Predicting salting time in dry-cured hams using linear mixed models
Industrial production of dry-cured hams involves three stages: reception and salting, post-salting, and drying maturation. We focus on the salting phase, essential for determining final salt content and reducing variability between pieces. We built a predictive linear mixed model (LMM) based on non-invasive X-ray measurements and longitudinal data to optimize salting duration and achieve a target salt percentage; salt reduction aligns with WHO recommendations. The LMM can be implemented directly on the production line, enabling autonomous decision-making based on data, following Industry 4.0 principles. The classical system determines salting time based on ham weight (1 day/kg), while the LMM allows setting a salt target and reduces variability. The classical system results in 0.72% variability, while LMM achieves 0.32%. LMM also achieves a 4.5% average salt without bias, vs. the biased 5.27% in the classical method. Overall, this leads to a more efficient industrial production process
Keywords: Linear mixed model dry-cured-ham Salt content Salting time