Is this project an undergraduate, graduate, or faculty project?

Undergraduate

Project Type

group

Campus

Daytona Beach

Authors' Class Standing

Josephine - Senior Caleb - Senior Lola - Sophomore Olivia - Sophomore Collin - Junior

Lead Presenter's Name

Josephine Parker

Lead Presenter's College

DB College of Arts and Sciences

Faculty Mentor Name

Dr. Mihhail Berezovski

Abstract

This research explores recent advancements in solving Partial Differential Equations (PDEs) through a fusion of data-driven methods and Physics-Informed Neural Networks (PINNs). Traditional numerical techniques encounter challenges with complex systems in modeling diverse physical phenomena. The study focuses on applying PINNs to efficiently solve elliptic PDEs, crucial equations in scientific and engineering disciplines. Elliptic PDEs, known for their steady-state nature, pose challenges for traditional methods due to high computational costs over complex domains. The research utilizes a comprehensive dataset, synthetically generated for PINN training and accuracy validation. The study introduces a novel approach to constructing a PINN that not only efficiently solves elliptic PDEs but also quantifies prediction uncertainties. This dual capability is crucial for applications where decision-making relies on understanding the reliability of model outputs. The results demonstrate the potential of PINNs to revolutionize elliptic PDE solving, offering a fast, accurate, and physically consistent method that inherently accounts for uncertainty, advancing computational science.

Did this research project receive funding support (Spark, SURF, Research Abroad, Student Internal Grants, Collaborative, Climbing, or Ignite Grants) from the Office of Undergraduate Research?

No

Share

COinS
 

Data-driven Methods for Partial Differential Equations

This research explores recent advancements in solving Partial Differential Equations (PDEs) through a fusion of data-driven methods and Physics-Informed Neural Networks (PINNs). Traditional numerical techniques encounter challenges with complex systems in modeling diverse physical phenomena. The study focuses on applying PINNs to efficiently solve elliptic PDEs, crucial equations in scientific and engineering disciplines. Elliptic PDEs, known for their steady-state nature, pose challenges for traditional methods due to high computational costs over complex domains. The research utilizes a comprehensive dataset, synthetically generated for PINN training and accuracy validation. The study introduces a novel approach to constructing a PINN that not only efficiently solves elliptic PDEs but also quantifies prediction uncertainties. This dual capability is crucial for applications where decision-making relies on understanding the reliability of model outputs. The results demonstrate the potential of PINNs to revolutionize elliptic PDE solving, offering a fast, accurate, and physically consistent method that inherently accounts for uncertainty, advancing computational science.

 

To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.