Automated Prediction of Compressor Performance Maps: Surrogate-Based Optimization with RNNs for Enhanced Extrapolation and Interpolation
This paper presents a surrogate-assisted hyperparameter tuning approach using the Sequential Parameter Optimization Toolbox (SPOT) to improve prediction of compressor performance maps. Accurate prediction is critical for efficiency and reliability in power generation, aerospace and other industries, but experimental data collection is costly and time-consuming. By employing a Kriging-based surrogate model for hyperparameter tuning, the evaluation of the objective function is significantly reduced.
Authors
- Thomas Bartz–Beielstein (ORCID: https://orcid.org/0000-0002-5938-5158)
- Alexander Hinterleitner (ORCID: https://orcid.org/0009-0002-7615-6952)
- Jens Uwe Brandt (ORCID: https://orcid.org/0009-0001-9713-8158)
- Richard Schulz (ORCID: https://orcid.org/0009-0007-0834-9927)
- N. Pütz (ORCID: https://orcid.org/0009-0002-9895-3661)
- Matthias M. Müller
Institutions
- TH Köln - University of Applied Sciences (DE)
- University Hospital Augsburg (DE)
Publication Details
- Journal
- Proceedings of the Genetic and Evolutionary Computation Conference Companion
- Published
- 2025-07-14
- DOI
- https://doi.org/10.1145/3712255.3734366
- Primary Topic
- Turbomachinery Performance and Optimization
- Type
- article
- Field-Weighted Citation Impact
- 0.00