State observer design for microchannel cooling system using extended Kalman filter

In this study, lumped physical models are developed for microchannel cooling systems, based on which extended Kalman filters are designed to estimate system states. Experimental validations indicate that the state estimate inaccuracies are within 3% under various operating conditions. Artificial white noise is added to system measurements to investigate the filtering effect of the extended Kalman filter, the fluctuation of temperature can be reduced by 50%. The filter is applied to a system with two microchannel heat exchangers experiencing unsynchronized heat loads. Numerical and experimental results prove that the application of an estimator is beneficial for improving thermal management efficiency by tracking the system states which are hard or impossible to be measured directly.