An Experimental Assessment of Numerical Predictive Accuracy for Electronic Component Heat Transfer in Forced Convection—Part I: Experimental Methods and Numerical Modeling

Abstract

Numerical predictive accuracy is assessed for component-printed circuit board (PCB) heat transfer in forced convection using a computational fluid dynamics (CFD) software for the thermal analysis of electronic equipment. This is achieved by comparing numerical predictions with experimental benchmark data for three different components, mounted individually on single-component PCBs, and collectively on a multi-component PCB. Benchmark criteria are based on measured steady-state component junction temperature and component-PCB surface temperature profiles. The benchmark strategy applied permits the impact of both aerodynamic conditions and component thermal interaction on predictive accuracy to be quantified. In the accompanying Part II of this paper, the experimental measurements are reported and numerical predictive accuracy is assessed.

An Experimental Assessment of Computational Fluid Dynamics Predictive Accuracy for Electronic Component Operational Temperature

“…Addressing this weakness, Rodgers et al [35,164] [169][170][171][172][173][174]. These studies have found CTM accuracy to be generally within 10% of the detailed component model predictions for both junction temperature and heat flows from the package external surfaces.…”

“…The two-equation standard k-s model is subsequently used for refined sub-system analysis, such as for PCB assemblies. However, this strategy has been shown to fail to yield accurate predictions o f component operating temperature in air-cooled PCBs [35], as neither a laminar or standard high-Reynolds number k-s flow model are specific for such flows.…”

“…No defect was found at other interfaces. Device A was free o f defect, and was therefore considered as a more realistic reference to assess predictive accuracy [35].…”

“…Numerical models for the S 08 and PQFP 208 test configurations were previously presented by Lohan et al [149,225] and Rodgers et al [35] respectively, incorporating detailed component models. Lohan et al analysed component heat transfer in both free convection [149] and 1 m/s airflow [225], and these analyses are extended here to both 2 and 5 m/s airflows.…”

“…While there is no consensus in the literature on maximum permissible prediction error for component junction temperature, the following guidelines have been proposed. Whereas in the early to intermediate design phase, an accuracy o f ±10°C or 20% to ±5°C or 10% is generally sufficient [35,36], for temperature predictions to be applied as boundary conditions in component Physics-ofFailure (PoF) based reliability prediction methods [10,[37][38][39][40], which rely on the accurate determination of electronic component operating parameters, a criterion o f ±3°C or 5%…”