OSE800: Analytical and Computational Dynamics (Spring, 2013) Lecture 3: D’Alembert’s Principle In this lecture we will learn: • D’Alembert’s Principle • The concept of virtual displacements • Why the reactions do disappear in the final equations of motions • How we can recover reaction forces (and moments, if any) from the Free-Body Diagrams, if needed. 3.1 D’Alembert’s contributions to mechanics D’Alembert’s own words: The principle of equilibrium, together with the principle of the force of inertia, and the principle of compound motion, therefore leads us to the solution of all problems which concern the motion of body in so far as it can be stopped by an impenetrable and immovable obstacle – that is, in general, by another body to which it must necessarily impart motion in order to keep at least a part of its own. From these (three) principle together can easily be deduced the laws of the motion of bodies that collide in any manner whatever, or which affect each other by means of some body placed between them and to which they are attached. 1 Fig. 1. Traité de Dynamique (par Jean D’Alembert), Preleminaire (page XV)) On the force of inertia: the property of bodies of remaining in their state of rest or motion. On time: The nature of time is to run uniformly, and mechanics supposes this uniformity. On equilibrium: Lagrange said, ”D’Alembert had reduced dynamics to statics by means of his principle.” On the center of oscillation: the concept of equilibrium in dynamical systems embodied in D’Alembert’s principle subsumed the work of Jacques Bernoulli (1654-1705), Jacob Herman (16781733) and Leonhard Euler(1707-1783). On the laws of impact: the velocity of two bodies after impact given by v= mu + M U m+M : The principle of momentum conservation! (3.1) On D’Alembert’s paradox: For incompressible and inviscid potentia lflow, the drag force is zero on a body moving with a constant velocity relative to the fluid (due to neglect of the effects of viscosity!). Finally, he derived the wave equation while working on the vibration of musical instruments. 2 3.2 D’Alembert’s Principle Principle of Virtual Work for Statics For statics, we have: N X f i · δri = 0 (3.2) i=1 where {f i } are impressed forces, and {δ} designates the virtual character of the instantaneous variations, as opposed to the differential symbol {d} designating the actual differential of the position vector {ri } taking place during the time interval {dt}. Principle of Virtual Work for Dynamics For dynamics á la d’Alembert: N X (Fi − mi r̈i ) · δri = 0 (3.3) i=1 where {r̈i } is the acceleration of particle mi . Generalized Coordinates Virtual displacements δri may be expressed in terms of the generalized virtual displacements δqk (k = 1, 2, 3, ..., n): δri = n X ∂ri δqk , i = 1, 2, 3, ..., N k=1 ∂qk (3.4) Let’s consider the following example: r = xi + yj + L(cos θi + sin θj) 3 (3.5) Fig. 2. A Spring-Mass System Observe that there are three independent variables, (x, y, θ), yielding the following virtual displacement: δr = ∂r ∂r ∂r δx + δy + δθ ∂x ∂y ∂θ (3.6) = δxi + δyj + L(− sin θi + cos θj) δθ Note that the dimensional unit of (x, y) is in meter(m) whereas that of θ is in radian (non-dimensional quantity), suggesting that the units of generalized coordinates can be different. 3.3 A Spring-Mass-Bar System Revisited: Let’s revisit the formulation of the equations of motion for the spring-mass-bar problem studied in the previous lecture. To this end, first, we must obtain the virtual displacements. Virtual Displacements: Virtual displacement for mass m: δrA0 = δx i 4 (3.7) Virtual displacement for bar M : δrC 0 = δrA0 + δrA0 C 0 L δrA0 C 0 = (cos θi + sin θj) δθ 2 (3.8) Virtual rotation for bar M : Here, one utilizes the angular velocity ω = θ̇k to obtain the virtual rotation δθ = δθk (3.9) δθ = δθx i + δθy j + δθz k (3.10) In general one has Acceleration Vectors: Acceleration vector for the sliding mass m: aA0 = v̇A0 = ẍi Acceleration vectors for the pendulum bar M : aC 0 = aA0 + aA0 C 0 aA0 C 0 = Lθ̈ (cos θi 2 + sin θj) − Lθ̇2 (sin θi 2 − cos θj) Equilibrium equation for the sliding mass m: f 0A = −kxi − mẍi − mgj + NA j + XAC i + YAC j = 0 Equilibrium equation for the pendulum bar M : f 0C = F i − M aC 0 − M gj − XAC i − YAC j = 0 5 Fig. 3. Free-body diagram for mass Fig. 4. Free-body diagram for the bar 2 L θ̈ MC 0 = − M12 k + rC 0 A0 × (−XAC i − YAC j) + rC 0 B 0 × F i = 0 6 Application of D’Alembert’s Principle for the springmass-bar problem 3.4 (Fi − mi r̈i ) · δri =f 0A · δr0A + f 0C · δr0C + M0C · δθC0 = 0 X ⇓ {−kxi − mẍi − mgj + NA j + XAC i + YAC j} · δr0A + {F i − M aC 0 − M gj − XAC i − YAC j} · δr0C M L2 θ̈ {− k + rC 0 B 0 × F i 12 + rC 0 A0 × (−XAC i − YAC j)} · δθ (3.11) Note that f 0A · δr0A 6= 0 and so other two terms! The first term in (3.11) becomes f 0A · δr0A = {−kxi − mẍi − mgj + NA j + XAC i + YAC j} · (δxi) ⇓ 0 0 f A · δrA = {−kx − mẍ + XAC } δx (3.12) Remark: Observe that the reactions forces YAC and NA have played no role in the resulting virtual work! 7 The second term in (3.11) becomes f 0C · δr0C = {F i − M aC 0 − M gj − XAC i − YAC j} · δr0C Lθ̈ Lθ̇2 (cos θi + sin θj) − (sin θi − cos θj)] 2 2 L − M gj − XAC i − YAC j} · {δxi + (cos θi + sin θj) δθ} 2 ⇓ = {F i − M [ẍi + Lθ̈ Lθ̇2 cos θ − sin θ] − XAC } δx 2 2 Lθ̈ Lθ̇2 L + {F − M [ẍ + cos θ − sin θ] − XAC } cos θ δθ 2 2 2 2 Lθ̈ Lθ̇ L + {−M [ sin θ − cos θ] − M g − YAC } sin θδθ 2 2 2 ⇓ f 0C · δr0C = {F − M [ẍ + f 0C · δr0C = {F − M [ẍ + Lθ̇2 Lθ̈ cos θ − sin θ] − XAC } δx 2 2 L cos θ M Lg ML M L2 θ̈ − F− cos θ ẍ − sin θ 2 2 4 2 L L − cos θ XAC − sin θYAC } δθ 2 2 +{ (3.13) In evaluating the third term of (3.11), first, carry out: L (sin θi − cos θj) × F i 2 FL cos θk = 2 rC 0 B 0 × F i = rC 0 A0 × (−XAC i − YAC j) = { L L cos θXAC + sin θYAC }k 2 2 (3.14) so that we obtain M C0 · δθ C0 M L2 θ̈ F L L L = {− + cos θ + [ cos θXAC + sin θYAC ]}δθ 12 2 2 2 (3.15) 8 Substituting (3.12), (3.13) and (3.15) into (3.11), we obtain N X ML ML cos θ θ̈ − sin θ θ̇2 − F } δx 2 2 2 ML ML M gL +{ θ̈ẍ + cos θ ẍ + sin θ − L cos θF }δθ = 0 3 2 2 (3.16) (Fi − mi r̈i ) · δri = {(M + m)ẍ + kx + i=1 3.5 Coupled equations for x and θ via d’Alembert’s principle Since δx and δθ are arbitrary, we obtain: (M + m)ẍ + kx + ML ML cos θ θ̈ − sin θ θ̇2 = F 2 2 (3.17) M L2 ML M gL θ̈ + cos θ ẍ + sin θ = L cos θF 3 2 2 Observations 1. When using Newton’s second law, (3.17) are obtained by eliminating the reaction forces XAC and YAC . 2. On the other hand, one does not need to consider reaction forces in applying d’Alembert’s principle! Only apparent forces including inertia forces need to be considered. This is a major advantage of applying energy principles over Newton’s method. 9