Coal Flotation Research Articles

The application of machine learning models based on particles characteristics during coal slime flotation

In this study, four different machine learning (ML) models were used to simulate the migration behavior of minerals during coal slime flotation based on particle characteristics (shape, size, compositions, and types): random forest (RF), logistic regression (LR), AdaBoosting (Ada), and k-nearest neighbors (KNN). For ML model development, 70% of the total data was used for the training phase, and 30% was used for the testing phase. F-score and area under the curve (AUC) were used as the most vital indicators for evaluating the different ML models. Compared to the other ML models, the RF model had the best accuracy for simulating particle migration behavior during flotation. Furthermore, the RF model avoided the drawback of having to be retrained when the feed conditions changed. The results revealed that particle size and particle composition play the most significant role in coal slime flotation.

Cleaning high ash coal waste from coking coal via froth flotation method

In this work, scanning electron microscope (SEM), X-ray photoelectron spectroscopy (XPS), X-ray diffraction (XRD) and laser particle size methods were employed. The surface of the particle obtained from the cyclonic-static microbubble flotation column (FCSMC) concentrate was coated with numerous small particles when compared to the clean coal collected from the Denver flotation cell. Even though the mass recovery of FCMSC concentrate (51.13%) was approximately twice the mass yield of Denver flotation cell concentrate (26.66%), the ash content in FCSMC concentrate (33.07%) was higher than that of the Denver flotation cell concentrate (14.88%). This result can be attributed to the advantages of FCSMC in small coal particle flotation, and this resulted in the misplacement of coarser particles with lower ash content in the tailing. The highest mass recovery of 0.250–0.500 mm in magnesium ion solution was 95.39%, which was slightly high than that (92.84%) in calcium ion solution. However, the concentrate ash content with 0.250–0.500 mm coarse size fraction in magnesium ion solution was 11.41%, which was higher than that (11.14%) in calcium ion solution. Flotation performances of different size fractions in rougher-cleaner flowsheet indicated that flotation responses in calcium ion solution were better than that in magnesium solution.

The interaction of frothers with hydrophobic and hydrophilic sites of coal particles in NaCl solution

Frothers are added in coal and mineral flotation to assist the formation of small air bubbles and stable froth. However, we have found that frothers can also adsorb on hydrophobic and hydrophilic sites of coal particles in flotation. In the current study, the interaction of two widely used frothers with coal particles in NaCl solution was investigated. While focused beam reflectance monitor was applied to measure the aggregation of coal particles, surface tension was measured to understand the aggregation of frother molecules. It was found that the adsorption of frothers on coal particles in NaCl solution was influenced by the NaCl-modified properties of frother molecules and coal particles. The presence of NaCl enhanced the intermolecular interaction between frothers and water and also facilitated the aggregation of coal particles. On hydrophobic sites, the hydrophobic interaction (CH/ π) between frothers and coal particles could be easily replaced by the stronger π/π interaction between coal particles, while on hydrophilic sites, the hydrophilic interaction between frothers and coal particles could be enhanced through ion-dipole attraction. Whether the hydrophobic interaction or the hydrophilic interaction played a dominant role in frother adsorption depended on the ratio of hydrophobic to hydrophilic sites and the formation of hemimicelles of frother molecules.

Destabilising persistent coal froth using silicone oil

Persistent froth is becoming more and more common in coal and mineral flotation plants and presents safety and operational challenges. No effective method has been developed to destabilise persistent froth. As a new initiative, this study examined the structural difference between persistent foam and coal froth, based on which a solution was developed to maximumly destabilise coal froth. Destabilisation test, oscillatory rheology measurement and scanning electron microscopy (SEM) analysis indicated that the coal froth was more stable than the foam due to the formation of thin capillaries and tightly arranged coal particles on bubble surfaces. Although 107 µm silicone oil droplet could completely destabilise the foam at 2 mmol/L concentration, it only destabilised less than 50% coal froth even at 6 mmol/L concentration. To maximumly destabilise the coal froth formed by −38 and −20 µm particles, 24 and 18 µm silicone oil droplets were required to pass through the thin capillaries and enter the bubble films, respectively. However, smaller silicone oil droplets could not bridge the bubble films to destabilise coal froth and therefore a critical droplet size occurred depending on the size of particles stabilising the froth.

Processing of finely dispersed technogenic raw materials for aluminum production in order to extract valuable components

The paper provides the results of experiments on the hydrometallurgical processing of finely dispersed technogenic raw materials for primary aluminum production in Soderberg cells (case study of the Irkutsk Aluminum Smelter) – aged sludge. The components of this sludge are dust from electrostatic precipitators (79.7 %), wet gas cleaning sludge (4.4 %) and coal froth flotation tailings (15.8 %). According to the grain-size analysis carried out, aged sludge sample particles have a size of –50 μm. According to the chemical composition analysis of the sludge sample, main components in it are carbon, cryolite, chiolite with a small amount of other compounds (corundum, ralstonite, spodumene, fluorite). Fluorine leaching experiments were carried out with a 2 % sodium hydroxide solution at a stirrer speed of ~1020 rpm. Using the mathematical planning of a three-factor experiment, it was found that the maximum concentration of fluorine in the solution (15.844 g/dm3) is achieved with the following optimal parameters of fluorine alkaline leaching: temperature of 90 °C, liquid-to-solid ratio of 9 : 1, and time of 90 min. The multidimensional polynomial equation was obtained for fluorine alkaline leaching from aged sludge. Cryolite was obtained from fluorine-containing solutions (by the reaction of sodium fluoride interaction with sodium bicarbonate and an aluminate solution), which was confirmed by X-ray phase analysis data.

The impact of water salinity and its interaction with flotation reagents on the quality of coal flotation products

Coal preparation plants are increasingly turning to alternative sources of water to reduce their consumption of fresh water, including recycled water. Some of these sources can vary greatly in inorganic electrolyte content. In this study, standard flotation tests were conducted with varying water chemistry. The concentrations of collector, frother and salt, which reflect the water quality, were varied to study the main effect of the reagents with water of varying purity as well as their interactions. The concentration of frother was clearly the most influencing factor for the yield, water recovery and ash. However, it was found that ionic content (or water quality) also had a profound effect. In particular, a higher concentration of salt tended to produce flotation products with a higher ash content. Thus, as regulation of coal preparation plants becomes more stringent and water of varying quality such as process water is used in flotation, the quality of the coal concentrates deteriorate. Thus, the environmentally beneficial practice of recycling water can lead to increased air pollution and waste generation due to an inferior product.

Investigation of adhesion behavior between reactive oily bubble and low-rank coal

Reactive oily bubble, defined as air bubble covered with a thin layer of diesel oil containing surfactants, was used in low-rank coal flotation. Both the cationic surfactant dodecyl amine (DDA) and the anionic surfactant palmitic acid (HPa) were used to generate the reactive oily bubble. The strengthening mechanism of the reactive oily bubble was investigated by a series of flotation tests and the adhesion force measurements. The results showed that at low concentrations (0.5, 2.5, and 5.0 mM), the effect of the reactive oily bubble with DDA to enhance adhesion and promote separation was not significant. When the concentration of DDA increased to 50 mM, the coalescence of bubbles deteriorated separation. In addition, the competitive adsorption between oil and DDA (reverse orientation) reduced the hydrophobicity of the coal surface and weakened the adhesion force between the bubble and coal surface. While for HPa, the synergistic adsorption between HPa and diesel oil played an important role in promoting separation and the concentration of HPa in the oily bubble and the flotation efficiency was positively correlated. The values of normalized contact line (x) could be used to represent the expansion of the three-phase contact line which indicated that the reactive bubble containing HPa could squeeze more water out from grooves of coal surface than that of bubble containing DDA. These findings suggest that using the reactive oily bubble is conducive to promoting the separation efficiency only when proper concentration and type surfactant is used to generate the reactive oily bubble.

Interpretable modeling of metallurgical responses for an industrial coal column flotation circuit by XGBoost and SHAP-A “conscious-lab” development

Surprisingly, no investigation has been explored relationships between operating variables and metallurgical responses of coal column flotation (CF) circuits based on industrial databases for under operation plants. As a novel approach, this study implemented a conscious-lab “CL” for filling this gap. In this approach, for developing the CL dedicated to an industrial CF circuit, SHapley Additive exPlanations (SHAP) and extreme gradient boosting (XGBoost) were powerful unique machine learning systems for the first time considered. These explainable artificial intelligence models could effectively convert the dataset to a basis that improves human capabilities for better understanding, reasoning, and planning the unit. SHAP could provide precise multivariable correlation assessments between the CF dataset by using the Tabas Parvadeh coal plant (Kerman, Iran), and showed the importance of solid percentage and washing water on the metallurgical responses of the coal CF circuit. XGBoost could predict metallurgical responses (R-square > 0.88) based on operating variables that showed quite higher accuracy than typical modeling methods (Random Forest and support vector regression).

Investigating the effect of lip froth washing on coal yield during flotation of a high ash South African coal

Investigating the effect of lip froth washing on coal yield during flotation of a high ash South African coal

Deep learning-based ash content prediction of coal flotation concentrate using convolutional neural network

Convolutional neural networks, as the current state-of-the-art in image classification, are regarded as a promising way for flotation soft sensors based on froth images. This paper proposes a flotation soft sensor solution that predicts the concentrate ash content of coal flotation by using froth images and convolutional neural network architecture. A froth image dataset from the coal flotation site was divided into seven interval classes based on concentrate ash content, and data augmentation was used for this dataset. Then, several state-of-the-art convolutional neural networks (AlexNet, VGG_16, VGG_19, ResNet_18, ResNet_34, ResNet_50, ResNet_101) are trained to classifying the froth images with different concentrate ash content interval (integer ± 0.5). The classification performance of each model, the relationship between model performance and hyperparameters, and the abstract pixel feature of the best model are examined and visualized. The highest classification accuracy (97.1%) is obtained by the ResNet_101 network after fine-tuning. CNN extracts the features into a multi-channel set of abstract pixel features (7 × 7) before entered into the fully-connected layer. The results show that the abstract pixel feature significantly outperformed the manually engineered features. Finally, it takes only 0.31 s for the model to classify a froth image, which indicates that the convolutional neural network has a good industrial performance.

The role of polyvinylpyrrolidone in the selective separation of coal from quartz and kaolinite minerals

High-effective flotation of valuable coal resource from gangue is more challenging due to the increase of ultra-fine minerals in coal slime. Despite the use of novel flotation equipment and process strengthening regulation, the method of flocculation of coal or gangue particles to enhance selective separation draws increasing attention recently. The great potential in selective flocculation of particles makes the development of novel polymer flocculants a research hotspot in this field. Particularly, as a kind of polymeric reagent, polyvinylpyrrolidone (PVP) has great significance in clay aggregation/dispersion whereas there are few related researches in coal flotation. Thus, the effect of PVP on coal flotation was systematically investigated in this study by flotation tests and the measurements of zeta potential, turbidity, FBRM, and X-ray photoelectron spectroscopy. The results proved that PVP would be preferentially adsorbed on quartz and kaolinite surfaces rather than coal. The adsorption of PVP with low dosage induced the aggregation of quartz and kaolinite, as indicated by turbidity and FBRM results, while it basically made no difference on coal. It was discovered that the selective flocculation of quartz and kaolinite by PVP promoted the selectivity of flotation. This study furnishes a novel and feasible polymer flocculant to improve coal flotation efficiency. • PVP as a flotation regulator could reduce the ash content of clean coal. • PVP could improve flotation selectivity by flocculation of gangue particles. • PVP could adsorb preferentially on quartz and kaolinite than coal.

The interactions of the biosurfactant surfactin in coal flotation

Surfactin, a biosurfactant, has the potential to replace synthetic surfactants in froth flotation. This research investigates the surface interactions between surfactin and coal which determine whether surfactin has collecting or depressant characteristics on coal using electrokinetic, FTIR, and micro-flotation experiments. At concentrations of 5 mg/L, surfactin removes species on the surface of coal, especially in the alkaline pH range. At surfactin concentration of 15 mg/L adsorption of surfactin onto coal surfaces are indicated and likely to be due to hydrophobic physisorption between aliphatic groups of surfactin and the hydrophobic carbonaceous surface functional groups of coal. Hydrophobicity of coal was observed to be enhanced at pH 3,6 and 10 except the intermediate pH of 8. Micro-flotation tests showed that hydrophobicity of coal was most significantly enhanced by surfactin presence at pH 3 and 15 mg/L from 42.9% to 74%, and least impacted when surfactin concentration was 5 mg/L at a pH of 10. The collecting or depressant characteristics of surfactin on coal as a function of pH and surfactin concentration have been demonstrated for the potential use of this novel biosurfactant as a Green Chemistry substitute for synthetic surfactants.

Spreading behavior of dodecane-oleic acid collector mixture in low-rank coal flotation

Spreading of water-insoluble collectors on the mineral surface is important in flotation. In this study, the spreading behavior of dodecane (D), oleic acid (OA), and dodecane-oleic acid (D-OA) mixture in low-rank coal flotation was investigated. The viscosity and oil/water interfacial tension was measured to calculate the spreading coefficient and speed. The relaxation of coal-oil–water contact angle and the movement of three- phase contact line during the collector droplet spreading process were interpreted using the molecular kinetic (MK) model. It was found that the order of absolute values of spreading coefficient were D-OA < OA < D, and the spreading speeds were in the order of OA < D-OA < D. D-OA had the smallest coal-oil–water contact angle and the longest contact line length. The addition of OA can enhance the interaction between D and the low-rank coal surface, but at the same time reduce the spreading speed. This was mainly because OA not only increased the spreading force of the droplet but also increased the spreading resistance. MK model fitting results showed that the length of molecular displacement increased in the order of D < D-OA < OA, indicating that the hydrophobic sites were more dense than hydrophilic ones on the coal surface and D-OA interacted with both sites simultaneously. The molecular substitution frequency and the free energy of wetting activation had an opposite order. High molecular substitution frequency meant fast-spreading velocity, and thus the MK model fitting parameters were consistent with the contact angle relaxation results.

Insight into the low-rank coal flotation using amino acid surfactant as a promoter

Low-rank coal flotation faces challenges due to its poor surface hydrophobicity. Massive oily collector, e.g. kerosene and diesel, are needed to improve its flotation performance, causing huge cost input. This study experimentally demonstrated that Sodium Cocoyl Glycinate (SCG), an amino acid surfactant, could help with obtaining higher efficiency when adopted in low-rank coal flotation. Results showed that compared with the condition using kerosene (5000 g/t) alone, a tiny amount of SCG (50 g/t) increased the concentrate yield by about 12%, while its ash content roughly remained the same. Effects of SCG on low-rank coal flotation was established from coal surface hydrophobicity, solution surface tension and the forth layer. The amphipathic groups of Sodium Cocoyl Glycinate would bond both with the hydrophilic and hydrophobic sites on the coal surface through hydrogen bonding, leaving the hydrophobic end exposed, which shorten the bubble-particle attachment time to about 100 ms. The decrease of solution’s surface tension was conducive to the emulsification of kerosene, making droplets diameter distribution into a smaller range of 0.26 to 10 μm. Meanwhile, lower surface tension offset the negative impact of oily collector, providing a more stable froth layer. Considering its hypotoxicity, biodegradability, environmental friendliness, Sodium Cocoyl Glycinate could be a green flotation promoter assisting with the conventional oily collectors.

Frother distribution in an industrial coal flotation circuit

ABSTRACT In reuse of process water in flotation plants, the inability to measure residual frother concentrations in both flotation and plant water circuits in real time has created challenges for flotation process control. We previously developed a simple and fast method to measure frother concentration based on the guest–host interactions between frother molecules and specially formulated liquid comprising cyclodextrin and pyrene. In the present study, the formulation and analysis procedure were upgraded to further improve the measurement sensitivity. The potential of the present method was demonstrated through methyl isobutyl carbinol (MIBC) concentration measurement at six different locations across an industrial coal flotation circuit. The measured MIBC distribution was in good agreement with that obtained using the colorimetric method developed by Finch and coworkers at McGill University. The implications of the established MIBC distribution in the coal flotation circuit for frother dosing and water reuse were discussed.

Evaluation of kinetics and thermodynamics of a naturally hydrophobic coal flotation in the presence of frother and regulated with pH

ABSTRACT Zeta potential of coal particles and kinetics of flotation along with the maximum recovery after a long time of fractionated flotation of naturally hydrophobic coals, both as a function of pH, were investigated. Empirical relations between the 1st order rate constant of flotation and zeta potential as well as the maximum recovery and zeta potential were established. Since fractionated flotation has features similar to chemical reactions, thermodynamic relations between maximum recovery and zeta potential as well as between rate constant and zeta potential were found. It was determined that the 1st order rate constant depends on two constants: flotation equilibrium constant and activity coefficient of surface electrical potential. It also depends on two variables: surface electrical potential (approximated with zeta potential) and the rate constant of the particle-bubble aggregate (flotulas) disintegration being a function of surface electrical potential. The values of the flotation equilibrium constant, activity coefficient of surface electrical potential and reverse process rate constant were determined. These parameters (especially flotation constant and activity coefficient of surface electrical potential) can be used for comparison of different flotation systems. Also useful for comparison purpose are kinetico-thermodynamic limits (Rmax vs k1) and Arrheniusan-type (ln k 1 vs ξ) plots.

Using emulsified and pre-dispersed hydrocarbon oil in waste coal reprocessing: A case study

Waste coal is a major environmental hazard produced in the traditional coal preparation unit operations. In this paper, a case study is presented for further scavenging the combustible matter in waste coal using flotation with a self-designed collector (Ksw) consisted of kerosene (K), span (S) and water. The said collector took advantage of both the emulsification and pre-dispersion technique so that the waste coal processability could be improved to a greater extent. Our flotation tests have shown comparable combustible matter recovery when using Ksw and emulsified kerosene (Ks), and both collectors significantly outperformed that of K. At 4 kg/t collector dosage, combustible matter recovery was at 54.9%, 64.9% and 65.3% for K, Ks and Ksw respectively. However, Ksw represented an economically superior option due to its specific composition features. The zeta potential measurements and interaction energy calculations have suggested that although there existed an electrostatic repulsion between the selected oil droplets and coal particles, flotation of waste coal was still feasible because of the hydrophobic force dominating the interaction process. Further, it was found that the interaction energy calculation results could not sufficiently account for the superior flotation performance of using Ksw as collector. Dynamic light scattering (DLS) measurement has shown that the other key instruments at play were the diameter and number of the oil droplets. At 400 mg/L collector suspensions, emulsification reduced the effective diameters of K from 4.23 to 1.62 μm (Ks), and further decreasing to 1.12 μm (Ksw) when dispersion was imposed. The outcomes of this work have suggested Ksw as an economically and technically superior reagent over the conventional hydrocarbon oil collectors for waste coal reprocessing.

Coarse-Grain molecular model development and dynamics simulations study of dodecane droplet spreading at the coal-water interface

The mesoscopic scale is usually used to realize the dynamics calculation on a larger spatial scale and time scale. However, there have been few dynamic simulation calculations attempted on the mesoscopic scale for coal flotation systems. In this study, based on the Martini force field, the Coarse-Grain (CG) models of coal, dodecane, and water were established, and the dynamics simulations were performed to study the collector droplet spreading at the coal-water interface. The CG dynamics simulations results showed that the interaction between dodecane and coal was a spontaneous process, and dodecane existed as a collector at the coal-water interface. The radial distribution function of dodecane around coal and water indicated that dodecane was more likely to be distributed around the coal, and as far away as possible from the water. The CG dynamics simulations also revealed the specific spreading stages of the dodecane droplet. The dodecane drop first moved in the form of droplets to the coal-water interface in the aqueous phase. Then at the coal-water interface, the circular droplet broke its shape and spread rapidly on the coal surface through the dodecane CG molecules. Finally, the dodecane CG molecules penetrated deeper into the coal surface layer and formed a dynamic and stable oil film. This provided a good attempt and reference for studying the spreading of collector droplets on the coal-water interface.

Flotation behavior of metallurgical coals from different seams under electrolytic process water

ABSTRACT Various studies have reported that electrolytes present in the plant’s process water influence the coal flotation performance. However, coal fines, which are fed into the flotation circuit are from different coal seams, and their surface properties usually vary considerably. In this study, flotation experiments were conducted using coal fines from three different coal seams of the West Bokaro coalfield in plant’s process water to explore the extent of change in flotation response of each coal. Besides, flotation experiments using deionized water were parallelly examined for comparison. XRD, SEM, FTIR, BET, contact angle, wetting rate, and zeta potential measurements were performed to understand and compare the surface properties of three coals. The flotation results indicated that compared to deionized water, process water increased flotation yield, combustible recovery as well as ash recovery for all three coals at all examined collector dosages.

Composition characteristics of siloxane derived waste lubricating oil and its effect on low-rank coal flotation

Waste motor oil/lubricating oil (WMO) is a hazardous and important resource. The waste siloxane-based motor oil/lubricating oil (WSMO) has a special composition based on siloxane. The WSMO has a large and increasing production worldwide. In this paper, the feasibility of WSMO as a collector for low-rank coal (LRC) flotation was discussed with a comparison to diesel oil (DO). The gas chromatography-mass spectrometry (GC/MS) and thermal gravimetric analysis (TGA) were used to analyze the chemical composition of WSMO. The flotation experiment was conducted to investigate the collectivity and selectivity of WSMO as an LRC collector. The interaction between WSMO and LRC was systematically analyzed using X-ray photoelectron spectroscopy (XPS), contact angle, zeta potential, and induction time measurements. The results indicated that WSMO majorly contained dimethylsiloxane and cyclodimethysiloxane with various chain lengths. High concentrates yield and low ash content was obtained in the flotation using WSMO. The adsorption of WSMO can mask the oxygen-containing groups (OCGs) on the LRC surface, leading to an increase of C content and a decrease in O content. This was beneficial to improve the hydrophobicity of the LRC surface so that the water-coal contact angle increased from 35.6° to 57.85° after WSMO treatment. Moreover, a significant increase in zeta potential was observed in the pH range of 4 to 10. LRC particles pretreated with WSMO were more easily adhere to the bubble surface in a short induction time of 40 ms. It was demonstrated that WSMO was an effective collector to promote LRC flotation.