FEMtools

FEMtools Correlation Analysis

FEMtools Correlation Analysis is a toolbox to quantitatively and qualitatively compare 2 sets of analysis results data. Usually this is a FEA and a test database that are imported in the FEMtools database. However, the tools can be used for FEA-to-FEA and test-to-test correlation as well.

FEMtools Correlation Analysis is an extension of the FEMtools Framework and includes the following tools:

Spatial correlation - Compares location in space of nodes and measurement points and results in a table of node-point pairs and DOF pairs. This may require changing orientation and scaling of the models. This can be done manually or using automated tools.

Visual shape correlation – Visually compare analytical and test shapes (static displacement shapes, mode shapes and operational shapes).

Global shape correlation - Compares global analytical and test shapes. This information is used for shape pairing. Correlation tools include Modal Assurance Criterion (MAC) and mode shape orthogonality checks.

Local shape correlation (error localization) – Spatial comparison of analytical and test mode shapes. Results can be interpreted to localize modeling errors.

Shape pairing - Creates a table of shape pairs (static, modal or dynamic).

FRF pairing - Creates a table of FRF pairs.

FRF correlation – Computes shape and amplitude correlation functions for a set of FRF pairs as function of frequency.

Correlation coefficients - Calculates values of error functions from the selection of responses.

FEMtools Correlation Analysis is used for FE model validation, design of optimal test conditions, evaluate different modeling strategies, identification of modeling errors, damage detection, ...

Results from correlation analysis are used as reference for model validation and updating. Another application is to provide the analyst with information that can only be measured. An example is modal damping, used in modal superposition methods. Modal damping can be obtained experimentally and applied to the analytical mode shape that, using correlation analysis, was found to best match the experimental one.

Unlike global correlation analysis, spatial methods can be used to identify areas of better or poorer correlation, which when linked to structural information, can be interpreted in terms of 'modeling error'. Depending on how these tools are used, the results help with the selection of updating variables (parameters), or are used to assess structural damage.