On-line Particle Size Analysis Research Articles

Online particle size analysis on conveyor belts with dense convolutional neural networks

Particle size distributions in ore feed systems, as well as the identification of particles in these feed systems can provide important information in the advanced control of unit operations in mineral processing, such as crushing and grinding circuits. Image analysis has long been considered a promising approach to achieve this, as it is an inexpensive, unobtrusive means of acquiring information rich measurements. It typically requires segmentation of images in order to identify individual particles. This is a challenging task to accomplish reliably, as variable lighting, fines adhering to larger particles or contiguous particles, as well as variable particle sizes and shapes can all compromise the accuracy of traditional algorithms. Image segmentation with deep learning methods have recently been investigated to surmount these difficulties. In this investigation, U-net and U-net with superpixel preprocessing with simple linear iterative clustering (SLIC) are proposed and compared with a traditional watershed algorithm. The U-net approaches were markedly more reliable than the watershed algorithm. In addition, preprocessing of the images with SLIC resulted in further improvement of the results.

Dry process for SO2 and NOx removal via gas-to-particle conversion with ozone and ammonia injection

Because of the stability of its production and cost-effectiveness, coal is expected to be a primary energy source in near future. However, SO2 and NOx emissions from the coal combustion have been considered a serious environmental issue in the last few decades; they react with water vapor and ammonia (NH3) in the atmosphere to form secondary aerosols. In this study, ozone (O3) was injected to oxidize NO into NO2. Thereafter, NH3 was injected to convert SO2 into (NH4)2SO3 particles, which further reacted with NO2 to produce N2 and (NH4)2SO4 particles. These reaction mechanisms were discussed employing different chemical analyses of the products. Moreover, the size of the product particles was analyzed by online particle size analysis and the scanning electron microscopy images. Further, the quantitative analysis of the product was performed by ion chromatography, and the atomic balance of sulfur was verified by gas analysis. A higher SO2 concentration increased the NOx removal efficiency because the product ((NH4)2SO3), which was obtained from SO2 removal, acted as a reactant for the removal of NOx. Therefore, a higher concentration of O3 favored the improvement of the NOx removal efficiency since O3 oxidizes NO into NO2, which reacts with (NH4)2SO3. Thus, the SO2 and NOx removal efficiencies were 93.6% and 69.6%, respectively, at SO2, NO, and O3 concentrations of 900, 30, and 30 ppm, respectively.

Beyond the minimum dimension for particle size control

Particle size control is important in efficient mineral processing, but today many milling operations are controlled using techniques which only consider the minimum particle dimension and assume that shapes of the particles remain constant. Many process operations suffer from variation in efficiency which cannot be explained by conventional techniques such as screening. This can be as a result of process changes such as ore source causing a change in particle morphology which is simply not detected when only the minimum particle dimension is considered.Image analysis data will be used to demonstrate the key principles behind particle size measurement and the reality of mineral slurry shape distributions. Unfortunately the application of image analysis techniques to on-line analysis of mineral slurries is not a practical proposition for reasons of sampling, statistics and cleanliness, therefore alternative techniques must be used on-line to obtain reliable and process sensitive data.Ultrasonic extinction fulfils many of the requirements for on-line particle size analysis, being able to monitor vast numbers of particles at high concentration and since the technique responds to the area of the particles, a more process sensitive measure may be obtained. The OPUS ultrasonic extinction system uses 31 frequencies to measure a full particle size distribution from approx. 1um to 3mm and for the first time allows for parameters such as distribution width, fine particle levels and solids concentration to be measured reliably on-line in a single unit. Data will be presented to show the potential for better process understanding through utilisation of the full particle size distribution information.

Processing nanoparticles with A4F-SAXS for toxicological studies: Iron oxide in cell-based assays

Nanoparticles are not typically ready-to-use for in vitro cell culture assays. Prior to their use in assays, powder samples containing nanoparticles must be dispersed, de-agglomerated, fractionated by size, and characterized with respect to size and size distribution. For this purpose we report exemplarily on polyphosphate-stabilized iron oxide nanoparticles in aqueous suspension. Fractionation and online particle size analysis was performed in a time-saving procedure lasting 50 min by combining asymmetrical flow field-flow fractionation (A4F) and small-angle X-ray scattering (SAXS). Narrowly distributed nanoparticle fractions with radii of gyration ( R g) from 7 to 21 nm were obtained from polydisperse samples. The A4F-SAXS combination is introduced for the preparation of well-characterized sample fractions originating from a highly polydisperse system as typically found in engineered nanoparticles. A4F-SAXS processed particles are ready-to-use for toxicological studies. The results of preliminary tests of the effects of fractionated iron oxide nanoparticles with a R g of 15 nm on a human colon model cell line are reported.

Simultaneous on-line size and chemical analysis of gas phase and particulate phase of cigarette mainstream smoke

This paper describes the combined set-up of on-line chemical analysis of gas phase by single-photon ionisation/resonance enhanced multiphoton ionisation-time-of-flight mass spectrometry (SPI/REMPI-TOFMS) and on-line particle size analysis by differential electrical mobility particle spectrometry (DMS 500) for the investigation of fresh cigarette mainstream smoke. SPI is well suited for the investigation of a great variety of organic species, whereas REMPI is highly sensitive for aromatic compounds. Gas phase measurements of filtered and unfiltered smoke are possible with the SPI/REMPI-TOFMS in order to determine the influence of the presence of particles on the chemical composition of the gas phase. Initial results are shown for the characterisation and comparison of three pure Virginia tobacco research cigarettes having filter ventilations of 0%, i.e. no filter ventilation, 35% and 70% ventilation. The three cigarette types are smoked under two different smoking regimes, a standard regime using puff parameters equivalent to the conventional International Standard Organisation regime and a more intense smoking regime. For the gas phase, qualitative puff-by-puff resolved yields of three selected compounds (acetaldehyde, phenol and styrene) are shown and compared. For particulate matter, particle number, count median diameter and total surface area are illustrated on a puff-by-puff basis. Yields of the chemicals analysed, puff number and surface area are in good agreement with the intensity of the smoking regime and the dilution of smoke by filter ventilation. However, gaseous compounds are influenced differently, depending whether an absolute particle filter is present or not, i.e. they can be totally removed (phenol), partially removed (styrene) or not affected (acetaldehyde). For particle analysis, the count median diameter decreases from puff to puff and is strongly dependent on the smoking regime and ventilation rate. Thereby, 0% ventilated cigarettes smoked under the intense regime result in the smallest count median diameters of ca. 180 nm, whereas 70% ventilated cigarettes smoked with a standard regime lead to the largest values of up to 280 nm. As particle diameter increases, particle number decreases as a consequence of increasing time for particle coagulation.

Insitec online particle size analysis cuts toner manufacturing costs

Insitec online particle size analysis cuts toner manufacturing costs

Improving the Robustness of Particle Size Analysis by Multivariate Statistical Process Control

AbstractThe robustness of online particle size analysis in wet processes is improved by applying data based modeling methods to the control of the sample preparation and measurement sequence of the particle size analyzer. The aim is to find a more accurate and reliable method of determining the end of the particle size integration period using multivariate statistical process control (MSPC).The studied approach is tested on analyzers installed at two mineral processing plant sites and validated using two validation tests. Research shows that the proposed method works with two very different slurry types. The main advantage of the adapted approach is that there are no adjustable parameters that have to be set by the user.

An improved estimation of size distribution from particle profile measurements

Several methods are available to measure particle size. The majority of them, such as sieving, are off-stream techniques where samples must first be separated from the main stream for analysis. Therefore, the search for on-line particle size analysis systems has provided the impetus for the introduction of image-based particle size analysers to the mineral industry in the past three decades. Generally, the estimation of particle size distribution on the basis of image analysis depends on measuring a single parameter of particle profile. For example the equivalent area diameter ( d A) or mean Feret's diameter ( d F) distributions, then transforming this data to the equivalent size distribution. However, due to the irregularity of particles being analysed, it is believed that this kind of analysis may increase the error in estimation of particle size distribution since profiles of irregular particles carry more information than can be represented by a single parameter. In this paper, a proposed technique which measures two parameters, equivalent area diameter ( d A) and mean Feret's diameter ( d F), for each particle profile has been developed. The accuracy of the technique has then been investigated in the laboratory by successfully estimating (unfolding) the size distribution, where size refers to sieve size, of three samples of different particle shapes with known size distribution.

Mineral Slurry On-Stream Particle Size Analysis

On-line particle size analysis of mineral slurries is important because of the potential impact of size distribution on downstream processs, such as filtration and slurry thickening. Matti Kongas, Outokumpu Technology, Finland, discusses existing on-line particle analysis technology and presents a new technique employing laser diffraction that is being successfully applied in the field.

Effects of Calcium and Other Ionic Impurities on the Primary Nucleation of Burkeite

Primary nucleation of burkeite, which is a crystalline solid solution of sodium carbonate and sodium sulfate having the stoichiometry Na2CO3·2Na2SO4, was monitored using an on-line particle-size analyzer. Observations of unusual nucleation behavior led to explorations that identified specific impurities in the system that affected nucleus formation. By varying the conditions under which nucleation occurred, calcium ions were found to inhibit nucleation, and further research showed that the substitution of calcium ions into the burkeite crystal lattice was the reason for the observed behavior.

Particle Size Distribution Determination from Spectral Extinction Using Neural Networks

The use of techniques based on spectral extinction to recover particle size distributions has become increasingly popular in recent years. However, they are time-consuming and are not always successful in practical applications. In this paper, a novel method is proposed to determine particle size distributions using neural networks from several spectral extinction measurements. Simulations and experiments have illustrated that it is feasible to use a neural network to obtain the parameters of a particle size distribution from turbidity measurements. Although the neural network was trained using log-normal distribution data, it can be used to recover some non-log-normal distributions. The method has advantages of simplicity of use, instantaneous delivery of results, and suitability for online particle size analysis.

Plant-site evaluations of the OPSA system for on-line particle size measurement from moving belt conveyors

The image-based On-line Particle Size Analyzer (OPSA) system has been developed at the University of Utah to measure particle size distributions under plant conditions where overlapped and segregated particles are being transported on a moving belt conveyor. Evaluation of the OPSA system at several plant sites has been completed. The results show a consistent agreement between the particle size distributions obtained from the OPSA system and those obtained from laboratory screen analysis of belt samples. These results and the stable performance of the OPSA system during the field tests suggest that continuous quality control of crushed products is possible using the OPSA system.

Estimation of Operating Ranges of Jet Mills using Dynamic Modeling

ABSTRACT In today's process industry there is an increasing demand for fine polymer powders with a narrow, well defined particle size distribution. Several types of jet mills are used on a semi-large scale, though their energy efficiency is very poor. The research project aimed at a 50% energy reduction of the jet milling process by developing a control system and an improved design of the grinding plant. An important feature of the plant was the on-line particle size analysis. The topic of this paper will be the dynamic modeling of a fluid energy driven mill, in this case the opposed type. A population balance model was developed for a mechanistic description of the plant behaviour. The model contains terms for size reduction, particle transfer between zones, and internal classification. Experimental data was used to verify and optimize the model. The model served as a tool for plant control and scale-up.

Development of a softsensor for particle size monitoring

Some key control variables of industrial processes, associated with product quality, often cannot be measured directly or frequently enough to establish adequate control. In such cases, it is possible to use available measurements to provide a prediction for these process variables and use them in a control strategy, thereby giving rise to what is now commonly called a softsensor. In some industrial grinding circuits, the on-line particle size analyzer is shared between various sampling points. Therefore, for a given location, the actual measurement is only available every 10 to 20 minutes, a delay which is unacceptable for automatic control purposes. To alleviate this problem, a softsensor based on an artificial neural network has been investigated. First, the structure of the neural network and different schemes for the training process are analyzed. Then, the performance of the neural network softsensor is compared with other inferential methods such as ARMA models and Kalman filters.

Control and optinisation of mineral flocculation and transport processes using on-line particle size analysis

The use of an on-line scanning laser microscope for measuring the size distribution of particle aggregates in concentrated suspensions is demonstrated. This method allows quantification of the aggregate size distribution under actual process hydrodynamic conditions. The kinetics of floc formation processes and the subsequent breakage of flocs (e.g. under the degradative influence of shear fields encountered during transportation) can be followed in real-time. This can be used to elucidate the optimal dosing protocol for flocculant addition (agitation rate, floc dosage, polymer type etc.). Here we report, for the first time, direct on-line measurement of silica floc breakage kinetics in a stirred vessel. The floc breakage functions can be deduced for a range of nominal velocity gradients and for flocs that have been formed in different ways. Hence it is possible to predict the manner in which a floc population will behave under different shear conditions in processing circuits (e.g. in pipelines, within separation equipment). Other applications for this measurement methodology for acheiving improved understanding and control of minerals processing plant are assessed.

The development of a PC, image-based, on-line particle-size analyzer

A PC, image-based, on-line instrument for measuring coarse particle size on a conveyor belt is under development. The hardware and software required to implement such a system are discussed, including the interface for the special image-acquisition hardware. The system concept is given with examples to illustrate the step-by-step procedure for the implementation of image-processing algorithms.

On‐Line Analysis, Its Potential and Its Problems

In the application and the development of on-line methods for particle size analysis one has to distinguish three principally different problem areas: a allocation and preparation of a representative sample, b analysis of the sample, c evaluation of the results. Methods and problems connected with these areas will be discussed in this paper. Some of the methods available for rapid response analysis of particle size distributions mark the beginning of a new era of particle size analysis. One has to recognize however that a uniform solution of the problems of on-line particle size analysis is extremely difficult if not impossible to find. The boundary conditions vary to such an extent that similar methods or combinations of methods can rarely be used in different cases. Since instrument manufacturers will in most cases not be able to adopt their instruments to different industrial applications, specialists will have to solve the problem. One has to recognize, however, that not all of the steps described in this paper have been investigated to an extent which renders possible appropriate solutions. There is still room for improvement.