Journal article · 2026 · Open access

Multi-fidelity Bayesian optimization framework for CFD-based non-premixed burner design

Patrick Souza Lima, Paulo Roberto Santana dos Reis, Alex Álisson Bandeira Santos, Ehecatl Antonio del Rio‐Chanona, Idelfonso B. R. Nogueira

Results in Engineering, 32, 112628 · Published 26 August 2026 · DOI 10.1016/j.rineng.2026.112628

Open full textCite this paper

In brief

“Multi-fidelity Bayesian optimization framework for CFD-based non-premixed burner design” is a 2026 journal article by Patrick Souza Lima and 4 co-authors, published in Results in Engineering. OpenAlex counts 2 citations as of 19 November 2026, which puts it in the top 6% of works published in 2026 by citations. It is open access under a CC BY-NC-ND licence: the full text is free to read at Results in Engineering.

Citations
2
Most cited in
2026 (2)
Field-weighted impact
3.32
References
44
Authors
5

Abstract

We propose a cost-aware variable-fidelity Bayesian optimization (MF-BO) framework that couples computational fluid dynamics (CFD) with Gaussian-process (GP) surrogates for the geometric design of a hydrogen-enriched non-premixed burner. Unlike hierarchical multi-fidelity approaches, the framework augments the two geometric design variables, burner height h and length l , with mesh element size s as a continuous fidelity variable, mapped to a normalized fidelity index Z ( s ) ∈ [0, 1] so that no discrete fidelity levels are required. A pre-fitted analytical runtime model t ^ ( h , l , s ) provides the relative cost signal used to penalize expensive queries (rank-preserving against the measured wall times of the executed campaign: Spearman ρ = 0.78 , Pearson r = 0.86 ), and a constrained noisy expected improvement (qNEI) drives sampling under a NO x mass-fraction cap of 2 × 10 − 6 . Each CFD evaluation solves a two-dimensional axisymmetric steady RANS problem (standard k –ε closure) with a steady laminar flamelet combustion model (GRI–Mech 3.0) and thermal/prompt NO x post-processing; this physical model was validated against the open-flame Sandia Flame D benchmark, and no experimental validation under confined conditions was performed. Both the MF-BO framework and an executed single-fidelity (SF-BO) baseline were run under identical conditions ( N = 50 evaluations, same BoTorch/GPyTorch infrastructure), enabling a direct empirical comparison. Over the MF campaign, the GP surrogates converged after approximately 20 iterations; leave-one-out cross-validation on thermally active designs ( T m ≥ 1000 K, N = 46 ) confirms surrogate accuracy: RMSE = 17.0 K and R 2 = 0.795 for T m . Automatic relevance determination identifies reactor length l as the dominant geometric variable for temperature. The best validated MF candidate was found at h * = 218.0 mm, l * = 1228.3 mm, reaching T m = 1 , 553 K when re-evaluated at full fidelity ( s = 0.350 mm, Z = 1.0 ) with the emission constraint satisfied. The SF baseline reached T m = 1 , 557 K at h * = 250.0 mm, l * = 1237.5 mm, a thermally equivalent result ( < 0.3% difference), but required 2, 715 min (45.2 h) versus 740 min (12.3 h) for the MF campaign, corresponding to an empirical wall-time saving of 73% (3.7 × speedup) between the two executed campaigns, derived entirely from measured wall-clock times. This figure reflects a single run of each campaign under the present benchmark configuration (matched N = 50 budgets, the same geometric Sobol initialization points, and β =…

Cite this paper

Formatted from the catalogue record. Check names, capitals and page numbers against the paper before you submit.

APA 7

Lima, P. S., dos Reis, P. R. S., Santos, A. Á. B., del Rio‐Chanona, E. A., & Nogueira, I. B. R. (2026). Multi-fidelity Bayesian optimization framework for CFD-based non-premixed burner design. Results in Engineering, 32, 112628. https://doi.org/10.1016/j.rineng.2026.112628

MLA 9

Lima, Patrick Souza, et al. “Multi-fidelity Bayesian optimization framework for CFD-based non-premixed burner design.” Results in Engineering, vol. 32, 2026, p. 112628. https://doi.org/10.1016/j.rineng.2026.112628.

Chicago

Lima, Patrick Souza, Paulo Roberto Santana dos Reis, Alex Álisson Bandeira Santos, Ehecatl Antonio del Rio‐Chanona, and Idelfonso B. R. Nogueira. 2026. “Multi-fidelity Bayesian optimization framework for CFD-based non-premixed burner design.” Results in Engineering 32: 112628. https://doi.org/10.1016/j.rineng.2026.112628.

Harvard

Lima, P. S. et al. (2026) ‘Multi-fidelity Bayesian optimization framework for CFD-based non-premixed burner design’, Results in Engineering, 32, p. 112628. doi:10.1016/j.rineng.2026.112628.

IEEE

P. S. Lima, P. R. S. dos Reis, A. Á. B. Santos, E. A. del Rio‐Chanona, and I. B. R. Nogueira, “Multi-fidelity Bayesian optimization framework for CFD-based non-premixed burner design,” Results in Engineering, vol. 32, p. 112628, 2026, doi: 10.1016/j.rineng.2026.112628.

BibTeX

@article{lima2026multi,
  title     = {{Multi-fidelity Bayesian optimization framework for CFD-based non-premixed burner design}},
  author    = {Lima, Patrick Souza and dos Reis, Paulo Roberto Santana and Santos, Alex Álisson Bandeira and del Rio‐Chanona, Ehecatl Antonio and Nogueira, Idelfonso B. R.},
  journal   = {Results in Engineering},
  year      = {2026},
  volume    = {32},
  pages     = {112628},
  doi       = {10.1016/j.rineng.2026.112628},
  url       = {https://doi.org/10.1016/j.rineng.2026.112628}
}

Download for a reference manager: BibTeX (.bib) · RIS for Zotero, Mendeley and EndNote (.ris)

Topic and keywords

Physical Sciences › Computer Science › Computational Theory and Mathematics › Advanced Multi-Objective Optimization Algorithms

  • multi-fidelity Bayesian optimization
  • Kriging surrogate model
  • fidelity index
  • Sandia Flame D
  • automatic relevance determination

Work with this paper

Most-cited papers in Computer Science

All highly cited Computer Science papers

Questions about this paper

How do I cite this paper in APA style?

Lima, P. S., dos Reis, P. R. S., Santos, A. Á. B., del Rio‐Chanona, E. A., & Nogueira, I. B. R. (2026). Multi-fidelity Bayesian optimization framework for CFD-based non-premixed burner design. Results in Engineering, 32, 112628. https://doi.org/10.1016/j.rineng.2026.112628

How many times has this paper been cited?

2 times, according to OpenAlex on 19 November 2026. Google Scholar and Scopus count different sets of sources, so their numbers differ.

Is this paper free to read?

Yes. It is open access at Results in Engineering: sign in and open the full text from this page.

Who wrote this paper?

Patrick Souza Lima, Paulo Roberto Santana dos Reis, Alex Álisson Bandeira Santos, Ehecatl Antonio del Rio‐Chanona, Idelfonso B. R. Nogueira.

What is the DOI of this paper?

10.1016/j.rineng.2026.112628, permanently at https://doi.org/10.1016/j.rineng.2026.112628. Sign in and open it from this page.

Bibliographic data from OpenAlex (CC0), last checked 19 November 2026. ResearchGuru does not host this paper; once you are signed in, the links open it at the publisher or the repository that does.