CN1566965A - Novel method for AC sampling calculation and true effective value sampling integrated circuit based on the same - Google Patents

Abstract

The measuring for alternating current quantity in early time often adopts direct current sampling whose essence is to disperse ac cycle signals into several spots and then to calculate root-mean-square value though integral scattering and which is time waste and needs to be proved by theory and practice. While this invention method is capable of increasing thousand times of speed and is new method with ac sampling. And according to this, a new type of root-mean-square converter can be made. This method is to accomplish the data process of sampling results though supercomputer and the single chip microprocessor or micro control unit only serve as to search the address or as calculation method for comparing and judging data.

Description

A kind of AC sampling New Calculation Method and based on " real effective " of the method sampling integrated circuit

1, field under the present invention

Single-chip microcomputer, integrated circuit, AC sampling, intelligence instrument is measured and control, RMS switching device

2, technical background of the present invention:

In electric system, to the measurement of ac electric, generally adopt the direct current sampling in early days, promptly the DC quantity behind overcommutation is carried out sampled measurements.The method software design is simple, and convenience of calculation only need be done transformation of scale to sampled value and can obtain measured numerical value.But there is certain problem in the direct current method of sampling: measuring accuracy directly is subjected to the influence of rectification circuit, and the rectification circuit difficulty of parameter tuning is influenced greatly by form-factor, especially measures and lags behind real-time difference etc.Owing to have these problems, limited the application of the direct current method of sampling greatly.

The method that another kind is used widely, it is ac sampling method, in ac sampling method, for how calculating effective value, several different methods is then arranged: one again, maximal value (if sampled point is abundant, can think that sampled value is a peak value) with sampled data in the one-period takes advantage of 0.707 to obtain effective value; Its two, obtain effective value with average on duty 1.11 of sampled data in the semiperiod.But development along with Power Electronic Technique, along with various speed governing in the electric system, the widespread use of Power Regulation equipment, the quality of electric network source begins to degenerate, wave form distortion, association's wave component increases, above-mentioned two kinds of methods can not accurately be calculated effective value, for this reason, cause extensive attention in the world, and begin to adopt a kind of method of new calculating effective value, promptly calculate the method for so-called " real effective ", the essence of this method is to calculate its root-mean-square value (RMS) or " real effective " by discrete integration again by the ac cycle signal discrete is become some point values.

Calculate the alternating current sampling method of root-mean-square value (RMS) or " real effective ", in recent years, along with improvement of computer science, though obtained increasing application, still there are serious problems in this method aspect sampling precision and processing speed.Determine the factor of this method sampling precision to mainly contain following 2 points: 1, the sampling precision of each discrete point; 2, the quantity of the sampled point that quilt disperses in each signal period.Previous restraining factors depend on the precision (figure place) of A/D conversion, and a back factor then depends on the travelling speed of microprocessor, and this is recognized by people.Yet increasing of the raising of the sampling precision of each discrete point or figure place needs to consume the more relatively time equally, and this illustrates this method aspect sampling precision and processing speed, or aspect accuracy and real-time, still has serious restriction.More time-consumingly in fact be also, carry out numerical value to a large amount of sampled result and handle that these processing comprise the square root of a large amount of discrete values being asked square and asked a number.Calculate the square root of a number, need especially to expend the plenty of time.A kind of be widely used and comparatively faster method is used to calculate the square root of a number arranged, i.e. newton-pressgang gloomy (Newton-Raphson) method, although this method can be very fast on mathematical computations, it need carry out a large amount of division arithmetics repeatedly.Exactly because there are these problems, people have to select for use the microprocessor with higher travelling speed, perhaps (RMS) switching device of the what is called " real effective " of Zhuan Yong digital processing device (DSP) and various special uses.

Theoretical and practical application proves that the inventive method can improve thousands of times of speed than existing method.

3, content of the present invention

One of feature of this method is that it is a kind of AC sampling New Calculation Method, and can produce a kind of novel " real effective " (RMS) switching device according to this method.This method is that a kind of (as personal computer, PC) numerical value of finishing sampled result is handled, and chip microprocessor or microcontroller are only made the index search of numerical value address, or only makes the computing method that value is relatively judged by high level computer.The principle of the inventive method is to recognize, although in fact the value of tested electric weight may be unlimited many, but for given one-chip computer or analog to digital conversion (A/D) device, the number of numerical value after its conversion then be completely specified can only be limited a plurality of, the number of this numerical value is 256 for 8 A/D, for 10 A/D is 1024, for 12 A/D, then be 4096, or the like.Equally, the number of its " real effective " also must be completely specified can only be limited a plurality of, and be equally respectively 256,1024 and 4096, or the like.

According to this method, under the situation about equating when the time interval of the adjacent double sampling of supposition, voltage or electric current effectively

$I_{\mathrm{RMS}}=\sqrt{\frac{1}{N}\underset{i=1}{\overset{N}{\mathrm{\Σ}}}{X}_i^2}$

Value can be expressed from the next (effective value with electric current is an example):

In the formula: the instantaneous value of i time interval current sample of Xi..........

N.......... the sampling number in the one-period

Algorithm according to proposed by the invention is rewritten into following formula

$I_{\mathrm{RMS}}=\sqrt{\frac{{\mathrm{<figure-callout\; id="1"\; label="X"\; filenames="A0312941300071.png"\; state="\{\{state\}\}">X</figure-callout>}}_1^2}{N}+\frac{{\mathrm{<figure-callout\; id="2"\; label="X"\; filenames="A0312941300071.png"\; state="\{\{state\}\}">X</figure-callout>}}_2^2}{N}+\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;+\frac{X_N^2}{N}}$

So, the present invention--------a kind of AC sampling New Calculation Method comprises following steps:

1, continuous sampling N point, and preserve the each point sampled value, X1, X2 ... .Xn.

2, obtain in the radical sign every, promptly each sampled instantaneous value square N/one.Because for the A/D converter spare of determining because its figure place determines, so X1, X2 ... the value of .Xn also must be completely specified, is example with 8 precision A/D, and they must be 0 to 255, or 16 systems 0 to FF.So the invention provides an auxiliary calculation procedure with the computing machine high level language, it can provide rapidly from each sampled instantaneous value X1, X2 ... Xn searches N/one of its square, promptly

$\frac{{\mathrm{<figure-callout\; id="1"\; label="X"\; filenames="A0312941300071.png"\; state="\{\{state\}\}">X</figure-callout>}}_1^2}{N},\frac{{\mathrm{<figure-callout\; id="2"\; label="X"\; filenames="A0312941300071.png"\; state="\{\{state\}\}">X</figure-callout>}}_2^2}{N},\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;,\frac{X_N^2}{N}$

3, single-chip microcomputer or other microcontroller according to step 2 by table look-up or value relatively obtain each sampled instantaneous value square N/value, and finish the addition summation of its N subitem.

4, single-chip microcomputer or other microcontroller obtain according to step 3 and and provided by the present invention one with the given number table of comparisons subduplicate of the auxiliary calculation procedure of computing machine high level language with it by tabling look-up or value is relatively obtained " real effective " or the root-mean-square value (RMS) of this current sample.

Two of the feature of this method is, it comprises a computer applied algorithm, and this program is write as (Application Program Interface is seen accompanying drawing 1) with VB or VC.This program provides an interface, figure place (8 when the selected A/D conversion of user, 10,12, ... etc.), and in the input one-period after the number (N) of sampled point, this program can provide respectively: 1, once the analog to digital conversion value Xi of sampling and this value square N/one's (1/N* (Xi*Xi)) the 16 hex value tables of comparisons; 2, the table of comparisons of a number and its subduplicate 16 hex value.This table of comparisons will be used to single-chip microcomputer or other microcontroller (MCU), and make it to obtain rapidly " real effective " or the root-mean-square value (RMS) of AC signal.

Three of the feature of this method is can make a kind of " real effective " or root-mean-square value (RMS) sampling integrated circuit with higher price-performance ratio based on this method.This integrated circuit is a kind of hydrid integrated circuit that comprises analog to digital conversion (A/D) module and be solidified with microcontroller (MCU) module of this patent algorithm.

Four of the feature of this method is, a kind of method of searching fast that the inventive method is implemented for the multibyte data table is promptly implemented the branch byte for the multiword joint number and searched, and the method for searching to low byte from high byte.With the square root of searching the double word joint number is example, and in the following table, right side one row are double word joint numbers of one 16 system, and left side one row are its corresponding roots:

0toFFH square (H)

0x01 ；0x1

0x02 ；0x4

0x03 ；0x9

0x04 ；0x10

0x05 ；0x19

0x06 ；0x24

0x07 ；0x31

0x08 ；0x40

0x09 ；0x51

0x0A ；0x64

0x0B ；0x79

0x0C ；0x90

0x0D ；0xA9

0x0E ；0xC4

0x0F ；0xE1

0x10 ；0x100

0x11 ；0x121

0x12 ；0x144

0x13 ；0x169

0x14 ；0x190

0x15 ；0x1B9

0x16 ；0x1E4

0x17 ；0x211

0x18 ；0x240

0x19 ；0x271

0x1A ；0x2A4

0x1B ；0x2D9

From above numerical table as can be seen, when high byte is zero, the scope of the value of its root be 0x01 to 0x0F, and when high byte is 1, the scope of the value of its root be 0x10 to 0x16, when high byte was 2, the scope of the value of its root then was that 0x17 is to 0x1B.

When if single-chip microcomputer adopts value relatively to search, at first than the scope of higher bit with definite root, relatively low byte finds definite root again.

Be further described below in conjunction with 1 pair of this patent method of accompanying drawing.

The combobox on accompanying drawing 1 " A/D figure place: " right side is a figure place of choosing A/D conversion for the user, and the user can select 8,10,12, etc." sampling number/cycle; " text box on right side imports for the user and the phase sampling number is set weekly uses.Several buttons in the lower left corner of program interface then can be finished different calculating respectively.Show in the big text box on program interface right side then be when the A/D figure place be 8, the phase sampling number is 40 o'clock weekly, by " Xi ←---→ (Xi*Xi)/and N " behind the button, the form that the little core of Gong the U.S. (MicroChip) PIC of company Series of MCU (MCU) chip that Automatic Program generates directly calls.Below be the partial content of this form:

8 A/D, the tables of data during N=40

； Xi*Xi/N(H) ；Xi(H) Xi(D) Xi*Xi/N(D)

retlw 0x00 ；0x01 1 .025

retlw 0x00 ；0x02 2 .1

retlw 0x00 ；0x03 3 .225

retlw 0x00 ；0x04 4 .4

retlw 0x01 ；0x05 5 .625

retlw 0x01 ；0x06 6 .9

retlw 0x01 ；0x07 7 1.225

retlw 0x02 ；0x08 8 1.6

retlw 0x02 ；0x09 9 2.025

retlw 0x02 ；0x0A 10 2.5

retlw 0x03 ；0x0B 11 3.025

retlw 0x04 ；0x0C 12 3.6

retlw 0x04 ；0x0D 13 4.225

retlw 0x05 ；0x0E 14 4.9

retlw 0x06 ；0x0F 15 5.625

retlw 0x06 ；0x10 16 6.4

retlw 0x07 ；0x11 17 7.225

retlw 0x08 ；0x12 18 8.1

retlw 0x09 ；0x13 19 9.025

retlw 0x0A ；0x14 20 10

retlw 0x0B ；0x15 21 11.025

retlw 0x0C ；0x16 22 12.1

retlw 0x0D ；0x17 23 13.225

retlw 0x5DD ；0xF5 245 1500.625

retlw 0x5E9 ；0xF6 246 1512.9

retlw 0x5F5 ；0xF7 247 1525.225

retlw 0x602 ；0xF8 248 1537.6

retlw 0x60E ；0xF9 249 1550.025

retlw 0x61A ；0xFA 250 1562.5

retlw 0x627 ；0xFB 251 1575.025

retlw 0x634 ；0xFC 252 1587.6

retlw 0x640 ；0xFD 253 1600.225

retlw 0x64D ；0xFE 254 1612.9

retlw 0x65A ；0xFF 255 1625.625

As can be seen from the table, when signal is smaller, if " Xi*Xi/N " shows with 16 system numerical tables, error is bigger owing to round, when noticing this point, Single Chip Microcomputer (SCM) program should read decimal system floating number when signal is smaller, converts thereof into 16 system numbers again after to its N item summation.

" all side and with its root " in the program interface be when single-chip microcomputer or other microcontroller according to step 2 by table look-up or value relatively obtain each sampled instantaneous value square N/value, and after finishing the addition summation of its N subitem, this equal side and remove to ask for its root as a number, this root is net result " real effective ", or root-mean-square value (RMS).

Claims (4)

1, a kind of AC sampling New Calculation Method and based on " real effective " of the method sampling integrated circuit, it is characterized in that, this method be a kind of by high level computer (as personal computer, PC) numerical value of finishing sampled result is handled, and chip microprocessor or microcontroller are only made the index search of numerical value address, or only make the computing method that value is relatively judged.

2, the described a kind of AC sampling New Calculation Method of claim 1 is characterized in that it comprises a computer applied algorithm, and this program is write as with VB or VC.This program provides an interface, figure place (8 when the selected A/D conversion of user, 10,12, ... etc.), and in the input one-period after the number (N) of sampled point, this program can provide respectively: 1, once the analog to digital conversion value Xi of sampling and this value square N/one's (1/N* (Xi*Xi)) the 16 hex value tables of comparisons; 2, the table of comparisons of a number and its subduplicate 16 hex value.This table of comparisons will be used to single-chip microcomputer or other microcontroller (MCU) and make it to obtain rapidly " real effective " or the root-mean-square value (RMS) of AC signal.

3, the described a kind of AC sampling New Calculation Method of claim 1 is characterized in that, can make a kind of " real effective " or root-mean-square value (RMS) sampling integrated circuit with higher price-performance ratio based on this method.This integrated circuit is a kind of hydrid integrated circuit that comprises analog to digital conversion (A/D) module and be solidified with microcontroller (MCU) module of this patent algorithm.

4, the described a kind of AC sampling New Calculation Method of claim 1 is characterized in that, for a kind of method of searching fast that the multibyte data table is implemented, promptly implement the branch byte and search for the multiword joint number, and the method for searching to low byte from high byte.

Images