| 21 | 0 | 29 |
| 下载次数 | 被引频次 | 阅读次数 |
目的:针对风电叶片激光除冰过程中易引发的叶片表面热损伤问题，提出一种基于CNN-HHO模型的激光除冰速度预测方法。方法:采用卷积神经网络(convolutional neural network, CNN)模型预测激光除冰速度，并引入哈里斯鹰优化(Harris hawk optimization, HHO)算法，优化CNN模型的超参数空间，形成CNN-HHO优化模型。结果:对比分析优化前后的CNN模型性能发现，经HHO算法优化后，模型平均绝对百分比误差MAPE(mean absolute percentage error, MAPE)减小了2.78%,决定系数R2增加了0.004。进一步将CNN-HHO预测模型与反向传播(back propagation, BP)、BP-HHO、极限学习机(extreme learning machine, ELM)、ELM-HHO和CNN预测模型进行对比分析，结合泰勒图可视化分析表明，CNN-HHO模型的各项性能评价指标均表现最佳，其MAPE=3.42%,R2=0.992。结论:所设计的CNN-HHO模型能够实现高精度的激光除冰速度预测，为风电叶片激光除冰工艺参数的智能调控提供了重要的理论支撑。
Abstract：Aims: To address the issue of thermal damage easily induced on the blade surface during laser de-icing of wind turbine blades, a laser de-icing speed prediction method based on a CNN-HHO model is proposed. Methods: A convolutional neural network(CNN) was employed to predict the laser de-icing speed, while the Harris hawk optimization(HHO) algorithm was introduced to optimize the CNN's hyperparameter space, resulting in a CNN-HHO optimization model. Results: A comparative analysis of the CNN model's performance before and after optimization showed that, following optimization using the Harris hawks optimization algorithm, the model's mean absolute percentage error(MAPE) decreased by 2.78%, while the coefficient of determination(R2) increased by 0.004. Furthermore, the CNN-HHO model was compared with several other prediction models, including the back propagation(BP) model, BP-HHO, the extreme learning machine(ELM), ELM-HHO, and the original CNN model. Based on performance metrics and visualization through a Taylor diagram, the CNN-HHO model demonstrated superior performance across all evaluation criteria, achieving the lowest MAPE of 3.42% and the highest R2 value of 0.992. Conclusions: The designed CNN-HHO model enables high-precision prediction of laser de-icing speed, providing significant theoretical support for the intelligent regulation of laser de-icing process parameters for wind turbine blades.
[1] SOGUKPINAR H,BOZKURT I,OYUCU S,et al.Unveiling the wind energy future of Turkiye with policies technologies and potential[J].Heliyon,2025,11(4):e42592.
[2] WEI K X,YANG Y,ZUO H Y,et al.A review on ice detection technology and ice elimination technology for wind turbine[J].Wind Energy,2020,23(3):433-457.
[3] YU S R,SONG M J,GAO R M,et al.A review of icing prediction techniques for four typical surfaces in low-temperature natural environments[J].Applied Thermal Engineering,2024,241:122418.
[4] REKUVIENE R,SAEIDIHARZAND S,MAZEIKA L,et al.A review on passive and active anti-icing and de-icing technologies[J].Applied Thermal Engineering,2024,250:123474.
[5] DANILIUK V,XU Y M,LIU R B,et al.Ultrasonic de-icing of wind turbine blades:Performance comparison of perspective transducers[J].Renewable Energy,2020,145:2005-2018.
[6] ZHAO Y,WANG X,ZHOU Q B,et al.Numerical study of lightning protection of wind turbine blade with de-icing electrical heating system[J].Energies,2020,13(3):691.
[7] 刘智颖，穆竺，王加科，等.激光除冰光学系统的设计与参数分析[J].光子学报，2018,47(8):147-153.LIU Z Y,MU Z,WANG J K,et al.Design and parameter analysis of laser deicing optical system[J].Acta Photonica Sinica,2018,47(8):147-153.
[8] 王一博.风电机组叶片的无人机除冰控制系统与优化试验研究[D].呼和浩特：内蒙古农业大学，2022:10-14.WANG Y B.Research on UAV Deicing Control System and Optimization Experiment of Wind Turbine Blade[D].Hohhot:Inner Mongolia Agricultural University,2022:10-14.
[9] SHEHAB M,MASHAL I,MOMANI Z,et al.Harris Hawks optimization algorithm:Variants and applications[J].Archives of Computational Methods in Engineering,2022,29(7):5579-5603.
[10] 刘骏鹏.哈里斯鹰算法的改进及应用研究[D].杭州：浙江大学，2021:2-8.LIU J P.Research on Improved Harris Hawks Optimization and Its Applications[D].Hangzhou:Zhejiang University,2021:2-8.
[11] DOKEROGLU T,OZDEMIR Y S.A new robust Harris hawk optimization algorithm for large quadratic assignment problems[J].Neural Computing & Applications,2023,35(17):12531-12544.
[12] GAO Z P,YI W J.Prediction of projectile interception point and interception time based on Harris hawk optimization-convolutional neural network-support vector regression algorithm[J].Mathematics,2025,13(3):338.
[13] GADEKALLU T R,SRIVASTAVA G,LIYANAGE M,et al.Hand gesture recognition based on a Harris hawks optimized convolution neural network[J].Computers & Electrical Engineering,2022,100:107836.
[14] DANGI R,LALWANI P.Harris hawks optimization based hybrid deep learning model for efficient network slicing in 5G network[J].Cluster Computing-the Journal of Networks Software Tools and Applications,2024,27(1):395-409.
[15] HEIDARI A A,MIRJALILI S,FARIS H,et al.Harris hawks optimization:Algorithm and applications[J].Future Generation Computer Systems-the International Journal of Escience,2019,97:849-872.
[16] ALABOOL H M,ALARABIAT D,ABUALIGAH L,et al.Harris hawks optimization:A comprehensive review of recent variants and applications[J].Neural Computing & Applications,2021,33(15):8939-8980.
[17] IMAM A A,ABUSORRAH A,SEEDAHMED M M A,et al.Accurate forecasting of global horizontal irradiance in Saudi Arabia:A comparative study of machine learning predictive models and feature selection techniques[J].Mathematics,2024,12(16):2600.
基本信息:
DOI：
中图分类号:TM315;TN249;TP18
引用信息:
[1]董科锋,金英连,吴善强等.基于CNN-HHO模型的风电叶片激光除冰速度预测研究[J].中国计量大学学报,2025,36(02):219-226.
基金信息:
浙江省重点研发计划项目(No.2023C03186)