Integrating Machine Learning Classification with Thermal Integrity Profiling for Concrete Pile Assessment
Please login to view abstract download link
Thermal Integrity Profiling (TIP) is a non-destructive monitoring technique used to assess the integrity of concrete piles and foundations by analysing hydration temperature data for defect detection. Currently, TIP is limited by manual interpretation, which can lead to inconsistencies and errors in distinguishing between defects and normal temperature variations. By integrating this new ML approach with TIP, defects can be more accurately classified for any concrete composition, improving the reliability of the assessment. The presented model demonstrates high predictive accuracy for defect detection within concrete piles in the context of TIP. Implementing the ANN in real-world projects can enable the detection of construction defects in near real-time, benefiting the safety of concrete foundations and the overall efficiency of construction.
