Machine learning based prediction of piezoelectric energy harvesting from wake galloping

Zhang, Chengyun; Hu, Gang; Yurchenko, Daniil; Lin, Pengfei; Gu, Shanghao; Song, Dongran; Peng, H.Y.; Wang, Junlei

doi:10.1016/j.ymssp.2021.107876

Cited by 63 publications

(19 citation statements)

References 63 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Finally, the weighted sum of the regression trees at each stage is added to obtain the final result. 40 This tree-based ensemble model is flexible, easy to train, and able to handle unnecessary or relevant features without over-fitting. 41 The GBRT was selected as the measurement model to accurately measure the substance content in water.…”

Section: Principle and Methodsmentioning

confidence: 99%

Color-deconvolution-based feature image extraction and application in water quality analysis

et al. 2022

View full text Add to dashboard Cite

show abstract

Section: Principle and Methodsmentioning

confidence: 99%

Color-deconvolution-based feature image extraction and application in water quality analysis

et al. 2022

View full text Add to dashboard Cite

show abstract

“…The authors explored the feasibility of training a machine learning model based on a series of experimental results for predicting the root mean square (RMS) voltage output from the wind energy harvester. 63 The major advantage of this scheme is that without knowing the underlying physics behind the multi-physics problem, the complicated effects of various factors on the energy harvesting performance of the wind energy harvester can be revealed by a well-trained machine learning model. However, from another Perspective, a well-trained machine learning model cannot provide any in-depth insight into the fundamental physics and mechanisms.…”

Section: A Machine-learning Approaches For Addressing the Fsi Problemmentioning

confidence: 99%

Perspectives in flow-induced vibration energy harvesting

Wang

Yurchenko

et al. 2021

Applied Physics Letters

Self Cite

View full text Add to dashboard Cite

Flow-induced vibration (FIV) energy harvesting has attracted extensive research interest in the past two decades. Remarkable research achievements and contributions from different aspects are briefly overviewed. Example applications of FIV energy harvesting techniques in the development of Internet of Things are mentioned. The challenges and difficulties in this field are summarized from two sides. First, the multi-physics coupling problem in FIV energy harvesting still cannot be well handled. There is a lack of system-level theoretical modeling that can accurately account for fluid-structure interaction, the electromechanical coupling, and complicated interface circuits. Second, the robustness of FIV energy harvesters needs to be further improved to adapt to the uncertainties in practical scenarios. To be more specific, the cut-in wind speed is expected to be further reduced and the power output to be increased. Finally, Perspectives on the future development in this direction are discussed. Machine-learning approaches, the versatility of metamaterials, and more advanced interface circuits should receive more attention from researchers, since these cutting-edge techniques may have the potential to address the multi-physics modeling problem of FIV energy harvesters and significantly improve the operation performance. In addition, in-depth collaborations between researchers from different disciplines are anticipated to promote the FIV energy harvesting technology to step out of the lab and into real applications.

show abstract

“…Flow-induced vibration (FIV) is mainly divided into vortex-induced vibration (VIV) [8,9,10], galloping [11,12], utter [13,14] and wake-induced vibration (WIV) [15,16]. The energy harvester based on VIV, galloping and WIV design has aroused the interest of scholars.…”

Section: Introductionmentioning

confidence: 99%

Energy harvesting of flow induced vibration enhanced by bionic non-smooth surfaces

Wang

Tang

Yang

et al. 2023

Preprint

View full text Add to dashboard Cite

Inspired by the shield scale on the shark surface, a D-type bionic fin with non-smooth surface is proposed and used in tandem cylinders piezoelectric energy harvesters (PEHs) for the utilization of wind energy on the roof of buildings. The repeating unit of D-type bionic fin is semicircle, and the corresponding center angle of each repeating unit is 7.2°. PEHs consist of a piezoelectric cantilever beam and a wind-interference cylinder connected to the beam tip. The influence of the spacing ratio on the amplitude of PEHs with D-type bionic fins added under elastic interference is studied through wind tunnel tests and three installation positions are designed: only installed upstream, only installed downstream, and not installed upstream and downstream (bare). It is found that the maximum amplitude response law of the upstream piezoelectric energy harvester (UPEH) is not affected by the D-type bionic fins, and the D-type bionic fins can make the downstream piezoelectric energy harvester (DPEH) realize the change of the maximum amplitude from small spacing ratio to large spacing ratio. In addition, the influence of the installation position of D-type bionic fins on the output voltage of upstream and downstream PEHs is also studied. The research shows that the addition of D-type bionic fins significantly changes the vibration behavior of PEHs. D-type bionic fins can enhance the energy harvest performance by coupling "coupled vortex-induced vibration" and wake induced galloping (WIG), and increasing the surface velocity of PEHs. D-type bionic fins can also expand the bandwidth of the voltage harvested by the PEHs. The analysis of the power under the experimental wind speed shows that the installation of D-type fins in PEHs can increase the output power of the upstream and downstream PEH by 392.28% and 13% respectively compared with the bare piezoelectric energy piezoelectric energy harvester (BARE-PEH). In addition, the computational fluid dynamics is used to analyze the flow pattern, wake structure and lift coefficient of PEHs, and the reason why the installation of D-type bionic fins in the upstream has an impact on the harvest performance of upstream and downstream PEHs at 1.5 spacing ratio is explained.

show abstract

Machine learning based prediction of piezoelectric energy harvesting from wake galloping

Cited by 63 publications

References 63 publications

Color-deconvolution-based feature image extraction and application in water quality analysis

Color-deconvolution-based feature image extraction and application in water quality analysis

Perspectives in flow-induced vibration energy harvesting

Energy harvesting of flow induced vibration enhanced by bionic non-smooth surfaces

Contact Info

Product

Resources

About