爬壁机器人风电叶片曲面路径规划设计与仿真  

DESIGN AND SIMULATION OF PATH LAYOUT FOR CLIMBING ROBOT ON ROTOR BLADE OF WIND TURBINE

作　　者：王斌锐[1] 骆浩华[1] 金英连[1] 冯伟博[1] 

机构地区：[1]中国计量学院机电工程学院,杭州310018

出　　处：《太阳能学报》2015年第8期1806-1811,共6页Acta Energiae Solaris Sinica

基　　金：国家高技术研究发展(863)计划(2015AA042302);浙江省自然科学基金(LY14F030021);科技部质检公益项目(201210076)

摘　　要：针对爬壁机器人在风电叶片叶中段表面的全局路径规划,构造路径点对空间曲面体的碰撞罚函数神经网络结构,以路径能量最优为目标设计路径点优化算法,并引入模拟退火法避免算法陷入局部极小值,设计温度更新函数。根据兆瓦级风电叶片曲率变化幅度,将风电叶片叶中段等效为由边界曲面及边界平面构成的曲面体。用Matlab仿真路径规划算法,得到不同速度系数和模拟退火权值下的可行路径曲线。结果表明较小的速度系数能提高路径点位置精度,模拟退火权值越小收敛速度越高,该算法可规划出沿着叶中段表面的可行路径。Aiming at overall path layout for climbing robot on the surface of middle leaf of wind turbine blade, the path point for the collision penalty function neural network structure of space surface body was constructed. The path point optimum algorithm was designed through setting the best path energy as the target, and introducing the simulation annealing method to avoid sinking local minimum, the temperature updating function was designed. The middle leaf of blade of wind turbine was equated to the space surface composed by the boundary surfaces and boundary plane according to curvature change of megawatt wind turbine blade. The path layout algorithm was simulated using Madab to get feasible path curve for different speed coefficient and simulation annealing weight. The results showed that the smaller speed coefficient can improve the position accuracy of path point, the smaller simulation annealing weight can quicken convergence speed. The algorithm can plan the feasible path along the surface of middle leaf of wind turbine blade.

关 键 词：风电叶片 爬壁机器人 神经网络 模拟退火 路径规划 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]