Improvements on the thermal analysis for system perturbed by micro-thermal fluctuations are presented: the method applies to any kind of (small) perturbation, in particular to the random ones.
Opposite to time domain conventional transient analysis, this method answers the need for frequency domain thermal analysis dictated by the newest scientific missions, with tight temperature stability requirements (expressed in the frequency domain).
The small temperature fluctuations allow for assuming any thermal systems a linear one; hence linear system theory holds, and powerful tools to calculate key parameters like frequency response can be successfully employed.
MIMO (Multi-Input-Multi-Output) systems theory is applied, the inputs being perturbations to the thermal system (boundary temperatures oscillations and power sources ripple of any shape: pulse, step, periodic, random, …), while the outputs are the temperatures of the sensible parts.
Main advantage of the presented method is that, this way, there is no longer need to stimulate the system with micro-perturbations in transient analysis, thus avoiding:
Moreover, this allows to have a synthetic definition how the effects of disturbances are propagated through the model, suitable for a graphical representation of the way the perturbation is damped by the heat transfer differential equations within the thermal model.