CN104156558A - Method for judging consistency of results obtained according to different chemical component analysis methods - Google Patents

Abstract

The invention belongs to chemical element analysis technologies and relates to a method for judging the consistency of results obtained according to the different chemical component analysis methods. The method comprises the steps of checking of outliers, checking of the consistency of precision degrees of two samples and checking of the consistency of average values of the two samples. According to the method for judging the consistency of the results obtained according to the different chemical component analysis methods, the characteristics of chemical analysis are fully considered based on the principle of statistics, the principle of statistics and chemical detection data analysis are combined organically, the analysis results, of specific elements in a material, obtained according to the two different methods are processed, and thus whether data obtained according to the two analysis methods are theoretically consistent is judged.

Description

The conforming determination methods of a kind of different chemical composition analytical approach result

Technical field

The invention belongs to analysis of chemical elements technology, relate to the conforming determination methods of a kind of different chemical composition analytical approach result.

Background technology

In chemical composition analysis process, different analytical approach data consistency detection problems are very important.Carrying out in new chemical composition analysis method formulation process, new system setting analysis method and existing methodical consistance judgement; In task reception process, the consistance judgement between the analytical approach that the analytical approach of customer requirement and actual detection data are used, step is all absolutely necessary.Under existence conditions, the judgement of method consistance comprises two aspects: method precision judgement and method accuracy judgement.At present, the judgment mode of relative standard deviation (RSD) is generally taked in method precision judgement.Method accuracy judgement is general to be adopted two groups of the new analytical approach of formulating and original analytical approach gained to detect the mode that data carry out directly contrasting to carry out.Concrete way of contrast is the detection data arithmetic mean value comparison that the analytical approach of newly formulating is detected to data and original analytical approach, if the analytical approach detection data of formulating new are all within the scope of former analytical approach tolerance, and the new analytical approach of formulating of judgement is consistent with former method accuracy.

Summary of the invention

The object of the invention is: to the different analytical approachs of certain element-specific in material through overtesting gained to data compare, analyze, in theory whether there is consistance with the data that judge different analytical approach gained.

Technical scheme of the present invention is: determination methods comprises test of outlier, the consistency check of two sample precision and three steps of two sample mean consistency check, and concrete steps are as follows:

(1), test of outlier:

(1.1) different chemical composition analytical approach result the data obtained should meet following requirement:

A) two groups of detection data of same material, identity element, different analytical approach gained;

B) detect data without technological deficiency;

C) detect the choice of data end and take the revision of the convention principle of " four give up six enters five Dan Shuan ", " four houses six enter five considerations, all zero look odd even after five rear non-zeros enter, five, and five is front for occasionally casting out, and five is front for very entering " 0 is considered as even number;

D) every group of detection data unit is identical;

E) every group of detection data significant figure are identical;

F) each group detects data amount check n >=3,

(1.2) extract two groups of one group of data detecting in data, carry out data processing:

(1.2.1) calculate the arithmetic mean value of these group data, formula is as follows:

$\stackrel{\‾}{x}=\frac{x_1+x_2+x_3+...+x_n}{n}=\frac{1}{n}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{x}_i$

In formula:

X ₁, x ₂... x _n-measured value;

The symbol of Σ-expression summation;

-represent from x ₁to x _nmeasured value summation;

(1.2.2), taking arithmetic mean value as true value, calculate the absolute deviation of each data;

Absolute deviation=measured value-arithmetic mean value

(1.2.3) calculate standard deviation

Standard deviation data expression formula is:

$s=\sqrt{\frac{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{(x_i-\stackrel{\‾}{x})}^2}{n-1}}$

In formula:

S-standard deviation

the arithmetic mean of-n observed reading

X _i-any single-measurement value

N-detects data amount check

(1.3) this group detects data detection

The detection data of (1.2) are pressed from small to large and arranged;

Choose respectively the maximal value and the minimum value that detect data, be calculated as follows respectively

$G_n=\frac{|x_i-\stackrel{\‾}{x}|}{s}$

X _i-be maximum or minimum measured value

Determine that level of significance is α=0.05, in the T of Grubbs method of inspection value table, find corresponding critical value, be labeled as G _table;

Judgement: work as G _n> G _tabletime, judge measured value x _ifor exceptional value; Otherwise there is no exceptional value;

(1.4) rejecting abnormalities value

If x _ifor exceptional value, need to reject, if detect data amount check n>=3 after rejecting abnormalities value, remaining data should be continued to the step process by (1.2)-(1.3), until without exceptional value; After rejecting abnormalities value, if data amount check n < 4 calculates and stops, can not carry out data consistent sex determination;

(1.5) undertaken another by (1.2)-(1.4) step and organize the inspection that detects data;

(2), two sample precision consistency check

(2.1) will check two groups of data of rear rejecting abnormalities value as the calculating data of this step by (1), and data amount check n meet n >=3;

(2.2), by formula in entry (1.2.3), calculate respectively the standard deviation S of two groups of data ₁and S ₂;

(2.3) compare S ₁and S ₂size;

(2.4) calculate two sample precision consistance F _meter:

If S ₁> S ₂, calculate with following formula

If S ₂> S ₁, calculate with following formula

(2.5) calculate respectively the degree of freedom V of every group of data variance _point, computing formula is:

V _point=n-1

(2.6) according to the degree of freedom numerical value of variance, the in the situation that of level of significance α=0.05, in F check table or F distribution table, find F value, be labeled as F _table

(2.7) draw judged result according to institute's value:

If F _meter≤ F _tablewhen value, there is not significant difference in two kinds of different chemical composition Standard of analytical methods deviations, and precision is consistent;

If F _meter> F _tablewhen value, there is significant difference in two kinds of different chemical composition Standard of analytical methods deviations, and precision is inconsistent, calculates and stop, and can not carry out data consistent sex determination;

(3), two sample mean consistency checks;

(3.1) will be by (1), (2) two groups of data after the assay was approved the calculating data as this step;

(3.2) calculate respectively the arithmetic mean value of two groups of data by formula in entry (1.2.1) with

(3.3) calculate respectively the standard deviation S of two groups of data by formula in entry (1.2.3) ₁and S ₂;

(3.4) determine degree of freedom V _always

V _always=n ₁+ n ₂-2

In formula:

N ₁-be the number of first group of data;

N ₂-be the number of second group of data;

(3.5) S _closecalculate

S _closecomputing formula is:

(3.6) two sample mean consistance T _closevalue is calculated

T _closevalue computing formula is:

(3.7) total degree of freedom V of two groups of detection data of foundation _alwaysnumerical value, the in the situation that of level of significance α=0.05, at t inspection tables of critical values or t distribution table, is labeled as T _table

(3.8) draw judged result according to institute's value:

If T _close>=T _table, when value, two kinds of different chemical composition analytical approach average results there are differences, and average result is inconsistent statistically, calculates and stops, and judges that these two kinds of analytical approach results are inconsistent;

If T _close< T _table, when value, two kinds of different chemical composition analytical approach average results not there are differences, and average result has consistance statistically, judges that these two kinds of analytical approach results are consistent.

Advantage of the present invention is:

1) the present invention considers the own characteristic of chemical analysis specialty, and Principle of Statistics and chemical detection data analysis are organically combined, and has formed and has judged two kinds of judgment criterion whether analytical approach is consistent;

2) the present invention adopts the treatment principle of " four house six enters five Dan Shuan " on data are accepted or rejected, and " rounding up " method is more reasonable, has avoided after the revision of the convention system of measured value higher;

3) in abnormality value removing, the consistency check of two sample data precision, the consistency check of two sample data mean value, all determined applicable level of significance, so both ensure the requirement of quality control in the measurement of normalization Epidemiological Analysis, avoided again the phenomenon without alternative method too strictly causing due to method Data Control;

4) in the consistency check of two sample data mean value, take two-sided test method, avoided the misjudgment causing due to small probability event;

5) the corresponding data otherness that the present invention obtains for different chemical analytical approach is carried out intuitively, is compared flexibly and analyze, thereby reaches the object of consistance judgement.

Embodiment

Determination methods comprises test of outlier, the consistency check of two sample precision and three steps of two sample mean consistency check, and concrete steps are as follows:

(1), test of outlier:

(1.1) different chemical composition analytical approach result the data obtained should meet following requirement:

A) two groups of detection data of same material, identity element, different analytical approach gained;

B) detect data without technological deficiency;

C) detect the choice of data end and take the revision of the convention principle of " four give up six enters five Dan Shuan ", " four houses six enter five considerations, all zero look odd even after five rear non-zeros enter, five, and five is front for occasionally casting out, and five is front for very entering " 0 is considered as even number;

D) every group of detection data unit is identical;

E) every group of detection data significant figure are identical;

F) each group detects data amount check n >=3,

(1.2) extract two groups of one group of data detecting in data, carry out data processing:

(1.2.1) calculate the arithmetic mean value of these group data, formula is as follows:

$\stackrel{\&OverBar;}{x}=\frac{x_1+x_2+x_3+...+x_n}{n}=\frac{1}{n}\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{x}_i$

In formula:

X ₁, x ₂... x _n-measured value;

The symbol of Σ-expression summation;

-represent from x ₁to x _nmeasured value summation;

(1.2.2), taking arithmetic mean value as true value, calculate the absolute deviation of each data;

Absolute deviation=measured value-arithmetic mean value

(1.2.3) calculate standard deviation

Standard deviation data expression formula is:

$s=\sqrt{\frac{\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{(x_i-\stackrel{\&OverBar;}{x})}^2}{n-1}}$

In formula:

S-standard deviation

the arithmetic mean of-n observed reading

X _i-any single-measurement value

N-detects data amount check

(1.3) this group detects data detection

The detection data of (1.2) are pressed from small to large and arranged;

Choose respectively the maximal value and the minimum value that detect data, be calculated as follows respectively

$G_n=\frac{|x_i-\stackrel{\&OverBar;}{x}|}{s}$

X _i-be maximum or minimum measured value

Determine that level of significance is α=0.05, in the T of Grubbs method of inspection value table, find corresponding critical value, be labeled as G _table;

Judgement: work as G _n> G _tabletime, judge measured value x _ifor exceptional value; Otherwise there is no exceptional value;

(1.4) rejecting abnormalities value

If x _ifor exceptional value, need to reject, if detect data amount check n>=3 after rejecting abnormalities value, remaining data should be continued to the step process by (1.2)-(1.3), until without exceptional value; After rejecting abnormalities value, if data amount check n < 4 calculates and stops, can not carry out data consistent sex determination;

(1.5) undertaken another by (1.2)-(1.4) step and organize the inspection that detects data;

(2), two sample precision consistency check

(2.1) will check two groups of data of rear rejecting abnormalities value as the calculating data of this step by (1), and data amount check n meet n >=3;

(2.2), by formula in entry (1.2.3), calculate respectively the standard deviation S of two groups of data ₁and S ₂;

(2.3) compare S ₁and S ₂size;

(2.4) calculate two sample precision consistance F _meter:

If S ₁> S ₂, calculate with following formula

If S ₂> S ₁, calculate with following formula

(2.5) calculate respectively the degree of freedom V of every group of data variance _point, computing formula is:

V _point=n-1

(2.6) according to the degree of freedom numerical value of variance, the in the situation that of level of significance α=0.05, in F check table or F distribution table, find F value, be labeled as F _table

(2.7) draw judged result according to institute's value:

If F _meter≤ F _tablewhen value, there is not significant difference in two kinds of different chemical composition Standard of analytical methods deviations, and precision is consistent;

If F _meter> F _tablewhen value, there is significant difference in two kinds of different chemical composition Standard of analytical methods deviations, and precision is inconsistent, calculates and stop, and can not carry out data consistent sex determination;

(3), two sample mean consistency checks;

(3.1) will be by (1), (2) two groups of data after the assay was approved the calculating data as this step;

(3.2) calculate respectively the arithmetic mean value of two groups of data by formula in entry (1.2.1) with

(3.3) calculate respectively the standard deviation S of two groups of data by formula in entry (1.2.3) ₁and S ₂;

(3.4) determine degree of freedom V _always

V _always=n ₁+ n ₂-2

In formula:

N ₁-be the number of first group of data;

N ₂-be the number of second group of data;

(3.5) S _closecalculate

S _closecomputing formula is:

(3.6) two sample mean consistance T _closevalue is calculated

T _closevalue computing formula is:

(3.7) total degree of freedom V of two groups of detection data of foundation _alwaysnumerical value, the in the situation that of level of significance α=0.05, at t inspection tables of critical values or t distribution table, is labeled as T _table

(3.8) draw judged result according to institute's value:

If T _close>=T _table, when value, two kinds of different chemical composition analytical approach average results there are differences, and average result is inconsistent statistically, calculates and stops, and judges that these two kinds of analytical approach results are inconsistent;

If T _close< T _table, when value, two kinds of different chemical composition analytical approach average results not there are differences, and average result has consistance statistically, judges that these two kinds of analytical approach results are consistent.

Embodiment mono-

Determination methods comprises test of outlier, the consistency check of two sample precision and three steps of two sample mean consistency check, and concrete steps are as follows:

(1), test of outlier:

(1.1) test the data obtained is as follows:

A) material trademark: K4169 high temperature alloy; Measure element: Al; Measuring method: cupferron, cupferron are divided EDTA volumetric determination aluminium content (HB5220.19-2008); ICP-AES method is measured aluminium content (Q/6S2205.4-2009) in Ni-based, Fe Ni matrix high temperature alloy;

B) detect data without technological deficiency, order ticket requires to retain 2 significant digits significant figure, and this detection data retain 3 position effective digitals, and data are taked the revision of the convention principle of " four houses six enter five Dan Shuan ".Detect data in table 1:

The analysis result of Al element in table 1:K4169 material

(1.2) one group of data of extraction HB5220.19-2008 analytical approach gained, carry out data processing;

(1.2.1) calculate the arithmetic mean value of these group data, formula is as follows:

$\begin{array}{c}\stackrel{\&OverBar;}{x}=\frac{x_1+x_2+x_3+...+x_n}{n}=\frac{1}{n}\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{x}_i\\ =\frac{0.612+0.641+0.617+0.623+0.634+.0631+0.627+0.625}{8}\\ =0.626\end{array}$

(1.2.2), taking arithmetic mean value as true value, calculate the absolute deviation of each data;

Absolute deviation=measured value-arithmetic mean value

Absolute deviation is followed successively by :-0.014,0.015 ,-0.009 ,-0.003,0.008,0.005,0.001 ,-0.001

(1.2.3) calculate standard deviation

The date expression of standard deviation is:

$\begin{array}{c}s=\sqrt{\frac{\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{(x_i-\stackrel{\&OverBar;}{x})}^2}{n-1}}\\ =\sqrt{\frac{{(-0.014)}^2+{0.015}^2+{(-0.009)}^2+{(-0.003)}^2+{0.008}^2+{0.005}^2+{0.001}^2+{(-0.001)}^2}{7}}\\ =0.0093\end{array}$

(1.3) this group detects data detection

The detection data of (1.2) are pressed from small to large and arranged: 0.612,0.617,0.623,0.625,0.627,0.631,0.634,0.641

First choose maximal value 0.641, be calculated as follows

$G_n=\frac{(x_8-\stackrel{\&OverBar;}{x})}{s}=\frac{0.641-0.626}{0.00926}=1.613$

Determine that level of significance is α=0.05, find n=8 in " chemurgy analysis " book 220 page table 9-2 of the publication Yuan Guangwu chief editor of petroleum industry publishing house time, corresponding critical value G _table=2.03;

Judgement: because 1.613 < 2.03, so X ₈it not exceptional value;

Choose again minimum value 0.612, be calculated as follows

$G_n=\frac{(\stackrel{\&OverBar;}{x}-x_1)}{s}=\frac{0.626-0.612}{0.00926}=1.505$

Determine that level of significance is α=0.05, find n=8 in " chemurgy analysis " book 220 page table 9-2 of the publication Yuan Guangwu chief editor of petroleum industry publishing house time, corresponding critical value G _table=2.03;

Judgement: because 1.505 < 2.03, so X ₁it not exceptional value;

(1.4) rejecting abnormalities value

Because x ₁and x ₈not all exceptional value, so do not need to reject, can carry out calculation procedure below.

(1.5) another group detection data are undertaken by (1.2)-(1.4) step, and result, without exceptional value, can be carried out two sample precision consistency checks.

(2), two sample precision consistency check

(2.1) will check two groups of data of rear rejecting abnormalities value as the calculating data of this step by (1);

(2.2), by formula in entry (1.2.3), calculate respectively the standard deviation S of two groups of data ₁and S ₂;

S ₁＝0.00926

S ₂＝0.01071

(2.3) compare S ₁and S ₂size: S ₂> S ₁:

(2.4) calculate two sample precision consistance F _meter:

Computing formula is as follows:

(2.5) calculate respectively the degree of freedom V of every group of data variance _point:

V ₁＝n-1＝8-1＝7

V ₂＝n-1＝8-1＝7

(2.6) according to the degree of freedom numerical value of variance, the in the situation that of level of significance α=0.05, in " chemurgy analysis " book 228 page table 9-5 of the publication Yuan Guangwu chief editor of petroleum industry publishing house, find corresponding F value, F _table=4.99

(2.7) draw judged result according to institute's value:

Because 1.3377 < 4.99, so two kinds of Standard of analytical methods deviation there was no significant differences, precision is consistent;

(3), two sample mean consistency checks;

(3.1) will be by two groups of data after the inspection of (1), (2) the calculating data as this step;

(3.2) calculate respectively the arithmetic mean value of two groups of data by formula in entry (1.2.1);

${\stackrel{\&OverBar;}{x}}_1=0.626$

${\stackrel{\&OverBar;}{x}}_2=0.638$

(3.3) calculate respectively the standard deviation S of two groups of data by formula in entry (1.2.3) ₁and S ₂;

S ₁＝0.00926；

S ₂＝0.01071

(3.4) determine degree of freedom V _always

V _always=n ₁+ n ₂-2=8+8-2=14

(3.5) S _closecalculate

S _closecomputing formula is:

(3.6) two sample mean consistance T _closevalue is calculated

T _closevalue computing formula is:

(3.7) total degree of freedom V of two groups of detection data of foundation _alwaysnumerical value the in the situation that of two-sided test level of significance α=0.05, is found degree of freedom and is the T value of 14 o'clock, T in petroleum industry publishing house publishes " chemurgy analysis " book 229 page table 9-9 of Yuan Guangwu chief editor _table=2.145

(3.8) draw judged result according to institute's value:

Because 2.4 > 2.145, so these two kinds of analytical approach the data obtained average results there are differences, average result is inconsistent statistically, calculates and stops, and judges that these two kinds of analytical approach results are inconsistent.

Embodiment bis-

Determination methods comprises test of outlier, the consistency check of two sample precision and three steps of two sample mean consistency check, and concrete steps are as follows:

(1), test of outlier:

(1.1) test the data obtained is as follows:

A) material trademark: ZL106 aluminium alloy; Measure element: Ti; Measuring method: two antipyrine methane spectrophotometry titanium amounts (GB/T6987.12-2001); ICP-AES is measured Cu, Mg, Zn, Cd, Fe, Mn, B, Ti, Zr, V, Ni, Cr content (HB6731.10-2005);

B) detect data without technological deficiency, order ticket requires to retain 2 significant digits significant figure, and this detection data retain 3 position effective digitals, and data are taked the revision of the convention principle of " four houses six enter five Dan Shuan ".Detect data in table 1:

The analysis result of Ti element in table 1:ZL106 aluminum alloy materials

(1.2) one group of data of extraction GB/T6987.12-2001 analytical approach gained, carry out data processing;

(1.2.1) calculate the arithmetic mean value of these group data, formula is as follows:

$\begin{array}{c}\stackrel{\&OverBar;}{x}=\frac{0.185+0.187+0.185+0.183+0.185+0.182+0.183+0.183}{8}\\ =0.184\end{array}$

(1.2.2), taking arithmetic mean value as true value, calculate the absolute deviation of each data;

Absolute deviation=measured value-arithmetic mean value

Absolute deviation is followed successively by: 0.001,0.003,0.001 ,-0.001,0.001 ,-0.002 ,-0.001 ,-0.0010

(1.2.3) calculate standard deviation

The date expression of standard deviation is:

$\begin{array}{c}s=\sqrt{\frac{\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{(x_i-\stackrel{\&OverBar;}{x})}^2}{n-1}}\\ =\sqrt{\frac{{0.001}^2+{0.003}^2+{0.001}^2+{(-0.001)}^2+{0.001}^2+{(-0.002)}^2+{(-0.001)}^2+{(-0.001)}^2}{7}}\\ =0.0016\end{array}$

(1.3) this group detects data detection

The detection data of (1.2) are pressed from small to large and arranged: 0.182,0.183,0.183,0.183,0.185,0.185,0.185,0.187

First choose maximal value 0.187, be calculated as follows

$G_n=\frac{(x_8-\stackrel{\&OverBar;}{x})}{s}=\frac{0.187-0.184}{0.0016}=1.875$

Determine that level of significance is α=0.05, find n=8 in " chemurgy analysis " book 220 page table 9-2 of the publication Yuan Guangwu chief editor of petroleum industry publishing house time, corresponding critical value G _table=2.03;

Judgement: because 1.875 < 2.03, so X ₈it not exceptional value;

Choose again minimum value 0.182, be calculated as follows

$G_n=\frac{(\stackrel{\&OverBar;}{x}-x_1)}{s}=\frac{0.184-0.182}{0.0016}=1.25$

Determine that level of significance is α=0.05, find n=8 in " chemurgy analysis " book 220 page table 9-2 of the publication Yuan Guangwu chief editor of petroleum industry publishing house time, corresponding critical value G _table=2.03;

Judgement: because 1.25 < 2.03, so X ₁it not exceptional value;

(1.4) rejecting abnormalities value

Because x ₁and x ₈not all exceptional value, so do not need to reject, can carry out calculation procedure below.

(1.5) another group detection data are undertaken by (1.2)-(1.4) step, and result, without exceptional value, can be carried out two sample precision consistency checks.

(2), two sample precision consistency check

(2.1) will check two groups of data of rear rejecting abnormalities value as the calculating data of this step by (1);

(2.2), by formula in entry (1.2.3), calculate respectively the standard deviation S of two groups of data ₁and S ₂;

S ₁＝0.0016

S ₂＝0.0008

(2.3) compare S ₁and S ₂size: S ₁> S ₂:

(2.4) calculate two sample precision consistance F _meter:

Computing formula is as follows:

(2.5) calculate respectively the degree of freedom V of every group of data variance _point:

V ₁＝n-1＝8-1＝7

V ₂＝n-1＝8-1＝7

(2.6) according to the degree of freedom numerical value of variance, the in the situation that of level of significance α=0.05, in " chemurgy analysis " book 228 page table 9-5 of the publication Yuan Guangwu chief editor of petroleum industry publishing house, find corresponding F value, F _table=4.99

(2.7) draw judged result according to institute's value:

Because 4.00 < 4.99, so two kinds of Standard of analytical methods deviation there was no significant differences, precision is consistent;

(3), two sample mean consistency checks;

(3.1) will be by two groups of data after the inspection of (1), (2) the calculating data as this step;

(3.2) calculate respectively the arithmetic mean value of two groups of data by formula in entry (1.2.1);

${\stackrel{\&OverBar;}{x}}_1=0.184$

${\stackrel{\&OverBar;}{x}}_2=0.185$

(3.3) calculate respectively the standard deviation S of two groups of data by formula in entry (1.2.3) ₁and S ₂;

S ₁＝0.0016；

S ₂＝0.0008

(3.4) determine degree of freedom V _always

V _always=n ₁+ n ₂-2=8+8-2=14

(3.5) S _closecalculate

S _closecomputing formula is:

(3.6) two sample mean consistance T _closevalue is calculated

T _closevalue computing formula is:

(3.7) total degree of freedom V of two groups of detection data of foundation _alwaysnumerical value the in the situation that of two-sided test level of significance α=0.05, is found degree of freedom and is the T value of 14 o'clock, T in petroleum industry publishing house publishes " chemurgy analysis " book 229 page table 9-9 of Yuan Guangwu chief editor _table=2.145

(3.8) draw judged result according to institute's value:

Because 1.538 < 2.145, so these two kinds of analytical approach the data obtained average results not there are differences, average result is consistent statistically, judges that these two kinds of analytical approach results are consistent.

Embodiment tri-

Determination methods comprises test of outlier, the consistency check of two sample precision and three steps of two sample mean consistency check, and concrete steps are as follows:

(1), test of outlier:

(1.1) test the data obtained is as follows:

A) material trademark: K465 high temperature alloy; Measure element: Mo; Measuring method: EDTA volumetric determination molybdenum content (HB5220.21-2008); Thiocyanate Their Determination by Spectrophotometry molybdenum content (HB5220.22-2008);

B) detect data without technological deficiency, order ticket requires to retain 2 significant digits significant figure, and this detection data retain 3 position effective digitals, and data are taked the revision of the convention principle of " four houses six enter five Dan Shuan ".Detect data in table 1:

The analysis result of Mo element in table 1:K465 high-temperature alloy material

(1.2) one group of data of extraction HB5220.21-2008 analytical approach gained, carry out data processing;

(1.2.1) calculate the arithmetic mean value of these group data:

$\stackrel{\&OverBar;}{x}=1.503$

(1.2.2), taking arithmetic mean value as true value, calculate the absolute deviation of each data;

Absolute deviation=measured value-arithmetic mean value

Absolute deviation is followed successively by :-0.068,0.029,0.015 ,-0.006 ,-0.081 ,-0.020,0.029,0.029

(1.2.3) calculate standard deviation

The date expression of standard deviation is:

$\begin{array}{c}s=\sqrt{\frac{\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{(x_i-\stackrel{\&OverBar;}{x})}^2}{n-1}}\\ =0.0494\end{array}$

(1.3) this group detects data detection

The detection data of (1.2) are pressed from small to large and arranged: 1.422,1.435,1.495,1.509,1.530,1.544,1.544,1.544

First choose maximal value 1.544, be calculated as follows

$G_n=\frac{(x_8-\stackrel{\&OverBar;}{x})}{s}=\frac{1.544-1.503}{0.0494}=0.829$

Determine that level of significance is α=0.05, find n=8 in " chemurgy analysis " book 220 page table 9-2 of the publication Yuan Guangwu chief editor of petroleum industry publishing house time, corresponding critical value G _table=2.03;

Judgement: because 0.829 < 2.03, so X ₈it not exceptional value;

Choose again minimum value 1.422, be calculated as follows

$G_n=\frac{(\stackrel{\&OverBar;}{x}-x_1)}{s}=\frac{1.503-1.422}{0.0494}=1.64$

Determine that level of significance is α=0.05, find n=8 in " chemurgy analysis " book 220 page table 9-2 of the publication Yuan Guangwu chief editor of petroleum industry publishing house time, corresponding critical value G _table=2.03;

Judgement: because 1.64 < 2.03, so X ₁it not exceptional value;

(1.4) rejecting abnormalities value

Because x ₁and x ₈not all exceptional value, so do not need to reject, can carry out calculation procedure below.

(1.5) another group detection data are undertaken by (1.2)-(1.4) step, and result, without exceptional value, can be carried out two sample precision consistency checks.

(2), two sample precision consistency check

(2.1) will check two groups of data of rear rejecting abnormalities value as the calculating data of this step by (1);

(2.2), by formula in entry (1.2.3), calculate respectively the standard deviation S of two groups of data ₁and S ₂;

S ₁＝0.0494

S ₂＝0.0201

(2.3) compare S ₁and S ₂size: S ₁> S ₂:

(2.4) calculate two sample precision consistance F _meter:

Computing formula is as follows:

(2.5) calculate respectively the degree of freedom V of every group of data variance _point:

V ₁＝n-1＝8-1＝7

V ₂＝n-1＝8-1＝7

(2.6) according to the degree of freedom numerical value of variance, the in the situation that of level of significance α=0.05, in " chemurgy analysis " book 228 page table 9-5 of the publication Yuan Guangwu chief editor of petroleum industry publishing house, find corresponding F value, F _table=4.99

(2.7) draw judged result according to institute's value:

Because 6.04 > 4.99, so two kinds of Standard of analytical methods deviations exist significant difference, precision is inconsistent; Now do not need to carry out t distribution inspection, calculate and stop, illustrate this in two detection method result inconsistent.

Claims (4)

1. the conforming determination methods of different chemical composition analytical approach result, is characterized in that, determination methods comprises test of outlier, the consistency check of two sample precision and three steps of two sample mean consistency check, and concrete steps are as follows:

(1), test of outlier:

(1.1) different chemical composition analytical approach result the data obtained should meet following requirement:

A) two groups of detection data of same material, identity element, different analytical approach gained;

B) detect data without technological deficiency;

C) detect the choice of data end and take the revision of the convention principle of " four give up six enters five Dan Shuan ", " four houses six enter five considerations, all zero look odd even after five rear non-zeros enter, five, and five is front for occasionally casting out, and five is front for very entering " 0 is considered as even number;

D) every group of detection data unit is identical;

E) every group of detection data significant figure are identical;

F) each group detects data amount check n >=3,

(1.2) extract two groups of one group of data detecting in data, carry out data processing:

(1.2.1) calculate the arithmetic mean value of these group data, formula is as follows:

In formula:

X ₁, x ₂... x _n-measured value;

The symbol of ∑-expression summation;

-represent from x ₁to x _nmeasured value summation;

(1.2.2), taking arithmetic mean value as true value, calculate the absolute deviation of each data;

Absolute deviation=measured value-arithmetic mean value

(1.2.3) calculate standard deviation

Standard deviation data expression formula is:

In formula:

S-standard deviation

the arithmetic mean of-n observed reading

X _i-any single-measurement value

N-detects data amount check

(1.3) this group detects data detection

The detection data of (1.2) are pressed from small to large and arranged;

Choose respectively the maximal value and the minimum value that detect data, be calculated as follows respectively

X _i-be maximum or minimum measured value

Determine that level of significance is α=0.05, in the T of Grubbs method of inspection value table, find corresponding critical value, be labeled as G _table;

Judgement: work as G _n> G _tabletime, judge measured value x _ifor exceptional value; Otherwise there is no exceptional value;

(1.4) rejecting abnormalities value

If x _ifor exceptional value, need to reject, if detect data amount check n>=3 after rejecting abnormalities value, remaining data should be continued to the step process by (1.2)-(1.3), until without exceptional value; After rejecting abnormalities value, if data amount check n < 4 calculates and stops, can not carry out data consistent sex determination;

(1.5) undertaken another by (1.2)-(1.4) step and organize the inspection that detects data;

(2), two sample precision consistency check

(2.1) will check two groups of data of rear rejecting abnormalities value as the calculating data of this step by (1), and data amount check n meet n >=3;

(2.2), by formula in entry (1.2.3), calculate respectively the standard deviation S of two groups of data ₁and S ₂;

(2.3) compare S ₁and S ₂size;

(2.4) calculate two sample precision consistance F _meter:

If S ₁> S ₂, calculate with following formula

If S ₂> S ₁, calculate with following formula

(2.5) calculate respectively the degree of freedom V of every group of data variance _point, computing formula is:

V _point=n-1

(2.6) according to the degree of freedom numerical value of variance, the in the situation that of level of significance α=0.05, in F check table or F distribution table, find F value, be labeled as F _table;

(2.7) draw judged result according to institute's value:

If F _meter≤ F _tablewhen value, there is not significant difference in two kinds of different chemical composition Standard of analytical methods deviations, and precision is consistent;

If F _meter> F _tablewhen value, there is significant difference in two kinds of different chemical composition Standard of analytical methods deviations, and precision is inconsistent, calculates and stop, and can not carry out data consistent sex determination;

(3), two sample mean consistency checks;

(3.1) will be by (1), (2) two groups of data after the assay was approved the calculating data as this step;

(3.2) calculate respectively the arithmetic mean value of two groups of data by formula in entry (1.2.1) with ;

(3.3) calculate respectively the standard deviation S of two groups of data by formula in entry (1.2.3) ₁and S ₂;

(3.4) determine degree of freedom V _always

V _always=n ₁+ n ₂-2

In formula:

N ₁-be the number of first group of data;

N ₂-be the number of second group of data;

(3.5) S _closecalculate

S _closecomputing formula is:

(3.6) two sample mean consistance T _closevalue is calculated

T _closevalue computing formula is:

(3.7) total degree of freedom V of two groups of detection data of foundation _alwaysnumerical value the in the situation that of level of significance α=0.05, is found T value in t inspection tables of critical values or t distribution table, is labeled as T _table

(3.8) draw judged result according to institute's value:

If T _close>=T _table,when value, two kinds of different chemical composition analytical approach average results there are differences, and average result is inconsistent statistically, calculate and stop, and judge that these two kinds of analytical approach results are inconsistent;

If T _close< T _table,when value, two kinds of different chemical composition analytical approach average results not there are differences, and average result has consistance statistically, judge that these two kinds of analytical approach results are consistent.

2. the conforming determination methods of a kind of different chemical composition analytical approach result according to claim 1, is characterized in that number n >=4 of every group of described experimental data.

3. the conforming determination methods of a kind of different chemical composition analytical approach result according to claim 1, it is characterized in that, described determination methods is chosen sample when being greater than 2, taking 2 as one judging unit, when one group of detection data in a judging unit exist exceptional value, after this exceptional value is disallowable, detect number n >=3 of data, this group detects data can detect a judging unit of sample reformulation with another.

4. the conforming determination methods of a kind of different chemical composition analytical approach result according to claim 1, it is characterized in that, described determination methods is chosen sample when being greater than 2, taking 2 as one judging unit, after one group of detection data in a judging unit exist this group of exceptional value data disallowable, another group in this judging unit detects data and another detects a judging unit of sample reformulation.