Control of Electrically Assisted Forming - Electrically Assisted Forming - page 172

the electrical current had on the material strength. Accordingly from CCD testing, it
was shown that the flow stress is further reduced as a result of the increased current
applied to the material with increasing strain. Moreover, increases in flow stress reduc-
tions were increased by approximately 30 % in certain cases. Additional experimental
results and data-driven modeling for both NCCD and CCD profiles are given in [ 3 ].
From an applications standpoint, this type of control technique could be used
to help maintain consistent material flow and strength levels through components
with varying cross sections such that the resulting output is a formed part with
more uniform strain/strength properties.
7.4 Model-Based Control Feasibility
MBC is a control method where the control system incorporates a process model in
the control algorithm. Within MBC, there have been numerous approaches devel-
oped and this work focuses on model predictive control (MPC). In MPC, the model
of the process is used to estimate the response of the system to apply control action
instead of waiting for feedback from the process. Specifically in MPC, a weighted
objection function is defined, the response of the system to the inputs is predicted
over a finite time horizon, the performance of the system is optimized with respect
to the objective function using design variables as system inputs, and the system
is driven toward the optimized state [ 4 ]. This type of strategy has two main advan-
tages over traditional control in that it (1) betters the performance as a result of an
understanding of the system physics instead of reactive compensation, and (2) the
process output can be optimized to any parameter(s) while the underlying model
may contain uncertainty [ 5 ]. A general MPC architecture is shown in Fig. 7.20 .
When considering this control strategy for EAF, the previous sections used
a PID controller which employed a compensation strategy instead of predictive
action. Additionally, the desired state was directly measurable or capable of being
directly calculated from the actual state of the process. For advanced control of EAF
Fig. 7.20 General MPC block diagram. The MPC controller optimizes the objective function
with information from the process model. Output is the control input to the process [ 1 ]