Electric power conversion systems – Frequency conversion without intermediate conversion to d.c. – By semiconductor converter
Reexamination Certificate
2003-05-22
2004-03-09
Sherry, Michael (Department: 2838)
Electric power conversion systems
Frequency conversion without intermediate conversion to d.c.
By semiconductor converter
C363S098000, C363S132000
Reexamination Certificate
active
06704215
ABSTRACT:
BACKGROUND OF THE INVENTION
The present invention relates to a method for controlling a matrix converter, in particular a method using space vector modulation.
A matrix converter is a self-commutated direct converter. It enables the conversion of a constant three-phase system into a system with variable voltage and frequency. Through the arrangement of the bidirectional power switches in a 3×3 switch matrix, each of the three output phases of the matrix converter can be electrically connected to any one input phase. One phase of the matrix converter includes an arrangement of three bidirectional power switches wherein each switch is connected, on the one hand, to an input phase and, on the other hand, to an output phase. An arrangement of this type is also referred to as a 3×1 switch matrix. The matrix converter does not require an intermediate circuit. Due to its topology, the self-commutated direct converter advantageously has a recovery capability and achieves sinusoidal mains currents through a suitably designed control.
Each of the bidirectional power switches of the matrix converter has two anti-serially connected semiconductor switches. Insulated Gate Bipolar Transistors (IGBT) are preferably used as semiconductor switches, which each include an antiparallel diode. Bidirectional power switches designed in this way are preferably used in converters for low and medium power. Through the control of these semiconductor switches of the bidirectional power switches, a continuous current path is established in a direction determined by the arrangement of the semi-conductor switches. If both semiconductor switches of a bidirectional power switch are controlled, the latter is bidirectionally activated and a current can flow in both directions. This creates a safe electrical connection between an input phase and an output phase of the matrix converter. If only one semiconductor switch of a bidirectional power switch is controlled, the latter is unidirectionally activated, creating an electrical connection between an input phase and an output phase of the matrix converter only for a preferred current direction.
Any desired time-averaged output voltage can be obtained—within certain limits—by a controlled temporal sequence of combinations of switch positions within a modulation period. A matrix converter includes a controller capable of computing a suitable switch combination based on information about the input voltage space vector and a desired value for the output voltage space vector.
Conventional control methods operate either according to a phase-oriented or a vector-oriented method.
The phase-oriented control method is described in the publication “Analysis and Design of Optimum-Amplitude Nine-Switch Direct AC—AC Converters”, by Alberto Alesina and Marco G. B. Venturini, IEEE Transactions on Power Electronics, Vol. 4, No. 1, January 1989, pp. 101-112. The space vector control method is described in “Space Vector Modulated Three-Phase to Three-Phase Matrix Converter with Input Power Factor Correction”, by Lászó Huber and Du{haeck over (s)}an Borejević, IEEE Transactions on Industrial Applications, Vol. 31, No. 6, November/December 1995, pp. 1234-1245.
The publication “Space Vector Modulated Matrix Converter with Minimized Number of Switchings and a Feedforward Compensation of Input Voltage Imbalance”, by P. Nielsen, F. Blaabjerg, and J. K. Pedersen, Proceedings of the 1996 International Conference on Power Electronics, Drives and Energy Systems for Industrial Growth, pp. 833-839, discloses a method for reducing the number of commutations. With his method, four active switching states and one switching state which generates at the output of the matrix converter a voltage space vector with zero amplitude, is calculated using a space vector modulation method. The switching states are referred to in this reference as active vectors and as Null vector. During the space vector modulation of a matrix converter, the input current vector and the output voltage vector can be located in the same sector or in neighboring sectors. Other combinations are possible in addition to the aforedescribed combinations. The pulse frequency, i.e. the voltage space vector sequence, is usually configured symmetrically, with the null vector being located in the center relative to the four active vectors. If the input current vector and the output voltage vector of the matrix converter are located in the same sector, then the pulse sequence results in eight commutations. Conversely, if the input current vector and the output voltage vector of the matrix converter are located in neighboring sectors, then the pulse sequence results in ten commutations, without optimization. By using the optimization proposed in the reference, a pulse sequence is generated which has also only eight commutations. The optimized pulse sequence is obtained by combining the calculated four active vectors and a null vector. The optimized pulse sequence differs from the non-optimized pulse sequence in that the pulse sequence of the active vectors is reversed in time and a suitable Null vector is selected. The null vector is selected from the three possible null vectors so that only one commutation takes place. With this optimized space vector modulation method, only eight commutations occur during each modulation period. Reducing the number of commutations per modulation period also reduces the switching losses of the matrix converter.
According to the last reference, eight commutations are always required when implementing the switching states calculated with a space vector modulation method in a modulation interval, also referred to as modulation period or half modulation period. Commutation takes place in several steps which have to be separated by a blocking time. As a result, commutation is subject to time constraints which result in a minimum on period. The commutation is controlled in a modulator connected after a controller. Output voltage errors result if on periods are calculated that cannot be realized with the commutation controller.
Because a difference between an on period calculated by the control set and an actual on period leads to a control error, such error should advantageously be eliminated, and the on periods should be calculated by the control set so as not to produce on periods that are shorter than a predetermined minimum on period.
Conventional controls based on space vector modulation accept the resulting errors. The controls are optimized only for attempting to keep the minimum on period as short as possible. This is achieved with a control method for a matrix converter as described in the publication “Semi Natural Two Steps Commutation Strategy for Matrix Converters”, by M. Ziegler, W. Hofman, in PESC, 1998, pp. 727-731. The method is based on a new definition of the 60° elec. sectors. The sector boundaries are here no longer defined by the zero crossings of the phase voltages of the input power grid, but rather by the zero crossings of the linked phase voltages. As a result, three voltage potentials V
P
, V
M
and V
N
occur in each 60° elec. sector, which can be easily determined. This method uses only two steps instead of four steps to switch from one main state to another, thereby significantly reducing the minimum on period. This commutation method, however, cannot prevent that switching states are calculated with a space vector modulation method that have minimum on periods less than a predetermined on period, since the space vector modulation method and the commutation control are performed in two different planes.
It would therefore be desirable and advantageous to provide an improved method for controlling a matrix converter, which obviates prior art shortcomings and eliminates output voltage errors on the matrix converter.
SUMMARY OF THE INVENTION
According to one aspect of the present invention, a method for controlling a matrix converter, with nine bidirectional power switches arranged in a 3×3 switch matrix, includes the steps of computing with a space vector modu
Feiereisen Henry M.
Laxton Gary L.
Sherry Michael
LandOfFree
Method for controlling a matrix converter does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method for controlling a matrix converter, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method for controlling a matrix converter will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3207127