CN103869053B - Regional geochemical survey sample analysis and abnormal point sampling inspection method - Google Patents

Abstract

The invention discloses a regional geochemical survey sample analysis and abnormal point sampling inspection method which mainly solves the problems that an abnormal point selected by a sampling inspection method in the prior art is lack of representativeness and the purpose of monitoring the analysis quality of the whole batch is not actually achieved. The regional geochemical survey sample analysis and abnormal point sampling inspection method comprises the following steps: calculating an expected number of a measurement value of a point, close to m, of a sample; calculating a residual error Yi of a measurement value xi of the sample, which deviates from an expected value omega i according to the measurement value xi of the sample and an expected value omega of the measurement value of a point close to the sample; calculating a relative deviation value of the sample, which deviates from the expected value, according to the expected value omega i and the residual error Yi in the step; sequentially sequencing according to a relative deviation value Zi; and performing sampling inspection on sequenced data according to a ratio or fixed quantity. By adopting the scheme, the regional geochemical survey sample analysis and abnormal point sampling inspection method has the advantages that errors in a monitoring and analyzing process are reliably and completely detected, the detection leak in the field is eliminated, the actual demand purpose is fulfilled, and high practical value and popularization value are achieved.

Description

Samples in regional geochemical survey analyzes abnormity point sampling observation method

Technical field

The present invention relates to a kind of sampling observation method, specifically, relate to a kind of samples in regional geochemical survey and analyze abnormity point sampling observation method.

Background technology

Geological experiment test is one of geochemical prospecting important means of looking for ore deposit, the reliability of its test result directly can have influence on the regional geochemistry basis establishment of map and the delineation of geochemical anomaly, according to the requirement of specification " geological and mineral laboratory test mass management regulation ", " for preventing the geochemical map illusion caused owing to analyzing accidental error, the high point of reply sudden change and sudden change low spot carry out the iterative testing of 3% ", its objective is and guarantee that test data accurately and reliably, the selection of catastrophe point seems very important, because samples in regional geochemical survey has quantity large, scope is wide, analytical element is many, the features such as constituent content fluctuation is large, make quality control officer operation easier and workload in the judgement of abnormity point quite large.

In actual applications, because " geological and mineral laboratory test mass management regulation " is not clearly determined the selective examination ratio of exception and selective examination method, the method that each laboratory is adopted also is not quite similar, and roughly has following two kinds:

One, the method for maximal value and minimum point is only spot-check

The sample data amount that ore deposit, district is adjusted is more, and some laboratory is due to manpower, and abnormal extraction can only according to " geological and mineral laboratory test mass management regulation " requirement, generally take high level 1.5%, the method of low value 1.5%, although this method is simple, is obviously incomplete, particularly when running into continuous print high level point and continuous print low value point, for sample by the gross, representative very poor, science is not strong, some laboratory also has the way only extracting high level, inadvisable especially;

Two, the sampling observation method of three times of standard deviations

This method is based on mathematical statistics, namely the mean value (ω) of sample is by the gross obtained, with standard deviation (S), emphasis extracts the sample higher than ω+3S, and this sampling observation method is seemingly reasonable, but it has obscured "abnormal" in analytical test and geologic "abnormal", miss out real significant "abnormal" in analytical test, larger problem is that the quantity extracted is uncontrollable, even if repeatedly calculate, quantity reaches requirement, but operation is very loaded down with trivial details.

In sum, all there are same serious problems in existing sampling observation method, they all have ignored the fluctuation situation of the constituent content in different background district, the low abnormity point of high background area and the high abnormity point of low background area cannot be monitored, the abnormity point selected lacks representative, does not really reach the object that quality is analyzed in monitoring by the gross.

Summary of the invention

A kind of samples in regional geochemical survey is the object of the present invention is to provide to analyze abnormity point sampling observation method, mainly solve the fluctuation situation that the sampling observation method existed in prior art have ignored the constituent content in different background district, the abnormity point selected lacks representative, does not really reach the problem that the object of quality is analyzed in monitoring by the gross.

To achieve these goals, the technical solution used in the present invention is as follows:

Samples in regional geochemical survey analyzes abnormity point sampling observation method, comprises the following steps:

(1) calculation sample closes on the expectation value ω of the measured value of m point _i, m > 1;

(2) according to the measured value x of this sample _iand the expectation value ω of the measured value of its point of proximity that step (1) draws _i, use formula Y _i=x _i-ω _icalculate the measured value x of this sample _idepart from expectation value ω _iresidual error Y _i;

(3) according to the expectation value ω that step (1) draws _iand the residual error Y that step (2) draws _i, use formula Z _i=Y _i/ ω _icalculate the relative standard deviation values that this sample departs from expectation value;

(4) according to relative standard deviation values Z _isize sort successively;

(5) proportionally or fixed qty from step (4) sequence after data inspect by random samples.

Specifically, in described step (1), sample expectation value is drawn by following formula: ${\mathrm{\ω}}_i=\frac{1}{2m}(x_{i-m}$ $+...+x_{i-2}+x_{i-1}+x_{i+1}+x_{i+2}+...+x_{i+m}).$

Compared with prior art, the present invention has following beneficial effect:

(1) in the present invention, introduce the concept of relative deviation dexterously, and utilize computerized algorithm, establish effective mathematical model, thus achieve the dynamic surveillance of situation that constituent content beated, and then to have had the error in monitoring analysis process and more reliably, comprehensively detect, effectively eliminate the detection leak in this field in prior art, more realistic demand.

(2) in the present invention, after each sample is resequenced by relative standard deviation values more in proportion or fixed qty extract, thus have and automatically screening capacity is removed for continuous high level point and continuous low value point, and effectively can extract the low value point under high background value and the high abnormity point under low background, there is higher recognition capability, extract quantity can also arbitrarily increase according to the actual requirements or reduce, and the data extracted more comprehensively, have more representativeness, effectively can solve extraction number quantity not sufficient, representative not enough problem.

(3) the present invention realizes more for convenience, and the scope of application is comparatively wide, can be applied to computer programming calculation, also can try to achieve in Microsoft Excel, effectively can reduce the artificial extraction time, there is outstanding substantive distinguishing features and marked improvement, be applicable to large-scale promotion application.

Accompanying drawing explanation

Fig. 1 is the abnormal sample point distribution plan adopting max min to extract.

Fig. 2 is the abnormal sample point distribution plan adopting ω+3S to extract.

Fig. 3 is the abnormal sample point distribution plan that the present invention extracts.

Embodiment

Below in conjunction with drawings and Examples, the invention will be further described, and embodiments of the present invention include but not limited to the following example.

Embodiment

The fluctuation situation of the constituent content in different background district is have ignored in order to solve the sampling observation method existed in prior art, the abnormity point selected lacks representative, really do not reach the problem that the object of quality is analyzed in monitoring by the gross, the invention discloses a kind of novel samples in regional geochemical survey and analyze abnormity point sampling observation method, in the method, introduce the concept of relative deviation, by sorting from big to small or from small to large to the relative standard deviation values of each sample measured value, in proportion or fixed qty extract, realize the situation of beating of dynamic monitoring constituent content, the object of the error better in monitoring analysis process.

In order to prove the effect of the sampling observation method in the present invention, as shown in table 1, give the analysis data of certain batch of As element:

Table 1

Data are beated comparatively large as can be seen from the above table, and occur continuous print high level at sample spot 107 ~ 113 and 139 ~ 154 content, continuous print low value region appears in sample spot 084 ~ 106, guarantees that the abnormity point of this batch data of selecting properly is tested, very important.

As shown in Figure 1, if according to the way that max min extracts, then obtain 090,091,094,007 respectively, 110,111,112,144 8 sample spot, as can be seen from Figure 1 have together with three abnormal high level have concentrated on three abnormal low value points, miss out suspicious data point, lack representative; As shown in Figure 2, if what extract according to the method for ω+3S is 107,108,109,110,111,112,144 7 abnormal high level points, obviously, there is the Railway Project that we expect, the i.e. problem of high level point continuous drawing and extraction number quantity not sufficient, and cannot extract out for abnormal low value point, selective examination emphasis has still been placed on the large several samples of content.

The relative deviation method in the present invention is used to extract abnormal, obtaining following result is 007, 042, 106, 138, 004, 066, 144, 178(is shown in Fig. 3), the abnormity point that this method extracts is relatively uniform, continuous print high level point is removed, for low background 001 ~ 100 between sample also extracted abnormal high point (004 out, 066), also abnormal low spot (138) has been extracted out between high background area 106 ~ 155, representative, effect is very good, and abnormal quantity can arbitrarily be increased and be reduced, if needed, 112 and 123 two sample spot can also include abnormity point in.

Because above-mentioned data are more, only for 1 ~ 20 sample spot in table 1, the sampling observation method in the present invention is described in the present embodiment.

According to formula ${\mathrm{\ω}}_i=\frac{1}{2m}(x_{i-m}+...+x_{i-2}+x_{i-1}+x_{i+1}+x_{i+2}+...+x_{i+m})$ Calculation sample closes on the expectation value ω of the measured value of m point _i; According to formula Z _i=(X _i-ω _i)/(ω _i) relative standard deviation values of calculation sample, herein, m value is 2, i is 1 ~ 20, then expectation value and relative deviation as shown in table 2, wherein, the measured value that it closes on sample got by end value such as sample 1,2,19,20 etc.

Table 2

According to relative standard deviation values Z _isize from big to small sample is sorted, then draw sequence as described in Table 3:

Table 3

Can draw from this sequence, if be taken as abnormal high point from maximal value, abnormal low spot is taken as from minimum value, then can show that yp004 and yp008 is for abnormal high point, yp006 and yp007 is abnormal low spot, concrete sampling observation quantity is determined by user, according to a certain percentage or fixed qty carry out inspecting comprehensive sampling observation that just can realize sample spot by random samples from the data after above-mentioned sequence.

According to above-described embodiment, just the present invention can be realized well.

Claims (2)

1. samples in regional geochemical survey analyzes abnormity point sampling observation method, it is characterized in that, comprises the following steps:

(1) with closing on ithe front and back of point are each mthe measured value of individual point calculates described the ithe expectation value of point ω i, m > 1;

(2) according to the measured value x of this sample _iand the expectation value ω of the measured value of its point of proximity that step (1) draws _i, use formula Y _i=x _i-ω _icalculate the measured value x of this sample _idepart from expectation value ω _iresidual error Y _i;

(3) according to the expectation value ω that step (1) draws _iand the residual error Y that step (2) draws _i, use formula Z _i=Y _i/ ω _icalculate the relative standard deviation values that this sample departs from expectation value;

(4) according to relative standard deviation values Z _isize sort successively;

(5) proportionally or fixed qty from step (4) sequence after data inspect by random samples.

2. samples in regional geochemical survey according to claim 1 analyzes abnormity point sampling observation method, and it is characterized in that, in described step (1), the expectation value of the measured value of sample is drawn by following formula: ω _i= (x _i-m+ ... + x _i-2+ x _i-1+ x _i+1+ x _i+2+ ... + x _i+m).