基于遗传支持向量机流量软测量研究

Soft Measurement of Flow Based on Genetic Algorithm Support Vector Machine

针对中国石油宁夏石化公司离心式CO2压缩机四段出口动态流量准确测量困难的问题,引入软测量方法对动态流量进行测量,实现了基于支持向量机的软测量技术对流量的间接测量.应用遗传算法对支持向量机的参数进行了优化,并将参数优化后的支持向量机得到的体积流量软测量模型的预测效果与实际测量值进行了对比,取得了良好的效果.

The paper provides an indirect measurement method of the flow, which is the soft sensor technique based on support vector machines for the measuring not accurately of the flow of the CO2 compressor for PetroChina Ningxia Petrochemical Company. The genetic algorithms and the particle swarm optimization, which belong to heuristic optimization algorithms, are provided for parameters optimization of support vector machines. The effects of soft sensor model of the measurement of flow, which are gained by using these two algorithms for optimizing the parameters of the support vector regression machine, are compared with the prediction result of the traditional non -heuristic cross validation algorithms.