View this PageEdit this PageAttachments to this PageHistory of this PageHomeRecent ChangesSearch the SwikiHelp Guide

Time Delay Inverse Response

Example Situation for Time Delay



Suppose an application of an input to a system is performed a length L away from the system, there will be a time delay.

Uploaded Image: tube.gif

The above system yields a transport delay as such:

Uploaded Image: timedealy1.gif


Proof that Time Delay Systems are Infinite Order



Now consider a more general pure time-delay process whose transfer function is:

Uploaded Image: transfer1.gif

Note the only affect a has on an input is a delay. In order to find a rational version of this equation, we will rewrite it as follows:

Uploaded Image: rationalize.gif

Now we can see that a pure time delay system is an infinite order system. A Taylor Series expansion gives a numerical approximation to g(s).

Uploaded Image: taylor.gif


Dynamic Behavior of Time Delay Systems



From Ogunnaike and Ray, consider Fig. 8.4 (from section 8.3)

Uploaded Image: reactorexample.gif

In this situation we consider a to be

Uploaded Image: alpha.gif

The relationship between the pipe inlet concentration co and the pipe outlet concentration is found to be as follows

Uploaded Image: concentrations.gif

We can write the deviation variables as

Uploaded Image: deviation.gif

We now take the Laplace transform and obtain,

Uploaded Image: laplace.gif

A block diagram representation is now shown as such. Note that the dynamic behavior of this system is actually that of two processes. The first is the long pipe, and the second is the reactor.

Uploaded Image: block.gif

Because we know that the reactor is a first order system, we designate the deviation of the reactor outlet concentration cA(t) from the steady state concentration cAss as y(t). The transfer function relationship between the input, w(s), and y(s) is given by:

Uploaded Image: transfer function.gif

Now we consider the overall transfer function between the SYSTEM input, u(s), and output y(s).

Uploaded Image: transfer function2.gif

This equation is the transfer function of a first order plus time delay system. When we take the inverse Laplace transform, we must consider e-as. In the text, table C1 (in Appendix C) is consulted to find a transfer function of the form,

Uploaded Image: transfer function3.gif

The inverse Laplace Transform of the above transfer function is given as such:

Uploaded Image: inverselaplace.gif

In this case b=a can be seen to be the time lag. The inverse Laplace Transform in this example is then given as follows:

Uploaded Image: inverselaplace2.gif

The catch to this is that for any value of ta the function has to be equal to zero. This makes sense, because we know nothing is happening before the response.

Consider the following simple graph. The value of a from the graph is easy to see.

Uploaded Image: graph1.gif

The horizontal part of this graph represents a and is the time delay.


Problem



Now try a time delay problem.

Consider the dynamic behavior of a process, represented as a first order system. The steady state gain is 0.5, the time constant is 10 minutes, and the time delay is 4 minutes. What is the transfer function representation of this system? What is the unit step response of this system.


Problem Solution



From the problem statement we know that the transfer function is as follows:

Uploaded Image: problemsoln1.gif

After taking the inverse Laplace Transform from Appendix C, we obtain the unit step response for this system.

Uploaded Image: problemsoln2.gif

INVERSE RESPONSE TO BE POSTED SOON!!

Links to this Page