Mill Load Research Articles

Multi-source unsupervised domain adaptive mill load forecasting method based on deep learning and fusion features

In the grinding industry, accurate prediction of the mill load is the key to enhancing plant profitability and reducing mill failures. After the condition changes, the model is usually mismatched, and the true label cannot be obtained in time. To solve this problem, we propose a multi-source unsupervised domain adaptation model based on deep learning and fusion features (MUDA_DF), which is suitable for predicting multiple mill load parameters. Considering that the information contained in the data of a single working condition is insufficient to completely cover the information of the unmodeled working condition, we introduce multiple source domains into the model, which can improve the generalization of the model. Since the special features of the source domain also contain useful information, we use the features after the fusion of special features and common features for regression prediction. By designing multiple loss functions, we achieve the prediction of three mill load parameters in unknown operating conditions: material-to-ball volume ratio (MBVR), pulp density (PD), and charge volume ratio (CVR). Experiments were carried out on data sets collected by a laboratory ball mill. Compared with the traditional mill load soft measurement method, the R2 of this method can reach above 0.9 and the highest can reach 0.99, which proves the effectiveness of this method.

A Case Study on Reliability Improvement of New Bar Mill Stand Gearboxes

Abstract: This paper describes the major work carried out in study & analyzing the critical reasons for failure of mill stand drive gearboxes that had occurred in New Bar Mill (Rebar Mill) in past 5 years. Suggest improvement measures and maintenance guidelines to decrease the probability of failure and hence develop an action plan for reliability improvement of New Bar Mill stand gearboxes. In the process, study on basics of gearbox and on rolling process of New Bar Mill has been done. Present production data analysis has been done for actual mill load condition. Based on present mill load data, current running service factor of gearboxes has been calculated. Many gearboxes of New bar mill stand found with low service factor than recommended as per AGMA standards. Accordingly, upgradation of mill drives gearboxes have been done for higher service factor and deployment of mill load alarm system to avoid any abnormal load on mill drive gearboxes has been carried out based on study

Developing a data-driven soft sensor to predict silicate impurity in iron ore flotation concentrate

Soft sensors are mathematical models that estimate the value of a process variable that is difficult or expensive to measure directly. They can be based on first principle models, data-based models, or a combination of both. These models are increasingly used in mineral processing to estimate and optimize important performance parameters such as mill load, mineral grades, and particle size. This study investigates the development of a data-driven soft sensor to predict the silicate content in iron ore reverse flotation concentrate, a crucial indicator of plant performance. The proposed soft sensor model employs a dataset obtained from Kaggle, which includes measurements of iron and silicate content in the feed to the plant, reagent dosages, weight and pH of pulp, as well as the amount of air and froth levels in the flotation units. To reduce the dimensionality of the dataset, Principal Component Analysis, an unsupervised machine learning method, was applied. The soft sensor model was developed using three machine learning algorithms, namely, Ridge Regression, Multi-Layer Perceptron, and Random Forest. The Random Forest model, created with non-reduced data, demonstrated superior performance, with an R-squared value of 96.5% and a mean absolute error of 0.089. The results suggest that the proposed soft sensor model can accurately predict the silicate content in the iron ore flotation concentrate using machine learning algorithms. Moreover, the study highlights the importance of selecting appropriate algorithms for soft sensor developments in mineral processing plants.

Control of ball mill operation depending on ball load and ore properties

Ball mills are used for the second stage of ore grinding in mining operations by proportioning ore piles. Controlling the operating conditions of ball mills is one of the key factors for optimal mill perform and high mill load turnover. One of the factors affecting the efficiency of the ball mill is the loading of the balls. Optimum distribution of ores according to the size of the mill, increasing the efficiency of the grinding process and reducing energy consumption are achieved. In addition, the ore properties are one of the important parameters for the optimal utilizing of the ball mill. Ore can have various properties, such as hardness, density, moisture content, etc., which can affect the ore’s productivity and grinding efficiency

Assessing the Effects of Material Properties on Load Behavior in Dry Ball Mills Using DEM

Understanding how the mill load behaves is crucial for enhancing ball mill effectiveness. This study aimed to create a discrete element method (DEM) model to simulate the motion charge in ball mills and to analyze how the simulation material properties affected the load behavior. The steel balls were modelled as a collection of distinct particles, each of which was subject to Newton’s laws of motion and tracked in a Lagrangian manner. Hertzian contact law was used to describe inter-particle collisions. Then, this numerical model was coded using the open-source C++ program LAMMPS Improved for General Granular and Granular Heat Transfer Simulation (LIGGGHTS) to mimic laboratory and pilot-scale ball mills. The load positions measured from the DEM simulations were compared to the published experimental data and empirical models of comparable laboratory and pilot-scale experiments to validate the findings. The angular shoulder position ranged between 137° to 154° for the range of Young’s modulus of 0.5 to 1000 MN/mm2. Angular shoulder and toe positions had a variation of less than 10% from the laboratory and pilot-scale experimental data. The outcomes demonstrated a significant relationship between load position and material characteristics such as Young's modulus in DEM simulation. This preliminary model can be used for choosing the appropriate material parameters for ball mills both with DEM and coupled CFD–DEM multiphase simulations. This assessment concluded that material properties affect the load behavior in computer simulations of ball mills.

Monitoring of grinding condition in drum mills based on resulting shaft torque

Grinding is the most energy-intensive process among all stages of raw material preparation and determines the course of subsequent ore beneficiation stages. Level of electricity consumption is determined in accordance with load characteristics forming as a result of ore destruction in the mill. Mill drum speed is one of process variables due to which it is possible to control ore destruction mechanisms when choosing speed operation mode of adjustable electric mill drive. This study on increasing energy efficiency due to using mill electric drive is based on integrated modelling of process equipment – grinding process and electromechanic equipment – electric drive of grinding process. Evaluating load torque by means of its decomposition into a spectrum, mill condition is identified by changing signs of frequency components of torque spectrum; and when studying electromagnetic torque of electric drive, grinding process is monitored. Evaluation and selection of efficient operation mode of electric drive is based on the obtained spectrum of electromagnetic torque. Research results showed that with increasing mill drum speed – increasing impact energy, load torque values are comparable for the assigned simulation parameters. From the spectra obtained, it is possible to identify mill load condition – speed and fill level. This approach allows evaluating the impact of changes in process variables of grinding process on parameters of electromechanical system. Changing speed operation mode will increase grinding productivity by reducing the time of ore grinding and will not lead to growth of energy consumption. Integration of digital models of the technological process and automated electric drive system allows forming the basis for developing integrated methods of monitoring and evaluation of energy efficiency of the entire technological chain of ore beneficiation.

Heterogeneous selective ensemble learning model for mill load parameters forecasting by using multiscale mechanical frequency spectrum

A ball mill is a heavy mechanical device and its safe operation affects the entire grinding process. Mill load is a key index in the optimum operation of the grinding process, but it cannot be measured directly. In industrial practice, operational experts normally estimate its value based on their experiences and the mechanical signals produced by the ball mill. In this paper, we proposed a heterogeneous selective ensemble method by using a multiscale mechanical frequency spectrum. The multicomponent adaptive decomposition algorithm is first used to decompose the original shell vibration and acoustic signals into sub-signals with different timescales. Then, selective ensemble (SEN) kernel projection to latent structure algorithm is used to model the spectral data of these sub-signals. Furthermore, the latent features of multiscale spectra are extracted to construct SEN models based on fuzzy inference. Finally, the two types of heterogeneous SEN models are fused by using information entropy. The main contribution of this study is that the proposed soft-sensing model has a dual-layer ensemble structure that can fuse multi-source information in different mechanical sub-signals with physical meaning. Moreover, the proposed model can simulate the fuzzy cognitive behavior of domain experts in the mineral grinding process. The effectiveness of the method is verified by the shell vibration and acoustical data of a laboratory-scale ball mill.

The influence of water-based nanolubrication on mill load and friction during hot rolling of 304 stainless steel

Using pure water in comparison to water-based lubricant containing 4% TiO2 nanoparticles (NPs), the hot rolling tests of 304 stainless steel were carried out at a rolling temperature of 1050 °C under varying rolling reductions and speeds. The effects of lubrication on rolling force, torque, power and contact friction were systematically investigated. The coefficient of friction (COF) during steady-state hot steel rolling was inversely calculated using a developed flow stress model. The COF models including the effects of rolling reduction and speed were proposed via multiple linear regression. The results indicated that the use of the nanolubricant enabled a reduction of rolling force up to 6.1% and decreases in rolling torque and power up to 21.6%, compared to that of water condition. The results obtained from the linear regression agreed well with those from the inverse calculation, suggesting the developed COF models had high accuracy. The lubrication mechanisms were derived from a boundary lubrication regime, owing to ball bearing and mending effects of TiO2 NPs, and formation of thin lubricant film under high rolling pressure.

Study of the Load Forecasting of a Wet Mill Based on the CEEMDAN-Refined Composite Multiscale Dispersion Entropy and LSTM Nerve Net

To address the problem of wet ball milling in a strong noise environment, it is difficult to accurately detect the internal load parameters of the cylinder during grinding. In this paper, a mill load parameter prediction method is proposed based on complementary ensemble empirical mode decomposition (CEEMDAN)-refined composite multiscale dispersion entropy (RCMDE) and-long and short-term memory (LSTM) neural networks. Using this method, the vibration signals of the mill barrel under strong noise were decomposed using the CEEMDAN algorithm, sensitive modal components with strong correlation with the original signal were selected for reconstruction using the correlation coefficient method, and features of the reconstructed signals under different load parameters were extracted through RCMDE. The load characteristic vector of an RCMDE mill was used as the input of LSTM neural networks, and the filling rate, material and ball ratio, and grinding concentration were used as the output to establish the internal load prediction model of wet mill. Experiment results show that the prediction method has a high accuracy, with average absolute percentage errors of the filling rate, feed-to-ball ratio, and grinding concentration of 6.08%, 3.50%, and 3.47%, and average absolute errors were of 0.0167, 0.0146, and 0.0146, respectively.

Extremum seeking control for optimization of an open-loop grinding mill using grind curves

A semi-autogenous grinding mill is simulated with gradient and non-gradient based extremum seeking controllers to maximize the mill performance using grind curves. Grind curves map the essential performance measures of a grinding mill to the mill load and rotational speed. The curves vary with the changes in the feed ore characteristic but show generic parabolic features with extremums. The extremum seeking controllers search along the unknown input–output map to steer the process towards an unknown optimum. In this study, a classical perturbation-based method, a time-varying parameter estimation-based method and the Nelder–Mead simplex method are employed as extremum seeking control (ESC) methods to search along the grind curves to either optimize the mill throughput or grind by means of manipulating the mill feed or rotational speed. The proposed extremum seeking controller could reduce the need for a plant operator to manually select the optimal operating conditions that maximize the performance measures of a grinding mill. Since the controller is agnostic to the process model, the grinding mill can be optimized without the need for a detailed process model. The simulated results show that the extremum seeking controllers steer the mill operating conditions toward the steady-state optimum and can be used to satisfy operational objectives. However, the slow grinding mill dynamics result in a long convergence rate when the initial conditions are far from the optimal operating conditions.

Soft measurement of ball mill load under variable working conditions based on deep transfer learning

In actual industrial production, labeled sample data of a ball mill is difficult to obtain under variable working conditions. Aiming to realize the soft measurement of ball mill load under variable working conditions, a joint discriminative high-order moment alignment network (JDMAN) is proposed, based on the deep transfer learning in this paper. With this method, discriminative features were learned through jointly training labeled samples belonging to the source domain and unlabeled samples belonging to the target domain. Simultaneously, the features learned by a deep convolution network were aligned and clustered through central moment discrepancy and center loss to accomplish transfer. The comparison with other transfer methods indicates that the proposed JDMAN can effectively promote the accuracy of soft measurement under variable working conditions.

СОСТОЯНИЕ И НАПРАВЛЕНИЯ РАЗВИТИЯ ТЕХНИКИ И ТЕХНОЛОГИИ ИЗМЕЛЬЧЕНИЯ

The analysis of the state and direction of development of equipment and technology for grinding materials in the Russian and foreign cement industry is carried out. The equipment used for grinding with ball mills, vertical mills and press-roll grinders is considered. The effectiveness of the use of ball mills and their advantage over other grinding units used in large-capacity production is shown. One of the main directions for the development of the efficiency of the grinding process in ball drum mills has been determined. It consists in the development of intra-mill devices that are aimed at destroying stagnant zones in the feed, intensifying the work of grinding bodies and creating conditions within the mill classification of the crushed material. The design and principle of operation of a ball drum mill equipped with an inclined inter-chamber partition is described. It ensures the transverse-longitudinal movement of grinding bodies in the mill body. The advantages of using an inclined inter-chamber partition are given, which include a reduction in the weight of the mill load and the hydraulic resistance of partition, improved aspiration, reduced power consumption and provision of transverse-longitudinal movement.

Assessing load in ball mill using instrumented grinding media

Monitoring mill load is vital for the optimization and control of grinding process. This study proposed the use of an instrumented grinding media to assess solid loading inside a ball mill, with size and density of the instrumented ball comparable to that of the ordinary grinding media. The acceleration signal was captured by an embedded triaxial accelerometer. The signal was first detrended by a complete ensemble empirical mode decomposition and then reconstructed using a correlation coefficient method. The filling ratio, particle to ball ratio, time domain features and sample entropy are features extracted from the signal, providing input to a support vector machine (SVM) learning model. Grinding experiments with different loads were conducted. The typical loading level was classified according to grinding efficiency index and associated power consumption. Different methods were adopted to determine the optimal values of parameters in the SVM model, including particle swarm optimizer (PSO), genetic algorithm (GA), and grid search (GS). The results showed that the accuracy of particle swarm optimizer can reach 96.67%. This study demonstrates the potential of using instrumented grinding media for real-time characterization of mill feed and operation monitoring.

Autogenous grinding mill process control practices

This paper considers pre-requisite conditions for developing grinding mill control systems and describes the main problems of control and existing constraints. The author selected a set of grinding equipment that consists of autogenous grinding mills and ball mills. The paper considers the selected mill types as control objects. It is shown that there exists a good correlation between the mill charge and its acoustic and vibration noise. The paper examines the relationship between the level of vibroacoustic noise and the mill motor power draw and defines what can be considered an extreme dependence, as well as the conditions in which a stable operation of the automatic system can be maintained excluding the mill overload mode. The paper specifies what hardware and software means would be necessary to implement such system and describes the mill charge analyzer VAZM-1M developed by Soyuztsvetmetavtomatika that was selected for this application. The author looks at certain downsides typical of the conventional control scheme when the head mill feed rate changes as the mill motor power draw changes without allowing for changing physical and mechanical properties of the ore material. The author also considers the capability of the VAZM-1M analyzer in terms of mill load estimation accuracy. This laid the basis for developing mill protection and optimization algorithms for the AG and Ball Mill comminution circuit. The paper features a block diagram of the control algorithm and its brief description. The algorithm consists of blocks, which are responsible for the following actions: they receive key process parameters from the process control system database, check them for validity, perform initial processing and filtering; after that they analyze the trends of the measured parameters and analyze if an overload condition is probable. As decided by the mill operator, the ore flow rate can be adjusted. The paper describes a case study of running an AG mill control system on the basis of the above described algorithm and using the VAZM-1M analyzer. It is noted that this algorithm can be implemented both as an adviser for the operator and for automatic control of the mill when running in overload mode.

On the problem of control over the primary grinding mill ball charging

Considering the fact that the costs related to grinding balls and electric power account for a significant share of operational costs incurred by concentrator plants, this paper highlights the relevance of using an optimized ball charge. The paper considers a conventional approach to ball charging, which involves doing calculations on the basis of a given throughput. The latter serves as a parameter for the specific ball consumption rate approval procedure. At the same time, no consideration is given for the ball wear rate as a parameter defining the grinding performance. It is noted that the ball charge dynamics can be analyzed based on the noise produced by the mill, or the vibroacoustic parameters of the mill. The paper examines some vibro-acoustic ball charge analysis techniques utilized by both domestic and international service providers. It is noted that in all these cases the technique uses just one physical parameter, which is not enough to monitor the total mill load. The VAZM-1M analyzer developed by Soyuztsvetmetavtomatika JSC analyzes and calculates an integral amplitude of the mill vibro-acoustic field. This ensures that all components of the grinding process are taken into account. The paper takes a detailed view of the findings obtained upon analysis of the primary grinding mill ball charge. The work was carried out at the Erdenet concentrator plant as part of contractual scope. A number of different ball charge options was considered, and is it noted that none of the options can resolve this problem completely. The authors describe a possibility to monitor the mill process load with the help of the VAZM-1M analyzer. The authors also analyzed the experiments that aimed at identifying the grinding parameters governed by the mill ball charge. It is shown that the VAZM-1M analyzer gives adequate readings of any ball charge deviations. The mill spectrum registered by the VAZM-1M analyzer contains a resonance peak, which is believed to correlate as a small mill ball charge spectrum. In this regard, a plan of further research was drafted that relies on the use of the VAZM-1M analyzer. The aim is to identify an exact frequency band that would adequately correlate with the mill ball charge.

A method to identify wet ball mill’s load based on CEEMDAN, RCMDE and SRNN classification

Ball mill plays a key role in mineral processing plant, and its load identification for optimal control has great significance for the energy consumption reduction and production efficiency improvement. The vibration signal of ball mill shell contains abundant load information, which can be used to identify ball mill load. However, due to the non-linear and non-stationary characteristics of vibration signals, as well as the heavy background noises, the load identification becomes a challenging task in practice. In this paper, a novel method based on complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN), refined composite multi-scale dispersion entropy (RCMDE), and stacked recurrent neural network (SRNN) is proposed. First, CEEMDAN algorithm is used to decompose the ball mill’s vibration signals and obtain the intrinsic mode function (IMF) components. Then, the sensitive IMF components are selected through the correlation coefficient method, and the signal is reconstructed with the sensitive IMF components. Secondly, the RCMDE of the reconstructed signal is calculated to obtain the load feature vector, and the dimension of the feature vector is reduced by principle component analysis (PCA). Thirdly, the SRNN is applied to establish a load recognition model, taking the feature vector as its input and the load state as its output. The experimental results show encouraging accuracy to apply this approach to recognize the wet ball mill’s load, with a recognition rate of 98.67%.

Integrated approach for ball mill load forecasting based on improved EWT, refined composite multi-scale dispersion entropy and fireworks algorithm optimized SVM

Shell vibration signals generated during grinding have useful information related to ball mill load, while usually contaminated by noises. It is a challenge to recognize load parameters with these signals. In this paper, a novel approach is proposed based on the improved empirical wavelet transform (EWT), refined composite multi-scale dispersion entropy (RCMDE) and fireworks algorithm (FWA) optimized SVM. Firstly, vibration signals are denoised by improved EWT, which uses cubic spline interpolation to calculate envelope spectrum for segmentation. Then, RCMDEs of the denoised signals are calculated as feature vectors. The vectors’ dimensionalities are reduced by principal component analysis (PCA). Finally, a mill load prediction model is established based on the FWA optimized SVM. The reduced feature vectors are fed to the model, thus material-to-ball ratio and filling rate being outputs. Grinding experiments show that the extracted features by RCMDE can effectively distinguish three load states. Meanwhile, experiments also show that FWA reduces the forecasting errors of material-to-balls ratio and filling rate by 1.9% and 2.9% compared with genetic algorithm (GA), as well as by 1.92% and 4.21% compared with particle swarm optimization (PSO) algorithm. It demonstrates that the proposed approach for ball mill load forecasting has high accuracy and stability.

Analysis on voltage characteristics of distribution network considering rolling mill load and Research on control measures

In view of the voltage fluctuation and flicker of distribution network caused by rolling mill type typical load in some distribution networks and the reduction of power quality of power grid, it is necessary to analyze the voltage characteristics of distribution network, and adopt an effective governance measure to suppress harmonics in distribution network and improve power quality of power grid. The mathematical models of several typical loads corrected based on the measured data are established, and the influence radiation range of the corrected rolling mill load model on the distribution network voltage is analyzed, and a SVC (static var compensator) governance measure with SOGI phase-locked strategy is adopted for governance. Finally, combined with the actual operation parameters of a substation of State Grid, the power system simulation model is built by using simulation software, and the effectiveness of the governance measures adopted in this paper is verified by simulation analysis.

Mill Load Identification Method for Ball milling Process Based on Grinding Signal

Mill Load Identification Method for Ball milling Process Based on Grinding Signal

State Recognition of Mill Load Based on Improved K-means Clustering Algorithm

In this work, a new identification method for the load identification of ball mill based on improved K-means clustering is proposed. Since the external device response signals of ball mill are closely related to its load during the operation, the appropriate parameters are selected. The number of clusters and initial clustering centers are meliorated based on their density and genetic algorithm. The dimensionality of input features are reduced with kernel principal component analysis (KPCA), and the final clustering centers are obtained by iterating in the space with reduced dimensionalities. Tests on datasets have demonstrated the higher accuracy and stability of our improved K-means method compared with other methods. The case-analysis results show that the mill load status can be identified clearly according to the displacement variations of cluster centers calculated by this method. Our research provides a new reference for mill load identification in the viewpoint of multi-source information fusion.