Understanding the model behavior of a thermal-mechanical simulation using ANSYS and SoS
November 7, 2018 | 9-10 am CET and 9-10 am PST
Machine learning, sensitivity analysis and statistical analysis help to understand the behavior of CAE models, the relationships between parameters, the variation patterns of the structural performance and more. SoS helps here to do such analysis on 3D FEM models visualizing the results for many different locations and also analyzing the relations between the different locations in space.
The webinar illustrates the data analysis with SoS using an electrical circuit model in ANSYS Mechanical. Results can be directly exported from ANSYS Mechanical to SoS using the new SoS plugin for ANSYS. Based on a Design of Experiments the FEM results data are visualized and analyzed in SoS. The questions to be answered are:
- Where are the largest variations of temperature and mechanical stresses?
- Where are the largest values with a specific probability?
- What are the most sensitive input parameters for individual locations?
- How can one create a nonlinear real-time approximation model for the complete FEM solution model?
- What are the relations between response fields like temperature and stress?
Signal analyses with optiSLang and Statistics on Structures
In CAE – based product development, computed results are often not only scalar numbers, but kind of XY-data. Examples are dynamic signals, spectra, sensor data, or stress-strain-curves. A typical application is the calibration of material law parameters to given measurements.
The functionality of SoS required for the analysis of signals is implemented as a simple, single node in optiSLang. With that, the Field-Metamodel of Optimal Prognosis (F-MOP) is fitted to the signal data. The post-processing displays the F-CoP (Field-Coefficient of Prognosis) and sensitivities of input parameters versus the abscissa of the signal. Thus it is easy to see where in the range the response signal is well explainable and where it is noisy, and to assess which input has dominant influence where in the range of the signal. The F-MOP can be used for optimization in an equally simple way.
During the webinar, the capabilities of optiSLang and Statistics on Structures will be demonstrated by several examples, such as calibration of non-linear material parameters for a concrete wedge splitting test, and other industrial applications dealing with time series and spectra.
Understanding the properties of porous materials using GeoDict combined with optiSLang
Nonwovens, foams, ceramics, and other porous materials become increasingly used as technical materials with unique properties (e.g. lightweight, insulating materials, electrodes). Composite structures often combine several, sometimes contradictory characteristics. Thus, determining an “optimal material“ is a challenging task. The properties assessed by GeoDict simulations are extremely complex and almost impossible to measure in a laboratory. The combination of GeoDict and optiSLang enables an easy and efficient way to create a Design of Experiment. Metamodels of Optimal Prognosis identifies the important input parameter for an optimization of the material properties. Due to the leading-edge Pareto optimization technology of optiSLang, this knowledge can be used to optimize the materials even if there are conflicting objectives (e.g. light and stable).
As an example, the analysis of the particle deposition in a fluid filter made out of fiber material will be shown. These filter will be optimized regarding the efficiency and the pressure drop. With these methodology new Designs can be found, which could be used in a further simulation in one of the ANSYS CFD Tools.
Optimal cooling and quality proof of electronic components using optiSLang and FloEFD
A linkage between FloEFD and optiSLang has been created to provide the customer with a leading-edge optimization procedure for cooling simulations of electronic components in an embedded industrial computer. Using sensitivity analysis, meta-modeling techniques and optimization on Response Surfaces will enable you to design electronic modules, which operate at optimal temperatures. The parametrized geometry is generated in FloEFD and the workflow is set up in optiSLang. During the webinar, the sensitivity analysis and a geometry optimization of a single board computer with case will be presented. Furthermore, the quality of the single board computer will be verified with a tolerance analysis.
Optimization and design of turbo machines using ANSYS optiSLang and CFturbo
Optimization in turbomachinery is a challenging task due to demanding numerical CFD simulations. Using sensitivity analysis, meta-modeling techniques and optimization on Response Surfaces will enable you to design turbo machines with acceptable numerical effort. The efficiency of the approach becomes even more relevant when several operating points should be taken into account. The basis for such an analysis is a well parametrized geometry and an automated workflow which will be both presented in this webinar. The parametrized geometry is provided by CFturbo and the workflow is set up in ANSYS optiSLang using different Solvers like PumpLinx and CFX in ANSYS Workbench. As an example, the analysis of a pump and a radial turbo compressor will be presented.
Optimal layout of electric machines using Motor-CAD and ANSYS optiSLang
A linkage between Motor-CAD and optiSLang has been created to provide a world leading optimization procedure for the design of rotating electric machines. Typically, optimization procedures for electric machines focus on evaluating each candidate design at a single operating condition considering only electromagnetic effects, this causes many challenges and often leads to a sub-optimal design.
The Motor-CAD-optiSLang workflow allows engineers to evaluate candidate designs across the full operating speed range, considering both peak and continuous torque/power characteristics, using state-of-the-art electromagnetic and thermal modeling techniques. In addition, the efficiency and performance of considered designs regarding drive cycles can be optimized. Multi-core analysis is used to provide a significant speed improvement to the solution, while the cutting-edge design space analysis techniques contained within optiSLang enable the optimum design to be found quickly and reliably.
Dimensioning and quality proof of a connector using ANSYS optiSLang
Connectors are used in a wide range of industrial products: eMobility, energy automation, automotive industry, etc. In order to design such connectors, in a first step, it is often necessary to calibrate the virtual material model to experimental measurements. A design improvement can be achieved by optimizing the geometry for a required insertion and pull-out process. Finally, the quality of the connector has to be proven. With the help of ANSYS optiSLang, the engineer can easily solve these challenges. Learn how ANSYS optiSLang helps you to perform a fully automatic calibration and optimization of a connector with respect to challenging environmental conditions. You will also learn how to set up and perform a tolerance analysis with ANSYS optiSLang.
Calibration of a Rocky DEM Simulation with ANSYS optiSLang
The automated calibration of a numerical simulation with experimental data can help to improve the credibility of the simulation results with all stakeholders. In this webinar, the setup of such an automated analysis is demonstrated for the angle-of-repose and drawn-down-angle simulated with Rocky DEM. It also shows how easy it is to use sensitivity analysis and meta-modeling techniques to find the best fit of particle specific data with the experimental data. The webinar demonstrates how to apply Rocky DEM inside optiSLang, however, you can apply the presented workflow with any FEM, CFD or electric magnetic simulation model.
Machine tool optimization with ANSYS optiSLang
The automated optimization of an engineering structure or system can economize material costs and planning time. During this webinar, the setup of such an automated analysis is demonstrated for a machine tool part simulated with ANSYS Mechanical. Using a sensitivity analysis and meta-modeling techniques, the best compromise between the structural mass and the deformations under several working conditions can be found very quickly. In this example, the optimization procedure using ANSYS optiSLang within the ANSYS Workbench could significantly reduce the mass with respect to the manual engineering design. You also learn how to apply easily the presented workflow to your FEM, CFD or electric magnetic simulation model.