Triboelectric Sensing Enables Tactile Perception and Material‐Adaptive Intelligent Roughness Detection for CNC Machine Tools

Abstract

Surface roughness is the core quality indicator for high‐end manufacturing, but traditional inspection technologies are limited by efficiency, universality and environmental robustness. In this study, an intelligent tactile probe (ITTP) based on TENG is proposed to realize the high‐precision online detection of cross material and continuous roughness. Through the biomimetic multi‐layer structure design, ITTP converts the surface morphology into dynamic electrical signals and combines physical mechanism analysis and signal decomposition algorithm to remove the interference of electronic affinity and contact condition fluctuation. In addition, a further innovative hybrid classification and regression dual neural network model is developed to achieve 100% identification on a combination of six engineering materials and five levels of discrete roughness, while the MLP regression model predicted an average error of <5% for a full range of continuous roughness (0.05–12.5 µm). ITTP is successfully embedded in the Computer Numerical Control (CNC) machine tools to build a “perception‐decision‐control” closed‐loop system, promote the transformation of surface inspection from offline sampling inspection to all‐area online inspection, and provide the core support for real‐time process optimization and zero‐defect production for intelligent manufacturing.