Modeling and Optimization of Triticale Wort Production Using an Artificial Neural Network and a Genetic Algorithm

Area of Science:

• Brewing Science
• Agricultural Chemistry
• Process Optimization

Background:

• Triticale, a wheat-rye hybrid, is a promising brewing adjunct due to its favorable characteristics.
• Optimizing triticale wort production requires understanding complex interactions between various processing parameters.
• Traditional methods for optimizing brewing processes can be time-consuming and may not capture intricate relationships.

Purpose of the Study:

• To model and optimize the triticale mashing process for improved wort production.
• To evaluate the impact of triticale variety, ratio, enzyme concentration, and mashing regime on wort characteristics.
• To implement artificial neural networks (ANNs) and genetic algorithms (GAs) for process optimization.

Main Methods:

• Investigated two triticale varieties (malted and unmalted) at different ratios.
• Applied varying mashing regimes and concentrations of Shearzyme® 500 L enzyme.
• Utilized ANNs to model the mashing process and a GA to optimize key variables.
• Validated ANN model predictions through laboratory-scale mashing experiments.

Main Results:

• The GA identified optimal input parameters for triticale ratio (23%), enzyme ratio (9%), mashing regime (1), and triticale variety (3).
• The ANN model predicted wort extract (8.65%), viscosity (1.52 mPa·s), and FAN content (148.32 mg/L) under optimized conditions.
• Experimental validation closely matched model predictions, with wort extract (8.63%), viscosity (1.51 mPa·s), and FAN (148.88 mg/L).

Conclusions:

• The integrated ANN and GA approach effectively models and optimizes triticale wort production.
• The study confirms the practical applicability of computational methods for enhancing brewing efficiency and quality.
• Optimized triticale wort production shows significant potential for the brewing industry.