Session: 4.1.4 - Renewable Energy Systems

Paper Number: 101973

101973 - Urban Airflow Analysis Using Reduced-Order Modeling 

The objective of this work is to examine publicly sourced weather data with multiple reduced order modeling techniques.  Namely, proper orthogonal decomposition (POD), spectral proper orthogonal decomposition (SPOD), and dynamic mode decomposition (DMD).  These data-driven models are able to extrapolate the dominant energies and frequencies embedded in turbulent flow that are otherwise not easily identified.  Computations of each method are based on the singular value decomposition (SVD) and therefore exploits a small subset of high dimensional data to reveal the dominate coherent structures or frequencies.  Evaluating weather data in this manner can help identify weather patterns relevant to the analysis of wind power generation, owing to its intermittent nature.  Furthermore, the methods of POD and DMD have been well established in many areas of data-driven sciences, while SPOD is a relatively new development.  Therefore, evaluating this specific type of data set provides an opportunity to examine the differences between the methods mentioned and expand upon the knowledge of where these methods are best suited.  Data collected will be computed and analyzed using MATLAB software. 

The objective of this work is to examine publicly sourced weather data with multiple reduced order modeling techniques.  Namely, proper orthogonal decomposition (POD), spectral proper orthogonal decomposition (SPOD), and dynamic mode decomposition (DMD).  These data-driven models are able to extrapolate the dominant energies and frequencies embedded in turbulent flow that are otherwise not easily identified.  Computations of each method are based on the singular value decomposition (SVD) and therefore exploits a small subset of high dimensional data to reveal the dominate coherent structures or frequencies.  Evaluating weather data in this manner can help identify weather patterns relevant to the analysis of wind power generation, owing to its intermittent nature.  Furthermore, the methods of POD and DMD have been well established in many areas of data-driven sciences, while SPOD is a relatively new development.  Therefore, evaluating this specific type of data set provides an opportunity to examine the differences between the methods mentioned and expand upon the knowledge of where these methods are best suited.  Data collected will be computed and analyzed using MATLAB software. 

The objective of this work is to examine publicly sourced weather data with multiple reduced order modeling techniques.  Namely, proper orthogonal decomposition (POD), spectral proper orthogonal decomposition (SPOD), and dynamic mode decomposition (DMD).  These data-driven models are able to extrapolate the dominant energies and frequencies embedded in turbulent flow that are otherwise not easily identified.  Computations of each method are based on the singular value decomposition (SVD) and therefore exploits a small subset of high dimensional data to reveal the dominate coherent structures or frequencies.  Evaluating weather data in this manner can help identify weather patterns relevant to the analysis of wind power generation, owing to its intermittent nature.  Furthermore, the methods of POD and DMD have been well established in many areas of data-driven sciences, while SPOD is a relatively new development.  Therefore, evaluating this specific type of data set provides an opportunity to examine the differences between the methods mentioned and expand upon the knowledge of where these methods are best suited.  Data collected will be computed and analyzed using MATLAB software. 

Presenting Author: Brad Warga Cleveland State University

Presenting Author Biography: Brad is a graduate research assistant in the ME program at CSU.