A comprehensive study of the long pulse Nd:YAG laser drilling of multi-layer carbon fibre composites

The results of an extensive experimental study of the free running Nd:YAG laser drilling of a multi-layer carbon fibre composite, where adjacent layers have differently orientated fibres, are reported. For holes drilled with the laser operating in fixed-Q mode at 1064 nm, parallel sections of blind holes illustrating discontinuities in the hole size along a given section direction will be shown to occur at the interface between adjacent layers. An explanation for this effect is proposed. Detailed single pulse drilling characteristics will be presented illustrating the exit hole diameter as a function of pulse energy and material thickness. These characteristics illustrate a ‘stable' drilling regime in which the exit hole diameters are least sensitive to changes in pulse energy or material thickness and a less ‘stable' regime in which they are more strongly dependent on these parameters. Drilling characteristics will be given for two different beam qualities, illustrating the greater drilling depth and reduced hole size achievable with an improved beam quality. Finally holes drilled through a 2 mm thick sample of material with multiple pulses are considered. Size distribution curves for entrance and exit holes will be presented. The total energy required (number of pulses × pulse energy) to drill through 2 mm thick material will be reported as a function of pulse energy in stationary air and argon atmospheres and in a partial vacuum, illustrating a threshold energy which is dependent upon the drilling atmosphere. The threshold energies will be discussed with reference to plasma formation and the reactivity of the drilling atmosphere.     

Epoxy based coatings on glass: strengthening mechanisms

Glass may be strengthened by epoxy coatings although the strengthening mechanisms remain unclear. Possible strengthening mechanisms are reviewed and are used to analyse strength data for both a solvent based and a water based coating system. The coatings either fill (solvent based coatings), or partially fill (water based coatings) surface cracks and it is shown that closure stresses arising from the thermal expansion mismatch of the coating within these cracks can account for the observed degrees of strengthening. It is also demonstrated that other suggested mechanisms such as flaw healing cannot fully account for the observed degree of strengthening.     

Synthesis and NMR-Spectroscopic Structure Determination of Novel 7,7′-Diphenyl-7,7′-diapocarotenoids

Wittig reaction of crocetindial ( 1 ) and benzylidenetripenylphosphorane ( 2 ) gave (7E, 7′Z)-7,7′-diphenyl-7-7′-diapocarotene ( 3 ), instead of the previously reported (7E, 7′E)-isomer. Similar reaction of 8,9-didehydrocrocetindial ( 4 ) with 2 yielded the three acetylenic isomers 5a–c which differ in the configuration of the terminal double bonds. Structures were established by 1D- and 2D-NMR studies. Illustrative spectra and their interpretation are presented. Most chemical shifts of corresponding protons in 3 and 5 and nearly identical, but ¹³C shifts differ considerably.     

Structural studies of iron in vitrified toxic wastes

The formation of iron carbides by reactive milling of α-Fe and C powders is reported. The products formed were analyzed by Mössbauer spectroscopy and X-ray diffraction. It was found that iron carbide phases start forming after an incubation period of about 3 h depending on the ball-to-powder weight ratio (BPR). Carbide amounts increased with increasing milling time while α-Fe content decreased. Energy transfer increased with increasing BPR and high BPR resulted in an increase in the reaction rate. Although it was not possible to selectively synthesise a specific Fe _x C phase, samples containing predominantly one type of carbide phase, either Hägg carbide or cementite, were successfully prepared. The formation of the different iron carbide phases is discussed within the context of the Fe–C phase diagram for non-equilibrium processes.     

Where are the large and difficult datasets?

A great many comparative performance assessments of classification rules have been undertaken, ranging from small ones involving just one or two methods, to large ones involving many tens of methods. We are undertaking a meta-analytic study of these studies, attempting to distil some overall conclusions. This paper describes just one of our observations. The dataset analysed in this paper contains 5,203 error rates taken from 45 articles and describing 146 datasets. One curious general relationship which was persistent in our data, despite the fact that we were looking at results mixed between distributions rather than conditional on distributions, was that error rate decreased with increasing dataset size. We believe this to be an artefact of the way datasets are collected by the research community.     

Likelihood-ratio changepoint features for consumer-behaviour models

Some predictive models for customer value management might benefit from information about certain changes in individual-consumer behaviour. We take changepoint methods as the first step in producing a model-input feature for this purpose. An unusual feature in the application of changepoint methods to consumer data is there are as many streams of data as there are customers. This property is used to help decide whether an individual has changed their behaviour by ordering likelihood-ratio statistics from the changepoint models. Following a review of changepoint methods, the approach is demonstrated on cash machine transactions. Models for the amount, location and time of transaction are used and accounts exhibiting large evidence of change are examined in detail. For the data set used the approach performs sensibly. The worth of likelihood-ratio statistics to rank evidence for change is considered more generally through some of the literature.     

Handling selection bias when choosing actions in retail credit applications

In many situations one needs to know which action one should take with a customer to yield the greatest response. Typically, estimates of the response functions of different actions will be based on the responses of customers previously assigned to each action. Often, however, the previous assignments will not have been random, so that estimates of the response functions will be biased. We examine the case of two possible actions. We look at the error arising from using the simple OLS estimate ignoring the selection bias, and also explore the possibility of using the Heckman model to allow for the sample selectivity. The performance of Heckman’s model is then compared with the simple OLS through simulation.