【作者】 高彬彬；

【导师】 王彤；

【作者基本信息】 上海交通大学 ， 动力工程， 2012， 硕士

【摘要】 直升机常常近地面飞行,包括低空悬停、起飞、降落等。在悬翼的影响下,地面沙尘等固体颗粒被扬起,不可避免地与气体工质一起被吸入发动机。这些颗粒的粒径范围分布较广,从微米量级到毫米量级,甚至到分米量级。其中较大尺度的颗粒与压气机叶片发生碰撞,会造成叶片破损,同时剥落的金属碎屑会掺杂在气体工质中进入下游部件中,由于这些金属碎屑硬度大、形状常带有尖角,会对下游的部件造成更大的破坏。较小尺度的颗粒可能会粘附于压气机叶片上,破坏其动平衡,甚至会改变叶片实际型线,降低压气机效率。另外,这些颗粒还可能会堵塞喷油口和气膜冷却孔,影响机组的性能和安全运行,缩短其使用寿命。为满足直升机的工作要求,尽可能地降低气体工质夹杂的颗粒给发动机带来的损害,需要在发动机入口安装颗粒分离器,考虑到直升机部件结构的限制、工作性能的要求,常采用惯性粒子分离器。为了分析惯性粒子分离器中的流场和性能参数,优化分离器的流道结构,本文首先开发出一套可自动生成分离器模型的参数化造型软件。在此基础上,本文对原惯性粒子分离器模型进行了建模和数值计算,对气相结果进行分析,得到该分离器的性能参数,并对不同粒径的颗粒在分离器内的运动规律进行了分析。接着本文选取了控制分离器进气流道弯曲度及喉部截面积、外缘线最大半径位置、舌部长度等的6个重要参数,对约150个分离器模型的性能参数进行了细致的数值分析和比较。在此基础上,利用神经网络建立分离器流道结构型线参数和分离器颗粒分离效率以及总压损失的非线性关系,通过遗传算法得到了分离器颗粒分离效率和总压损失的近似最优解。和原分离器模型相比,优化模型对大尺度颗粒的分离效率有很大改善,在总压损失略有增加的情况下,从原来的20%增加到了90%;并且,优化后分离器的高效工作范围也大大扩展,在进气速度为30m/s～80m/s、清除流量比为0.12～0.18的范围内,优化模型对大尺度颗粒均有较高的分离效率。因此,本文的研究为惯性粒子分离器的优化设计奠定了良好的理论基础。更多还原

【Abstract】 Helicopter is often near the ground flight, including low-altitude hover, takeoff,landing and so on. Ground dust and other solid particles which are raised under theinfluence of suspended wing will be sucked into the engine along with the gasrefrigerant inevitably. The range of diameter distribution of these particles is wide,from micron magnitude to millimeter magnitude, even arrives decimeter magnitude.Large-scale particles collide with the compressor blades and can damage them,while the metal debris peeling off from the blades can get into the downstreamcomponents, which are doped in the gas refrigerant. Because of high hardness andsharp corners of the metal debris, they can cause more damage to the downstreamcomponents. Small-scale particles which may adhere to the compressor blades candamage their dynamic balance,and even can change the actual type line of theblades, which will reduce efficiency of compressor.In addition, these particles mayjam injection mouth and film cooling hole, which will affects the performance andsafe operation of engine unit and shorten its service life.To meet the work requirements of helicopter and to reduce the damage which iscaused by the particles doped in the gas refrigerant, it’s necessary to install particleseparators in engine entrance. Considering the limits of helicopter componentstructure and the requirements of its work performance, an inertial particle separatorwas usually selected. In order to analyze the flow field and performance parametersof inertial particle separator and optimize its flow structure, this paper developed a set of parameterized automatic modeling software for separator firstly. Based on thesoftware, we finished the numerical computation for original separator model, andanalyzed its result of gas phase. We got the performance parameters of the separator,and analyzed the motion law of different diameter’s particles within separator. Thensix important parameters were selected, which controlled bending degree of separatorinlet port, section area of throat, position of the biggest radius of outside wall, lengthof tongue department and so on. Then we did detailed numerical analysis andcomparison for the performance of about 150 separators. Based on the results, thenonlinear relationship of parameters of separator flow structure line and particleseparation efficiency and total pressure loss of separator was established by the neuralnetwork. we got the approximate optimal solution of separation efficiency and totalpressure loss of separator by the genetic algorithm. Compared with the original model,the separation efficiency of large-scale particles was increased for the improvedmodel, from less than 20% to about 90%,while the total pressure loss was increasedslightly. The efficient work scope of separator was also greatly expended afteroptimization, the separation efficiency of large-scale particles is all high for optimizedmodel when inlet velocity is in the range of 30m/s to 70m/s and clear flow ratio wasin the range of 0.12 to 0.20. This study has laid a good theoretical basis for optimaldesign of inertial particle separator.更多还原