Electric power conversion systems – Current conversion – Using semiconductor-type converter
Reexamination Certificate
2002-06-03
2004-03-23
Riley, Shawn (Department: 2838)
Electric power conversion systems
Current conversion
Using semiconductor-type converter
Reexamination Certificate
active
06711038
ABSTRACT:
This application claims the priority of German Patent Application Serial No. 199 58 041.3, filed Dec. 3, 1999, pursuant to 35 U.S.C. 119(a)-(d), the subject matter of which is incorporated herein by reference.
BACKGROUND OF THE INVENTION
The present invention relates to a method for controlling bi-directional switches in converters, in particular 3×3 matrix converters, using separate control signals for the two current directions.
Various methods for current commutation in matrix converters extend the basic topology by incorporating additional components which carry a load current during operation. For example, C. T. Pan (“
A zero switching loss matrix con verter”,
1993, IEEE PESC, p.545-550) and J. G. Cho (“
Soft switched Matrix converter for high frequency direct AC to AC power conversion”,
1991, EPE, p.196-201) propose the use of resonant circuits operating at higher frequencies; W. Söhner (“
The self-commutated direct converter and its use as a matrix converter for supplying asynchronous machine drives whose rotation speed is regulated”.
Dissertation, University Karlsruhe, 1993) suggests freewheeling paths, while T. Svensson (“
The modulation and control of a Matrix Converter—Synchronous Machine Drive”,
1991. EPE p. 469-476) suggests the use of capacitors connected in parallel.
Systems incorporating higher-frequency resonant circuits disadvantageously have additional components which carry a load current during operation, so that accurate switching at the time of the zero crossing of the high-frequency current or voltage is required. It is not possible to select arbitrary switching times, which can cause dead times. Only a limited amount of damping of the oscillations in the input filters which are typical of matrix converters is possible. The solutions proposed by Söhner and Svensson disadvantageously also include additional components which carry current during operation, such as B6 bridges, an intermediate circuit capacitance with devices for voltage limiting, for example retarding choppers, or capacitors connected in parallel with the switches, as well as the additional associated losses.
However, four methods are known which do not require additional components to carry the load current during operation and use separate control signals for the two current directions (bi-directional switches, also known as four-quadrant switches, comprising a back-to-back parallel circuit formed by two two-quadrant switches, also referred to as unidirectional switches): two four-step methods that measure the commutation voltage and/or of the load current direction before each switching step (N. Burany “
Safe control of
4
Quadrant Switches”,
1989. IEEE-industry Application Society, p. 1190-1194), a method which allows commutation between only two input phases in two steps (R. Cittadini et al.: “
A matrix converter switching controller for low losses operation without snubber circuits”,
1997. EPE p.4.199-4.203) and a method which operates using two steps (M. Ziegler et al.: “
Methods for controlling bidirectional switches in converters”,
1997, German Patent Application 19746797.0-32).
The latter methods employ a switching algorithm in two or four steps. The switching sequence starts either by determining the polarity of the commutation voltage between the two phases involved in the switching process, or by determining the polarity of the current in the switch which is switched on at that time. Safety times, which are governed essentially by the switching times of the power semiconductors and their drive devices, must be observed between the steps.
The major disadvantage of the last-mentioned four methods is the necessity to precisely detect the current direction or the commutation voltage. This is particularly difficult when the values are small. In the case of a current measurement, this relates to the current zero crossing, and in the case of a voltage measurement, it relates to the zero crossing of the concatenated voltages. An incorrect measurement, for example as a result of offset errors due to residual magnetism in the measurement transformer, interference or fluctuations/oscillations in the input phase voltages, cause a brief short-circuit between two line phases or to an interruption in the output current. Both can destroy switch elements in the converter, particularly when using the method that detects the current direction. The trend to components with a low threshold voltage, such as Cool MOS for avoiding high forward losses, can lead to high short-circuit currents, even with low commutation voltages.
With the two four-step methods proposed by Burany, two unidirectional switches are switched off and two unidirectional switches are switched on. Furthermore, Burany's methods disadvantageously require four switching steps, which makes these methods less suitable for the rapid switching processes which are desirable for matrix converters. Matrix converters demand rapid switching, particularly for active damping of line filter oscillations and owing to the lack of any energy store.
A further disadvantage is that, once the current direction/commutation voltage has/have been detected, the subsequent switching steps are executed automatically before the first switching step, without the possibility to react to a change in the polarity of the current direction/commutation voltage, which can briefly result in a short-circuit or an interruption in the current during the entire four-step switching process.
The method proposed by R. Cittadini et al. inherently causes brief short-circuits. Furthermore, only commutation between two input phases is considered.
It would therefore be desirable and advantageous to control bidirectional switches with separate control signals for the two current directions, preferably in 3×3 matrix converters, with a minimum number of unidirectional switches that need to be switched in each commutation process, in such a way that:
no interruption can occur in the load/output currents, even when the current is very small,
no short-circuit can occur, even when the concatenated input voltages are in the vicinity of a zero crossing,
commutation is possible in as few switching steps as possible, without any additional components which carry load current during operation,
the switching times can be defined within wide limits,
the load current can be commutated for all input phases,
a freewheeling path is provided for both current directions, at all times, even during switching.
SUMMARY OF THE INVENTION
According to one aspect of the invention, a method is provided for controlling bidirectional switches in converters, preferably 3×3 matrix converters, without the use of additional components that carry a load current during operation, wherein each bi-directional switch is composed of two back-to-back connected unidirectional switches that receive separate control signals for forward and reverse current directions. The method includes switching between a first main state and a second main state by a voltage-controlled two-step process which includes in a first step, switching off all unidirectional switches except for a subset of the unidirectional switches that include the first main state and the second main state, and in a second step, switching on all unidirectional switches for the second main state, wherein the first main state is a basic main state and the second main state is a secondary main state, or the first main state is a secondary main state and the first main state is a basic main state.
Advantageously, in the basic main state, unidirectional switches are redundantly closed in addition to the unidirectional switches that provide a bi-directional connection between a nominal input phase and a corresponding output phase, and in an associated secondary main state, one of the unidirectional switches is switched on between the corresponding output phase and the basic input phase. The latter unidirectional switch is in category N a unidirectional switch in a forward direction, and in category P a unidirectional switch in the rev
Hofmann Wilfried
Ziegler Marcus
Feiereisen Henry M.
Riley Shawn
Siemens Aktiengesellschaft
LandOfFree
Method for controlling bi-directional switches in matrix... 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 bi-directional switches in matrix..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method for controlling bi-directional switches in matrix... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3281527