In addition to the control of the instantaneous values of the variables, it is possible to include requirements in the cost function for the spectral content of the variables.
The general implementation of predictive control presented in this book does not impose any pattern on the switching signals. The maximum switching frequency is limited by the sampling frequency but the optimal switching state could be maintained by several sampling periods. This results in variable switching frequency and a spread spectrum of the controlled variables.
Taking the predictive current control scheme presented in Chapter 4 as an example, it can be observed that the load currents display a spread spectrum like the ones shown in Figure 10.9. Here, two different sampling frequencies have been considered, presenting different ranges and magnitudes of the spectral content in each case.
In some applications, such a spread spectrum is not desirable because it can produce oscillations and make the design of passive filters difficult. In order to overcome these problems it is possible to shape the spectrum by introducing a filter in the cost function. In this way, different frequencies have different weights in the cost ...