Fast Simulation of steady-state temperature distributions in electronic components using multidimensional model reduction

Celo, D.; Gunupudi, P.; Khazaka, Roni; Walkey, D.J.; Smy, T.; Nakhla, M.

doi:10.1109/tcapt.2004.838877

Cited by 16 publications

(15 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…-B = of (7) is "equivalent" to the Laplace variable s. Hence, the first idea explored by several groups [10] - [13] and based on generalization of moment matching, was to make multivariate expansions of (8) with respect to the Laplace variable s and parameters p i which are thus simultaneously preserved in symbolic form.…”

Section: Parametric Model Order Reductionmentioning

confidence: 99%

See 1 more Smart Citation

Efficient solution of inverse thermal problem via parametric model order reduction

Bechtold

Hohlfeld

Rudnyi³

2009

EuroSimE 2009 - 10th International Conference on Thermal, Mechanical and Multi-Physics Simulation and Experiments in Microelect

View full text Add to dashboard Cite

In this paper we present a novel approach to determine material thermal properties of thin film materials. As a testcase we have chosen silicon nitride employed in a micro-hotplate structure. We use ANSYS to build a threedimensional finite element (FE) model of the test structure, which considers heat conduction and convection effects. It is further parameterized and subjected to parametric model order reduction (pMOR). The parameters of the reduced model are the unknown material thermal properties and the heat transfer coefficient, which are automatically adjusted within the optimization loop, so that the simulation results of the model are in agreement with the transient temperature measurements. The use of parameterized reduced order models within the optimization iterations speeds up the transient solution time by several orders of magnitude, while retaining almost the same precision as the full FE model.

show abstract

Section: Parametric Model Order Reductionmentioning

confidence: 99%

“…We have used the multivariate-moment-matching approach from [13] and were able to reduce the time for transient integration of the FE model by a factor of 100. A parametric reduced model has only 117 degrees of freedom (DOF) but still provides high accuracy (see Fig.…”

Section: Parametric Model Order Reductionmentioning

confidence: 99%

Efficient solution of inverse thermal problem via parametric model order reduction

Bechtold

Hohlfeld

Rudnyi³

2009

EuroSimE 2009 - 10th International Conference on Thermal, Mechanical and Multi-Physics Simulation and Experiments in Microelect

View full text Add to dashboard Cite

show abstract

“…We have used the multivariate-moment-matching approach from [9] and were able to reduce the time for transient integration of the FE model by a factor of 330. Parametric reduced models have only 60 degrees of freedom (DOF) but still provide high accuracy (see Figure 5).…”

Section: Parametric Model Order Reductionmentioning

confidence: 99%

Efficient extraction of thin film thermal properties via parametric model order reduction and optimization

Bechtold

Hohlfeld

Rudnyi³

2009

TRANSDUCERS 2009 - 2009 International Solid-State Sensors, Actuators and Microsystems Conference

View full text Add to dashboard Cite

In this paper we present a novel highly efficient approach to determine material properties from measurement results. We apply our method to thermal properties of thin-film multilayers with three different materials, amorphous silicon, silicon-nitride and silicon-oxide. The individual material properties are identified by solving an optimization problem. For this purpose, we build a parameterized reduced-order model from a finite element (FE) model and fit it to the measurement results. The use of parameterized reduced order models within the optimization iterations speeds up the transient solution time by several orders of magnitude, while retaining almost the same precision as the full-scale FE model.

show abstract

“…By replacing a fully detailed thermal component model with a simplified one, simulation times can be greatly reduced. Depending on the procedure adopted, a replacement model can be in the form of a simple resistive network, such as a boundary condition independent compact thermal model [1]- [8], or in the form of a mathematical (reduced) model [9]- [11]. A significant amount of work has been done in developing methodologies to generate and construct these simplified models; however, their application in modeling a board-level environment has seen limited study [12].…”

Section: Introductionmentioning

confidence: 99%

“…In general, any linear component model can be described by the attachment (7), including boundary condition independent models created by a parametrized model reduction technique [11] or the use of an optimized compact model [22]- [24]. The use of these techniques would, however, introduce error into the component model itself which is undesirable for this work as the error due to the port configuration could not be isolated.…”

Section: Introductionmentioning

confidence: 99%

A General Method for the Connection of a Component Thermal Model to a Board

Gao

Celo

Walkey

et al. 2006

IEEE Trans. Adv. Packag.

Self Cite

View full text Add to dashboard Cite

In this paper, a generalized method for the connection of a thermal component model in board-and system-level thermal simulations is presented. The method allows for the definition of uniform heat flow connections as well as the standard uniform temperature interface regions. The use of uniform heat flow ports will be shown to better handle cases where large temperature gradients are present in the base model. The two methods of connecting the component model will be evaluated using two different models. First, a simple example will be presented to illustrate the nonphysical behavior introduced by the use of uniform temperature connections. Second, a model of an electronic package will be used to evaluate the relative merits of the two connection methods with respect to board thermal conductivity and boundary conditions present on the board and the package. It will be shown that the results from use of uniform heat flow connections are generally better than from use of uniform temperature regions with respect to predicting junction and board temperatures.

show abstract

Fast Simulation of steady-state temperature distributions in electronic components using multidimensional model reduction

Cited by 16 publications

References 17 publications

Efficient solution of inverse thermal problem via parametric model order reduction

Efficient solution of inverse thermal problem via parametric model order reduction

Efficient extraction of thin film thermal properties via parametric model order reduction and optimization

A General Method for the Connection of a Component Thermal Model to a Board

Contact Info

Product

Resources

About