Field sensing of temperature and degree of cure for composites manufacture using parameterized physical‐informed neural networks and sparse FBG sensors

Abstract Monitoring the curing process of polymer composite parts stands imperative for improving the quality and efficiency of manufacturing processes. This paper proposes a field sensing framework that fuses parameterized physics‐informed neural networks (PPINN) and sparse fiber Bragg grating sensors to provide real‐time estimates of temperatures and degree of cure field in composite parts. Moreover, a novel separation network is developed for constructing the PPINN predictor, which integrates the coupled partial differential equation of the thermochemical curing process, achieving data‐free training and real‐time performance. In the simulation experiment, temperature accuracy reached 99.7%, while in laboratory experiments, the accuracy reached 96.1%. Both simulation and experiment demonstrated the successful field sensing of the temperature and degree of cure field in a negligible time with noise resistance ability. Highlights Proposed framework integrates PPINN and sparse FBG sensors for field sensing. PPINN solves thermo‐chemical equations without simulation data required. Simulation accuracy reaches 99.7%, and experimental accuracy reaches 96.1%. The framework shows strong resistance to noise interference in experiments.