AE-BP Based Prediction Algorithm for Vibration Anomaly Detection and Intelligent Control of Civil Structures

Abstract

Intelligent structures utilize neural networks for intelligent controller prediction, allowing vibrating structures to adaptively adjust their state under dynamic loading. Intelligent structure technology plays a crucial role in preventing the loss of life and the destruction of structures, especially for large structures with hundreds or thousands of members, and their contents. Therefore, to address the problems of unknown time-varying characteristics of the structural system and uncertainty of the environmental loading of the structural system in the vibration control of large nonlinear structures, this paper designs a smart structure technique based on a self-encoder and BP neural network. The technique aims to detect vibration anomalies in civil structures and predict them for intelligent control. This paper firstly introduces the concept of structural vibration control, selects intelligent control for prediction, and proposes an anomaly detection algorithm based on AE to detect structural vibration anomalies. Through comparison experiments of BP neural networks and LSTM, the BP neural network is finally selected for the intelligent control prediction of civil structure vibration.