Combined effects of pre-harvest drought stress and drying temperature on rice milling quality using multivariate analysis and fuzzy clustering

Abstract Climate-driven drought events, together with post-harvest processing constraints, increasingly threaten rice milling quality, a critical determinant of market value, process efficiency, and food security. Although the individual effects of pre-harvest drought stress and post-harvest drying temperature have been reported, their combined impact on milling performance across rice genotypes has not been systematically evaluated from an engineering perspective. This study demonstrates, for the first time, that the interaction between drought stress and thermal drying induces genotype-specific trade-offs between drying efficiency and milling quality in a fluidized-bed drying system. Six rice genotypes were cultivated under normal and drought irrigation regimes over two growing seasons and subsequently dried at 40, 50, and 60°C in a controlled fluidized-bed dryer. Important process and milling quality indicators, including milling yield (MY), broken rice percentage (BRP), crack percentage (CP), degree of milling, and drying time (DT), were quantified to evaluate system performance. Multivariate statistical analyses (ANOVA, correlation analysis, and PCA) were integrated with deep fuzzy clustering (DFC) to resolve complex response patterns and support decision-oriented evaluation. Drought stress reduced the degree of milling by 12–15% and MY by 3–5%, while increasing BRP and CP by 10–25%. Increasing drying temperature shortened DT by up to 50% but significantly elevated kernel damage, particularly at 60°C, highlighting a critical process trade-off. DFC identified drought-tolerant genotypes that maintained higher MY and lower CP under combined stress. These findings provide an integrated framework for optimizing genotype selection and drying temperature, supporting resilient and resource-efficient rice processing systems.