User`s guide

8 Model Analysis
model for co ntro l design. In this case, you can import your unstable model
into Simulink or Control System Toolbox products.
One way to check if a nonlinear model is unstable is to plot the simulated
model output on top of the validation data. If the simulated output diverges
from measured output, the model is unstable. However, agree m ent between
model output and measured output does not guarantee stability.
In the case of linear models, if you believe that your system is stable, but your
model is unstable, the n you can estimate the model again with a
Focus se tting
that guarantees stability. For example, set
Focus to Stability to find the best
stable model. This s etting might result in a reduced model quality. For more
information about
Focus,seetheAlgorithm Properties reference page.
A more advanced approach to achieving a stable m odel is by setting the
stability threshold p roperty to allow a m argin of error. The threshold model
property is accessed as a field in the
algorithm structure:
• For continuous-time models, set the value of
model.algorithm.advanced.sstability .Themodelisconsideredstable
ifthepoleonthefarrightistotheleftof
sstability threshold.
• For discrete-time models, set the value of
model.algorithm.advanced.zstability .Themodelisconsideredstable
if all poles inside the circle centered at the origin and with a radius
zstability.
For more information abou t
Threshold fields for linear models, see the
Algorithm Pr oper ties reference page.
Missing Input Variables
If the Model Output plot and Residual Analysis plot shows a poor fitand
you have already tried different struct ures and orders and modeled noise, it
might be that there are one or more missing inputs that have a significant
effect on the output.
Try including other measured signals in your input data, and then estimating
the models again.
8-70