CN105973272A - Method for improving electric energy measuring precision - Google Patents

Abstract

The invention discloses a method for improving the electric energy measuring and sampling precision. The method comprises the following steps that multi-periodic-wave sampling is conducted on alternating-current voltage and current signals, sampling values are accumulated to acquire an average value, fast Fourier transform (FFT) is conducted according to the sampling average value, a real part value and an imaginary part value of a voltage phasor and a real part value and an imaginary part value of a current phasor are calculated, the square of the real part value of the voltage phasor and the square of the imaginary part value of the voltage phasor are added, the square of the real part value of the current phasor and the square of the imaginary part value of the current phasor are added, software gain multiple amplification and extraction are conducted to obtain a Um value and an Im value, 4<n+2> Um values and 4<n+2> Im values are continuously sampled and then subjected to bubble sorting from large to small to acquire 4<n> values, the 4<n> values are accumulated and then shifted rightwards by n bits, a M-bit oversampling value is obtained, and then the ADC sampling resolution ratio precision is improved. According to the method, an expensive ADC does not need to be adopted, the hardware cost is low, and the ADC sampling resolution ratio bit number can be increased; the anti-disturbance performance of the system is high, and the measuring precision is high; the method is easy to achieve, accurate and reliable in data and capable of meeting the requirements of an industrial electric energy meter.

Description

A kind of method improving electric energy metrical precision

Technical field

The invention belongs to technical field of electric power, more particularly, to a kind of method improving electric energy metrical precision.

Background technology

In power industry field, a lot of device products are in order to meet the error criterion of industry, frequently with high bit In the off-chip of number or sheet, ADC carries out sampling and changing the signal such as voltage, electric current, and expensive ADC chip increases Add product cost, the most do not improve signal to noise ratio and quantization error.The certain methods that industry is conventional at present measures effect The most not very good, it is impossible to increase certainty of measurement and improve quantization error.

Summary of the invention

For the defect of prior art, it is an object of the invention to provide a kind of side improving electric energy metrical precision Method, it is intended to solve to use the voltage of ac signal, electric current to cause measuring essence to measure ADC in prior art Spend low and that quantization error is big technical problem.

The invention provides a kind of method improving electric energy metrical precision, comprise the steps:

(1) voltage signal and current signal are carried out the sampling of many cycles, and carry out sampled value cumulative asking it to put down Obtain average voltage and current average the most afterwards；

(2) described average voltage and described current average are carried out FFT fast fourier transform respectively, Obtain the real part of voltage phasor, imaginary values and the real part of electric current phasor, imaginary values；

(3) it is added after the value of real part of described voltage phasor and imaginary values being carried out square and obtains U_n, and to U_nEnter After row amplifies, evolution obtains U_m；It is added after the value of real part of described electric current phasor and imaginary values being carried out square and obtains I_n, and to I_nIt is amplified rear evolution and obtains I_m；

(4) step (1)-(3) continuous sampling 4 is repeated⁽ⁿ⁺²⁾Individual U_mAnd I_m, and by 4⁽ⁿ⁺²⁾Individual U_mAnd I_mValue Middle 4 are obtained the most from big to small after sequenceⁿIndividual value, cumulative 4ⁿIndividual value also moves to right n position and obtains M position over sampled values.

Further, described average voltageDescribed current averageWherein, i is a cycle voltage x current sampling number, and N is continuous sampling cycle Number；K is the sequence number of sampling cycle, and U [N] [i] is continuous sampling N number of cycle voltage signal, and I [N] [i] is continuous Sample N number of week signal wave current.

Further, in step (2), described voltage phasor real partVoltage Phasor imaginary part

Further, in step (3), described U_n=Ux*Ux+Uy*Uy, described I_n=Ix*Ix+Iy*Iy, DescribedDescribed

Further, step (4) particularly as follows:

(4.1) continuous sampling 4⁽ⁿ⁺²⁾Individual U_mAnd I_mValue, by 4⁽ⁿ⁺²⁾Individual U_mAnd I_mValue is arranged the most from big to small Sequence, removes the highest by 4ⁿ/ 2 and minimum 4ⁿ/ 2 U_mAnd I_mValue, takes out middle 4ⁿIndividual value；

(4.2) cumulative 4ⁿIndividual U_mAnd I_mValue, moves to right n position, obtains M position over sampled values.

Further, n is the positive integer of integral multiple of 2.

The present invention is applicable to improve ADC precision, and computing is convenient and simple, it is not necessary to increase too many MCU and meter Evaluation time expense, reduces product cost.Use repeatedly the result of accumulation calculating averagely as final result, essence Degree height, eliminates random error；Bubble sort from big to small is used to take intermediate value method, strong anti-interference performance, Remove corrupt data；Move to right n position and improve the accuracy resolution of n position, eliminate the amount in ADC sampling process Change error and SNR (signal to noise ratio)；This method implements simply, and the anti-interference reliability of data is high, real Strong by property.

Accompanying drawing explanation

Fig. 1 is the flowchart of the method improving electric energy metrical precision that the embodiment of the present invention provides；

Fig. 2 is the realization being averaging step in the method improving electric energy metrical precision that the embodiment of the present invention provides Flow chart；

Fig. 3 is FFT computing and amplification step in the method improving electric energy metrical precision that the embodiment of the present invention provides Rapid flowchart；

Fig. 4 is the realization stream of accumulation step in the method improving electric energy metrical precision that the embodiment of the present invention provides Cheng Tu.

Detailed description of the invention

In order to make the purpose of the present invention, technical scheme and advantage clearer, below in conjunction with accompanying drawing and reality Execute example, the present invention is further elaborated.Only should be appreciated that specific embodiment described herein Only in order to explain the present invention, it is not intended to limit the present invention.

The present invention is directed to the existing voltage of ac signal, the inadequate shortcoming of current measurement ADC precision in the art, A kind of a kind of method improving electric energy metrical precision in the case of not being greatly improved MCU and computing overhead is proposed.

A kind of method improving electric energy metrical precision that the present invention provides, as it is shown in figure 1, comprise the following steps:

Step (1): voltage and current signal is carried out that the sampling of many cycles is cumulative asks it average；

According to formulaObtain average voltage；According to formulaObtain current average；

Wherein, i is a cycle voltage x current sampling number, times integer of minimum 32, perfect could describe and hand over Stream cycle；N is continuous sampling cycle；Minimum be 4 integer.U [N] [i] is the N number of cycle of continuous sampling Voltage signal, I [N] [i] is continuous sampling N number of week signal wave current.One cycle adopts i sampled point；Ut[i] Average for N number of voltage cycle is cumulative；It [i] is that N number of electric current cycle is cumulative to average.

Step (2): average voltage and current average are carried out FFT fast fourier transform respectively, Obtain voltage, the real part of electric current phasor, imaginary values；

Wherein, Ut [i], It [i] signal being carried out FFT fast fourier transform, calculating is tried to achieve sampling and is obtained The real part of voltage phasor and imaginary part.Voltage phasor real partVoltage phasor imaginary part

Electric current phasor real part, imaginary part computing formula such as voltage are similar.Wherein Ut [i] is AC signal sampled value, FFT fast fourier transform has the strongest filter capacity, reduces accumulative quantization error, and operation efficiency is high, MCU operational capability expense can be reduced.

Step (3) is run according to FFT and is drawn summed square after voltage, electric current phasor real part, imaginary values, Carry out gain factor amplification evolution again and obtain U_mAnd I_mValue:

(3.1) according to formula U_n=Ux*Ux+Uy*Uy is squared to voltage phasor real part, imaginary values, and According to formula I_n=Ix*Ix+Iy*Iy is squared to electric current phasor real part, imaginary values；

(3.2) carry out again obtaining after gain factor amplifies evolutionWith

Wherein, G is gain factor amplification, the integral multiple of minimum 2, compensating sampling data, reduces Range error；U_mFor exchange fundamental voltage amplitude, I_mFor exchange fundamental current amplitude.

Step (4) continuous sampling 4⁽ⁿ⁺²⁾Individual U_mValue, by 4⁽ⁿ⁺²⁾Individual I_mValue is steeped from big to small and is emitted sequence and ask it 4ⁿIndividual value, cumulative 4ⁿIndividual value moves to right n position and obtains M position over sampled values, improves AD sampling resolution precision.

(4.1) continuous sampling 4⁽ⁿ⁺²⁾Individual U_mAnd I_mValue, by 4⁽ⁿ⁺²⁾Individual U_mAnd I_mValue emits the most from big to small Bubble sequence, removes the highest by 4ⁿ/ 2 and minimum 4ⁿ/ 2 U_mAnd I_mValue, takes out middle 4ⁿIndividual value.Bubbling is arranged Sequence method can eliminate accidental error, and the n value taken is the biggest, and data are the most stable, and anti-interference is the best, but meeting Increase MCU computing overhead and time.N takes the integral multiple of minimum 2.

(4.2) cumulative 4ⁿIndividual U_mAnd I_mValue, moves to right n position, obtains M position over sampled values.

$Ut={\mathrm{\&Sigma;}}_{i=0}^{t-1}{U}_m\&lsqb;t\&rsqb;>>n;$

$It={\mathrm{\&Sigma;}}_{i=0}^{t-1}\mathrm{Im}\&lsqb;t\&rsqb;>>n;$

t=4ⁿ；

M=12+n；

Wherein, the value of Ut, It is M position ADC accuracy value.N is desirable to the additional accuracy figure place obtained, n Being worth the biggest, the precision obtained is the highest, but higher n value can make the too fast and high consumption MCU of ADC sample rate Internal memory, is unfavorable for stability of sampling, so reasonably n value is the most important, typically takes 2 or 4.

The present invention use above method with existing cumulative be averaging method compared with, have following technical effect that

(1) cumulative it be averaging method and be equivalent to low pass filter, the jitter value of randomness can be effectively eliminated, Make sampled signal average, but the precision of conversion will not be improved, and use above method not only to have low-pass filtering Device function, it is also possible to improve ADC conversion accuracy value, without using expensive ADC chip.

(2) the cumulative method that is averaging will not improve ADC quantization error and improve SNR, and uses with top Method can improve the quantization error of ADC sampling, improves SNR (signal to noise ratio).

(3) this method facilitates simple and direct when calculating, and the computing overhead taking MCU is little, is suitable for electricity Power acquisition system uses.

The operation principle of the present invention is to be said by ADC Sample AC voltage, the multiple cycle of current signal Bright, as in figure 2 it is shown, ADC is sampled by RC low pass filter, after sampling, data are read by MCU, It is stored in relief area, 32 some integral multiples of each periodic sampling, sampling time interval 20ms/ (32*n), adopts Sample number of times is the most, and waveform describes the most perfect, but takies the expense of MCU and core buffer is the biggest.

Idiographic flow of the present invention is as follows:

(1) MCU is to be spaced 20ms/ (32*n) time, by the ADC Sample AC electricity of in sheet 12 Pressure, 32 some integer multiple data of electric current one cycle, sampled data stores in MCU array relief area, even Continuous tired note adopts N number of cycle, and by the data accumulation of 32 some integral multiples in N number of cycle, accumulated value is divided by N Obtain average data, to eliminate random error interference.Implement flow process to be referred to shown in Fig. 2.Flat Mean value computation formula includes: voltage is averaged computing formula:Electric current asks flat Mean value computation formula:Wherein, i is 32 some integral multiples of a cycle, at least adopts 32 points；Ut [i] is the meansigma methods array relief area of the N number of cycle of voltage accumulation；It is N number of that It [i] is that electric current adds up The meansigma methods Lou group relief area of cycle.

(2) alternating voltage, current average Ut [i], It [i] are carried out FFT fast fourier transform, obtain Voltage, the real part of electric current phasor, imaginary values.Voltage, electric current phasor real part, imaginary values each respectively square, Voltage, electric current phasor real part, imaginary part are each added, and obtain the signal value after voltage, electric current FFT computing, Signal value is multiplied by software gain amplifier multiple evolution, draws voltage, current amplitude signal.The tool of step (2) Body realizes flow process as shown in Figure 3.

Part computing formula is as follows:

(2.1) to voltage, electric current phasor real part, the squared formula of imaginary values:

Voltage U_n=Ux*Ux+Uy*Uy；Electric current I_n=Ix*Ix+Iy*Iy；

(2.2) gain factor amplification evolution must be worth U_mAnd I_m；Formula: Wherein, G is gain factor amplification, the integral multiple of minimum 2, compensating sampling data, reduces range by mistake Difference.

(3) voltage, current signal amplitude U are obtained by Fig. 2, Fig. 3 steps flow chart_mAnd I_m.Weight continuously Multiple above-mentioned Fig. 2, Fig. 3 steps flow chart, tired note obtains 4⁽ⁿ⁺²⁾Individual U_mAnd I_m, by 4⁽ⁿ⁺²⁾Individual U_mAnd I_mValue Carry out bubble sort the most from big to small, remove the highest and minimum 4ⁿ/ 2 data values, take out 4ⁿIndividual centre Value is cumulative, and (noise jamming such as ADC itself causes for namely shake, power supply to eliminate sampling random error value Shake), accumulated value is moved to right n position, namely improves n position accuracy resolution value.

In sum, a kind of method improving electric energy metrical precision of the present invention, it is adaptable to improve ADC precision, Computing is convenient and simple, it is not necessary to increases too many MCU and calculates time overhead, reducing product cost.Employing is many The result of secondary accumulation calculating is averagely as final result, and precision is high, eliminates random error；Use from greatly to Little bubble sort takes intermediate value method, strong anti-interference performance, removes corrupt data；Move to right n position and improve n position Accuracy resolution, eliminate the quantization error in ADC sampling process and SNR (signal to noise ratio)；This method is real Now get up simple, and the anti-interference reliability of data is high, practical.

As it will be easily appreciated by one skilled in the art that and the foregoing is only presently preferred embodiments of the present invention, and Not in order to limit the present invention, all made within the spirit and principles in the present invention any amendment, equivalent With improvement etc., should be included within the scope of the present invention.

Claims (6)

1. the method improving electric energy metrical precision, it is characterised in that comprise the steps:

(1) voltage signal and current signal are carried out the sampling of many cycles, and carry out sampled value cumulative asking it to put down Obtain average voltage and current average the most afterwards；

(2) described average voltage and described current average are carried out FFT fast fourier transform respectively, Obtain the real part of voltage phasor, imaginary values and the real part of electric current phasor, imaginary values；

(3) it is added after the value of real part of described voltage phasor and imaginary values being carried out square and obtains U_n, and to U_nEnter After row amplifies, evolution obtains U_m；

It is added after the value of real part of described electric current phasor and imaginary values being carried out square and obtains I_n, and to I_nIt is amplified Rear evolution obtains I_m；

(4) step (1)-(3) continuous sampling 4 is repeated⁽ⁿ⁺²⁾Individual U_mAnd I_m, and by 4⁽ⁿ⁺²⁾Individual U_mAnd I_mValue Middle 4 are obtained the most from big to small after sequenceⁿIndividual value, cumulative 4ⁿIndividual value also moves to right n position and obtains M position over sampled values.

2. the method for claim 1, it is characterised in that described average voltageDescribed current average

Wherein, i is a cycle voltage x current sampling number, and N is continuous sampling cycle；K is sampling cycle Sequence number, U [N] [i] is continuous sampling N number of cycle voltage signal, I [N] [i] be continuous sampling N number of cycle electricity Stream signal.

3. the method for claim 1, it is characterised in that in step (2), described voltage phasor Real partVoltage phasor imaginary part

4. the method as described in any one of claim 1-3, it is characterised in that in step (3), described U_n=Ux*Ux+Uy*Uy, described I_n=Ix*Ix+Iy*Iy, describedDescribed G is gain factor amplification.

5. the method as described in any one of claim 1-3, it is characterised in that step (4) particularly as follows:

(4.1) continuous sampling 4⁽ⁿ⁺²⁾Individual U_mAnd I_mValue, by 4⁽ⁿ⁺²⁾Individual U_mAnd I_mValue is arranged the most from big to small Sequence, removes the highest by 4ⁿ/ 2 and minimum 4ⁿ/ 2 U_mAnd I_mValue, takes out middle 4ⁿIndividual value；

(4.2) cumulative 4ⁿIndividual U_mAnd I_mValue, moves to right n position, obtains M position over sampled values.

6. method as claimed in claim 5, it is characterised in that n is the positive integer of the integral multiple of 2.