Research of flotation dosing control method based on SVM and image processing technology

• Published in: Proceedings of the 31st Chinese Control Conference pp. 7067 - 7070
• Main Authors: Wang YuKun, Zhang Yong, Wang JieSheng, Chen Ying
• Format: Conference Proceeding
• Language: Chinese; English
• Published: IEEE 01-07-2012
• Subjects: Abstracts; Accuracy; Educational institutions; Electronic mail; Flotation Dosing; Froth image; Image processing; Mathematical model; Support vector machines; SVM; Texture characteristic
• ISBN: 1467325813; 9781467325813
• ISSN: 1934-1768; 2161-2927

Flotation process is nonlinear, strong coupling and large time delay. It is difficult to control the flotation dosage precisely. The concentrate grade fluctuations are large and the product quality is poor. A Flotation dosing intelligent control methods is proposed based on the actual production of the concentrator. Using SVM technology to establish dosing mathematical model to determine the initial flotation dosage quantum; by extracting features parameters of flotation froth image to determine the flotation state, and combined with the dosing mathematical model adjust flotation dosage quantum, improve control accuracy. The simulation result shows that the method can effectively improve the control accuracy of flotation reagent and stable the concentrate grade.