Assuming that there is a particle size range it will be difficult to determine the average composition without taking a number of samples. Provided the truck is always loaded in a similar manner it may be useful to construct a pattern of deposition from multiple samples and derive a segregation pattern, taking account of the volumes involved relative to these. From this it may be possible to assess the number and location of samples required to give an acceptable aproximation of the overall composition. ■

Filling up to 40 cubic metres of product through two or three loading points will, in itself, present some difficulty in securing representative samples from the truck. Quite importantly, there is also the uncertainty of the consistency of the feed stream, as this will depend on how it was previously stored and discharged. It would seem probable that this amount of material will have been held in a storage silo awaiting discharge to the truck and with therefore have suffered some degree of prior segregation that would reflect on the uniformity of the loading. Also, unless the total contents of the holding vessel were emptied into the truck at every cycle, this would result in much uncertainty as to the condition of the material that is being discharged. A particularly insidious form of segregation takes place when a funnel flow storage vessel is reloaded before total discharge. Outflow develops a ‘drained repose cone’ from which the surface layers slough off to flow down a narrow flow channel through the rest of the static bulk. As this process collects a uniform layer of material from the whole cross section of the silo, it remixes any radial segregation that occurred during filling and gives a well-balanced size distribution. Fresh material entering the silo from a single flow stream deposits a positive repose pile that usually tends to lay coarse fractions on the surface of this drained repose cone as it refills. The problem is that as the silo discharges later, and the content level falls below that of the previous discharge period, this whole layer of coarse fractions comes together to run out as a concentrated ‘flash’ of coarse fractions within a period that the previous and following material was very representative of the average composition. This concentration can be ‘lost’ in subsequent storage and handling, but is especially unfortunate if going to a continuous process or packing.

The most effective way to determine the mean composition of the truck contents would therefore to take periodic samples of the in-feed stream, rather than try to establish the overall size distribution from the deposited bulk. This would require a detailed examination of the prior storage and discharge process, but it may be useful to review the overall prior flow route to assess if and how segregation may have taken place, as this could point to the form of variation that may be expected and provide a guide to the number and timing of sample extraction that should be taken during the fill period. Samples should take a ‘slice’ of the feed stream cross section. If too much is extracted for analysis it should be reduced by means of a spinning riffler. This situation exemplifies the value of undertaking a flow route audit when interested in the effects of segregation. ■