pyforce-1.0.0: Python Framework for data-driven model Order Reduction of multi-physiCs problEms

Stefano Riva, Yantao Luo, Carolina Introini, Antonio Cammi

arXiv:2605.18082·cs.LG·Published 2026-05-18

pyforce is a Python package implementing Data-Driven Reduced Order Modelling techniques for applications to multi-physics problems, mainly set in the Nuclear Engineering world. The package is part of the ROSE (Reduced Order modelling with data-driven techniques for multi-phySics problEms): mathematical algorithms aimed at reducing the complexity of multi-physics models (for nuclear reactors applications), at searching for optimal sensor positions and at integrating real measures to improve the knowledge on the physical systems. With respect to the previous original implementation based on dolfinx package (v0.6.0), version 1.0.0 of pyforce has been completely re-written using pyvista as backend for mesh importing, computing integrals, and visualisation of results; in addition, functions are stored as numpy arrays, improving the ease of use of the package. This choice allows to use pyforce with any software solver able to export results in VTK format.

TopicsProcess Modeling & System Identification

arXiv categoriescs.LG

arXiv abstract pagePDF