
The basic principle behind thermography is quite simple. An IR camera records the surface temperature of a sample as it is heated with a light pulse lasting only a few milliseconds. The sample cools as heat diffuses from the surface to the interior. However, internal features or flaws affect the diffusion process. For example, a subsurface void obstructs the flow of heat and causes a transient temperature increase to occur in the vicinity of the flaw. Additionally, conditions such as porosity affect the thermal properties of the sample and the rate at which heat diffuses. Although these changes can be subtle, dedicated software can analyze the data from the IR camera and isolate likely defects. The software can also measure part or coating thickness, flaw depth and size, and thermal diffusivity.

Many of the advances in thermography can be attributed in part to the evolution of the underlying technologies, including the IR camera, PC and communications protocols, and hardware that allows real-time transfer of data between camera and computer. These enabling advances have allowed researchers to acquire and study high-precision data, and they've facilitated the development of advanced models for the heat transfer process in complex material systems and geometries.
Signal processing has played a major role in this evolution. In early thermography systems, data from the IR camera was treated similar to a movie, which the operator viewed in slow motion to visually identify anomalous features that might indicate subsurface defects.
In modern systems, each camera pixel is treated as an independent entity that represents the behavior of a point on the sample. In one patented technique,* the time-dependent temperature response of each pixel to thermal excitation is expressed as an equation that represents the degree to which that point on the sample conforms to "ideal" behavior. In effect, the entire image sequence is reduced to a set of equations in which time varying noise from the camera or the immediate environment is eliminated.
The sequence, which might comprise hundreds-or even thousands-of discrete frames, can be reduced to a set of noise-free equations, enabling rapid mathematical manipulation of the sequence and advanced analysis capabilities. Put simply, the pixel signal acquired by the IR camera is processed mathematically to enhance weak signals that would normally be undetectable.
*Thermographic Signal Reconstruction® (TSR), developed by Thermal Wave Imaging, Inc., Ferndale, Mich.
After the accident, NASA undertook a comprehensive survey of NDE technologies for the inspection of the RCC leading edge. As a result, an infrared NDE system* was selected as the primary inspection system. The system offers fast, non-contact, wide-area inspection of flat or curved structures, and it can measure the length, depth and area of subsurface defects. Additionally, automation features allow the system to communicate with other programs and devices, and the systems are optimized for ease of use and for handling the large volume of data generated by each inspection.
*EcoTherm®, available from Thermal Wave Imaging, Inc., Ferndale, Mich.

Figure 2 illustrates the difference between an optical image (left) and a thickness map generated through infrared NDE (right) for a TBC combustor liner. Measurement accuracy of modern systems is on the order of 1 mil. Once the blade is in service, thermography is used during periodic inspections to identify thin TBC areas or delaminated regions where spalling is likely to occur (see Figure 3).

For more information about thermographic NDE, contact Thermal Wave Imaging, Inc., 845 Livernois St., Ferndale, MI 48220; (248) 414-3730; fax (248) 414-3764; e-mail sales@thermalwave.com ; http://www.thermalwave.com .