En
期刊文章
文章精选
当期目录
过刊浏览
高被引
高下载
特邀文章
专刊合集
待刊文章
Volume 113
您当前的位置：首页 > 期刊文章 > 过刊浏览 > Volumes 108-119 (2025) > Volume 113
High-fidelity CFD analysis of catalyst regeneration in a pilot-scale MTO fluidized bed reactor: A comparative assessment of LES and RANS models
Mahdieh Garmsirian, Reza Mosayebbi Behbahani *, Mohammad Reza Khosravi-Nikou
Department of Gas Engineering, Ahvaz Faculty of Petroleum, Petroleum University of Technology (PUT), Ahvaz, Iran
10.1016/j.partic.2026.02.033
Volume 113, June 2026, Pages 281-297
Received 9 December 2025, Revised 26 January 2026, Accepted 14 February 2026, Available online 21 March 2026, Version of Record 8 April 2026.
E-mail: behbahani@put.ac.ir
Highlights

• Direct comparison of LES and RANS for pilot-scale MTO catalyst regeneration.

• LES predicts temperature, velocity, and species more accurately than RANS.

• RANS shows higher deviations, missing transient phenomena like bubbles.

• LES requires ∼9 × more computational resources than RANS.

• Guidelines for model selection and hybrid LES/RANS strategies for industry.


Abstract

The regeneration of catalysts in fluidized bed reactors is a critical process in methanol-to-olefins (MTO) technology, where computational fluid dynamics (CFD) plays a pivotal role in optimizing performance. This study presents a comprehensive comparison of two turbulence modeling approaches Large Eddy Simulation (LES) and Reynolds-Averaged Navier-Stokes (RANS) for simulating the catalyst regeneration zone in a pilot-scale MTO reactor. Using ANSYS Fluent 2024 R2, both models were validated against published experimental data from the literature, evaluating their accuracy in predicting temperature distributions, velocity fields, gas-solid volume fractions, and species transport.

These results highlight the specific impact of turbulence modeling on temperature, velocity, etc. predictions., with deviations below 8% for critical parameters such as peak temperatures (669 °C vs. experimental 670–690 °C) and velocity profiles (12.7 m/s vs. PIV data). In contrast, RANS exhibits higher deviations (10–18%) due to its time-averaged nature, particularly in resolving transient phenomena like bubble dynamics and localized hot spots. However, LES demands significantly higher computational resources (∼1600 CPU-hours vs. RANS's ∼180 CPU-hours), highlighting a trade-off between fidelity and efficiency.

The study further proposes hybrid modeling strategies and design optimizations, such as refined gas distributor geometries and secondary air injection, to enhance combustion uniformity. These insights bridge the gap between academic research and industrial application, offering actionable guidelines for model selection in MTO reactor design and scale-up. The results underscore LES as the preferred choice for detailed analysis, while RANS remains viable for preliminary simulations where computational cost is a constraint.


Graphical abstract
Keywords
Catalyst regeneration; Fluidized bed reactor; CFD validation; Turbulence modeling; Combustion dynamics
文章导读