CN1149128A - Apparatus and method for measuring quasi-value - Google Patents

Abstract

A quasi-true value measurer features that during measurement the drift of reference is real-time measured at first and the measured parameter is then corrected according to change of reference, so greatly increasing measurement accuracy. The measurer suitable for measuring DC voltage, frequency, temp and mass is composed of reference voltage generator, D/A converter, inverter, voltage transform synthesizers, A/D converter, computer, control keyboard and display.

Description

Quasi-value measurement mechanism and measuring method

The present invention relates to a kind of measuring method of quasi-value, is a kind ofly can be converted to the new measuring method of the various physical quantitys of voltage to quality and voltage and all.

Current measuring methods is an at first selected known standard amount, goes to measure measured with this known standard then.Weigh tested value according to the measured ratio that accounts for the known standard amount.This measuring method at first is that the variation of assert known standard volume can be ignored, and it can be called absolute method of measurement, assert that promptly standard volume is absolute constant.In fact, selected standard volume also is a variable in test process.

The objective of the invention is at above-mentioned shortcoming, study a kind of brand-new method of testing, this method is applicable to the test of various physical parameters such as voltage, quality, tries one's best near the measuring method of actual value on this method evaluation.

The solution of the present invention is as follows: the present invention proposes method at first will give real-time measurement to the drift of selected standard volume, secondly also will be revised tested parameter according to the variation of standard volume.Measurement mechanism of the present invention is composed as follows: 1. standard capacitance box IC1 is made up of one or more, produce one or more independent standard voltage Va (t) or Va (t), Va ' (t), Va " (t) ...; look the input variable difference, the ternary system is one road Va (t) only.Variable of every increase just increases by one tunnel output; 2. digital to analog converter IC2 under the control of computer circuits, produce one with the variation of measuring number of times the ladder triangular wave voltage Vs of increasing or decreasing, this system also can be replaced by ultralow frequency triangle wave voltage generator circuit; 3. inverter ic 3 is with the ladder triangular wave voltage inversion, to form one and the antipodal ladder triangular wave voltage of IC2 change direction Vs; 4. voltage transformation compositor IC4, in order to Va (t) voltage and Vs voltage are synthesized a new voltage Vb (t), it at first gives enough decay with Vs voltage, and Va (t) voltage given after the trace decay, the stack of two-way voltage is formed Vb (t) voltage, satisfy the condition of Va (t)＞Vb (t); 5. voltage transformation compositor IC5, Va (t) voltage and Vs voltage are carried out synthetic Vc (t) voltage of conversion, this voltage satisfies Va (t)＞Vc (t) condition, can make Vb (t), Vc (t) as far as possible and Va (t) approaching, be condition not produce overflow error.Because Vs and Vs are anti-phase variations, so Vb of Xing Chenging (t) and Vc (t) two voltages have anisotropic variation characteristic thus, Va (t), Vb (t), these three variablees of Vc (t), to such an extent as to n variable also satisfies anisotropic change condition, 2., 3., 4., 5. part is also available anyly has each other the voltage generator that oppositely slowly changes and replaces; 6. analog to digital converter IC6 is used for analog quantity is transformed into corresponding with it digital quantity circuit part, and this circuit can be done standard with any one amount among Va (t), Vb (t), the Vc (t).When detecting, both can carve i at a time detects Vb (t) and Vc (t), Va (t) simultaneously, also can adopt the asynchronous detection mode, add the timesharing converting system at the analog-digital converter circuit input end, be about to Va (t) voltage and all the other n-1 voltage timesharing relatively, require the summation of timesharing detection time will be far smaller than the cycle length that twice adjacent timesharing detects, different detection is required to have proposed the requirement of the detection speed of A-D converter, and this analog to digital converter IC6 tests unknown voltage Vx (t) in the interval that twice timesharing detects; 7. computer circuits IC7, task is: 1. receive the sampled data of analog to digital converter IC6, sampling process is finished under computer control; 2. and control digital to analog converter IC2 circuit and produce ladder triangular wave voltage; 3. data processing, the input data that are about to analog to digital converter IC6 are handled by certain procedure; 4. receive keyboard order and decision measurement and display mode; 8. display circuit IC8 shows measured value and standard value; 9. keyboard IC9 is used for input instruction; 10. power circuit IC10-changes with the accuracy requirement of detection system, and as improving accuracy of detection, then the degree of stability of feed circuit and noise all are controlled on the corresponding grade.Above-mentioned standard capacitance box IC1 or IC1 ', IC2 ' ... IC6 joins with analog to digital converter, and analog to digital converter IC6 and unknown voltage Vx (t) join.Digital to analog converter IC2 input signal and computer circuits IC7 join, and the IC4 of one of output terminal and voltage transformation compositor joins, and the other end of IC4 and analog to digital converter IC6 join, and the common point of voltage transformation compositor IC4 and IC1, IC6 joins.Inverter ic 3 one end IC2 and the common point of putting C4 join, and two IC5 of the other end and voltage transformation compositor joins, and the other end and the IC6 of IC5 join, and IC5 also joins with the common point of IC1 and IC4.Computer circuits IC7 except that joining with supervisory keyboard IC9 and display circuit IC8, also and analog to digital converter IC6 join.

In the method that the present invention proposes, at first must measure in real time the drift of selected standard volume, according to the variation of standard volume, measured parameter be revised then, its sampling and data handling procedure are as follows:

One. selected standard volume is measured in real time:

A. measure for the i+1 time with respect to the i time measurement under the anisotropy condition, sampling process should satisfy Condition.

B. data handling procedure: under the condition that satisfies sampling, the full scale value of n variable measuring system is M.Measure for the first time and remove to measure Vb (1), Vc (1), Va ' (1), Va by Va (1) " (1) ..., measured value is Vb1, Vc1 ..., the value that Va (1) is recorded by another system is Va1.Va1, Vb1, Vc1, Va ' 1, Va " 1 ... identical last digit value unit is arranged.

Set up and measure equation and ask particular solution for the first time:

Va(1)－Vb(1)＝Y1(1)＝M－Vb1

Va(1)－Vc(1)＝Y2(1)＝M－Vc1

Va(1)－Va′(1)＝Y3(1)＝M－V′a1

Va(1)－Va″(1)＝Y4(1)＝M－V″a1

： ：

： ：

Va(1)＋Vb(1)＋Vc(1)＋Va′(1)＋Va″(1)＋…＝R(1)；

Make R (1)=0, ask particular solution to try to achieve its particular solution to be:

Ask

Value: $\widehat{\mathrm{\ϵ}}\left(1\right)=\mathrm{Va}1-\hat{V}a\left(1\right)$ The quasi-value of measuring is respectively for the first time: $\mathrm{Vb}\left(1\right)=\hat{V}b\left(1\right)+\widehat{\mathrm{\ϵ}}\left(1\right);\mathrm{Vc}\left(1\right)=\hat{V}c\left(1\right)+\widehat{\mathrm{\ϵ}}\left(1\right);$ $V{a}^{\′}\left(1\right)=\hat{V}{a}^{\′}\left(1\right)+\widehat{\mathrm{\ϵ}}\left(1\right);V{a}^{\″}\left(1\right)=\hat{V}{a}^{\″}\left(1\right)+\widehat{\mathrm{\ϵ}}\left(1\right);$ , carry out the second time then and measure; Set up measure equation successively, Va (2)-Vb (2)=Y1 (2)=M-Vb2Va (2)-Vc (2)=Y2 (2)=M-Vc2Va (2)-Va ' (2)=Y3 (2)=M-Va ' 2Va (2)-Va " (2)=Y4 (2)=M-Va " 2

： ：

:: Va (2)+Vb (2)+Vc (2)+Va ' (2)+Va " (2)+...=R (2) makes R (2)=0, asks particular solution to try to achieve its particular solution to be:

Ask Value: $\widehat{\mathrm{\ϵ}}\left(2\right)=a\left(2\right)-a^{\′}\left(2\right)$ (the integral multiple part of a (2) expression last digit value unit, a ' (2) expression is less than last digit value unit part) Δ=R (2)-R (1)=n[a (2)-a (1) again]+n[a ' (2)-a ' (1)]=n ξ (1)+its ξ of ψ (1) (1)=[a (2)-a (1)], ψ (1)=n[a ' (2)-a ' (1)] satisfying under the sampling condition:

Wherein ξ (1) should satisfy | Δ | and be the condition of minimum, utilize formula: $\widehat{\mathrm{\ϵ}}\left(2\right)=a\left(2\right)+a^{\′}\left(2\right),$ Wherein a ' (2),

All known, so the Δ value is known, ξ (1) value is known, and according to ξ (1)=a (2)-a (1), then a (2) can ask, again basis $\widehat{\mathrm{\ϵ}}\left(2\right)=a\left(2\right)+a^{\′}\left(2\right),$ Then

Value is obtained.Drawing the quasi-value of measuring for the second time is respectively: $\mathrm{Va}\left(2\right)=\hat{V}a\left(2\right)+\widehat{\mathrm{\ϵ}}\left(2\right);$ $\mathrm{Va}\left(2\right)=\hat{V}b\left(2\right)+\widehat{\mathrm{\ϵ}}\left(2\right);$ $\mathrm{Vc}\left(2\right)=\hat{V}c\left(2\right)+\widehat{\mathrm{\ϵ}}\left(2\right);V{a}^{\′}\left(2\right)=\hat{V}{a}^{\′}\left(2\right)+\widehat{\mathrm{\ϵ}}\left(2\right);$ $V{a}^{\″}\left(2\right)=\hat{V}{a}^{\″}\left(2\right)+\widehat{\mathrm{\ϵ}}\left(2\right)\…\…$

Measure for the third time then, set up formula, ask particular solution, ask according to measuring method for the second time $\widehat{\mathrm{\ϵ}}\left(3\right)$ Value is asked the quasi-value of measuring for the third time ..., the 4th, the 5th then ... until the unlimited quasi-value of repeatedly measuring.

Two. the measured correction given in the variation of establishing criteria amount, goes to measure tested voltage Vx (t) value according to Va (t), measured value Vx (survey), the quasi-value of Va this moment (t) records Va (i) by preceding method, according to Δ Va (i)=Va (i)-M, according to formula

Used formula and derive and theoretically be that the present invention proposes alone in measuring method that the present invention proposes and the data handling procedure will be introduced with the appendix form behind embodiment, so that support method of the present invention.

The present invention is with new method and device, measure as far as possible quasi-value near actual value, drift to selected standard volume gives real-time measurement, changed the traditional measurement method that variation is ignored to standard volume in measuring in the past, thereby improved measuring accuracy greatly, this method has the meaning that measured and standard volume is placed on same equality position, and its test result has equal measuring error transmission.Be applicable to quality, physical quantity that dc voltage measurement and all can be converted into DC voltage is as the measurement of temperature, power etc.

Below in conjunction with accompanying drawing embodiment is illustrated the present invention:

Fig. 1-Fig. 4 is the system chart of several embodiments of the present invention.Wherein Fig. 1 is ternary test block scheme; Fig. 2 is five variable test block schemes; Fig. 3 is five variable test block schemes with voltage debugging functions; Fig. 4 is the block scheme of difference mensuration.

Embodiment 1 (ternary test macro):

In the ternary measuring system: the IC1-standard capacitance box, export the accurate voltage Va of a road sign (t).The IC2-digital to analog converter produces one and changes and the ladder triangular wave voltage Vs of increasing or decreasing with measuring number of times under the control of computing machine, this system can be by ultralow frequency triangle wave voltage generator generation; The IC3-phase inverter with the ladder triangular wave voltage inversion that IC2 produces, forms the complete reverse ladder triangular wave Vs with ladder triangular wave voltage Vs; One of IC4-voltage transformation compositor, receive Va (t) voltage and Vs voltage and form one tunnel new voltage Vb (t), in circuit, this circuit is at first given Vs voltage the decay of q.s, Va (t) voltage is given the decay (deciding) of trace on circuit theory, then the stack of two-way voltage is formed Vb (t) and satisfy Va (t)＞Vb (t), and Vb (t) the faint variation along with the variation of Vs ladder triangular wave voltage again; Two of IC5-voltage compositor, the circuit formation can be identical with IC4, but its effect is that Va (t) voltage Vs voltage is carried out conversion is synthetic and form Vc (t), satisfies Va (t)＞Vc (t) condition, and Va (t), Vb (t), Vc (t) ternary have satisfied the condition of anisotropy variation; The IC6-analog to digital converter, analog quantity is transformed into it corresponding digital quantity circuit can Va (t), Vb (t), arbitrary amount is done reference standard among the Vc (t), and often selecting Va (t) is reference standard.Adopt the synchronous detection mode during detection, constantly simultaneously Va (t), Vb (t) are detected at i.Also can adopt the asynchronous detection mode, as adopt the asynchronous detection mode in analog to digital conversion circuit, to be added with timesharing conversion input system, be about to Va (t) voltage timesharing and all the other n-1 voltage and compare, and analog to digital converter IC6 can detect unknown voltage Vx (t) in twice timesharing assay intervals phase; The IC7-computer circuits act as: a. receives the sampled data of analog to digital converter IC6 and controls sampling process; B. export the digital to analog converter data, the control d convertor circuit produces triangle wave voltage; C. data processing; D. receiving the keyboard commands decision measures and display mode; E. transmit data presentation; The IC8-display circuit shows measured value and standard voltage value; The IC9-keyboard; The IC10-power circuit.

The sampling of this ternary measuring system and data handling procedure: one. sampling process: sampling process should satisfy

Condition, so measure for the i+1 time with respect to the i time measurement under anisotropic condition | Δ |≤1 (last digit value unit is 1 o'clock).

Two. data processing: under the condition that satisfies sampling, the full scale value of establishing the ternary measuring system is M.Measured value removes to measure Vb (t) and Vc (t) by Va (t) for the first time, and measured value is respectively Vb1, Vc1.Va this moment (t) is Va1 by the value that another system records, and gets Va1 and Vb1, Vc1 has with last digit value unit, sets up equation, and asks particular solution.

Va(1)－Vb(1)＝Y1＝M－Vb1

Va(1)－Vb(1)＝Y2＝M－Vc1

Make Va (t)+Vb (1)+Vc (t)=0

Solving this system of equations particular solution is:

Ask Value: $\widehat{\mathrm{\ϵ}}\left(1\right)=\mathrm{Va}1-\hat{V}a\left(1\right)$ Ask Vb (1), Vc (1) $\mathrm{Vb}\left(1\right)=\hat{V}$ $\left(1\right)+\widehat{\mathrm{\ϵ}}\left(1\right);\mathrm{Vc}\left(1\right)=\hat{V}c\left(1\right)+\widehat{\mathrm{\ϵ}}\left(1\right);$ Establish again for the second time that measured value is respectively Vb2, Vc2 asks particular solution value for the second time, and Va (2)-Vb (2)=Y1 (2)=M-Vb2Va (2)-Vc (2)=Y2 (2)=M-Vc2 makes Va (2)+Vb (2)+Vc (2)=0 try to achieve its particular solution being: $\hat{V}a\left(2\right),\hat{V}b\left(2\right),\hat{V}c\left(2\right)$ Ask again

Value, according to With

Its character less than the value complementation of last digit value unit of jack per line can be asked less than the value of last digit value unit exists its absolute value of contrary sign to equate in the value $\widehat{\mathrm{\ϵ}}\left(2\right)=a\left(2\right)+a^{\′}\left(2\right)$ In a ' (2), Known its of value can be decomposed into $\widehat{\mathrm{\ϵ}}\left(1\right)=$ A (1)+a ' (1) then has: Δ=R (2)-R (1)=3[a (2)-a (1)]+3[a ' (2)-a ' (1)]=3 ξ (1)+ψ (1) because

Known, so a ' (1) is known, a ' (2) is known, so ψ (1) is known, and order | Δ |=minimal condition is set up, and promptly has | 3 ξ (1)+ψ (1) |=minimum the establishment, therefrom ask ξ (1) value, according to ξ (1)=a (2)-a (1), therefrom solve a (2) value, then $\widehat{\mathrm{\ϵ}}\left(2\right)=a\left(2\right)+a^{\′}\left(2\right),$ Then

Can ask, $\mathrm{Va}\left(2\right)=\hat{V}a\left(2\right)+\widehat{\mathrm{\ϵ}}\left(2\right);$ $\mathrm{Vb}\left(2\right)=\hat{V}b\left(2\right)+\widehat{\mathrm{\ϵ}}\left(2\right);$ $\mathrm{Vc}\left(2\right)=\widehat{\mathrm{Vc}}\left(2\right)+\widehat{\mathrm{\ϵ}}\left(2\right)$ 。

Set up according to above-mentioned the 3rd measurement, the 3rd time measured value can be asked, and n sub-value like this can be asked.

To the n variable test macro can the rest may be inferred.

Above-mentioned measuring process requires Vb (t) and Vc (t) value to be less than Va (t) and the very condition establishment of convergence Va (t) value (testing not produce that to overflow be condition).

When removing to measure free voltage Vx (t) by Va (t), can be calculated as follows the value of Vx (t), if doing standard by Va (t) carves at a time and measures Vx (t) voltage to get measured value be Vx (survey), the magnitude of voltage of this moment Va (t) is that this value of Va (i) is tried to achieve by the front theory again, Δ Va (i)=Va (i)-M then, according to (45) formula, have this moment:

Embodiment 2 (five variable measuring systems):

As shown in Figure 2.

Select the difference of ternary and five variablees to be: the i+1 time measurement measured with respect to the i time, the variation difference of ternary or five variate-value sum R, ternary allows the R value to change a last digit value unit, and five variablees allow the R value to change two last digit value units.The ternary system need survey two measured values, and five variablees need be surveyed the four measuring value, and the n variable system then is a n-1 measured value.

System forms:

IC1-1-standard capacitance box outputting standard voltage Va (t); IC1 '-standard capacitance box, outputting standard voltage Va ' (t), IC1 " standard capacitance box, outputting standard voltage Va " (t), counts the output of the accurate voltage of three road signs; The IC2-digital to analog converter, function is with the ternary measuring system; The IC3-phase inverter; One of IC4-voltage transformation compositor; Two of IC5-voltage transformation compositor; The IC6-analog to digital converter; The IC7-computer circuits; The IC8-display circuit; The IC9-keyboard; The IC10-power circuit is with the ternary measuring system.

Embodiment 3 (five variable measuring systems) with voltage debugging functions:

As shown in Figure 3.

Digital to analog converter IC2 '-and computer circuits IC7 and standard capacitance box IC1-1 join, level off to null value in order to the change in voltage of control IC1-1.Other IC1 ', IC1 ", IC2, IC3, IC4, IC5, IC6, IC7, IC8, IC9, IC10 and five variable measuring systems figure shown in Figure 2.

Embodiment 4:

As shown in Figure 4.

Should comprise the variables that satisfies the anisotropy change condition among the n variable standard difference combiner IC11, difference signal is sent to analog to digital converter through the difference voltage that is converted into corresponding physical quantity behind the sensing circuit IC12 or current signal and carries out the numerical value conversion process in twos.

This measuring method is the transfer process that need not when numerical value is handled through full scale value doing with the difference of Fig. 1, Fig. 2 and Fig. 3, and reason is that measured value itself promptly is the difference signal of input canonical variable, therefore can directly utilize measured value substitution formula to ask particular solution.This method is applicable to the detection of normal voltage standard quality.

Appendix

The source of the formula that the present invention is used and derivation

Traditional measurement method is: an at first selected known standard amount, go to measure measured then with this known standard amount, and weigh tested value according to the measured ratio that accounts for the known standard amount.This measuring method assert that at first the variation of known standard amount ignores.This measuring method can be described as the absolute method of measurement, assert that promptly the known standard amount is absolute constant, and measurement result is the relative changing value with respect to the known standard amount.

The quasi-value mensuration is based on up-to-date mathematical method and the method for sampling and the brand-new measuring method that proposes, the essential distinction of itself and traditional measurement method is: it thinks that all amounts that participate in observation all are variable (comprising the known standard amount), in the time domain of measuring, not only can provide measured variation, also can provide the variation of standard volume, and measured and measuring error standard volume has same order, therefore can claim that this kind measuring method is a relative measurement method, and the value of the accurate absolute change that test result is represented with the Finite Number word bit.This measuring method has solved the detection of known standard value drift on theoretical and method, it has carried out organic inner link with stochastic error and systematic error.For explanation the method is at first illustrated mathematical principle:

One. the equation expression of measuring method

1. the mathematical expression that detects of normal voltage

Be provided with two normal voltages, compared: make V _1i(greatly)=V+ Δ V _Li(greatly) V _2i(greatly)=V+ Δ V _2i(greatly) be V then _1i(greatly)-V _2i(greatly)=Δ V _1i(greatly)-Δ V _2i(greatly) (1) can have according to formula (1): V if there be n normal voltage to compare _1i(greatly)-V _2i(greatly)=Δ _Vi(greatly)-Δ V _2i(greatly) be V 1. _Li(greatly)-V _3i(greatly)=Δ V _1i(greatly)-Δ V _3i(greatly) 2. ... (2) V _1i(greatly)-V _Ni(greatly)=Δ _1i(greatly)-Δ V _Ni(greatly)

V _2i(greatly)-V _Ni(greatly)=Δ V _2i(greatly)-Δ V _Ni(greatly)

If record its difference then have again: Δ V _1i(greatly)-Δ V _2i(greatly)=y _1i1. Δ V _1i(greatly)-Δ V _3i(greatly)=y _2i2.

(3) Δ V _1i(greatly)-Δ V _Ni(greatly)=y _{(n-1) i}

Δ V _2i(greatly)-Δ V _Ni(greatly)=y _Ni System of linear equations (3) is a difference of representing n variable normal voltage with the difference of n variable normal voltage error term.N dimensional linear system of equations (3) linear dependence is separated so have infinite many groups, and reading system of equations is that dimensionality reduction is linear independent.Any one amount is known in system of linear equations (3), and then all the other n-1 amount all can be asked.

2. the mathematical expression of mass measurement

Be provided with n standard test weight, on equal-arm balance, compare, in like manner can have:

Δ m _1i(greatly)-Δ m _2i(greatly)=y _1i1.

Δ m _2i(greatly)-Δ m _3i(greatly)=y _2i2.

……

Δ m _{(n-1) i}(greatly)-Δ m _Ni(greatly)=y _{(n-1) i}

(4)

Δ m _Ni(greatly)-Δ m _1iGreatly)=y _Ni

3. the mathematical expression of frequency measurement

Be provided with n independent standard frequency source separately, its frequency nominal value is respectively: f1, and f2 ... fn.If carry out forming behind the frequency division standard time amount by frequency f i, and time quantum removes survey frequency fj (wherein can comprise processing such as frequency multiplication, frequency division, difference frequency) thus, then have following relational expression set up (this slightly in digital measurement ± 1 word error)

m＝fjN/fi (5)

N represents divider ratio in the formula (5), and m represents the measured value of counter.

As everyone knows, the frequency of each frequency source is all changing constantly, so measured value m also will change thereupon.The funtcional relationship that they change can be provided by following formula according to partial differential equation $\mathrm{dm}=\frac{N}{f_i}\∂f_j-\frac{f_{j}N}{{f_i}^2}{\∂f}_i$ (6) if Δ fi and Δ fj one change very little amount then measured value after changing can be expressed as: $m+\mathrm{\Δm}=\frac{f_{j}N}{f_i}+\frac{N}{f_i}({\mathrm{\Δf}}_i-\frac{f_j}{f_i}{\mathrm{\Δf}}_i)---\left(7\right)$ If fi=fj is arranged; There is the divider ratio of N to equate then have again again: m+ Δ m=N+ (Δ fj with frequency nominal value fi _-Δ f _i) (8) can set up following system of equations according to formula (8) again when n standard frequency source surveyed mutually: Δ f _2i-Δ f _1i=Δ m _1i1. Δ f _3i-Δ f _1i=Δ m _2i2. ... (9) Δ f _Ni-Δ f _1i=Δ m _{(n-1) i}

Δ f _Ni-Δ f _2i=Δ m _Ni

System of linear equations (9) has identical expression-form with system of linear equations (3) and (4), and they are linear dependence all.Can also find that their important common ground are, the coefficient Algebra of each equation and be zero in each system of equations.

With regard to linear algebra, the expression-form of system of equations (3), (4), (9) is also not exclusive, for example in mass measurement, fully can be by two groups of counterweights, and the number of every group of counterweight is identical, on equal-arm balance, measure, thereby form new system of linear equations, thereby have reason to set up more that the general linear system of equations is: a ₁₁x ₁+ a ₁₂x ₂+ ... + a _1nx _n=y ₁1. a ₂₁x ₁+ a ₂₂x ₂+ ... + a _2nx _n=y ₂2.

…… (10)a _n1x ₁＋a _n2x ₂＋……＋a _nnx _n＝y _n

In system of linear equations (10), form n linear equation by n unknown quantity, there is on all four character its collinearity system of equations (3), (4), (9).So have: $\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{a}_{\mathrm{ji}}=0---\left(11\right)$

For reaching the character of truly separating, the rule of separating of studying system of linear equations (10) to do following reasoning.

According to the character of stochastic process, the change procedure of any one group of stochastic variable is being crossed the Cheng Qian, and the variation of its amount is unpredictable, but each amount all has definite value corresponding with it after process, so but the row following expression:

x ₁＋x ₂＋……＋x _n＝R (12)

R value be the n variable algebraic sum after each stochastic process is observed in the formula (12), as make remove arbitrary equation in itself and the system of linear equations (10) after, with the new system of linear equations of (12) composition, then it is independently system of equations of n dimensional linear, that is: a ₁₁x ₁+ a ₁₂x ₂+ ... + a _1nx _n=y ₁1. a ₂₁x ₁+ a ₂₂x ₂+ ... + a _2nx _n=y ₂2. a _{(n-1) 1}x ₁+ a _{(n-1) 2}x ₂+ ... + a _{(n-1) n}x _n=y _(n-1) x ₁+ x ₂+ ... + x _n=R

(13)

Because the R value is a unknown quantity that constantly changes with measurement, so the following equation order of examination: x ₁+ x ₂+ ... + x _n=0 (14)

And by (14) formula with remove in (10) form behind arbitrary equation new linearity independently system of equations be

a ₁₁x ₁＋a ₁₂x ₂＋……＋a _1nx _n＝y ₁ ①

a ₂₁x ₁+ a ₂₂x ₂+ ... + a _2nx _n=y ₂2. ... (15) a _{(n-1) 1}x ₁+ a ( _{N-1) 2}x ₂+ ... + a _{(n-1) n}x _n=y _(n-1)

x ₁+ x ₂+ ... + x _n=0

System of linear equations (15) be linearity independently, it has unique solution, make its separate into: ${(x_1,x_2,\…\…x_n)}^T={({\hat{x}}_1,{\hat{x}}_{2,}\…\…{\hat{x}}_n)}^T---\left(16\right)$ Claim this group to separate particular solution, consider into system of linear equations (10) $\underset{i=1}{\overset{n}{\mathrm{\Σ}}}\mathrm{aji}=0$ Character, the general solution that then can provide system of linear equations (10) is:

Formula (17) expression whether be that the general solution of system of linear equations (10) can be done following proof, with separating in the substitution system of linear equations (10) of (17), putting in order then has $a_{11}{\hat{x}}_1+a_{12}{\hat{x}}_2+\…\…+a_{1n}{\hat{x}}_n+\mathrm{\ϵ}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{a}_{1i}=y_1$ $a_{21}{\hat{x}}_1+a_{22}{\hat{x}}_2+\…\…+a_{2n}{\hat{x}}_n+\mathrm{\ϵ}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{a}_{2i}=y_2$ …… (18) $a_{n1}{\hat{x}}_1+a_{n2}{\hat{x}}_2+\…\…+a_{\mathrm{nn}}{\hat{x}}_n+\mathrm{\ϵ}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{a}_{\mathrm{ni}}=y_n$

Clearly the condition of separating of formula (17) satisfies the condition of system of linear equations (10), establish again system of linear equations (10) except that (17) formula represent separate also have all the other to separate existence, then must have the unequal condition of ε to exist, establish ε ₁, ε ₂ε _nOnly rare one with remaining is unequal, then the system of equations of (18) will become: $a_{11}{\hat{x}}_1+a_{12}{\hat{x}}_2+\…\…{+a}_{1n}{\hat{x}}_n+\underset{i=1}{\overset{n}{\mathrm{\Σ}}}\left(a_{1i}{\mathrm{\ϵ}}_i\right){\≠y}_1$ $a_{21}{\hat{x}}_1+a_{22}{\hat{x}}_2+\…\…+a_{2n}{\hat{x}}_n+\underset{i=1}{\overset{n}{\mathrm{\Σ}}}\left(a_{2j}{\mathrm{\ϵ}}_i\right)\≠y_2$ $a_{n1}{\hat{x}}_1+a_{n2}{\hat{x}}_2+\…\…+a_{\mathrm{nn}}{\hat{x}}_n+\underset{i=1}{\overset{n}{\mathrm{\Σ}}}\left(a_{\mathrm{ni}}{\mathrm{\ϵ}}_i\right){\≠y}_n$

The condition contradiction of separating of this result and system of linear equations (10), therefore must demonstrate,prove does not have remaining again the condition of separating that system of linear equations (10) removes the expression of (17) formula and separates existence, claim that separating of (17) formula expression is the general solution of system of linear equations (10), according to the character of general solution as can be known: truly separate and also must be included in the general solution expression formula, truly separate as known, then be known R value, then have this moment: $\widehat{\mathrm{\ϵ}}=\frac{R}{n}\·\·\·\·\·\left(19\right)$

Clearly

Be the ε value of representing under the condition of truly separating, it is the amount of determining, the relation of it and R value is its n five equilibrium.

If undertaken measuring for m time then its expression formula of truly separating can be by the measuring system of any n-1 equation in the system of equations (10): M can be any positive integer value in formula (20), comprises m=n; In fact in (20) formula Be unknown,

Be the single-valued function of time t, it carries out translation on rectangular coordinate system $\frac{1}{n}\underset{i=1}{\overset{m}{{\mathrm{\Σ}\widehat{\mathrm{\ϵ}}}_i}}$ After the value, following characteristics are arranged and form ε i: ${\stackrel{-}{\mathrm{\ϵ}}}_i={\widehat{\mathrm{\ϵ}}}_i-\frac{1}{n}\underset{i=1}{\overset{m}{\mathrm{\Σ}}}{\widehat{\mathrm{\ϵ}}}_i---\left(21\right)$ Order $\stackrel{=}{x}=\frac{1}{n}\underset{i=1}{\overset{m}{\mathrm{\Σ}}}{\widehat{\mathrm{\ϵ}}}_i---\left(22\right)$ $\underset{i=1}{\overset{m}{\mathrm{\Σ}}}{\stackrel{-}{\mathrm{\ϵ}}}_i=0---\left(23\right)$

Can set up fully and the getting in touch of traditional variance theory according to (21), (22) and (23) formula.

Two. the effect of last digit value unit in random number is handled

As everybody knows, all amounts that are observed and variation thereof all are there to be spacing digital quantity to represent in the world.The numerical value unit of its position, end of any limited bit digital quantity all is 1 word.Yet any one measures its true value (theoretical true value is set true value and reached except the value of representing with natural number) can be a value of representing with the infinite place digital quantity, otherwise the error axiom does not promptly exist.

Yet in real society, do not have anyone and go to represent a certain objective value with the digital quantity of infinite place because this is impossible, Here it is in the world all amounts that are observed and tool to change all be the reason of representing with limited bit digital.For same observation, numerical value figure place that can observation is many more, shows that then the precision of this observation is high more, thereby to the observation of any amount, its accuracy of observation all depends on the last digit value unit of this observation, and this has just shown the vital role of last digit value unit in random number is handled.

Same recording geometry for the n variable, its accuracy of observation is equal to, thereby this n variable should have identical last digit value unit, and will be less than the value of last digit value unit, being submerged in fully in the last digit value unit. yet each amount of the n variable of occurring in nature participation observation is all at time to time change, its the i+1 time observation in an enough little time city is very little with respect to the variation of the i time observed quantity to the variable summation in n the stochastic variable and each variable, and the incorgruous variation characteristic (the guiding variation characteristic of so-called amount be the n variable have an amount less with respect to all the other n-1 the constant characteristics of measuring of change direction phase XOR with respect to observing for the i time the i+1 time observation) of establishing the n variable amount of having again then can solve the accurate absolute change of n variable in the observation time city.

If the recording geometry of certain n variable is in observation process, each amount all has surface properties under the Xie Biyou that identical last digit value unit then provides by formula (16):

Less than the value of last digit value unit, symbol is identical, and numerical value equates, the different numerical value complementation of symbol, otherwise this time observed reading of n variable has different last digit value units.

Proving of this character is easy, and this is that any n variable translation in rectangular coordinate system with identical last digit value unit all has this character.

Observe formula (20) again, once observe any, the last digit value unit that separates is identical (because be by limited bit representation numerical value), so

With

Its value less than last digit value unit exists its absolute value of contrary sign to equate again, and its value complementation less than last digit value unit of jack per line.This character is not as existing, and then (20) formula does not promptly satisfy the condition of identical last digit value unit, promptly do not satisfy the condition of equal observation.

Known again n variable recording geometry is equal observation, so they have identical last digit value unit, the algebraic sum R value of n variable also is the amount that has identical last digit value unit with the n variable so, and this conclusion is self-evident, so variable number R value and

The relation of value can be represented that the last digit value unit that establishes certain n variable is the worker, is then provided the relation value of 3,4,5,6 variablees by table one by table one.After having understood last digit value unit and be 1 variable parameter table, then the position, end all can provide immediately for the table of any value, and the variation of this value is radix point mobile only.

In the recording geometry of n variable, n can be any more than or equal to 3 positive integer.Generally speaking advise that n gets odd number value.

One group of measured value is satisfying under the condition with identical last digit value unit, and the particular solution value of this group measured value satisfies value less than last digit value unit to be had jack per line and equate that the condition of contrary sign complementation, this conclusion are the popularizations of the middle ε of formula (17)=0.

Table one last digit value unit is 1 variable parameter table

Three. anisotropy n variable stochastic process data sampling method and optimum solution value are estimated

Can have according to equation (12)

R(t)＝X ₁(t)＋X ₂(t)＋……＋X _n(t) (24)

If X ₁(t), X ₂(t) ... X _n(t) be the continuous function of n variable, but because the restriction of scope resolution makes that R (t) can only be a discrete spectrum function, and its discreteness depends on last digit value unit, if known:

R(t)＝R(i) t＝i (25)

When the time changes to i+1 by i (this moment, i can regard the time shaft variable as), then have:

R(t)＝R(i＋1) t＝i＋1 (26)

Make Δ=R (i+1)-R (i) (27)

If it is enough little that time t changes to the i+1 time period constantly constantly by i, variation of function is also corresponding enough little constantly then to have R (t) to change to i+1 constantly by i, so the Δ value also should be one and change in enough little scope, and be the value of increment unit with last digit value unit, basis again: $\widehat{\mathrm{\ϵ}}\left(t\right)=\frac{R\left(t\right)}{n}---\left(28\right)$ $\widehat{\mathrm{\ϵ}}\left(t\right)=a+a^{\′}---\left(29\right)$ A represents in the formula (29) The part of last digit value unit integral multiple, and a ' be expression less than last digit value unit value part, so have again: Δ=R (i+1)-R (i)=n (a (i+1)-a (i))+n (a ' (i+1)-a ' is (i)) makes: (a (i+1)-a (i))=ξ (i); N (a ' (i+1)-a ' (i))=ψ (i)

Δ＝nξ(i)＋ψ(i) (30)

ψ (i) also is the integral multiple value of last digit value unit in formula (30), its span dependent variable number and different, example is an example with the ternary, ψ (i) value can only be (0,1,2,-1,-2), corresponding this class value then can provide three progression ordered series of numbers (value of ψ (i) is to be 1 o'clock conclusion with last digit value unit) by formula (30).

-9 ,-6 ,-3,0,3,6,9 ... ordered series of numbers one

-8 ,-5 ,-2,1,4,7,10 ... ordered series of numbers two

-10 ,-7 ,-4 ,-1,2,5,8 ... ordered series of numbers three

Satisfy the Δ absolute value and be minimum condition, then the Δ value can only correspondence be got (0,1 ,-1) with this understanding, and the corresponding value of ξ (i) is (0,1 ,-1), decides on ψ (i).

When R (i) value that known i measures constantly, so according to (30) formula, the R (i+1) that i+1 measures can ask, so

Can ask, then i+1 moment X ₁(t), X ₂(t), X ₃(greatly) each variate-value can be asked.

If variable is 4, then the value of ψ (i) can only be (0,1,2,3 ,-1 ,-2 ,-3), can imitate this for 5 variablees and even n variable and analogize.When last digit value unit was 0.1, as long as radix point is moved to the left one, the rest may be inferred by analogy.

Further observe ordered series of numbers one, ordered series of numbers two and ordered series of numbers three again, can find that its value rule is to provide the value of separating every 3 numbers, then 3 to be called last digit value unit be 1 the spacing of separating, clearly 3 is variable numbers, in like manner the spacing of 4 variablees is 4, and promptly the spacing of n variable is n, and is big more at interval, the anti-jamming means for test system ability is strong more, but system is huge more.According to as above reasoning following conclusion can be arranged: to having the test macro of precision such as anisotropic n variable, if measure with respect to the i time and measure for the i+1 time, the variable quantity of twice n variable algebraic sum is partly long-pending with last digit value unit less than spacing, and then this recording geometry can be the variation of magnitude variations unit apart from each amount of track with its last digit value unit in real time.

The correctness of formula (31) can enumerate 4,6,8 respectively ... proving Deng the pairing ordered series of numbers situation of even number variable number, is example with 4 variablees, as

, ordered series of numbers is:

……-6，-2，2，6，10……

Clearly the value of Δ about " 0 " equates and opposite in sign, with this understanding, differentiate direction and be under the impartial condition, so the possibility of the misjudgement of producing is arranged, and when variable number is odd number, promptly do not exist

Formula (31) has promptly determined sampling condition.From formula (31) as seen, the time interval of sampling condition and double sampling is irrelevant, and it only depends on variable number and last digit value unit.This formula has also shown the reason that n can not get two variablees.

When the Δ value of any twice measured value can be asked (under the sampling thheorem condition that satisfies (31) formula), then according to (30) formula, promptly ψ (i) value is known, and ξ (i) value is known, and then R (i+1) value can be asked, thereby (i+1) value can be asked, and according to (20) formula, can ask at the measured value of i+1 n variable constantly.According to as above reasoning as can be known: after having carried out continuous measure for m+1 time, the variable quantity of n variable summation is:

The physical significance of the variable quantity of n variable summation average is the side-play amount of central value in rectangular coordinate system of n variable.Then the value of each amount is: $x(m+1)=\hat{x}(m+1)+\frac{1}{n}\underset{i=1}{\overset{m}{\mathrm{\Σ}}}\mathrm{\Δi}+\frac{R\left(1\right)}{n}---\left(34\right)$ (34) (m+1) expression is measured measured value constantly the m+1 time in the formula. $\widehat{\mathrm{\ϵ}}(m+1)=\frac{1}{n}\underset{i=1}{\overset{m}{\mathrm{\Σ}}}\mathrm{\Δi}+\frac{R\left(1\right)}{n}---\left(25\right)$ $\widehat{\mathrm{\ϵ}}\left(1\right)=\frac{R\left(1\right)}{n}---\left(36\right)$ (36) (1) expression is measured for the 1st time in the formula, (1) expression is measured for the 1st time

Value.

In actual measurement,

(1) value can be tested from higher leveled standard and be got or get with test in the grade standard, then (1) there is measuring error for the first time, under this theory, this error is a certain amount of, it is to measure constantly, the function of measuring system can be called systematic error, if measuring system satisfies the condition of (31) formula fully in whole measurement, this systematic error is a constant in this measuring system, and the error term that is submerged in the last figure unit is the stochastic error of system.

The derivation of above-mentioned theory has provided the method for sampling, method of testing and the data processing method that waits n standard volume of accuracy test.

Four. analog to digital conversion system value detects

If the normal voltage of certain analog to digital conversion system is Va (t), nominal value is Va, and the full scale value corresponding with this voltage is M, then quantizes magnitude of voltage in any measurement unit constantly to be:

Establish thus normal voltage again and remove to measure tested voltage Vb (t), nominal value is Vb, then measures measured value constantly at certain to be:

S (i) is that i measures measured value constantly, when following condition: conclusion below Va (t) ≈ Vb (t) (39) then can provide according to partial differential equation:

Condition according to (39) formula Va (t) ≈ Vb (t) makes Va=Vb, and then (40) formula can be rewritten as:

What clearly equal sign left side S+ Δ S (t) showed in the physical sense in (41) formula is the measured value of A-D converter, and What represent is the inverse that unit quantizes magnitude of voltage, takes out the physical concept that quantizes magnitude of voltage, and then unit quantification magnitude of voltage only shows a measured value in test process, and the inverse of 1 measured value still is 1, thereby (41) formula can be rewritten as again:

Δ Va (t)-Δ Vb (t)=M-measured value (42)

Since Va (t)=Va+ Δ Va (t), Vb (t)=Vb+ Δ Vb (t)

Known Va=Vb, so have again:

Va (t)-Vb (t)=M-measured value (43)

Under satisfied (39) formula condition, the A-D converter test macro satisfies this theoretical condition according to (42) formula and (43) formula.

When Vb (t) is the free voltage value, at this moment mainly consider because the existence of Δ Va (t) to the influence of test value, then can be derived by (38) formula

(44) condition of formula establishment is that Δ Va (t) is a very little amount, and this condition all is satisfied in general test.[S+ Δ S (t)] represented measured value in formula (44), and

The expression be

Actual value under Δ Va (t)=0 condition, so have:

ΔVa(t)＝Va(t)－M (46)

Va (t) expression is that standard value is being measured actual value constantly in (46) formula, and Va (t) promptly is the accurate actual value of the standard value that records in this article.

Claims (3)

1. device of measuring quasi-value, comprise display circuit IC8, supervisory keyboard IC9, power circuit IC10, the device that it is characterized in that above-mentioned measurement quasi-value mainly is by the IC1-standard capacitance box, or 1C ' 1, IC " 1 can produce one or more normal voltage Va (t) or Va (t), V ' a (t), V " a (t) The IC2-digital to analog converter, the ladder triangular wave voltage VS that generation one is successively decreased or increased progressively with the variation of measuring number of times under the control of computing machine, IC2 also can be a ultralow frequency triangle wave voltage generator; The IC3-phase inverter, ladder triangular wave is anti-phase, produce one and the antipodal ladder triangular wave voltage of above-mentioned IC2 direction Vs; IC4-voltage transformation compositor one carries out the synthetic new voltage Vb (t) of conversion with Va (t) voltage and Vs voltage; IC5-voltage transformation compositor two, Va (t) voltage and Vs voltage are carried out the synthetic new voltage Vc (t) of conversion, Va (t), Vb (t), Vc (t), Va ' (t), Va " (t) ... satisfy anisotropic change condition, also can adopt the method for Positive and Negative Coefficient Temperature to satisfy the anisotropic change condition of voltage; The IC6-analog to digital converter, analog quantity is transformed into corresponding with it digital circuit part, this circuit can be used Va (t), Vb (t), Vc (t), Va ' (t) ... in arbitrary amount do standard, when detecting, simultaneously Vb (t), Vc (t), Va (t) are detected during at i, above-mentioned analog to digital converter IC6 can add the timesharing converting system, Va (t) voltage and all the other n-1 voltage are compared, carry out timesharing and detect; Above-mentioned standard capacitance box IC1, IC1 ', IC1 " output terminal all join with analog to digital converter IC6; above-mentioned digital to analog converter IC2 signal input part and aforementioned calculation machine IC7 join; the IC4 of one of output terminal and voltage transformation compositor joins; an end of the common point of voltage transformation compositor IC4 and above-mentioned IC2 and above-mentioned inverter ic 3 joins; the end of two IC5 of the other end of inverter ic 3 and above-mentioned voltage transformation compositor joins; the other end and the analog to digital converter IC6 of two IC5 of voltage transformation compositor join, two IC5 of above-mentioned voltage transformation compositor also has the common point of the IC4 of an end and standard capacitance box IC1 and voltage transformation compositor to join, direct and the above-mentioned analog to digital conversion circuit IC6 of tested unknown voltage Vx joins, and above-mentioned analog to digital conversion circuit IC6 also has two ends and computer circuits IC7 to join.

2. quasi-value measuring method, it is characterized in that said method must give real-time measurement to the drift of selected standard volume Va (t), according to the variation of standard volume tested parameter is revised gained as far as possible near the quasi-value of actual value then, said method institute must adopt the quasi-value measurement mechanism, and its data processing and sampling process are as follows:

One. earlier selected standard volume is measured (for n variable measuring system) in real time

A. measure for the i+1 time with respect to the i time measurement under the condition that anisotropy changes, sampling process should satisfy Condition.

B. data handling procedure: under the condition that satisfies sampling, the full scale value of n variable measuring system is M.Measure for the first time and remove to measure Vb (1), Vc (1), Va ' (1), Va with Va (1) " (1) ...; measured value is Vb1, Vc1 ...; the value that Va (1) is recorded by another system is Va1, Va1, Vb1, Vc1, Va ' 1, Va " 1 ... identical last digit value unit is arranged.

Set up and measure equation and ask particular solution for the first time:

Va(1)－Vb(1)＝Y1(1)＝M－Vb1

Va(1)－Vc(1)＝Y2(1)＝M－Vc1

Va(1)－Va′(1)＝Y3(1)＝M－Va′1

Va(1)－Va″(1)＝Y4(1)＝M－Va″1

： ：

： ：

Va(1)＋Vb(1)＋Vc(1)＋Va′(1)＋Va″(1)＋…＝R(1)，

Make R (1)=0, ask particular solution to try to achieve its particular solution to be:

Ask Value: $\widehat{\mathrm{\ϵ}}\left(1\right)=\mathrm{Va}1-\hat{V}a\left(1\right)$ The quasi-value of measuring is respectively for the first time: $\mathrm{Vb}\left(1\right)=\hat{V}b\left(1\right)+\widehat{\mathrm{\ϵ}}\left(1\right);\mathrm{Vc}\left(1\right)=\hat{V}c\left(1\right)+\widehat{\mathrm{\ϵ}}\left(1\right),$ $V{a}^{\′}\left(1\right)=\hat{V}{a}^{\′}\left(1\right)+\widehat{\mathrm{\ϵ}}\left(1\right);V{a}^{\″}\left(1\right)=\hat{V}{a}^{\″}\left(1\right)+\widehat{\mathrm{\ϵ}}\left(1\right);$ , carry out the second time then and measure; Set up measure equation successively, Va (2)-Vb (2)=Y1 (2)=M-Vb2Va (2)-Vc (2)=Y2 (2)=M-Vc2Va (2)-Va ' (2)=Y3 (2)=M-Va ' 2Va (2)-Va " (2)=Y4 (2)=M-Va " 2

： ：

:: Va (2)+Vb (2)+Vc (2)+Va ' (2)+Va " (2)+...=R (2) makes R (2)=0 try to achieve its particular solution being:

Ask Value: $\widehat{\mathrm{\ϵ}}\left(2\right)=a\left(2\right)-a^{\′}\left(2\right)$ (the integral multiple part of a (i) expression last digit value unit, a ' (i) represents the part less than last digit value unit) Δ=R (2)-R (1)=n[a (2)-a (1) again]+[a ' (2)-a ' (1)]=n ξ (1)+ψ (1) ξ (1)=[a (2)-a (1)] wherein, ψ (1)=n (a ' (2)-a ' (1)) is satisfying under the sampling condition: Wherein ξ (1) should satisfy | Δ | and be the condition of minimum, utilize formula: $\widehat{\mathrm{\ϵ}}\left(2\right)=a\left(2\right)+a^{\′}\left(2\right),$ Wherein a ' (2), All known, so the Δ value is known, ξ (1) value is known, and according to (1)=a (2)-a (1), then a (2) can ask, again basis $\widehat{\mathrm{\ϵ}}\left(2\right)=a\left(2\right)+a^{\′}\left(2\right),$ Then

Value is obtained.Drawing the quasi-value of measuring for the second time is respectively: $\mathrm{Vb}\left(2\right)=\hat{V}b\left(2\right)+\widehat{\mathrm{\ϵ}}\left(2\right);\mathrm{Vc}\left(2\right)=\widehat{\mathrm{Vc}}\left(2\right)+\widehat{\mathrm{\ϵ}}\left(2\right);\mathrm{Va}\left(2\right)=\hat{V}a\left(2\right)+\widehat{\mathrm{\ϵ}}\left(2\right)$ $V{a}^{\′}\left(2\right)=\hat{V}{a}^{\′}\left(2\right)+\widehat{\mathrm{\ϵ}}\left(2\right);V{a}^{\″}\left(2\right)=\hat{V}{a}^{\″}\left(2\right)+\widehat{\mathrm{\ϵ}}\left(2\right);$ Measure for the third time then, set up formula, ask particular solution, ask according to measuring method for the second time

Value is asked the quasi-value of measuring for the third time ..., the 4th, the 5th then ... until the unlimited quasi-value of repeatedly measuring.

Two. the measured correction given in the variation of establishing criteria amount.Go to measure tested voltage Vx (t) value according to Va (t), measured value Vx (survey), the quasi-value of Va this moment (t) records Va (i) by preceding method, and Δ Va (i)=Va (i)-M is according to formula

3. quasi-value measurement mechanism according to claim 1, comprise analog to digital converter IC6, computer circuits IC7, supervisory keyboard IC9, display circuit IC8 etc., it is characterized in that above-mentioned measurement mechanism also comprises n variable standard difference combiner IC11, difference signal sensing circuit IC12 in twos, the above-mentioned sensing circuit of difference signal in twos IC12 one end and n variable standard difference combiner IC11 join, the other end and analog to digital converter IC6 join, and said n variable standard difference combiner IC11 should comprise the variables that satisfies the anisotropy variation.

Images