Hybrid Intelligence for Fouling Prediction and Adaptive Clean-in-Place Optimization in Food Processing Heat...

Fouling remains one of the principal challenges limiting the thermal efficiency, hygienic performance, and sustainability of food processing heat exchangers. The accumulation of protein, mineral, biofilm, and particulate deposits increases thermal resistance, energy consumption, production losses, and the frequency of cleaning-in-place (CIP) operations, thereby affecting process economics and food safety. This systematic review critically evaluates advances in hybrid intelligence for fouling prediction and adaptive CIP optimization in plate and tubular heat exchangers used in dairy, beverage, and liquid food processing. A structured literature review of publications from 2000 to 2026 was conducted using Web of Science, Scopus, PubMed, IEEE Xplore, and Google Scholar. Unlike previous reviews that address these topics separately, this review integrates fouling mechanisms, predictive modeling, intelligent sensing, explainable artificial intelligence (AI), adaptive CIP, and digital twins within a unified engineering framework for food processing. The evidence indicates that hybrid intelligence approaches outperform purely mechanistic or data-driven models by improving predictive accuracy, robustness under variable operating conditions, physical consistency, and model interpretability. AI-enabled condition-based CIP strategies also demonstrate considerable potential to reduce water, chemical, and energy consumption while minimizing production downtime without compromising hygienic performance. However, industrial implementation remains constrained by limited food-specific datasets, insufficient large-scale validation, uncertainty quantification, model explainability, and integration with hazard analysis and critical control point-based food safety systems. Future research should prioritize physics-informed, food-specific hybrid intelligence frameworks integrating real-time sensing, digital twins, adaptive CIP control, and explainable decision support to enable sustainable, resilient, and intelligent food manufacturing systems.