Validation of size estimation of debonds in external wall’s composite finishes via passive Infrared thermography and a gradient algorithm

• Passive infrared thermography in external wall’s building inspection was validated.
• In render, accuracy is at most 15% for cover depths of debond up to 20 mm. Accuracies are dependent on and can vary with different debond sizes.
• In tile-render, accuracy is at most 30% for cover depths of debond up to 9 mm. Accuracies are dependent on and can vary with different debond sizes.
• Sizes of debond deeper than the above cover depths cannot be quantified.

This paper aims to present a study on validating a passive and quantitative infrared thermography (PQIRT) methodology to estimate the size of debond in rendered and tile composites in external walls of high-rise buildings. The accuracies of the PQIRT methodology were evaluated under four conditions, including (i) nine depths of debonds, (ii) three temperature ranges, (iii) two types of composites and (iv) changes of heat flux during a 2-h diurnal and continuous monitoring. The method processed infrared images with a simple spatial pixel differentiation algorithm to determine temperature gradients, where the absolute temperature value was no longer required. The methodology was validated by comparing the estimated and the actual sizes of the debonds by statistical error analysis. In the rendered composites, the percentage error was at most 15% for cover depths of debond up to 20 mm for all temperature ranges. In the tile composites, the error was at most 30% for cover depths of debond up to 9 mm at all temperature conditions, while the estimation was erroneous for cover depths beyond 9 mm. These results will form the basis to develop guidelines for using PQIRT for inspection of high-rise building envelops in densely populated cities, which supplement the current qualitative inspection adopted by a handful of few accredited laboratories in Hong Kong and elsewhere.