基于FOA-GRNN油井计量原油含水率的预测  

Application of FOA-GRNN to Prediction of Moisture Content in Crude Oil of Wellheat Metering

作　　者：刘翠玲[1] 张路路[1] 王进旗[2] 李亮亮[1] 

机构地区：[1]北京工商大学计算机与信息工程学院,北京100048 [2]东北石油大学地球科学院,黑龙江大庆163318

出　　处：《计算机仿真》2012年第11期243-246,259,共5页Computer Simulation

摘　　要：研究原油含水率准确预测问题,提供高精度的原油含水率数据在油井计量中具有重要意义。针对原油含水率预测受到多因素影响,由于原油中存在复杂的非线性关系,传统的预测方法无法满足预测精度要求。为了提高原油含水率的预测精度,提出了果蝇优化广义回归神经网络的原油含水率预测方法,果蝇优化算法用于广义回归神经网络的参数优化。通过同轴线相位法含水率计的测量系统对原油含水率有影响的多个参量进行测定,建立果蝇算法优化广义回归神经网络的原油含水率预测模型。仿真结果表明:相对于广泛应用的BPNN预测模型,果蝇算法优化的广义回归神经网络预测精度高,是一种实用有效的原油含水率预测方法。Accurate prediction in crude oil moisture content and the high accuracy data of crude oil moisture con- tent has vital importance in oil well measurement. The moisture content prediction can be affected by many factors, and there is complicated nonlinear relationship between the prediction and factors, so the traditional forecast methods can not meet the requirements of prediction accuracy. In order to improve the prediction precision, a crude oil mois- ture content measure method named fruit fly optimization algorithm of the generalized regression neural network fore- cast method was put forward, which was used for generalized regression neural network parameters optimization. By determing several parameters which can affect crude oil moisture content using the coaxial line phase indicator meas- urement system, we established fruit fly optimization algorithm of the generalized regression neural network predicting model of crude oil moisture content. Simulation and experimental results show that： relative to the widely used Back Propagation Neural Network（BPNN） prediction model, the ffruit fly optimization algorithm of the generalized regres- sion neural network, has higher prediction accuracy. It is a practical and effective prediction method of crude oil moisture content.

关 键 词：油井计量 原油含水率 果蝇算法 广义回归神经网络 预测模型 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]