Het belong van automatische monsterneming; mogelijkheden en toepas‐singen

Automatic sampling; its possibilities and applications. The problem considered is that of sampling a lot of coal or ore in order to determine the average content of some essential component; as a particular example the sampling of coal for ash content is taken. Three sources of error are considered, namely: 1 the sampling error, depending on the size of, and the distribution of ash content amongst, the grains; this error can be reduced by taking a large sample; 2 the error due to inhomogeneities in the lot, which can be cut down by using a gross sample composed of a large number of increments; and 3 the error in the analysis, including the errors introduced by grinding and dividing the gross sample into the final sample of about 1 gram used for the determination of ash content. The accuracy (a) corresponding to 99% confidence limits is expressed in a simple formula, (1), in which the influence of each of the sources of error is represented by a separate term. Taking as a lot the coal produced at a mine during a shift, the formula shows that, to achieve a reasonable accuracy of a =i/0 ash, at least some 150 to 200 increments have to be taken, the second sourse of error (inhomogeneity of the lot) being by far the most important one. The size of the increments can be relatively small so that the weight of the gross sample stays within acceptable limits. This large number of increments necessitates the use of an automatic sampling procedure. In § 3 a sampler developed and installed at one of the Dutch State Mines is described together with the precautions required to prevent bias in the samples due tj segregation of the coal across the conveyor belt. For further details the English reader may be referred to Fuel 19, 101–105,1950.