A Functional Principal Component Analysis Framework for Vibration‐Based Structural Condition Assessment of Highway Bridge Spans

The management of lifespan and safety performance, aimed at ensuring the structural integrity and durability of highway infrastructure, has become an increasingly important area of research. In civil engineering, traditional methods for evaluating dynamic behavior, such as Operational Modal Analysis and the Peak Picking technique, often fall short in detecting subtle signs of damage. This study proposes a novel approach based on Functional Principal Component Analysis (FPCA) to enhance the sensitivity of vibration‐based structural condition and anomaly assessment. A preliminary analysis based on classical PCA applied to the same dataset was previously presented in Agrò et al. (2025a). That study highlighted the capability of PCA to capture dominant variance patterns in the vibration signals. However, PCA operates on discretized time series and does not explicitly account for their functional nature. In contrast, the FPCA framework proposed in this study models the signals as continuous functions, enabling a more refined characterization of temporal variability and improving sensitivity to subtle structural differences. By comparing the optimal functional subspaces derived from FPCA for both intact and potentially compromised bridge spans, we quantify structural differences through the cosine of principal angles. The approach is applied to a real case involving a viaduct on the Palermo–Catania highway, demonstrating its potential to reveal anomalies not evident through conventional modal analysis.