AI to optimize raspberry aroma production through fermentation by Phytowelt

Create a fast and reliable analysis system to minimize human errors during quality control

Phytowelt's main challenge lay in its reliance on human evaluations for quality control—an approach that, while traditional, significantly increased the risk of subjective errors and inconsistencies in detecting aroma variations. Moreover, to ensure consistent quality and optimal yield, it was essential to implement precise, real-time management of the fermentation process.

Identify correlations between various parameters and aggregate data to guide the next steps.

To enhance the control and efficiency of the fermentation process, a methodical approach was adopted, comprising two key steps:

• Identifying correlations between fermentation parameters – This initial evaluation revealed significant relationships between various fermentation parameters. These correlations provided valuable insights into how different variables directly influence the quality and yield of the produced aroma, enabling targeted adjustments to optimize the overall process.
• Leveraging data aggregation and clustering – By analyzing different fermentation process states, we simplified the understanding of production patterns and laid the foundation for automated monitoring. This approach improved real-time production efficiency and accuracy, ensuring a more consistent and optimized process.