Huangjiu lees are rich in flavor compounds, but traditional methods of utilizing them remain inefficient. Developing an efficient process for the rapid release and extraction of these flavor compounds from Huangjiu lees to produce lees-flavored seasonings holds significant promise. To tackle this complex process, a multi-round, progressive optimization strategy was devised. First, four separate orthogonal experiments identified the key factors for pre-fermentation, acid hydrolysis, enzyme hydrolysis, and debittering/decolorization, which were determined to be pre-fermentation time, HCl concentration, flavor enzyme concentration, and activated carbon ratio, respectively. Subsequently, response surface methodology was employed to rank the importance of these four factors in influencing the sensory score of the seasonings. The optimal conditions were determined as follows: pre-fermentation time of 7.785 days, HCl concentration of 14.978 %, flavor enzyme concentration of 6.276 %, activated carbon ratio of 7.496 %, with an expected sensory score of 45.117. Finally, the AI-based ANN-PSO method was applied for further refinement, yielding optimized conditions of a pre-fermentation time of 7.417 days, HCl concentration of 12.569 %, flavor enzyme concentration of 6.843 %, activated carbon ratio of 6.946 %, and an improved expected sensory score of 47.838. Verification experiments confirmed the superiority of the latter conditions. This study not only establishes a novel method for the efficient utilization of the abundant flavor resources in Huangjiu lees, which has substantial market potential, but also demonstrates that combining multiple optimization methods in a stepwise, multi-round approach is effective in addressing complex manufacturing processes involving multiple stages and factors.
Read full abstract