Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
May 27, 2026Engineering Applications of Computational Fluid MechanicsOpen Access

Active learning framework for surrogate modelling of centrifugal pumps performance

View Full Paper
Ask AI
Bookmark
Share

Authors

DMDaniel Morantes-MoralesJVJuan Pablo ValdésPPPaula Pico

Discussion

Loading...

Member takes

Overview

Randomized trial evaluates pump performance modelling through active learning, indicating improved predictive accuracy.

Key Points

  • The aim is to create efficient surrogate models for predicting centrifugal pump performance using active learning and CFD.
  • Developed a hybrid framework integrating active learning and computational fluid dynamics.
  • Utilized synthetic fluid datasets and sampling strategies for model training.
  • Trained two machine learning models, XGBoost and GPR, evaluated against a fixed test set of real pump data.
  • GPR-based variational method achieved more accurate and stable improvements compared to random sampling.
  • Demonstrated that informed sampling reduces the computational demands of simulations while maintaining model accuracy.

Cite This Study

Morantes-Morales et al. (2026) studied this question.

synapsesocial.com/papers/6a16891d0c924ddd1bd57eeehttps://doi.org/10.1080/19942060.2026.2668183
View Full Paper
Ask AI
Bookmark
Share

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

  1. 1Construction of collaborative system for predicting and optimizing the performance of centrifugal pumps based on machine learning, and experimental verification2025
  2. 2CFD-driven machine learning surrogate modeling for thermo-hydraulic prediction in circular pipe flows2026
  3. 3Multi-fidelity surrogate modeling with low-fidelity auxiliary sampling for fluid flow prediction2026
  4. 4Optimizing Hydraulic Performance of a Jet Centrifugal Pump: A Hybrid Approach Combining Continuous Adjoint Method and Machine Learning2026
  5. 5Active learning for efficient data-driven aerodynamic modeling in spaceplane design2024 · 4 citations