 | |  | |  | |  | |  | |  | |  | |  | |  |
Modeling Philosophy
All models start from simple hand built schematic models. The simple model has as few nodes as
possible in order to capture important features and gain insight. Complexity is added to the models until the
desired refinement and accuracy is obtained. The final model may have only a dozen nodes or it may have,
in the case of some finite element models, 100’s of thousands of nodes. The goal is to reach answers
quickly in a cost-effective manner while having assurance of the validity of the results. A series of simple
models gives one assurance that the answer is right. An analyst who plunges into a large model with all the
incumbent complexity tends to become overwhelmed by the logistics and manipulations and lose sight of
what are reasonable results. A legacy of simple models provides a reality check.
The level of complexity required for a particular model is determined by the desired results. If, for example,
one needs to size a fan for an electronics box, modeling individual components on each board is
unnecessary. Overall board temperatures and powers will suffice, however it maybe necessary to know
component temperatures at a later stage in the design process. The analysis is taken to a level that
provides an answer that is sufficient, this reduces time and cost.
Engineering judgment must be made in each case to decide what level of complexity is required at each
step in the design process. In most cases modeling complexity and refinement proceeds as the design
becomes finalized. A possible exception occurs when CAD geometry is used. Often the thermal analyst is
handed finalized CAD geometry that is overly detailed for thermal analysis. One can simplify the geometry
(this requires special skills) or one can create new geometry. However, it may be cost effective to take the
geometry as is, mesh it, and perform the analysis. Simplification will almost always be required when
radiation is involved. Current Monte Carlo radiation programs take unduly long when a large number of
surfaces are involved.
At PEI you can be assured that your project will be modeled in a logical progression that will give clarity
and useable results at a cost effective price.
possible in order to capture important features and gain insight. Complexity is added to the models until the
desired refinement and accuracy is obtained. The final model may have only a dozen nodes or it may have,
in the case of some finite element models, 100’s of thousands of nodes. The goal is to reach answers
quickly in a cost-effective manner while having assurance of the validity of the results. A series of simple
models gives one assurance that the answer is right. An analyst who plunges into a large model with all the
incumbent complexity tends to become overwhelmed by the logistics and manipulations and lose sight of
what are reasonable results. A legacy of simple models provides a reality check.
The level of complexity required for a particular model is determined by the desired results. If, for example,
one needs to size a fan for an electronics box, modeling individual components on each board is
unnecessary. Overall board temperatures and powers will suffice, however it maybe necessary to know
component temperatures at a later stage in the design process. The analysis is taken to a level that
provides an answer that is sufficient, this reduces time and cost.
Engineering judgment must be made in each case to decide what level of complexity is required at each
step in the design process. In most cases modeling complexity and refinement proceeds as the design
becomes finalized. A possible exception occurs when CAD geometry is used. Often the thermal analyst is
handed finalized CAD geometry that is overly detailed for thermal analysis. One can simplify the geometry
(this requires special skills) or one can create new geometry. However, it may be cost effective to take the
geometry as is, mesh it, and perform the analysis. Simplification will almost always be required when
radiation is involved. Current Monte Carlo radiation programs take unduly long when a large number of
surfaces are involved.
At PEI you can be assured that your project will be modeled in a logical progression that will give clarity
and useable results at a cost effective price.
Quality Assurance
All models built by PEI undergo rigorous quality checking:
- Each model is peer reviewed by one or more consultant using PEI quality procedures.
Reports
- Introduction - The report will contain an introduction explaining the nature of the devices and
processes being modeled and the goals of the consulting project. It will also outline the analysis
methods used. - Body - The body of the report will contain elaboration of devices and processes and modeling details.
It will state assumptions, boundary conditions, materials, etc. Results will be stated here with any
graphs and post-processing pictures. All assumptions are enumerated and caveats are discussed. - Conclusion and Recommendations - General conclusions from analysis results as well as
recommendations for further analysis or design changes as required. - Appendices – material that is important but need not be included in the body of the report such as
custom code used in the model or derivation of formulas. - References - References to textbooks, papers, other reports.