【作者】 王健；

【导师】 刘厚林； Matevz Dular；

【作者基本信息】 江苏大学 ， 流体机械及工程， 2015， 博士

【摘要】 本文在国家自然科学基金项目（51239005,51309120）、国家留学基金委公派联合培养博士生项目、欧洲太空总署（European Space Agency,ES A）项目"Cavitation in Thermosensible Fluids"和江苏省高校博士研究生创新基金（CXLX12₀640）的资助下开展工作。随着科技的进步以及工业制造精度的提高,水力机械得到了迅猛的发展,然而伴随着其大型化高速化而产生的空化问题则更加凸显。空化的存在不仅会降低水泵等水力机械的运行效率,还会产生振动和噪声,甚至会对水力机械固体表面造成破坏,严重影响系统的安全稳定运行。另一方面,空化由于具有复杂的物理化学相变流动特征与特殊的非定常动力特性,已成为污水处理、水下兵器减阻和管道清理等领域的一种高效低能耗新兴技术手段。本文首先对文丘里管和孔板两种结构简单的水力装置进行空化空蚀研究,采用理论分析、数值计算和实验测量相结合的方法研究空化空蚀形成机理与相互关系。随后以此为基础,对应用更为广泛、结构更为复杂的旋转类水力机械——离心泵进行空化空蚀的研究,改进离心泵空化数值计算模型,揭示离心泵内空化的非定常特性,建立离心泵空蚀预测方法。最后对一种水力空化发生器进行系统研究,为空化应用的研究提供一定的参考。本文的主要工作和创新性成果有：1.针对水力装置与机械空化数值模拟过程中需人为不断调整边界条件的问题,基于批处理（Batch）和ANSYS-CFX建立了一种空化数值模拟自动运行方法,实现了CFD计算软件的自动调用和边界条件的自动修改,使计算能够直接执行下一步。该方法保证了数值模拟过程的无缝衔接,缩短了计算周期。2.通过对文丘里管空化空蚀的研究,建立了一套试验研究—数值计算—图像后处理相结合的空化空蚀研究方法,具体研究了空泡各发展阶段与空蚀的内在联系。基于RNG k-ε模型提出了一种考虑空化可压缩性及湍流粘度修正的CDM湍流模型,有效地解决了常用湍流模型对湍流粘度过预测的问题：采用两台高速相机同时对文丘里管空泡结构演变与空蚀过程进行记录,实现了空化与空蚀的同步测量；采用蚀点计算法对空化空蚀数据进行分析,并辅以可压缩数值模拟方法,研究了空泡不同发展阶段时的空蚀破坏力。研究结果表明：空泡团的脱落与溃灭过程均会释放高能量的冲击波,并对固壁表面造成破坏,且后者为主要原因；蚀点的数量与面积的增长规律并非线性,而是呈阶梯状递增,空蚀大多在空泡溃灭或脱落时发生。3.基于孔板射流的空化空蚀试验研究,建立了一种基于空泡图像的空蚀预测方法。采用MATLAB软件对空化试验数据,即高速相机采集的连续空泡图像进行标准方差与平均值处理,预测出了潜在的空蚀区域以及相应的破坏程度,并反映出空泡的分布情况。空蚀试验结果证明了该方法的有效性与准确性,孔板射流空化的空蚀区域位于脱落空泡的溃灭点位置呈环状形态,其半径与采用空泡图像空蚀预测方法的半径相同。试验研究还揭示了流场速度与空蚀的关系,在空化数相同的前提下,流场流速越高,空蚀的程度越严重。4.以本文建立的CDM湍流模型为基础,考虑离心泵的旋转效应与大曲率结构特征,建立了一种适用于离心泵空化数值模拟的RCD湍流模型,并与试验结果进行了对比,结果表明：相比标准κ-ε模型与SST k-ω模型,RCD模型有效地抑制了传统湍流模型对湍流粘度的过预测,并准确地捕捉到更为细致的泵内湍流涡团与速度分布,使得数值计算能够准确地模拟出离心泵内空化流非定常脱落与溃灭的现象。当空化数不变时,在一个旋转周期内每一流道的空泡会依次经历发展、局部脱落、缩小再发展的周期性非定常过程。5.研究了ZGB空化模型中空泡半径、汽化与凝结经验系数对离心泵空化扬程下降曲线预测精度的影响,结果表明：空泡半径与汽化经验系数同时影响空泡的长度及含气量大小,凝结系数主要影响空泡的长度；高空泡含气量区域主要影响叶片吸力面进口前缘低压区的范围,低压区的长度与高空泡含气量的长度相当,而高空泡含气量区域的大小对离心泵扬程的影响较小,影响扬程预测精度的主要因素为空泡长度。6.考虑离心泵的旋转运动特性和流体湍动能与空泡结构演变的内在联系,基于ZGB空化模型发展了一种适用于离心泵空化数值模拟的RZGB空化模型,建立了离心泵转速和几何结构与空泡半径的函数关系。对RZGB模型对不同比转速离心泵的预测精度进行了评估,结果表明：对于低、中比转速泵,RZGB模型的汽化凝结项经验系数分别为5000和0.001时预测精度更高；对于高比转速泵采用汽化凝结项经验系数为50和0.01计算更为准确。与可视化空化试验结果对比表明,RZGB模型能够更准确地捕捉到离心泵进口空泡的三维特殊结构以及空穴尾端的不稳定现象。7.以本文建立的空泡图像空蚀预测方法为基础,建立了适用于旋转水力机械的空蚀预测方法,并通过空蚀试验证明了该方法的准确性与可靠性。该方法对连续的高速摄影图像或数值计算图像进行一系列的图形变换,保证各叶轮流道固定不变,再以图像标准方差与平均值处理方法进行分析,从而获得了离心泵叶片表面的空蚀区域与空泡分布。结果表明空蚀位置与空泡区域基本相当,但主体偏向空穴尾端,即离心泵中附着空泡尾端的高湍流不稳定现象同样会对固壁表面造成破坏。8.基于RCD湍流模型与RZGB空化模型,并结合试验对一种转子一定子型水力空化发生器的性能进行了系统的研究。研究结果表明,空化发生器内主要有三种空化类型,即转子与定子相对运动形成的楔形槽空化、转子高速旋转形成的转子叶片前缘空化和尾端低压区空化。在相同转速下空化发生器内的压力脉动随着流量的增加而增大；在相同流量下空化发生器内的压力脉动随着转速的增加而增大；增大转子与定子的间距将小幅度地降低腔内的压力幅值；空化发生器内的主要空蚀区域为转子叶片尾端,定子叶片前缘空泡附着部分及其尾端；空蚀主要在转子与定子完全交错时发生,此时形成的空化强度高且贴近固壁表面,这表明空泡在水力机械近壁面处的溃灭是导致空蚀发生的必要条件。更多还原

【Abstract】 With the progress of technology and industrial manufacturing precision, the hydraulic machinery gained a significant development in the past few years. However, since they are becoming much faster and larger, the issue of cavitation cannot be ignored any more. The exsistence of cavitation would degrade the hydraulic efficiency, induce vibration and noise and even damage the solid wall, which severely affects the system operation safty. On the other hand, despite of such disadvantages caused by the cavitation, its special physical-chemical phase transformation property and high unsteady dynamic characteristic bring us a fantastic water treatment approach, with outstanding efficiency and low energy cost. In order to figure out the cavitation problems in typical hydraulic machinery, the investigation from simple hydraulic apparatus, such as venturi and orifice plate, to complex ones, like centrifugal pump and cavitation generator, were conducted via combined numerical and experimental method. The main research contents and creative achievements arrived are as follows:1. An automatic running program for cavitating flow simulation in hydraulic machinery was developed on the basis of windows Batch and ANSYS-CFX. Within the Batch language, the program can automaticly create a.bat file to run ANSYS-CFX and simultaneously adjust the boundary conditions. It enables the calculation move on to another step without any time consuming, successfully reducing the entire simulation duration.2. A combined experimental — numerical — image post-processing prediction approach for cavitation erosion was established. The relataionship between cavitation developing process and its induced damges was analyzed. A simultaneously testing method for cavitation structures and cavitation erosion was achieved by applying two high speed cameras aside a venturi section, and the aluminum foil was used as a damage measurement sample. The numerical simulation and image post-processing methods were employed to enrich the analysis. Considering the compressible property of cavitating flow the eddy viscosity over-prediction problem of the original turbulence models, an improved Compressible—Density corrected turbulence model (CDM) was presented. The results show that both shedding off and collapse processes of the cavitation cloud would cause damages, and the collapse comes at the first place for cavitation erosion. Furthermore, it is found that the increase of the damge pits number and area is not linear, but step-wise, which means the pits are only generated as cavitation cloud shedding or collapse and the gerneration is not transient.3. On the basis of the investigation on orifice plates, an image based post-processing cavitation erosion method was proposed. A commercial code MATLAB was employed to process the images under the frame of standard deviation method. It is proved the continuous cavitation evolution images can be used as a cavitation erosion prediction method as well. The erosion area is annular, located in the sites where bubbles collapse. The diameter of the area is identical with the prediction result. The cavitation shedding frequency increases with decreasing cavitation number. While the cavitation stays constant, higher flow velocity contributes higher cavitation intensity, and then leads to severe cavitation erosion.4. In the terms of the rotating movement characteristic of centrifugal pump and based on the CDM model, a modified Rotating—compressible—Density corrected turbulence model (RCD) was proposed, which is suitable of handling the cavitating flow in centrifugal pump. Comparison had been made with the original k-ε model and SST k-ω model, and aslo the experiment results. It is indicated that the RCD model has better performance on predicting the pump head and its cavitation chacteristic. Moreover, by applying both the visualization experiment and unsteady simulation, the cavtitation evolution mechanism was found. The RCD model can successfully reduce the over-prediction of eddy viscosity, and in the meanwhile accurately capture the unsteady cavitation shedding and collapse processes, as well as the variation of the fluid density.5. The ZGB cavitation model was investigated by analyzing the influences of the bubble radius, evaporation coefficient and condensation coefficient. The results show that the bubble radius and evaportation coefficient affect the cavity length and vapor volume fraction at the same time, while the condensation coefficient only impacts the cavity length. High vapor volume fraction expands the low pressure region on the leading edge of the blade, but hardly affects the head of the pump. The cavity length is the essential factor which would influence the predition result of the pump head.6. Considering the influence of rotating movement characteristic of centrifugal pump on bubbles, and the interaction between turbulent dynamic energy, a rotating based ZGB cavitation model (RZGB) was presented specially for cavitating flow in centrifugal pump. This model effectively connects the pump rotating speed, turbulent dynamic energy and bubble radius. In order to evaluate the RZGB model, three centrifugal pumps with different specific rotating speed were taken into accout as the research objects. The results show that, compared with original ZGB model, when the evaporation and condensation coefficients are 5000 and 0.001 respectively, the RZGB model obtains more accurate predictions on the pumps of both low and medium specific rotating speed pump. For the high specific ratating spedd pump, the RZGB model with coefficients of 50 and 0.01 is more suitable. Furthermore, compared with the visualization experiment, it demonstrates that the RZGB model can capture the cavitation structures more accurate than ZGB model. The simulated result matches the experiment very well:the triangle shape is obtained as in the visualization.7. By applying the previous research in the current dissertation, an image based post-processing cavitation erosion prediction method for rotating hydraulic machinery was established. Both the numersical simulation images and experimental visualization images can be used to predict the erosion area and its extent of the damage. The experimental test was also conducted to validate the method. The results show that the sites of the damages are identical with the cavitaion region, while the most potential erosion area is close to the trailing part of the attached cavity. It is implied that the unsteady trailing part of the attached cavity could also be a trigger inducing damages.8. Based on the visualization experiment and numerical simulation with the proposed RCD turbulence model and RZGB cavitation model, a rotor—stator hydraulic cavitation generator was studied detailedly. It is found that there are three kinds of cavitation generating mechanism in the generator:one is produced by the interaction movement of rotor and stator, which forms a nozzle tube at the moment. The others are caused by the high rotating rotor, producing the cavitation at the leading edge and the rear part of the blade. The pressure pulsation in the generator increases as the flow rate increases while keeping the rotating speed constant. When the flow rate remains the same, the pressure pulsation rises with increasing ratating speed. Increasing the distance between rotor and stator would slightly decrease the pressure. Furthermore, the oil ink method was employed to investigate the erosion area of the cavitation generator. It indicates that the rear part of the rotor blade, the leading edge and the rear part of the stator blade are the potential erosive area. According to the previous conclusion, the damages are generated by high intensity cavitation when the rotor and stator are crossing each other, leading to nozzle cavitating flow. And it reveals that the cavitation erosion in hydraulic machinery is primarily caused by tiny bubbles close to the solid wall, not big cavitation cloud.更多还原