Concrete Dam Behavior Prediction Using Multivariate Adaptive Regression Splines with Measured Air Temperature

摘要

This paper presents a dam health monitoring model using long-term air temperature based on multivariate adaptive regression splines (MARS). MARS is an intelligent machine learning technique that has been successfully applied to deal with nonlinear function approximation and complex regression problems. The proposed long-term air temperature-based dam health monitoring model was verified on a real concrete gravity dam with efficient safety monitoring data. Results show that the proposed approach is promising for concrete dam behavior modeling considering the prediction error is much reduced.