Stress Imaging works because solid materials will exhibit a temperature change when subjected to load. These temperature changes are a direct measure of the stress on the material but are too small to measure with traditional techniques. See How Does it Work page to understand how we are able to tease out these small temperature responses and generate high-fidelity stress images.
Frequently Asked Questions
What is stress sensing, and why is it important?
Also known as Thermoelastic Stress Analysis (TSA), we use the term Stress Imaging to describe more simply what our technology does. As the name suggests, this is an imaging technique, which employs an infrared, thermal imaging sensor and advanced image processing to visualise stress on an object under dynamic load.
Stress Imaging is important as it allows for comprehensive validation of digital stress simulations (Finite Element Models). This allows for improved fine-tuning of the simulations, leading to more efficient product development cycles.
Stress Imaging can also be used for many other applications, such as monitoring stress on operating plant/equipment, tracking and characterising fatigue cracks and conducting quality control on structurally critical parts.
How accurate are the stress images?
We get asked this question a lot as people wonder whether our images rely on simulation data or calibration data to get the images to look the way they do. No external data is required as our cameras simply measure the thermoelastic response of the material under stress, which is a small reversible change in temperature that occurs in a solid when it experiences compression or tension see How Does it Work. This temperature change is proportional to the change in bulk stress and therefore can be used to accurately infer stress levels – a process known as Thermoelastic Stress Analysis (TSA) or Stress Imaging. Because the relationship between temperature and stress is well known, and because we can measure temperature to exceedingly low levels, Stress Imaging can be very accurate.
How sensitive are the stress images?
The sensitivity comes down to two key factors: the material properties of the subject [See How Does it Work] and the time spent observing it. Generally, image quality improves as a function of the square-root of the observation time, e.g. if the processing time is quadrupled, the signal to noise ratio of the measurement improves by a factor of two. Our cameras can sustain this rate of improvement for a very long time, producing stress imagery that can be easily mistaken for a numerical simulation (but without all the uncertainty!)
How did you validate that this works?
TSA has a solid theoretical foundation that provides a ground truth for checking that experimental measurements are correct. Experimental validation of 1MILLIKELVIN products is undertaken using specially designed test coupons for which theoretical solutions are readily available. We also compare measurements to finite element predictions but only for relatively simple geometries where the uncertainties of Finite Element Analysis (FEA) can be minimised.
Why use a thermal (uncooled) detector instead of a more sensitive photonic (cooled) detector?
From the inception of TSA and until recently the practical implementation has relied exclusively on cooled detectors. This changed rather quickly when it was realised that uncooled detectors have some compelling advantages for stress imaging applications. While a cooled detector easily outperforms an uncooled detector in a raw imaging application (just compare the Noise Equivalent Delta Temperatures (NEDT)!), there is a quirk in the detector noise morphology that does not translate to better stress imagery in practice. Recognising this, and the significantly higher cost, bulk and shorter life of cooled detectors, 1MILLIKELVIN saw an opportunity to make TSA significantly more practical, versatile, affordable and ultimately more sensitive.
Does this work on complex geometries?
Yes. This is one of the key advantages of Stress Imaging over other stress/strain imaging methods. Providing the temperature at a location of interest can be measured, so too can the stress, whether that’s in a geometrically complex transition zone, i.e. a fillet radius in a beam, or the flat face of a tensile coupon.
Why do I need a dynamic load?
Because temperature is subject to the laws of heat diffusion, the stress-induced temperature gradients caused by a static load would quickly dissipate and be unmeasurable. Modulating the load, i.e. making it dynamic, allows these gradients to persist.
How fast do I need to apply the load?
The rate at which the load needs to change to allow gradients to persist depends mainly on two factors: the severity of the stress gradient and the thermal diffusivity of the material. A relatively low-diffusivity material, e.g. Titanium or steel, containing a relatively minor stress gradient, e.g. a beam subject to pure bending, would not require as high a loading rate as a higher diffusivity material, e.g. aluminium, containing a more severe stress gradient, e.g. a small circular hole.
What type of stress are you measuring?
A thermoelastic response is triggered by the hydrostatic (or volumetric) stress component, known as the bulk stress or first stress invariant, which is simply the sum of the principal stresses, this is what we are measuring.
In contrast, a pure shear stress state does not generate a thermoelastic response. Consequently, Stress Imaging cannot detect shear stresses. When comparing Finite Element predictions with Stress Imaging measurements, make sure to set the Finite Element output as the sum of the principal stress components.
Can this be used in very hot or cold environments?
Yes. The thermoelastic response increases with absolute temperature. For example, all else being equal, the thermoelastic response at a subject temperature of 50 deg C will be approximately 18% higher than at a subject temperature of 0 deg C.
What units are the results given in?
By default, imagery produced by the system is presented in a temperature related unit. Should calibration to stress or strain be required, this can be easily applied during image acquisition or afterwards.
How do you calibrate the images from temperature to strain/stress?
While it is possible to determine a calibration factor theoretically, the preferred approach is to use a strain measurement obtained from a biaxial or rosette strain gauge placed in an area with relatively low strain gradient.
Does this only work on some materials?
Stress Imaging is applicable to all solid materials. What about concrete, isn’t that a sold material?
What if the object I am looking at is made up of different materials?
Different materials have different thermoelastic coefficients, it is important that this is taken into account when comparing signals across the constituent materials. This is especially important when calibrating thermoelastic response measurements to strain or stress.
What surface preparation is required?
If the subject is unpainted or has a reflective surface, i.e. high gloss coating or bare metal, a thin coating of high emissivity (non-reflective) paint will be required. Often, an existing coating may have sufficiently high emissivity to be imaged without any additional preparation, i.e. most airframe primers have an excellent infrared emissivity and can be imaged without supplementary coating.