Data-Driven Identification of Key Water Quality Parameters for Beverage Production Control

This paper proposes a data-driven approach to improving water quality monitoring and control in the production of drinks. Water parameters, level of pH and chlorine, taste, turbidity, color, hardness, total dissolved solids (TDS), conductivity, oxygen content, temperature, solubility, and flow rate consist of a dataset analyzed with MATLAB. The analysis of the inter-relationships between the parameters was performed using Pearson correlation analysis, and the following calculations of Average Absolute Correlation (AAC) achieved the ranking of the impact of each variable as a whole. The most prominent parameters were determined using a rigorous correlation criterion of |r| ≥ 0.7. The powerful parameters were then graphed with a bar type chart, where the strongest impact and controllable parameters are displayed. The findings showed that the AAC of TDS (0.739), Hardness (0.738), Conductivity (0.737), Color (0.726), Solubility (0.712), Temperature (0.692), Turbidity (0.678) and pH (0.674) had the highest levels, meaning that these parameters were most interdependent and that they could control each other. Conversely, Chlorine (0.569), Oxygen (0.521), Taste (0.453) and Flow Rate (0.231), revealed lesser influences. The results can be used as a basis to apply intelligent control strategies that target the most important parameters, thus enhancing consistency, cost and quality of beverage production systems. Automated systems also enable more rapid responses to deviations, more stable operations, and better long-term process reliability and operational efficiency, by enabling control of the most crucial parameters.