Computer Simulation Models for Consideration of Seasonal Trends Influence on the Structural Dynamics of Bridges

Abstract

Elimination of seasonality temperature trends in a structural health monitoring of bridges is considered in the article. On example of a beam bridge, eigenfrequencies are plotted against seasonal fluctuations of environmental temperature. Next, analysis of a time series, formed from a data of frequency changes in the bridge, was made. To describe the time series, two different methods were implemented: ARIMA model based on a statistical relationship between the data and LSTM model of a recurrent neural network.

Keywords

• computer simulation
• seasonality elimination
• structural health monitoring
• time series
• forecast
• shm
• arima
• recurrent neural network