Controlling a Dynamic System - Part 3

Calculus Level pending

Consider an arbitrary physical system, the behaviour of which is determined by the following differential equation:

d x d t = x + u \frac{dx}{dt} = x + u

Here, x x is a time-varying quantity of the system and u u is a time-varying input to the system. The given differential equation is converted to a discrete difference equation of the form:

x ( T + h ) = a x ( T ) + b u ( T ) x(T+h) = a x(T) + b u(T)

Where T T is an arbitrary time instant and h = 0.1 h = 0.1 s. Also, a a and b b are real numbers. A solver is asked to refer to the following note:

Conversion of Differential Equation to Difference Equation

This discrete dynamic system is given a time-varying input dependent on x ( T ) x(T) which is:

u ( T ) = a x ( T ) u(T) = -a x(T)

Moreover, x ( 0 ) = 10 x(0) = 10

Assertion: The input ensures that the energy of the system always decreases with time.

Is this assertion true or false? Explain why.

The energy of the system is:

E ( T ) = ( x ( T ) ) 2 2 E(T) = \frac{(x(T))^2}{2}

Also try:

Controlling a Dynamic System

Controlling a Dynamic System - Part 2

Converting Differential to Difference Equation

True False

This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try refreshing the page, (b) enabling javascript if it is disabled on your browser and, finally, (c) loading the non-javascript version of this page . We're sorry about the hassle.

1 solution

Mark Hennings
Jul 25, 2019

Since a = e h a = e^h and b = e h 1 b = e^h-1 , we obtain the relation x ( T + h ) = a ( 1 b ) x ( T ) x(T+h) \; =\; a(1-b)x(T) and hence that x ( j h ) = 10 [ a ( 1 b ) ] j j 0 x(jh) \; = \; 10\big[a(1-b)\big]^j \hspace{2cm} j \ge 0 Thus x ( j h ) x(jh) , and hence E ( j h ) E(jh) , will be a decreasing sequence provided that 0 < a ( 1 b ) = 2 e h e 2 h = 1 ( 1 e h ) 2 < 1 0 < a(1-b) = 2e^h - e^{2h} = 1 - (1 - e^h)^2 < 1 . This is certainly true for any 0 < h < ln 2 0 < h < \ln2 .

A more normal (there are many others) method of creating a discrete version of a differential equation is to make the approximation (using the first two terms of the Taylor series for x ( T + h ) x(T+h) ): x ( T + h ) x ( T ) + h x ( T ) x(T + h) \; \approx x(T) + hx'(T) which leads to the equation x ( T + h ) = ( 1 + h ) x ( T ) + h u ( T ) x(T+h) \; = \; (1+h)x(T) + hu(T) The method of this question is a little strange, since it makes the control term u ( T ) = a x ( T ) u(T) = -ax(T) dependent on the size the discrete time interval h h , since a = e h a = e^h .

Thanks for posting the solution. I wanted to show that simple linear feedback of the form u ( T ) = C x ( T ) u(T) = -Cx(T) can ensure that the energy monotonically decreases, provided 0 < ( a b C ) < 1 0<(a-bC)<1 . The choice C = a C = a met this requirement. I had not noticed that the control action was step-size dependent until you pointed it out.

Karan Chatrath - 1 year, 10 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...