� 121 24. The Laplace transform and more general systems This section gives a hint of how flexible a device the Laplace trans-form is in engineering applications. 24.1. Zeros of the Laplace transform: stillness in motion. The mathematical theory of functions of a complex variable shows that the zeros of F (s)—the values r of s for which F (r) = 0—are just as important to our understanding of it as are the poles. This symmetry is reflected in engineering as well; the location of the zeros of the transfer function has just as much significance as the location of the poles. Instead of recording resonance, they reflect stillness. We envision the following double spring system: there is an object with mass m1 suspended by a spring with spring constant k1. A second object with mass m2 is suspended from this first object by a second spring with constant k2. The system is driven by motion of the top of the top spring according to a function f(t). Pick coordinates so that x1 is the position of the first object and x2 is the position of the second, both increasing in the downward direction, and such that when f(t) = x1 = x2 = 0 the springs exert no force. f(t) x 1 x 2 m1 m 2 k1 k 2 Figure 16. Two spring system The equations of motion are m1x = (1) ¨ 1 k1(f(t) − x1) − k2(x1 − x2) m2x¨ 2 = k2(x1 − x2) This is a system of second order equations, and as you can imagine mechanical engineering is full of similar systems.� � � 122 Suppose that our main interest is in x1. Let’s take Laplace trans-forms, and assume rest initial conditions. (m1s2 + (k1 + k2))X1 = k2X2 + k1F (m2s2 + k2)X2 = k2X1. Use the second equation to express X2 in terms of X1, and substitute this value into the first equation. Then solve for X1 to get: m2s2 + k2X1(s) = k1F (s). (m1s2 + (k1 + k2))(m2s2 + k2) − k2 · 2 The “transfer function” W (s) is then the ratio of the LT of the system response, X1, and the LT of the input signal, F : k1(m2s2 + k2)W (s) = . (m1s2 + (k1 + k2))(m2s2 + k2) − k22 It is still the case that W (r) is the multiple of ert which occurs as x1 in a solution to the equations (1) when we take f (t) = ert . Thus the zeros of W (s) at s = ±i k2/m2—the values of s for which W (s) = 0— reflect a “neutralizing” circular frequency of � = k2/m2. If f(t) is sinusoidal of this circular frequency then x1 = 0 is a solution. The suspended weight oscillates with (k1/k2) times the amplitude of f (t) and reversed in phase (independent of the masses!), and exactly cancels the impressed force. Check it out! 24.2. General LTI systems. The weight function w(t), or its Laplace transform, the transfer function W (s), completely determine the sys-tem. The transfer function of an ODE has a very restricted form—it is the reciprocal of a polynomial; but the mechanism for determin-ing the system response makes sense for much more general complex functions W (t), and, correspondingly, much more general “weight func-tions” w(t): given a very general function w(t), we can define an LTI system by declaring that a signal f(t) results in a system response (with null initial condition, though in fact nontrivial initial conditions can be handled too, by absorbing them into the signal using delta functions) given by the convolution f(t) w(t). The apparatus of the Laplace ↑transform helps us, too, since we can compute this system response as the inverse Laplace transform of F (s)W (s). This mechanism allows us to represent the system, the signal, and the system response, all three, using functions (of t, or of s). Differential operators have vanished from the scene. This flexibility results in a tool of tremendous power.MIT OpenCourseWare http://ocw.mit.edu 18.03 Differential Equations���� Spring 2010 For information about citing these materials or our Terms of Use, visit:
View Full Document
Unlocking...