Page 145 - Decision Making Applications in Modern Power Systems
P. 145
110 Decision Making Applications in Modern Power Systems
system and has been popular with power converters’ control recently for its
prominent characteristics, such as robustness, fast and precise dynamic
response, multiobjective control ability, and capability of application to non-
linear systems. MPC is classified into two main groups: continuous control
set MPC and finite control set MPC (FCS MPC). The first classification of
MPC uses a modulator to generate switching signals based on a continuous
output of the predictive control. The main advantage of this controller is the
fixed switching frequency besides the common advantages of MPC control-
ler. The first one has the advantage of dealing with a finite number of states
in the optimization problem, which will lead to a lower amount of computa-
tional burden and a good solution for the control systems with limited
computational abilities. Another advantage of FCS MPC is the direct con-
trol of the converter without need for a modulation step that decreases the
computational burden but has the drawback of variable switching frequency;
if the control case is not sensitive to variable switching frequency, then
FCS MPC is a good solution. To apply the multiobjective MPC control, it
is much convenient to use FCS MPC because of the lower computational
burden requirement; since it is the case in this study to have a multiobjective
control over smart grids, the focused MPC method will be FCS MPC
[44,45]. The main idea behind FCS MPC is to generate a discrete model of
system to forecast the behavior of the system, then a cost function is formed,
which best fits the control objectives. After forecasting the system behavior,
it would be applied to the defined cost function; the control actions that min-
imize the cost function will be the selected control actions in each cycle. To
ensure the optimum function of the system, this cycle would be done repeat-
edly; the flowchart view of this control method is shown in Fig. 4.10. In the
next section, MPC would be applied to a prototype microgrid including an
MFDG to validate the PQI characteristics as well as multiobjective operation
capability of the control method. The simplest cost function for an MPC con-
troller could be a current reference tracking such as
g½K 1 1 5 i L ½k 1 1 2 i L ½k 1 1 ð4:8Þ
which could be easily applied to an MFDG interfacing converter to track the
current reference, which could represent a CCM as discussed in
Section 4.3.4.4.1. Similar to PR-CCM, the current reference would be fed
into the controller, so that the controller could track it with the minimum
error. The advantage of simple MPC to PR-CCM is that, in MPC, there is no
need for harmonic extractor block and the local load current could be directly
used as a current reference when compensating local load harmonics.
4.3.4.4.3 Multiobjective model-based predictive control
The main difference between multiobjective model predictive control (MOMPC)
and single-objective model predictive control lies in defining the cost function
