A damage recognition method for civil engineering structures bsed on MTL-1DCNN

Key Points

• Effective damage recognition was observed with a high accuracy rate of 92% in testing across various conditions.
• The study employs a multi-task learning approach, utilizing the MTL-1DCNN model optimized for spatial-temporal feature extraction.
• A comprehensive analysis was conducted on multiple engineering structures subjected to various stress tests.
• This method shows promise for advancing civil engineering techniques, highlighting the need for further real-world validation.