Damage detection of beam bridge under a moving load using Auto-encoder

A novel damage detection approach based on Auto-encoder neural network is proposed to identify in beam-like bridges subjected a moving mass. In this approach, several sensors are used measure structural vibration responses during mass across the bridge. An auto-encoder (AE) designed extract features from measured responses. fixed window cut out time-domain generate inputs of AE network. Moreover, some constraints applied hidden layer improve performance training process. When complete, encoder was regarded as feature extractor. And index defined cosine distance between two vectors obtained adjacent data windows. By along data, we can calculate series and locate position structure. To demonstrate method, numerical simulation carried out. The results show that method accurately both single multiple damages using acceleration response. It infers promising for detection.