支持向量机在压滤脱水过程优化中的应用

THE APPLICATION OF SUPPORT VECTOR MACHINE IN DEWATERING PROCESS OPTIMIZATION OF PRESSURE FILTERING

为提高自动压滤机在精矿脱水过程中的作业效率,提出了压滤脱水过程控制参数的优化。研究了自动压滤机脱水过程的优化机理。将SVM(支持向量机)方法应用到压滤脱水过程的建模仿真研究中,并建立了压滤脱水过程的SVM模型。在此基础上提出了一套"循序寻优"的压滤脱水过程控制参数寻优方法。结果表明,采用支持向量机方法建立的工业压滤脱水过程仿真模型仿真精度高,对水分和处理能力的仿真相对误差分别是1.57%和3.81%;利用"循序寻优"方法获得的工业压滤脱水过程最优控制参数,不但可以保证生产指标的稳定,而且将压滤周期缩短到了原来的85%以下。

In order to obtain the higher operating efficiency of concentrate pressure filtering, the optimization of pressure filter dewatering process control parameters was proposed. The principle and optimization-mechanism of pressure filter dewatering processes were studied. Support Vector Machines （SVM） was studied to simulate the dewatering process of pressure filter, and the pressure filter dewatering SVM simulation models were established. A ＂progressive optimization＂ pressure filter dewatering control parameters optimization methods was proposed. The results show that the accuracy of pressure filtering industry SVM models is high, and the simulation relative errors of cake moisture and capacity are 1.57% and 3.81%. Using the control parameters obtained by ＂progressive optimization＂ method. It can not only guarantee the stability of operating data, but also shorten the period of per pressure filtering cycle lower to 85% of the previous.