一种基于改进的ANFIS橡胶聚合转化率软测量模型  

A Soft Sensor Model for Rubber Polymerization Conversion Based on Improved ANFIS

作　　者：高世伟[1] 赵力[2] 

机构地区：[1]兰州石化职业技术学院电子电气工程学院,甘肃兰州730060 [2]中国石油兰州石化公司自动化研究院,甘肃兰州730060

出　　处：《特种橡胶制品》2017年第5期43-47,共5页Special Purpose Rubber Products

基　　金：国家自然科学基金资助(批准号00000000)

摘　　要：针对合成橡胶聚合转化率难以进行有效在线测量的问题,提出一种基于自适应神经模糊推理系统ANFIS的转化率软测量预测模型。模型引入了闭环反馈信息,提高了模型的收敛速度和准确度。利用SBR装置的现场数据进行仿真,结果表明,提出的软测量模型能实现转化率的在线预测,可以满足工业现场的生产要求,对于实现装置的优化运行具有重要的意义。In order to solve the problem that it was difficult to measure the polymerization conversion of synthetic rubber effectively,a soft sensor model based on Adaptive Network-based Fuzzy Inference System was proposed.The closed-loop feedback information was introduced to improve the convergence speed and accuracy of the model.Simulation was carried out using field data of styrenebutadiene rubber plant.The simulation results showed that the proposed soft sensor model could realize the on-line prediction of the conversion,which could meet the requirements of industrial production,and had great significance for the realizing the optimal operation of the device.

关 键 词：SBR 聚合转化率 模糊逻辑系统 神经网络系统 神经模糊系统 

分 类 号：TQ330.493[化学工程—橡胶工业] TQ333.1