Nonuniformity in Multi‑Physical Field Drying of Fruits and Vegetables: Mechanisms, Cascade Amplification, and Artificial Intelligence–Driven Control

Nonuniformity is the most critical yet unresolved bottleneck limiting the industrial adoption of multi-physical field drying (microwave, radio frequency, infrared, and ultrasound) for fruits and vegetables. While individual physical field technologies and AI applications have been reviewed separately, a conceptual framework that explains how nonuniformity originates, amplifies, and can be intelligently controlled is currently lacking. This review provides a problem-driven analysis of nonuniformity in physical field drying, proposing a three-stage cascade framework (Initiation → Amplification → Termination) to elucidate the multi-scale coupling between microscopic energy defects and macroscopic quality failure. We further critically evaluate the transition from black-box artificial intelligence (AI) modeling to advanced closed-loop intelligent control strategies involving physics-informed neural networks (PINNs) and digital twins, structured along a perception-prediction-control paradigm. The cascade framework reveals that initial minor heterogeneities are self-reinforced through positive feedback loops, necessitating early-warning prediction. AI-driven adaptive control, coupled with synergistic cross-field designs, may provide opportunities to interrupt this amplification process at early stages. We provide a strategic roadmap emphasizing multimodal sensing and physics-based hybrid modeling to achieve precision, uniformity, and sustainability in next-generation industrial drying systems. This review offers a distinctive problem‑oriented perspective that complements prior technology‑centric surveys by placing nonuniformity as the central narrative and offering a conceptual framework for its prediction and control.