Statistical Models for Temperature Prediction in Civil Structures

Abstract

This paper presents two models for reconstructing the thermal field of civil structures without relying on thermocouple data, using instead information from weather stations located near the structure under analysis. The first model estimates temperature through a linear correlation with air temperature, while the second adopts a physics-based approach that also accounts for the effects of solar radiation. Both models were applied to the bell tower of Portogruaro. The physics-based model demonstrated high accuracy, with a prediction uncertainty of 1.50 °C and a correlation coefficient of 0.99 between predicted and measured temperatures.