This study assesses a lumped parameter modelling strategy for the fast response, layer-resolved prediction of the thermal ield within the die of a selected power module device. A benchmark 1D lumped parameter model was initially deined, providing a simple description of the thermal ield through the identiication of a low number of lumped parameters but retaining information only on the junction temperature. A reined analytical model was therefore developed by adding the internal temperature nodes at the interfaces between the device layers. Thus, with a moderate increase of the lumped parameters, the model was able to predict the thermal behaviour of the inner layers of the die, which is the fundamental information required by reliability analysis. The set of model parameters were determined from the step-response thermal impedances of the power module, simulated by means of a numerical model purposely assessed under a boundary condition of forced liquid cooling. The lumped parameter identiication consisted in a curve itting procedure calculating the couples of thermal resistance and conductance. Both models were implemented in Matlab- Simulink, and simulated in order to predict the thermal response of the device to conditions of both step-wise and alternating current power inputs.
Lumped Parameter Modelling of Internal Temperature Dynamics in Multichip Power Modules
Alberto Fichera;Marta Lo Faro;Arturo Pagano
;Rosaria Volpe
2025-01-01
Abstract
This study assesses a lumped parameter modelling strategy for the fast response, layer-resolved prediction of the thermal ield within the die of a selected power module device. A benchmark 1D lumped parameter model was initially deined, providing a simple description of the thermal ield through the identiication of a low number of lumped parameters but retaining information only on the junction temperature. A reined analytical model was therefore developed by adding the internal temperature nodes at the interfaces between the device layers. Thus, with a moderate increase of the lumped parameters, the model was able to predict the thermal behaviour of the inner layers of the die, which is the fundamental information required by reliability analysis. The set of model parameters were determined from the step-response thermal impedances of the power module, simulated by means of a numerical model purposely assessed under a boundary condition of forced liquid cooling. The lumped parameter identiication consisted in a curve itting procedure calculating the couples of thermal resistance and conductance. Both models were implemented in Matlab- Simulink, and simulated in order to predict the thermal response of the device to conditions of both step-wise and alternating current power inputs.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.