R² = 0.854, p = 0.0010) where pitching rate was displayed as a significant the attenuation was negatively impacted by pitching rate (p = 0.0163). On the other hand, the ABV model illustrated poor explanatory power due to the limited variability of ABV values produced batches (R² = 0.545, p = 0.084). Strong multicollinearity was also observed between pitching rate and oxygenation leading to the conclusion that these two parameters should be considered together (r = -0.902). Type of beer had a significant influence on the fermentation behavior in industrial conditions which was explained by the clear separation of black and golden beer in the clustering.
Due to the small size of the dataset and the study being observational, optimization was done by interpreting regression results rather than using formal optimization techniques. Some practical operating ranges are suggested to help inform process improvement under industrial conditions. Findings here are initial ones, and it is advisable to carry out further research with bigger datasets.