QUALITATIVE AND QUANTITATIVE ANALYSIS OF STOCHASTIC PROCESSES BASED ON MEASURED DATA, II: APPLICATIONS TO EXPERIMENTAL DATA

Analysis of stochastic processes governed by the Langevin equation is discussed. The analysis is based on a general method for non-parametric estimation of deterministic and random terms of the Langevin equation directly from given data. Separate estimation of the terms corresponds to the decomposition of process dynamics into deterministic and random components. Part I of the paper presented several possibilities for qualitative and quantitative analysis of process dynamics based on such decomposition. In Part II, some of these analysis possibilities are applied to experimental datasets from metal cutting and laser-beam welding.     

Needle Variations and Almost Lower Semicontinuous Differential Inclusions

It is shown that the construction of needle variations can be carried out for almost lower semicontinuous differential inclusions rather than for the case of ordinary single-valued continuously differentiable vector fields usually considered in the literature. The construction leads to needle variations whose flows are in general set-valued but still have good differentiability properties. The variations are constructed by using single-valued selections that are not necessarily continuous with respect to the state variable, but have instead a much weaker 'integral continuity' property, somewhat more general that the 'directional continuity' considered in previous work by A. Cambini and S. Querci, A. Pucci, and A. Bressan. The existence of many such selections is proved by slightly adapting an argument due to Bressan, extending it from the lower semicontinuous to the almost lower semicontinuous case, and strengthening it to yield not only directional continuity at all points but also full continuity at a specified point.     

Galactic model of element formation

A cosmological model is presented that produces He, C, O, D, Li, Be, and B in their observed abundances without a Big Bang. The elements are produced during the 1.5-Gy formation period of a galaxy, with C, O, and other heavy elements produced by stars of M>12 M₃ , He by stars of 6 M_s<M<12 M_s, and D, Li, Be, and B by cosmic-ray reactions in a nearly pure H plasma halo. Thus, the model shows that the major elements used in astrophysical studies can be produced during galactic formation by exactly the same processes known to function today at much lower rates     

Adaptive local grid refinement algorithms for finite-element collocation

An adaptive grid refinement procedure allows accurate solutions to advection-dominated, time-dependent flows using finite-element collocation. The technique relies on a data structure that is readily amenable to parallel computing. The paper discusses computational aspects of the method.     

On the excitation of a closely spaced array by a line source

An infinite row of periodically spaced, identical rigid circularcylinders is excited by an acoustic line source which is parallelto the generators of the cylinders. A method for calculatingthe scattered field accurately and efficiently is presented.When the cylinders are sufficiently close together, Rayleigh–Blochsurface waves that propagate energy to infinity along the arrayare excited. An expression is derived which enables the amplitudesof these surface waves to be computed without requiring thesolution to the full scattering problem.     

Self-broadening in the millimeter-wave spectrum of ozone

Self broadened widths of seven lines of ozone between 75 and 120 GHz have been measured at room temperature. The variation with temperature of the self-broadening parameter for three of these lines has been studied in the range 245–292°K.     

Reductions in overshoot following intense sound exposures

Overshoot refers to the poorer detectability of brief signals presented soon after the onset of a masking noise compared to those presented after longer delays. In the present experiment, brief tonal signals were presented 2 or 190 ms following the onset of a broadband masker that was 200 ms in duration. These two conditions of signal delay were tested before and after a series of exposures to a tone intense enough to induce temporary threshold shift (TTS). The magnitude of the overshoot was reduced after the exposure when a TTS of at least 10 dB was induced, but not when smaller amounts of TTS were induced. The reduction in overshoot was due to a decrease in the masked thresholds with the 2-ms delay; masked thresholds with the 190-ms delay were not different pre- and post-exposure. The implication is that the mechanisms responsible for the normal overshoot effect are temporarily inactivated by the same stimulus manipulations that produce a mild exposure-induced hearing loss. Thus the result is the paradox that exposure to intense sounds can produce a loss of signal detectability in certain stimulus conditions and a simultaneous improvement in detectability in other stimulus conditions.