This article describes potential projects for MISG 1996, to be held from 29 January to 2 February 1996 at The University of Melbourne.
Eight projects are all that can be managed during the week. The projects below look the most likely to be studied, although there are still a few other companies who are anxious to present projects. The final list and further details of the projects will be available in mid-December. The list below is in company alphabetic order.
If you need a registration form and details regarding accommodation please contact Kerry Landman. There is no registration fee for students and delegates from academic and government laboratories. Partial financial support to out-of-town students and mathematicians participating at MISG 1996 will be offered.
Warehouse optimisation (Australian Paper): Australian Paper supplies a dozen box plants around Australia with various grades of paper from 6 paper machines. Most grades are made in fortnightly cycles, but the storage capacity at most box plants is not large enough to take the total production. Some plants may be only able to immediately accept 30% of a production run of a particular grade of paper. The rest will have to be temporarily stored elsewhere, at a significant cost. The problem is to optimise the storage of reels in the box plant so as to minimise the need to use an outside store. An algorithm is needed so that a decision can be made daily about how many reels of each specification should be either directly accepted from the paper mill (when available) or drawn from the outside store.
Methodologies for decision-making in tree breeding programs (Various industries involved in the CRC Temperate Hardwood Forestry): Tree breeding operations aim to maximise the rate of genetic progress per year. To be efficient, tree breeders have to take a number of key decisions based on various tree characteristics. The companies wish to know how to deal with the complexity and uncertainty associated with these decisions in a systematic manner.
Microwave heating of a slurry Part 2 (Kodak): As a direct result of MISG 1995, Kodak needs to investigate enlarging the cylindrical reactor tube which transports the following slurry through the microwave chamber. Now the electric field can no longer be assumed to be uniform. Non uniformities in this field will give rise to hot and cold spots in the reactor tube. Kodak requires an assessment of how the temperature difference between the hot and cold spots depends on the various parameters, in particular the ratio of the reactor tube diameter to the chamber size.
Homogenisation of photographic dispersions (Kodak): After the heating process described at MISG 1995, the material is a oil-in-water emulsion. The oil phase consists of droplets between 1-2 micron in size. These droplets need to be broken to a tenth of this size in order to make high quality photographic paper. Under high pressure the emulsion flows down a tube and meets a plate with an orifice. The oil droplets must be broken up within this orifice to produce smaller oil droplets on the outlet side. Kodak requires an understanding of how the resulting size of the droplet depends on orifice shape, thickness of plate, viscosities, flows speeds and surface tension. Also, both an individual plate or a series of plates can be analysed.
Unsteady-state cooling of jarred cheese spreads (Kraft): The project is to develop a model of the overall process used to cool cheese spread products to help optimise performance and hence minimise costs associated with residual heat damaged product. The jarred product is produced at high volumetric rates. A zoned cooling tunnel uses water sprays to force cool the product. It is then placed in a cardboard shipper and closed stacked in forced draft ambient ventilation racks prior to shipment offsite. A model of the cooling process would enable the cooling rate to be optimised with the system constraints and enable key parameters to be identified. The model could also be used to determine the impact of changing packaging, for example, producing a larger jar.
Modelling optical fibre cable (MM Cables): The typical optical fibre cable is made in two stages: 6 optical fibres are enclosed by a loose fitting sheath, then 6 to 12 of these loose tubes are wrapped around a central strength member. These tubes are wrapped helically for a number of turns and then the rotation is reversed. Hence reversal sections are between helical sections of opposite rotation. When the cable is elongated under tension or compressed by low temperatures, where do the fibres go? What are the critical parameters at the stranding process and what tolerances should be applied to these parameters to ensure that the cable will comply to the customer specifications for tension and temperature?
Labelling of wine bottles (Southcorp):
(a) Pressure sensitive labelling: Creasing of pressure sensitive labels is
frequently experienced during application to wine bottles. The creasing
shows up as vertical and horizontal lines, where the paper has been brushed
or rolled by mechanical devices to ensure firm adhesion. Uneven glass
surfaces, the size of the label and the type of adhesive used all
contribute to the defect. Small labels do not show this defect. What are
the maximum label dimensions before wrinkling occurs?
(b) Wet gum labelling: Under adverse weather conditions 'air bubbles' appear
randomly distributed on wet gum labels. When a bottle sample was removed
from the conveyer belt immediately after label application, the label
seemed evenly affixed. However, some 10 minutes later, the quality
deteriorates, as moisture is adsorbed by the paper and bubbles form. (There
is glue underneath these bubble prone areas.) The defects are not permanent
and reverse when the paper dries out. The bubbles reappear under humid or
wet conditions. How can these defects be eliminated - how do they depend
on glue setting times, paper properties, machine operating speeds?
Modelling the cooking process of a single cereal grain (Uncle Toby's): The process of `cooking' involves hydrating the grain contents followed by heating to a temperature in excess of the gelatinisation temperature. Currently the grain is cooked by applying the hydration and heating process simultaneously - the possibility of hydrating and heating as sequential processes needs to be explored. The hydration and heating rates as a function of various process controls is required.
Cereals manufacturing plant production scheduling (Uncle Toby's): Cereal products are manufactured in a facility which uses several complex but interdependent processes. This means that planning to manufacture some products must be scheduled when others are not being processed on adjacent lines. A planning system which recognised these interdependencies is needed to enable rapid changes to the production schedule at short notice.
Blending methodologies in talc operations (Western Mining Corp.): Each daily production run of lump talc (300-500 tonne) is automatically sampled every 90 seconds and that sample is crushed and split to obtain reflectance data (for whiteness levels) and chemical data (for contaminant levels) for that batch. This batch data is tracked to the port where large stockpiles are ready for shipping. These stockpiles are segregated into defined products and trainloads are added to stockpiles with grades being weighted by the tonnages. At the mine, the batches from each day are stockpiled separately. The material is `manually' blended onto trucks to be put onto the train, or loaded directly to fit a product specification. WMC is concerned about this methodology and wants suggestions on sampling rates, blending algorithms, efficient utilisation of lower grade materials and optimising the highest grade materials.
Department of Mathematics, University of Melbourne
Parkville, Vic 3052
ph (613) 9344 6762, fax: (613) 9344 4599, email:
misg@maths.mu.oz.au