Abstract:Due to the unknown system structure of the froth flotation process and frequent fluctuations in production conditions, design of control strategy is a challenging problem. As a result, manual operation is still widely applied in practice by observing froth image features. However, since the manual observation is subjective and the production conditions are time-varying, the manual operation cannot make decisions quickly and accurately. In this paper, a data-driven-based adaptive fuzzy neural network control strategy is developed to implement the automatic control of the antimony flotation process. The strategy is composed of fuzzy neural network (FNN) controllers, a data-driven model, and an on-line adaptive algorithm. The FNN is constructed to derive the control laws of the reagent dosages. The parameters of the FNN controllers are tuned by gradient descent algorithm. To obtain the real-time error feedback information, the data-driven model is established, which integrates the long short term memory (LSTM) network and radial basis function neural network (RBFNN). The LSTM network is utilized as a primary model, and the RBFNN is used as an error compensation model. To handle the challenges of the frequent fluctuations in the production conditions, the on-line adaptive algorithm is proposed to tune the parameters of the FNN controllers. Simulations and experiments are carried out in a real-world antimony flotation plant in China. The results demonstrate that the proposed adaptive fuzzy neural network control strategy produces better control performance than the other two existing methods.

Flotation process with model free adaptive control

Fuzzy Association Rule-Based Set-Point Adaptive Optimization and Control for the Flotation Process

Data-driven Adaptive Modeling Method for Industrial Processes and Its Application in Flotation Reagent Control.

Motion Control of Docking Auv in Currents

Adaptive nonlinear model predictive control for a class of multivariable chemical processes

Fuzzy Adaptive Generalized Predictive Control for Wastewater Treatment Process

Data-Driven Adaptive Optimal Control for Flotation Processes with Delayed Feedback and Disturbance

Nonlinear adaptive control based on RBF networks and multi-model method

Learning-Based Adaptive Optimal Control for Flotation Processes Subject to Input Constraints

Flotation state classification applying dynamic textures analysis

Optimal Reagents Control for Flotation Processes: an Adaptive Dynamic Programming Approach

Data-driven-based Adaptive Fuzzy Neural Network Control for the Antimony Flotation Plant

Deep Learning Feature-Based Setpoint Generation and Optimal Control for Flotation Processes.

Machine-vision-based Online Measuring and Controlling Technologies for Mineral Flotation — A Review

Ph Control in Flotation Process Based on Prediction Model

A Fuzzy Control Approach Based On Multi-Sensor Management And Its Application

Dual-Rate Operational Optimal Control for Flotation Industrial Process With Unknown Operational Model

Reagent Predictive Control Using Joint Froth Image Feature for Antimony Flotation Process

Fuzzy Association Rule Based Froth Surface Behavior Control In Zinc Froth Flotation

Flotation classification approach by controlling interactions between particles

Toward Flotation Process Operation-State Identification via Statistical Modeling of Biologically Inspired Gabor Filtering Responses