Improving Creation and Use of Material Models Derived from Noisy Data

This training module focuses on techniques to clean-up messy nonlinear material measurements as well as improve the post-processing of models that contain failure. Today’s advanced nonlinear simulation models are used to simulate a large range of materials and behaviors that are much more challenging to measure and characterize. Often multiple nonlinear data curves that contain non-ideal characteristics (noise, toe-in start-up, large spread of behavior), are measured, and then this data needs to be converted into a representative curve for material law creation. In the case of modeling failure, excessive noise in transient simulations is also a common problem. This class focuses on DSP techniques and other techniques to properly handle such cases, ultimately resulting in significantly improved results. This training is recommended those taking such measurements or working with models that utilize such challenging data.

Course Objectives & Benefits

• Learn how to apply DSP techniques and other approaches to clean-up messy stress/strain (or force/displacement) data curves
• Learn how to objectively judge outlier curves using measurements of merit
• Learn how to compute a proper average curve from multiple nonlinear curves
• Explore how to significantly improve interpretations of transient outputs for models containing failure

Course Outline

• Turning Noisy and Messy Measurements into Useful FEA Material Data
• Smoothing Noisy Measurements Pushing the Noise-Floor of Testing Systems
• Removing Start-Up Data Curve Distortions Via Trimming, Curve-Fitting and Extrapolation
• Assessing Outlier Curves via Measures of Merit
• Averaging Multiple Nonlinear Stress/Strain Curves
• Improving the Interpretations of Models Containing Failure
• T-bracket Failure Example
• Workshops
• Cleaning and Processing Messy Tension Data
• Analyzing Failure and Crack Propagation

Course Duration

• One session (3 hours total)

Prerequisites

• Some previous exposure to material measurements & characterization and/or nonlinear FEA simulations is beneficial, but not required
• Basic understanding of digital signal processing similar to that taught in DSP Essentials module is recommended

Training Delivery Options

• Public Web Training via Zoom
• Cost: $280
• Request a customized course for your company/organization.
• Can be delivered LIVE over the web or in-person at your site.
• You can customize content, including having use utilize one or more of YOUR datasets in workshop examples.
• Click here to request a quote.