_—< PART I Single-Degree-of-Freedom SystemsEquations of Motion, Problem Statement, and Solution Methods PREVIEW In this opening chapter, the structural dynamics problem is formulated for simple strac- tures that can be idealized as a system with a lumped mass and a massless. supporting Structure. Linearly elastic structures as well as i tic structures subjected to applied dynamic force or earthquake-induced Then four methods for solving the differential equation governin the structure are reviewed briefly. The chapter ends with an « mic response of single-degree-of-freedom systems is 1.1 SIMPLE STRUCTURES imple structures. such as the pergola Lo and the elevated water tank of Fig, 1.1.2. We are imerested in under- standing the vibration of these structures when subjected to a lateral (or horizontal) force at the top or horizontal ground motion due to an carthquake, We call these structures simpie because they can be idealized as a concentrated or Jumped mass a supported by a massless structure with stiffness & in the lateral direction. Such an idealization is appropriate for this pergola with a heavy concrete roof supported by light-sicel-pipe columns, which can be assumed as massless, The concrete raat is very SUiff and the Mexibility of the structure in literal for horizontal) motion is provided ly by the columns, The idealized system is shown in Fig. 1.1.38 with a ‘Supporting the tributary length of the concrete rool, ‘This4 Equasons of Motion, Problem Staiement, and Solution Math Chap. he mass of the roof shown, and its lateral stiffness & is equal to the sum of the stifffresses of individual pipe columns. A similar idealization, shown in Fig. 1.1.3b, is appropriate for the tank when it is full of water. With sloshing of water not possible in a full tank, it is a lumped mass m supported by a relatively light tower that can be assumed as massless. The cantilever tower supporting the water tank provides lateral stiffness k to the structure. For the moment we will assume that the lateral motion of these structures is small in the sense that the supporting structures deform within their linear elastic limit, ‘We shall see later in this chapter that the differential equation governing the lateral displacement u(r) of these idealized structures without any external excitation—applied force or ground motion—is mii +ku= 0 (LL.D) us ii denotes the velocity of the mass and 4 its acceleration. The solution of this equation, presented in Chapter 2, will ‘show that if the mass of the idealized systems of Fij Jis displaced through some initial displacement (0), then released and permitted to vibrate freely, the sfructure will oscillate ‘of vibrate back and forth about its initial equilibrium position, Ax shown in Fig, 1.1.3e, the same maximum displacement occurs oscillation after oscillation; these oscillations con- tinue forever and these idealized systems would never come to rest. This is unrealistic,Figure 1.1.2 This reinforced-concrete tank on a 40-ft-tall siigle concrete cohen, located meas the Vabdivia Airport, was undamaged by the Chilean earthquakes ‘of May 1960, When the tank is fall of Mo ‘water, the structure can be analyzed as a ssingle-degree-of freedom system. (From K. V. Steinbrugge Collection. courtesy of the Eanhquake Enginoering Research Center, University of California at Berkeley.) Suggests that if the roof of the pergola or the top of the water tank were Pulled laterally by a rope and the rope were suddenly cut, the structure would oscillate with ever-decreasing amplitude and eventually come to rest. Such experiments were performed on laboratory models of one-story frames. and measured records of their free vibration Rigid slab - — Ss Tributary ve ft > 0) length fon ~@—« s a | - Massless columns, Masses tower (a b) eo ; q Figure 1.1.3 (3) Idealized pergola, (h) idealized water tank: 4) initial displacement, iaNAAAHNVAUUUTUVHTTTTTUUUE 2 Bie = Wil J ’ 52 Bingle-Degroe-of-Freedom System ‘The process by which vibration steadily diminishes in amplitude is called In damping the kinetic energy and strain every of the vibrating system are dissipated by ‘Yariows mechanisms that we shall memion Laer. For the moment, we simply recognize ‘that an energy-dlissipating mechanism should be included in the structural idealization in order to incorporate the feature of decaying motion observed during free vibwation tests of a structure. ‘The most commonly used damping element is the viscous damper, in part because it is the simplest to deal with mathematically. In Chapters 2 and 3 we introduce other energy-dissipating mechanisms. 1.2 SINGLE-DEGREE-OF-FREEDOM SYSTEM The system considered is shown schematically in Fig, 1.2.1, It consists of a mass m con- centrated at the roof level, a massless frame that provides stiffness 1 the system, and a viscous damper (also known as a dashpot) that dissipates vibyratio ‘The beam and columns are assumed to be inextensible axis This system may be considered as an idealization of 4 one-story structure. Each Structural member (beam, column, wall, ete.) of the actual structure contributes 10 the inertial (mass), elastic (stif flexibility), and energy dissipation (damping) properties of the structure. In the idealized system, however, cach of these properties is concentrated in three separate, pure components ponent, stiffness compement, and damping component ‘The number of independent displ tions of all the n ‘of freedom (DOFs) for the stiffness peopertic al energy of the system. red to define the displaced posi- alled the number grees typically necessary to det « DOFs necessary fo ied ts y tse formulated with e the lateral stiff- ‘one ses relative one-story {ra tion. The st three DOFs—lateral displacement and tw ness of the frame (see Section 1.3). In 0 detern ast, the structure s concentrated at one gree-af freedom (SDP d with Mass i L211 Singhe-lnpree-cf-freedaen system: (ab applied farce it); (ts) eartbaqualke- ed ground enn,Equations of Mation, Problem Statement, and Solution Methods = Chap. 1 “Two types of dynamic excitation will be comsidered: (i pexternal force p(1) in the lat- ‘eral direction (Fig. 1.2.La), and (2) earthaake-induced ground motion p(t) (Fig. 1.2 1b), Tn both cases w denotes the relative displacement between the mass and the base of the ssiracture 1.3 FORCE-DISPLACEMENT RELATION Consider the system shown in Fig. 1.3.Ja with no dynamic excitation subjected to an externally applied static force fs along the DOF was shown. The internal force resist~ ing the displacement w is equal and opposite to the external force fs (Fig. 1.3.1b). I is desired to determine the relationship between the force fs and the relative displace- meat u associated with deformations inthe structure, This force-displacement relation would be linear at small deformations but would become nonlinear at larger deforma- Wioas (Fig. 1.3.le%, both nonlinear and linear relations are considered (Fig. 1.3.16 and d). To determine the relationship between fc and w is a standard problem in static stnac- tural analysis, and we assume that the reader is familiar with such analyses. Thus the presentation here is brief and limited to those aspects that are essential.Sec. 13 Force—Displacement Relation 1.3.1 Linearly Elastic Systems For a linear system the relationship between the Lateral force fs and resulting deformation w is linear, that is, fomku a3 ) ‘where & isthe lateral stiffness of the system: its units are fosce/length, Implicit in Eq, (1.3.1) is the assumption that the linear /,—w relationship determined for small deformations of ‘the structure is also valid for larger deformations, This lingar relationship implies that fi isa single-valued function of w (i.e., the loading and unloading curves are identical), Such aa system is said t be elastic: hence we use the term linearly elastic system to emphasiae both properties. Consider the frame of Fig. 1.3.20 with bay width £, height A, elastic modulus E, and second moment of the cross-section: of moment of inertia)’ about the axis of bending = fs and J. for the beam and columns, respectively; the columns are clamped (or fixed) at the base. The lateral stiffness of the frame can readily be determined for the two ‘extreme cases: If the beam is rigid [i.c., flexural rigidity EJ, = 00 (Fig. 1.3.2b)), sy ~ IE), El. 5 le ‘On the other hi aay Observe that for the two extr lateral stiffness of the frame is independent of £., the The lateral stiffness « can be calculated by star trix of the frame is forn beam lengt alistic stiffness of the beam ral analysis. The stiff vlaied with respect to three DOFs: the Ei, Fl,=0 tb) te Figure 14.2 ‘a this book the preferred term for Fix sevond momen of wre instead of the coenmonly used meme ey inertia, the lator will be reserved foe defining inertial efccis aes inied with sedationsl mation of rid bauies,Equations of Motion, Problem Statement, and Solution Methods = Chap 4 amd the rotations of the two beam-columa joints (Fig. 1.3.2a). By static condensation gp slimination of Uhe rotational DOFs, the lateral foree-displacement relation of Eq. (1.3.4) ig determined. Applying this procedure to a frame with £ = 2h and Ely = El, its latera) stiffness is obtained (see Example 1.1): 96 El, ta (134) The lateral stiffness of the frame can be computed similarly for any values of fy, and 4, using the stiffness coeflicients for a uniform flexural element presented in Appendix |, If shear deformations in clements are neglected, the result can be written in the form ers 35) where = 1/41, is the beam-ro-column stiffixess ratio (to be elaborated in Section 18.1.1), For p = 0,00, and }. Eg, (1.3.5) reduces to the results of Eqs. (1.3.3), (1.3. respectively, The lateral stiffness is plotted as a function of p in Fig. | a factor of 4 as p increases from zero to infinity. m4 EAGEL SIP) 1092 6407 tot Pig “a0! 1 e Figure 13.3 ‘Variation of lateral stiffness, £, with Iheam-to-columm stiffness ratio, p. Example 11 Calculate the lateral stiffness for the frame shown in Fig. Ella, assuming the elements to be axially rigid. .os eee metic’ locations and formation of plastic hinges to obtain the Sec. 1.3 Force-Displacoment Relation Ww ‘Solution This structure can be analycod by any'oF the standard methods, inchiding moment slistributins, Hore we use the d fof stiffness influcace eefficients ti solve the probiem, ‘The systern has the three DOFs shown in Fig. E1.1a. Tosobtain the first columa of the 353 stiffness maurix, we imparse unit si 1. with wey = yO. The fares 4) roquired 10 miaindain this deflected shewn in Fig, AIR, These are deterred using the stiffness coefficients for a ural clement presented in Appendix 1, The elements 4, in the second cobun are determined by impasing a) = | ‘with wy) = ary mm 0; see Fig. El te stiffisess matrix can be dete “Thais the 8 = 3 saith ‘writtca, Fos a frase w a [ss “ eye] (?] ; From the second and thi sms of Laveral by Thus This procedure to elimin 1.3.2 Inelastic Systems the force—detorw Determined by ¢ nent unde mms is shown fger amplitudes of deform nd the un differ from the initial loading branch; such a system is suid to be ithe force fs correspond ory of the deformat This implies that valued and depen asing (positive velocity) force can be expressed as fy = fot, a) (1.3.6) ‘The force-deformation relation for the idealized one-story frame (Fig, 1.3.14) deforming ‘into the inelastic range can be determined in one of two w ys. One approach is methods of nonlinear static structural analysis, For example, in analyzing a steel structure with an assumed stress—sirain Law, the analysis keeps track of the initiation and spreading ial boasting. use= @ Equations of Motion, Probiem Statement, and Solution Methods = Chap y Figure 134 Foece-deformation relation for a structural sizel component. (From HL. Krawinkiler, V. ¥, Bertera, and EF. Popov, “Inelastic Behavior of Stee! Beam-io- ‘Column Subassemblages.” Report No. FERC 71-7, University of California, Berkeley, Calif, 1971.) ‘curve (oa) shown in Fig. k3.1c. The unloading (a—c) and reloading (e-a) curves can be ‘computed similarly of can be defined from the initial loading curve using existing hypothe- ses. Another approach is to define the inelastic force—deformation relation as an idealized version of the experimental data, such as in Fig. 1.3.4. ‘We are interested in studying the dynamic response of inelastic systems because ‘many structures are designed with the expectation that they will undergo some cracking. yielding, and damage during intense ground shaking caused by earthquakes.amt friction between the structure itself and nonstructural clement: sch as jo _ It seems impossible wo identify or describe mathematically each of these rey ting mechanisms in un actual building. As a result, the damping in actual structures is usually represented ina highly ideal- ited manner, For many purposes the actual damping ina SDF structure can be idealized satisfactorily by a linear viscous damper or dashpot. ‘The damping coefficient is selected ‘so that the vibrational energy it dissipates is equivalent to the energy dissipated in all the damping mechanisms, combined, present in the actual structure. This idealization is there- fore called equivalent viscous damping, a concept developed further in Chapter Figure 1.4.la shows linear viscous damper subjected toa force fp along the DOF a, ‘The internal force in the damper is equal and opposite to the external force fy (Fig. 1.4.1), force fp is related vo the velocity d across the linear : fo = cit (4) where the constant c is the viscous damping ce it has units of force x time/length, Unlike the stiffness of a structure, the damping coefficient cannot be calculated from the dimensions of the structure and the sizes of the structural clements. This should not be surprising because, as we noted earlier, it is not feasible to identify all the mechanisms that dissipate vibrational energy of actual structures. Thus vibration experiments on actual structures provide the data for evaluating the damping coctiicient. These may be free vibration experiments that bead to data such as those shown 2. 1.1.4; the measured rate ‘at which motion decays in free vibration will provide a basis for evaluating the damping coefficient, as we shall see in Chapter 2. The damping property may also be determined * from forced vibration experimemts, a topic that we study in Chapter 3 ‘The equivalent viscous damper is intended to model the energy dissipation at defor- ‘mation amplitudes within the linear clastic limit structure, Over this range of deformations, the damping coefficient « deter ) esperiments may wary with the deformation amplitude. This nonlinea property is usually not consid- i j indirectly by selecting a value for Jeformation amplitude, usually ity of the darn the damping coefficient that is appropriate for the expec taken as the deformation associated with the linearly elastic limit of the structure, ; External force fn errr nn —— te fo Resisting fowweEquations of Motion, Problem Statement, and Solution Methods — Chap Additional energy is diss tions, Under cyclic Fores force—deformati: duc to inelastic behavior of the structure at ta deformations, this behavior implies formation sop (Fig. Late). The daumy defor deform a hyste g energy dissipated during ‘one deformation cycle betwee n limits ta, is given by the area within the hysteresis loop abeda (Fig. I.3.lc). ‘This energy dissipation is usually mot modeled by a viscous damper, especially if the excitation is earthquake gr 7. Instead, the w accurate appre ‘count for the energy dissipation thro recognize the inelastic relationship between re tion, such as shown in Figs. 1.3.1eand 1.3.4, in solving the ¢ 0 (Chapter 5). Such force-deformation relationsh from experiments on structures or structural components 1 slow F ate-dependent effects. of deform: note in Chi stic behavior is te ps are obtained excludi The usua by the sume viscous dumper ions within the linearly elastic ran; g any energy diss sing fi isto model this damping in the incl that was defined cartier for si 1.5 EQUATION OF MOTION: EXTERNAL FORCE Figure 1.5.14 shows th idealized one-story frame introduced earlier subjected to-an exter: nally applied dynamic force p(s) in the direction of the DOF a. This notation indicates that the force p varies with time /. The resulting displacement of the mass also varies with time; it is denoted by w(t). In Sections 1.5.1 and 1.5.2 we derive the differential equation governing the displacement wis) by two methods using (1) Newton's second law of motion, and (2) dynamic equilibrium. An altemative poim of view for the derivation is presented in Section 1.5.3 1.5.1 Using Newton's Second Law of Motion The forces acting on the mass at some instant of time are show include the external force p(t), the elastic (or inelastic) resisting force fs (Fig. 1.3.1). and the damping resisting force fy (F The external force is wken to be positive in the direction of the x-axis, and the displacement u(/), velocity irr), and acceleration itt) are also: positive in the direction of the x-axis. The elastic and damping forces are shown Fy ™ "1 -- pn pu th an gee gene a oe (ay th) te Figure 1.84 in Fig. 1.5.1b. TheseSec. 1.5 Equation of Motion: External Force 18 Acting in the opposite direction because they ure internal forces that resist the deformation and velocity, respectively, The resultant force along the x-axis isp — fs — fr and Newton's second taw of motion gives P-Ss=fo= mi on mii + fot fs = ply This equation after substituting Eqs, (1.3.1) and (1.4.1) becomes mié +c + ku = pir) 5.2) This is the equation of motion governing the deformation or displacement w(t) of the idealized structure of Fig. 1.5.14, assumed to be linearly elastic, subjected to an external dynamic force p(1). The units of mass are force/acceleration. ‘This derivation can readily be extended to inelastic systems. Equation (1.5.1) is still valid and all that needs to be done is to replace Eq. (1.3.1), restricted to linear systems, by Bg, (1.3.6), valid for inelastic systems. For such systems, therefore, the equation of motion is mii + cit + fst, 1) = pity 41.5.3) 1.5.2 Dynamic Equilibrium Having been trained to think in terms of equilibrium of forces, structural engineers may find D'Alemben’s principle of dynamic equilibrium particularly appealing. This principle is based on the notion of a fictitious inertia force, a force equal to the product of mass fimes its acceleration and acting in a direction opposite to the acceleration. It states that with inertia forces included, a system is in equilibrium at each time instant. Thus a free- body diagram of a moving mass can be drawn, and principles of statics can be used to develop the equation of motion. Figure 1.5.1c is the free-body diagram at time ¢ with the mass replaced by its inertia force, which is shown by a dashed line to distinguish this fictitious force from the real forces. Seiting the sum of all the forces equal to zero gives Eq. (1.5.1b),! which was derived carlier by using Newton's second law of motion, 1.5.3 Stiffness, Damping, and Mass Components In this section the governing equation for the idealized one-story frame is formulated based ‘on an alternative viewpoint. Under the action of external force p(t), the state of the system ne [ aD, velocity 1i(¢), and acceleration ii(r); see Fig. 15.20,I EEE lll = Equations of Motion, Problem Statement, and Solution Mathods Chap, 4 fo mn fe ae +— = = i A eoaeee —— = 1 cei ~E- - ee. me mee simile wie ohn os Displacement « Displacement * Velocity Acceleration @ Velocity a Acceleration a @ (oo ©) ry ampeng conmponent; 4} mass component Piguew LS. {a} System: (bh) The external force fy on the stiffness component is related to the displacement u by Eq. (1.3.1) iif the system is linearly elastic, the external force fp on the damping com- ponent is related to the velocity @ by Eq. (1.4.1), and the external force /, on the mass The external force pir) applied to the complete system may therefore be visualized as distributed among the three com- ponents of the structure, and fs + fm + fi must equal the applied force pir) leading to Eq. (1.5.1b). Although this alternative viewpoint may seem unnecessary for the simple system of Fig. 1.5.2a, it is useful for complex systems (Chapter 9), Example 1.2 A small one-story industrial building. 20 by 30 f1 in plan. is shown in Fig. E12 with moment frames in the north-south direction and braced frames in the east-west direction. The weight of the structure can be idealized as 301b/f? lumped at the roof level. The hoe- ivomtal cross bracing is at the bottom chord of the roof trusses. All colunins are WB x 24 Sections: their second moments of cross-sectional area about the x and y axes are dy = 82.8 int and J, = 18.3 in‘, respectively; for steel, £ = 29,000 ksi. ‘The vertical cross- bbracings are made of I-in.-chameter rods. Formulate the equation governing free vibratioe in (a) the north-south direction and (Ib) the east-west direction.Sec. 1.5 Equation of Motion: External Force Solution The mass humped at the rf ix ee 86.63 Ibesee Fin, = 0.08603 kip-see? fin, Because of the horizontal cross-bracing, the roof can be treated as a rigid diaphragm. ) Novth—somih direction. — Be clamped-clamped column and the 1 is (ee ) i a apd the equation of motion is se of the oof truss, each column behaves as a IH siifTness of the (wo moment frames (Fig. E12) fous = 34.54 kips/in, init + (ky-shw = 0 fa) (b) East-west direction, Braced frames, such as those ally designed as two superiaposed systems: an ondina own in Fig. E.2c, are usu- rigid frame that supports vertical (dead and live) loads, plus a vertical bracing system, generally regarded as a pin-connecied ‘uss that resists lateral forces. ‘Thus the lateral stiffness of a braced Frame can be estimated as the sum the lateral stilfinesses of individual braces, The siiffiness of a brace (Fig. El.2d) is AE/L) cos? 4, This can be derived as fallaws. We stant with the axial force-deformation relation for a brace: Pp ) By statics fy = pcos, and by kinematics u« = 38/cos®. Substinting p = fs/cos@ and § = wcosd in Eq, (b) gives f= time Kimwe = AE cnt ro) For the brace in Fig. EL.2c, cos? = 20/y/'T £=23.3 fi. and O.8575, A = 0,785 in? 108) (0.8573)? = 59.8 kips/in. ‘Although cach frame has two cross-beaces, oaly the one in tension will provide lateral resise tance; the one in compression will buckle at small axial force and wall contribute little to the Lateral stiffness, Considering the two frames ew = 2 59.8 = 119.6 kips/in,‘Sec. 1.6 Mass-Spring-Damper System 1% is 125 (2, The weight af the bridge is idealized as lumped at the deck level; the nat weight af ‘concrete is 150 Ibvit", The weight af the bewts may be neglected. Each hem co 2$-f-tall columns of circular cross le= meruate the in the longitudinal direction, ‘The elastic sdalan, of concrete is = 3000 ksi Solution ‘The weight perunitilength The total weight lumped at the dew at the deck level is (123% 1)150 = 18.45 hips, w = 18.48 x 375 = 6919 kips and the comesponding mass is m= = 24.9 kip-seo™M & The longitudinal stiffness of the bridge is computed assuming the bridge deck to displace rigidly as shown im E ‘column of a bent behaves as a.clamped-clamped column. ‘The longitudinal stiffness provided by each bem is 23 3 :) [ae] = 12,940 kipytt ay ‘Two bents provide a total stiffness of b= 2% Keon = 2 12040 = 25,880 kipwht The equation governing the longitudinal displacement 1 is mii + kw = 0 1.6 MASS-SPRING-DAMPER SYSTEM ‘We have introduced the SDF system by idealizing a one-story structure (Fig. 1.5.1), an Approach that should appeal to structural engineering students. However, the classic SDF sysem is the mass—spring-damper system of Fig. 1.6. 1a. ‘The dynamics of this system is developed in textbooks on mechanical vibration and elementary physics. If we consider ‘the spring and damper to be massless, the mass to be rigid, and all mation to be in the direction of the x-axis, we have an SDF system. Figure 1.6.1b shows the forces acting 7 mi the mass; these include the elastic resisting force, fg = ku, exerted by a linear spring and = damping resisting force, fy = cui, due to a linear viscous damper. 0 ‘then gives Bg. (1.5.1). Alternatively, the same equation= (Equations of Motion, Problem Statement, and Solution Methods = Chap. 1 ( i) te - me € fa mel Friction-free surface“ = Figure 1&1 Mace-xpring-damper system. Example 1 Derive the equation of motion of the weight w suspended from a spring at the free end of a cantilever steel beam shown in Fig. Elda. For stcel, £ = 29,000 ksi. Neglect the mass of the beans and spring. Undleformed position pe ‘Static equilibriam: i re 1 it i * tb) co) @ Figure E14 (a) System; (b) undeformed, deformed, (eee gra wring and cam fence “rtm posiOe ‘Sec. 1.6 Mass-Spring—Damper System a where fs = he (hy and the effective stiffness 4, of the systems remains to be determined, ‘The equation of motion init + kell = w+ play te ‘The displacement if can be expressed as edt a where &y is the static displacement duc to weight w and a is measured from the position of Matic equilibrium, Substituting Eq, (d) in Bg. (a) and noting that (1) 41 = ii because Sy does mot wary with time, and (2) Key = a gives mii + ket = ptr) fe) ‘Observe that this is the same as Eq. (1.5.2) with c= 0 for a spring-mass system oriented in the horizontal direction (Fig. 1.6.1), Also note that the equation of motion (e} governing 1, ; measured from the static equilibrium position, is unaffected by gravity forces. For this reason we usually formulate @ dynamic analysis problem for a linear system with its static equilibrium position as the reference position. The displacement w(t) and as- sociated internal forces in the system will represent the dynamic response of the system, The tolal displacements and forces are obtained by adding the corresponding static quantities to the dynamic response. ‘The effective stiffness ky remains to be determined. I relates the static Force fs to the resulting displacement ii by fs = heli wo where a = baping + Socom rr) ‘where Sicam is the deflection of the right end of the beam and Spiny i the deformation in the Spring. With reference to Fig. El.td. * 5 = kbspeieg = Kream Beam iby In Eg, (g), substitute for @ from Eq. (f) and the 8°s from Eq. (h) to-obtain2 Equations of Motion, Problem Statement, and Solution Methods = Chap 1 ‘Example 1.5 Derive the equation governing the free motion of a simpke pendulum (Fig. E1Sa), which consists of a point mass mi suspended by a light string of length f. ar : fing Figure ES (a) simple pendlum: fay fb) ch) froe-boxdy diagram. Solution Figure E1_Sa shows the displaced position of the pendulum defined by the angle & ‘measured from the vertical position, and Fig. E1.5b shows the free-bedy diagram of the mass, The forces acting are the weight nrg, tension T in the string, and D’ Alembert's fictitious inemtia force fy = ml Equilibrium of the moments of forces about O gives ut78 + mglsin ? = 0 (aa This és a nonlinear ifferential equation governing #, For small rotations, sin# = @ and the equation of motion [Eq. (ai can be rewritten as a8 e+e o ib) Example 16 ‘The system of Fig. E1.6 consists of a weight w attached to a rigid massless bar of length L joined to its support by a rotational spring of stiffness &. Derive the equation of motion. ‘Neglect rotational inertia and assume small deflections, What is the buckling weight?Sec. 1.7 Equation of Motion: Earthquake Excitation 2 Equilibria of the monents about O gives Fatt fis = mpl. sine peti we sin ® w For small retations sin ® ~ # and Eq. (a) can be rewritten as guitu ~ whi = 0 ty ‘Observe that the gravity load reduces the effective stiffitess of the system. [the weight w = 4/L, the eflective stiffness is zere and the system becomes unstable under its own weight, ‘Thus, the buckling load (or weight) is & oom (eh 1.7 EQUATION OF MOTION: EARTHQUAKE EXCITATION In carthquake-prone regions, the principal problem of structural dynamics that concen Mructural engineers is the behavior of structures subjected to earthquake-induced motion cof the base of the structure, The displacement of the ground is denoted by a,, the total (or absolute) displacement of the mass by a’, and the relative displacement between the mass ‘and ground by (Fig. 1.7.1), Ateach instant of time these displacements are related by w(t) = u(t) + uplt) (7.1) Both w' and w, refer to the same inertial frame of reference and their positive directions coincide. ‘The equation of motion for the idealized one-story system of Fig. 1.7.La subjected to earthquake excitation can be derived by any one of the approaches introduced in Section 1.5. Here we choose to use the concept of dynamic equilibrium, From the free-body ‘diagram including the inertia force Jr, shown in Fig. 1.7.1b, the equation of dynamic Titdat t= D a7 ee™ Equations of Motion, Problem Statement, and Solution Methods = Chap, 4 ‘Only the relative motion w between the mass and the base due to structiral produces elastic and damping forces (i.c., the rigid-body component of the displacement of the structure produces no internal forces). Thus for a linear system Eqs. (1.3.1) and (1.4.1) are still valid. The inertia force f; is related te the acceleration @ of the mass by fi = mid! (173) Substituting Eqs. (1.3.1), (1.4.1), and (1.7.3) in Bq, (1.7.2) and using Eq. (1.7.1) gives mii + et + ku = —mil,(r) (174) This is the equation of motion governing the relative displacement or deformation w(t) of the linear structure of Fig. 1.7.la subjected to ground acceleration ii (1). For inelastic systems, Eq. (1.7.2) is valid, but Eg. (1.3.1) should be replaced by Eq. (1.3.6). The resulting equation of motion is mii + cit + faa, it) = — mil g(t) (1.75) ‘Comparison of Bgs. (1.5.2) and (1.7.4), or of Eqs. (1.5.3) and (1.7.5), shows that the equations of motion for structure subjected to two separate excitations—ground acceleration di ,(1) and external force = —mii,(1)—are one and the same. Thus the relative displacement or deformation u(t) of the structure duc to ground acceleration i, () will be identical to the displacement w(1) of the structure if its base were stationary and if it were subjected to an external force = —mii,(r). As shown in Fig. 1.7.2, the ground motion can therefore be replaced by the effective earthquake force (indicated by the subscript “eff™): Peart) = —wnii,(r) (1.7.6) This force is equal to mass times the ground acceleration, acting opposite to the accelera~ tion. It is important to recognize that the effective earthquake force is proportional to the mass of the structure. Thus the structural designer increases the effective earthquake force if the structural mass is increased. Although the rotational components of ground motion are not measured during carth- quakes, they can be estimated from the measured translational components and it is of ‘interest to apply the preceding concepts to this excitation. For this purpose, consider the cantilever tower of Fig. 1.7.3a, which may be considered as an idealization of the water tank of Fig. 1.1.2, subjected to base rotation @,. The total displacement 1’ of the mass isSec. 1.7 Equation of Motion: Earthquake Excitation 25 rea Ele + 5 Peakt) = =mbiieny 8} | = —r9, % Stationary base (a (b) Figure 1.7.3 Effective cartayuake foece: tational ground motion. made up of two parts: w associated with structural deformation and a rigid-body compo- nent ff. where A is the height of the mass above the base. At each instant of time these displacements are related by u(t) = w(t) + AB, (rt) (7.7) Equations (1.7.2) and (1.7.3) are still valid, but the total acceleration i(r) must now be determined from Eq. (1.7.7), Putting all these equations together leads to mii + ci + ku = mh, (0) (1.7.8) The effective carthquake force associated with ground rotation is Pear(t) = —mhd,(r) (17.9) Example 1.7 A uniform rigid slab of total mass m ix supported on four columns of height A rigidly con- ‘nected to the top slab and to the foundation slab (Fig. £1.74). Each column has a rectangular ‘cross section with second moments of area ¢, and /, for bending about the x and y axes, ‘respectively. Determine the equation of motion for this system subjected to rotation wpa of the foundation about a vertical axis, Neglect the mass of the columms, ‘Solution The clastic resisting torque or torsional moment fs acting on the mass is shown ins Fig. £1.76, and Newton's second law gives —fy = doit,Figure E1.7 where ky is the torsional stiffness, To determine ke, we imroduce a unit rotation, wy = 1, and identify the resisting forces in each column (Fig. El.7c). For a column with both ends clamped, ky, = 12E4,/h* and ky = 12E1,/43, The torque required to equilibrate these resisting forces is dd bb + tons (0.35) +4655) =tat p ae? a) ‘Substituting Eqs. (c), (d), and (b} in (a) gives Jolie + (hr? + kb?) = =Ioitge fe) ‘This is the equation governing the relative rotation uy of the roof slab due to rotational acceleration fige of the foundation slab, - * iSec. 1.8 — Problem Statement and Element Forces, a is used in a general sense to include any response quantity, such as displacement, velocity, er acceleration of the mass; also, an intemal forve or internal stress in the structure, When the excitation is an external force, the respotise quantities of interest are the displacement or deformation #(1), velocity ior), and acceleration ii(r) of the mass, For earthquake exci tion, both the total (or absolute) and the relative values of these quamtities may be needed, The relative displacement u(r) associated with deformations of the structure is the most important since the internal forces in the structure are directly related to w(t), 1.6.2 Element Forces ‘Gnee the deformation response history u(r) has been evi by dynamic analysis of the structure (ie.. by solving the equation of mation}, the clement forces—bending moments, shears, and forces—and stresses needed for structural design can bbe determined by static analysis of the structure at cach instant in time (i.c., no additional dynamic analysis is necessary). This static analysis of a one-story frame can be visualized in two ways 1. At cach instamt, the Lateral displacement w is known to which joint rotations are related and hence they can be determined; see E mple 1.1. From the known displacement and rotation of each end of a structural element (beam and column) the ele- ment forces (bending moments and shears} can be determined through the clement stiffiness Properties (Appendix 1); and stresses can be obtained from element forces. 2. The second approach is to introduce the equivalens static force, a central concept in earthquake response of structures, as we shall see in Chapter 6. At any instant of time ¢ this force ffs is the static (slowly applied) external force that will produce the deformation u determined by dynamic analysis, Thus Jett) = kuti) (1B. where & is the lateral stiffness of the structure, Altematively, fs cam be interpreted as the external force that will produce the same deformation « in the stiffness component of the structure [i.e., the system without mass or damping (Fig. 1.5.2b)] as that determined by dynamic analysis of the structure (i.e. the system with mass, stiffness, and damping (Fig. 1.5.2a)]. Element forces or stresses can be determined al cach time instant by static ‘analysis of the structure subjected to the force fs determined from Eq. (1.8.1). Itis unnec= essary to introduce the equivalent static force concept for the mass—spring—damper system because the spring force, also given by Eq. (1.8.1), can readily be visualized. "i cpsac: olin ana earl seer| 28 Equations of Motion, Problem Statement, and Solution Methods = Chap y time f; are combined with their increments over the time step to determine their values ay tar Why is extemal force im the second approach defined as (1) and not as /)(r)7 From Eq. (1.7.2), — fey = fete + fott) = karte) + ent! appropriate to include the velocity-dependent damping force because for structural design the computed clement siresses are tw be compared with allowable stresses that are specified based on static tests, on materials (i.c., tests conducted at slow loading rates), 1.9 COMBINING STATIC AND DYNAMIC RESPONSES In practical application we need to determine the total forces in a structure, including those existing before dynamic excitation of the structure and those resulting from the dynamic excitation. For a linear system the total forces can be determined by combining the results ‘of two separate analyses: (1) static analysis of the structure due to dead and live loads, temperature changes, and so on; and (2) dynamic analysis of the structure subjected to the time-varying excit ‘This direct superposition of the results of two analyses is valid only for linear systems. The analysis of nonlinear systems cannot, however, be separated into twa indepen dent analyses. The dynamic analysis of such a system must recognize the forces and deformations already existing in the structure before the onset of dynamic excitation, This is necessary, in part, to establish the initial stiffness property of the structure required to start the dynamic analysis. 1.10 METHODS OF SOLUTION OF THE DIFFERENTIAL EQUATION __The equation of motion for a linear SDF system subjected to external force is the second= order differential equation derived earlier: mii + cii + ku = ple) 1.10.1) ‘The initial displacement u(0) and initial velocity (0) at time zero must be specified 10 define the problem completely. ‘Typically, the structure is at rest before the onset of dy- ‘namic excitation, so that the initial velocity and displacement are zero. A brief review of four methods of solution is given in the following sections, , iSeo. 1.10 Methods of Solution of the Diflerential Equation 2 Example 1 ‘Consider a step force: p(t) = pr damping 0, fe this case, the differential equation of motion for a system without al+ep ia) The particular solution for Ba, (a) is p. apr = ( nya b) and the complementary solution is welt) = Acoscigt + B sin aryt te) where A and 8 are constants of integration and any == /E 7m ‘The complete solution is given by the sum of Eqs. (b) and (c): We) = Acosent + Bsinant + 2 () If the system is initially at rest, (0) = O and (0) = 0 ats = 0. For these initial conditions the constants A and 8 can be determined: Po A=-2 #=0 te) ‘Substituting Eq. (e) in Eq. (d) gives att) = ma = O08 yt) i The classical solution will be the principal method we will use in solving the dif- ‘ferential equation for free vibration and for excitations that can be described analytically, such as harmonic, step, and pulse forces. 1.10.2 Duhamel's Integral Another well-known approach to the solution of linear differential equations,”such as the equation of motion of an SDF system, is based on representing the applied force asx ‘8 sequence of infinitesimally short impulses, The response of the system to an applied force, pir), at time F is obtained by adding the responses to all impulses up to that time. aos leading to the following result for an undamped weap de aa a fm sin[w, (f= yd (e102)Equations of Motion, Problem Statement, and Solution Methods == Chap, 4 specializes t * elon) me w= &[ inf — olde = [se] = Lt — conan) mon Joy 10 ‘This result is the same as that obtained in Section 1.10.1 by the classical solution of the differ. ential equation Duhamel’s integral provides an alternative method to the classical solution if the applied force p(?) is defined analytically by a simple function that permits analytical eval. uation of the integral. For complex excitations that are defined only by numerical values of pit) at discrete time instants, Dubamel's integral can be evaluated by numerical methods, Such methods are not included in this book, however, because more efficient numerical procedures are available to determine dynamic response; some of these are presemied in ‘Chapter 5. 1.10.3 Frequency-Domain Method ‘The Laplace and Fourier transforms provide powerful tools for the solution of linear dif ferential equations, in particular the equation of motion for a linear SDF system. Because the Ewo transform methods are similar in concept, here we mention only the use of Fourier transform, which leads to the frequency-domain method of dynamic analysis. ‘The Fourier transform P(q) of the excitation function p(1) is defined by - Por = Fein = f pine" de (1.103) eee ‘The Fourier transform L/(c:) of the solution u(r) of the differential equation is then given by Uw) = H (oo) Pio) (1.10.4) ‘where the complex frequency-response function H (w) describes the response of the system to harmonic excitation. Finally, the desired solution w(r) is given by the inverse Fouriet transform of U (co): a0) = fe Hie) Plwe™ dav (1.10.5) Straightforward integration can be used to evaluate the integral of| (1.10.3) but contour intepration in the complex plane is necessary for Eg, (1.15), Clectd tory scale conte maly if p(r) is a simple function, method esr ‘untilSec, 1.10 Methods of Se ia) Equation a0 Figure 1.10.1 ‘The frequency-domain method of dynamic analysis is symbolized by Eqs. (1.10.3) and (1.10.5). The first gives the amplitudes P(a) of all the harmonic components that make up the excitation p(s). The second equation can be interpreted as evaluating the har= monic response of the syste ach component of the excitation and then superposing the harmonic responses to obtain the response w(1). The frequency-domain method, which is an altemative to the fime-dorain method symbolized by Duhamel's integral, is espe- cially useful and powerful for dynamic analysis of structures interacting with unbounded media. Examples are (1) the carthquake response analysis of a structure where the ef- fects of interaction between the structure and the unbounded underlying soil are signifi- cant (Fig. 1.10.1), and (2) the carthquake response analysis of concrete dams interacting with the water impounded in the reservoir that extends to great distances in the upstreamMethods — Chap 4 = m Statoment, and Solution direction (Fig. 1.10.2). Because earthquake an: structure—fluid sy the frequency -dor tion to th lysis of such complex structure-soil and this book, a comprehensive presentation of in method of dynamic analysis is not included. However, an introduc ethod is presented in Appendix A. 1.10.4 Numerical Methods ‘The preceding three dynamic analysis methods are Festricte consider the inelastic behavior of structures anticipated during earthquakes if the ground shaking is intense, The only practical approach for Stich systems involves numerical tim Mepping methods, which are presented in Chapter §. These methods are also useful for evaluating the response of linear systems itation—applied force p(r) or ground maori SP anes 's too complicated 10 be defined analytically and is described of¥ d to linear systems and cannotAppendix 1: Stifness Coafficiants tor a Flexural Element STUDY OF SDF SYSTEMS: ORGANIZATION ‘We will study the dynamic response of linearly clastic SDF systems in free vibration (Chapter 2), to harmonic and periodic excitations (Chapter 3), to step and pulse excita- tions (Chapter 4}, and to earthquake ground motion (Chapter 6). Because most structures are designed with the expectation that they will deform beyond the linearly elastic limit during major, infrequent earthquakes, the inelastic response of SDF systems is studied in Chapter 7. The time variation of response r(r) to these various excitations will be of inter- est, For structural design purposes, the maximum value (over time) of response r contains the crucial information, for it is related to the maximum forces and deformations that a structure must be able to withstand. We will be especially interested in the peak value of response, or for brevity, peak response, defined as the maximum of the absolute value of the response quantity: Fo mm max [KEL (LTE By definition the peak response is positive; the algebraic sign is dropped because it is usually ielevant for design. Note that the subscript @ attached to a response quantity denotes its peak value. FURTHER READING Clough, R. W., and Penzien, J., Dynamics of Structures, McGraw-Hill, New York, 1993, Sections 4-3, 6-2. 6-3, and 12-6. ‘Hamar, J. L. Dynamics of Structures, Prentice Hall, Enghewood Cliffs, NJ.. 1990, Chapter 9 and Section 13.5. 1: STIFFNESS COEFFICIENTS FOR A FLEXURAL ‘To compute bending moments and shears in a flexural structural clement—beam or column—the stiffness coefficients for the clement are required. These are presented in Fig. Al.1 for a uniform clement of length £, second moment of area J, and elastic34 Equations at Motion, Problem Statement, and Solution Methods Chap, ‘modulus £. The stiffness coefficients for joint rotwtion are shown in part (a) and thase for joint translation in part (b) of the figure. Now consider the uniform element shown in Fig. Al. Lc with its (Wo nodes identifieg asa and b that is assumed tw be axially inextensible, Its four degrees of freedom are the nodal displacements u, and uw), and nodal rotations @, and Gy. The bending moments at the two nodes are 4EI 2é1 bE! Mam — 8, + Oe + Ha — (ALA) 2g1 4ér 6EI My = 8a + Os + ry — 4AL2) The shearing forces at the two nodes are 12ET 12ET 6E/ 6EI ere at a te als " WE! 12ET 6El 6EI Vow tet ye — eh (ALA) Ateach instant of time, the nodal forces Mf,, My, V,, and Vj, are calculated from i, us, 4,, and 8; The bending moment and shear at any other location along. the element are determined by statics applied to the element of Fig. Al.te. PROBLEMS L1- Starting from the basic definition of stiffness, determine the effective stiffness of the combined L.3 spring and write the equation of motion for the spring—mass systems shown in Figs. PI.I to PLB.ed ee a a “Figure PLA LS Consider the free motion in the xy plane of a compound pendulum which consists of a rigid rod suspended from a point (Fig P15). The length of the rod is &. and its mass.m is uniformly sistributed, The width of the uniform rod is b and the thickness is r. The angular di of the centerline of the pendulum measured from the y-axis is denoted by (1). (a) Derive the equation governing (1). Ab) Lincarize the equation for small 6. (e) Determine the natural frequency of small oscillations, o Figure PLS Figure PL 1.6 Repeat Problem 1.5 for the system shown in Fig, P1.6, which differs in only one sense: its width varies from zero at O to b at the free end. LT Develop the equation governing the longitudinal motion of the system at Fig. PI.7. The rod is ‘made of an elastic material with clastic modulus £; its cross-sectional area is A and its length ‘is L. Ignore the mass of the rod and measure w from the static equilibrium position.PLA). Neglecting the 18 A right disk of mass m is mounted at the end of a exible shaft (Fig- weight ofthe shaft and neglecting damping, derive the equation of free orskonal vibration of the sick, The shear moxtulus (of rigidity) of the shaft is G. x Figure PLS 1.9. Write the equation governing the free vibration of the systems shown in Figs. PL9 to PLU, LIL Assuming the beam to be massless, each system has a single DOF defined as the vertical deflection under the weight w.. The flexural rigidity of the beam is EV and the length is 2.‘Chap. 1 Problems a LAL Derive the cquation of motion for the frame shown in Fig. PILI. ‘The Besural rigidity of the beams aie! columns is a8 nied, ‘The mass lurmpedt at the beam ts wr; otherwise, assume the frame tobe massless and Beplect damping. Hy comparing the result wish Iq, (1.3.2), comment ‘om the effect of base fixity + . Figure PL.13 LIM) Write the equation of motion for the one-story, one-bay frame shown in Fig. PILI. The flexural rigidity of the beam and colunins is ax ooted, ‘The mass lumped at the beam is. m: ‘otherwise, assume the frame to be massless and neglect damping. By comparing this equation ‘of motion with the one for Example 1.1, comment on the effect off base fixity. El. m Figure Pi LALS— Write the equation of motion of the one-story, one-bay frame shown in Figs. P1.1S and P11, 1.16 The flexural rigidity of the beam and columns is as noted. The mass lumped at the beam is im; otherwise, assume the frame te be massless and neglect damping, Check your result from, Problem 1.15 against Eq. (1.3.5), Comment on the effect of base fixity by comparing the two __ equations of mation.Equations of Motion, Problem Statement, and Solution Methods = Chap, Fig. P1.17. Bach diagonal wire is pretensioned tw a high stress; ifs cross-sectional area is 4 andl clastic modulus is £, Neglecting the mass of the columns and wires, derive the equation af motion governing free vibration in (a) the x-direction, and (lb) the y-direction, (Him: Because of high pretension, all wires contribute to the structural stiffness, unlike Example 1 ; the braces in compression do not provide stiffiness.) Figure P17 1.18 Derive the equation of motion governing the torsional vibration of the system of Fig. P1.IT about the vertical axis passing through the center of the platform. 1.19 An automobile is crudely idealized ax a lumped mass om supported on a spring-damper sys- tem as shown in Fig. P1.19, The automobile travels at constant speed v over a road whose roughness is known as a function of position along the road. Derive the equation of motion,