Airfoils optimization based on deep reinforcement learning to improve the aerodynamic performance of rotors

Liu, Jiaqi; Chen, Rongqian; Lou, Jinhua; Wu, Hao; You, Yancheng; Chen, Zhengwu

doi:10.1016/j.ast.2023.108737

Cited by 6 publications

(3 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“… Reinforcement learning (RL): RL techniques, such as Q‐learning or deep reinforcement learning, optimize decision‐making processes through trial and error learning. RL can be employed for optimizing complex systems where actions influence future states and outcomes 59–63 …”

Section: Bo Of Dynamic Systemsmentioning

confidence: 99%

A data‐driven Bayesian approach for optimal dynamic product transitions

Flores‐Tlacuahuac,

Fuentes‐Cortés

2024

AIChE Journal

View full text Add to dashboard Cite

In the processing industry, dynamic product transitions are essential for achieving high product quality, minimizing the use of raw materials and energy and reducing production costs. However, optimizing dynamic product transitions is a challenging task due to the complex dynamics of the process and the uncertainty in the measurements. In this article, a data‐driven Bayesian approach for optimal dynamic product transitions is proposed. The proposed approach is based on a dynamic optimization problem that is solved using a Bayesian optimization algorithm. One of the advantages of this approach for process optimization tasks is that it does not require a first‐principles dynamic mathematical model for drawing optimal solutions. The approach is applied to three case studies, and the results are comparable in performance quality with those obtained using a traditional gradient‐based optimization approach. The results show that the proposed approach is able to find optimal transition trajectories that meet the product composition requirements using smooth control actions. The approach is also able to cope with noisy measurements, which is an important feature in real‐world applications. The proposed approach has several advantages over traditional optimization approaches, including being data driven, able to cope with noisy measurements, computationally efficient, and it requires modest computational effort. Complex online optimal control problems can benefit from adopting a data‐driven Bayesian optimization scheme.

show abstract

Section: Bo Of Dynamic Systemsmentioning

confidence: 99%

A data‐driven Bayesian approach for optimal dynamic product transitions

Flores‐Tlacuahuac,

Fuentes‐Cortés

2024

AIChE Journal

View full text Add to dashboard Cite

show abstract

“…An alternative approach is based on the cascade of airfoils [6][7][8][9][10], which is based on the strip theory. The blade strip theory is a theoretical model used to analyze the aerodynamic behavior of an axial-flow fan blade by dividing it into a series of blade strips along the spanwise direction.…”

Section: Introductionmentioning

confidence: 99%

“…The DDES revealed detailed vortex structures, including shedding streamwise vortices, impacting wake behavior. Liu et al [7] introduced a deep reinforcement learning (DRL) framework for rotor airfoil optimization to address issues with conventional methods lacking interpretability and generalizability. Using neural networks and reinforcement learning, it enhances the rotor airfoil dynamic stall characteristics.…”

Section: Introductionmentioning

confidence: 99%

Development of a Performance-Based Design Technique for an Axial-Flow Fan Unit Using Airfoil Cascades Based on the Blade Strip Theory

Ryu,

Cheong,

Kim

et al. 2024

Applied Sciences

View full text Add to dashboard Cite

Axial-flow fans are widely used as cooling fans in the outdoor units of split-type air conditioners. The design of an axial-flow fan blade involves stacking several airfoils that can be differently designed for each spanwise section. However, the complex flow field around the fan blade, including circumferential and axial flows, presents challenges when applying the single airfoil theory. This study proposed a systematic performance-based design method for axial-flow fans using a cascade of airfoils based on the blade strip theory. The theory characterized the complex three-dimensional flow field driven by an axial-flow fan in terms of a two-dimensional cascade of airfoil flows. Computational fluid dynamics based on finite volume methods were used to predict the flow field and aerodynamic sound sources of an existing low-pressure axial-flow fan partially covered by a fan shroud, and the results were validated against experimental measurements. Three radial locations in the spanwise region from the hub to the blade tip that have a significant impact on aerodynamic performance were selected, and the two-dimensional flow field on a cylindrical surface with a constant radius was extracted from the three-dimensional flow field to characterize the performance of an axial fan. Then, the airfoils at the targeted span locations were optimized for a higher flow rate and greater efficiency via two-dimensional simulations using the cascades of the airfoil, and the selected optimized airfoils were applied to existing fan blades. The effectiveness of the proposed performance-based design method for low-pressure axial-flow fans was validated by the results, which showed that the redesigned fan blades with cascades of airfoils performed as predicted, increasing the intended higher flow rate by about 1%, improving power consumption by 8%, and lowering the overall sound pressure level by 1.5 dBA.

show abstract