窄矩形通道两相流动沸腾压降特性实验研究

Experimental Study on Two-phase Flow Boiling Pressure Drop Characteristics in Narrow Rectangular Channel

实验研究了横向均匀和非均匀(多项式和正弦分布)加热条件下垂直矩形通道(2 mm×60 mm×1000 mm)的沸腾压降特性,实验段为双面加热,有效加热面尺寸为56 mm×700 mm。工作流体为去离子水,通过改变入口压力和流量边界开展不同参数工况下的实验研究。结果表明,两相压降梯度随饱和压力的增加而减小,随质量流速的增加而增大,含气率对两相压降的影响与质量流速有关,横向功率分布形式对流动沸腾压降也有重要影响。基于均匀加热实验数据对现有的两相压降预测模型进行了评价,发现使用等效黏度假设的均相模型极大低估了实验值,且预测结果的分散度较大;分相模型中Müller-Steinhagen和Heck、Li和Wu关系式预测效果最好,平均绝对误差分别为11.8%和12.3%,且大多数预测值在±20%误差带内。本文基于Müller-Steinhagen和Heck关系式形式引入邦德数Bo考虑表面张力的影响,拟合得到新的预测关系式,该关系式对实验数据的预测误差在±8%的误差范围内。

The flow boiling pressure drop characteristics of a vertical rectangular channel(2 mm×60 mm×1 000 mm) were studied under transverse uniform and non-uniform heating. The test section was double-sided heating with an effective heating area of 56 mm×700 mm. Three kinds of heating power distributions(uniform, polynomial and sinusoidal) were selected, and the working fluid was deionized water. A wide range of operating conditions was obtained by varying the inlet pressure and mass flux. The results show that the pressure drop gradient decreases with the increase of saturation pressure and increases with mass flux. The influence of vapor quality on two-phase pressure drop is related to mass flux, and the transverse power distribution has an important effect on the flow boiling pressure drop. Based on the experimental data of uniform heating, the existing two-phase pressure drop prediction model was evaluated. The homogeneous model using the equivalent viscosity assumption significantly underestimated the experimental values, and the data dispersion is relatively large. The Müller-Steinhagen and Heck and Li and Wu formulas are found to be the best with mean absolute errors of 11.8% and 12.3%, and most of predicted values are within ±20% error band. In this study, a new correlation based on the form of Müller-Steinhagen and Heck formula was proposed, and the Bond number Bo was introduced to consider the effect of surface tension. The prediction error of the new correlation for experimental data is within 8% error band.