风电叶片自动铺层设备位置姿态估计算法研究  

RESEARCH ON LOCATION AND ATTITUDE ESTIMATION OF WIND TURBINE BLADE AUTOMATIC LAYERING EQUIPMENT

作　　者：张华强[1] 陈传训 杨先海[1] 刘卫生 王国栋 吕云飞 Zhang Huaqiang;Chen Chuanxun;Yang Xianhai;Liu Weisheng;Wang Guodong;Lyu Yunfei(School of Mechanical Engineering,Shangdong University of Technology,Zibo 255049,China;Lianyungang Zhongfu Lianzhong Composite Group Company Limited,Lianyungang 222000,China)

机构地区：[1]山东理工大学机械工程学院,淄博255049 [2]中复连众复合材料集团有限公司,连云港222000

出　　处：《太阳能学报》2021年第9期224-230,共7页Acta Energiae Solaris Sinica

基　　金：国家重点研发计划(可再生能源与氢能专项)(2018YFB1501203);山东省重点研发计划(2018GGX100304);山东省自然科学基金(ZR2019MEE076)。

摘　　要：为解决风电叶片自动铺层设备位置姿态估计问题,采用位置姿态传感器进行信息融合,提出杂草算法支持的粒子群优化算法(IWO-PSO),该算法以支持向量回归机(SVR)为基础算法并用于SVR参数寻优。首先利用高斯核函数将铺层设备位置姿态进行高维映射;然后利用IWO-PSO增强算法全局搜索能力,通过IWO的入侵特性优化PSO寻找适应度能力,最终获得SVR的最优参数,与同类算法相比算法收敛速度提高20%;实验验证表明该算法可高效快速完成对铺层设备位置姿态的估计,IWO-PSO-SVR算法求出铺层设备逆运动学数值解无需其他要求,姿态误差小于0.2 rad,位置误差小于0.03 m提高了姿态预测的精度,具有很高的工程应用价值和经济价值。To solve the problem of position and posture estimation of wind turbine blade automatic laminating equipment,the position and posture sensor is used for information fusion,and the Invasive Weed Optimization-Particle Swarm Optimization Algorithm(IWOPSO)is proposed to support the algorithm.Support Vector Regression(SVR)is the basic algorithm and is used for SVR parameter optimization.Firstly,the Gaussian kernel function is used to carry out the high-dimensional mapping of the position and posture of the layering equipment.Then the IWO-PSO is used to enhance the global search ability of the algorithm,and the ability of PSO to find fitness is optimized by IWO’s intrusion features.,and finally the optimal parameters of the SVR are obtained.Compared with the convergence speed,the speed is increased by 20%.The experimental results show that the algorithm can estimate the position and posture of the layering equipment efficiently and quickly.The new algorithm can find the inverse kinematic numerical solution of the layering equipment without other requirements.The posture error is less than 0.2 rad.The position error is less than 0.03 m,which improves the accuracy of posture prediction and has high engineering application value and economic value.

关 键 词：风电叶片 支持向量回归机 姿态估计 粒子群优化算法 自动铺层设备 

分 类 号：TP278[自动化与计算机技术—检测技术与自动化装置]