最小二乘支持向量机在重质燃料油调合中的应用  

Application of Least Squares Support Vector Machine in Heavy fuel Oil Blending

作　　者：刘名瑞[1] 肖文涛[1] 张雨[2] 李遵照[1] 薛倩[1] Liu Mingrui Xiao Wentao Zhang Yu Li Zunzhao Xue Qiant(Sinopec Fushun Research Institute of Petroleum and Petrochemicals, Fushun, 113001, China Sinopec Fuel Oil Sales Co. Ltd. , Beijing, 100029, China)

机构地区：[1]中国石油化工股份有限公司抚顺石油化工研究院,辽宁抚顺113001 [2]中国石油化工股份有限公司燃料油销售有限公司,北京100029

出　　处：《石油化工自动化》2017年第1期33-36,共4页Automation in Petro-chemical Industry

摘　　要：为解决重质船用燃料油调合中非线性指标预测准确度低的问题,提出了基于最小二乘支持向量机(LSSVM)的非线性指标预测模型的应用。介绍了LSSVM的基本算法以及构建了预测模型,以重质燃料油调合中黏度和密度指标为例,对比一般计算模型,进行指标数据的计算。应用结果表明:基于LSSVM的预测模型,通用性强、预测精度高、自适应能力强,具有较高的应用前景。To solve low accuracy problem of nonlinear predictor in heavy fuel oil blending, application of non-linear index predictive model based on least square support vector machine （LSSVM） is proposed. Basic algorithms of LSSVM is introduced. Predictive model is constructed. Index data is calculated with viscosity and density as case and with comparison with general computing model. Results indicate the predictive model based on LSSVM is of strong universality, high prediction accuracy and strong adaptability. It has a significant future prospects.

关 键 词：最小二乘支持向量机 非线性 重质燃料油 调合 

分 类 号：U677[交通运输工程—船舶及航道工程] TP18[交通运输工程—船舶与海洋工程]