Review of Artificial Neural Networks for Wind Turbine Fatigue Prediction

Wind turbines have emerged as a prominent renewable energy source globally.Efficient monitoring and detection methods are crucial to enhance their operational effectiveness,particularly in identifying fatigue-related issues.This review focuses on leveraging artificial neural networks(ANNs)for wind turbine monitoring and fatigue detection,aiming to provide a valuable reference for researchers in this domain and related areas.Employing various ANN techniques,including General Regression Neural Network(GRNN),Support Vector Machine(SVM),Cuckoo Search Neural Network(CSNN),Backpropagation Neural Network(BPNN),Particle Swarm Optimization Artificial Neural Network(PSO-ANN),Convolutional Neural Network(CNN),and nonlinear autoregressive networks with exogenous inputs(NARX),we investigate the impact of average wind speed on stress transfer function and fatigue damage in wind turbine structures.Our findings indicate significant precision levels exhibited by GRNN and SVM,making them suitable for practical implementation.CSNN demonstrates superiority over BPNN and PSO-ANN in predicting blade fatigue life,showcasing enhanced accuracy,computational speed,precision,and convergence rate towards the global minimum.Furthermore,CNN and NARX models display exceptional accuracy in classification tasks.These results underscore the potential of ANNs in addressing challenges in wind turbine monitoring and fatigue detection.However,it’s important to acknowledge limitations such as data availability and model complexity.Future research should explore integrating real-time data and advanced optimization techniques to improve prediction accuracy and applicability in real-world scenarios.In summary,this review contributes to advancing the understanding of ANNs’efficacy in wind turbine monitoring and fatigue detection,offering insights and methodologies that can inform future research and practical applications in renewable energy systems.

Assessing aerodynamic loads on low-rise buildings considering Reynolds number and turbulence effects: a review

This paper presents an extensive review of existing techniques used in estimating design wind pressures considering Reynolds number and turbulence effects,as well as a case study of a reference building investigated experimentally.We shed light on the limitations of current aerodynamic testing techniques,provisions in design standards,and computational fluid dynamics(CFD)methods to predict wind-induced pressures.The paper highlights the reasons for obstructing the standardization of the wind tunnel method.Moreover,we introduce improved experimental and CFD techniques to tackle the identified challenges.CFD provides superior and efficient performance by employing wall-modeled large-eddy simulation(WMLES)and hybrid RANS-LES models.In addition,we tested a large-scale building model and compared the results with published small-scale data.The findings reinforce our hypothesis concerning the scaling issues and Reynolds number effects in aerodynamic testing.

Experimental investigation of the aerodynamics of a large industrial building with parapet

The aerodynamic performance of a roof depends significantly on its shape and size,among other factors.For instance,large roofs of industrial low-rise buildings may behave differently compared to those of residential homes.The main objective of this study is to experimentally investigate how perimeter solid parapets can alter the flow pattern around a low-rise building with a large aspect ratio of width/height of about 7.6,the case of industrial buildings/shopping centers.Solid parapets of varied sizes are added to the roof and tested in an open-jet simulator in a comparative study to understand their impact on roof pressure coefficients.Roof pressures were measured in the laboratory for cases with and without parapets under different wind direction angles(representative of straight-line winds under open terrain conditions).The results show that using a parapet can alter wind pressures on large roofs.Parapets can modify the flow pattern around buildings and change the mean and peak pressures.The mean pressure pattern shows a reduction in the length of the separation bubble due to the parapet.The parapet of 14%of the building’s roof height is the most efficient at reducing mean and peak pressures compared to other parapet heights.