Analysis of multi-exponential relaxation data with very short components using linear regularization

Linear regularization is a common and robust technique for fitting multi-exponential relaxation decay data to obtain a distribution of relaxation times. The regularization algorithms employed by the Uniform-Penalty inversion (UPEN) and CONTIN computer programs have been compared using simulated transverse (T2) relaxation data derived from a typical bimodal distribution observed in cartilage tissue which contain a component shorter than t(0), the time of the first decay sample. We examined the reliability of detecting sub-t(0) relaxation components and the accuracy of statistical estimates of T2 distribution parameters. When the integrated area of the sub-t(0) component relative to that of the total distribution was greater than 0.25, our results indicated a signal-to-noise threshold of about 300 for detecting the presence of the sub-t(0) component with a probability of 0.9 or greater. This threshold was obtained using both the UPEN and CONTIN algorithms. In addition, when using the second-derivative-squared regularizer, UPEN solutions provided statistical estimates of T2 distribution parameters which were substantially free of the biasing effect of the regularizer observed in analagous CONTIN solutions.     

斜流/离心压缩机初步设计和分析方法研究

依托两区模型开展了单级斜流/离心压缩机的初步设计和分析方法研究,验证了基于SOA-TEIS混合扩压度模型的改进ODP问题迭代方法在斜流/离心压缩机初步设计和分析中的准确性。采用本文方法对Krain高比转速叶轮和低比转速叶轮模型进行了一维设计和分析,并将结果与实验数据进行了对比,结果表明:本文所述方法可在减少对工程经验参数选取依赖性的同时,实现较准确的斜流/离心压缩机的初步设计计算和非设计工况性能预测,且易于程序化实现,采用被动滑移因子计算方法获得的设计点附近工况的等熵效率预测偏差在1%以内。     

基于声发射和GAN-CNN的铝合金管道法兰 连接松动泄漏检测*

面向管道法兰连接松动引起的泄漏检测需求,为解决数据样本不足和减少特征指标手动选取的繁琐环节。本文,考虑到生成性对抗网络（GAN）作为数据扩充工具,已被证明能够生成与真实数据相似的样本数据。同时,卷积神经网络（CNN）作为一种深度学习方法,为自动提取数据的特征提供了一种有效的方法。开展了基于GAN和CNN的铝合金管道法兰连接松动泄漏检测研究。首先,搭建管道泄漏标定和数据采集实验台,利用声发射技术获取不同等级的原始泄漏信号。其次,采用GAN生成样本数据扩充原始数据。同时,为了评估生成模型的性能,引入统计特评估生成质量。最后,将生成的样本数据与原始数据设置为不同训练集,基于卷积神经网络构建智能分类检测模型,应用于管道泄漏检测。同时,分类检测结果与小样本智能分类方法SVM进行了比较,实验结果表明,基于GAN和CNN构建的智能分类模型可显著提高管道法兰连接松动泄漏检测精度。     

Determination of Particle Size Distribution in Fluids using photon correlation spectroscopy

Photon correlation spectroscopy (PCS) has been applied to various systems of particles suspended in a fluid. Computer simulations were used to analyse CONTIN, a program for the numerical inversion of the experimental data. Measurements were performed on a spectrometer equipped with a digital correlator. Various suspensions of polysterene latex spheres differing both in size and size distribution were prepared in double distilled water. Further some samples provided by an external institution without prior specification were analysed. Diffusivity measurements on submicron DES (di-2-ethylhexyl sebacate) particles in nitrogen were also performed. The results obtained demonstrate the strengths and the limitations of PCS as a method of determining particle size distributions in particle fluid systems.     

Particle Size Distributions from Static Light Scattering with Regularized Non‐Negative Least Squares Constraints

Simulated data from static light scattering produced by several particle size distributions (PSD) of spherical particles in dilute solution is analyzed with a regularized non‐negative least squares method (r‐NNLS). Strong fluctuations in broad PSD's obtained from direct application of NNLS are supressed through an averaging procedure, as introduced long ago in the inversion problem in dynamic light scattering. A positive correlation between the best PSD obtained from several averaging schemes and the condition number of the respective data transfer matrices was obtained. The performance of the method is found to be similar to that of constrained regularization (CONTIN), which uses also NNLS as a starting solution, but incorporates another regularizing strategy.     

Energy dissipation prediction of particle dampers

This paper presents initial work on developing models for predicting particle dampers (PDs) behaviour using the Discrete Element Method (DEM). In the DEM approach, individual particles are typically represented as elements with mass and rotational inertia. Contacts between particles and with walls are represented using springs, dampers and sliding friction interfaces. In order to use DEM to predict damper behaviour adequately, it is important to identify representative models of the contact conditions. It is particularly important to get the appropriate trade-off between accuracy and computational efficiency as PDs have so many individual elements. In order to understand appropriate models, experimental work was carried out to understand interactions between the typically small (～1.5–3 mm diameter) particles used. Measurements were made of coefficient of restitution and interface friction. These were used to give an indication of the level of uncertainty that the simplest (linear) models might assume. These data were used to predict energy dissipation in a PD via a DEM simulation. The results were compared with that of an experiment.     

Determination of Particle Mass Density Distribution

A new method for the determination of two characteristics of a sample of particles, the correlation of particle mass density with particle size and the distribution of particle mass density, is presented. The new method to meet these requirements is based on the combination of two optical particle sizing techniques where the measured particle characteristics have a different dependence on particle mass density. Photosedimentation and laser diffraction were chosen as suitable techniques for this goal. The determination of particle mass density by photosedimentation and laser diffraction is based on the fact that for each particle size class of the sample to be analysed the mass density can be calculated by an application of the Lambert-Beer law. The particle size distribution of the sample has to be known for the determination of particle mass density, and it is measured by laser diffraction. From this, two particle characteristics, the relationship of particle mass density and particle size and the distribution of particle mass density, are obtained. The capacity of the algorithm and its limitations are demonstrated by computational simulations including an error propagation analysis. Experimental results are shown for homogeneous and heterogeneous materials.     

The reconstruction of particle size distributions from dynamic light scattering data using particle swarm optimization techniques with different objective functions

In this paper, the reconstruction of particle size distributions (PSDs) using particle swarm optimization (PSO) techniques from dynamic light scattering (DLS) data was established. Three different objective functions containing non-smooth constrained objective function, smooth functional objective function of Tikhonov regularization and L objective function, were employed. Simulated results of unimodal, bimodal and bi-dispersed particles show that the PSO technique with non-smooth constrained objective function produces narrower PSDs focusing on peak position in the presence of random noise, the PSO technique with smooth functional of Tikhonov regularization creates relative smooth PSDs, which could be successfully applied to the broad particles inversion, and the PSO technique with L objective function yields smooth PSDs, which saves calculation amount. Experimental results as well as comparisons with CONTIN algorithm and Cumulants method demonstrate the performance of our algorithms. Therefore, the PSO techniques employing the three different objective functions, which only require objective function and need a few initial guesses, may be applied to the reconstruction of PSDs from DLS data.     

Testing a new photocentrifuge for submicron particles with silica spheres

A theory has been developed for spectral photosedimentation in a centrifuge. The use of several wavelengths allows particle size analyses in the submicron range without a knowledge of the optical material properties and is not restricted to spherical particles. As a result of this theory the extinction efficiency and the particle size distribution of the analysed particles can be calculated even for non-spherical particles from an experiment. A photocentrifuge has been built according to the developed theory. The theory has been verified by the measurement of spherical particles for which Mie theory is applicable. The experimentally obtained extinction efficiency agrees well with computations according to Mie theory. The accuracy of measured particle size distributions is comparable to those measured with other sedimentometers.     

Characterization of the Composition of Solid Mixtures Using Particle Shape Distributions Obtained by Image Analysis

There are many applications of particle technology in which solid mixtures consisting of particles having the same chemical, mineralogical and physical properties are used. The particles are then distinguishable only by shape. The particle shape significantly influences the technological behaviour of the mixture and the properties of the materials produced. Therefore, mathematical procedures are necessary in order to determine the composition of such a mixture according to the proportions of certain particle shape classes. These classes are the respective components of the mixture. In this work, various statistical methods were tested to classify the particles by shape analysis. A two-component mixture of quartz and muscovite was used as a reference material. The results obtained by different theoretical methods were compared with each other. Alternative methods such as histogram fit and the EM algorithm provided better results than the conventional method of discriminant analysis, as expected. Because the statistical behaviour of the components can be estimated only using reference samples, a statistical error of about 5–10% depending on the respective sample size was met. However, this is a satisfactory result. It turned out that it is possible to conclude exclusively on the basis of the particle shape distribution about the composition of a mixture. The methodical knowledge obtained can be used in industrial applications such as the ceramic industry, the production of paints and the bulk solids technology.     

颗粒系的可见消光光谱分析及最佳波长的选择

在光全散射法颗粒粒径测量中,被测颗粒系的消光光谱包含有颗粒粒径、折射率等信息。对常用的单峰及双峰R-R分布的可见消光光谱进行模拟,分析了可见消光光谱随粒径和相对折射率变化的规律。选取二阶微分可见消光光谱不连续点对应的波长作为测量波长,并将可见光边界波长同时作为测量波长,在非独立模式下,采用遗传优化算法反演粒径分布。仿真计算与实验验证结果表明,通过消光光谱分析,可以预先确定被测颗粒系的分布状况,缩小优化算法中反演参数的搜索范围。结合最佳波长选择方法,使测量的精度和可靠性明显提高。对测量光谱加入1%随机噪声时,单峰及多峰分布粒径反演均能得到满意的结果。     

An investigation of the thermal and catalytic behaviour of potassium in biomass combustion

Potassium, a key nutrient in biomass growth, contributes to problematic ash chemistry and corrosion in combustion. This study seeks to examine the behaviour and fate of potassium in biomass combustion under high temperature flame conditions. A model to predict potassium release is presented. Short rotation willow coppice was treated to reduce metals, by water-washing, and remove them, by demineralisation, and then potassium was doped into the demineralised sample. The resultant fuels have been studied for their combustion behaviours in methane–air flames, both as suspended, moving particles, and as stationary, supported particles, using high speed digital video. In the latter case, potassium release was measured simultaneously by emission spectroscopy. In both experiments, potassium was seen to catalyse devolatilisation, and for the stationary particles it was possible to detect potassium catalysis in the char burn-out rates. Demineralised willow was seen to melt in the flame and combustion resembled heavy oil combustion, rather than solid fuel combustion. The residual char was extremely slow to burn-out. In the potassium-doped particles, potassium was seen to evolve over three regimes, devolatilisation, char burn-out and, less significantly, during ash cooking. The first two evolution processes have been modelled using an apparent first order devolatilisation rate for the first stage, and a KOH evaporation model for the second stage.     

Fly Ash Size Distributions: Use of coulter multisizer and fitting to truncated lognormal distributions

Accurate information on the size distribution of fly ash is needed to determine its role in the radiation transfer process in pulverized coal combustors. The Coulter Multisizer was used to determine the size distribution in the particle diameter range 1–200 μm. To size over such large diameter ranges, data must be obtained using several orifices, and then combined. In order to use the smaller orifices, the larger particles have to be removed from the sample. A wet-sieving apparatus, designed for accurate separation of the particles by size, is described. A scheme for combining data obtained using orifices of different diameters is presented. It appears from this study that the lower limit of size measurement using an orifice is set by sensitivity, rather than by signal/noise. A lognormal distribution function, truncated outside the measurement limits, fits the size distribution data well. This function allows detailed size information to be stored compactly using four parameters. Size parameters for six fly ashes representative of U. S. coals are presented.     

Experimental gas amplification study in boron-lined proportional counters for neutron detection

Based on theoretical considerations related to the microscopic behaviour of electrons in gases in conjunction with a macroscopic hypothesis, a semi-microscopic gas gain formula for conventional proportional counters has been already developed and was afterwards successfully tested on gas gain data obtained in Ar–CH₄- and He–isobutane-filled proportional counters, using an X-ray radiation source. In the present work, it is verified that the semi-microscopic formula works as well in case of the detection of spread energy-spectrum of ionizing charged particles, such as involved within boron lined proportional counters, commonly used for thermal neutron detection. Indeed, we investigate experimentally the validity of the formula for the parameterization of the gas amplification factor data obtained in two boron lined proportional counters, filled with Ar–CH₄ (10%) and Ar–CO₂ (5%). The obtained results show clearly the ability of the gas gain formula to perfectly describe gas gain characteristics in such counters independently of the filling gas mixture used. Consequently, this provides further evidence of generality of the underlying gas amplification semi-microscopic model.     

Structuring from nanoparticles in oil-based ferrofluids

The effect of magnetic field on the structure formation in an oil-based magnetic fluid with various concentrations of magnetite particles was studied. The evaluation of the experimental data obtained from small-angle X-ray scattering and ultrasonic attenuation indicates the formation of chain-like aggregates composed of magnetite particles. The experimental data obtained from ultrasonic spectroscopy fit well with the recent theoretical model by Shliomis, Mond and Morozov but only for a diluted magnetic fluid. In this model it is assumed that a dimer is the main building block of a B -field-induced chain-like structure, thus the estimation of the nematic order parameter does not depend on the actual length of the structure. The scattering method used reveals information about the aggregated structure size and relative changes in the degree of anisotropy in qualitative terms. The coupling constant l \lambda , concentrations f \phi , average particle size d and its polydispersity s \sigma were initially obtained using the vibrating sample magnetometry and these results were further confirmed by rheometry and scattering methods. Both the particles’ orientational distribution and the nematic order parameter S were inferred from the ultrasonic measurements. The investigation of SAXS patterns reveals the orientation and sizes of aggregated structures under application of different magnetic-field strengths. In addition, the magnetic-field-dependent yield stress was measured, and a relationship between the yield stress and magnetic-field strength up to 0.5T was established.     

Time-resolved fluorescence polarization anisotropy of multimodal samples: the asphaltene case

In this work we describe an application of the time-resolved fluorescence polarization anisotropy (TRFPA) technique to the analysis of asphaltenes, complex mixtures of high-molecular weight compounds, typically present in petroleum oils. Our asphaltene samples consist of nanometer-sized polydispersed particles, whose lighter fraction showed a relatively high fluorescence quantum yield. Most of the fluorescence intensity observed from the complex sample originated from a well defined sample fraction presenting a large fluorescence yield. Consequently, the TRFPA analysis only provided the average size of more fluorescing particles, that, in our case, were the smaller ones. Larger and less fluorescing aggregates did not significantly contribute to the TRFPA signal. Hence, to overcome intrinsic limitations of the TRFPA technique in characterizing multimodal samples, we preliminarily fractionated our complex samples by means of size exclusion chromatography (SEC), thus obtaining nearly monomodal fractions of the original samples. This procedure allowed to estimate also the size of less fluorescing and larger particles. A comparison of particle size estimate by means of TRFPA and SEC methods was also used to acquire information about occurrence of aggregation phenomena, and about the kind and strength of the chemical bonds linking chromophores to each other or to their parent particle. PACS 82.53.Uv; 82.80.Bg; 47.57.E-     

Reduced-order modeling for mistuned centrifugal impellers with crack damages

An efficient method for nonlinear vibration analysis of mistuned centrifugal impellers with crack damages is presented. The main objective is to investigate the effects of mistuning and cracks on the vibration features of centrifugal impellers and to explore effective techniques for crack detection. Firstly, in order to reduce the input information needed for component mode synthesis (CMS), the whole model of an impeller is obtained by rotation transformation based on the finite element model of a sector model. Then, a hybrid-interface method of CMS is employed to generate a reduced-order model (ROM) for the cracked impeller. The degrees of freedom on the crack surfaces are retained in the ROM to simulate the crack breathing effects. A novel approach for computing the inversion of large sparse matrix is proposed to save memory space during model order reduction by partitioning the matrix into many smaller blocks. Moreover, to investigate the effects of mistuning and cracks on the resonant frequencies, the bilinear frequency approximation is used to estimate the resonant frequencies of the mistuned impeller with a crack. Additionally, statistical analysis is performed using the Monte Carlo simulation to study the statistical characteristics of the resonant frequencies versus crack length at different mistuning levels. The results show that the most significant effect of mistuning and cracks on the vibration response is the shift and split of the two resonant frequencies with the same nodal diameters. Finally, potential quantitative indicators for detection of crack of centrifugal impellers are discussed.     

最小角回归结合竞争性自适应重加权采样的近红外光谱波长选择

近红外光谱分析技术对检测样品无损伤且检测速度快、精度高,因此被广泛应用在了药品检测、石油化工等领域,尤其近年来机器学习和深度学习建模方法的深入应用使其具备了更准确的检测性能。然而,样品的近红外光谱数据具有比较高的维度且存在谱间重合、共线性和噪声等问题,对近红外光谱模型的性能产生消极影响,此时样品有效特征波长的筛选极为重要。为了提高近红外光谱定量和定性分析模型的准确性和可靠性,提出了一种近红外光谱变量选择方法,其结合了最小角回归（LAR）和竞争性自适应重加权采样（CARS）的优点,具有更优的性能。该方法利用LAR初步筛选样品全谱区的特征波长,接着利用CARS对筛选出来的特征波长进一步选择,从而有效去除无关特征波长。为验证该方法的有效性,从定量和定性分析两个方面评价该方法。在定量分析实验中,以FULL,LAR,CARS,SPA和UVE作为对比方法,以药品样品数据集为实例建立PLS回归分析模型,经LAR-CARS筛选出的变量建立的PLS模型在药品数据集表现出较高的预测决定系数和较低的预测标准偏差。在定性分析实验中,以SVM,ELM,SWELM和BP作为对比方法、不同比例训练集的药品数据集为实例建立分类模型,经LAR-CARS筛选出的变量建立的SVM分类模型精度最高达100%。从实验结果可见,LAR-CARS可有效的筛选出表征样品特征的波长,利用其筛选出的波长建立的定量、定性分析模型具有更好的鲁棒性,可用于样品光谱的特征波长筛选。     

Fatigue in precipitation hardened materials: a three-dimensional discrete dislocation dynamics modelling of the early cycles

Three-dimensional dislocation dynamics (DD) fatigue simulations in precipitation hardened metals is a major challenge in terms of numerical development. Several precipitate configurations have been investigated with an original treatment of precipitate–dislocation interactions and a parallelized DD code. In grains containing single-sized shearable particles (r _p?=?160?nm), strain is localized in clear bands where the precipitates are totally sheared-off. The fatigue behaviour involves an initial hardening followed by severe cyclic softening and significant surface slip irreversibility. In the presence of large single-sized particles (r _p?=?400?nm), the persistent slip band (PSB) structure is accompanied by highly reversible surface displacements. Slip dispersion originates from Orowan loops that have little effect on the mechanical response. The mechanical behaviour of a bimodal distribution is intermediate between the two above cases with the above microstructural features coexisting within the same grain. Unlike in the monomodal large-particle case, where all the particles retain their initial strength, some of the large particles of the bimodal distribution undergo a significant strength reduction.     

扫描近场光学显微镜的光耦合偶极子模型

为了研究扫描近场光学显微镜中探针和粗糙样品表面的耦合相互作用,提出了一种光耦合偶极子模型.在该模型中,探针和样品突起都由光极化偶极子表示,在准静态电磁场近似的情况下样品表面的诱导极化效应由影像偶极子表示,应用偶极子辐射理论可以得到系统的自洽场方程.此模型提供了一种直观分析扫描近场光学显微镜中探针和样品相互作用机理的方法.在此基础上,进一步讨论了金属样品的近场成像特点和其特有的局域光学共振现象.数值结果表明：不同于一般的介质样品,金属样品的近场图像与入射光频率直接相关,改变入射光的频率,获得的样品近场图像的形状和对比度都会发生变化.特别是当入射光频率处于样品极化共振范围内时,金属纳米粒子的极化率会出现光极化共振,这样就可以获得样品粒子的最大有效尺寸,为提高系统的分辨率提供了一条重要途径.