JP3518978B2 - Converter control device for obtaining DC from polyphase AC - Google Patents

Description

Description: BACKGROUND OF THE INVENTION [0001] 1. Field of the Invention [0002] The present invention relates to a switching control device for a converter that converts polyphase alternating current to direct current by pulse width control. 2. Description of the Related Art FIG. 4 is a system block diagram of a converter control most conventionally used in industry. FIG.
, 1 is a polyphase converter constituting a main circuit of the converter, 2 is a polyphase AC power supply for supplying power, 3 is an AC controller, and 4 is a DC controller. In FIG. 1, a DC controller 4 amplifies the error by an amplifier so that the DC amount of the converter output matches a given DC command value, and outputs an AC command value to an AC control loop. The AC controller 3 compares the AC command value with the polyphase AC of the AC power supply 2,
The converter is controlled by amplifying the error and outputting the switching signal to the semiconductor switch element constituting the multi-phase converter 1. Conventionally, two types of AC control in the converter control shown in FIG. 4 are mainly used: a carrier modulation system and a tracking system. A control block diagram of the carrier modulation scheme is shown in FIG. That is, the AC command value is ^* is compared with the AC amount is, and the error is
To generate a switching input command value u ′.
The carrier modulator 6 compares the input command value u ′ with a triangular carrier signal to determine an open / close signal for the polyphase converter 1.
The switching timing of the open / close signal occurs when the input command value u 'and the triangular wave carrier signal are in a compare match. A control block diagram of the tracking system is shown in FIG. That is, the AC command value is ^* is compared with the AC amount is, and the switching signal is determined by the vector selection unit 7 according to the polarity of the error. The switching timing of the switching signal is determined by the hysteresis comparison unit 8 in which the error is set in advance. Occurs when the width is exceeded. In the carrier modulation method, the switching timing can be finely determined and the AC distortion can be suppressed to some extent. However, the switching signal, that is, the switching vector does not directly participate in the control. In addition, a useless switching operation occurs, and the following performance is limited. On the other hand, in the tracking method, since the switching vector is generated by the polarity of the error, the tracking performance is expected by directly controlling the error. However, since the switching timing of the vector is determined by the hysteresis width, it cannot be determined as finely as in the carrier modulation method, resulting in a large distortion. In particular, in the case of digital realization, the switching vector is output at intervals of the control cycle, so that the control cycle must be sufficiently short in order to prevent the error from exceeding the hysteresis width.
It is not economical in terms of cost. As described above, both methods have advantages and disadvantages, and it is considered that a control method that combines the advantages of both methods, that is, the control performance of the tracking method and the fineness of the switching timing of the carrier modulation method, is ideal. SUMMARY OF THE INVENTION The present invention has been made in view of the above-mentioned problems, and has a control device for a converter capable of determining a switching timing of a switching vector for minimizing an AC current distortion rate while ensuring follow-up performance. The purpose is to provide. SUMMARY OF THE INVENTION The present invention has been made in view of the above-described circumstances, and is intended to determine the above-described vector switching timing in a converter control device for obtaining DC from polyphase AC. Means are provided. In addition,
In the following description, the determination of the switching timing when switching from the first vector to the second vector will be described, and the selection of the first and second vectors will not be described, but the selection of the vector may be obtained by various methods. it can. In the present invention , an effective AC current id and a reactive AC current iq are obtained by dq conversion of a three-phase AC current, and the effective AC current id and the reactive AC current iq are respectively converted into an effective AC current command value id ^* and an invalid AC current command value id ^*. The converter is configured such that two switching vectors are sequentially selected in one control cycle T so as to match the AC current command value iq ^*, and the switching is switched from the first switching vector to the second switching vector at an appropriate timing α. When controlling the semiconductor switch element, in order to determine the switching timing, the distortion generated in one control cycle is used as an evaluation function, and the vector switching timing is determined so that the evaluation function is minimized. As an example, assuming that the instantaneous AC current following error is s (t) = i (t) ^* -i (t), the evaluation function D can be expressed by the following equation (1). [0008] [0009] In the invention of claim 1 of the present invention, in determining the switching timing, as described above, the alternating current command of the first detection value or the estimated value of the alternating current of the start point of the control cycle is outputted from the DC controller compared with the values and calculate the initial error s _0. Subsequently, the change amount β ₁ of the error vector when the _first vector is continuously selected for one control cycle T
Is calculated from the first switching vector and the dynamics of the system, and the variation β ₂ of the error vector with respect to the second vector is similarly calculated. Then, the switching timing α ^* from the first vector to the second vector is determined using the calculated initial error s ₀ and the error change amounts β ₁ and β ₂ corresponding to the first and second switching vectors. Calculate as follows. A = (β ₂ −2β ₁ ) ^T (β ₂ −β ₁ ) B = (β ₂ + 2s ₀ ) ^T (β ₂ −β ₁ ) α ₀ = B / A (2) where x ^T y Represents an inner product operation of a vector, and A represents α ₀
Is the denominator of the division for calculating, and B is the numerator of the division. Next, based on the calculated result, the optimum switching timing α ^* is determined as follows in several cases. If A> 0: if α ₀ ≦ 0, then α ^* = 0 if 0 <α0 <1, then α ^* = α ₀ if α0 ≧ 1, if α ^* = 1, if A ≦ 0: α _{If 0} ≦ １ ／, α ^* = 1 if α ₀ > １ ／, α ^* = 0 (3) That is, based on the sampling time,
The time of switching from the first vector to the second vector is t ^* = α ^* T. In the present invention, since the first vector is switched to the second vector at the timing determined as described above, the AC current distortion factor can be minimized without impairing the followability. FIG. 1 is a diagram showing a configuration of an AC controller according to an embodiment of the present invention. In the figure, 1 is the three-phase voltage type converter bridge, 3 is an AC controller, 4 is a DC controller, 12 is a dq converter for calculating an effective current id and a reactive current iq from an AC current, and 13 is A power factor controller 14 for providing tan (φ ^* ) for power factor control, the first of which causes the active current id and the reactive current iq to follow the active current command value id ^* and the reactive current command value iq ^* , A means for selecting the second vector. 15 is a means for calculating the change amounts β ₁ and β ₂ of the error vector in one control cycle, and 16 is a means for calculating the optimum switching timing α ^* based on the equations (2) and (3). FIG. 2 is a main circuit configuration diagram of the three-phase voltage type converter shown in FIG. 1. In FIG. 2, 1 is a three-phase voltage type converter bridge, 2 is a three-phase AC power supply, 9 is an AC reactor, and 10 is an AC reactor. DC capacitor, 11 is a DC load,
L indicates the value of the AC reactor, and C indicates the value of the DC capacitor. Hereinafter, the operation principle of the AC controller shown in FIG. 1 will be described. A. Calculation of Error Vector Variations β ₁ , β ₂ The one-period error variation calculation means 15 in FIG. 1 calculates the above β ₁ , β ₂ as follows. In FIG. 2, Vs = [vd, vq] ^T (4) is the power supply voltage obtained by dq conversion from the three-phase amount, and is = [id, iq] ^T (5) is the power supply current obtained by dq conversion. Assuming that the DC voltage is V _D , the circuit equation can be expressed by the following equations (6) and (7). L (did / dt) = vd + ωLiq -V D ud (6) L (did / dt) = vq -ωLid -V D uq (7) where, omega power angular frequency, also ud, uq is a three-phase volume Are the elements obtained by dq conversion of the switching vector defined by the following equation (8). u = [u ₁ , u ₂ , u ₃ ] ^T where In order to pay attention to the movement within one control cycle, the time axis is normalized by the control cycle T so that τ = t / T.
The circuit equations of equations (6) and (7) can be expressed as the following equations (9) and (10), respectively. did / dτ = (vd + ωLiq -V D ud) T / L (9) diq / dτ = (vq -ωLid -V D uq) T / L (10) alternating current command value is ^* (τ) = [id ^* (Τ), iq
^* (Τ)] ^T , the current error (s = [sd, sq]
^T ) is defined as s (τ) = is ^* (τ) −is (τ) (11) By differentiating the current error s with respect to τ and substituting the equations (9) and (10), the following equations (12) and (13) are obtained. ^{dsd / dτ = id * '-id} ' = id * '+ (V D ud -vd -ωLiq) T / L (12) dsq / dτ = iq *' -iq '= iq *' + (V D uq - vq + ωLid) T / L (13) where id ^* ', iq ^* ', id ', iq' represent the derivative of id ^* , iq ^* , id, iq, respectively. Here, if a new variable β = (βd, βq) is defined as in the following equations (14) and (15), the error equation can be expressed as in the following equations (16) and (17). ^{βd = id * '+ (V} D ud -vd -ωLiq) T / L (14) βq = iq *' + (V D ud -vq + ωLid) T / L (15) dsd / dτ = βd (16) dsq / Dτ = βq (17) Here, if the control cycle T is sufficiently short, the motion of the current error can be approximated to a linear motion.
8). s (τ) = s (0) + τβ (18) That is, sd (τ) = sd (0) + τβd (19) sq (τ) = sq (0) + τβq (20) Here, s (0) is an initial error when the switching vector u starts to operate, and τ represents a time during which the vector lasts. When a certain switching vector u lasts for one control cycle (τ = 1), the change amount of the error s becomes exactly equal to β from the above equation (18). That is, in the current control of the voltage-type converter, when the selected vector is output for one control cycle, the change amount of the error vector can be obtained by the above equations (14) and (15). B. Calculation of optimal switching timing Next, the optimal switching timing calculating means 16 in FIG.
Will be described. The following explanation is
Applicable regardless of three-phase or two-phase. In addition, it is assumed that the error change amounts (β ₁ , β ₂ ) corresponding to the first and second vectors are obtained by the equations (14) and (15). In this embodiment, an appropriate evaluation function is selected in order to determine the switching timing of the vector.
An optimal switching timing that minimizes the switching timing is determined. Specifically, the switching timing is determined directly using the definition of the total distortion factor as an evaluation function. Hereinafter, details of the method of determining the switching timing of the present embodiment will be described. According to the definition of the total strain rate, the strain D generated during one control cycle is represented by the following equation (21). [Equation 2] In the above, s (τ) is the above-mentioned error variable. Now, assuming that the switching time is α, the error change amount β ₁ acts in the α time zone of the section [0, α],
The error change amount β ₂ operates in the 1-α time zone of the section [α, 1]. Therefore, while 0 ≦ τ ≦ α, the instantaneous value of the error vector can be expressed by the following equation (22). s (τ) = s ₀ + τβ ₁ (0 ≦ τ ≦ α) (22) The error s ₁ at the end point α of the above time zone is s ₁ =
s ₀ + αβ ₁ . Next, the error change amount β ₂ acts, and the instantaneous value of the error vector during α <τ ≦ 1 is expressed by the following equation (23). s (τ) = s ₁ + (τ−α) β ₂ = s ₀ + αβ ₁ + (τ−α) β ₂ (α ≦ τ ≦ 1) (23) The error variables of equations (22) and (23) are By substituting into equation (21), the strain can be expanded as in equation (24) below. [Equation 3] As can be seen from equation (24), the strain D is α
Of the third order polynomial. Since the range of α is limited to the closed section [0, 1], the minimum value of D is the minimum value between (0, 1) or the boundary value (the value of D when α = 0 and α = 1) ). α
The value of D when = 0 is given by the above equation (25) from equation (24). Similarly, the value of D when α = 1 is given by the above equation (26)
It is required as follows. The value of α which is an extreme value (minimum or maximum) is
The equation (24) is differentiated with respect to α, and the differentiated value becomes a value α ₀ at which ₀ . By differentiating equation (24) with respect to α, the following equation (27) is obtained. ∂D / ∂α = (1−α) [α (β ₂ −2β ₁ ) − (β ₂ + 2s ₀ )] ^T (β ₂ −β ₁ ) (27) From equation (27), ∂D / ∂α There are two solutions where = 0, one is 1 and the other is given by the following equation (28). (Equation 4) Whether α _{0 in} equation (28) is a minimum value or a maximum value can be mathematically determined, but this is not necessary as shown below. That is, to summarize the above, the optimal timing α ^*
Is any one of {0, α ₀ , 1}. In order to determine the optimum timing α ^*, it may be considered in several cases while paying attention to the equation (24) for calculating the distortion factor and the features of the solution of the extreme value. As can be seen from equation (24), the distortion D is a third-order polynomial of the argument α. Ignoring the range of α,
If α is an arbitrary variable, the cubic polynomial function generally has two extrema (or no extrema), one being a local maximum and one being a local minimum. (When both overlap, there is no extremum.) Further, it can be determined which of the minimal value or the maximal value appears first from the polarity (overall slope) of the coefficient of the cubic term of the polynomial. Incidentally, when the coefficient of the third order term is negative, the minimum value appears first, and when the coefficient of the tertiary term is positive, the maximum value appears first. Note that one extreme value of the strain D always appears when α = 1. According to the polarity of the third order coefficient and the value of α ₀ , FIG.
It can be seen that there are no cases other than the six cases (a) to (f) shown in FIG. In FIG. 3, (a), (b) and (c) show cases where the third-order coefficient is negative. In this case, FIG.
As is clear from the above, α ^* is as follows. 0 <α ₀ <1, α ^* = α ₀ α ₀ ≦ 0, α ^* = 0 α ₀ ≧ 1, α ^* = 1 On the other hand, in FIG. 3, (d), (d) e),
(F) shows the case where the third order coefficient is positive. In the case of (e) and (f) of FIG. 3, either 0 or 1 is a solution, and α ^* is as follows. When α ₀ ≦ 0, α ^* = 1, when α ₀ > 1, α ^* = 0. In the case of FIG. 3D, it is difficult to intuitively determine which one, but the distortion D (1 ) And strain D at 0
Compare (0). From the above equation (22), D (1)
−D (0) can be arranged as in Expression (29). (Equation 5) The large parenthesis (1 / 3-α ₀ ) in equation (29)
Note that the expression before is negative with the sign of the coefficient of the cubic term of (24) reversed: α ₀ = 1/3, α ^* = 1 α ₀ > 1/3, α ^It can be seen that ^* = 0. This conclusion is shown in FIGS.
The same applies to the case. That is, when the coefficient of the third order term is positive, if α is larger than 1/3, 0 is a solution, and α is 1 /
If it is less than 3, 1 is the solution. In this embodiment, as described above, the corresponding error change amount is calculated for the selected vector and used together with the initial error to determine an appropriate switching timing so as to minimize the total distortion. Things. α ^* is 0
If the solution is between 1, two vectors are output in one control cycle, but in other cases, only one vector is output.
That is, when α ^* is 1, the first vector continues for one control cycle, and the output of the second vector is stopped. When α ^* is 0, the second vector is output directly without outputting the first vector. As a result, useless switching operation that does not contribute to reduction of distortion can be omitted. As described in detail above, according to the present invention,
The following effects can be obtained. (1) Since the total distortion is directly used as an evaluation function to determine the switching timing and its minimization is attempted, a system with a low distortion rate can be obtained. (2) Since a switching operation that does not contribute to the reduction of the distortion factor can be naturally omitted, an efficient system can be obtained in principle. (3) Just by further evaluating the selected vector, it is possible to achieve both high tracking performance and low distortion with the same system without impairing the tracking performance of the original switching vector.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a block diagram showing a configuration of an embodiment of the present invention. FIG. 2 is a diagram showing a main circuit configuration of a three-phase voltage type converter. FIG. 3 is a diagram shown for explaining the principle of the optimum switching timing of the present invention. FIG. 4 is a system block diagram shown for explanation of a conventional technique. FIG. 5 is a control block diagram of a conventional carrier modulation scheme. FIG. 6 is a control block diagram of a conventional tracking system. [Description of Signs] 1 Multi-phase converter bridge 2 Multi-phase AC power supply 3 AC controller 4 DC controller 5 AC amplifier 6 Carrier modulation type carrier modulation unit 7 Tracking type vector selection unit 8 Tracking type hysteresis comparison unit 9 AC reactor DESCRIPTION OF SYMBOLS 10 DC capacitor 11 DC load 12 dq converter 13 Power factor controller 14 Means for selecting two vectors for the present invention 15 Calculation of the amount of changing the error in one cycle corresponding to the vector of the present invention Means 16 Optimal switching timing calculation means of the present invention is polyphase AC current id active current iq reactive current Vs multiphase AC power supply voltage V _D DC output voltage is ^* AC current command value id ^* active current command value iq ^* invalid current command value L Inductance value of AC reactor C DC capacitor value T Control cycle t Time variable τ Time variable β normalized by control cycle Amount s to change corresponding error vector s AC current control error s ₀ AC current control error at start of control cycle

Continuation of the front page (58) Field surveyed (Int.Cl. ⁷ , DB name) H02M 7/21 H02M 7/219

Claims (1)

(1) A converter for converting AC supplied from a polyphase AC power supply into DC, and an effective AC current command value id ^* from a DC voltage output and a DC voltage command value of the converter ^. , A dq converter for dq converting the three-phase AC current to obtain an effective AC current id and a reactive AC current iq, and an effective AC current command value id ^* obtained from the DC controller ^.
And calculates a power factor command disables the AC current command value on the basis of the value iq ^* and calculator, the effective AC current id and the reactive alternating current iq and invalid alternating current command value i respectively the effective AC current command value id ^*
q ^* , two switching vectors are sequentially selected in one control cycle T, and 1 is selected at an appropriate timing α.
A pulse width control converter, comprising: an AC controller that outputs an open / close signal of a semiconductor switch element that constitutes the converter so as to switch from the second switching vector to the second switching vector. The timing α is compared with the AC current command value and the AC current control amount at each sampling point, and a first means for obtaining an error vector s ₀ thereof, and for the selected first and second vectors, When each vector is output during one control cycle, a second means for obtaining the change amount (β ₁ , β ₂ ) of the error vector and a third means for obtaining the determination timing α ^* of the switching timing α And the third means determines the decision timing α ^* of the switching timing α as follows: A = (β ₂ −β ₁ ) ^T (β ₂ −β ₁ ) B = (Β ₂ + 2s ₀ ) ^T (β ₂ −β ₁ ) When α ₀ = B / A, A> 0: If α ₀ ≦ 0, then α ^* = 0 0 <α ₀ <1 , Α ^* = α _{0 If} α ₀ ≧ 1, α ^* = 1 If A ≦ 0: If α ₀ ≦ １ ／, α ^* = 1 If α ₀ > １ ／, then α ^* = 0 ( Where α ^* is the ratio of the switching time to the control cycle T, and the subscript [ ^T ] represents the transposition of the vector).