Mineral particle size is an important parameter in the mineral beneficiation process. In industrial processes, the grinding process produces pulp with qualified particle size for subsequent flotation processes. In this paper, a hierarchical intelligent control method for mineral particle size based on machine learning is proposed. In the machine learning layer, artificial intelligence technologies such as long and short memory neural networks (LSTM) and convolution neural networks (CNN) are used to solve the multi-source ore blending prediction and intelligent classification of dry and rainy season conditions, and then the ore-feeding intelligent expert control system and grinding process intelligent expert system are used to coordinate the production of semi-autogenous mill and Ball mill and Hydrocyclone (SAB) process and intelligently adjust the control parameters of DCS layer. This paper presents the practical application of the method in the SAB production process of an international mine to realize automation and intelligence. The process throughput is increased by 6.05%, the power consumption is reduced by 7.25%, and the annual economic benefit has been significantly improved.

The process operating performance assessment (POPA) of froth flotation under multimode is an important means to improve production efficiency and product quality. In the plant, froth flotation presents multimode characteristics due to the uncontrollable of the objective conditions of production. In this article, a method based on graph representation learning is studied to solve the POPA problem of a multimode froth flotation process. Based on the qualitative and quantitative information represented by the froth images and process variables of the flotation process, an appropriate convolutional neural network is selected to fuse the multisource heterogeneous information. The powerful representation capabilities of deep clustering methods are used to divide working modes. A deep embedded graph clustering network is proposed to learn the node representation, on which the clustering effect of the model is improved by reconstructing the topological structure and node content of the graph data. In addition, a dual self-monitoring mechanism is designed to guide the model parameter update. Finally, a soft recognition method is designed to fuse the performance grades under different modes by using Bayesian fusion theory, which improves the accuracy of assessment. The proposed method is applied to a gold flotation process to illustrate the feasibility of the method.

Operating performance assessment of gold flotation process plays an important role in improving the metallurgical performances and pursue the best comprehensive economic benefits. The appearance of froth is a good indicator of flotation performance, however, labeling the data is costly. To assess the operational performance of flotation process with a small amount of labeled data, this study proposes a semi-supervised cluster generative adversarial network (SSClusterGAN). First, latent variables are sampled from one-hot encoded variables and continuous normal distribution variables. While projecting all data into the latent variable space, the specific types of clustering in the latent space is realized by controlling the type of distribution of the learned latent codes, so that the training objectives of labeled data and unlabeled data are consistent. Then, by adapting the generative adversarial networks (GAN) framework, a pair of stacked discriminators are used to learn the conditional distribution of attributes for labeled and unlabeled data, respectively. In addition, using semi-supervised learning (SSL) as the basic learning technology, the high-confidence pseudo-labels obtained by training with unlabeled data are combined with the labeled training data to increase the training samples and effectively improve the accuracy of the model. Finally, the method was applied to the gold flotation process. The experimental results show that the method effectively utilizes unlabeled samples to expand the labeled dataset, combines the advantages of semi-supervised learning and GAN, obtains accurate assessment results, and effectively improves the comprehensive economic benefits of the gold flotation process.

The process operating performance assessment (POPA) of the flotation process is of great significance for monitoring the flotation operating performance and improving the product quality. Flotation process is a typical industry process with multiple subprocesses, and the subprocesses are interconnected. In addition, due to the feed characteristics, the flotation process presents various working modes, and the standards of operating behavior under various modes are different. Therefore, the POPA problem of the plant-wide flotation needs to take into account the coupling relationship of the subprocesses under various working modes. In this article, a novel distributed multigraph deep embedding graph clustering (DMG-DEGC) network is proposed for the flotation process, which uses graph representation to fully mine the structural features related to data and process. A distributed structure is adopted based on the multiple subprocesses characteristics of the plant-wide process. Among them, the strong influence subprocess is used for the division of multiple working modes. The adjustment of several variables in the weak influence subprocess indirectly affects the performance, and the variable augmentation method is used to guide the performance assessment of the subprocess. DMG-DEGC uses spatial domain, temporal domain, and prior knowledge to construct multigraph. The information between multigraph is aggregated, and the consistency and complementarity information between multigraph is retained. Finally, the proposed method is applied to the actual flotation process to illustrate its effectiveness and superiority.

The grinding process is a key link in mineral processing production and a typical complex, controlled process. The steady-state model is limited by its model structure and thus difficult to applyied in a control system. A hybrid modeling method driven by multi-source compensation is proposed in this paper based on the mechanism model using key equipment in the grinding and classification process, addressing the uncertainties which affect the stability of the control systems. This method combines the relevant multi-source signals with uncertainties by using a priori knowledge, extracts the nonlinear feature vector in the signal through an unsupervised depth network, and constructs a compensation model based on dynamic radial basis function network to realize the integration of mechanism modeling and data-driven compensation modeling. The simulation results show that the model has a high degree of fit to the real physical system; the industrial validation was conducted at a gold concentrator, the grinding product quality was predicted and controlled with the model.

The accurate and efficient modeling of the overall process of complex industrial processes is an essential step in the automation and intelligence of such processes. Nevertheless, in today’s Internet era, the environment of complex industrial processes is complicated and variable, the data sources and structures are more complex, and the industrial equipment used in similar complex industrial processes is not guaranteed to be free of differences. This problem makes it difficult to construct models for centralized method training. In order to solve the problem of how to use multisource complex industrial process data to assist new process modeling in the industrial Internet while avoiding data leakage, this article proposes a building block modeling method for complex industrial processes based on federated learning. In particular, this article establishes a cloud data center, delegating multiple servers corresponding to a typical link, each of which owns a specific model of a typical link, and utilizes federated learning to enable the typical links under scattered similar complex industrial processes to collaboratively train the typical links’ models, thereby building a global model of typical links of complex industrial processes with high generalizability. Next, the local base client implements local modeling of typical links using long short-term memory (LSTM) network. Subsequently, based on the construction of a global model of typical links, typical links’ models are locally fine-tuned and linked according to the new process, thus enabling the individual and rapid construction of the required complex industrial process models. Finally, the effectiveness of this method is verified by the coal slurry flotation process and dense medium cyclone process.