基于改进SFLA-LSSVM的矿浆管道临界淤积流速预测

发布时间：2018-05-20 05:03

本文选题：临界淤积流速 + 矿浆管道　；参考：《云南大学学报(自然科学版)》2017年01期

【摘要】：由于矿浆管道的临界淤积流速受被输送物料的性质、浆体的物理特性等因素的影响,计算较为复杂,现有各种类型的经验计算公式无论在适用面还是精度上都难以满足实际需求.为此,引入了改进的混合蛙跳算法(SFLA)优化最小二乘支持向量机(LS-SVM)的新方法对临界淤积流速值进行了预测.针对该算法容易陷入早熟收敛和局部最优等问题,所改进的混合蛙跳算法通过参考PSO算法的更新方式,引入了非线性递减的惯性权值.此外,还融合了基于平均值思想的局部搜索更新策略和基于群体适应度方差的局部最优的判定方式.仿真结果表明该算法所取得的预测效果优于常规的方法,同时与所选的临界淤积流速的经验公式相比有着较高的精度.
[Abstract]:Because the critical deposition velocity of the slurry pipeline is affected by the properties of the material to be transported and the physical characteristics of the slurry, the calculation is more complicated. All kinds of empirical formulas are difficult to meet the practical requirements in terms of application surface and accuracy. For this reason, a new method of optimizing LS-SVM based on the improved hybrid leapfrog algorithm (SFLAs) is introduced to predict the critical deposition velocity. In order to solve the problem of premature convergence and local optimization the improved hybrid leapfrog algorithm introduces the nonlinear decreasing inertia weight by referring to the updating method of PSO algorithm. In addition, the local search update strategy based on average value and the local optimal decision method based on the variance of population fitness are combined. The simulation results show that the prediction effect of this algorithm is better than that of the conventional method, and it has a higher accuracy than the empirical formula of critical deposition velocity.
【作者单位】：昆明理工大学信息工程与自动化学院;云南省矿物管道输送工程技术研究中心;
【基金】：国家自然科学基金(51169007) 云南省科技计划(2013DH034)
【分类号】：TD50;TP18

【相似文献】