摘要
本文利用12电极ECT系统提供的电容测量信息,基于主成分回归(PCR)和偏最小二乘回归(PLSR)技术,提出了一种油气两相流空隙率测量的新方法。该方法针对不同的油气两相流流型,分别基于PCR和PLSR建立两组空隙率测量模型。实际测量时,根据流型辨识结果分别选用相应的基于PCR的空隙率测量模型和基于PLSR的空隙率测量模型,通过计算获得空隙率值并进行对比。研究结果表明,本文提出的空隙率测量方法是有效的,两组空隙率测量模型的测量结果精度相当,在典型流型下的测量误差均在5%以内,但基于PLSR技术所获空隙率测量模型的降维效果优于基于PCR技术的相应模型。同时,由于该方法直接利用ECT系统提供的电容测量值来建立空隙率测量模型,省去了以往ECT系统测量空隙率所必需的复杂费时的图像重建过程,提高了空隙率测量的实时性,整个测量时间小于0.1s。
A new method is proposed for the voidage measurement of gas-oil two-phase flow, based on principle component regression （PCR） and partial least squares regression （PLSR）, with the capacitances from 12-electrode Electrical Capacitance Tomography （ECT） system. The PCR and PLSR are used respectively to establish two groups of regression model between the capacitances and the voidage of the two-phase flow. In the voidage measurement process, according to the flow pattern identification result, a relative PCR based model and a relative PLSR based model are selected to calculate the voidage, and the voidage measurement results of these two groups of models are compared. Re- search work results show that the proposed method is effective. The maximum measurement errors are both less than 5%. Nevertheless, the dimensionality reduction of PLSR is more effective than that of PCR. The research work also shows that the real-time performance of the proposed voidage measurement method is satisfactory, because it implements the voidage measurement without the time-consuming image reconstruction process. The total voidage measurement time is less than 0.1s.
出处
《仪器仪表学报》
EI
CAS
CSCD
北大核心
2008年第4期705-710,共6页
Chinese Journal of Scientific Instrument
基金
国家自然科学基金(50576084
60532020)资助项目
关键词
两相流
空隙率
电容
主成分回归
偏最小二乘回归
two-phase flow
voidage
capacitance
principle component regression
partial least squares regression
