Article
Engineering, Marine
Islam Abdelghafar, Emeel Kerikous, Stefan Hoerner, Dominique Thevenin
Summary: Wind turbines are a possible solution for harnessing the kinetic energy in wind and converting it to useful energy. In this study, a bionic blade design inspired by sandeels is proposed to optimize the performance of the Savonius rotor. Experimental results show a significant increase in turbine efficiency compared to conventional designs.
Article
Computer Science, Interdisciplinary Applications
Anthony Aubry, Hamid R. Karbasian, Brian C. Vermeire
Summary: This paper demonstrates the use of Large Eddy Simulation (LES) and the Mesh Adaptive Direct Search (MADS) optimization algorithm for shape optimization of Low Pressure Turbine (LPT) cascades. The results show that using LES for LPT cascade optimization is feasible and can lead to significant improvements in performance.
COMPUTERS & FLUIDS
(2022)
Article
Engineering, Mechanical
Marco Gambitta, Arnold Kuehhorn, Bernd Beirow, Sven Schrape
Summary: This study investigates the impact of geometrical uncertainties due to the manufacturing process on the modal forcing of axial compressor blades. A stochastic model is created to represent the geometrical variability and a data reduction method is proposed to evaluate its effect on vibration behavior. Uncertainty quantification is performed to identify potential increase in low engine orders.
JOURNAL OF TURBOMACHINERY-TRANSACTIONS OF THE ASME
(2022)
Article
Computer Science, Artificial Intelligence
Lucas R. C. de Farias, Aluizio F. R. Araujo
Summary: This paper introduces a MOEA/D-UR algorithm based on decomposition, which utilizes a metric to detect improvements and a procedure to increase diversity in the objective space. Experimental results suggest that MOEA/D-UR is more effective in handling real-world problems and multi-objective scenarios compared to other algorithms.
SWARM AND EVOLUTIONARY COMPUTATION
(2022)
Article
Energy & Fuels
Rania M. Himeur, Sofiane Khelladi, Mohamed Abdessamed Ait Chikh, Hamid Reza Vanaei, Idir Belaidi, Farid Bakir
Summary: This article describes a new methodology that uses analytical modeling and experimental data to deduce the performance of cross-flow fans in turbomachinery. By conducting multidisciplinary studies, CFD simulations, and optimization algorithms, the efficiency of cross-flow fans can be evaluated and their global performance predicted.
Article
Engineering, Mechanical
Luying Zhang, Loukia Kritioti, Peng Wang, Jiangnan Zhang, Mehrdad Zangeneh
Summary: In this paper, an evaluation method based on entropy production has been developed to quantify the loss generation from different flow mechanisms inside a turbomachine. The method is applicable to different machine types and has been applied to centrifugal compressors. The impact of design modification on loss generation has also been assessed.
JOURNAL OF TURBOMACHINERY-TRANSACTIONS OF THE ASME
(2022)
Review
Thermodynamics
M. Sreekanth, R. Sivakumar, M. Sai Santosh Pavan Kumar, K. Karunamurthy, M. B. Shyam Kumar, R. Harish
Summary: This paper provides a detailed and objective review of regenerative flow turbomachines, including pumps, blowers, and compressors, covering aspects such as design, working principles, and performance parameters. It also includes experimental work, consolidated plots, industrial outlook, and suggestions for future research. Targeted at designer engineers needing quantitative data.
PROCEEDINGS OF THE INSTITUTION OF MECHANICAL ENGINEERS PART A-JOURNAL OF POWER AND ENERGY
(2021)
Article
Energy & Fuels
Sergio Gonzalez Horcas, Nestor Ramos-Garcia, Ang Li, Georg Pirrung, Thanasis Barlas
Summary: Curved tip extensions are an innovative concept in horizontal axis wind turbines that can improve performance and reduce costs. This study provides an overview of different computational aerodynamic models used in the design stage, comparing their predictions for various load cases. The study highlights the limitations of certain models in capturing the effects of curved tips, while also noting the accurate results of other methods in most load cases.
Article
Chemistry, Physical
Mengshan Li, Ming Zeng, Bingsheng Chen, Lixin Guan, Yan Wu, Nan Wang
Summary: A novel particle dynamics evolutionary algorithm (DP-PD-EA) is developed to simulate the movement of molecules in the process of dissolution, showing better prediction performances for the solubility of supercritical carbon dioxide (SCCO2) in three polymers than other models.
JOURNAL OF MOLECULAR LIQUIDS
(2022)
Article
Computer Science, Artificial Intelligence
Shuai Wang, Jing Liu, Yaochu Jin
Summary: The robustness of complex networks is crucial for their stability, and multiobjective robustness optimization is gaining increasing attention. However, challenges remain, including different computational complexities, insufficient network diversity, and high computational costs. By introducing a computationally efficient multiobjective optimization algorithm, a unique feature-based fitness evaluation method, and a surrogate ensemble based on graph embedding information, these challenges have been successfully addressed.
IEEE TRANSACTIONS ON EVOLUTIONARY COMPUTATION
(2021)
Article
Computer Science, Artificial Intelligence
Majdi Radaideh, Koroush Shirvan
Summary: In this study, a rule-based RL method is proposed to guide evolutionary algorithms in constrained optimization, showing significant improvement in continuous optimization over standalone algorithms. RL-guided EA outperforms standalone algorithms by more than 10 times in exploration capabilities and computational efficiency, indicating the effectiveness of RL in focusing the search space in areas where expert knowledge has merit.
KNOWLEDGE-BASED SYSTEMS
(2021)
Article
Computer Science, Artificial Intelligence
Karl Grantham, Muhetaer Mukaidaisi, Hsu Kiang Ooi, Mohammad Sajjad Ghaemi, Alain Tchagang, Yifeng Li
Summary: This paper proposes a prototypical deep evolutionary learning (DEL) process that integrates deep generative model and multi-objective evolutionary computation for molecular design. DEL enables evolutionary operations in the latent space of the generative model, generates promising novel molecular structures, and improves the generative model learning through fine-tuning. Experimental results show that DEL achieves improvement in property distributions and outperforms other baseline molecular optimization algorithms in generating samples. Additionally, comparisons with various deep generative models demonstrate the benefits of DEL in improving sample populations.
IEEE COMPUTATIONAL INTELLIGENCE MAGAZINE
(2022)
Article
Engineering, Mechanical
Ben Mohankumar, Cesare A. Hall, Mark J. Wilson
Summary: Sweep is conventionally used to reduce losses in transonic fans, but its effectiveness diminishes at low pressure ratios. However, sweep can still improve performance at critical high Mach number off-design conditions. In this study, computational fluid dynamics was used to compare two transonic low pressure ratio fans at high angle of attack. Contrary to expectations, the swept fan increased rotor losses due to rotor choking and rotor-separation interaction.
JOURNAL OF TURBOMACHINERY-TRANSACTIONS OF THE ASME
(2022)
Article
Engineering, Mechanical
Chun Yui Wong, Pranay Seshadri, Ashley Scillitoe, Andrew B. Duncan, Geoffrey Parks
Summary: By leveraging ideas from dimension reduction, low-dimensional representations of aerodynamic performance metrics can be constructed to evaluate the impact of geometric variability on aerodynamic performance before manufacturing, as well as the amplified impact of degradation during service. Techniques for sampling within an inactive subspace can be designed to draft manufacturing tolerances and quantify whether a scanned component should be used or scrapped. The blade envelope is introduced as a computational manufacturing guide and qualitative visualization.
JOURNAL OF TURBOMACHINERY-TRANSACTIONS OF THE ASME
(2022)
Article
Computer Science, Artificial Intelligence
Dawei Zhan, Huanlai Xing
Summary: An incremental Kriging model for high-dimensional surrogate-assisted evolutionary computation is proposed in this study to reduce time complexity by updating the model step by step, suitable for online SAEAs. The algorithm achieves competitive optimization results on test problems and is significantly faster than the standard Kriging approach.
IEEE TRANSACTIONS ON EVOLUTIONARY COMPUTATION
(2021)