Crack tracking on heat exchanger structures using EikoTwin DIC: A practical case of DIC in the field of thermal fatigue testing @ Liebherr Aerospace

In the aerospace industry, the reliability of components is crucial, especially the study and tracking of cracks. The well-known heat exchanger manufacturer Liebherr Aerospace, with the expertise of EikoSim conducted in-depth research on crack development under extreme thermal fatigue loads. This article provides a comprehensive overview of the methods, results, and conclusions of this collaboration.

Issue: Crack detection and monitoring in fatigue structures

The main goal of this study was to track the propagation of existing cracks and detect any newly appearing cracks. Additionally, to better understand the behavior of the heat exchanger under repeated thermal loads, it is essential to obtain displacement and strain fields with sufficient resolution for each thermal cycle.

Method: Digital Image Correlation (DIC) equipment approach

The study began with thermal stress testing of the exchanger, with temperature cycles ranging from 20 ° C to 600 ° C. The instrumentation included the use of two dual-camera systems for digital image stereo correlation (DIC) to target two critical areas in the exchanger where cracks might appear.

Preparation of the instrumentation simulation

The implementation of the instrumentation system was carefully prepared, fully considering the optimal positioning of the cameras to best capture deformations. The camera layout plan was virtually prepared using EikoTwin Virtual software to ensure the program's feasibility in the crowded area of the test frame. This process included checking the field of view, clarity, camera calibration quality, and estimating measurement accuracy and related uncertainties. For this, virtual images were generated using finite element meshing and numerical simulation of the heat exchanger, then analyzed using EikoTwin DIC image correlation software.

The experimental setup consisted of high-resolution cameras and powerful, uniform lighting positioned to avoid reflections. The lighting had to be strong enough to ensure that external lighting would not affect image brightness, as the test would last hundreds of hours during which ambient lighting might change.

Implementation

Additionally, a speckle coating was applied to the parts to measure deformation during the image correlation process. To ensure accurate deformation measurement through Digital Image Correlation (DIC), a random contrast texture (called speckles) must be applied to the parts. This texture was applied to the exchanger surface using special filters and high-temperature resistant thermal coatings. The measurement area was only 50 x 50 millimeters, requiring about 1 millimeter spatial resolution, corresponding to the size of finite element mesh elements, meaning the speckle size had to be between 0.1 millimeters and 0.2 millimeters.

During testing, the cameras were synchronized with the temperature controller and captured images at each stage of the thermal cycle over thousands of cycles, enabling continuous monitoring of the exchanger's behavior. However, the test also faced significant challenges, especially thermal fog generated at high temperatures, an optical phenomenon caused by beam deflection that could lead to displacement measurement errors. To address this, an air gap was installed and positioned at the location where the exchanger wall was smoothed, greatly reducing the fog's impact.

The reliability of measurements through stereo image correlation depends on precise camera calibration. The calibration process includes determining the camera's focal length, position, and orientation. The specific steps are: first using markers on the parts and the finite element model, then completing calculations with the speckle texture. To ensure calibration accuracy, this calibration was performed in the actual test environment before the thermal test began.

EikoTwin DIC software was used to process the captured images. This digital image correlation software was used to calibrate the cameras and track the displacement and deformation of the exchanger structure by analyzing differences between images captured at different stages of the thermal cycle.

After completing camera calibration, the software used the stereo vision system to identify deviations between the nominal geometry and the observed shape in the measurement area. This measurement deviation characterizes the internal (welded) structure of the exchanger.

The first displacement measurements on the cracked heat exchanger showed that existing cracks were clearly distinguishable in the displacement field, characterized by a distinct gradient from negative to positive displacement. This monitoring confirmed the progression of existing cracks during the thermal cycling process. Data analysis also showed higher strain in the components where existing cracks were located, confirming deformation concentration around these critical areas.

Results: Rapid crack localization

For measurement areas where cracks initially did not appear, no new cracks were detected, but diffuse deformation concentration was measured in locations similar to where cracks are prone to appear. This area may become a site for new crack initiation and should be carefully monitored.

Nevertheless, using image correlation residuals allows for the finest observations, enabling pixel-level (20 µm) crack observation. Through analysis of virtual extensometers, we can quantify the evolution of crack opening under different propagation mechanisms: crack opening formation, extension, and stabilization.

Additionally, because the crack opening along the crack line was specially post-processed from image correlation measurements, this residual area can be used for automatic crack tracking. Such post-processing at each stage enables crack tracking over time.

In summary, this comprehensive study demonstrated the effectiveness of digital image stereo correlation in accurately tracking crack propagation during thermal fatigue loading. The ability to accurately detect and track existing cracks offers important prospects for improving component reliability in harsh thermal environments. No new cracks were found during the study. Future work includes continuing crack tracking over more cycles to analyze the shell's behavior when new cracks form. Furthermore, examining deformation thresholds where new cracks may appear is also of interest, enabling better understanding of failure mechanisms under repeated thermal loads. Finally, this proof of concept allowed the Liebherr team to better understand the possibilities and limitations of using image correlation systems and to commit themselves to future autonomous use.