为提高油水两相流含油率的测量精度，提出基于最小二乘支持向量机(least squares support vector machine，LSSVM)和改进的粒子群算法(particle swarm optimization，PSO)的含油率建模方法。该方法将测量的油水总流量和加热器上下游温差作为LSSVM输入，含油率作为输出，对含油率与温差和总流量的关系进行训练，通过改进的PSO优化LSSVM的参数，建立了含油率的优化模型，并用测试数据对含油率的模型进行了比较。实验结果表明，基于改进的 PSO-LSSVM含油率模型比PSO-LSSVM和遗传算法–最小二乘支持向量机模型运算速度快，比理论修正模型测量精度高，含油率在4%~60%时，平均测量误差为0.93%。

In order to improve the measurement accuracy of oil holdup of oil-water two phase flow, a modeling methodology of oil holdup based on least squares support vector machine (LSSVM) and modified particle swarm optimization(PSO) was proposed. The measured total flux of oil-water and temperature difference between upstream and downstream of the heater were employed as inputs, and the oil holdup was used as output of LSSVM, the relation between input and output was trained and the ideal model of oil holdups was obtained by applying a modified PSO to determine the optimal parameters of LSSVM automatically. The Modified PSO-LSSVM model was compared with genetic arithmetic (GA)-LSSVM model and theory amendatory model using test datum. The experimental results indicate that the modified PSO-LSSVM model performs better than the other model in term of measurement accuracy and calculating speed. The average measurement errors is 0.93% in the range of 4% to 60% oil holdup.

国家863高技术基金项目(2007AA04Z180，08JCYBJC 11800)。