Process Control: Modeling, Design, and Simulation

8.4. Model Uncertainty and Disturbances

Thus far we have assumed that the model is perfect when designing the open-loop control system. In practice, it is impossible to have a perfect model. Often, the process gain can be off by 20–50% (or even more, for highly nonlinear processes). The next example shows the problems associated with a purely open-loop control strategy when there are uncertainties or disturbances.

Example 8.4: First-Order Process with Model Uncertainty

The model and process are represented by the following transfer functions:

There are no noninvertible elements, so the control design procedure results in the following:

Notice ...