Statistical mechanics of mutual information maximization

An unsupervised learning procedure based on maximizing the mutual information between the outputs of two networks receiving different but statistically dependent inputs is analyzed (Becker S. and Hinton G., Nature, 355 (1992) 161). By exploiting a formal analogy to supervised learning in parity machines, the theory of zero-temperature Gibbs learning for the unsupervised procedure is presented for the case that the networks are perceptrons and for the case of fully connected committees.

The boundary lubricated friction and wear of low alloy steel

Pin on disc wear machines were used to study the boundary lubricated friction and wear of AISI 52100 steel sliding partners. Boundary conditions were obtained by using speed and load combinations which resulted in friction coefficients in excess of 0.1. Lubrication was achieved using zero, 15 and 1000 ppm concentrations of an organic dimeric acid additive in a hydrocarbon base stock. Experiments were performed for sliding speeds of 0.2, 0.35 and 0.5 m/s for a range of loads up to 220 N. Wear rate, frictional force and pin temperature were continually monitored throughout tests and where possible complementary methods of measurement were used to improve accuracy. A number of analytical techniques were used to examine wear surfaces, debris and lubricants, namely: Scanning Electron Microscopy (SEM), Auger Electron Spectroscopy (AES), Powder X-ray Diffraction (XRD), X-ray Photoelectron Spectroscopy (XPS), optical microscopy, Back scattered Electron Detection (BSED) and several metallographic techniques. Friction forces and wear rates were found to vary linearly with load for any given combination of speed and additive concentration. The additive itself was found to act as a surface oxidation inhibitor and as a lubricity enhancer, particularly in the case of the higher (1000 ppm) concentration. Wear was found to be due to a mild oxidational mechanism at low additive concentrations and a more severe metallic mechanism at higher concentrations with evidence of metallic delamination in the latter case. Scuffing loads were found to increase with increasing additive concentration and decrease with increasing speed as would be predicted by classical models of additive behaviour as an organo-metallic soap film. Heat flow considerations tended to suggest that surface temperature was not the overriding controlling factor in oxidational wear and a model is proposed which suggests oxygen concentration in the lubricant is the controlling factor in oxide growth and wear.

Automobile friction materials: a test instrument and technique for determining dynamic mechanical properties in relation to brake squeal

This study is primarily concerned with the problem of break-squeal in disc brakes, using moulded organic disc pads. Moulded organic friction materials are complex composites and due to this complexity it was thought that they are unlikely to be of uniform composition. Variation in composition would under certain conditions of the braking system, cause slight changes in its vibrational characteristics thus causing resonance in the high audio-frequency range. Dynamic mechanical propertes appear the most likely parameters to be related to a given composition's tendency to promote squeal. Since it was necessary to test under service conditions a review was made of all the available commercial test instruments but as none were suitable it was necessary to design and develop a new instrument. The final instrument design, based on longitudinal resonance, enabled modulus and damping to be determined over a wide range of temperatures and frequencies. This apparatus has commercial value since it is not restricted to friction material testing. Both used and unused pads were tested and although the cause of brake squeal was not definitely established, the results enabled formulation of a tentative theory of the possible conditions for brake squeal. The presence of a temperature of minimum damping was indicated which may be of use to braking design engineers. Some auxilIary testing was also performed to establish the effect of water, oil and brake fluid and also to determine the effect of the various components of friction materials.