基于改进型LS-SVM技术的煤泥浮选智能优化控制方法
2022,30(12):119-124
摘要：针对现有的煤泥浮选控制算法公式复杂、评估时间长等问题,提出了一种基于多重最小二乘支持向量机(Least Squares-Support Vector Machine,LS-SVM)的浮选精煤灰分综合评估模型。首先,建立了基于LS-SVM的单一煤种的单一估计模型,并利用引力搜索算法对其内部参数进行了优化。其次,设计了模型更新策略,解决了单一模型精度下降的问题。此外,为了解决模型失配问题,还研究了由多个单一模型组成的多个LS-SVM模型以及模型切换机制。最后,进行了工业试验和评价,评估值与实际值的平均相对误差为3.32%,综合模型的估计精度和适应性能够满足工业要求。
关键词：煤泥浮选;最小二乘支持向量机;引力搜索算法;综合评估模型;洁净煤灰含量
Intelligent Optimization Control Method Of Slime Flotation Based On Improved LS-SVM Technology
Abstract：Aiming at the problems of complex formulas and long evaluation time of existing slime flotation control algorithms, a comprehensive evaluation model of ash content of flotation cleaned coal based on multiple least squares support vector machine (LS-SVM) is proposed. Firstly, a single estimation model of single coal type based on LS-SVM is established, and its internal parameters are optimized by gravity search algorithm. The first mock exam is to update the model updating strategy, which solves the problem of the single model"s accuracy degradation. In addition, in the first mock exam, we also study several LS-SVM models composed of multiple single models and model switching mechanism. Finally, the industrial test and evaluation are carried out. The average relative error between the evaluation value and the actual value is 3.32%. The estimation accuracy and adaptability of the comprehensive model can meet the industrial requirements.
Key words：Slime flotation; Least squares support vector machine; Gravity search algorithm; Comprehensive evaluation model; Clean coal ash content
收稿日期：2022-01-28
基金项目：
