The MIIS Eprints Archive: No conditions. Results ordered -Date Deposited. 2021-08-05T15:43:10ZEPrintshttp://miis.maths.ox.ac.uk/images/sitelogo.gifhttp://miis.maths.ox.ac.uk/miis/2012-02-03T21:53:41Z2015-05-29T20:11:12Zhttp://miis.maths.ox.ac.uk/miis/id/eprint/565This item is in the repository with the URL: http://miis.maths.ox.ac.uk/miis/id/eprint/5652012-02-03T21:53:41ZUsing fractals and power laws to predict the location of mineral depositsAround the world the mineral exploration industry is interested in getting that small increase in probability measure on the earth's surface of where the next large undiscovered deposit might be found. In particular WMC Resources Ltd has operations world wide looking for just that edge in the detection of very large deposits of, for example, gold. Since the pioneering work of Mandelbrot, geologists have been familiar with the concept of fractals and self similarity over a few orders of magnitude for geological features. This includes the location and size of deposits within a particular mineral province. Fractal dimensions have been computed for such provinces and similarities of these aggregated measures between provinces have been noted. This paper explores the possibility of making use of known information to attempt the inverse process. That is, from lesser dimensional measures of a mineral province, for example, fractal dimension or more generally multi-fractal measures, is it possible to infer, even with small increase in probability, where the unknown (preferably large) deposits might be located.M. BroadgateQ. ChengN. HaywardL. Jennings2012-02-03T21:51:29Z2015-05-29T20:11:09Zhttp://miis.maths.ox.ac.uk/miis/id/eprint/564This item is in the repository with the URL: http://miis.maths.ox.ac.uk/miis/id/eprint/5642012-02-03T21:51:29ZThe application of pesticides to grape bunchesThe application of pesticides to grape bunches is complicated by the different shapes and forms of the grape bunch during growth. Initially the grape bunch has a very porous structure, while in later stages the grapes are closely packed. We consider estimates of the flow velocity through the grape bunch, droplet density within the spray, probability of droplet impaction on a bunch or individual grape and the maximum size of drop that can adhere to a grape surface.S.I. BarryR.O. Weber2012-02-03T21:49:33Z2015-05-29T20:11:06Zhttp://miis.maths.ox.ac.uk/miis/id/eprint/563This item is in the repository with the URL: http://miis.maths.ox.ac.uk/miis/id/eprint/5632012-02-03T21:49:33ZDrying and curing of stains and lacquers used in furniture finishingMechanisms for both the blooming and blistering defects of concern to Nexus have been suggested, but much more work is needed to establish the exact causes of these problems.Y.M. StokesP. Pendleton2012-02-03T21:47:03Z2015-05-29T20:11:03Zhttp://miis.maths.ox.ac.uk/miis/id/eprint/562This item is in the repository with the URL: http://miis.maths.ox.ac.uk/miis/id/eprint/5622012-02-03T21:47:03ZInferring eye movements on the basis of head and visual target positionThe modern optical lens design process relies on ray-tracing that requires the geometry of the visual task to be simulated in order to derive the optical parameters. Such measurements must take into account distant, intermediate and near visual tasks. SOLA has developed a non-intrusive, low cost system which tracks eye movements during the reading process. This report analyses their procedure in order to determine the accuracy of the tracking system. It concludes that some assumptions in the existing algorithm are overly restrictive, while, overall, this is an effective tracking method. In addition, a Fourier analysis of the sampling rate demonstrates that 10 Hz is a sufficiently high rate to use, and that lossy compression is adequate for their needs.L.M. BattenB. Whiten2012-02-03T21:43:32Z2015-05-29T20:11:00Zhttp://miis.maths.ox.ac.uk/miis/id/eprint/561This item is in the repository with the URL: http://miis.maths.ox.ac.uk/miis/id/eprint/5612012-02-03T21:43:32ZEfficient loading of intermodal container trainsEfficient loading and unloading of trains is crucial if rail transport is to compete with road transport. The cost of rail transport can be reduced by improving the placement of containers on trains so that the number of wagons required is minimised and so that the train can be operated safely and efficiently. The cost of terminal operations, and the delays experienced by customers' trucks, can be reduced by improving the procedures used to load and unload trains.

We divided the problem into three subtasks. The first task is to develop a load plan for a train based on the expected mix of containers. An ideal plan specifies positions on the train for various container types in a way that minimises the number of wagons required while meeting a number of packing, safety and aerodynamic constraints. The second task is truck dispatching - when a truck arrives at the terminal, where should it be sent? This problem is not straightforward because trucks arrive randomly and loading and unloading occur simultaneously, and so the ideal position for a container may not yet be available. The third task is to dispatch the stackers that transfer containers between trains and trucks in such a way that truck waiting time is minimised.

The train planning problem can be formulated as an integer programming problem, but it can also be solved using the OASIS software already owned by National Rail Corporation. Train plans can be made more flexible by using container classes rather than specific container details.

The study group suggested several truck dispatch schemes, but did not have sufficient data to evaluate the schemes.A. ErnstP. Pudney2012-02-03T20:59:27Z2015-05-29T20:10:53Zhttp://miis.maths.ox.ac.uk/miis/id/eprint/559This item is in the repository with the URL: http://miis.maths.ox.ac.uk/miis/id/eprint/5592012-02-03T20:59:27ZA comparison of bearing life in new and refurbished railway axle boxesA simple linear dynamical model shows that at normal running speeds of freight wagons, forced oscillations due to periodic track compliance are transferred to the overlying unsprung mass and significantly amplified. Due to these oscillations, a small gap opens and closes between the collar of a journal bearing and the axle box many times every second. The forces between these components reach peaks of over 10 tonnes. This is an environment in which wear of the soft spherical graphite iron of the axle box will eventually take place.

Due to repeated unloadings of the weight on the bearing during oscillations, the bearing collar may slowly slip against the axle box wall. Although our calculations show that abrasive wear due to this slippage is negligible, the calculation raises general principles that apply to other possible wear mechanisms. If lifetime is proportional to hardness, we can estimate relative lifetimes of refurbished and new boxes. Although the resleeve material is softer than the original, the cost to lifetime ratio would favour refurbishment under this assumption.

Important unanswered questions are identified and a specific integrated program of field, laboratory, and theoretical study is suggested.P. BroadbridgeG. Wake2012-02-03T20:56:45Z2015-05-29T20:10:57Zhttp://miis.maths.ox.ac.uk/miis/id/eprint/560This item is in the repository with the URL: http://miis.maths.ox.ac.uk/miis/id/eprint/5602012-02-03T20:56:45ZEfficient design of tall tapered feedersToowoomba Foundry is seeking help with the design of tall tapered feeders, which supply extra molten iron to poured castings as they cool. There is a problem with the reliability of the feeders, particularly for thicker castings of the newer spheroidal graphite irons. An effective feeder remains molten until the casting has set. The setting of the melt in the feeder is delayed by making it large enough to retain its heat for longer than the casting, and by placing it close to the thermal centre of the casting. So a larger feeder is more reliable. But any metal remaining in the feeder is recycled, and has an associated energy cost. If a feeder is too small, it will set too soon, and the casting will have unwanted holes in it that may require the entire casting to be recycled. Thus there is a tension between making the feeder smaller so as to minimise recycled metal, and having the feeder large enough that the casting is good.

Existing design methods use purely conductive models of heat transport. We investigate the relevance of convection in the cooling feeder, and set up a boundary-layer model of flow driven by density differences. We find that convection is a significant factor in the design of a feeder, effectively maintaining constant temperature across it. The height of the feeder is important mainly in providing the driving force for this flow.M.J. McGuinnessA.J. Roberts