IEEE Electrification Magazine - December 2019 - 15

reference signals have to be manipu-
lated to achieve power sharing.

conventional droop-control method
can provide stable power sharing
between the generators and other
energy carriers. Figure 13 illustrates a
simplified communication-less volt-
age reference generation for power
sharing in the ac system that is based
on the droop-control principle. The
voltage and current are measured to
generate the average power signal.
The active power (P) is shared through
the speed droop, whereas the reactive
power is distributed through voltage
droop. The voltage amplitude and
phase angle (which are generated
through voltage and speed droops,
respectively) are then combined to
generate the voltage reference signal

system and stabilize the network's
behavior with a dwell time to reduce
the effects of jumps.

Model Predictive Control
A converter control structure that
implements model predictive control
(MPC) is presented in Figure 11,
where the objective is to regulate the
output voltage. The future error sig-
nal between the reference and the
estimated (and filtered) future volt-
age output of the observer or predic-
tion model is fed to the optimizer as
an input. The objective for the MPC is
to generate the optimal control sig-
nal for the converter.
MPC is a control algorithm where
the optimization problem (control
objective) is solved at each time step
while satisfying the given constraints.
A converter forecasts the future pro-
cess states and output over a predic-
tion horizon that is based on the
control and process outputs at the
present time step. An optimization
problem is formulated and solved at
each time step using model predic-
tions and the process reference. Since
the future optimal control output
necessary for attaining the future pro-
cess output is generated at each time
step, it is also called the receding, or
sliding horizon, control strategy.

Load Sharing
In an onboard power system with
multiple energy carriers, the convert-
ers' voltage and power set point can
be generated and transferred using
communication-based or communi-
cation-less control. With a communi-
cation network, there are different
strategies to set the references for
each converter, including the circular
chain, central limit control, and paral-
lelism control bus (with a voltage ref-
erence). As a more practical approach
that has no communication, the

MPC
Voltage
Set Point

Constraints

Future
Error

+

Optimizer

Control
Output

Converter

Voltage
Output

Control
Objective

-
Future
Output

Prediction
Model

Figure 11. The converter-control structure using the MPC.

Hybrid Control
Hybrid control combines the behav-
iors of continuous- and discrete-time
dynamical systems, such as switched
electrical circuits, voltages, and cur-
rents. The discrete-time state can
consist of logical parameters under
constraints and the state of the
switches (on or off) that triggers
jumps in the system. Moreover, if
there are uncertain parameters and
perturbations in the system model,
these uncertainties can be included
with a parametric interval where
there are several controllers, and
each controller has the best perfor-
mance at each interval. Then the
supervisory controller, as seen in Fig-
ure 12, can switch the controllers
based on the current state of the

Analog
to Digital

Supervisory Control
Switching Signal
+
Vref

-

Controller 1
Controller 2

Digital
to Analog

Converter

Vout

Controller n

Figure 12. The converter controller based on the hybrid-control approach.

	

IEEE Elec trific ation Magazine / D EC EM BE R 2 0 1 9

15
IEEE Electrification Magazine - December 2019

Table of Contents for the Digital Edition of IEEE Electrification Magazine - December 2019
