Belt conveyor support structures are under corrosive conditions because of dust accumulation. To assess the structural condition, the use of Cross-Sectional vibration Mode (CSM) has been studied. CSM, which can be estimated even under the plant operation, has been shown to have sensitivity to damage to a member, but the quantitative evaluation of the damage is yet to do. In this paper, the minimum cross-sectional plate thickness, which determines the load-carrying capacity of the tension member, is estimated using a two-step method. By using machine learning in the first step and modifying the estimated value in the second step, the method can estimate the minimum cross-sectional plate thickness with high accuracy and stability.
