Sample increments

Samples from municipal wastewater treatment plants and samples of industrial discharges often are collected as 24-h composites. Samples are obtained using an automatic sampler that periodically removes individual grab samples. The volume of each sample increment and the frequency of sampling may be constant or may vary in response to changes in flow rate. 

Extractions performed with the conveyor stopped allow more assured accuracy by the certainty of including fines in the sample increment. Sampler design to extract increments from a flat belt or rotaiw table sampler while the conveyor is stopped minimizes potential for residual fine particles remaining on the conveyor surface in carrying out extractions. See Fig. 19-6 for rotary table sampler extraction diagram. 

The sampling variance, V, is inversely proportional to the actual number of sampling increments (n) and may be expressed as ... 

Therefore, the challenge in sampling solids for environmental analysis is to collect a relatively small portion of the sample that accurately represents the composition of the whole. This requires that sample increments be collected such that no piece, regardless of position (or size) relative to the sampling position and implement, is selectively collected or rejected. Optimization of solids sampling is a function of the many variable constituents of coal and is reflected in the methods by which an unbiased sample can be obtained, as is required by coal sampling (ASTM D197). 

A statistically valid sampling plan requires careful design emd execution so that generalizations based on mathematical probability can be drawn from a small number of test portions. Guidelines are given for estimation of the minimum number and size of sample increments needed to achieve a given level of confidence in chemical analyses. 

If the sample increment is larger than the amount (test portion) needed per measurement, subsampling is necessary. 

Major considerations" in any sampling plan are the size and number as well as the location of the sampling increments. The following sections consider aspects of these points. 

A second factor to consider in a valid sampling plan is the collection of enough individual sample increments to ensure that heterogeneity on a large scale does not bias the results. Estimation of this number can be made straightforwardly if the component of interest is distributed... 

The minimum number of sample increments required can be calculated from either Equation 2 or 5. From Equation 2, assuming that a 50% level of confidence is desired and that an acceptable percent relative standard deviation R is 50, n... 

From direct inspection of any lot portion alone, it is not possible to ascertain whether it is a representative sample, or a specimen only. Focus must instead be exclusively on the sampling process. The basic prerequisite for a representative sampling process is that all potential increments from a batch, container, or pipeline have the same probability P of being selected, and that the elements selected are not altered in any way after the sample/increment has been taken. Furthermore, all elements that do not belong to the batch or container must have zero probability of being selected, for example, leftovers at the bottom of the lot, or leftovers in the sampling equipment from previous increments/batches (cross-pollution). 

Correct process sampling increment delimitation/extraction... 

The samples collected can either be assayed individually or combined, and then a subset of the gross sample is assayed as depicted in Figure 1 and described below. Sample increments should be combined on a clean, dry surface or in a suitable container or bag. All containers that the sample comes into contact with should be inert and not chemically or physically react with the sample. In addition, accurate sample labeling and records must be kept. If the sample is divided, then the retained portion should be kept for possible future analysis and labeled with at least the date, gross field sample number, lot number, and/or reference number from bill of lading. The material to be tested should be labeled with at least the date, gross field sample... 

Transportation (by belt, rail, or truck) can initiate (due to movement of the coal) processes that result in size and density segregation. Thus, variations from one side of a conveyor belt to the other, from side-to-side, end-to-end, and top-to-bottom locations in individual cars or trucks, and between one location and another in a coal pile, must be anticipated (ASTM D-346 ASTM D-2234 ASTM D-4182 ASTM D-4702 ASTM D-4915 ASTM D-4916 ASTM D-6315 ASTM D-6518 ASTM D-6543 ISO 1988). Therefore, the challenge in sampling coal from a source or shipment is to collect a relatively small portion of the coal that accurately represents the composition of the coal. This requires that sample increments be collected such that no piece, regardless of position (or size) relative to the sampling position and implement, is collected or rejected selectively. Thus, the coal sample must be representative of the composition of the whole coal (i.e., coal in a pile or coal in a railcar or truck) as represented by the properties or quality of the sample. 

Stream sampling and flow sampling are terms usually reserved for the collection of sample increments from a free-falling stream of coal as opposed to the collection of increments from a motionless (stopped) conveyor belt. Coal that passes from one belt to another at an angle tends to become segregated because... 

Some general problems associated with the determination of sulfur in coal are nonuniform distribution of pyrite particles, failure to convert all the sulfur to sulfate, and loss of sulfur as sulfur dioxide during the analysis. The nonuniform distribution of pyrite necessitates the collection of many sample increments to ensure that the gross sample is representative of the lot of coal in question. Pyrite particles are both hard and heavy and have a tendency to segregate during the preparation and handling of samples. Because the particles are harder, they are more difficult to crush and pulverize and tend to concentrate in the last portion of material that remains from these processes. 

Thompson et al.13 have shown quality control in sampling to be practicable in sampling procedures requiring a combination of sample increments to form a composite sample. It is very important, however, that the approach proposed does not require any extra time and resources beyond those used for the normal sampling procedure that precedes analysis. 

Scheme a Take five sample increments and blend them. Perform a duplicate analysis of the blended sample. 

Scheme b Take three sample increments and perform a duplicate analysis on each. 

Characteristic value of the Gaussian standard deviation a is 10 cm (range 10-20 cm ) and of the sampling increment I is 5 cm (range 2-20 cm ), resulting in 800 descriptor variables. 

---