DOC PREVIEW
MIT 18 03 - Impulse and step responses

This preview shows page 1-2 out of 6 pages.

Save
View full document
View full document
Premium Document
Do you want full access? Go Premium and unlock all 6 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 6 pages.
Access to all documents
Download any document
Ad free experience
Premium Document
Do you want full access? Go Premium and unlock all 6 pages.
Access to all documents
Download any document
Ad free experience

Unformatted text preview:

� 93 18. Impulse and step responses In real life, we often do not know the parameters of a system (e.g. the spring constant, the mass, and the damping constant, in a spring-mass-dashpot system) in advance. We may not even know the order of the system—there may be many interconnected springs (or diodes). (We will, however, suppose that all the systems we consider are linear and time independent, LTI.) Instead, we often learn about a system by watching how it responds to various input signals. The simpler the signal, the clearer we should expect the signature of the system parameters to be, and the easier it should be to predict how the system will respond to other more complicated signals. To simplify things we will always begin the system from “rest.” In section we will study the response of a system from rest initial conditions to two standard and very simple signals: the unit impulse β(t) and the unit step function u(t). The theory of the convolution integral, Section 19, gives a method of determining the response of a system to any input signal, given its unit impulse response. 18.1. Impulse response. In engineering one often tries to understand a system by studying its responses to known signals. Suppose for defi-niteness that the system is given by a first order left hand side ˙x+p(t)x. (The right hand side q(t), isn’t part of the “system”; it is the “input signal.”) The variable x will be called the “system response,” and in solving the ODE we are calculating that response. The analysis pro-ceeds by starting “at rest,” by which is meant x(t) = 0 for t less than the moment at which the signals occur. One then feeds the system various signals and watches the system response. In a certain sense the simplest signal it can receive is a delta function concentrated at some time t0: β(t − t0). This signal is entirely concentrated at a single instant of time, but it has an effect nevertheless. In the case of a first order system, we have seen what that effect is, by thinking about what happens when I contribute a windfall to my bank account: for t < t0, x(t) = 0; and for t > t0, x(t) is the solution to x˙ + p(t)x = 0 subject to the initial condition x(t0) = 1. (Thus x(t0−) = 0 and x(t0+) = 1.) If p(t) = a is constant, for example, this amounts to 0 if t < t0 x(t) = −a(t−t0 )e if t > t0.� � 94 This system response depends upon t0, but if the system is LTI, as it is in this example, its dependence is very simple: The response to a unit impulse at t = 0 is called the weight function or unit impulse response of the system, or written w(t). If the system is given by x˙ + ax, the weight function is given by 0 for t < 0 w(t) = −ate for t > 0. In terms of it, the response to a unit impulse at any time t0 is x(t) = w(t − t0). 18.2. Impulses in second order equations. The word “impulse” comes from the interpretation of the delta function as a component of the driving term q(t) in a second order system: (1) mx¨ + bx˙ + cx = q(t). In the mechanical interpretation of this equation, q(t) is regarded as an external force acting on a spring-mass-dashpot system. Force affects acceleration, so the cumulative total of force, that is the time integral, affects velocity. If we have a very large force exerted over a very small time, the acceleration becomes very large for a short time, and the velocity increases sharply. In the limit we have an impulse, also known as a good swift kick. If q(t) = aβ(t − t0), the system response is that the velocity ˙x increases abruptly at t = t0 by the quantity a/m. This produces a corner in the graph of x as a function of t, but not a break; the position does not change abruptly. Thus the system response, w(t), to a unit impulse at t = 0 is given for t < 0 by w(t) = 0, and for t > 0 by the solution to (1) subject to the initial condition x(0) = 0, ˙x(0) = 1/m. For example, if the system is governed by the homogeneous LTI equation x¨ + 2 ˙x + 5x = 0, an independent set of real solutions is {e−t cos(2t), e−t sin(2t)}, and the solution to the initial value problem with x(0) = 0, ˙x(0) = 1, is (1/2)e−t sin(2t). Thus 0 for t < 0 w(t) = (1/2)e−t sin(2t) for t > 0. This is illustrated in Figure 13. Note the aspect in this display: the vertical has been inflated by a factor of more than 10. In fact the slope w˙ (0+) is 1. The unit impulse response needs to be defined in two parts; it’s zero for t < 0. This is a characteristic of causal systems: the impulse at� 95 0.3 0.25 0.2 0.15 0.1 0.05 0 −0.05 −0.1 −1 0 1 2 3 4 5 6 Figure 13. The weight function for x¨ + 2 ˙x + 5x t = 0 has no effect on the system when t < 0. In a causal system the unit impulse response is always zero for negative time. 18.3. Singularity matching. Differentiation increases the order of singularity of a function. For example, the “ramp” function 0 for t < 0 ramp(t) = t for t > 0. is not differentiable at t = 0 but it is continuous. Its derivative is the step function u(t), which is not continuous at t = 0 but it is a genuine function; its singular part is zero. But its derivative is the delta function. (This can be made to continue; one can define an even more singular type of generalized function, of which β→(t), often called a doublet, is an example, but we will not enter into this here.) Suppose a function satisfies an ODE, say mx¨ + bx˙ + cx = q(t),96 in which q(t) may have a singular part. Whatever singularities x may have get accentuated by the process of differentiation, so the most singular part of q(t) must match up with the most singular part of mx¨. This then forces x to be not too very singular; otherwise its second derivative would be more singular than q(t). To be more precise, if q(t) is a generalized function in our sense, then its singular part must occur as the singular part of mx¨. The result is that x˙ does not have a singular part, but does have discontinuities at the locations at which q(t) has delta components. Similarly, x is continuous, but has jumps in its derivative at those locations. This makes physical sense: a second order system response to a generalized function is continuous but shows sudden jumps in velocity where the signal


View Full Document

MIT 18 03 - Impulse and step responses

Documents in this Course
Exam II

Exam II

7 pages

Exam 3

Exam 3

8 pages

Load more
Download Impulse and step responses
Our administrator received your request to download this document. We will send you the file to your email shortly.
Loading Unlocking...
Login

Join to view Impulse and step responses and access 3M+ class-specific study document.

or
We will never post anything without your permission.
Don't have an account?
Sign Up

Join to view Impulse and step responses 2 2 and access 3M+ class-specific study document.

or

By creating an account you agree to our Privacy Policy and Terms Of Use

Already a member?