Finding the Best Solutions for Powertrain, CFD and Solid Element-Based Assemblies

Guna Krishna

Guna Krishna

Vice President, SimLab Development at Altair
Guna Joined Altar in 2010, and currently is the Vice President of SimLab development. He has over 25 years of experience in CAE. Prior to Altair, Guna worked at PDA Engineering and MSC before he helped incorporate SimLab Corporation. His current role includes directing development and product specification for SimLab. He holds a doctoral degree in Engineering Mechanics from Iowa State University, Ames, Iowa, USA.
Guna Krishna

Multiple iterations can help realize a design that is optimal for a given problem. An optimal design can lower cost to manufacture, make it lighter, reduce noise, prevent overheating and more. Yet, multiple design iterations are more a wish, hindered by many factors, such as cost, lack of expertise, and the complexity of the process.

SimLab was developed with automation as the core for modeling and analysis of a design. It provides an environment where users can start with CAD, create an FE model, submit the job to a solver and interpret results. The steps are defined by templates and the process can be recorded.  Together, templates and process recordings can realize the goal of multiple design iterations.

The details that affect a design and the solver’s requirements for the FE model determine the modeling process and the cost. Typically, this is achieved by cleaning up the geometry to exclude noises, such as cracks, overlaps, sliver surfaces, edges, and unnecessary details. Then, a mesh is created and modified to achieve the desired quality and maintain associativity with geometry. These two steps are manual and become expensive as the model size becomes big.

The advanced meshing technology in SimLab eliminates the need for geometry cleanup. The mesh patterns, density, and quality of a mesh can be specified prior to its creation. This new approach eliminates the manual step and therefore becomes suitable for automation to iterate through multiple design iterations.

CAD assemblies are a collection of parts that interact with each other. SimLab can create a conformal mesh between the parts and define the interactions between them. The interactions can be problem specific. They can be a detailed contact definition between parts, capturing engineering details like press fit and friction coefficients, or they can be a simplified representation, such as a bolt represented by a spider and beam elements. The interaction between parts can be defined in a template and reused over many design iterations. Complex assemblies, such as engines, transmissions, gearboxes, motors, pumps, head lamps, cellphone, printers, circuit boards, and more can be modeled in SimLab.

In many cases design changes are localized. Regenerating a ready to run solver deck from scratch for each design iteration can be expensive. SimLab provides a mechanism to update the model locally without the need to regenerate it. This makes SimLab suitable for design studies of large, complex models, including legacy models.

SimLab is CAD and solver neutral. The modeling process is independent of the geometry source or the solver used. This is important in workplaces that use multiple solvers. SimLab reads CAD models created in Creo, Catia, UG and others based on Parasolid. SimLab can export the model to several solvers including OptiStruct, Abaqus, Nastran, ANSYS, PERMAS, AcuSolve, Fluent and more.

SimLab provides an environment to create structural and CFD models. To study the flow of hot gases through an exhaust manifold, the manifold with structural and thermal loads and solid material can be defined. Inside of the manifold, the gas flow, the temperature, and the fluid properties can be defined. The structural and the fluid problem can be solved independently or together capturing the interaction between the two solutions.

Thermal results on an exhaust manifold model

SimLab allows users to confirm if the results are close to the true answer. It relies on the principal that results converge as the mesh is refined. Instead of refining the mesh everywhere, SimLab can read the results and identify regions that requires refinement. SimLab, though an automated process, can then refine the mesh locally and re-run the analysis. This is repeated multiple times until the result between successive iteration is within a limit. This provides a mechanism that allows users to have confidence in the analysis they perform.

SimLab is both a pre- and post-processor for a solver. Multiple results file (structural and fluid results) can be imported and overlaid. In addition, the results from one analysis can be used to update the FE model, by locally refining the mesh, or use the results as boundary conditions for another analysis. For example, the results from a CFD analysis (e.g. temperature and heat transfer coefficient) can be assigned to a solid surface for durability analysis.

Optimization uses results from the current iteration to decide the parameters for the next iteration. SimLab automates the steps in one iteration while HyperStudy invokes SimLab to optimize a design. The interaction of SimLab and HyperStudy has been improved in a way that it enables optimization runs by simply dropping a SimLab script file into the HyperStudy user interface. This allows automatic import of input and output variables to HyperStudy, significantly simplifying the study setup process for the user.

SimLab is part of the HyperWorks package and provides the tools for designing powertrain components like engine assemblies, gear boxes, transmission systems, and more. It is also used for solid modeling in other products like cell phones, motors, and volumes that define a fluid region as in CFD applications. SimLab interfaces with Hyper Works solvers like OptiStuct, AcuSolve, and nanoFluidX. It can also use PBS Works to submit the jobs to a cluster and HyperStudy for design optimization.

Detailed modeling of crankshafts, welds and bearings is simplified in SimLab. For those who are not familiar with detailed modeling of features like welds, bearings, crank shafts, bolts, etc., can benefit from SimLab as the modeling practices is built into the simplified UI.

Template based definition of bearings

SimLab is a highly automated pre- and post-processor for structural and CFD meshes, which has its roots in the powertrain domain. Reading native CAD geometry without the need of geometry simplification and cleanup, the ability to define mesh patterns for model regions and features, and the efficient way of creating assemblies via contact or bolted connections, helps with huge timesaving’s in model setup. Templates for creation of gaskets, bearings, bolts and more help further to quickly find the best solution in a product development cycle.

SimLab provides efficient parameter based weld creation

Want to learn more? The Altair Learning Center can help you quickly learn how to use SimLab: http://web2.altairhyperworks.com/simlab-learning-center

Watch an on-demand webinar on the latest SimLab functionality: https://altairhyperworks.com/ResourceLibraryDetail.aspx?title=HyperWorks+2017:+Pre-Processing+with+SimLab

Guna Krishna

About Guna Krishna

Guna Joined Altar in 2010, and currently is the Vice President of SimLab development. He has over 25 years of experience in CAE. Prior to Altair, Guna worked at PDA Engineering and MSC before he helped incorporate SimLab Corporation. His current role includes directing development and product specification for SimLab. He holds a doctoral degree in Engineering Mechanics from Iowa State University, Ames, Iowa, USA.