Ash content is an important production indicator of flotation performance, reflecting the current operating conditions of the flotation system and the recovery rate of clean coal. It also holds significant importance for the intelligent control of flotation. In recent years, the development of machine vision and deep learning has made it possible to detect ash content in flotation-recovered clean coal. Therefore, a prediction method for ash content in flotation-recovered clean coal based on image processing of the surface characteristics of flotation froth is studied. A convolutional neural network –long short-term memory (CNN-LSTM) model optimized by Newton–Raphson is proposed for predicting the ash content of flotation froth. Initially, the collected flotation froth video is preprocessed to extract the feature dataset of flotation froth images. Subsequently, a hybrid CNN-LSTM network architecture is constructed. Convolutional neural networks are employed to extract image features, while long short-term memory networks capture time series information, enabling the prediction of ash content. Experimental results indicate that the prediction accuracy on the training set achieves an R value of 0.9958, mean squared error (MSE) of 0.0012, root mean square error (RMSE) of 0.0346, and mean absolute error (MAE) of 0.0251. On the test set, the prediction accuracy attains an R value of 0.9726, MSE of 0.0028, RMSE of 0.0530, and MAE of 0.0415. The proposed model effectively extracts flotation froth features and accurately predicts ash content. This study provides a new approach for the intelligent control of the flotation process and holds broad application prospects.

In froth flotation, one of the pivotal metrics employed to evaluate the flotation efficacy is the clean ash content, given its widely acknowledged status as a paramount gauge of coal quality. Leveraging deep learning and computer vision, our study achieved the dynamic recognition of coal flotation froth, a key element for predicting and controlling the ash content in coal concentrate. A comprehensive dataset, assembled from 90 froth flotation videos, provided 16,200 images for analysis. These images revealed key froth characteristics including bubble diameter, quantity, brightness, and bursting rate. We employed Keras to build a comprehensive deep neural network model, incorporating multiple features and mixed data inputs, and subsequently trained it with a rigorous 10-fold cross-validation strategy. Our model was evaluated using robust metrics including the mean squared error, mean absolute error, and root mean squared error, demonstrating a high precision with respective values of 0.003017%, 0.053385%, and 0.042640%. With this innovative approach, our work significantly enhances the accuracy of ash content prediction and provides an important breakthrough for the intelligent advancement and efficiency of froth flotation processes in the coal industry.

The online measurement of coal ash has overcome the shortcomings of chemical tests. However, there could be large fluctuations and errors in the results of online ash monitors because of the transient change in coal quality resulting from different geological conditions in the mining process. In this study, to resolve the problems of the dual-energy γ-ray online ash monitor in the Linhuan Coal Preparation Plant, we investigated the internal factors, such as the composition of multimineral and multicoal, and external factors, such as the moisture and impurities, which affect the measurement results of the coal ash monitor. Furthermore, we developed a mathematical model to determine the effect of relevant factors on the coefficient of the online ash monitor, which revealed the relationship between coal composition and the parameters of the ash monitor, ensuring the stable and accurate measurement of ash in clean coal. The method of determining parameters used in the case of coal blending has been applied in the Linhuan Coal Preparation Plant. By comparing with tested ash content, the average absolute error and relative error for daily ash content measured in April are 0.21 and 2.18%, respectively. Meanwhile, it shows certain accuracy and reproducibility while opposed to the daily average absolute error of 0.22 and relative error of 2.39% in May.

This study successfully achieved high-precision detection of the clean coal ash content in the coal froth flotation domain by integrating deep learning with the likelihood function. Methodologically, a novel data processing and prediction framework was established by combining a deep learning Keras neural network with the likelihood function from probability statistics. The SIFT algorithm was utilized to extract key feature points and descriptors from the images, and keypoint matching and mean-shift clustering algorithms were employed to accurately obtain information on foam motion trajectories and velocities. For parameter optimization, the maximum likelihood estimation was applied to find the optimal parameter estimates of the likelihood function, ensuring enhanced model accuracy. By incorporating the optimized likelihood function parameters into the Keras deep neural network, an efficient prediction model was constructed for the dosage of flotation reagents, froth velocity, and clean coal ash content. The model’s evaluation involved six performance metrics. The experimental results were highly significant, with R² at 0.99997%, RMSE at 0.04458%, MAE at 0.00170%, MAPE at 0.02329%, RRSE at 0.00994%, and MAAPE at 0.00067%.

This study addresses the timely and accurate measurement of coal ash content by proposing a detection model based on nuclear science technology, which is validated using FLUKA 4-4.0 simulation software. The background provided highlights the fact that coal ash content is a critical sales indicator, and its precise measurement is essential for adjusting production parameters in coal preparation plants. In terms of methodology, this study employs the widely used FLUKA4-4.0 software in the field of nuclear physics to simulate X-ray transmission through coal, investigating the impact of changes in coal type on the accuracy of ash measurements. The results indicate that, when the proportions of high-atomic-number elements in coal remain constant, the ash measurement results are accurate and reliable. However, significant fluctuations occur when these proportions change. The conclusion emphasizes the fact that variations in coal type are the primary cause of inaccuracies in ash measurement, particularly when the ratios of high-atomic-number elements are altered. This research provides a new perspective on the online measurement of coal ash content and offers theoretical support for improving measurement accuracy.

Three products hyrdrocyclone screen (TPHS) has been proposed for particle separation based on size. In TPHS, a cylindrical screen was embedded in a conventional hydrocyclone (CH) to combine the centrifugal classification and screening to particle separation based on size. The industrial application of TPHS indicates its better device performance than CH. Although, the earlier studies reveal some common understanding for TPHS, the information of the absent air column remains unknown. Hence, the combination of physical experiment and numerical simulation was considered involving a 75 mm TPHS for this knowledge gap. First, both the computational fluid dynamics (CFD) simulation with Reynolds stress mode and the physical experiment with a high-definition camera illustrate the development process details of a flow field in TPHS. That is, the water was imported along the tangential inlet into TPHS; then under the effects of the feed chamber wall and gravity, the liquid phase spiraled downward until the cylindrical screen passed through the sieve; as the liquid moved to the spigot, it could be discharged in time due to the small underflow port, thus the volume fraction of air rapidly reduced from 1 to 0; subsequently the water filled the TPHS and the absent air column could be observed. Furthermore, the distribution comparisons of air volume fraction and static pressure show that TPHS displayed the absent air core with the negative static pressure in the center region along the z-axis, while CH displayed the opposite features. In addition, despite the different inlet velocity, TPHS consistently presented the vanished air column which could be ascribed to the fact that the present cylindrical screen resulted in positive static pressure distribution inside TPHS.