【作者】 刘凯；

【导师】 楚武利；

【作者基本信息】 西北工业大学 ， 流体机械及工程， 2019， 硕士

【摘要】 而今,为朝着更高压比的方向发展,压气机级负荷逐渐提高,由此带来的是压气机内部的复杂多变的二次流结构,尤以叶根与叶尖端区的复杂流动现象更为突出。而近几年的研究发现,叶根区域的角区分离现象,不但能造成大程度的能量损失,还将诱发叶根失速。因此,本文对某低速叶栅及某通风机静叶,借助CFD及拓扑分析手段,就其叶根角区分离的原因以及损失分布进行探究。同时,结合目前叶片表面粗糙度研究热点,研究了叶片表面粗糙度对角区分离的程度以及损失的影响。针对某低速风机,分析了其内部二次流结构,并借助优化手段,通过调节叶片不同截面的安装角,以优化其内部流动,提升静压效率。主要研究内容如下:(1)借助CFD手段,研究该扩压叶栅内的角区分离现象。通过对壁面极限流线的拓扑分析,研究了该叶栅内的奇点以及相应的分离线和附着线分布,探究了角区内各涡系的发展形成过程,并借助三维流线分析提出了通道涡产生的三源结构,同时建立了对应的扩压叶栅涡结构模型。通过以上分析提出建立了通道涡的位置分离方法,探究了通道涡的发展特征,并借助耗散函数对角区分离下的叶栅内的损失展开探究。(2)为了探究叶表粗糙度变化是否会促进角区分离的产生,以及粗糙度变化对压气机内损失类型的影响,借助CFX商用软件对某低雷诺数扩压叶栅展开数值计算研究。同时,还引入Gamma模型来研究粗糙度变化对通道内转捩现象的影响。研究发现,叶片表面粗糙度的增加将使得分离转捩和旁路转捩加强,但对逆转捩影响较小。此外,借助损失源分析方法,将叶栅内的损失分为前缘损失,叶型摩擦损失,二次流损失和尾迹损失四部分。结果表明,在角区分离严重且表面等效砂砾粗糙度增加到50μm时,相比于光滑情况,其总损失增加了9.6%。借助拓扑分析,可以发现随粗糙度增加,前缘分离泡不断前移,扰乱前缘部分流动,由此导致的前缘损失随粗糙度变化最为敏感。(3)针对某低速通风机展开数值计算,发现在所研究的流量范围内,其二次流主要存在位于吸力面轮毂与机匣处的螺旋点所诱导的涡结构,以及叶中区域的压力面分离。且随着流量增大,该压力面分离急剧增大,直至覆盖整个叶高范围,导致熵增加剧,效率明显下降。为提升其静压效率,利用神经网络优化算法,以静叶不同展向高度的叶型安装角为优化变量,静压效率为优化目标,最终使得该风机总压效率以及静压效率均提升0.6%。同时,对最优解的流场对比发现其主要在于抑制了低叶展区域的压力面分离现象。更多还原

【Abstract】 Nowadays,in order to develop the aerodynamic performance of the compressor,especially the high pressure ratio,it is needed to increase the stage pressure ratio,which will create more complex secondary flow in the compressor,particularly in the regions near the blade tip and hub corner.And according to the recent researches,the corner separation phenomenon near the hub will cause a large scale of loss and even make the compressor come into stall situation,which is more serious.Thus,in this paper,the development of the corner separation and the loss distribution will be investigated with the help of CFD method.Meanwhile,combining with the hotspot about the blade surface roughness,the effect of blade surface roughness on the corner separation and loss distribution to the cascade was studied.Apart from this,with the help of the numerical optimization means,the aerodynamic performance of a low-speed fan was studied by optimizing the installation angle of different sections.Brief introduction is listed below.(1)the corner separation in the hub was studied by means of the CFD,which was based on a low-speed and high-load axial flow compressor cascade.Based on the topological analysis of the streamline on the wall,the distribution of the singular points and the corresponding separation line as well as attachment line were studied.And this paper also investigated the development of all kinds of vortex in the cascade.Based on the three-dimensional streamlines,the threesource theory about how passage vortex generated was put forward,and the corresponding model of the cascade vortex structure was established.Through the above analysis,the position extraction method of channel vortex is established,and the development characteristics of channel vortex were explored.At the same time,the losses in the cascade under the diagonal separation of the dissipative function were investigated.(2)Low-Reynolds compressor cascade was studied to investigate that whether the roughness on the blade can promote the 3D separation and make a difference on all type of the loss in the cascade or not by CFX.In this study,the Gamma transition model was introduced to investigate the influence of blade surface roughness to transition phenomenon.It can be concluded that the blade surface roughness has more effect on the separation transition and the bypass transition than the reverse transition.Furthermore,a detailed loss sources analysis model was used to study the loss influenced by the surface roughness,which divided the total loss in the cascade into the leading edge loss,the surface friction loss,the secondary flow loss and the wake flow loss.The results show that the total loss increases to 9.6% at large scale of corner separation condition when the equivalent sand grain roughness in the blades increases to 50μm.And the leading edge loss is most sensitive to the surface roughness.With the help of the topology analysis method,it can be found that the movement of the separation bubble is the most obvious phenomenon in the field as the roughness increasing.(3)According to the numerical results,it’s found that there are three kinds of typical secondary flow in a low-speed fan,the vortex structure induced by the spiral points at the hub and casing on the suction surface and the separation around the pressure side on the blade.With the increase of the flow rate,the separation in the pressure blade increases sharply until the entire blade height is covered,which leads to an increase in entropy.Moreover,neural network optimization algorithm was used to improve the aerodynamic performance of the fan,which divided the stator blade into eight sections along the span direction,and treats the stagger anger of each section as the optimization variable and static pressure efficiency as the optimization objective.At the same time,the flow field of the optimal result was compared with the base stator,and it was found that the pressure surface separation was inhibited.更多还原