This paper presents a detailed multivariable structural health monitoring (SHM) analysis focused on an industrial conveyor belt system operating under dynamic environmental and mechanical conditions. The main objective is to explore how time- and frequency-domain features, derived from strain, acceleration, and temperature data, can provide insights into failure risks and dynamic behavior. Notably, the Fast Fourier Transform (FFT) is applied to strain gauge signals, a relatively underexplored method in this field, to identify high-frequency vibration modes correlated with operational loading. Additionally, the study quantifies the influence of thermal cycles on stress variations and integrates a probabilistic model to estimate failure risk based on alternating stress amplitude and excitation presence. The findings highlight the critical value of retaining high-frequency vibrational components, often dismissed as noise, which are shown to carry structurally relevant information. This integrative approach not only improves understanding of system dynamics but also supports predictive maintenance practices in harsh industrial environments. The study concludes with discussions on applicability, scalability, and limitations.
