Data and Code underlying the publication: Accelerated and data-efficient flow prediction in stirred tanks via physics-informed learning

This project introduces a data-efficient framework for accelerating fluid flow predictions in stirred tanks using physics-informed neural networks. By leveraging a custom-generated dataset of steady-state RANS ($k$–$\omega$) simulations across varying liquid heights and stirring rates, the models predict high-fidelity velocity, pressure, and turbulence fields. Our work provides a systematic evaluation of the data-accuracy trade-off, demonstrating that the integration of PDE constraints improves prediction reliability and mixing estimates, particularly in low-data regimes. This repository provides the complete implementation and controlled dataset to support further research into coupling learned neural fields with traditional CFD solvers for accelerated industrial design.