Damage detection under environmental and outlier effects using regression coefficients

Damage detection of civil engineering structures using the vibration-based approach is affected by the environmental and operational conditions the structures face. This is because these conditions also affect the vibration properties of the structures, which are commonly analysed to detect damage. Moreover, the presence of outlier measurements also affects the performance of the damage detection methods. Outliers are measurements with abnormal values, and they can create masking effects where small levels of damage are not identified. Therefore, a method is proposed in this paper to detect damage under the effects of environmental and operational conditions, and outlier effects. It is proposed to analyse the regression coefficients of the regression model of the database of vibration properties obtained from a structure for damage detection. The regression coefficients are sensitive to the presence of abnormalities in the database, which can be due to damage and outliers. To test the proposed method, an experimental wooden bridge is analysed, and the results obtained demonstrate that damage can be detected under the influence of environmental and operational conditions, and outlier effects.