CN103257271A - Device and method for detecting micro grid harmonic wave and inter-harmonics based on STM32F107VCT6 - Google Patents

Abstract

The invention discloses a device and method for detecting a micro grid harmonic wave and inter-harmonics based on STM32F107VCT6, and belongs to the technical field of power quality detection. The device comprises a signal acquisition module, a signal processing module and a signal display output module. A Hall voltage current sensor is adopted in the signal acquisition module to serve as a collector for collecting power grid three-phase voltage electric current. The signal processing module comprises a switching circuit of voltage and current, an anti-aliasing filter circuit, an electrical level regulator circuit, a core processor STM32F107VCT6 and a processor peripheral circuit. The output end of the electrical level regulator circuit is connected with the input end of an A/D converter of the core processor STM32F107VCT6. The method comprises the processes of collection and processing of signals. The device and method meets the requirement of computational accuracy and computational speed, and meets flexibility and convenience of the structure.

Description

A kind of little mains by harmonics and harmonic wave pick-up unit and detection method based on STM32F107VCT6

Technical field

The invention belongs to the Power Quality Detection technical field, specifically, relate to a kind of device and detection method that can detect microgrid system harmonics, a harmonic wave, more particularly, relate to a kind of little mains by harmonics based on STM32F107VCT6 and harmonic wave pick-up unit and detection method.

Background technology

At present, known harmonic wave pick-up unit is mostly at big electrical network, and is mostly to detect harmonic wave, for the pick-up unit of a harmonic wave and few.IEC6100-4-7 standard and most existing harmonic wave pick-up unit all adopt FFT(Fast Fourier Transform, fast fourier transform to the detection of a harmonic wave at present) algorithm.Yet traditional fft algorithm is more accurately when detecting integral frequency harmonizing wave, but when there was a large amount of non-integer harmonics in electrical network, algorithm will occur than mistake, even can't detect, and causes result of calculation very insincere.Therefore, be badly in need of the detection that a kind of more accurate algorithm improves harmonic wave and a harmonic wave.

Along with carrying forward vigorously of intelligent grid construction, little electric power network technique development is more and more rapider, and little electrical network power quality problem also more and more causes people's attention.Yet harmonic wave is the key index of weighing the electrical network quality of power supply, in order to administer harmonic wave, avoids harmonic wave to inject big electrical network, and it is very necessary that the harmonic wave of little electrical network is detected.The frequent throwing of the distributed power source generation harmonic current of withdrawing the membership in little electrical network, the nonlinear characteristic of power electronic equipment also can produce a large amount of harmonic voltages, batch (-type) energy output power instability can cause voltage fluctuation and flickering, and the little load of little net capacity is changeable.In addition, because little electrical network has two kinds of steady operational status, namely be incorporated into the power networks state and islet operation state, when being incorporated into the power networks, the fundamental frequency of little electrical network can change along with the change of external power grid fundamental frequency; During islet operation, if distributed power source is exerted oneself and load power consumption does not match, also will cause the change of microgrid fundamental frequency.These characteristics of microgrid make the electric current and voltage in the microgrid very easily fluctuate, specified fundamental frequency is not a steady state value, make that electric current and voltage is the non-stationary quantity of state, except the integral frequency harmonizing wave component, also exist a large amount of between harmonic component, totally unfavorable for power grid operation.Power quality problem is the gordian technique of being badly in need of solution in little electrical network, but the present pick-up unit that does not have effective characteristics at little mains by harmonics and design.Therefore, be badly in need of a harmonic wave pick-up unit at microgrid, and along with the continuous popularization of little electrical network, market also presses for this harmonic wave pick-up unit.

China Patent No.: 200910043655.6, open day on November 25th, 2009, the patent document that a name is called the detection method of a kind of harmonic wave and a harmonic parameters is disclosed, it is the deficiency that overcomes existing harmonic wave, harmonic detection method, this invention provides harmonic wave and a harmonic parameters detection method, its main performing step is: 1) according to characteristic sinusoidal, cosine function, by mathematic(al) manipulation the fundametal compoment in voltage or the current signal and each harmonic component are transformed into DC component respectively; 2) extract DC component with low-pass filtering, parameters such as the amplitude of calculating fundametal compoment and each harmonic component and phasing degree; 3) from voltage signal, deduct first-harmonic and harmonic component, only contained voltage or the current signal of each harmonic wave; 4) search element by the amplitude spectrum maximal value and obtain each harmonic parameters.But this method exists accuracy of detection not high, practical operation problem such as have difficulties, and be not suitable for the detection of harmonic wave between little electrical network.

Traditional fft algorithm detect little electrical network between the harmonic wave reason that will certainly produce than mistake be: the spectral leakage of fft algorithm can produce harmonic wave between puppet, thereby harmonic wave between flooding really, FFT is only applicable to the detection of fixed cycle amount, but in fact little line voltage current fluctuation does not present the fixed cycle regularity.Therefore, be badly in need of the accurate detection that a kind of improved little electrical network harmonic detection method is realized a harmonic wave.

Defective on the above-mentioned algorithm makes also and does not occur at harmonic wave pick-up unit between little electrical network, therefore is badly in need of researching and solving.The technical barrier that will solve harmonic wave detects between little electrical network has: how (1) realizes separating of harmonic wave and a harmonic wave; (2) how to suppress between the harmonic spectrum the phase mutual interference and to the interference of harmonic spectrum: how (3) suppress harmonic components floods a harmonic components.In addition, existing harmonic wave pick-up unit mostly adopts 8 single-chip microcomputers, and minority does not also adopt the pick-up unit of 32 single-chip microcomputers with 16 single-chip microcomputers.8 single-chip microcomputer characteristics are that cost is low, use flexibly, but that shortcoming is arithmetic speed are slow, and generally also need expansion RAM and ROM, hardware circuit is complicated.

Summary of the invention

1, the problem that will solve

At prior art to the detection of little mains by harmonics exist arithmetic speed slow, need expansion RAM and ROM and hardware circuit than complicated problems, the invention provides a kind of little mains by harmonics based on STM32F107VCT6 and harmonic wave pick-up unit and detection method, it structurally satisfies dirigibility and convenience again when satisfying computational accuracy, computing velocity.

2, technical scheme

For addressing the above problem, the present invention adopts following technical scheme.

A kind of little mains by harmonics and a harmonic wave pick-up unit based on STM32F107VCT6, the demonstration output module that comprises signal acquisition module, signal processing module and signal adopts the Hall voltage current sensor as the collector of gathering electrical network three-phase voltage electric current in the described signal acquisition module; Described signal processing module comprises change-over circuit, anti-aliasing filter circuit, level adjusting circuit, core processor STM32F107VCT6 and the processor peripheral circuit of electric current and voltage; The collection terminal of the Hall voltage current sensor of described signal acquisition module is connected with electrical network, and output termination signal processing module passes through current-to-voltage converting circuit, anti-aliasing filter circuit and level adjusting circuit successively in signal processing module; The A/D converter input end of the output termination core processor STM32F107VCT6 of described level adjusting circuit, LCD display is inserted in the periphery of STM32F107VCT6.

Preferably, described processor peripheral circuit comprises power circuit, reset circuit and crystal oscillating circuit; Described power circuit provides the power supply of the required different voltages of total system work by voltage transformation; Adopt the chip hand-reset mode that resets in the described reset circuit; Described crystal oscillating circuit provides clock signal for single-chip microcomputer.

Preferably, described output display module comprises LCD display, keyboard, PC communication module.

Preferably, also include keyboard circuit in the described output display module, keyboard is to connect by keyboard circuit with being connected of core processor STM32F107VCT6.

In order to be used device of the present invention, harmonic wave pick-up unit designed system software of the present invention mainly comprises data sampling, data processing, man-machine interface and communication, mainly adopts Keil development environment C language development instrument.The algorithm aspect proposes the harmonic detection method based on accurate synchronized sampling reconstruction signal on the basis of fft algorithm, this algorithm is by harmonic wave and a harmonic components in the separation signal, reach and suppress the purpose that both disturb mutually, can realize the accurate measurement of harmonic wave and a harmonic parameters simultaneously.Since between the existence of harmonic wave, the system voltage current signal no longer is cycle stably, but non-stationary signal.The traditional FFT algorithm has very mistake for the analysis of non-stationary signal, and " elimination method " detects harmonic wave and a harmonic wave also can occur than mistake in microgrid, harmonic content is many than electrical network between in the reason microgrid, can not ignore a harmonic wave during calculating to influence of harmonic.The present invention proposes the detection algorithm based on accurate synchronized sampling reconstruction signal, and this method is applicable to the non-synchronous sampling situation, so the harmonic wave pick-up unit can adopt fixed sampling frequency that electric power signal is sampled, thereby guarantees the validity of each sampled data.

A kind of little mains by harmonics and harmonic detection method the steps include:

(1) satisfying under the situation of nyquist sampling theorem, with fixed sampling frequency power system signal is carried out equal interval sampling, equation is namely arranged

K·T _s＝P·T ₀ （1-1）

Set up, wherein T _sBe the sampling period, T ₀Be the signal period, K is the sampled point number, and P is the signal period number, and for non-synchronous sampling, obviously P is not integer;

(2) adopt adaptive tracing digital notch filter elimination sample sequence x (KT _s) in fundametal compoment, obtain new sample sequence x'(KT _s), therefore from x (KT _s) in deduct x'(KT _s), can obtain fundametal compoment x ₀(KT _s); Compare with traditional notch filter, the adaptive tracing digital notch filter makes also accurately filtering when fundamental frequency fluctuates owing to combine adaptive filtering mode near 50Hz, and the frequency response of ideal digital notch filter is

$H\left(e^{\mathrm{j\ωt}}\right)=\left\{\begin{array}{cc}1& {\mathrm{\ω}\≠\mathrm{\ω}}_0\\ 0& \mathrm{\ω}={\mathrm{\ω}}_0\end{array}\right.---(1-2)$

Wherein e is natural constant, and j is imaginary unit, and ω is frequency, and t is the time, ω ₀Be trap frequency;

For its transfer function H of single trap frequency (z ^-1) have following symmetrical mirror as form

$H\left(z^{-1}\right)=\frac{1+a_1{z}^{-1}+z^{-2}}{1+a_1{\mathrm{pz}}^{-1}+p^2{z}^{-2}}---(1-3)$

In the formula, z is plural variable, p be one near 1 but be slightly less than 1 constant, a ₁Be parameter, determined by trap frequency,

The trap frequency expression formula is

$f_0=\frac{\arccos (-a_1/2)}{2\mathrm{\πT}}---(1-4)$

Wherein T is the sampling period, is obtained a that changes with sampled point by the adaptive least square method ₁(n), then its corresponding trap frequency also is time dependent, is p (n) in order to realize that adaptivity p changes with sampled point; Thus, x (KT _s) in the H (z of each sampled point signal correspondence ^-1) all can be different, realized the self-adaptation adjustment, namely obtain fundamental signal by original signal and filtered signal subtraction;

(3) utilize linear interpolation method to calculate the primitive period of signal, be located at node a=t ₀＜t ₁＜t ₂＜t _nThe functional value at=b place is y ₀, y ₁, y ₂..., y _n, at each minizone [x _j, x _J+1] in replace curve with straight line, then signal passes the time t of threshold value a for the i time _iFor

$t_i=\frac{(a-y_j)}{y_{j+1}-y_j}+t_j---(1-5)$

In like manner can get the time t that signal passes threshold value a for the i+20 time _I+20, the mean value T* in the sampling time of 20 Cycle Lengths of sample sequence x (n) is:

$T*=\frac{1}{20}\left[(t_{i+20}-t_i\right]---(1-6)$

Then T* is the primitive period of signal, and adopting 20 cycles in addition here is in order to calculate the primitive period more accurately;

(4) according to the primitive period of calculating gained, the harmonic component in the former sample sequence is carried out accurate synchronization, obtain reproducing sequence x ^*(k λ _s), P is not integer during non-synchronous sampling, can produce spectral leakage, in order to realize the accurate synchronization of non-synchronous sampling sequence, to sampling period T _sAdjust, utilize T* to calculate accurate synchronized sampling cycle λ _s

${\mathrm{\λ}}_s=\frac{20T^{*}}{L^{*}}---(1-7)$

L wherein ^*Be sampled point number in 20 cycles of new accurate synchronized sampling sequence, get L but can be similar to ^*=L, reducing error, L is the sampled point numbers of crude sampling sequence in 20 signal periods, obtains each sampled point in the accurate synchronized sampling sequence by cubic spline interpolation, is interpolation knot with the sampled point in the former sample sequence, T _sBe the interpolation knot interval, make the starting point of a certain signal period be the sampled point starting point simultaneously, because cubic spline difference functions K (t) is at each interval [x _j, x _J+1] on be the segmentation cubic polynomial, know the second derivative K''(t of cubic spline function K (t)) on each minizone, be an order polynomial, establish K''(t _j)=M _j, K''(t _J+1)=M _J+1, K''(t then) and expression formula is

$K\″\left(t\right)=M_j\frac{t_{j+1}-t}{h_j}+M_{j+1}\frac{t-t_j}{h_j}---(1-8)$

Wherein, h _j=t _J+1-t _j

With K''(t) integration twice, and bring the expression formula that boundary condition gets K (t) into and be

$K\left(t\right)=M_j\frac{{(t_{j+1}-t)}^3}{{6h}_j}+M_{j+1}\frac{{(t-t_j)}^3}{{6h}_j}+(y_j-\frac{1}{6}{M}_j{h}_j^2)\frac{t_{j+1}-t}{h_j}+(y_{j+1}-\frac{1}{6}{M}_{j+1}{h}_j^2)\frac{t-t_j}{h_j}$

(t∈[t _j,t _j+1];j＝0,1,…,n-1) （1-9）

If m accurate synchronous sampling point is between j synchronous sampling point and j+1 synchronous sampling point, then this accurate synchronous sampling point can be calculated by following formula

$K\left({\mathrm{m\λ}}_s\right)=M_j\frac{{(t_{j+1}-{\mathrm{m\λ}}_s)}^3}{{6h}_j}+M_{j+1}\frac{{({\mathrm{m\λ}}_s-t_j)}^3}{{6h}_j}+(y_j-\frac{1}{6}{M}_j{h}_j^2)\frac{t_{j+1}-{\mathrm{m\λ}}_s}{h_j}+(y_{j+1}-\frac{1}{6}{M}_{j+1}{h}_j^2)\frac{{\mathrm{m\λ}}_s-t_j}{h_j}$

(t∈[t _j,t _j+1];j＝0,1,…,n-1) （1-10）

Repeat K (m λ _s) computation process, can obtain L ^*-1 calculated value, this L ^*It is T that-1 accurate synchronous sampling point and starting point have been formed length jointly ^*Accurate synchronized sampling sequence, so far, obtained reconstructed sample sequence x ^*(k λ _s) full detail, i.e. the position of reconstruction signal and sampled value;

(5) utilize the FIR notch filter to separate x ^*(k λ _s) in harmonic wave and a harmonic component, wave filter arranges a plurality of trap frequencies and forms pectinations; Because sample frequency be the integral multiple of this wave filter between cog spacing frequency, so the FIR notch filter has good effect for synchronized sampling, can produce power not lose, and is applicable to accurate synchronized sampling sequence; Because the trap frequency of wave filter is the integral multiple of first-harmonic, therefore can filtering first-harmonic and all integral frequency harmonizing waves;

(6) to harmonic components x between obtaining after the filtering of FIR notch filter ^{* I}(k λ _s), use the parameter of harmonic wave between the two interpolation FFT algorithms calculating add the Hanning window;

(7) from accurate synchronized sampling sequence x ^*(k λ _s) in deduct a harmonic components x ^{* I}(k λ _s), obtain harmonic component x ^{* H}(k λ _s);

(8) to x ^{* H}(k λ _s) carry out DFT or FFT computing, calculate the each harmonic parameter by its result.

Further, described adaptive tracing digital notch filter is for to increase adaptive control on the basis of digital notch filter, and its acceptance has the input signal of trap frequency, obtains a by the least square adaptive algorithm ₁, utilize a ₁Constitute digital notch filter H (z ^-1) fall into except the signal with this trap frequency.

On present detection technique basis, the more superior key problem in technology of this device mainly contains 2 points:

1) effectively limits the leakage of harmonic wave, avoid the loss of harmonic energy.

2) by separating harmonic wave and a harmonic wave, reduce influencing each other between harmonic wave and a harmonic wave.

Above-mentioned key problem in technology point 1 in analyzing big electrical network) put 2 than key problem in technology) even more important, still increase key problem in technology point 2 to some extent in view of harmonic content in the middle of little electrical network) and key problem in technology point 1) be the technical barrier that mainly need solve equally.

For aforementioned techniques key point 1), because harmonic wave and a harmonic wave after separating all are synchronized samplings, therefore eliminate fence effect and the spectral leakage effect that causes owing to the FFT non-synchronous sampling, thereby reached the leakage of effective restriction harmonic wave, reduced the technical requirement of harmonic energy loss as far as possible.

For aforementioned techniques key point 2), proposition is based on the detection algorithm of accurate synchronized sampling reconstruction signal, with the accurate synchronization of non-synchronous sampling signal, reach the purpose that suppresses spectrum leakage and fence effect, algorithm at first adopts the adaptive tracing digital notch filter that the signal that obtains under the non-synchronous sampling is carried out pre-service, other frequency component beyond the filtering fundamental frequency, use with the method for threshold values comparison to filtering signal then and utilize linear interpolation method to obtain the primitive period of signal, according to this primitive period, adopt cubic spline interpolation algorithm reconstruct crude sampling sequence, make reconstruction signal be similar to synchronous sampling signal, reconstruction signal is realized separating of harmonic wave and a harmonic wave by the FIR notch filter subsequently.Because spectrum leakage and fence effect have obtained remarkable inhibition, just can obtain the accurate parameter of each harmonic according to the FFT result of reconstruction signal.This method reaches the purpose that suppresses both phase mutual interference by harmonic wave and a harmonic components in the separation signal, can realize the accurate measurement of harmonic wave and a harmonic parameters simultaneously.

3, beneficial effect

Than prior art, the present invention has following beneficial effect:

(1) the present invention is a kind of little mains by harmonics based on STM32F107VCT6 and a harmonic wave pick-up unit, microgrid self distributed power source structure and major network have been taken into full account to the influence of microgrid, particularly microgrid and major network off-the-line or microgrid are kept the very big situation of unbalance voltage operation microgrid harmonic content of major network, solve an above-mentioned difficult problem well, it can detection system 51 times with interior harmonic wave, between harmonic wave, and can calculate demonstration single harmonic component aberration rate, total percent harmonic distortion, crest factor, active power, numerical value such as reactive power, this pick-up unit size is very suitable for handing in addition, is easy to carry;

(2) the present invention is a kind of little mains by harmonics based on STM32F107VCT6 and a harmonic wave pick-up unit, use on hardware is based on 32 single-chip microcomputer STM32F107VCT6 of Cortex-M3 nuclear, the fast cost of its processing speed is low, in the harmonic wave context of detection superperformance is arranged;

(3) the present invention is a kind of little mains by harmonics and harmonic detection method, the new detection algorithm of accurate synchronized sampling reconstruction signal has been proposed, its design adaptive tracing digital notch filter calculates the primitive period, this design is very targeted for the characteristics of little power network signal cyclic swing, it uses the method for cubic spline functions reconstruction signal to realize the accurate synchronization of nonsynchronous signal, when reducing cost, the hardware lock circuitry phase avoided spectral leakage and fence effect saving, its design FIR notch filter has well been realized separating of harmonic wave and a harmonic wave, avoid the phase mutual interference between harmonic wave and a harmonic wave, improved computational accuracy.

Description of drawings

Fig. 1 is harmonic wave pick-up unit general structure synoptic diagram between the present invention;

Fig. 2 is harmonic wave of the present invention and a harmonic wave algorithm flow synoptic diagram.

Embodiment

Describe the present invention below in conjunction with accompanying drawing.

As accompanying drawing 1 device workflow be: the present invention adopts following technical scheme.

A kind of little mains by harmonics and a harmonic wave pick-up unit based on STM32F107VCT6, the demonstration output module that comprises signal acquisition module, signal processing module and signal adopts the Hall voltage current sensor as the collector of gathering electrical network three-phase voltage electric current in the signal acquisition module; Signal processing module comprises change-over circuit, anti-aliasing filter circuit, level adjusting circuit, core processor STM32F107VCT6 and the processor peripheral circuit of electric current and voltage; The collection terminal of the Hall voltage current sensor of signal acquisition module is connected with electrical network, and output termination signal processing module passes through current-to-voltage converting circuit, anti-aliasing filter circuit and level adjusting circuit successively in signal processing module; The A/D converter input end of the output termination core processor STM32F107VCT6 of level adjusting circuit, LCD display is inserted in the periphery of STM32F107VCT6.Core processor STM32F107VCT6 is produced by group of STMicw Electronics (ST).

The processor peripheral circuit comprises power circuit, reset circuit and crystal oscillating circuit; Power circuit provides the power supply of the required different voltages of total system work by voltage transformation; Adopt the chip hand-reset mode that resets in the reset circuit; Crystal oscillating circuit provides clock signal for single-chip microcomputer.The output display module comprises LCD display, keyboard, PC communication module.Also include keyboard circuit in the output display module, keyboard is to connect by keyboard circuit with being connected of core processor STM32F107VCT6.

The collection terminal of the Hall voltage current sensor of signal acquisition module is connected with electrical network, output termination signal processing module, at first through current-to-voltage converting circuit current signal is transformed into voltage signal, this circuit uses adjustable resistance and voltage stabilizing diode to realize conversion and the voltage clamper of electric current and voltage.Signal enters the anti-aliasing filter circuit after being come out by current-to-voltage converting circuit, taking all factors into consideration this circuit of filter effect and delay time adopts the second order Butterworth filter to carry out analog filtering, filter cutoff frequency is that 2500Hz can elimination be higher than 51 times harmonic wave, the purpose of this place's filtering is to satisfy sampling thheorem, frequency of occurrences aliasing when avoiding sampling during for data acquisition in the back.Enter level adjusting circuit behind the signal process anti-aliasing filter circuit, this circuit is made of two amplifiers and resistance, circuit is supplied with the 1.6V reference voltage by the outside, signal is through behind this circuit, because the ratio circuit that amplifier constitutes and the adjusting of reference voltage, make signal amplitude become original half and waveform integral body is raised.In addition, because the effect of two-way amplifier has formed differential signal, signal is sent into the ADC pin of STM32F107VCT6 subsequently, utilizes the A/D converter of STM32F107VCT6 inside to carry out analog to digital conversion.STM32F107VCT6 has two 12 A/D converters, and totally 18 passages can be worked by two passages simultaneously.Remove in order to meet the requirement of IEC standard, the width that is analysis window is 200ms, and the frequency resolution of DFT is 5Hz, for accuracy in computation and the harmonic wave frequency resolution that improves the primitive period, the sample frequency of setting device is 10240Hz, and each sample sequence length of handling is 4096 points.Signal carries out analytical calculation in STM32F107VCT6 after sample conversion becomes digital signal.

Concrete performing step such as accompanying drawing 2:

A kind of little mains by harmonics and harmonic detection method the steps include:

(1) satisfying under the situation of nyquist sampling theorem, with fixed sampling frequency power system signal is carried out equal interval sampling, equation is namely arranged

K·T _s＝P·T ₀ （1-1）

Set up, wherein T _sBe the sampling period, T ₀Be the signal period, K is the sampled point number, and P is the signal period number, and for non-synchronous sampling, obviously P is not integer;

(2) adopt adaptive tracing digital notch filter elimination sample sequence x (KT _s) in fundametal compoment, obtain new sample sequence x'(KT _s), therefore from x (KT _s) in deduct x'(KT _s), can obtain fundametal compoment x ₀(KT _s); Compare with traditional notch filter, the adaptive tracing digital notch filter makes also accurately filtering when fundamental frequency fluctuates owing to combine adaptive filtering mode near 50Hz.The adaptive tracing digital notch filter at this place is for to increase adaptive control on the basis of digital notch filter, and its acceptance has the input signal of trap frequency, obtains a by the least square adaptive algorithm ₁, utilize a ₁Constitute digital notch filter H (z ^-1) fall into except the signal with this trap frequency.The frequency response of ideal digital notch filter is

$H\left(e^{\mathrm{j\ωt}}\right)=\left\{\begin{array}{cc}1& {\mathrm{\ω}\≠\mathrm{\ω}}_0\\ 0& \mathrm{\ω}={\mathrm{\ω}}_0\end{array}\right.---(1-2)$

Wherein e is natural constant, and j is imaginary unit, and ω is frequency, and t is the time, ω ₀Be trap frequency;

For its transfer function H of single trap frequency (z ^-1) have following symmetrical mirror as form

$H\left(z^{-1}\right)=\frac{1+a_1{z}^{-1}+z^{-2}}{1+a_1{\mathrm{pz}}^{-1}+p^2{z}^{-2}}---(1-3)$

In the formula, z is plural variable, p be one near 1 but be slightly less than 1 constant, a ₁Be parameter, determine that by trap frequency the trap frequency expression formula is

$f_0=\frac{\arccos (-a_1/2)}{2\mathrm{\πT}}---(1-4)$

Wherein T is the sampling period, is obtained a that changes with sampled point by the adaptive least square method ₁(n), then its corresponding trap frequency also is time dependent, and in order to realize adaptivity, p is p (n) with the sampled point variation, thus, and x (KT _s) in the H (z of each sampled point signal correspondence ^-1) all can be different, realized the self-adaptation adjustment, namely obtain fundamental signal by original signal and filtered signal subtraction;

(3) utilize linear interpolation method to calculate the primitive period of signal, be located at node a=t ₀＜t ₁＜t ₂＜t _nThe functional value at=b place is y ₀, y ₁, y ₂..., y _n, at each minizone [x _j, x _J+1] in replace curve with straight line, then signal passes the time t of threshold value a for the i time _iFor

$t_i=\frac{(a-y_j)}{y_{j+1}-y_j}+t_j---(1-5)$

In like manner can get the time t that signal passes threshold value a for the i+20 time _I+20, the mean value T* in the sampling time of 20 Cycle Lengths of sample sequence x (n) is:

$T*=\frac{1}{20}\left[(t_{i+20}-t_i\right]---(1-6)$

Then T* is the primitive period of signal, and adopting 20 cycles in addition here is in order to calculate the primitive period more accurately;

(4) according to the primitive period of calculating gained, the harmonic component in the former sample sequence is carried out accurate synchronization, obtain reproducing sequence x ^*(k λ _s), P is not integer during non-synchronous sampling, can produce spectral leakage, in order to realize the accurate synchronization of non-synchronous sampling sequence, to sampling period T _sAdjust, utilize T* to calculate accurate synchronized sampling cycle λ _s

${\mathrm{\λ}}_s=\frac{20T^{*}}{L^{*}}---(1-7)$

L wherein ^*Be sampled point number in 20 cycles of new accurate synchronized sampling sequence, get L but can be similar to ^*=L, reducing error, L is the sampled point numbers of crude sampling sequence in 20 signal periods, obtains each sampled point in the accurate synchronized sampling sequence by cubic spline interpolation, is interpolation knot with the sampled point in the former sample sequence, T _sBe the interpolation knot interval, make the starting point of a certain signal period be the sampled point starting point simultaneously, because cubic spline difference functions K (t) is at each interval [x _j, x _J+1] on be the segmentation cubic polynomial, know the second derivative K''(t of cubic spline function K (t)) on each minizone, be an order polynomial, establish K''(t _j)=M _j, K''(t _J+1)=M _J+1, K''(t then) and expression formula is

$K\″\left(t\right)=M_j\frac{t_{j+1}-t}{h_j}+M_{j+1}\frac{t-t_j}{h_j}---(1-8)$

Wherein, h _j=t _J+1-t _j

With K''(t) integration twice, and bring the expression formula that boundary condition gets K (t) into and be

$K\left(t\right)=M_j\frac{{(t_{j+1}-t)}^3}{{6h}_j}+M_{j+1}\frac{{(t-t_j)}^3}{{6h}_j}+(y_j-\frac{1}{6}{M}_j{h}_j^2)\frac{t_{j+1}-t}{h_j}+(y_{j+1}-\frac{1}{6}{M}_{j+1}{h}_j^2)\frac{t-t_j}{h_j}$

(t∈[t _j,t _j+1];j＝0,1,…,n-1) （1-9）

If m accurate synchronous sampling point is between j synchronous sampling point and j+1 synchronous sampling point, then this accurate synchronous sampling point can be calculated by following formula

$K\left({\mathrm{m\λ}}_s\right)=M_j\frac{{(t_{j+1}-{\mathrm{m\λ}}_s)}^3}{{6h}_j}+M_{j+1}\frac{{({\mathrm{m\λ}}_s-t_j)}^3}{{6h}_j}+(y_j-\frac{1}{6}{M}_j{h}_j^2)\frac{t_{j+1}-{\mathrm{m\λ}}_s}{h_j}+(y_{j+1}-\frac{1}{6}{M}_{j+1}{h}_j^2)\frac{{\mathrm{m\λ}}_s-t_j}{h_j}$

(t∈[t _j,t _j+1];j＝0,1,…,n-1) （1-10）

Repeat K (m λ _s) computation process, can obtain L ^*-1 calculated value, this L ^*It is T that-1 accurate synchronous sampling point and starting point have been formed length jointly ^*Accurate synchronized sampling sequence, so far, obtained reconstructed sample sequence x ^*(k λ _s) full detail, i.e. the position of reconstruction signal and sampled value;

(5) utilize the FIR notch filter to separate x ^*(k λ _s) in harmonic wave and a harmonic component, wave filter arranges a plurality of trap frequencies and forms pectinations; Because sample frequency be the integral multiple of this wave filter between cog spacing frequency, so the FIR notch filter has good effect for synchronized sampling, can produce power not lose, and is applicable to accurate synchronized sampling sequence; Because the trap frequency of wave filter is the integral multiple of first-harmonic, therefore can filtering first-harmonic and all integral frequency harmonizing waves;

(6) to harmonic components x between obtaining after the filtering of FIR notch filter ^{* I}(k λ _s), use the parameter of harmonic wave between the two interpolation FFT algorithms calculating add the Hanning window;

(7) from accurate synchronized sampling sequence x ^*(k λ _s) in deduct a harmonic components x ^{* I}(k λ _s), obtain harmonic component x ^{* H}(k λ _s);

(8) to x ^{* H}(k λ _s) carry out DFT or FFT computing, calculate the each harmonic parameter by its result.

By embodiment as can be seen the present invention have following advantage:

1) design of harmonic wave pick-up unit has taken into full account the characteristics of little mains by harmonics between.

2) simple in structure, overall cost is lower, is easy to realize, owing to adopt the signal reconstruction technology on the algorithm, has saved the phase lock circuitry on the hardware, not only makes the testing result of a harmonic wave more accurate, has also saved cost simultaneously.

3) selected 32 interconnected type single-chip microcomputer STM32F107 that examine based on Cortex-M3 for use at core processor, the high computing velocity of this processor can be given full play to the superiority of signal reconstruction algorithm, and implementation algorithm cooperates with the effective of hardware.

4) adopt the method for signal reconstruction in detection method, overcome the mutual interference mutually of harmonic wave and a harmonic wave and the spectral leakage of traditional FFT algorithm.

5) carry out separating of harmonic wave and a harmonic wave by design FIR wave filter, harmonic wave and a harmonic wave can be calculated separately, avoid the phase mutual interference to improve computational accuracy.

Claims (6)

1. little mains by harmonics and a harmonic wave pick-up unit based on a STM32F107VCT6, comprise the demonstration output module of signal acquisition module, signal processing module and signal, it is characterized in that: adopt the Hall voltage current sensor as the collector of gathering electrical network three-phase voltage electric current in the described signal acquisition module; Described signal processing module comprises change-over circuit, anti-aliasing filter circuit, level adjusting circuit, core processor STM32F107VCT6 and the processor peripheral circuit of electric current and voltage; The collection terminal of the Hall voltage current sensor of described signal acquisition module is connected with electrical network, and output termination signal processing module passes through current-to-voltage converting circuit, anti-aliasing filter circuit and level adjusting circuit successively in signal processing module; The A/D converter input end of the output termination core processor STM32F107VCT6 of described level adjusting circuit, LCD display is inserted in the periphery of STM32F107VCT6.

2. a kind of little mains by harmonics and a harmonic wave pick-up unit based on STM32F107VCT6 according to claim 1, it is characterized in that: described processor peripheral circuit comprises power circuit, reset circuit and crystal oscillating circuit; Described power circuit provides the power supply of the required different voltages of total system work by voltage transformation; Adopt the chip hand-reset mode that resets in the described reset circuit; Described crystal oscillating circuit provides clock signal for single-chip microcomputer.

3. a kind of little mains by harmonics and a harmonic wave pick-up unit based on STM32F107VCT6 according to claim 1 and 2, it is characterized in that: described output display module comprises LCD display, keyboard, PC communication module.

4. a kind of little mains by harmonics and a harmonic wave pick-up unit based on STM32F107VCT6 according to claim 3, it is characterized in that: also include keyboard circuit in the described output display module, keyboard is to connect by keyboard circuit with being connected of core processor STM32F107VCT6.

5. a little mains by harmonics and harmonic detection method the steps include:

(1) satisfying under the situation of nyquist sampling theorem, with fixed sampling frequency power system signal is carried out equal interval sampling, equation is namely arranged

K·T _s＝P·T ₀ （1-1）

Set up, wherein T _sBe the sampling period, T ₀Be the signal period, K is the sampled point number, and P is the signal period number, and for non-synchronous sampling, obviously P is not integer;

(2) adopt adaptive tracing digital notch filter elimination sample sequence x (KT _s) in fundametal compoment, obtain new sample sequence x'(KT _s), therefore from x (KT _s) in deduct x'(KT _s), can obtain fundametal compoment x ₀(KT _s); Compare with traditional notch filter, the adaptive tracing digital notch filter makes also accurately filtering when fundamental frequency fluctuates owing to combine adaptive filtering mode near 50Hz, and the frequency response of ideal digital notch filter is

$H\left(e^{\mathrm{j\ωt}}\right)=\left\{\begin{array}{cc}1& {\mathrm{\ω}\≠\mathrm{\ω}}_0\\ 0& \mathrm{\ω}={\mathrm{\ω}}_0\end{array}\right.---(1-2)$

Wherein e is natural constant, and j is imaginary unit, and ω is frequency, and t is the time, ω ₀Be trap frequency;

For its transfer function H of single trap frequency (z ^-1) have following symmetrical mirror as form

$H\left(z^{-1}\right)=\frac{1+a_1{z}^{-1}+z^{-2}}{1+a_1{\mathrm{pz}}^{-1}+p^2{z}^{-2}}---(1-3)$

In the formula, z is plural variable, p be one near 1 but be slightly less than 1 constant, a ₁Be parameter, determine that by trap frequency the trap frequency expression formula is

$f_0=\frac{\arccos (-a_1/2)}{2\mathrm{\πT}}---(1-4)$

Wherein T is the sampling period, is obtained a that changes with sampled point by the adaptive least square method ₁(n), then its corresponding trap frequency also is time dependent, and in order to realize adaptivity, p is p (n) with the sampled point variation;

(3) utilize linear interpolation method to calculate the primitive period of signal, be located at node a=t ₀＜t ₁＜t ₂＜t _nThe functional value at=b place is y ₀, y ₁, y ₂..., y _n, at each minizone [x _j, x _J+1] in replace curve with straight line, then signal passes the time t of threshold value a for the i time _iFor

$t_i=\frac{(a-y_j)}{y_{j+1}-y_j}+t_j---(1-5)$

In like manner can get the time t that signal passes threshold value a for the i+20 time _I+20, the mean value T* in the sampling time of 20 Cycle Lengths of sample sequence x (n) is:

$T*=\frac{1}{20}\left[(t_{i+20}-t_i\right]---(1-6)$

Then T* is the primitive period of signal;

(4) according to the primitive period of calculating gained, the harmonic component in the former sample sequence is carried out accurate synchronization, obtain reproducing sequence x ^*(k λ _s), P is not integer during non-synchronous sampling, can produce spectral leakage, in order to realize the accurate synchronization of non-synchronous sampling sequence, to sampling period T _sAdjust, utilize T* to calculate accurate synchronized sampling cycle λ _s

${\mathrm{\λ}}_s=\frac{20T^{*}}{L^{*}}---(1-7)$

L wherein ^*Be sampled point number in 20 cycles of new accurate synchronized sampling sequence, get L but can be similar to ^*=L, reducing error, L is the sampled point numbers of crude sampling sequence in 20 signal periods, obtains each sampled point in the accurate synchronized sampling sequence by cubic spline interpolation, is interpolation knot with the sampled point in the former sample sequence, T _sBe the interpolation knot interval, make the starting point of a certain signal period be the sampled point starting point simultaneously, because cubic spline difference functions K (t) is at each interval [x _j, x _J+1] on be the segmentation cubic polynomial, know the second derivative K''(t of cubic spline function K (t)) on each minizone, be an order polynomial, establish K''(t _j)=M _j, K''(t _J+1)=M _J+1, K''(t then) and expression formula is

$K\″\left(t\right)=M_j\frac{t_{j+1}-t}{h_j}+M_{j+1}\frac{t-t_j}{h_j}---(1-8)$

Wherein, h _j=t _J+1-t _j

With K''(t) integration twice, and bring the expression formula that boundary condition gets K (t) into and be

$K\left(t\right)=M_j\frac{{(t_{j+1}-t)}^3}{{6h}_j}+M_{j+1}\frac{{(t-t_j)}^3}{{6h}_j}+(y_j-\frac{1}{6}{M}_j{h}_j^2)\frac{t_{j+1}-t}{h_j}+(y_{j+1}-\frac{1}{6}{M}_{j+1}{h}_j^2)\frac{t-t_j}{h_j}$

(t∈[t _j,t _j+1];j＝0,1,…,n-1) （1-9）

If m accurate synchronous sampling point is between j synchronous sampling point and j+1 synchronous sampling point, then this accurate synchronous sampling point can be calculated by following formula

$K\left({\mathrm{m\λ}}_s\right)=M_j\frac{{(t_{j+1}-{\mathrm{m\λ}}_s)}^3}{{6h}_j}+M_{j+1}\frac{{({\mathrm{m\λ}}_s-t_j)}^3}{{6h}_j}+(y_j-\frac{1}{6}{M}_j{h}_j^2)\frac{t_{j+1}-{\mathrm{m\λ}}_s}{h_j}+(y_{j+1}-\frac{1}{6}{M}_{j+1}{h}_j^2)\frac{{\mathrm{m\λ}}_s-t_j}{h_j}$

(t∈[t _j,t _j+1];j＝0,1,…,n-1) （1-10）

Repeat K (m λ _s) computation process, can obtain L ^*-1 calculated value, this L ^*It is T that-1 accurate synchronous sampling point and starting point have been formed length jointly ^*Accurate synchronized sampling sequence, so far, obtained reconstructed sample sequence x ^*(k λ _s) full detail, i.e. the position of reconstruction signal and sampled value;

(5) utilize the FIR notch filter to separate x ^*(k λ _s) in harmonic wave and a harmonic component, wave filter arranges a plurality of trap frequencies and forms pectinations;

(6) to harmonic components x between obtaining after the filtering of FIR notch filter ^{* I}(k λ _s), use the parameter of harmonic wave between the two interpolation FFT algorithms calculating add the Hanning window;

(7) from accurate synchronized sampling sequence x ^*(k λ _s) in deduct a harmonic components x ^{* I}(k λ _s), obtain harmonic component x ^{* H}(k λ _s);

(8) to x ^{* H}(k λ _s) carry out DFT or FFT computing, calculate the each harmonic parameter by its result.

6. a kind of little mains by harmonics according to claim 1 and harmonic detection method, it is characterized in that: described adaptive tracing digital notch filter is for increasing adaptive control on the basis of digital notch filter, its acceptance has the input signal of trap frequency, obtains a by the least square adaptive algorithm ₁, utilize a ₁Constitute digital notch filter H (z ^-1) fall into except the signal with this trap frequency.

Images