Return to search

Direct torque control for brushless doubly-fed machines

The Brushless Doubly-Fed Machine (BDFM) has recently become an
important research topic in the field of variable-speed AC drives. In
recent studies, the BDFM has shown significant potential for improving the
reliability and performance of AC drive systems, as well as reducing total
system cost. While the BDFM offers several advantages over existing AC
drives in steady-state operation, it suffers from dynamic instabilities
and slow response times, and a feedback control system is necessary. The
mathematics of the BDFM are much more complicated than those of a singly-fed
machine, and thus traditional control methods can't be applied. In
this thesis, a control method known as "Direct Torque Control" has been
adapted from that of a singly-fed induction machine and successfully
applied to the BDFM.
The thesis begins by discussing the background of the BDFM, its open-loop
operating characteristics, and some of the control considerations.
The reduced-order system differential equations are introduced, and it is
noted that they are coupled and nonlinear. Furthermore, all state
variables are time-varying (but periodic), even in steady-state operation.
In the controller development, it is found that a linear relationship
exists between the desired torque/flux-level change and the d-q voltages
to be applied to the control winding of the machine via the power-electronic
converter. This linear relationship, together with a one-step-ahead
predictor to compensate for computational delay, is successfully
used to control the speed and efficiency of the machine, for a wide range
of speeds and load torques. Numerous open- vs. closed-loop simulations
are compared and summarized, and it is found that the performance of the
BDFM is greatly improved in the closed-loop, with faster response and
reduced oscillation. Further simulations investigating the robustness of
the controller are summarized, and it is found that the controller is
reasonably insensitive to errors in most of the the static machine
parameters. Hardware implementation is briefly discussed but is not
complete; laboratory results are not yet available but should be soon.
Future controller considerations are then discussed; included among the
recommendations are an on-line parameter estimator for use in adaptive
control, and a controller for generator applications of the BDFM. / Graduation date: 1993

Identiferoai:union.ndltd.org:ORGSU/oai:ir.library.oregonstate.edu:1957/36482
Date31 March 1993
CreatorsBrassfield, William R.
ContributorsSpee, Rene
Source SetsOregon State University
Languageen_US
Detected LanguageEnglish
TypeThesis/Dissertation

Page generated in 0.0076 seconds