节点文献
基于机械臂的月壤挖掘采样及力学参数辨识的研究
Study of Lunar Soil Excavation Acquisition and Mechanical Parameters Identification Based on Robot Arm
【作者】 李操;
【作者基本信息】 哈尔滨工业大学 , 机械电子工程, 2014, 博士
【摘要】 对未知领域的认知,对茫茫宇宙的探索,是人类发展的永恒动力,是人类的不懈追求。美国、前苏联、俄罗斯等均对地外星体进行过探测和采样分析。作为新晋崛起的航天大国——中国,目前也规划的“绕”、“落”、“回”三个阶段的月球探测计划。而在探测过程中,节能和对月球土壤(月壤)参数进行辨识则是两大焦点问题。本文在探月三期计划中的“表取采样机构月壤参数反演”项目支持下,旨在研制出一套适用于月球探测、采样返回的机械臂末端采样器,并且研究出一套能够可靠实现的月壤参数辨识的方法,同时依据辨识结果指导最优挖掘轨迹的制定,进而达到减少能耗的目的。考虑到月球表面采样环境的不确定性及低重力环境下月壤颗粒运动的不可知性,本文围绕着低重力环境下月壤颗粒的流动性,运用颗粒物质力学软件EDEM进行仿真分析,获得了月壤内摩擦角及月壤与工具间的摩擦角对其流动性的影响规律,同时得到了低重力环境下采样器敲击机构的最优激振频率及方向。在结构设计方面应用仿生学设计原理,以减阻为目标对末端采样器挖掘铲的铲形进行优选设计。在此基础上研制出一种集采样、激振、贮存、锁定等功能于一体的月壤采样器。本文以月壤—铲斗的土力学作为理论分析基础,对月壤发生剪切强度失效的判定方法、挡土墙的压力分布以及土层极限平衡的受力状况进行了深入的分析,并对理想状态下的土层发生剪切滑移进行了定量分析,确定了主动、被动土压力作用下的滑移面形态与位置,以此为基础建立了月壤—铲斗挖掘力模型;由于本文中采用仿生学设计的铲体为曲面构型,与常用的平面型铲体不同,本文提出一种基于微元法的隔离体极限平衡受力分析方法,通过对微元模型的二维受力分析分别建立了主动、被动土压状态下的土压力合力计算模型,进而获得铲斗的挖掘力峰值与月壤参数的函数关系;通过整合月壤—铲斗土力学研究成果,提出一种基于Newton加速迭代算法的月壤参数辨识算法,以实现对月壤重度(或密度)、土层内摩擦角及铲斗—土层摩擦角的参数辨识。本文建立了包含挖掘速率效应、铲斗边缘效应以及铲斗形态效应的挖掘阻力数学模型,并对关键的挖掘参数进行了定量分析,分别获得了挖掘角度、挖掘速率对挖掘阻力的影响规律。同时对挖掘过程中入铲、挖掘和装载三个阶段进行分析,在此基础上借鉴分层优化的思想提出了外层以挖掘能耗为指标的铲斗末端挖掘路径优化、内层以铲斗挖掘合外力为指标的挖掘角度优化的双层优化结构,并应用遗传算法进行求解。进而,以挖掘过程中的系统挖掘时间、能耗为综合考量指标建立了目标函数,通过变量转化将连续型非线性优化问题转化为一个标准的离散型凸优化问题并加以求解,最终获得了基于时间—能耗最优的机械臂挖掘轨迹。本文利用实验室现有设备搭建了实验平台,进行了末端采样器功能性验证实验、月壤参数辨识实验和挖掘轨迹优化实验。从而验证了末端采样器、参数辨识算法以及挖掘轨迹优化方法的可行性和可靠性。
【Abstract】 Exploring the unknown domain, such as universe, might be the eternalmotivation of human advancement. United States, the former Soviet Union,Russia and etc. have done some exploration and sampling analysis ofextraterrestrial planets. As a budding rising space powers, China, is also planning“circling”,“landing” and “returning” as the three-stage of the lunar explorationprogram. In the exploration process, energy comsumption and plantary soilparameter identification are two major focus. With the support of “surfacesampling mechanism and plantary soil parameters identification” for the ChineseLunar Exploration Plan Stage III, this disserstion aims to design a set of robotarm end-effecot which can complete the plantary exploration, sampling,acquisition and return tasks, develop a reliable plantary parameters identificationalgorithm and optimize the excavation strategy which based on the parametersidentification results to reduce the energy consumption.Due to the uncertainty of the lunar surface and unpredictability of theparticle movement in the low-gravity environment, the simulation analysis wascarried out via EDEM software aiming at the fluidity of lunar soil particle underlow-gravity environment. The influence rule of inner frction angle and soil-toolfriction angle on the fluidity of lunar particle was acquired as well as the theoptimal percussion frequency and direction of the percussion mechanism on theend-effector. Meanwhile, applying the bionic design theory, the shape of diggingshovel on the end-effector was optimized to ensure the lowest excavationresistance. Based on these studies, a highly integrated end-effector with multi-funtions, such as acquisition, cathing, locking and percussion, was designed.This paper took the soil mechanics of soil-bucket as the foundation oftheoretical analysis. Based on this theory, this paper analyzed the determinationmethod of the plantary soil strength failure, the force stauts of retaining wallpressure distribution and soil limit equilibrium and quantitative analyzed the soilshear slip under the ideal conditions to make sure the location of slip surfaceunder the initiative and passive soil pressure, in order to establish the soil-bucketdigging force model. Different from the common bucket, the shape of this bionic designed bucket is curved. Consequently, a method which based on the micro-element method to analyze the isolated body limit equilibrium force was appliedin this paper. Through the two-dimensional stress analysis of the micro-elementmodel, the earth pressure force calculation model under the initiative and passiveearth pressure conditions were established to obtain the function between thepeak digging resistive force and the plantary soil parameters. Based on the aboveoutcomes, a Newton Accelerate the iterative algorithm based plantary soilparameters identification algorithm was introduced to achieve the identificationof the soil density, the soil internal friction angle and the friction angle betweensoil and bucket.This paper established a mathematical model which contains the excavationvelocity effect, the bucket edge effect and bucket shape effect, quantitativelyanalyzed the key excavation parameters, obtained the effect of angle of attackand excavation velocity on excavation resistance. Also, based on the analysis ofthe three stage of excavation process, penetrate, drag and curl, this paperintroduced the thought of hierarchical optimization that contains the outer andinner layers. The outerlayer set the excavation energy consumption as target tooptimize the excavation path. The inner layer set the digging force as target tooptimize the angle of attack. The double layers were calculated by the geneticalgorithm. Then, an objective function was established which takes theexcavation time and energy consumption as the comprehensive specification.Continuous nonlinear optimization problem was transformed to a standarddiscrete convex optimization problem by the variable transformation to calculate.After this, the optimal excavation trajectory that based on the time-energyconsumption could be achieved.Finally, this paper used the existing equipments of the laboratory to bulid upa testbed to verify the feasibility and reliability of the end-effector function, thesoil parameters identification algorithm and optimal excavation strategy.