Study On Intelligent Control System For A Mixed Separation Process

China has rich source of hematite ore. Due to the characteristics of low grade, fine-grained and non-homogenous distribution particles of hematite ore, the mixed separation process which consist of grinding process, thickening process and flotation process, must be adopted so as to remove the gangue effectively and obtain higher concentrate grade. The mixed separation process is the key production processes and its aim is to improve concentrate grade and lower tailing grade. In this context, the concentrate grade and the tailing grade represent the actual metal composition in usable ores, and they are key technical indexes for the product quality and production efficiency in mixed separation process. The higher concentrate grade and lower tailing grade are important to improve product quality, reduce energy consumption, and increase enterprise economic benefit.The system dynamics between the technical indexes and the slurry levels of flotation machine are subjected to strong nonlinearity and strong coupling, which is difficult to describe using mathematical model. Such technical indexes vary with the production boundary conditions such as underflow slurry concentration, underflow slurry flow-rate, particle size of feed ore. The flotation process consists of three flotation machines connected in a cascade way. A control action implemented at any slurry level of flotation machine tends to be transmitted to both upstream and downstream units. The control of flotation machine should not only consider the inlet and outlet flow rates, but also pay attention to the strong coupling and strong nonlinearity. Therefore the optimal control of technical indexes cannot be realized using the conventional control method. The mixed separation thickening process of hematite beneficiation, with underflow slurry pump speed as input, underflow slurry flow-rate as output of inner loop, and underflow slurry concentration as output of outer loop, is a nonlinear cascade process that is difficult to model and subject to large random disturbances generated from the flotation middling. It requires the outer-loop output, the inner output and its rate of change to be within their target ranges simultaneously. Therefore, the existing cascade control method is difficult to be applied in the hematite thickening process. In industrial process, manual operation on the mixed separation is mainly used. However on-site operators cannot timely and accurately recognize the operational conditions, and cannot adjust the outlet valve opening of flotation machine and underflow slurry pump speed. Such a manual control would cause the concentration and the flow-rate of slurry often exceeding their target ranges and then affect the final concentration grade and the metal recovery rate.The work conducted in this thesis was supported by National Science and Technology Support Program of China (2012BAF19G01) of "Advanced control technology of mineral processing". The objective of this thesis is to control the concentrate grade and tailing grade within their required ranges and in the meantime to improve concentrate grade and reduce tailing grade. This thesis focuses on the research on the intelligent control system for the mixed separation process. An intelligent control method is proposed. An intelligent control software based on above intelligent conrol method is developed. This intelligent control software has been successfully applied in Jiugang mineral processing factory in China. The main research of this paper concludes the following contributions:1. For the comprehensive complexity of the mixed separation process, an intelligent operation control method which consists of setting control of flotation machine levels and the loop control of flotation machine levels is proposed. Among them, the setting control of flotation machine levels contains case-based reasoning based presetting model, a soft measurement model for concentrate grade and tailing grade, a feedforward compensation model and a feedback compensation model. The loop control of flotation machine level contains linear adaptive decoupling controller, nonlinear dcoupling controller and switching machnism. The linear adaptive decoupling controller can ensure the stability of the inputs and outputs. And the nonlinear adaptive decoupling controller can improve transient performance of the control system.2. For the above strong nonlinear cascade process that is difficult to model and subject to large random disturbance, by combining the fuzzy control, rule-based reasoning, an interval cascade intelligent switching control method which consists of the presetting based upon steady state model, the compensator based upon fuzzy reasoning, the maintainer and the switching mechanism based upon rule-based reasoning is proposed. Among them, the preset ing module generates the initial set-point for the slurry flow-rate. The compensator generates the compensation value of the flow-rate set-point using the errors of concentration and flow-rate as the inputs. The switching mechanism realizes the effective switching between the compensator and the maintainer using the error of concentration and its rate of change.3. The intelligent control software based on the intelligent control method of the mixed separation process has been designed. The intelligent control software consists of the intelligent operation control software, the intelligent switching control for the intervals of concentration and flow-rate of underflow slurry, the process control software, and the monitoring software.4. The intelligent control system of mixed separation process has been applied in Jiugang mineral processing factory in china. The experimental contrast of intelligent operation control method with manual operation has shown that the concentrate geade and tailing grade is improved with 0.31% and 2.18% respectively. For the control of flotation machine slurry levels, two kinds of experimental contrast are adopted, containing the disturbance of feed pulp and the set-point changes. The experiment results have shown that control performances of the proposed method greatly excel the manual operation. For the control of the underflow slurry concentration and flow-rate, the experimental of the large random disturbance is adopted. The experimental contrast results show the proposed intelligent switching control method of USC and USF is superior to the manual control. The concentrate grade is improved with 0.13%.