%0 Computer Program %A Ghasemi Khourinia, Alireza %A Kok, Jim %D 2022 %T PARAMOUNT: parallel modal analysis of large datasets %U https://data.4tu.nl/articles/software/PARAMOUNT_parallel_modal_analysis_of_large_datasets/20089760/1 %R 10.4121/20089760.v1 %K proper orthogonal decomposition (POD) %K singular value decomposition (SVD) %K Spectral Analysis %K Parallel Processing %K Unsupervised Machine Learning %X

PARAMOUNT: parallel modal analysis of large datasets

PARAMOUNT is a python package developed at University of Twente to perform modal analysis of large numerical and experimental datasets. Brief video introduction into the theory and methodology is presented  here.

Features
 

- Distributed processing of data on local machines or clusters using Dask Distributed
- Reading CSV files in glob format from specified folders
- Extracting relevant columns from CSV files and writing Parquet database for each specified variable
- Distributed computation of Proper Orthogonal Decomposition (POD)
- Writing U, S and V matrices into Parquet database for further analysis
- Visualizing POD modes and coefficients using pyplot


Using  PARAMOUNT

Make sure to install the dependencies by running `pip install -r requirements.txt`

 

Refer to csv_example to see how to use PARAMOUNT to read CSV files, write the variables of interest into Parquet datasets and inspect the final datasets.

Refer to svd_example to see how to read Parquet datasets, compute the Singular Value Decomposition, and store the results in Parquet format.

To visualize the results you can simply read the U, S and V parquet files and your plotting tool of choice. Examples are provided in viz_example.

Author and Acknowledgements

This package is developed by Alireza Ghasemi (alireza.ghasemi@utwente.nl) at University of Twente under the MAGISTER (https://www.magister-itn.eu/) project. This project has received funding from the European Union’s Horizon 2020 research and innovation program under the Marie Skłodowska-Curie grant agreement No. 766264.

%I 4TU.ResearchData