5. Conclusions
In this paper, an extensive study and comparison of five
algorithms for the active control of tonal disturbances
was made. The similarities and differences between the
algorithms were pointed out. By rough generalization,
the algorithms can be divided into two sets having the
same performance if the design parameters are chosen
in a suitable way, although this is a nontrivial task. The
DOFC algorithm provided the best overall performance,
having at least an adequate performance in
every task – yet not necessarily the best performance.
The remaining four algorithms all proved to yield a
very good performance and very poor performance,
again depending on the task. The choice of design parameters
is problem dependent for all methods except
for the DOFC method, which has a generic parameterization,
yielding similar results regardless of the underlying
process, with a minimal number of free
parameters. In essence, it is obvious that the controller
should be chosen according to the task at hand. If computational
power is scarce or if the process model is
not readily available, the IHC algorithm should be
preferred. If performance and stability are the
preferred quantities, the model-based algorithms
should be preferred. In the study, tables providing the
required information for choosing a suitable controller
were given.
5. ConclusionsIn this paper, an extensive study and comparison of fivealgorithms for the active control of tonal disturbanceswas made. The similarities and differences between thealgorithms were pointed out. By rough generalization,the algorithms can be divided into two sets having thesame performance if the design parameters are chosenin a suitable way, although this is a nontrivial task. TheDOFC algorithm provided the best overall performance,having at least an adequate performance inevery task – yet not necessarily the best performance.The remaining four algorithms all proved to yield avery good performance and very poor performance,again depending on the task. The choice of design parametersis problem dependent for all methods exceptfor the DOFC method, which has a generic parameterization,yielding similar results regardless of the underlyingprocess, with a minimal number of freeparameters. In essence, it is obvious that the controllershould be chosen according to the task at hand. If computationalpower is scarce or if the process model isnot readily available, the IHC algorithm should bepreferred. If performance and stability are thepreferred quantities, the model-based algorithmsshould be preferred. In the study, tables providing therequired information for choosing a suitable controllerwere given.
การแปล กรุณารอสักครู่..