Three-Dimensional Printed Subsurface Defect Detection by Active Thermography Data-Processing Algorithm.

Current methods for detecting subsurface flaws in 3D-printed materials struggle with depth sensitivity and noise. Traditional thermography techniques like TSR provide useful data but lack precision for complex geometries. Prior research has shown that heat transfer models can predict defect locations, but these models often fail in real-world AM applications. No prior work had resolved the issue of noise interference in thermography images. This gap motivated the search for a more accurate algorithm. Existing studies have used variance-based metrics, but these can obscure subtle thermal differences. The need for better contrast and sensitivity remains unmet. This paper addresses that need by proposing a new data-processing approach. The goal is to improve detection accuracy in AM materials.

The aim is to evaluate a new thermography algorithm for detecting subsurface flaws in 3D-printed materials. The study focuses on improving contrast and reducing noise in thermal images. It addresses the challenge of identifying defects with varying infill, depth, and size. The motivation comes from the limitations of current TSR techniques in AM quality control. The study seeks to enhance sensitivity to defect depth, a key limitation in existing methods. It also aims to provide clearer images for easier defect localization. The proposed algorithm uses absolute difference instead of variance to improve clarity. The goal is to validate this approach against established methods.

The study uses active thermography with a custom algorithm. It applies thermographic signal reconstruction and thermal contrast principles. The algorithm compares absolute differences in thermal data instead of variance. It analyzes heat transfer dynamics to detect subsurface flaws. The method is tested on materials with defects of different infill, depth, and size. The algorithm is benchmarked against the standard TSR technique. Image quality is assessed based on contrast, sensitivity, and noise levels. The results are compared to determine the algorithm's performance improvements.

The proposed algorithm outperforms standard TSR in detecting subsurface defects. It achieves higher contrast and better sensitivity to defect depth. The images produced are clearer and more reliable for defect localization. The use of absolute difference instead of variance improves thermal clarity. The algorithm reduces noise levels significantly compared to conventional methods. It successfully identifies defects with varying infill and depth characteristics. The results show consistent improvement across multiple test cases. The algorithm provides more accurate and reliable thermal imaging data.

The proposed algorithm demonstrates superior performance in detecting subsurface flaws. It offers better contrast and sensitivity than traditional TSR methods. The use of absolute difference enhances thermal image clarity. The results support the algorithm's effectiveness in AM defect detection. The study confirms the algorithm's ability to reduce noise interference. It provides a reliable alternative to existing thermography techniques. The findings suggest potential for broader application in AM quality control. The authors propose further validation in industrial settings.