Aerosol lidar intercomparison in the framework of the EARLINET project. 1. Instruments

In the framework of the European Aerosol Research Lidar Network to Establish an Aerosol Climatology (EARLINET), 19 aerosol lidar systems from 11 European countries were compared. Aerosol extinction or backscatter coefficient profiles were measured by at least two systems for each comparison. Aerosol extinction coefficients were derived from Raman lidar measurements in the UV (351 or 355 nm), and aerosol backscatter profiles were calculated from pure elastic backscatter measurements at 351 or 355, 532, or 1064 nm. The results were compared for height ranges with high and low aerosol content. Some systems were additionally compared with sunphotometers and starphotometers. Predefined maximum deviations were used for quality control of the results. Lidar systems with results outside those limits could not meet the quality assurance criterion. The algorithms for deriving aerosol backscatter profiles from elastic lidar measurements were tested separately, and the results are described in Part 2 of this series of papers [Appl. Opt. 43, 977-989 (2004)]. In the end, all systems were quality assured, although some had to be modified to improve their performance. Typical deviations between aerosol backscatter profiles were 10% in the planetary boundary layer and 0.1 x 10(-6) m(-1) sr(-1) in the free troposphere.     

From data topology to a modular classifier

This article describes an approach to designing a distributed and modular neural classifier. This approach introduces a new hierarchical clustering that enables one to determine reliable regions in the representation space by exploiting supervised information. A multilayer perceptron is then associated with each of these detected clusters and charged with recognizing elements of the associated cluster while rejecting all others. The obtained global classifier is comprised of a set of cooperating neural networks and completed by a K-nearest neighbor classifier charged with treating elements rejected by all the neural networks. Experimental results for the handwritten digit recognition problem and comparison with neural and statistical nonmodular classifiers are given.Received: 1 October 2002, Accepted: 21 November 2002, Published online: 6 June 2003     

A hybrid method for detecting the onset of local necking by monitoring the bulge forming load

Strain-based Forming Limit Diagrams (FLD), which are typically obtained under linear or quasi-linear loading conditions, describe the limiting strains in terms of the major and minor in-plane strains before the onset of necking or the final failure (FFD). These strains can be detected by analysing the strain field in the vicinity of necking or cracking defects. It has generally been agreed that the loading versus time signal is not suitable for detecting necking processes. A novel hybrid method of detecting the onset of necking based on the experimental and simulated bulging load is presented in this paper. This method consists mainly in comparing the experimental forming load, i.e., a load showing plastic instability, with the numerical predictions obtained by performing finite element simulation. The simulation of the bulging process does not include any damage or failure criteria. A homogeneous forming load can therefore be simulated without requiring any information about the localization. This method was applied to detecting the onset of local necking in circular and elliptic quasistatic bulge tests on sheet material, with a diameter of 200 mm. Two materials were tested, a 0.8 mm thick DP450 Dual Phase steel sheet and a 1 mm thick AA6016-T4 aluminium sheet. The onset of necking observed with our method was compared with the results obtained by performing Hogström’s analysis based on the measured strain field over time and similar necking strains were obtained. Beside, the Bressian Williams Hill (BWH) shear criterion was identified for each test from experimental results. A slight scattering of the shear stress values was observed.     

Analysis of photocurrents at the semiconductor—eletrolyte junction

The photocurrent vs potential characateristics for three different electrolyte—semiconductor junctions, representative of those generally found in semiconductor photoelectrochemistry, are analyzed in detail using parameters which define both the semiconductor and the electrolyte. It is shown that, in general, the behaviour of junctions including semiconductors with sufficiently wide energy pags and large free carriers densities, may be accurately described using the Gärtner model in the potential region which does not include the onset of the photocurrent. In this case, it is the characteristics of the semiconductor which control the photoresponse and the electrolyte does not induce limiting steps in the charge transfer across the interface. If certain restrictive conditions are fulfilled concerning the relative orders of magnitude of the semiconductor space charge region, diffusion length of minority carriers and penetration depth of light into the semiconductor, the value of the flat band potential may be easily determined by ploting i²_phvs V and extrapolating to i²_ph = 0. In the other cases, the relation established by Gärtner between i_ph and V has been verified provided that the free carriers density be sufficiently large.     

Linearly bounded infinite graphs

Linearly bounded Turing machines have been mainly studied as acceptors for context-sensitive languages. We define a natural class of infinite automata representing their observable computational behavior, called linearly bounded graphs. These automata naturally accept the same languages as the linearly bounded machines defining them. We present some of their structural properties as well as alternative characterizations in terms of rewriting systems and context-sensitive transductions. Finally, we compare these graphs to rational graphs, which are another class of automata accepting the context-sensitive languages, and prove that in the bounded-degree case, rational graphs are a strict sub-class of linearly bounded graphs.A preliminary version of this article appeared in MFCS 2005.     

Laser sintering of coated polyamide 12: a new way to improve flammability

Journal of Materials Science - In this study coatings of kaolin and talc particles were successfully applied on the surface of polyamide 12 powder intended for laser sintering (LS). Microscopic...     

Learning hierarchical probabilistic logic programs

Machine Learning - Probabilistic logic programming (PLP) combines logic programs and probabilities. Due to its expressiveness and simplicity, it has been considered as a powerful tool for learning...