摘要
针对风力发电机叶片覆冰问题,提出一种改进Adaboost-SVM组合算法的风力发电机叶片覆冰检测方法。该方法使用Adaboost架构对一组差异化的SVM模型分类结果进行集成。通过改进集成策略,改善了Adaboost-SVM算法对不平衡数据集成速度慢、集成效果差的问题。该算法对风力发电机叶片覆冰故障检测的精确度达到92.12%,故障的查全率达到76.54%,具有很好的泛化性和实用价值。
Aimed at the problem of wind turbine blades icing,an improved Adaboost-SVM combinatorial algorithm for wind turbine blade icing detection method is proposed.This method used the Adaboost architecture to integrate a set of differentiated SVM model classification results.By improving the integration strategy,the problem of slow integration of unbalanced data and poor integration effect of the Adaboost-SVM algorithm was improved.The accuracy of this algorithm for wind turbine blade icing fault detection reached 92.12%,and the fault recall rate reached 76.54%.It had good generalization and practical value.
作者
冉浦东
范磊
张军
张子凡
庞成鑫
黄墀志
Ran Pudong;Fan Lei;Zhang Jun;Zhang Zifan;Pang Chengxin;Huang Chizhi(School of Electronic and Information Engineering,Shanghai University of Electric Power,Shanghai 201306,China;NARI-TECH Nanjing Control Systems Co.,Ltd.,Nanjing 210000,Jiangsu,China)
出处
《计算机应用与软件》
北大核心
2023年第5期110-114,共5页
Computer Applications and Software
基金
国家电网有限公司科学技术项目(SGSCJY00GHJS2000014)。
作者简介
冉浦东,硕士,主研领域:机器学习在电力设备故障检测中的应用;范磊,硕士;张军,高工;张子凡,硕士;庞成鑫,教授;黄墀志,工程师。
