JP4893151B2 - Space vector modulation method for AC-AC direct conversion device - Google Patents

Description

  The present invention relates to an AC-AC direct conversion device (matrix converter) that obtains a multi-phase output converted into an arbitrary voltage or frequency from a multi-phase AC power source, and particularly, a space vector whose size and phase change from moment to moment. The present invention relates to a space vector modulation method that expresses each input / output and selects a basic vector to be used and calculates a duty.

  This type of AC-AC direct conversion device that has existed in the past switches a bidirectional switch using a self-extinguishing semiconductor element at high speed, and converts a single-phase or multi-phase AC input to power of an arbitrary voltage or frequency. FIG. 1 shows a basic configuration of a conversion device for conversion. An input filter (InputFilter) 2 and an AC-AC direct conversion circuit 3 having bidirectional switches S1 to S9 are inserted in the R, S, and T phases of the three-phase AC power source 1, and both are controlled by a controller (controller) 4. By performing PWM control of the direction switch at a frequency sufficiently higher than the power supply frequency, an AC output of U, V, and W controlled to an arbitrary voltage or frequency is obtained while directly applying an input voltage to a load Load such as a motor. The bidirectional switch may be configured by using a plurality of unidirectional switches as shown in the figure.

  Here, the control method of the AC-AC direct conversion device is roughly divided into two systems, a virtual DC link type (indirect conversion method) and a direct AC-AC conversion type. The virtual DC link system is devised so that the virtual input converter and the virtual output inverter can be controlled independently considering a virtual DC link, and is similar to the configuration of a conventional current source PWM converter + voltage source PWM inverter. , The idea of control becomes easier. On the other hand, there is a constraint that six switching patterns in which the phases on the input side and the output side are all 1: 1 and are connected to different phases do not occur. In the direct AC-AC conversion type, there is no restriction on the above switching pattern.

  In addition, as a modulation method for generating a switching pattern for PWM control, there are mainly a carrier comparison method and a space vector method. The carrier comparison method generates a switching pattern by comparing the size of a triangular wave carrier and a sine wave. As a carrier comparison method applied to the virtual DC link method, a virtual inverter carrier is generated from a virtual converter carrier and a virtual PWM pulse. Thus, there has been proposed a technique in which the number of times of PWM control switching is reduced to the same number to reduce switching loss and noise and improve the output voltage control accuracy (see, for example, Patent Document 1).

  This is an AC-AC direct conversion device that uses a virtual DC link type modulation method based on carrier comparison, and adjusts the magnitude of the output voltage by controlling the magnitude of the virtual DC voltage of the virtual converter using the PAM (Pulse Amplitude Modulation) method. In the virtual inverter, a method of controlling only the output frequency has also been proposed (for example, see Non-Patent Document 1). In this control method, when the output voltage is in a low output region, a pulse with a small voltage height difference is used, so that the pulse width can be made wider than a pulse with a large voltage height difference. Also, the common mode voltage varies with reference to the input intermediate phase voltage. As a result, harmonics of the output voltage and common mode voltage can be reduced.

  The space vector method is a method of selecting an instantaneous space vector according to the output voltage command value of the AC-AC direct conversion device, and the switching pattern is determined by this selection. An AC-AC converter that employs this space vector method has also been proposed (see, for example, Non-Patent Document 2). In this space vector method, a harmonic vector is selected by combining the rotation vector, maximum single vibration vector, and zero vector so as to approach the command value locus of the magnetic flux linkage number vector integrated over time. An output voltage waveform with a small component can be obtained, and magnetic noise and torque ripple when driving an induction motor can be reduced during high voltage output.

Furthermore, although the modulation method is not a space vector, the direct current AC / AC conversion type AC-AC direct conversion device reduces the waveform distortion by appropriately selecting 27 switching patterns using the space vector ( For example, refer nonpatent literature 3).
JP 2005-168198 A PAM control method of matrix converter based on virtual AC / DC / AC conversion system, 2005 IEEJ Industrial Application Conference, 1-43, 1-203-1-206 Akio Ishiguro, Takeshi Furuhashi, Muneaki Ishida, Shigeru Okuma, Yoshiki Uchikawa: “Method of controlling the output voltage of a PWM controlled cycloconverter using a space vector”, Electrical Theory D, Vol. 110, no. 6, pp. 655-663 (1990) P. Mutschler, M.M. Marks: “A Direct Control Method for Matrix Converters” IEEE trans. on Industrial Electronics. Vol 49, No. 2, p362- (2002)

  For example, in Patent Document 1 and Non-Patent Document 1, the modulation method is the carrier comparison method, which improves the control accuracy of the output voltage, or reduces the harmonics of the output voltage or the common mode voltage. The process of generating a switching pattern is different in the space vector method in which the transition of the size can be grasped by the behavior of the space vector, and cannot be applied.

  In Non-Patent Document 2, although the output voltage harmonics can be reduced and the torque ripple of the motor load at high output can be reduced, the input current cannot be controlled to an arbitrary sine wave. The waves become very big. Therefore, this method is limited to applications where the input is not connected to the system power supply. Also, the common mode voltage cannot be reduced.

  On the other hand, although Non-Patent Document 3 uses direct torque control / direct power control / switching control rule, etc., it is necessary to detect not only the output current but also the input current in order to perform direct switching control of the input / output current. In addition, there are switching patterns in which the voltage error and phase difference of the basic vector become large. In this case, the control is slow, the number of times of switching and the switching sequence are wasted, and it is not suitable for high-speed control. Cannot be reduced.

  An object of the present invention is to use a space vector type modulation method in which a process for generating a switching pattern is different from the carrier comparison method, and can reduce harmonics of an input / output voltage and a common mode voltage, or an input / output magnetic flux. An object of the present invention is to provide a space vector modulation method for an AC-AC direct conversion device capable of suppressing pulsation of vectors.

  The present invention for solving the above problems is characterized by the following method.

(1) A space vector modulation method for an AC-AC direct conversion device that PWM-controls a bidirectional switch of an AC-AC direct converter from a multiphase AC power supply by modulation using a direct AC / AC conversion type space vector,
The vector state in which the line voltage of the polyphase AC output is expanded on the two-phase stationary αβ axis, the single vibration vector axis in which the phase of the sector where the output voltage command value vector Vo * exists is delayed on the X axis The simple vibration vector axis is defined as the Y axis, and the maximum vectors VXmax and VYmax, the intermediate vectors VXmid and VYmid, the minimum vectors VXmin and VYmin, and the zero that is the intermediate voltage of the phase voltages. The vector Vz and one rotation vector Vrot existing in the sector are used as basic vectors, and when the output line voltage command value Vol * exists in the low output voltage region, four simple vibration vectors VXmid, VXmin, VYmid, VYmin And one zero vector Vz are selected, and the output line voltage command value Vol * exists in the high output voltage region. Select four simple vibration vectors VXmax, VXmid, VYmax, VYmid and one zero vector Vz,
The selected five basic vectors, the input current command value Ii * and the output line voltage command value Vol * on the stationary αβ axis, are input current command values Iiα * and Iiβ *, and the output line voltage command value Volα *. , Volβ *, the input voltage detection value Vi, and the output current detection value Io, the duty of the five basic vectors is obtained and the input and output waveforms are simultaneously converted into a sine wave.

(2) When the line voltage command value Vol * exists in the high output voltage region,
・ The input waveform and output waveform can be converted into sine waves at the same time.
-There is no direct commutation between the maximum and minimum phases in the switching transition,
・ Switching transitions for each phase are possible,
7 patterns of combinations that can satisfy these three conditions are determined, and the five basic vectors are selected from these seven patterns.

  (3) The duty is calculated using a Moore-Penrose generalized inverse matrix.

  (4) The duty is calculated when a solution of the generalized inverse matrix of Moore-Penrose is negative or when the sum of the duties of the five basic vectors is not 1. The selection of the basic vector is switched until it is determined.

  (5) The duty is calculated by excluding a pattern that cannot be generated in an arbitrary input / output sector state from the input / output sector phase information in advance among the seven patterns. .

  (6) The duty is calculated by calculating the input current command value Iiα * by comparing the output current detection value Io with the input current command values Iiα * and Iiβ * instead of the output current detection value Io. , Iiβ * is obtained.

  (7) In the duty calculation, when an unexpected command value that gives unstable behavior to the system is input, a vector similar to the virtual DC link method is selected as the fail-safe mode, and a duty solution is always obtained. And

  As described above, according to the present invention, the input / output waveform can be controlled to a sine wave, and even when the output line voltage command value exists in the high output and low output voltage regions, the input / output voltage harmonics can be reduced and the common. There is an effect that mode voltage can be reduced and pulsation can be suppressed.

The three-phase input and three-phase output AC-AC direct conversion device shown in FIG. 1 will be described as an example. Considering the switching condition that does not cause short circuit of the input power supply and discontinuity of the output current, the nine bidirectional switches are limited to the combination of 27 (3 ³ ) patterns shown in Table 1.

  In Table 1, each vector (hereinafter referred to as a basic vector) existing on a space vector will be described. S1, S2, and S3 are groups of simple vibration vectors that are switched using only two phases out of three phases, R1. Is a group of counterclockwise rotation vectors, R2 is a group of clockwise rotation vectors, and Z is a group of zero vectors whose output voltage is always zero.

  Therefore, in order to operate the AC-AC direct conversion device of FIG. 1 normally, it is necessary to select and control an arbitrary state from the 27 patterns. Therefore, when these 27 switching patterns are developed on a stationary αβ axis of two phases from three-phase AC by three-phase / two-phase conversion, the space vector of the output voltage can be expressed as shown in FIG. Example of phase voltage phase θ = 15 degrees), a modulation method using a direct AC-AC conversion type space vector applied to the present invention will be briefly described below. .

  With reference to the line voltage Vuv between the outputs UV of the AC-AC direct conversion device as a stationary α-axis direction, a space vector of the output voltage as shown in FIG. 2 is constructed. FIG. 2 shows an example in which the input phase voltage phase θ is 15 degrees. In the AC-AC direct conversion device, 27 basic vectors fluctuate depending on the phase state and magnitude of the instantaneous input voltage, and the input voltage is three-phase AC. In the case of a power supply, the space vector also fluctuates in synchronization with the power supply frequency (for example, 50 Hz / 60 Hz). This point is different from normal inverter control (the space vector used in the normal inverter is a hexagon having a fixed length and phase because the input voltage is DC).

  Further, in the method of virtually controlling the direct current link in the alternating current to alternating current direct conversion device, it can be considered separately from the virtual converter and the virtual inverter, so the hexagonal fixed length is individually set on the input side and the output side. A fixed phase space vector can be used. Therefore, although the control is simplified and facilitated as before, the virtual DC link needs to connect the input phase and the output phase with virtually two lines, so all three input phases and three output phases are used. The connection states (six switching states of STATEs 19 to 24 in Table 1) cannot be expressed. Therefore, in the present invention, in order to make effective use of these six switching states, control is considered by a direct AC / AC conversion type modulation method using a space vector.

The basic vectors described above are divided into six groups as shown in Table 1, and the group of simple vibration vectors with the phase angle of 30 degrees as the positive axis is the simple vibration vector S1, and the phase angle of 150 degrees as the positive axis. A single vibration vector S2, a single vibration vector S3 having a phase angle of 270 degrees as a positive axis, a rotation vector R1 having a maximum length and rotating counterclockwise, and a rotation vector R2 having a constant length and rotating clockwise And the zero vector Z fixed at the hexagonal center zero, divided into the above six groups. Each of these basic vectors depends on the phase θ of the input voltage, that is, varies in synchronization with the angular velocity ω _{i of the} input voltage. The length of the vector (hexagonal size) corresponds to the magnitude of the input line voltage.

On the other hand, the space vector of the input current can be defined in the same way. FIG. 3 shows a space vector of the input current when the output current phase φ = 15 degrees, and the input R-phase current is based on the stationary α axis. AC-AC direct conversion devices (such as matrix converters that control switching of switches at a frequency sufficiently higher than the power supply frequency) perform output voltage control in the manner of a voltage-type inverter that performs PWM control by chopping up the input voltage. The control is similar to a current source converter that performs PWM control by chopping an output current (load current) assuming an inductive load. Accordingly, the space vector of the input current is expressed by a basic vector that varies depending on the output current phase φ, that is, the angular velocity ω _{o of the} output current. Further, the length of the vector depends on the load (the magnitude of the output current) at that time.

  Here, attention is paid to the difference between the space vector of the output voltage in FIG. 2 and the space vector of the input current in FIG. The grouping shown in Table 1 corresponds to the types of basic vectors forming the output voltage space vector of FIG. 2, and does not correspond to the input current of FIG. FIG. 4 shows an example in which the space vector (left) of the input current is compared with the space vector (right) of the output voltage (when the input voltage phase θ = 15 degrees and the output current phase φ = 15 degrees). The group of basic vectors that oscillate with the output voltage space vector in FIG. 4B is expanded into the same length of the basic vector in the input current space vector in FIG. 4A. In the load condition / phase condition of FIG. 4, the simple vector S1 axis in the direction of 30 degrees in the basic vector of the output voltage is expanded to the maximum length basic vector in each of the six directions on the input side (see FIG. 4). IRTT, iSTT, iSRR, iTRR, iTSS, iRSS). The rotation vector is represented by a fixed-length vector as in the output side although it rotates according to the output current phase and the axis reference is shifted.

  A direct AC-AC conversion type space vector modulation system according to an embodiment of the present invention will be described below.

(Embodiment 1)
When a technique such as Non-Patent Document 2 is used to control the improvement of the output line voltage waveform and the switching frequency and the common mode voltage, the input current is a sine wave although the harmonics are suppressed by the switching frequency reduction. If it is a grid interconnection system, it will adversely affect the power system. Therefore, the method is limited to an application that is not a grid-connected system or a system that requires some device that reduces harmonics separately.

  In the present embodiment, a method is proposed in which the input and output waveforms are simultaneously converted into a sine wave by the direct AC / AC conversion type space vector modulation method.

  In Non-Patent Document 2, only the output is subjected to space vector modulation using three basic vectors. By adjusting the basic vectors of the three output voltages, the components of the output voltage in the static α-axis direction and β-axis direction are extracted, and the pulse output time (duty addition value) of the three basic vectors becomes equal to the calculation cycle time T. So that it is controlled.

  On the other hand, in order to control the input current, two degrees of freedom for controlling the α-axis direction and the β-axis direction on the input side are further necessary. Therefore, in order to arbitrarily control both input and output waveforms, it is necessary to control a total of five basic vectors. In the definition of sector division of the input and output space vectors shown in FIG. 5, the sector where the input phase command value θIi * exists is “1”, and the sector where the output line voltage command value Vol * exists is “1”. This will be described below as a representative example.

  FIG. 6 defines a total of eight simple vibration vectors, rotation vectors, and zero vectors for the output-side space vector sector “1”. As shown in the figure, a single vibration vector X axis and a Y axis are defined, and a rotation vector direction axis is an R axis. With respect to the X axis and the Y axis, there are three vectors in one sector, so the subscripts of max, mid, and min are given in descending order of instantaneous values (VXmax, VXmid, VXmin, VYmax, VYmid, VYmin). Only one rotation vector direction axis (R axis) exists in one sector and is set to Vrot. The zero vector Vz is expressed using an intermediate phase of the input phase voltage (for example, if the relationship of the magnitude of the input phase voltage is a relationship of R> S> T, the switching of Table 1 is performed because the S phase is the intermediate phase). State Z2: The connection state of the bidirectional switch uses SSS), and the intermediate phase used for the zero vector Vz is automatically determined by determining the input sector “1-12” where the input phase command value θIi * exists (input) Sectors 3, 4, 9, 10 → RRR, 1, 2, 7, 8 → SSS, 5, 6, 11, 12 → TTT). By setting the intermediate phase of the input phase voltage to the zero vector, the phase serving as the voltage reference for the waveform control on the output side is always the intermediate phase, so that the common mode voltage can be reduced. In addition, for the simple vibration vector and the rotation vector, the switching state and the magnitude / phase of the vector are determined from the input sector region where the input phase command value θIi * exists and its phase state.

Incidentally, in order to convert the input / output waveform into a sine wave, it is necessary to select five out of the eight basic vectors defined in FIG. 6 and perform space vector modulation. ₈ C ₅ = 56 patterns exist as basic vector selection patterns within one sector, and an optimum pattern in accordance with the control policy is appropriately selected from these, and modulation control is performed.

  As the selection method, the applicant of the present application has proposed a control method using one rotation vector, one zero vector, and three simple vibration vectors. In this method, both the input space vector and the output space vector are composed of 5 basic vectors within a phase difference of 60 degrees, and the sine wave of the input / output waveform is reduced, the number of times of switching is reduced, and direct commutation between the maximum and minimum phases is prevented. Can be realized.

  The present embodiment proposes a technique for reducing the PWM pulse drop of the output voltage while making the input / output waveform sine wave like the above selection technique. First, a case where the output line voltage command value Vol * exists in the low output voltage region will be described.

  The voltage error reduction will be described. FIG. 7 defines a low output voltage region and a high output voltage region in the output side space vector. When the output line voltage command value Vol * is inside the boundary line BL (that is, a region having an absolute value smaller than the boundary), the region outside the boundary line BL is a high output voltage region. In addition, in the case of the boundary line BL, it is arbitrarily determined whether the low output area or the high output area is set.

  The space vector modulation in the present embodiment is a method in which the duty of five basic vectors is determined and PWM control is performed so that both the input and output are integrated with the voltage command value within one control cycle. Here, attention is paid to the harmonics of the space vector modulation on the output side. FIG. 8 shows the output sector “1” when the input phase θ = 15 °. FIG. 9 shows the relationship between the output line voltage command value Vuv * and the PWM pulse within one control cycle in the state of FIG. In both cases, the same voltage command value is input, and the integrated value of the PWM pulse within one control cycle matches the magnitude of the voltage command value.

  In FIG. 8, in the conventional virtual DC link PWM system, control is performed using the maximum value of the single vibration vector having the best voltage utilization factor and the second largest single vibration vector. In the example of FIG. 8, RTR, RSR, RTT, RSS, and zero voltage SSS are used. Therefore, as shown in FIG. 9, the input line voltages Vrt and Vrs are compared with the output line voltage command value Vuv *. A pulse is output. The input line voltage Vrt is the maximum input line voltage at that moment. When this is used, the voltage utilization rate can be increased, but in the low output voltage region, the pulse has a large difference with respect to the voltage command value. . Further, since the pulse width is narrowed by outputting a pulse with a large deviation, the ratio of the pulse output being reduced due to dead time or the like is increased, and the voltage error is easily affected.

  Therefore, in the present embodiment, in a region where the voltage command value is small, instead of the maximum value of the simple vibration vector (that is, RTR or RTT in FIG. 8), the minimum value of the simple vibration vector (STT or STS in FIG. 8) is used. By using this, a pulse having a value closer to the voltage command value is output. FIG. 9 shows the state of the pulse output at that time. The input line voltage which is close to the output line voltage command value Vuv * without outputting the conventional input line voltage Vrt. By outputting Vst, the pulse output time is generally increased as compared with the conventional method (time when no pulse is output: zero voltage output time is reduced). As a result, the rate at which the pulse output is reduced due to the dead time is reduced, and the degree of influence on the voltage error can be reduced, thereby enabling more accurate control. In particular, in the extremely low voltage region, improvement of the minimum on-pulse width (pulse missing) can be expected.

  As described above, five basic vectors to be selected from the output-side space vectors are preferentially determined, and the input-side space vectors are used correspondingly. The duty calculation of the five basic vectors can be calculated by a geometrically solving method using a trigonometric formula or the like or a method of calculating an inverse matrix. In this embodiment, a generalized inverse matrix of Moore-Penrose is used. An example of calculating and making both input and output waveforms sine waves will be described below.

  After selecting the five basic vectors as described above, the input current command value Ii * and the output line voltage command value Vol * are developed on the stationary αβ axis, Iiα *, Iiβ *, and Volα *, Volβ *. give. For convenience, the five basic vectors “VXmid, VXmin, VYmid, VYmin, Vz” are “V1, V2, V3, V4, V5” on the output side and “I1, I2, I3, I4, I5” on the input side. Redefine (example: FIG. 10). For each input / output, the five basic vectors are decomposed into stationary αβ axes, and “I1α, I2α, I3α, I4α, I5α”, “I1β, I2β, I3β, I4β, I5β”, “V1α, V2α, V3α, V4α, V5α. “V1β, V2β, V3β, V4β, V5β”. The duty of the five basic vectors to be obtained is “d1, d2, d3, d4, d5”, and these added values are d1 + d2 + d3 + d4 + d5 = 1. Therefore, the equation (1) can be derived from the above.

  Here, the output line voltage command value Vol * may be arbitrarily given according to the load to be driven, but the input current command value Ii * is output in accordance with the principle of the AC-AC direct conversion device. It depends on the load current to be determined. Therefore, the magnitude of the input current command value Ii * cannot be controlled independently of the magnitude of the output voltage command value Vol *. Therefore, from the relationship that the three-phase instantaneous active power between the input and output of the AC-AC direct conversion device is equal, and the condition for adjusting the input reactive power by the input phase command θ *, Iiα *, The simultaneous equations for Iiβ * are derived.

  Note that Vi is an input phase voltage, Vo * is a phase voltage of an output voltage command, Io is an output current detection value, and each is subjected to αβ axis conversion. The reactive power adjustment input phase command θ * is a difference in input current phase with respect to the input phase voltage phase (θIi−θVi).

  Solving equation (2) yields equation (3). Information on input phase voltage detection value Vi, output voltage command value Vol *, and output current detection value Io is used to obtain Iiα * on the left side of equation (1). , Iiβ * can be calculated computationally.

  Vil is an input line voltage.

  The duties d1 to d5 of the five basic vectors can be obtained as shown in equation (4) by calculating the inverse matrix of equation (1).

  As described above, in the present embodiment, in the low output voltage region defined in FIG. 7, five basic vectors of four simple vibration vectors VXmid, VXmin, VYmid, VYmin, and zero vector Vz are selected, and the output current detection value Io and The output times of the five basic vectors are determined from the duty calculation formulas (1) to (4).

  According to this embodiment, a DC AC-AC having means capable of simultaneously reducing the output voltage harmonic by minimizing the PWM pulse drop of the output voltage in the low output voltage region by simultaneously converting the input / output waveform into a sine wave. It becomes a conversion type AC-AC direct conversion device. Further, in the extremely low voltage region, the influence of the missing pulse due to the dead time is reduced, and a more accurate voltage output can be obtained.

(Embodiment 2)
In the first embodiment, the input current command value Ii * is calculated using the output current detection value Io in the equation (3). Therefore, it is essential to detect the output current, and it is necessary to use detection means such as a current sensor. Therefore, this embodiment proposes a method that can calculate the duty (output time) of five basic vectors without detecting the output current.

  The procedure up to the duty calculation is the same as that in the first embodiment, except that the output current detection values Ioα and Ioβ in the equation (3) are not used. First, “I1α, I2α, I3α, I4α, I5α” and “I1β, I2β, I3β, I4β, I5β” in the coefficient matrix of the equation (1) are all converted into relational expressions of Ioα and Ioβ, and the coefficients are compared. For example, paying attention to the expression related to the third row Ioα in the expression (1), the state of the space vector in FIG. 11 (I1 = RSS, I2 = RSR, I3 = RTT, I4 = RTR, I5 = SSS) is represented. explain. When formula (1) in this state is expanded, it can be expressed as formula (5).

  It should be noted that any switching state other than this example (whatever the switching state is applied to I1 to I5) can be expressed by the coefficients of Ioα and Ioβ as shown in equation (5). Moreover, if all the duty factors I1α to I5α are developed in a table, it is only necessary to select them according to the basic vector.

  By substituting Iiα * in equation (3) for the left side Iiα * in equation (5), the left and right sides can be compared with coefficients of Ioα and Ioβ, respectively, so that the output current detection values for Ioα and Ioβ can be eliminated. That is, equation (6) can be derived from the necessary and sufficient conditions for coefficient comparison.

  Similarly for Iiβ *, an arbitrary duty relational expression is derived by coefficient comparison. From the above, when generalized in matrix form including the relational expression of Vo1 * and duty, equation (7) is obtained.

  Here, Kva1 to Kva5 are Kvaan (coefficient: n = 1 to 5) · Vi1a + Kvabn (coefficient: n = 1 to 5) · Vi1b, respectively. Kvb1 to Kvb5 are Kvban (coefficient: n = 1 to 5) · Vi1a + Kvbbn (coefficient: n = 1 to 5) · Vi1b, respectively. n is a vector number 1 to 5, d1 to d5 are space vector duties using five basic vectors, Volα * and Volβ * are output line voltage commands, and Iixx (Iia: Ioα in the equation for obtaining Iiα *) The coefficient to be multiplied, Iiab: the coefficient to be multiplied by the Ioβ term of the expression for obtaining Iiα *, Iiba: the coefficient to be multiplied by the Ioα term of the expression for obtaining Iiβ *, and Iibb: the coefficient to be multiplied to the Ioβ term of the expression for obtaining Iiβ *) The current command calculated value (substitute from equation (3)), Kixx1 to Kixx5 (xx is aa, ab, ba, bb) are coefficients.

  As shown in the equation (7), the duty factor matrix is a 7 × 5 non-square matrix, and therefore, using the Moore-Penrose generalized inverse matrix calculation method, the duties d1 to d5 are expressed as in the equation (8). Calculate.

  Note that “+” in the duty factor inverse matrix means a generalized inverse matrix of Moore-Penrose.

  The above is the procedure of duty calculation (space vector modulation). According to the present embodiment, the input / output waveform can be arbitrarily sine wave without detecting the output current, and the output voltage error (PWM pulse height) in the low output voltage region can be reduced as much as possible. The direct-current AC-AC conversion type AC-AC direct conversion device is provided. Further, since it is not necessary to detect the magnitude of the output current with high accuracy, an output current sensor for duty control is not required, and cost reduction and simple open loop control are possible.

  The simulation result of this embodiment is shown in FIG. For comparison, FIG. 12A also shows a simulation result of a conventional virtual DC link method. It can be seen that FIG. 13A does not output the maximum phase as compared to the output line voltage of FIG.

(Embodiment 3)
The present embodiment relates to a 5-vector selection method in the high output voltage region and its duty calculation as defined in FIG. In the high output voltage region, unlike the low output voltage region of the first and second embodiments, the voltage command value between the output lines is large, so that the voltage difference from the zero voltage vector is large. Therefore, in order to construct a PWM pulse only with a voltage vector close to the magnitude of the command value, harmonics of the output voltage (voltage error reduction) can be reduced without using the zero vector Vz having a large drop.

As a method for selecting 5 vectors from 7 vectors other than the zero voltage vector Vz defined in FIG. 6, there are ₇ C ₅ = 21 patterns. Among these, the constraint conditions are as follows.

  (1) The input waveform and the output waveform can be made sinusoidal at the same time (the duty solution exists).

  (2) There is no direct commutation between the maximum and minimum phases in the switching transition.

  (3) A switching transition for each phase is possible.

  The combinations that can satisfy the above three conditions are limited to 7 patterns shown in Table 2 out of 21 patterns.

  Therefore, in the present embodiment, the five basic vectors are determined by selecting combinations that meet the conditions from the seven patterns in the high output voltage region. As with the first embodiment, the duty calculation method is calculated using equations (1) to (4). At this time, the calculation is repeatedly performed until a valid duty solution is obtained for all the patterns for selecting the five basic vectors from the above seven vectors. If the above conditions (1) to (3) are satisfied, a duty solution can be obtained for one of the seven combinations. Therefore, the combination and the duty are applied to the final output.

  In the present embodiment, by calculating the five basic vectors and their output times in the above procedure, the input / output waveform can be controlled without outputting the zero voltage vector Vz having a large drop in the high output voltage region. A direct-current AC-AC conversion type AC-AC direct conversion device having Furthermore, by combining with the first embodiment, output harmonics can be reduced in the entire region.

(Embodiment 4)
In the present embodiment, as in the third embodiment, one of the seven patterns in Table 2 is selected in the high output voltage region. However, it differs from the third embodiment in that the output current detection value Io is not used when calculating the duty. This process is the same as in the second embodiment, and the output current values Ioα * and Ioβ * are deleted by coefficient comparison as shown in the equations (5) to (7). Since the coefficient matrix of the duty calculation matrix is a 7 × 5 non-square matrix as shown in Expression (7), the duty solution is calculated from the Moore-Penrose generalized inverse matrix calculation as shown in Expression (8). Thereafter, as in the third embodiment, since one duty solution is always obtained, a valid solution is derived by repeated calculation.

  According to the present embodiment, in addition to the effects of the third embodiment, a direct current AC-AC conversion type AC-AC direct conversion device including means that can be realized without detecting an output current value is obtained.

  An example of the simulation result is shown in FIG. For comparison, the waveform of the conventional virtual DC link system is also shown in FIG. Comparing the output line voltage, it can be seen that this embodiment does not output a zero voltage having a large drop from the command value when the output voltage is high. Therefore, it is possible to expect a reduction in harmonics of the output voltage as compared with the conventional method.

(Embodiment 5)
In the methods of the third and fourth embodiments, the duty solution using the Moore-Penrose generalized inverse matrix for all seven patterns that satisfy the three constraints shown in the third embodiment among the switching states in the high output voltage region. Always wanted. Among them, a correct solution that satisfies the condition that the duty does not finally become negative and the added value of the duty solution is 1 is derived and applied. In this procedure, all 7 patterns are obtained. Requires computation time. In addition, in order to uniquely determine the solution instantaneously from the seven patterns, it is necessary to further subdivide the sector information in the input / output space (including the command value phase and size), as in an AC-AC direct conversion device. This is a very difficult discriminant when the space vector itself changes every moment.

  Therefore, in the present embodiment, pattern information that cannot be generated in an arbitrary input / output sector state is tabulated in advance from only the phase information of the input / output sectors based on the methods of the third and fourth embodiments, and the table is always displayed. This is a direct current AC-AC conversion type AC-AC direct conversion device provided with means for reducing the number of patterns to be calculated to less than 7.

  For example, when the sector definition in FIG. 5 is redefined as shown in FIG. 14 and the case of the input sector “1” and the output sector “1” is considered, High-mode 3 and High-mode 6 in Table 2 are In this case, the constraint condition cannot always be satisfied, so that it can be excluded from the arithmetic processing.

  Therefore, the Moore-Penrose generalized inverse matrix operation can be shortened twice. Similarly for other sector conditions, such conditions always exist, and can be calculated by excluding them in advance.

  From the above, there is an effect that the arithmetic processing can be reduced and the speed can be increased as compared with the third and fourth embodiments.

(Embodiment 6)
In Embodiments 3 to 5, the Moore-Penrose generalized inverse matrix is processed in real time by the singular value decomposition method or the like, but the 7 × 5 coefficient matrix shown in the equation (8) has a maximum column rank ( rank = 5), it has already been proved that the equation (9) holds from the Moore-Penrose condition, and is converted into the equation (10) using this.

  A is a duty factor matrix, AT is a transpose matrix of A, and A + is a Moore-Penrose generalized inverse matrix of A.

  From the input / output sector information, the selection status of the five basic vectors is known in advance in each of the seven patterns in the high output voltage region. Therefore, by developing equation (10) in that state, a complicated matrix operation is also developed. Gives the duty equation. For example, in the input / output sector definition of FIG. 14, the input sector “1” and the output sector “1”, and taking the case of High-mode 1 in Table 2 as an example, the 5-vector selection situation is “RSR, STS, RSS” , STT, RTS ". The duty factor matrix when these five basic vectors are expanded on the stationary αβ axis is expressed by equation (11).

  Since the rank of the matrix is 5 (maximum column rank), each duty can be calculated by substituting into the equation (10) as in the equation (12).

  For other vector states, it is not necessary to process the inverse matrix operation in real time by deriving a table by deriving similar mathematical expressions.

  As described above, in the present embodiment, the direct current AC-AC conversion type AC-AC direct conversion device including means capable of simplifying the calculation itself by the calculation formula based on the formula (10) and speeding up the processing is obtained.

(Embodiment 7)
By combining the methods of the embodiments described above and performing space vector modulation, the input / output waveform can be converted into a sine wave and the harmonics of the output voltage can be reduced in the entire voltage range. In this embodiment, when the means of the first to sixth embodiments are used in the definition of the low output voltage region and the high output voltage region in FIG. Apply safe ideas.

  Table 3 summarizes the selection patterns of basic vectors that can be considered in all areas. By selecting five vectors VXmax, VXmid, VYmax, VYmid, and Vz as in the conventional virtual DC link method as the fail-safe mode, a valid duty solution is always obtained even if an unexpected command value is input. That is, processing is performed according to the procedure shown in FIG.

  The procedure of FIG. 15 discriminates the low output voltage region and the high output voltage region based on FIG. 7 (S1). If it is the high voltage region, the process proceeds to 7 pattern vector selection processing. If it is the low voltage region, one vector is determined. Therefore, the process proceeds to the duty calculation as it is (S2), and in the high voltage mode, among the seven combinations, the loop is repeated until a valid solution is obtained (S3, S4), and low when there is no valid solution in the high voltage mode. The process shifts to the voltage mode process (S5), and the process shifts to the fail-safe mode when a valid solution cannot be obtained even in the low voltage mode process (S6 to S8).

  According to the present embodiment, even if the input / output command value is in an unexpected range, there is provided a means capable of deriving some duty command and preventing unstable behavior to system control. It becomes a direct current AC-AC conversion type AC-AC direct conversion device.

The basic block diagram of an alternating current-alternating current direct conversion apparatus. The basic space vector diagram of output voltage. The basic space vector diagram of input current. Input side space vector diagram (a) and output side space vector diagram (b). Definition example of “input side” sector and “output side” sector of space vector. The state diagram of the basic space vector in the output sector “1”. Definition example of low output voltage region and high output voltage region of output side space vector. The state diagram of the basic space vector in the output sector “1”. The relationship diagram of the output voltage command value and PWM pulse within one control cycle. An example of redefining the space vector on the input and output sides. The state diagram of the space vector on the input side and output side. Simulation result of virtual DC link method. The simulation result of embodiment. An example of redefining I / O sectors. Example of duty calculation procedure.

Explanation of symbols

1 AC power supply 2 Input LC filter 3 AC-AC direct conversion circuit 4 Control device

Claims (7)

A space vector modulation method for an AC-AC direct conversion device that PWM-controls a bidirectional switch of an AC-AC direct converter from a polyphase AC power source by modulation with a direct AC / AC conversion type space vector,
The vector state in which the line voltage of the polyphase AC output is expanded on the two-phase stationary αβ axis, the single vibration vector axis in which the phase of the sector where the output voltage command value vector Vo * exists is delayed on the X axis The simple vibration vector axis is defined as the Y axis, and the maximum vectors VXmax and VYmax, the intermediate vectors VXmid and VYmid, the minimum vectors VXmin and VYmin, and the zero that is the intermediate voltage of the phase voltages. The vector Vz and one rotation vector Vrot existing in the sector are used as basic vectors, and when the output line voltage command value Vol * exists in the low output voltage region, four simple vibration vectors VXmid, VXmin, VYmid, VYmin And one zero vector Vz are selected, and the output line voltage command value Vol * exists in the high output voltage region. Select four simple vibration vectors VXmax, VXmid, VYmax, VYmid and one zero vector Vz,
The selected five basic vectors, the input current command value Ii * and the output line voltage command value Vol * on the stationary αβ axis, are input current command values Iiα * and Iiβ *, and the output line voltage command value Volα *. , Volβ *, input voltage detection value Vi, and output current detection value Io, the duty of the five basic vectors is obtained, and the input and output waveforms are simultaneously converted into a sine wave. Space vector modulation method. When the line voltage command value Vol * exists in the high output voltage region,
・ The input waveform and output waveform can be converted into sine waves at the same time.
-There is no direct commutation between the maximum and minimum phases in the switching transition,
・ Switching transitions for each phase are possible,
The space of the AC-AC direct conversion device according to claim 1, wherein seven patterns of combinations that can satisfy the three conditions are determined, and the five basic vectors are selected from the seven patterns. Vector modulation method.   3. The space vector modulation method for an AC / AC direct conversion device according to claim 1, wherein the duty is calculated using a generalized inverse matrix of Moore-Penrose.   The calculation of the duty is basically performed until a valid solution is obtained when one of the solutions of the Moore-Penrose generalized inverse matrix is negative, or when the added value of the duties of the five basic vectors is not 1. 4. The space vector modulation method for an AC-AC direct conversion device according to claim 1, wherein selection of vectors is switched.   2. The duty is calculated by excluding a pattern that cannot be generated in an arbitrary input / output sector state from the phase information of the input / output sectors in advance among the seven patterns. The space vector modulation method of the alternating current-alternating current direct conversion apparatus of any one of -4.   The calculation of the duty is performed by calculating the input current command values Iiα * and Iiβ * by comparing the output current detection value Io and the input current command values Iiα * and Iiβ * instead of the output current detection value Io. The space vector modulation method for an AC-AC direct conversion device according to any one of claims 1 to 5, wherein: In the duty calculation, when an unexpected command value that gives unstable behavior to the system is input, a vector similar to the virtual DC link method is selected as the fail-safe mode, and a duty solution is always obtained. Item 7. A space vector modulation method for an AC-AC direct conversion device according to any one of Items 1 to 6.

Images