文章快速检索 高级检索
  力学学报  2016, Vol. 48 Issue (4): 972-975  DOI: 10.6052/0459-1879-15-392
0

页岩气专题论文

引用本文 [复制中英文]

陈菊, 吴惠彬, 梅凤翔. 有多余坐标完整系统的自由运动[J]. 力学学报, 2016, 48(4): 972-975. DOI: 10.6052/0459-1879-15-392.
[复制中文]
Chen Ju , Wu Huibin , Mei Fengxiang . FREE MOTION OF HOLONOMIC SYSTEM WITH REDUNDANT COORDINATES[J]. Chinese Journal of Ship Research, 2016, 48(4): 972-975. DOI: 10.6052/0459-1879-15-392.
[复制英文]

基金项目

国家自然科学基金资助项目(10932002, 11272050, 11572034).

通讯作者

梅凤翔, 教授, 主要研究方向:分析力学.E-mail:meifx@bit.edu.cn

文章历史

2015-10-27 收稿
2016-06-01 录用
2016-06-06网络版发表
有多余坐标完整系统的自由运动
陈菊1, 吴惠彬2, 梅凤翔1     
1. 北京理工大学宇航学院, 北京 100081 ;
2. 北京理工大学数学学院, 北京 100081
摘要: 对于完整力学系统,若选取的参数不是完全独立的,则称为有多余坐标的完整系统. 由于完整力学系统的第二类Lagrange 方程中没有约束力,故为研究完整力学系统的约束力,需采用有多余坐标的带乘子的Lagrange方程或第一类Lagrange 方程. 一些动力学问题要求约束力不能为零,而另一些问题要求约束力很小. 如果约束力为零,则称为系统的自由运动问题. 本文提出并研究了有多余坐标完整系统的自由运动问题. 为研究系统的自由运动,首先,由d'Alembert-Lagrange 原理, 利用Lagrange 乘子法建立有多余坐标完整系统的运动微分方程;其次,由多余坐标完整系统的运动方程和约束方程建立乘子满足的代数方程并得到约束力的表达式;最后,由约束系统自由运动的定义,令所有乘子为零,得到系统实现自由运动的条件. 这些条件的个数等于约束方程的个数,它们依赖于系统的动能、广义力和约束方程,给出其中任意两个条件,均可以得到实现自由运动时对另一个条件的限制. 即当给定动能和约束方程,这些条件会给出实现自由运动时广义力之间的关系. 当给定动能和广义力,这些条件会给出实现自由运动时对约束方程的限制. 当给定广义力和约束方程,这些条件会给出实现自由运动时对动能的限制. 文末,举例并说明方法和结果的应用.
关键词: 多余坐标    完整系统    约束力    自由运动    
0 引言

完整系统的第二类Lagrange方程是受约束质点系且不包含约束力的运动微分方程[1].故为研究有多余坐标完整约束系统,需用第一类Lagrange方程或有多余坐标带乘子形式的Lagrange方程[2-4].对于完整力学系统通常采用第二类Lagrange方程来组建系统的运动微分方程,其中的坐标是彼此独立的.但是,有多余坐标的完整系统力学不仅在运动学描述上有重要意义,而且在诸多动力学问题中,如四连杆机构的运动学描述[5-7],工程中振动仪器[7],多体系统动力学[8-10]等有重要意义.在多体系统的动力学中更多地采用微分---代数方程,即有多余坐标完整系统的方程,以便更好地实施计算[11-13].故有时选多余坐标反而会带来方便.

有多余坐标完整系统的自由运动是指系统的约束力为零的运动. 关于约束系统的自由运动问题取得了一些成果.文献[14]给出了非完整系统的自由运动,研究了Chaplygin雪橇的自由运动,并讨论了有外力时实现非完整系统自由运动的可能性,具有一定的实际意义.2005年,Yushkov利用Maggi方程实现了Chaplygin雪橇问题的自由运动,并研究了在有主动力时非完整系统自由运动的可能性[15-16]. 在此基础之上,推广并研究了非完整力学的可控运动[17].关于自由运动问题,一些学者从不同方面进行了讨论研究[18-19].本文主要提出并研究有多余坐标完整系统的一类特殊运动,即自由运动,给出实现自由运动的条件.

1 系统的运动微分方程

设系统的位形由$r$个广义坐标$q_d \left( {d = 1,2,\cdots,r} \right)$来确定. 由于某些要求,需选$g$个多余坐标$q_{r + 1}$,$q_{r + 2}$,$\cdots,q_{r + g} \left( {g = n - r} \right)$,并有$g$个双面理想完整约束

$ f_\beta \left( {q_s ,t} \right) = 0 \quad \left( {\beta = 1,2,\cdots ,g ; s = 1,2,\cdots,n} \right)$ (1)

d'Alembert-Lagrange原理有形式

$ \left( {\dfrac {d }{d }\dfrac {\partial T}{\partial \dot {q}_s } - \dfrac {\partial T}{\partial q_s } - Q_s } \right)\delta q_s = 0 \quad \left( {s = 1,2,\cdots ,n} \right) $ (2)

这里及以后,同一项中相同指标表示求和. 式(1)加在虚位移$\delta q_s $上的限制为

$ \dfrac {\partial f_\beta }{\partial q_s }\delta q_s = 0 $ (3)

由式(2)和式(3),利用Lagrange乘子法,得到方程

$ \dfrac {d }{d }\dfrac {\partial T}{\partial \dot {q}_s } - \dfrac {\partial T}{\partial q_s } = Q_s + \lambda _\beta \dfrac {\partial f_\beta }{\partial q_s } \\ \left( {\beta = 1,2,\cdots ,g ; \ s = 1,2,\cdots,n} \right) $ (4)

式(4)右端带乘子的项$\lambda _\beta $代表约束力. 由式(4)和式(1)可求系统的运动,并且可求约束力.

2 约束力的确定

将系统动能写成

$ T = T_2 + T_1 + T_0 $ (5)

其中

$ T_2 = \dfrac {1}{2}A_{sk} \dot {q}_s \dot {q}_k ,T_1 = B_s \dot {q}_s $ (6)

则有

$ \dfrac {d }{d t }\dfrac {\partial T_2 }{\partial \dot {q}_s } - \dfrac {\partial T_2 }{\partial q_s } = \\ A_{sk} \ddot {q}_k + \left( {\dfrac {\partial A_{sk} }{\partial q_m } - \dfrac {1}{2}\dfrac {\partial A_{km} }{\partial q_s }} \right)\dot {q}_k \dot {q}_m + \dfrac {\partial A_{ks} }{\partial t}\dot {q}_k =\\ A_{sk} \ddot {q}_k + \left[{k,m;s} \right] \dot {q}_k \dot {q}_m + \dfrac {\partial A_{ks} }{\partial t}\dot {q}_k $ (7)

其中

$ \left[{k,m;s} \right] = \dfrac {1}{2}\left( {\dfrac {\partial A_{ks} }{\partial q_m } + \dfrac {\partial A_{ms} }{\partial q_k } - \dfrac {\partial A_{km} }{\partial q_s }} \right) $ (8)

为系数矩阵$\left( {A_{sk} } \right)$的第一类Christoffel记号. 又

$\dfrac {d }{d t }\dfrac {\partial T_1 }{\partial \dot {q}_s } - \dfrac {\partial T_1 }{\partial q_s } = \dfrac {\partial B_s }{\partial t} - \left( {\dfrac {\partial B_k }{\partial q_s } - \dfrac {\partial B_s }{\partial q_k }} \right)\dot {q}_k $ (9)
$\dfrac {d }{d t }\dfrac {\partial T_0 }{\partial \dot {q}_s } - \dfrac {\partial T_0 }{\partial q_s } = - \dfrac {\partial T_0 }{\partial q_s } $ (10)

将式(7)、 式(9)和式(10)代入式(4),得到

$ A_{sk} \ddot {q}_k + \left[{k,m; s} \right] \dot {q}_k \dot {q}_m = \\ \left( {\dfrac {\partial B_k }{\partial q_s } - \dfrac {\partial B_s }{\partial q_k }} \right)\dot {q}_k + Q_s - \dfrac {\partial B_s }{\partial t} + \dfrac {\partial T_0 }{\partial q_s } -\\ \dfrac {\partial A_{sk} }{\partial t}\dot {q}_k + \lambda _\beta \dfrac {\partial f_\beta }{\partial q_s } \quad \left( {s = 1,2,\cdots ,n} \right) $ (11)

${\rm{det}}\left( {{A_{sk}}} \right) \ne 0$,故可由式(11)求出所有广义加速度

$ \ddot {q}_l = A^{ls}\left[{ - \left[{k,m; s} \right] \dot {q}_k \dot {q}_m + \left( {\dfrac {\partial B_k }{\partial q_s } - \dfrac {\partial B_s }{\partial q_k }} \right)\dot {q}_k + Q_s } \right. -$

$ \left. { \dfrac {\partial B_s }{\partial t} + \dfrac {\partial T_0 }{\partial q_s } - \dfrac {\partial A_{sk} }{\partial t}\dot {q}_k + \lambda _\beta \dfrac {\partial f_\beta }{\partial q_s }} \right] $ (12)

其中

$ A^{ls}A_{sk} = \delta _k^l $ (13)

为求得约束力,将式(1)对$t$求两次导数,得

$ \dfrac {\partial ^2f_\beta }{\partial t^2} + 2\dfrac {\partial ^2f_\beta }{\partial t\partial q_s }\dot {q}_s + \dfrac {\partial ^2f_\beta }{\partial q_s \partial q_k }\dot {q}_s \dot {q}_k + \dfrac {\partial f_\beta }{\partial q_s }\ddot {q}_s = 0 $ (14)

将式(12)代入式(14),消去广义加速度,得到

$\eqalign{ & {{{\partial ^2}{f_\beta }} \over {\partial {t^2}}} + 2{{{\partial ^2}{f_\beta }} \over {\partial t\partial {q_s}}}{{\dot q}_s} + {{{\partial ^2}{f_\beta }} \over {\partial {q_s}\partial {q_k}}}{{\dot q}_s}{{\dot q}_k} + {{\partial {f_\beta }} \over {\partial {q_l}}}{A^{ls}} \cr & \left[ { - \left[ {k,m;s} \right]{{\dot q}_k}{{\dot q}_m} + \left( {{{\partial {B_k}} \over {\partial {q_s}}} - {{\partial {B_s}} \over {\partial {q_k}}}} \right){{\dot q}_k} + {Q_s}} \right. - \left. {{{\partial {B_s}} \over {\partial t}} + {{\partial {T_0}} \over {\partial {q_s}}} - {{\partial {A_{sk}}} \over {\partial t}}{{\dot q}_k} + {\lambda _\gamma }{{\partial {f_\gamma }} \over {\partial {q_s}}}} \right] \cr & = 0\left( {\beta = 1,2, \cdots ,g} \right) \cr} $ (15)

${\rm{det}}\left( {{{\partial {f_\beta }} \over {\partial {q_l}}}{A^{ls}}{{\partial {f_\gamma }} \over {\partial {q_s}}}} \right) \ne 0$ (16)

时,可由式(15)求出所有$\lambda _\gamma \left( {\gamma = 1,2,\cdots ,g}\right)$,进而可求出约束力

$ \varLambda _s = \lambda _\gamma \dfrac {\partial f_\gamma }{\partial q_s } \quad \left( {\gamma = 1,2,\cdots ,g;s = 1,2,\cdots ,n} \right) $ (17)
3 自由运动的条件

系统的自由运动是指约束力为零的运动,将$\lambda _\gamma = 0 \left( {\gamma =1,2,\cdots ,g} \right)$代入式(15),得到

$ \dfrac {\partial ^2f_\beta }{\partial t^2} + 2\dfrac {\partial ^2f_\beta }{\partial t\partial q_s }\dot {q}_s + \dfrac {\partial ^2f_\beta }{\partial q_k \partial q_s }\dot {q}_k \dot {q}_s +\\ \dfrac {\partial f_\beta }{\partial q_l }A^{ls}\Bigg [- \left[{k,m;s} \right] \dot {q}_k \dot {q}_m + Q_s - \\ \dfrac {\partial B_s }{\partial t} + \dfrac {\partial T_0 }{\partial q_s } - \dfrac {\partial A_{sk} }{\partial t}\dot {q}_k \Bigg] = 0 \\ \left( {\beta = 1,2,\cdots ,g} \right) $ (18)

这就是系统发生自由运动的条件. 特别地,如果

$ \left.\!\!\begin{array}{l} B_s = 0 \left( {s = 1,2,\cdots ,n} \right) \\ T_0 = 0 \\ \dfrac {\partial A_{ks} }{\partial t} = 0 \left( {k,s = 1,2,\cdots ,n} \right) \end{array} \right\} $ (19)

则式(18)成为

$ \dfrac {\partial ^2f_\beta }{\partial t^2} + 2\dfrac {\partial ^2f_\beta }{\partial t\partial q_s }\dot {q}_s + \dfrac {\partial ^2f_\beta }{\partial q_s \partial q_k }\dot {q}_s \dot {q}_k + \\ \dfrac {\partial f_\beta }{\partial q_l }A^{ls}\left( { - \left[{k,m; s} \right] \dot {q}_k \dot {q}_m + Q_s } \right) = 0 $ (20)

进而,如果$A_{sk} = {\rm const.} \left( {s,k = 1,2,\cdots ,n}\right)$,则式(20)成为

$ \dfrac {\partial ^2f_\beta }{\partial t^2} + 2\dfrac {\partial ^2f_\beta }{\partial t\partial q_s }\dot {q}_s + \dfrac {\partial ^2f_\beta }{\partial q_s \partial q_k }\dot {q}_s \dot {q}_k + \\ \dfrac {\partial f_\beta }{\partial q_l }A^{ls}Q_s = 0 \quad \left( {\beta = 1,2,\cdots ,g} \right) $ (21)

当系统发生自由运动时,式(4)成为

$ \dfrac {d }{d }\dfrac {\partial T}{\partial \dot {q}_s } - \dfrac {\partial T}{\partial q_s } = Q_s \quad \left( {s = 1,2,\cdots ,n} \right) $ (22)
4 算例

例1 系统的动能和约束分别为

$ T = \dfrac {1}{2}\left( {\dot {q}_1^2 + \dot {q}_2^2 + \dot {q}_3^2 } \right) ,f = tq_1 + 2q_2 - 3q_3 = 0 $ (23)

其中的量已无量纲化,试求实现自由运动所应施加的主动力之间的关系.

解:式(21)给出

$ 2\dot {q}_1 + tQ_1 + 2Q_2 - 3Q_3 = 0 $ (24)

如不施加主动力,即$Q_1 = Q_2 = Q_3 = 0$,则不能实现自由运动. 式(24)给出主动力应满足的条件,例如,取$Q_1 = 0$,$Q_2 = - \dot {q}_1$,$Q_3 = 0$,则可实现自由运动,此时约束力为零. 当然,还有其他选择.

例2 系统动能和约束分别为

$\left. {\matrix{ {T = {1 \over 2}\left[ {\dot q_1^2\left( {2 + \sin t} \right) + \dot q_2^2 + \dot q_3^2} \right]} \cr {f = {t^2}{q_1} - {q_2} - {q_3} = 0} \cr } } \right\}$ (25)

其中量已无量纲化,试求系统实现自由运动的条件.

解:由$T$表达式知

$ A_{11} = 2 + \sin t ,A_{22} = 1 ,A_{33} = 1 $

于是有

$ A_{11} = \dfrac {1}{2 + \sin t} ,A_{22} = 1 ,A_{33} = 1 $

以及

$ \left[{k,m;s} \right] = 0 \quad \left( {k,m,s = 1,2} \right)$

式(18)给出

$ 2q_1 + 4t\dot {q}_1 + \dfrac {t^2}{2 + \sin t}\left( {Q_1 - \dot {q}_1 \cos t} \right) - \\ Q_2 - Q_3 = 0 $ (26)

这就是系统实现自由运动对主动力$Q_1 ,Q_2 ,Q_3 $ 的限制. 显然,不施加主动力是不能实现自由运动的.

5 结论

本文研究了有多余坐标完整力学系统的自由运动,得到实现自由运动的条件. 如果需要自由运动,必须满足这些条件;如果不需要自由运动,必须避开这些条件. 同时,正如文献[15]所指出的,如果系统的运动对自由运动有小的偏离,那么就有小的约束力,可将其作为干扰来研究以实现规划运动.

参考文献
[1] Lagrange JL. Mécanique Analytique. Paris: Jaques Gabay, 2006 (0)
[2] 梅凤翔. 分析力学[M]. 北京: 北京理工大学出版社, 2013 . ( Mei Fengxiang. Analytical Mechanics[M]. Beijing: Beijing Institute of Techanology Press, 2013 . (in Chinese) ) (0)
[3] 朱照宣, 周起钊, 殷金生. 理论力学[M]. 北京: 北京大学出版社, 1982 . ( Zhu Zhaoxuan, Zhou Qizhao, Yin Jinsheng. Theoretical Mechanics[M]. Beijing: Peking University Press, 1982 . (in Chinese) ) (0)
[4] Hurtado JE, Sinclair AJ. Lagrangian mechanics of overparameterized systems[J]. Nonlinear Dynamics,2011, 66 : 201-212. DOI: 10.1007/s11071-010-9921-1. (0)
[5] Лypъе A И. Aналитическая Механика. Москва: ФИЗМAТГИЗ,1961 (0)
[6] 陈滨. 分析力学[M]. 北京: 北京大学出版社, 2012 . ( Chen Bin. Analytical Mechanics[M]. 2nd edn. Beijing: Peking University Press, 2012 . (in Chinese) ) (0)
[7] 梅凤翔, 刘桂林. 分析力学基础[M]. 西安: 西安交通大学出版社, 1987 . ( Mei Fengxiang, Liu Guilin. The Foundations of Analytical Mechanics[M]. Xi’an: Xi’an Jiaotong University Press, 1987 . (in Chinese) ) (0)
[8] Brogliato B, Goeleven D. Singular mass matrix and redundant constraints in unilaterally constrained Lagrangian and Hamiltonian systems[J]. Multibody System Dynamics,2015, 35 : 39-61. DOI: 10.1007/s11044-014-9437-4. (0)
[9] Wojtyra M, Fraczek J. Solvability of reactions in rigid multibody systems with redundant nonholonomic constraints[J]. Multibody System Dynamics,2013, 30 : 153-171. DOI: 10.1007/s11044-013-9352-0. (0)
[10] Whittaker ET. A Treatise on the Analytical Dynamics of Particles and Rigid Bodies, 4th edn[M]. Cambridge: Cambridge University Press, 1970 . (0)
[11] 刘延柱. 高等动力学[M]. 北京: 高等教育出版社, 2001 . ( Liu Yanzhu. Advanced Dynamics[M]. Beijing: Higher Education Press, 2001 . (in Chinese) ) (0)
[12] 尚玫. 高等动力学[M]. 北京: 机械工业出版社, 2013 . ( Shang Mei. Advanced Dynamics[M]. Beijing: China Mechine Press, 2013 . (in Chinese) ) (0)
[13] Jungnickel U. Differential-algebraic equations in Riemannian spaces and applications to multibody system dynamics[J]. ZAMM,1994, 74 : 409-415. DOI: 10.1002/(ISSN)1521-4001. (0)
[14] 梅凤翔. 非完整系统的自由运动和非完整性的消失[J]. 力学学报,1994, 26 (6) : 470-476. ( Mei Fengxiang. The freedom motion of nonholonomic system and disappearance of nonholonomic property[J]. Chinese Journal of Theoretical and Mechanics,1994, 26 (6) : 470-476. (in Chinese) ) (0)
[15] Зегжда С А,Солтаханов ШХ,Юшков МП. Уравнения Движения Неголономых Систем и Вариационные Принципы Механики. Новый Класс Задач Управления. Москва: ФИЗМАТЛИТ, 2005 (0)
[16] 杰格日达CA,索尔塔哈诺夫X H, 尤士科夫M P. 梅凤翔译. 非完整系统的运动方程和力学的变分原理,新一类的控制问题. 北京:北京理工大学出版社, 2007 ( Zegzda SA, Soltakhanov Sh Kh, Yushkov MP. Equations of Motion of Nonholonomic Systems and Variational Principle of Mechanics. Beijing: Beijing Institute of Techanology Press, 2007 (in Chinese) ) (0)
[17] Терmычныŭ Даурu B. Ю. Адаптивная механика [М]: Наука, 1998 (0)
[18] 丁光涛. 状态空间中约束系统的运动方程. 动力学与控制学报[M]. 2015 : 250 -255. ( Ding Guangtao. The equation motion of constrained systems in state space. Journal of Dynamics and Control[M]. 2015 : 250 -255. (in Chinese) ) (0)
[19] 郭仲衡. 从冰橇问题谈起[M]. 徐州: 中国矿业大 学出版社, 1993 : 451 -453. ( Guo Zhongheng. From the problem of sledge[M]. The Modern Mathematics and Mechanics. Xuzhou: China Mining University Press, 1993 : 451 -453. (in Chinese) ) (0)
FREE MOTION OF HOLONOMIC SYSTEM WITH REDUNDANT COORDINATES
Chen Ju1, Wu Huibin2, Mei Fengxiang1     
1. School of Aerospace, Beijing Institute of Technology, Beijing 100081, China ;
2. School of Mathematics, Beijing Institute of Technology, Beijing 100081, China
Abstract: If the parameters are not completely independent for holonomic systems, it is called holonomic systems with redundant coordinates. In order to study the forces of constraints for holonomic systems, we use the Lagrange equations with multiplicators of redundant coordinates or the first kind of Lagrange equations. Because there are no forces of constraints in the second kind of Lagrange equations. In some mechanical problems, the forces of constraints should not be equal to zero. In other conditions, the forces of constraints are very tiny. However, if the forces of constraints are all equal to zero, we called the free motion of constraints mechanical systems. This paper presents the free motion of holonomic system with redundant coordinates. At first, the differential equations of motion of the system are established according to d'Alembert-Lagrange principle. Secondly, the form of forces of constraints is determined by using the equations of constraints and the equations of motion. Finally, the condition under which the system has a free motion is obtained. The number of this conditions is equal to the constraints equation's, its depend on the kinetic energy, generalized forces and constraints equations. If the two arbitrary conditions are given, the third one should be obtained when the system becomes free motion. At the end, some examples are given to illustrate the application of the methods and results.
Key words: redundant coordinate    holonomic system    force of constraints    free motion