JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS Narrow failure planes are caused in some bulk handling equipment such as screw hopper dischargers, powder blenders and some types of scraper conveyor. Power requirements are sensitive to loading normal to the plane and starting loads are frequently many times greater than forces operating at working speed. Within this narrow band of failure, attrition is a function of load, and segregation due to percolation is a feature of dilation. Bulk shear takes place in screw conveyors and elevators, blenders, scraper conveyors and in "flowing material" in similar classes of equipment. Segregational effects are usually consequential to differing sizes of particles in the mass, the separation of these in the bulk being mainly due to a percolating or preferential locating mechanism for the smaller particles. Dilate conditions favour the ease of particle movement and as in bulk shear some re-orientation is taking place, the hazard of segregation tends to in- crease. The depth of bed section is an obvious limit to the depth of per- colation so a continuous "live" shallow stream will not incur the residual effects of batch plant operation or sustained flow over a "dead" layer. FREE SURFACE ACTIVITY In unconfined conditions the boundaries of the solids are acted on by gravity only and the surface will display behaviour which depends on the dilation of the solid. "Angle of repose" is only meaningful in specified conditions for most materials, and has no value other than its defined quality in specific circumstances (32). In dynamic conditions such as rolling down a slope or chute, the tendency is to entrain air due to particle separ- ation in their agitated condition. This effect leading to "flooding" may be pronounced with fine, less dense powders in conditions of initial turbulence. Segregation due to a preferential "capture" mechanism of small par- ticles operating on the surface layers is particularly effective in the building up of a pile of material (33, 34). CHUTES AND CHANNELS Chutes may be classified as "slow" or "fast" depending on whether or not the flow channel is fully occupied by the bulk material (35). Fast chutes operate with clearance between the top surface of the powder stream and the upper boundary wall, and although suitable for high rates of transit, such circumstances favour subsequent "flooding" and segre- gation. In principal, chutes form part of the flow path of the stream and

SOME ASPECTS OF HANDLING POWDERS IN MECHANICAL EQUIPMENT 699 transitional points of the route into and out of machines should receive detailed attention (36) with reference to the powder change of state, e.g. "slow" chutes passing a live plug of matehal should be provided with room for expansion of the mass at a change of direction {37) even if subsequent reduction of the flow channel is required. This feature is often given unconsciously by "wear boxes" designed to avoid abrasion of chute matehal at points of change in direction of chutes. It is common practice to consider chutes as mere sealed communicating channels between sections of plant whereas vertical falls, projecting ledges, wall irregularities, crevices and impingement points all contribute to the conditioning of the powder stream and apparently minor features can lead to flow stoppage (38). Once a chute even of modest length is blocked, in- creasing the feed pressure invariably fails to promote flow due to regener- ative wall friction and if the feed in is by mechanical means, overload or failure of some part is incurred. LIMITATIONS OF PRESENT THEORY 1. It appears obvious that no refinement of theory can accommodate a lack of fundamental data relating to variations of moisture content, tem- perature, organic or chemical reactions, variation in product, surface finish, etc., but in many cases precision is not possible for reasons of un- certainty, or confounding of variables. It may not be economic or possible to design for the worst combination of conditions but the effect of ex- ceptional conditions on both plant and production should be assessed in budget and contingency planning. 2. On the smaller scale, a large number of detail problems may not be economic to solve or even investigate. This is a practical rather than negative approach for example, the flow properties of small hoppers can frequently be established most economically by full scale test. Whether to do this by a "bare" trial hopper or manufacture the final item of equip- ment, is an economic calculation taking account of a subjective judgement of risk and cost based on experience. 3. Various forms of "imperfect" materials create testing problems, fibrous, rod shaped, flocculent, interlocking, materials with transient properties, indeterminate moisture distribution or slip-stick character- istics. 4. Safety factors included in theoretical designs may lead to conser- vative solutions incompatible with other limiting considerations of site or