JP2006317435A - Phase/amplitude detector, and method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a phase/amplitude detector capable of detecting a phase and/or an amplitude of an AC electric quantity within one cycle of the AC electric quantity. <P>SOLUTION: The AC electric quantity is read by an input part 11 at predetermined sampling intervals, and time-series data of the AC electric quantity read by the input part 11 is stored in a time-series data storage part 12. When data of a half cycle of the AC electric quantity is obtained, a double speed discrete Fourier transform means 17 changes the sign of the time-series data for the half cycle of the AC electric quantity to obtain the time-series data for the remaining half cycle to create the time-series data for one cycle of the AC electric quantity. Then it is Fourier-transformed to obtain a Fourier transform signal. A phase/amplitude calculation means 14 calculates the phase and/or the amplitude of the AC electric quantity based on the Fourier transform signal obtained by the means 17. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a phase / amplitude detection apparatus and method for detecting the phase and / or amplitude of an AC electric quantity (voltage, current).

  For example, when an inverter-type distributed power supply is connected to a power system for operation, the output current of the inverter is controlled in synchronization with the voltage phase on the system side. For this reason, it is required to accurately detect the voltage phase and amplitude of the power system. As a method for detecting the phase of the AC voltage, there are a method using an integration circuit that resets at a voltage zero cross point, a method using a PLL (Phase Locked Loop), a method using a discrete Fourier transform DFT (Discrete Fourier Transform), and the like. When the discrete Fourier transform DFT is used, the amplitude can be detected simultaneously.

  FIG. 18 is a configuration diagram illustrating an example of a phase / amplitude detection apparatus using a discrete Fourier transform DFT. FIG. 18 shows a case where an AC voltage signal v (t) is input as an AC electricity amount. Hereinafter, the case of the AC voltage signal v (t) as the AC electricity amount will be described. The AC voltage signal v (t), which is the AC electricity quantity of the power system, is read at a predetermined sampling interval by the input unit 11 and sequentially stored in the time-series data storage unit 12.

  Assuming that there are M pieces of time-series data for one period of AC electricity, the time-series data storage unit 12 has a capacity to store at least M pieces of time-series data for one period of AC electricity. is doing. FIG. 18 shows a case where M pieces of data are updated and stored in the time-series data storage unit 12. That is, when the latest sampling data v (k) is input, the time-series data is sequentially shifted one by one, the oldest sampling data v {k− (M−1)} is discarded, and the sampling data v { k- (M-2)} becomes the oldest sampling data v {k- (M-1)}.

The discrete Fourier transform means 13 takes in time-series data v (k) to v {k− (M−1)} for one period of the AC voltage signal v (t), performs Fourier transform, and generates the AC voltage signal v Real-time DFT processing synchronized with one cycle of (t) is performed. Real-time DFT processing is formulated as shown in equation (1).

As shown in equation (1), the Fourier transform signal _{a n (k) + jb n} (k) is expressed by a complex number. Therefore, the phase and amplitude of the AC voltage signal v (t) can be obtained from this. Phase and amplitude calculation unit 14 obtains a phase θ and an amplitude V _m of the AC electric quantity on the basis of the discrete values Fourier transform means 13 obtained in the Fourier transform signal _{a n (k) + jb n} (k).

In addition, during the sampling interval Δt after the AC voltage signal v (t) is taken in, the phase / amplitude calculation means 14 executes calculation processing and outputs the final phase θ _m at the next sampling timing. Therefore, during the sampling interval Δt, the phase θ is advanced by the phase Δθ corresponding to the sampling interval Δt. Therefore, the phase correction unit 24 performs phase correction in anticipation of this, and at the time when the arithmetic processing is completed. determination of the phase θ _m.

  19 shows a conventional phase / amplitude detection with respect to a change in the phase of the AC voltage signal v (t) of the power system, and FIG. 20 shows a conventional phase / amplitude detection with respect to a change in the amplitude of the AC voltage signal v (t) of the power system. It is a characteristic view at the time of detecting a phase and amplitude with an apparatus. 19 and 20 show characteristics in the case of the fundamental wave component (n = 1). The AC voltage signal v1 (t) by the real-time DFT processing follows the change of the phase or amplitude of the AC voltage signal v (t) with a delay of one cycle of the AC voltage signal v (t). As shown in the equation (1), the real-time DFT processing is performed using M time-series data v (k) to v {k− (M−1)} for one cycle of the AC voltage signal v (t). This is because the phase and amplitude are obtained by using. The same applies when the phase and amplitude change simultaneously.

Next, FIG. 21 shows an example of a conventional phase / amplitude detection apparatus that employs a recursive algorithm for reducing the amount of calculation. When employing a recursive algorithm to the phase and amplitude detecting apparatus shown in FIG. 18, 1 Fourier transform signal before sampling interval _{a n (k-1) +} jb n (k-1) conversion signal for storing the storage unit 15 is provided, with one sampling interval before the Fourier transform signal _{a n (k-1) +} jb n (k-1), the recursive discrete Fourier transform unit 16 by the Fourier transform signal a _n (k) + jb _n (k) is obtained. That is, when the formula (1) is modified, the formula (2) is obtained.

Fourier transform signal _{a n (k) + jb n} (k) is the (2) As can be seen from equation 1 the Fourier transform signal before sampling interval _{a n (k-1) +} jb n (k-1), It can be calculated using the latest sampling data v (k) and the sampling data v (kM) one cycle before. Here, the third term 2 / M × exp (j2nπ (kM) / M) * v (kM) in equation (2) is the second term 2 / M × exp (j2nπk / M) * v (k) Since the data is M before, storing the second term of equation (2) for each sampling eliminates the need to calculate the third term of equation (2) each time after M sampling. That is, in equation (1), M product-sum operations are required for M time-series data for each sampling. In equation (2), one product-sum operation is performed on the latest sampling data v (k). Since the calculation is sufficient, the product-sum calculation amount can be reduced to 1 / M with respect to the normal DFT processing.

Note that when starting up the phase / amplitude detection device, since no time-series data is accumulated, an initial value (for example, 0) is first set in the converted signal storage unit 15. The recursive discrete Fourier transform signal obtained by the Fourier transform unit 16 a _n (k) + jb _n (k) of saving the converted signal storage unit 15, one sampling interval before the Fourier transform signal a at the next sampling By setting _n (k-1) + jb _n (k-1), a correct Fourier transform signal can be obtained after one cycle.

Here, since the third term of the equation (2) is M data before the second term, the second term of the equation (2) is used instead of the time-series data storage unit 12 as shown in FIG. An integration result data storage unit 23 for storing the time series data of the term is provided, and the time series data of the second term of the formula (2) is stored for each sampling, so that after the M sampling, the formula (2) The third term is obtained. Therefore, it is not necessary to calculate the third term of the equation (2) each time. The calculation multiplying the factor 2 / M is them eventually (2) performed when calculating the Fourier transform signal _{a n (k) + jb n} (k) in the recursive discrete Fourier transform means 16 based on the equation However, when v (k) is obtained in the input unit 11, it can be executed prior to all operations, and an optimum method is selected so that the calculation error is minimized according to the calculation processing system. can do.

  Thus, since the reduction of the product-sum calculation amount in the calculation of the Fourier transform signal is a key in speeding up the real-time DFT processing, the product-sum calculation processing is also greatly reduced by a recursive algorithm. In addition, there is an algorithm that can reduce the product-sum calculation amount without using a recursive algorithm (see, for example, Non-Patent Document 1).

  The calculation algorithm for reducing the product-sum calculation amount of the DFT shown in this non-patent document uses the symmetry and similarity of the positive and cosine waveforms, which are Fourier transform coefficients, and calculates the Fourier component every ¼ period. Using the product-sum calculation results of 3/4 cycles ago, 2/4 cycles ago, and 1/4 cycles ago, the newly calculated product-sum calculation results for 1/4 cycle and these previous The product-sum calculation results are added to calculate a Fourier component for one period.

In the case of employing a recursive algorithm, (2) As can be seen from equation for the Fourier transform signal before one sampling interval _{a n (k-1) +} jb n (k-1), latest sampling data A new Fourier transform signal a _n (k) + jb _n (k) is calculated by adding / subtracting the product of v (k) and the multiplier and the integration result based on the sampling data v (kM) of the previous cycle. For this reason, numerical errors accumulate by repeating addition and subtraction, and some of them are reset at appropriate intervals (see, for example, Non-Patent Document 2).

FIG. 23 is the same as the phase / amplitude detection apparatus shown in FIG. The sequential product-sum calculation unit 26 of the accumulation error reset unit 25 calculates 2 / M × exp (j2nπk / M) * v (k) with respect to the AC electric quantity v (k) obtained at each sampling interval Δt. Performs a sequential multiply-add operation to add. That is, for each sampling, the calculation is performed by accumulating 2 / M × exp (j2nπk / M) * v (k) based on the equation (1) and sequentially adding. The sum obtained by the successive product-sum operation means 26 is stored in the reset Fourier transform signal storage unit 27 as a successive Fourier transform signal _anRES (k) + _jbnRES (k). Then, the reset Fourier transform signal storage unit 27 when the number of the AC electric quantity v (t) is the M-number of one period sequentially Fourier transformed signals a _NRES at that time _{(k) + jb nRES (k} ) is accumulated Since the Fourier transform signal does not include an error, it is output to the recursive discrete Fourier transform means 16 as a correction Fourier transform signal.

Thereafter, the successive product-sum operation means 26 _{resets the} successive Fourier transform signal _anRES (k) + _jbnRES (k) (0 + j0) and again obtains the alternating current obtained at each sampling interval Δt as shown in the equation (1). Performs sequential product-sum operation that calculates and adds 2 / M x exp (j2nπk / M) * v (k) to the electric quantity v (k), and the number of AC electric quantity v (t) is one cycle When the M number of successive Fourier transform signals _anRES (k) + _jbnRES (k), that is, Fourier transform signals not including an accumulation error, are reset, the same operation is repeated _thereafter .

Recursive Discrete Fourier transform means 16 Fourier correction from reset Fourier transform signal storage unit 27 converts the signal _{a nRES (k) + jb nRES} (k) inputting a the Fourier transform signal _{a n (k) + jb n} (k) _Is replaced with a correction Fourier transform signal _anRES (k) + _jbnRES (k). Thereby, the recursive discrete Fourier transform means 16 correction Fourier transform signal a _NRES (k) + jb Fourier transform _NRES at (k) for each M sampling (one period) signal _{a n (k) + jb n} (k ) Will be corrected. Therefore, the numerical error due to addition / subtraction can be reset.
2000 IEEJ Power and Energy Conference Document, No.330 (2000), “Computation Reduction Algorithm Suitable for DFT Real-Time Phase Detection Error Correction”, Tsuyoshi Funaki, Atsushi Matsuura, Shunsuke Tanaka Funaki et al. "Countermeasures for errors in real-time DFT phase detection using recursive algorithm", IEEJ Transaction B, vol.121, No.9, pp.1085-1093, 2001.9)

  In a power system in which inverter-type distributed power sources are connected, when the system voltage phase or amplitude changes suddenly, when the inverter control device uses a PLL for system voltage phase detection, two to three cycles for phase detection In the case of using DFT (Discrete Fourier Transform), the amplitude can be detected simultaneously, but it takes one cycle to detect the phase and amplitude. As a result, overcurrent and overvoltage may occur at the connection point between the inverter-type distributed power source and the electric power system and within the inverter until synchronization is achieved. When an overcurrent or overvoltage occurs, for example, the inverter type distributed power supply may have to stop the inverter to protect the conversion element of the inverter.

  Here, in the DFT of Non-Patent Document 1, the amount of product-sum calculation is reduced using the symmetry and similarity of the positive and cosine waveforms, and the Fourier component for one period is calculated at high speed. Since the calculation of the AC voltage phase and amplitude is performed based on the data for the past one cycle of the AC voltage, there is no difference in that it takes one cycle to detect the AC voltage phase and amplitude.

  In addition, the DFT using the recursive algorithm can further reduce the amount of calculation, but the AC voltage phase and amplitude are still calculated based on the data for the past AC voltage cycle. It takes one cycle to detect the voltage phase and amplitude. This is the same even when the detection accuracy is raised by the method of Non-Patent Document 2.

 An object of the present invention is to provide a phase / amplitude detection apparatus and method capable of detecting the phase and amplitude of an AC electric quantity in less than one cycle of the AC electric quantity.

  The phase / amplitude detection device according to the first aspect of the present invention includes an input unit that reads AC electric quantity at a predetermined sampling interval, and a time series data storage unit that stores time series data of AC electric quantity read by the input unit; When the time series data for the half cycle of the AC electricity amount is obtained, the time series data for the half cycle of the AC electricity amount is sign-inverted to estimate the time series data for the remaining half cycle, Based on the double-speed discrete Fourier transform means for creating a time-series data for one period of the AC electric quantity and performing Fourier transform to obtain a Fourier transform signal, and the Fourier transform signal obtained by the double-speed discrete Fourier transform means And a phase / amplitude calculating means for obtaining the phase and / or amplitude of the AC electric quantity.

  The phase / amplitude detection device according to the invention of claim 2 includes an input unit that reads an alternating current electric quantity at a predetermined sampling interval, a conversion signal storage unit that stores a Fourier transform signal before one sampling interval, the latest sampling data, Based on the sampling data before half a cycle and the Fourier transform signal one sampling interval before stored in the converted signal storage unit, the latest sampling data is sign-inverted to estimate the sampling data ahead of the half cycle, Double-speed recursive discrete Fourier transform means for sequentially obtaining a Fourier transform signal, and phase / amplitude calculation for obtaining the phase and / or amplitude of the AC electric quantity based on the Fourier transform signal obtained by the double-speed recursive discrete Fourier transform means. Means.

  A phase / amplitude detection device according to a third aspect of the invention includes an input unit that reads an alternating current electric quantity at a predetermined sampling interval, and a time series data storage unit that saves time series data of the alternating current electric quantity read by the input unit; When the time series data corresponding to 1/4 period of the AC electricity amount is obtained, the phase is shifted by π / 2 based on part or all of the time series data corresponding to 1/4 period of the AC electricity amount. In addition to the complementary data calculation means for calculating time series data for a quarter cycle of the AC electricity quantity, in addition to the time series data for a quarter period of the AC electricity quantity, a quarter period of the AC electricity quantity The time series data obtained by reversing the time series data of minutes, the time series data obtained by the complementary data computing means, and the time series data obtained by sign inverting the time series data obtained by the complementary data computing means, / 4 time series Obtained by a quadruple speed discrete Fourier transform means for generating a time series data for one period of the alternating current electric quantity and performing a Fourier transform to obtain a Fourier transform signal, and the quadruple speed discrete Fourier transform means. And a phase / amplitude calculation means for obtaining the phase and / or amplitude of the AC electric quantity based on the Fourier transform signal.

  A phase / amplitude detection apparatus according to a fourth aspect of the invention comprises an input unit for reading an alternating current electric quantity at a predetermined sampling interval, complementary data calculation means for calculating sampling data having a π / 2 phase shift of sampling data, A transform signal storage unit for storing a Fourier transform signal before the sampling interval; the latest sampling data; the sampling data before ¼ period; the latest sampling data with a phase shift of π / 2 phase; Based on sampling data with two phase shifts and a Fourier transform signal stored one time before the sampling interval stored in the conversion signal storage unit, data obtained by sign-inversion of the latest sampling data, and the latest π / 2 phase shift Sampling data and half-cycle ahead, 1 / 4-cycle ahead, 3 / 4-cycle ahead sampled data A quadruple-speed recursive discrete Fourier transform means for sequentially obtaining Fourier transform signals while estimating the data, and the phase and / or amplitude of the AC electric quantity based on the Fourier transform signals obtained by the quadruple-speed recursive discrete Fourier transform means And a phase / amplitude calculation means for obtaining the value.

  According to a fifth aspect of the invention, in the phase / amplitude detection device according to the third or fourth aspect of the invention, the output signal of the complementary data calculation means suppresses an excessive signal associated with a sudden change in the phase or amplitude of the AC electric quantity. Output through a limiter.

  A phase / amplitude detection apparatus according to a sixth aspect of the present invention is the phase / amplitude detection apparatus according to the second or fourth aspect, wherein the phase / amplitude detection apparatus stores the Fourier transform signal obtained by the double speed recursive discrete Fourier transform means or the quadruple speed recursive discrete Fourier transform means. And a storage error reset means for resetting the numerical error.

  According to a seventh aspect of the present invention, in the phase / amplitude detection device according to any one of the third to fifth aspects, the complementary data calculation means is a time corresponding to a quarter period of the AC electric quantity by a central difference method. Time series data for ¼ period of the AC electric quantity shifted by π / 2 phase is calculated based on part or all of the series data.

  In the phase / amplitude detection method according to the invention of claim 8, the alternating current electricity quantity is read at a predetermined sampling interval, the time series data of the read alternating current electricity quantity is stored, and the time series data for a half cycle of the alternating current electricity quantity is stored. Take out, invert the time series data for the half cycle of the extracted AC electricity amount, estimate the time series data for the remaining half cycle, create time series data for one cycle of AC electricity amount, and perform Fourier transform To obtain a Fourier transform signal, and obtain the phase and / or amplitude of the AC electric quantity based on the obtained Fourier transform signal.

  A phase / amplitude detection method according to a ninth aspect of the present invention is the method of claim 8, wherein the Fourier transform signal before one sampling interval is stored, the latest sampling data, the sampling data before a half cycle, and the one sampling interval. Based on the previous Fourier transform signal, the latest sampling data is subjected to sign inversion to estimate the half-cycle ahead sampling data, and the Fourier transform signal is sequentially obtained.

  The phase / amplitude detection method according to the invention of claim 10 reads an alternating current electric quantity at a predetermined sampling interval, stores time series data of the read alternating electric quantity, and a time series corresponding to a quarter period of the alternating electric quantity. The data is extracted and converted into time-series data corresponding to a quarter period of the first section, and a π / 2 phase calculated based on a part or all of the time-series data corresponding to a quarter period of the extracted AC electric quantity. The shifted time-series data is estimated as time-series data for the quarter period of the second or fourth section, and the sign of the time-series data for the quarter period of the first section is inverted to obtain the first of the third section. Estimated as time-series data for / 4 periods, and time-series data for quarter periods of the second or fourth section are inverted, and the time series for quarter periods of the fourth or second section Estimated as data, each of the first to fourth divisions Based on the time series data for 1/4 period, time series data for one period of AC electricity is created, and Fourier transform is obtained by applying Fourier transform to the created time series data for one period of AC electricity. The phase and / or amplitude of the AC electric quantity is obtained based on the obtained Fourier transform signal.

  The phase / amplitude detection method according to the invention of claim 11 is the method of claim 10, wherein the Fourier transform signal before one sampling interval is stored, the latest sampling data, the sampling data before ¼ period, the latest π / 2 phase-shifted sampling data, π / 2 phase-shifted sampling data before ¼ period, and data obtained by inverting the sign of the latest sampling data based on the Fourier transform signal before the one sampling interval, It is possible to obtain Fourier transform signals sequentially while estimating sampling data of half cycle ahead, 1/4 cycle ahead, and 3/4 cycle ahead from the latest sampling data with a phase shift of π / 2 and data obtained by inverting this sign. Features.

  A phase / amplitude detection method according to a twelfth aspect of the present invention is the method according to the tenth or eleventh aspect, wherein the phase or amplitude of the AC electric quantity is abruptly changed with respect to time series data or sampling data having a π / 2 phase shift. The present invention is characterized by applying a limiter that suppresses an excessive signal associated with.

  The phase / amplitude detection method according to claim 13 is characterized in that a numerical error accumulated in the Fourier transform signal is reset at a predetermined period.

  A phase / amplitude detection method according to a fourteenth aspect of the present invention is the method according to any one of the tenth to twelfth aspects, wherein the phase / amplitude detection method is based on a part or all of the time-series data corresponding to a quarter period of the AC electric quantity. A time-series data for a quarter period of the AC electric quantity with a phase shift of / 2 is calculated by a central difference method.

  According to the present invention, when the phase or amplitude of the AC electrical quantity changes suddenly, the phase and amplitude can be detected at high speed in a half cycle or a quarter cycle, so that control for achieving synchronization can be performed at an early stage. Therefore, it is possible to suppress the occurrence of overcurrent and overvoltage at the connection point between the inverter-type distributed power source and the power system and within the inverter, and to improve the operation continuation performance of the inverter connected to the power system. Thereby, it is possible to reduce the influence on the system due to the sudden dropout of the inverter type distributed power source when the system is shaken such as a sudden phase change at the time of system switching or an instantaneous voltage drop due to a system fault.

  Embodiments of the present invention will be described below. FIG. 1 is a block diagram of a phase / amplitude detection apparatus according to the first embodiment of the present invention. In the first embodiment, a double speed discrete Fourier transform means 17 is provided in place of the discrete Fourier transform means 13 with respect to the conventional example shown in FIG. From the time series data v (k) to v {k- (M / 2-1)} of the AC voltage signal v (t), the remaining half cycle time series data v {k- (M / 2)} to v { k- (M-1)} is estimated and calculated, and the phase of the AC voltage signal v (t) is detected in a half cycle. The same elements as those shown in FIG. 18 are denoted by the same reference numerals, and redundant description is omitted.

  The AC voltage signal v (t), which is the AC electricity quantity of the power system, is read at a predetermined sampling interval by the input unit 11 and is sequentially stored in the time series data storage unit 12 as time series data. Assuming that there are M pieces of time-series data for one period of AC electricity, the time-series data storage unit 12 updates and stores at least M / 2 pieces of time-series data for half periods of AC electricity. In the first embodiment, the time series data of the remaining half cycle is estimated from the time series data of the AC voltage signal v (t) for the half cycle to obtain the time series data for one cycle. Therefore, the necessary data is M / 2 time-series data corresponding to at least a half cycle of the AC electric quantity.

The double-speed discrete Fourier transform means 17 generates time series data v (k) to v {k− (M / 2-1) corresponding to a half cycle of the AC voltage signal every time the latest sampling data v (t) is obtained. }, And the time series data v (k) to v {k- (M / 2-1)} for the half cycle of the AC voltage signal v (t) is inverted and the time series data for the remaining half cycle v {k- (M / 2)} to v {k- (M-1)} are estimated and obtained, and time series data v (k) to v {k for one period of the AC voltage signal v (t) - performing a Fourier transform to create the (M-1)}, obtaining the Fourier transform signal _{a n (k) + jb n} (k). The phase and amplitude calculation unit 14, double speed discrete Fourier transform means Fourier obtained in 17 converts the signal _{a n (k) + jb n} (k) AC voltage signal on the basis of the v phase and amplitude of the (t) Ask.

Next, the contents of the arithmetic processing in the double speed discrete Fourier transform means 17 will be described. First, the basic formula of the real-time DFT processing shown in the equation (1) can be modified as shown in the equation (3) when one cycle is divided into two.

That is, the time series data v (k) to v {k− (M−1)} for one period of the AC voltage signal v (t) is the time series data v (k) for the half period shown in the first term. ~ V {k- (M / 2-1)} and the time series data v {k- (M / 2)} to v {k- (M-1)} for the remaining half period shown in the second term And can be divided into If m ′ = mM / 2 is set in the second term, equation (3) is expressed by equation (4).

Here, if the AC voltage signal v (t) does not include a DC component and even harmonics, the remaining half cycle of the AC voltage signal v (t) The signal is obtained by inverting the sign of a half cycle. When the AC voltage signal v (t) does not include a DC component and even harmonics, the equation (5) is established.

Now, assuming that the AC voltage signal v (t) does not include a DC component and even harmonics, substituting equation (5) into equation (4) yields equation (6).

When replacing m ′ in the second term of the equation (6) with m, the equation (7) is obtained.

Since it is assumed that the AC voltage signal v (t) does not include a DC component and even harmonics, Expression (7) is expressed by Expression (8) except when n is an even number.

The Fourier transform signal a _n (k) + jb _n (k) is represented by a complex number as shown in Equation (8), and the real axis component a is derived from the signal corresponding to the half cycle of the fundamental wave of the AC voltage signal v (t). _n (k) and the imaginary axis component b _n (k) can be estimated. Therefore, it is possible to detect the phase and amplitude at double the speed of normal DFT processing under the condition that the AC voltage signal v (t) does not include a DC component and even harmonics.

  FIG. 2 is a characteristic diagram when the phase and amplitude are detected by the phase / amplitude detection apparatus of the first embodiment with respect to the phase change of the AC voltage signal v (t) of the power system. FIG. 2 shows the characteristics in the case of the fundamental wave component (n = 1), and the AC voltage by the double speed DFT processing according to the first embodiment with respect to the phase change of the AC voltage signal v (t). The signal v2 (t) follows in the half cycle of the AC voltage signal v (t). As shown in the equation (8), the double speed DFT processing according to the first embodiment performs M / 2 time-series data v (k) to v {k for a half cycle of the AC voltage signal v (t). Since the phase and amplitude are obtained using-(M / 2-1)}, it is possible to detect the phase and amplitude at double the speed of normal DFT processing.

 FIG. 3 is a characteristic diagram in the case where the phase and amplitude are detected by the phase / amplitude detection device of the first embodiment with respect to the change in the amplitude of the AC voltage signal v (t) of the power system. The AC voltage signal v2 (t) by the double speed DFT processing according to the first embodiment follows the half cycle of the AC voltage signal v (t) even when the amplitude changes as in the case of the shown phase change. To do.

Further, the double speed DFT processing according to the first embodiment can be used in combination with a recursive algorithm. FIG. 4 is a configuration diagram illustrating an example of the phase / amplitude detection apparatus according to the first embodiment that employs a recursive algorithm for reducing the amount of calculation. When employing a recursive algorithm to the phase and amplitude detecting apparatus shown in FIG. 1, 1 Fourier transform signal before sampling interval _{a n (k-1) +} jb n (k-1) conversion signal for storing the storage unit 15 is provided, the first prior sampling interval Fourier transform signal _{a n (k-1) +} jb using _n (k-1), the Fourier transform by double speed recursive discrete Fourier transform means 21 signals a _n (k ) + jb _n (k).

That is, the recursive discrete Fourier transform means 16 of the conventional phase / amplitude detection device shown in FIG. 21 is replaced with a double speed recursive discrete Fourier transform means 21, and the time-series data storage unit 12 includes at least M / 2 pieces of time-series data corresponding to a half cycle of the AC electricity amount are updated and stored. Double speed recursive discrete Fourier transform means 21 using equation (9), 1 of the previous sampling interval Fourier transform signal _{a n (k-1) +} jb and _n (k-1), latest sampling data v (k ) and, leading to calculation of a Fourier transform signal _{a n (k) + jb n} (k) by using the half cycle preceding sampling data v (kM / 2).

In this case, when starting up the phase / amplitude detection device, since time series data is not accumulated, an initial value (for example, 0 + j0) is first set in the converted signal storage unit 15. Then, the Fourier transform signal an (k) + jbn (k) obtained by the double speed recursive discrete Fourier transform means 21 is stored in the transform signal storage unit 15 and the Fourier transform signal a one sampling interval before the next sampling is performed. By setting _n (k-1) + jb _n (k-1), a correct Fourier transform signal can be obtained after a half cycle.

  Here, the third term 2 × 2 / M × exp (j2nπ (kM / 2) / M) * v (kM / 2) in the equation (9) is expressed by the second term 2 × 2 / M × exp (j2nπk / M) * v (k) M / 2 previous data, so that the time-series data in the second term of equation (9) is stored in place of the time-series data storage unit 12 as shown in FIG. And the time series data of the second term of the equation (9) is stored for each sampling, so that the third term of the equation (9) is obtained after M / 2 sampling. It is done. Therefore, it is not necessary to calculate the third term of the equation (9) each time.

  When a recursive algorithm is adopted, numerical errors associated with addition / subtraction are accumulated, and it is necessary to reset the accumulation of numerical errors associated with the addition / subtraction. Therefore, for example, the accumulation of numerical errors is reset at a predetermined cycle by the method described in Non-Patent Document 2.

FIG. 6 is the same as the phase / amplitude detector shown in FIG. 4 except that an accumulation error resetting means 25 for resetting numerical errors is added. The double speed sequential product-sum calculation means 28 of the accumulation error reset means 25 calculates 4 / M × exp (j2nπk / M) * v (k) with respect to the AC electric quantity v (k) obtained at every sampling interval Δt. Then, a sequential product-sum operation is performed. That is, for each sampling, an operation is performed in which 4 / M × exp (j2nπk / M) * v (k) is integrated and sequentially added based on the equation (8). The sum obtained by the double-speed sequential product-sum operation means 28 is stored in the reset double-speed Fourier transform signal storage unit 29 as a sequential Fourier transform signal _anRES (k) + _jbnRES (k). Then, the reset double-speed Fourier transform signal storage unit 29, when the number of alternating current electric quantity v (t) is M / 2 for a half cycle, the successive Fourier transform signal _anRES (k) + _jbnRES ( Since k) is a double speed Fourier transform signal that does not include an accumulation error, it is output to the double speed recursive discrete Fourier transform means 21 as a correction Fourier transform signal.

Thereafter, the double-speed sequential product-sum operation means 28 _{resets the} sequential Fourier transform signal _anRES (k) + _jbnRES (k) (0 + j0) and obtains it again every sampling interval Δt as shown in equation (8). 4 / M × exp (j2nπk / M) * v (k) is calculated and added to the AC electrical quantity v (t), and the number of AC electrical quantities v (t) is reduced by half. When a sequential Fourier transform signal _anRES (k) + _jbnRES (k) _{corresponding to} M / 2 for a period, that is, a double-speed Fourier transform signal that does not include an accumulation error is obtained, the reset is performed, and _thereafter the same calculation is repeated.

When the double-speed recursive discrete Fourier transform unit 21 inputs the correction Fourier transform signal a _NRES from resetting double speed Fourier transform signal storage section _{29 (k) + jb nRES (} k) Fourier transform signal a _n (k) + jb _n (k) is replaced with a correction Fourier transform signal _anRES (k) + _jbnRES (k). Thus, double speed recursive discrete Fourier transform unit 21 correcting the Fourier transform signal a _NRES (k) + jb Fourier transform _NRES at (k) for each M / 2 sampling (half cycle) signals a _n (k) + jb _n (k) is corrected. Thus, in the case of the double speed DFT processing, the predetermined cycle to be reset is an integral multiple of a half cycle of the AC voltage signal v (t), and the numerical error due to addition / subtraction can be reset in that cycle.

  When a recursive algorithm is adopted for the double speed DFT processing, the product-sum operation amount can be reduced to 2 / M compared to the double speed DFT processing that does not employ the recursive algorithm. Note that the product-sum calculation amount can be reduced to 1 / M with respect to the normal DFT processing.

  According to the first embodiment, when the double speed discrete Fourier transform means 17 or the recursive algorithm is adopted, the AC voltage signal for a half cycle is obtained by the double speed recursive discrete Fourier transform means 21 by the double speed discrete Fourier transform. The time series data of the remaining half cycle is estimated from the time series data v (k) to v {k- (M / 2-1)}, and the phase and amplitude of the AC voltage signal v (t) in the half cycle Therefore, it is possible to detect the phase and amplitude at double the speed of normal DFT processing. In addition, when the recursive algorithm is employed and the accumulation error resetting means 25 is additionally provided, the numerical error associated with addition / subtraction can be reset, so that the calculation accuracy is improved.

  Next, a second embodiment of the present invention will be described. FIG. 7 is a block diagram of a phase / amplitude detection apparatus according to the second embodiment of the present invention. This second embodiment is different from the first embodiment shown in FIG. 1 in that it replaces the double speed discrete Fourier transform means 17 with a complementary data calculation means 18, a time series complementary data storage unit 30 and a quadruple speed. A discrete Fourier transform means 19 is provided, and a part or all of the time-series data v (k) to v {k− (M / 4-1)} of the AC voltage signal v (t) for ¼ period is left to the rest. 3/4 period time series data v {k- (M / 4)} to v {k- (M-1)} are estimated and calculated to detect the phase and amplitude of the AC voltage signal v (t) It is a thing. The same elements as those shown in FIG. 1 are denoted by the same reference numerals and redundant description will be omitted.

  The AC voltage signal v (t), which is the AC electricity quantity of the power system, is read at a predetermined sampling interval by the input unit 11 and is sequentially stored in the time series data storage unit 12 as time series data. Assuming that there are M time series data for one period of AC electricity, the time series data storage unit 12 updates and stores at least M / 4 time series data for 1/4 period of AC electricity. Let In the second embodiment, the remaining 3/4 period time series data is estimated from a part or all of the 1/4 period AC voltage signal v (t) time series data, This is because the time series data for one cycle is obtained, and the necessary data is at least M / 4 time series data for a half cycle of the AC electricity amount.

  When the complementary data calculation means 18 obtains the time series data v (k) to v {k− (M / 4-1)} for ¼ period of the AC voltage signal v (t), the AC voltage signal v (t) Time-series data having a π / 2 phase shift of time-series data v (k) to v {k− (M / 4-1)} corresponding to a quarter period of the signal v (t) is calculated. The time-series data of the AC voltage signal v ′ (t) with a π / 2 phase shift obtained by the complementary data calculation means 18 is stored in the time-series complementary data storage unit 30. The time-series data of the AC voltage signal v ′ (t) shifted by π / 2 phase is obtained from the time-series data (v {k− (3M / 4)) to v {k for the next ¼ period. -(M-1)}), and time series data for the last quarter period (equivalent to v {k- (M / 4)} to v {k- (M / 2-1)}) This is for use in estimation calculation.

  In addition, the π / 2 phase shift of the time series data v (k) to v {k− (M / 4-1)} for ¼ period of the AC voltage signal v (t) in the complementary data calculation means 18. In generating the components, methods such as differential calculation (= difference calculation), integration calculation, and Hilbert transform can be used. When using differential operation or Hilbert transform, a π / 2 phase is obtained using a part of 1/4 of the time series data v (k) to v {k- (M / 4-1)}. The shifted time-series data can be calculated.

Each time the latest sampling data v (t) is obtained, the quadruple-speed discrete Fourier transform means 19 generates time-series data v (k) to v {k− (1/4) of the AC voltage signal v (t). In addition to M / 4-1)}, when the time series data v (k) to v {k- (M / 4-1)} for ¼ period of the AC voltage signal v (t) is inverted The time series data for the remaining 3/4 period is estimated from the series data, the time series data obtained by the complementary data computing means 18 and the time series data obtained by inverting the sign of the time series data obtained by the complementary data computing means 18. to, Fourier-transform to create the time series data for one period of the AC voltage signal v (t), determining the Fourier transform signal _{a n (k) + jb n} (k). The phase and amplitude calculation unit 14, the phase and amplitude of the AC voltage signal v (t) based on the 4-speed discrete Fourier transform means 19 obtained in the Fourier transform signal _{a n (k) + jb n} (k) Ask.

Next, the contents of the arithmetic processing in the quadruple speed discrete Fourier transform means 19 will be described. First, the basic formula of the real-time DFT processing shown in the equation (1) can be modified as shown in the equation (10) when one cycle is divided into four.

That is, the time-series data v (k) to v {k− (M−1)} for one period of the AC voltage signal v (t) is the time-series data v (1) for the ¼ period shown in the first term. k) to v {k- (M / 4-1)}, the time series data v {k- (M / 4)} to v {k- () for the second quarter of the second section shown in the second term. M / 2-1)}, time series data v {k- (M / 2)} to v {k- (3M / 4-1)} for the quarter period of the third section shown in the third term, It can be divided into time series data v {k− (3M / 4)} to v {k− (M−1)} for the fourth period of the fourth division shown in the fourth term. Here, the fourth term corresponding to the fourth division is used for estimation of time series data for the next quarter period, and the third term corresponding to the third division is the next quarter period (1 The second term corresponding to the second section is used for estimating the time series data for the last quarter period. If m ′ = mM / 4 in the second term, m ″ = mM / 2 in the third term and m ′ ″ = m−3M / 4 in the fourth term, equation (10) is expressed by equation (11). It is.

Here, when the AC voltage signal v (t) does not include a DC component and even harmonics, the third term corresponding to the third segment is 1 of the AC voltage signal v (t) of the first segment. A signal obtained by inverting the sign of / 4 periods. Accordingly, when the AC voltage signal v (t) does not include a DC component and even harmonics, the equation (12) is established.

Now, assuming that the AC voltage signal v (t) does not include a DC component and even harmonics, substituting equation (12) into the third term of equation (11) yields equation (13).

 This makes it possible to estimate time-series data that is 1/2 cycle ahead (for the next 1/4 cycle).

  Also, suppose that the π / 2 phase lead component of the AC voltage signal v (t) is v ′ (t), and the AC voltage signal v (t) does not include a DC component, even harmonics, or 4n−1 harmonics. The second term corresponding to the second section is a signal obtained by inverting the sign of the π / 2 phase advance component v ′ (t) of the AC voltage signal v (t), and Equation (14) is established.

Here, 4n-1 harmonics are not included because the 4n-1 harmonics of the π / 2 phase advance component v ′ (t) of the AC voltage signal v (t) including 4n−1 harmonics are This is because the sign inversion method differs with respect to the fundamental wave. This is because in order to maintain the symmetry by reversing the sign, it is necessary that the harmonics have the same way of reversing the sign with respect to the fundamental wave. Since the 4n + 1 harmonic has the same sign inversion method as the fundamental wave, symmetry can be maintained even if the sign is inverted.

Similarly, the fourth term corresponding to the fourth section is a signal corresponding to ¼ period of the π / 2 phase advance component v ′ (t) of the AC voltage signal v (t), and Equation (15) is established. .

Therefore, when the equation (14) is substituted into the second term of the equation (11), the second term is expressed by the equation (16).

This makes it possible to estimate time-series data that is 3/4 cycles ahead (for the last 1/4 cycle). Further, when the equation (15) is substituted into the fourth term of the equation (11), the fourth term is expressed by the equation (17).

 This makes it possible to estimate time-series data that is 1/4 cycle ahead (for the next 1/4 cycle).

(13), are substituted into (16) and (17) equation (11), the Fourier transform signal _{a n (k) + jb n} (k) is represented by equation (18).

When the second term m ′, the third term m ″, and the fourth term m ′ ″ in the equation (18) are replaced with m, the equation (19) is obtained.

Since it is assumed that the AC voltage signal v (t) does not include a DC component and even harmonics, Expression (19) is expressed by Expression (20) except when n is an even number.

The Fourier transform signal a _n (k) + jb _n (k) is represented by a complex number as shown in the equation (20), and the real axis is derived from a signal corresponding to ¼ period of the fundamental wave of the AC voltage signal v (t). The component a _n (k) and the imaginary axis component b _n (k) can be estimated. Therefore, it is possible to detect the phase and amplitude at a quadruple speed of the normal DFT processing under the condition that the AC voltage signal v (t) does not include the DC component, even harmonics, and 4n-1 harmonics. In the above description, the complementary data calculation unit 18 has been described with respect to the case where the π / 2 phase advance component v ′ (t) of the AC voltage signal v ′ (t) is used. A / 2 phase delay component v ″ (t) can be used in the same manner. In that case, the signs of the equations (14) and (15) are merely inverted.

  Thus, in the second embodiment, among the time series data divided into four in the formula (10), the time series data of the second quarter period is the original signal (input AC voltage). The signal v (t) is estimated to be the same as the π / 2 phase advance component (or π / 2 phase lag component) of the π / 2 phase advance component of the ¼ period time sequence data), and this is 3/4 cycle ahead Time series data (for the last quarter cycle) is assumed, and the time series data for the third quarter cycle is estimated to have been inverted in sign of the original signal, and this is the next cycle (next next) Time series data). The quarter period of the fourth segment is estimated to be the same as the π / 2 phase advance component (or the sign inverted version of the π / 2 phase lag component) of the original signal, and this is the 1/4 cycle ahead (next) Time series data). Therefore, the time series data for one period of the AC voltage signal v (t) can be estimated and calculated from the time series data for a quarter period of the AC voltage signal v (t). Therefore, the phase and amplitude can be detected in a quarter cycle of the AC voltage signal v (t), and the phase and amplitude can be detected at a normal quadruple speed.

  FIG. 8 is a characteristic diagram when the phase and amplitude are detected by the phase / amplitude detection device of the second embodiment with respect to the phase change of the AC voltage signal v (t) of the power system. FIG. 8 shows the characteristics in the case of the fundamental wave component (n = 1), and the AC voltage by the quadruple speed DFT processing according to the second embodiment with respect to the phase change of the AC voltage signal v (t). The signal v4 (t) follows in a quarter cycle of the AC voltage signal v (t). As shown in the equation (20), the quadruple-speed DFT processing according to the second embodiment performs M / 4 time series data v (k) to v for 1/4 period of the AC voltage signal v (t). Since the phase is obtained using {k− (M / 4-1)}, it is possible to detect the phase and amplitude at a quadruple speed of normal DFT processing.

  FIG. 9 is a characteristic diagram in the case where the phase and amplitude are detected by the phase / amplitude detection apparatus of the second embodiment with respect to the change in the amplitude of the AC voltage signal v (t) of the power system. As in the case of, the AC voltage signal v4 (t) obtained by the quadruple-speed DFT processing according to the second embodiment follows the ¼ period of the AC voltage signal v (t) even when the amplitude changes.

The quadruple speed DFT processing according to the second embodiment can also be used in combination with a recursive algorithm. FIG. 10 is a configuration diagram illustrating an example of a phase / amplitude detection apparatus according to the second embodiment that employs a recursive algorithm for reducing the amount of calculation. When employing a recursive algorithm to the phase and amplitude detecting apparatus shown in FIG. 7, 1 Fourier transform signal before sampling interval _{a n (k-1) +} jb n (k-1) conversion signal for storing the storage unit 15 is provided, the first prior sampling interval Fourier transform signal _{a n (k-1) +} jb using _n (k-1), the Fourier transform signal a _n (k by 4x recursive discrete Fourier transform means 22 ) + jb _n (k).

That is, the double-speed recursive discrete Fourier transform means 21 of the phase / amplitude detection apparatus of the first embodiment shown in FIG. The series complement data storage unit 30 is added. 4x Recursive Discrete Fourier transform unit 22 by using the equation (21), leading to calculation of a Fourier transform signal _{a n (k) + jb n} (k).

That is, the Fourier transform signal a _n (k−1) + jb _n (k−1) before one sampling interval, the latest sampling data v (k), and the sampling data v (kM / 4) before ¼ period. ), Π / 2 phase shifted signal v ′ (k) of latest sampling data v (k), and π / 2 phase shifted signal of sampling data v (kM / 4) ¼ period ago v 'to calculate the Fourier transform signal _{a n (k) + jb n} (k) by using the (kM / 4).

In this case, when starting up the phase / amplitude detection device, since time series data is not accumulated, an initial value (for example, 0 + j0) is first set in the converted signal storage unit 15. Then, save Fourier transform signal a _n (k) + jb _n obtained in 4x recursive discrete Fourier transform means 22 (k) to the conversion signal storage unit 15, before one sampling interval at the time of the next sampling Fourier transform with signal _{a n (k-1) +} jb n (k-1), it is possible to obtain a correct Fourier transform signal after 1/4 period. In this case, the complementary data calculation means 18 samples the sampling data with the π / 2 phase shift of the latest sampling data, or the sampling data before the ¼ cycle with a π / 2 phase shift of the sampling data before the 1/4 cycle. Will be calculated.

  Here, the fourth term 2 × 2 / M × exp (j2nπ (kM / 4) / M) * v (kM / 4) in equation (21) is the second term 2 × 2 / M × exp (j2nπk / M ) * v (k) M / 4 previous data. In addition, the fifth term 2 × 2 / M × exp (j2nπ (kM / 2) / M) * v ′ (kM / 4) in equation (21) is the third term 2 × 2 / M × exp (j2nπ (kM / 4) / M) * v ′ (k) M / 4 previous data. Therefore, as shown in FIG. 11, instead of the time-series data storage unit 12, an integration result data storage unit 23 for storing the time-series data of the second term and the time-series data of the third term of the equation (21) is provided. By storing the time-series data of the second and third terms of equation (21) for each sampling, the fourth and fifth terms of equation (21) are obtained after M / 4 sampling. . Therefore, it is not necessary to calculate the fourth term and the fifth term of the equation (21) each time.

  When a recursive algorithm is adopted, numerical errors associated with addition / subtraction are accumulated, and it is necessary to reset the accumulation of numerical errors associated with the addition / subtraction. Therefore, for example, the accumulation of numerical errors is reset at a predetermined cycle by the method described in Non-Patent Document 2.

FIG. 12 is the same as the phase / amplitude detection device shown in FIG. 11 except that an accumulation error resetting means 25 for resetting numerical errors is added. The quadruple-speed sequential product-sum operation means 31 of the accumulation error reset means 25 is 4 for AC electric quantity v (t) obtained at every sampling interval Δt and AC electric quantity v ′ (k) shifted by π / 2 phase. A sequential product-sum operation is performed in which / M {exp (j2nπk / M) * v (k) −exp {j2nπ (kM / 4) / M} * v ′ (k)} is calculated and added. That is, for each sampling, the integration of 4 / M {exp (j2nπk / M) * v (k) −exp {j2nπ (kM / 4) / M} * v ′ (k)} based on equation (20) and Performs subtraction and sequential addition. The sum obtained by the _quadruple- speed sequential product-sum operation means 31 is stored in the reset _quadruple- speed Fourier transform signal storage unit 32 as a sequential Fourier transform signal _anRES (k) + _jbnRES (k). Then, the reset _quadruple- speed Fourier transform signal storage unit 32, when the number of AC electric quantities v (t) is M / 4 for a quarter period, the successive Fourier transform signal _anRES (k) + jb at that time. _{Since nRES} (k) is a _quadruple- speed Fourier transform signal that does not include an accumulation error, it is _{output to the quadruple-} speed recursive discrete Fourier transform means 22 as a correction Fourier transform signal.

Thereafter, the _quadruple- speed sequential product-sum operation means 31 _{resets the} sequential Fourier transform signal _anRES (k) + _jbnRES (k) (0 + j0) and obtains it again at every sampling interval Δt as shown in the equation (20). 4 / M {exp (j2nπk / M) * v (k) -exp {j2nπ (kM / 4) / M} * v '(k)} is added to the AC electric quantity v (t) Incremental product-sum operation is performed, and the number of AC electrical quantities v (t) becomes M / 4 for a quarter period. Sequential Fourier transform signal _anRES (k) + _jbnRES (k), that is, an accumulation error is included If no quadruple-speed Fourier transform signal is found, it is reset, and the same calculation is repeated thereafter.

4x recursive discrete Fourier transform means 22 by entering a correction Fourier transform signal a _NRES from reset quadruple speed Fourier transform signal storage section _{32 (k) + jb nRES (} k) Fourier transform signal a _n (k) + jb _n (k) is replaced with a correction Fourier transform signal _anRES (k) + _jbnRES (k). Thus, 4x recursive discrete Fourier transform means 22 correction Fourier transform signal a _NRES (k) + jb Fourier transform _NRES at (k) for each M / 4 sampling (half cycle) signals a _n (k) + jb _n (k) is corrected. Thus, in the case of quadruple speed DFT processing, the predetermined cycle to be reset is an integral multiple of a quarter cycle of the AC voltage signal v (t), and numerical errors due to addition and subtraction can be reset in that cycle.

  When the recursive algorithm is adopted for the quadruple speed DFT processing, the product-sum operation amount can be reduced to 4 / M with respect to the quadruple speed DFT processing that does not employ the recursive algorithm. Note that the quadruple-speed DFT processing requires a separate algorithm for generating a signal with a π / 2 phase shift of the time series data v (k) compared to the normal DFT. The reduction effect is not so great. Therefore, this method is effective when it is necessary to improve the speed of phase detection.

  Here, the π / 2 phase advance of the time-series data v (k) to v {k− (M / 4-1)} for ¼ period of the AC voltage signal v (t) in the complementary data calculation means 18. When the (delay) component is obtained by difference calculation, there are a backward difference method and a center difference method.

FIG. 13 is a block diagram showing the processing contents of the π / 2 phase advance component by the backward difference method. In the case of the backward difference method, the complementary data calculation means 18 calculates time series data v ′ (k) of the AC voltage signal of the π / 2 phase advance component using the equation (22).

  That is, the complementary data calculation means 18 reads the AC voltage signal v (k) from the time series data of the AC voltage signal v (t) that is read by the input unit 11 at a predetermined sampling interval and sequentially stored in the storage unit 12. The AC voltage signal v (k−1) before one sampling is input, the time series data v ′ (k) of the π / 2 phase advance component is calculated using the equation (22), and the time series complementary data storage unit 30 Save to.

  In the case of this backward difference method, since it is a derivative of the time series data v (k) and the time series data v (k-1) before one sampling, in actuality, Δt from the time of the time series data v (k). This corresponds to a differential coefficient at a time delayed by / 2, so when quadruple-speed DFT is performed based on this, the contraction / extension of the sine wave waveform based on the estimated time series data is repeated, and the calculated Fourier coefficient Vibrates at twice the period of the fundamental wave.

FIG. 14 is a block diagram showing the processing contents of the π / 2 phase advance component by the center difference method. In the case of the center difference method, the complementary data calculation means 18 calculates time series data v ′ (k ′) = v ′ (k−1) of the π / 2 phase advance component using the equation (23).

  That is, the time series data v (k) of the AC voltage signal v (t) read by the input unit 11 at a predetermined sampling interval is input to the complementary data calculation unit 18 and is also input to the data temporary storage unit 33. The data temporary storage unit 33 temporarily stores two continuous time series data v (k-1) and v (k-2). Now, k ′ = k−1. The time series data v (k ′) is sequentially stored in the time series data storage unit 12 and the time series data v (k−2) is output to the complementary data calculation means 18. The complementary data calculation means 18 uses the time-series data v (k) and the time-series data v (k-2), and the time-series data v ′ (k−1) of the π / 2 phase advance component according to the equation (23). ) And is stored in the time-series complementary data storage unit 30.

In the case of this central difference method, since it corresponds to the derivative at the time delayed by the sampling interval Δt, the derivative obtained by the equation (23) is set as v ′ (k ′) = v ′ (k−1). And v (k−1), a quadruple speed DFT is performed to calculate a Fourier coefficient at a time delayed by the sampling interval Δt. Since these arithmetic processes are executed during the sampling interval Δt after the AC signal v (t) is taken in, it is considered that the phase θ is advanced by the phase Δθ corresponding to the sampling interval Δt during this period. The phase correction means 24 (see FIGS. 7, 10, 11, and 12) performs phase correction (correction of 2 × Δθ) corresponding to 2 × Δt, which is twice the sampling interval. Can be detected. That is, when the center difference method is used in the complementary data calculation means 18, the phase correction means 24, instead of correcting the phase Δθ corresponding to the sampling interval Δt that is normally performed, twice the phase correction of 2 × Δθ. However, it is necessary to obtain the phase θ _m at the time when the arithmetic processing is completed.

  According to the second embodiment, the complementary voltage calculation means 18, the quadruple-speed discrete Fourier transform means 19 and the quadruple-speed recursive discrete Fourier transform means 22 using the quadruple-speed discrete Fourier transform make the AC voltage signal for ¼ period. Time series data v (k) to v {k- (M / 4-1)} are used to estimate and calculate the remaining 3/4 period time series data, and the AC voltage signal v (t) is 1/4 period. Therefore, the phase and amplitude can be detected at a speed four times that of normal DFT processing. In addition, when the recursive algorithm is employed and the accumulation error resetting means 25 is additionally provided, the numerical error associated with addition / subtraction can be reset, so that the calculation accuracy is improved.

  Next, a third embodiment of the present invention will be described. FIG. 15 is a block diagram of a phase / amplitude detection apparatus according to the third embodiment of the present invention. In the third embodiment, a limiter 20 is additionally provided between the complementary data calculation means 18 and the time series complementary data storage unit 30 with respect to the second embodiment shown in FIG. is there. The same elements as those in FIG.

  When calculating the component of the alternating voltage signal v (t) shifted by π / 2 phase in the complementary data calculation means 18, if differential calculation is used, if the phase or amplitude of the AC electrical quantity changes suddenly, The differential value becomes an excessive signal, and an error of a component having a phase shift of π / 2 phase of the AC voltage signal v (t) temporarily increases. As a countermeasure, a limiter 20 is provided at the output section of the complementary data calculation means 18 to limit the component of the AC voltage signal v (t) shifted by π / 2 phase by upper and lower limits. The time series data shifted by π / 2 phase of the AC voltage signal v (t) that has passed through the limiter 20 is stored in the time series complementary data storage unit 30. The quadruple-speed discrete Fourier transform means 19 performs quadruple-speed DFT processing based on a component having a phase shift of π / 2 of the AC voltage signal v (t) that has passed through the limiter 20.

  FIG. 16 is a characteristic diagram when the phase and amplitude are detected by the phase / amplitude detection device of the third embodiment with respect to the phase change of the AC voltage signal v (t) of the power system. FIG. 16 shows the characteristics in the case of the fundamental wave component (n = 1), which is obtained by the quadruple speed DFT processing with limiter according to the third embodiment with respect to the phase change of the AC voltage signal v (t). The AC voltage signal v4 ′ (t) smoothly follows the quarter cycle of the AC voltage signal v (t) as compared with the quadruple speed DFT processing without limiter of the second embodiment shown in FIG. Yes.

  FIG. 17 is a characteristic diagram when the phase and amplitude are detected by the phase / amplitude detection device of the third embodiment with respect to the change in the amplitude of the AC voltage signal v (t) of the power system. Similarly to the case of FIG. 9, the AC voltage signal v4 ′ (t) obtained by the quadruple-speed DFT processing with the limiter according to the third embodiment is also applied to the change in the amplitude according to the limiter of the second embodiment shown in FIG. Compared with none 4x DFT processing, it follows smoothly with 1/4 period of AC voltage signal v (t).

 Also in the quadruple speed DFT processing with limiter according to the third embodiment, a recursive algorithm can be used in combination to reduce the amount of calculation.

 According to the third embodiment, since the limiter 20 suppresses the output of an excessive signal, an excessive signal accompanying a sudden change in the phase or amplitude of the AC voltage signal v (t) can be suppressed. It is possible to reduce the error of the component shifted by π / 2 phase of v (t). Thereby, the phase and amplitude can be detected with higher accuracy.

  In each of the embodiments described above, the case of the AC voltage signal v (t) as the AC electric quantity has been described, but it goes without saying that the same applies to the case of the AC current signal i (t).

1 is a configuration diagram of a phase / amplitude detection apparatus according to a first embodiment of the present invention. The characteristic view at the time of detecting a phase and an amplitude with respect to the change of the phase of the alternating voltage signal v (t) of an electric power grid | system with the phase and amplitude detection apparatus of the 1st Embodiment of this invention. The characteristic view at the time of detecting a phase and an amplitude with respect to the change of the amplitude of the AC voltage signal v (t) of the power system by the phase / amplitude detection device of the first exemplary embodiment of the present invention. The block diagram which shows another example of the phase and amplitude detection apparatus of the 1st Embodiment of this invention. The block diagram which shows another example of the phase and amplitude detection apparatus of the 1st Embodiment of this invention. The block diagram which shows another another example of the phase and amplitude detection apparatus of the 1st Embodiment of this invention. The block diagram of the phase and amplitude detection apparatus concerning the 2nd Embodiment of this invention. The characteristic view at the time of detecting a phase and an amplitude with respect to the change of the phase of the alternating voltage signal v (t) of an electric power grid | system with the phase and amplitude detection apparatus of the 2nd Embodiment of this invention. The characteristic view at the time of detecting a phase and amplitude with respect to the change of the amplitude of the alternating voltage signal v (t) of an electric power grid | system with the phase and amplitude detection apparatus of the 2nd Embodiment of this invention. The block diagram which shows another example of the phase and amplitude detection apparatus of the 2nd Embodiment of this invention. The block diagram which shows another another example of the phase and amplitude detection apparatus of the 2nd Embodiment of this invention. The block diagram which shows another another example of the phase and amplitude detection apparatus of the 2nd Embodiment of this invention. The block block diagram which shows the calculation processing content of the advance (delay) component of (pi) / 2 by the backward difference method in the complementary data calculating means in the 2nd Embodiment of this invention. The block block diagram which shows the calculation processing content of the advance (delay) component of (pi) / 2 by the center difference method in the complementary data calculating means in the 2nd Embodiment of this invention. The block diagram of the phase and amplitude detection apparatus concerning the 3rd Embodiment of this invention. The characteristic view at the time of detecting the phase and amplitude with respect to the change of the phase of the alternating voltage signal v (t) of an electric power grid | system with the phase and amplitude detection apparatus of the 3rd Embodiment of this invention. The characteristic view at the time of detecting a phase and amplitude with respect to the change of the amplitude of the alternating voltage signal v (t) of an electric power grid | system with the phase and amplitude detection apparatus of the 3rd Embodiment of this invention. The block diagram which shows an example of the conventional phase and amplitude detection apparatus using discrete type Fourier-transform DFT. The characteristic view at the time of detecting a phase and an amplitude with the conventional phase and amplitude detection apparatus with respect to the change of the phase of the alternating voltage signal v (t) of an electric power system. The characteristic view at the time of detecting a phase and amplitude with the conventional phase and amplitude detection apparatus with respect to the change of the amplitude of the alternating voltage signal v (t) of an electric power system. The block diagram which shows an example of the conventional phase and amplitude detection apparatus which employ | adopted the recursive algorithm for the calculation amount reduction. The block diagram which shows another example of the conventional phase and amplitude detection apparatus which employ | adopted the recursive algorithm for the calculation amount reduction. The block diagram which shows another example of the conventional phase and amplitude detection apparatus which employ | adopted the recursive algorithm for the calculation amount reduction.

Explanation of symbols

DESCRIPTION OF SYMBOLS 11 ... Input part, 12 ... Time series data preservation | save part, 13 ... Discrete Fourier transform means, 14 ... Phase / amplitude calculation means, 15 ... Conversion signal preservation | save part, 16 ... Recursive discrete Fourier transform means, 17 ... Double speed discrete Fourier Conversion means, 18 ... complementary data calculation means, 19 ... quadruple speed discrete Fourier transform means, 20 ... limiter, 21 ... double speed recursive discrete Fourier transform means, 22 ... quadruple speed recursive discrete Fourier transform means, 23 ... integration result data Storage unit 24 ... Phase correction unit 25 ... Accumulation error reset unit 26 ... Sequential product-sum calculation unit 27 ... Fourier transform signal storage unit for resetting 28 ... Double speed sequential product-sum calculation unit 29 ... Double speed for resetting Fourier transform signal storage unit, 30... Time series complementary data storage unit, 31... Quadruple speed sequential product-sum operation means, 32... Reset quadruple speed Fourier transform signal storage unit, 33 Data temporary storage unit

Claims (14)

An input unit that reads AC electric quantity at a predetermined sampling interval, a time series data storage unit that saves time series data of AC electric quantity read by the input unit, and time series data for a half cycle of the AC electric quantity When obtained, the time series data for the half period of the AC electric quantity is inverted and the time series data for the remaining half period is estimated to create the time series data for one period of the AC quantity. A double speed discrete Fourier transform means for obtaining a Fourier transform signal by performing Fourier transform, and a phase / amplitude for obtaining the phase and / or amplitude of the AC electric quantity based on the Fourier transform signal obtained by the double speed discrete Fourier transform means. A phase / amplitude detection apparatus comprising a calculation means. An input unit that reads AC electric quantity at a predetermined sampling interval, a conversion signal storage unit that stores a Fourier transform signal before one sampling interval, the latest sampling data, sampling data before a half cycle, and the conversion signal storage unit Double-speed recursive discrete Fourier transform means for sequentially obtaining Fourier transform signals while inverting the sign of the latest sampling data and estimating sampling data ahead of a half cycle based on the stored Fourier transform signals before one sampling interval And phase / amplitude calculation means for obtaining the phase and / or amplitude of the AC electric quantity based on the Fourier transform signal obtained by the double speed recursive discrete Fourier transform means apparatus. An input unit that reads AC electric quantity at a predetermined sampling interval, a time series data storage unit that saves time series data of AC electric quantity read by the input unit, and a time series for a quarter period of the AC electric quantity When data is obtained, time-series data corresponding to a quarter period of the alternating current quantity shifted by a π / 2 phase based on part or all of the time-series data corresponding to a quarter period of the alternating current quantity. Complementary data calculation means for calculating the time series data, in addition to the time series data for a quarter cycle of the AC electricity amount, the time series data obtained by sign-inversion of the time series data for a quarter cycle of the AC electricity amount, Estimating the time series data for the remaining 3/4 period by the time series data obtained by the complementary data computing means and the time series data obtained by sign-inverting the time series data obtained by the complementary data computing means, One cycle of AC electricity A quadruple speed discrete Fourier transform means for creating a series data and applying a Fourier transform to obtain a Fourier transform signal, and the phase and / or amplitude of the AC electric quantity based on the Fourier transform signal obtained by the quadruple speed discrete Fourier transform means A phase / amplitude detection device comprising a phase / amplitude calculation means to be obtained. Input unit for reading AC electric quantity at a predetermined sampling interval, complementary data calculating means for calculating sampling data with a π / 2 phase shift of sampling data, and a conversion signal storage unit for storing a Fourier transform signal before one sampling interval The latest sampling data, the sampling data before ¼ period, the latest sampling data with a phase shift of π / 2, the sampling data with a phase shift of π / 2 before the quarter period, and the conversion signal storage unit Based on the Fourier transform signal stored one time before the sampling interval, the data obtained by sign-inversion of the latest sampling data, the latest sampling data shifted by π / 2 phase, and the data obtained by sign-inversion of the data are half-cycle ahead. Estimate the sampling data of 1/4 cycle ahead and 3/4 cycle ahead, and forward the Fourier transform signal A quadruple speed recursive discrete Fourier transform means to be obtained, and a phase / amplitude calculation means for obtaining the phase and / or amplitude of the AC electric quantity based on the Fourier transform signal obtained by the quadruple speed recursive discrete Fourier transform means. A phase / amplitude detection device. The output signal of the complementary data calculation means is output via a limiter for suppressing an excessive signal accompanying a sudden change in the phase or amplitude of the AC electric quantity. Amplitude detector. 5. An accumulation error resetting unit for resetting a numerical error accumulated in a Fourier transform signal obtained by the double speed recursive discrete Fourier transform means or the quadruple speed recursive discrete Fourier transform means. The phase / amplitude detection device described. The complementary data calculation means is a quarter period of the AC electric quantity shifted by a π / 2 phase based on a part or all of the time-series data for a quarter period of the AC electric quantity by a center difference method. 6. The phase / amplitude detection apparatus according to claim 3, wherein the time-series data is calculated. Reads the AC electricity amount at a predetermined sampling interval, saves the time series data of the read AC electricity amount, extracts the time series data for the half cycle of the AC electricity amount, and extracts the time series for the half cycle of the extracted AC electricity amount. Reversing the sign of the data, estimating the time series data for the remaining half cycle, creating time series data for one cycle of AC electricity, applying Fourier transform, and obtaining the Fourier transform signal, the resulting Fourier transform signal A phase / amplitude detection method characterized in that the phase and / or amplitude of an AC electric quantity is obtained based on The Fourier transform signal before one sampling interval is stored, and the latest sampling data is inverted based on the latest sampling data, the sampling data before the half cycle, and the Fourier transform signal before the one sampling interval, and the half cycle is performed. 9. The phase / amplitude detection method according to claim 8, wherein Fourier transform signals are sequentially obtained while estimating the previous sampling data. Reads the AC electricity amount at a predetermined sampling interval, saves the read AC electricity time-series data, extracts the AC electricity amount 1/4 period time-series data, and extracts the first quarter period Time-series data, and time-series data with a π / 2 phase shift calculated based on part or all of the time-series data corresponding to 1/4 period of the extracted AC electricity amount are classified into the second or fourth segment. Estimated as time-series data for ¼ period, reverses sign of time-series data for ¼ period of the first section, and estimates it as time-series data for ¼ period of the third section. The time series data for the quarter period of the fourth or fourth segment is inverted and estimated as the time series data for the fourth period of the fourth or second quarter, and each of the first to fourth sections AC electricity based on 1/4 time period of time series data The time series data for one period is created, Fourier transform is performed on the created time series data for one period of AC electricity to obtain a Fourier transform signal, and the phase of the AC electricity quantity is obtained based on the obtained Fourier transform signal. And / or a phase / amplitude detection method, wherein the amplitude is obtained. Stores Fourier transform signal one sampling interval before, latest sampling data, sampling data before 1/4 cycle, latest sampling data with π / 2 phase shift, π / 2 phase shift before 1/4 cycle Based on the sampling data obtained and the Fourier transform signal before the one sampling interval, the latest sampling data is obtained by inverting the sign, the latest π / 2 phase-shifted sampling data, and the sign inverting data for a half cycle. 11. The phase / amplitude detection method according to claim 10, wherein Fourier transform signals are sequentially obtained while estimating sampling data of the first quarter period ahead and the third quarter period ahead. 12. The phase according to claim 10, wherein a limiter is applied to time series data or sampling data having a phase shift of π / 2, which suppresses an excessive signal due to a sudden change in the phase or amplitude of the AC electric quantity.・ Amplitude detection method. The phase / amplitude detection method according to claim 9 or 11, wherein a numerical error accumulated in the Fourier transform signal is reset at a predetermined period. Based on a part or all of the time series data for 1/4 period of the AC electricity amount, time series data for 1/4 period of the AC electricity amount shifted by π / 2 phase is calculated by the central difference method. The phase / amplitude detection method according to any one of claims 10 to 12.

Images