K-Suit
Physics-driven AI Tools
for Material Reliability

Karax LLC offers an array of computational tools to meet your predictive modeling needs. K-Suite, our modular computational framework, can help guide your materials decisions with predictive tools for quantifying the effects of various damage processes on the mechanical behavior of polymeric components. K-Suite combines physics-based constitutive models and a machine learning engine to deliver reliable estimates for remaining useful life and survivability prospects of polymeric components subject to one-time, catastrophic events, in addition to real-time prediction of mechanical behavior from accumulated radiation, hydrolytic, and thermo-oxidative damage.  Our modeling suite is comprised of four main modules:

1. K-Load: is a physically-based predictive tool for modeling the mechanical behavior of polymeric components that have been subjected to environmental damage. The modular formulation allows seamless integration of multiple damage sources: radiation, thermal oxidation, and hydrolysis, for example, that can be incorporated into an FE model to assess the effects of individual damage sources on the structural integrity of the polymeric component. Alternatively, the damage sources can be coupled in various combinations to provide a synergistic understanding of environmental damage on the net mechanical behavior.

2. K-Extreme: is a machine-learned engine that provides conservative estimates of survivability by calculating the minimum damage capacity needed to survive single-event effects (SEE) such as: loss-of-coolant accidents in nuclear reactors, solar flares for space applications and nuclear detonation simulations for strategic defense applications.

3. K-Fail: Allows for the quantification of the remaining useful life of a polymeric component from the i) synergistic effects of radiation and thermal oxidation during service and ii) strength loss due to one-time extreme events. K-Extreme also takes into account the available data on different SEE and uses them to predict survivability in case of a new SEE.

4. K-Sense: K-Sense is a machined-learned statistical model developed to quantify the correlation between damage and optical parameters. In single-stressor aging condition, e.g. accelerated thermal aging, K-Sense can be as simple as a basic linear model with one predictor (stressor) and one response. In field applications with multiple stressors and multiple parameters in, K-Sense should be considered as a multi-variable mixed-effects regression model that can account for differences between material grades, exposure types, multi-stressor parameters and interactions among these factors.