National Instruments 370753C-01

Chapter 4 System Analysis

Often it is desirable to run several simulations with different inputs. In this

case, you can define a PDM whose columns contain the input vectors for

the different simulations. Then u will be ny×q×Nsamp, where q is the

number of different simulations to be run. The resulting y will be

ny×q×Nsamp, with each column of the PDM corresponding to a different

simulation.

The input PDM must have a regular domain—that is, the interval between

each domain value and the one succeeding it must be the same over all

points in the domain. If the system is discrete, the domain intervals must be

equal to the system’s sampling period. If the system is continuous, it is

discretized using the exponential (zero-order hold) method, with the

sampling interval set equal to the input domain interval spacing.

Note For accurate results, you need to make sure this sampling interval is small enough

that discretization effects are negligible.

The next step is to create a general signal and store it as a PDM where

domain is time as shown in Example 4-5. Because you are using a SISO

system, this input is a single-channel PDM.

Example 4-5 Performing a General Time-Domain Simulation

t = 0:0.1:15;

osig = ones(1,30);

sig = [0*osig,0.5*osig,osig,0.5*osig,0*ones(1,31)];

U = pdm(sig,t);

Create the system:

Sys = system([-2.3,0.01,5.1;0,-0.35,-2;0,2,-.35],

[1,.25,.25]',[1.34,0,0],0);

and perform the simulation:

Y = Sys*U;

To see how well the system tracks the input signal, plot the input,

as follows, and the system’s response, shown in Figure4-3.

plot ([U,Y], {legend = ["Input Signal",

"System Response"],line_color = "black",

xlab = "Time (sec)", ylab = "Amplitude"})