摘要
采用多圈环形管用于气液两相流参数的测量,对环形管上升段水平方向内外侧差压波动信号进行了分析,采用无因次分析方法获得与差压波动信号均方根相关的特征量,得到了此特征量与容积含气率的关系模型,并在此基础上进行了实验。利用支持向量机优良的非线性映射和强大的泛化能力,建立了一个基于最小二乘法支持向量机的含气率软测量模型,给出了相应的系统结构和算法,针对LS-SVM方法参数选取困难的特点,采用遗传算法进行优化,以提高软测量的精度。仿真和实际运行结果表明,基于LS-SVM的气液两相流含气率软测量模型具有较高的估算精度与泛化能力,为气液两相流含气率的测量提供了一种简单、可靠的新方法。
A multi-turn annulus tube was used for measuring the parameters of a gas-liquid two-phase flow,and the pressure difference fluctuation signals at both inside and outside of the riser section of the annulus tube along the horizontal direction were analyzed.A non-dimensional analytic method was used to obtain the characteristic value related to the mean square root of the pressure difference fluctuation signals and a model governing the relationship between the characteristic value in question and the volumetric gas content.On this basis,an experiment was performed.By making use of the excellent nonlinear mapping and strong generalization capacity of a supportive vector machine,a model for softly measuring the gas content was established based on the least square supportive vector machine and corresponding systematic configuration and algorithm were given.In the light of the specific feature that it is difficult to choose parameters by using the LS-SVM method,the genetic algorithm was used to perform optimization to enhance the precision of the soft measurement.The simulation and practical operation results show that the model for soft measurement of the gas content in a gas-liquid two-phase flow based on the LS-SVM has a relatively high estimation precision and generalization capacity,thus providing a new simple and reliable method for measuring a gas content in a gas-liquid two-phase flow.
出处
《热能动力工程》
CAS
CSCD
北大核心
2011年第1期58-62,122-123,共5页
Journal of Engineering for Thermal Energy and Power
基金
航空科学基金资助项目(2008ZC03005)
关键词
软测量
最小二乘法支持向量机
遗传算法
两相流
差压波动信号
含气率
soft measurement,least square supportive vector machine,genetic algorithm,two phase flow,pressure difference fluctuation signal,gas content
