Acoustic emission monitoring research on mural relic pigment-layer disruption under dynamic temperature and humidity loads

The ancient murals show multiple types of damage, which include pigment-layer disruption that endangers building stability yet proves difficult to fix. A controlled temperature-humidity cycling device was employed to simulate accelerated mural aging while enabling synchronized multi-sensor monitoring. The collected signals were processed through an integrated pipeline comprising autoencoder-based feature extraction, Gaussian anomaly detection, and fuzzy comprehensive evaluation. The results demonstrate three key findings: (1) acoustic emission effectively monitors the complete process of salt dissolution and crystallization; (2) the autoencoder-Gaussian anomaly detection system with fuzzy grading effectively reduces noise to detect faint events which enables the tracking of gradual damage progression; (3) salt migration and damage development occur at different rates based on temperature and humidity fluctuations yet stable environmental conditions reduce the rate of damage progression. The framework enables fast diagnosis and precise severity measurement, allowing rapid identification of early-stage deterioration changes associated with salt activity.