【作者】 刘浩；

【导师】 张雷；

【作者基本信息】 吉林大学 ， 机械制造及自动化， 2016， 博士

【摘要】 叶片是航空发动机、汽轮机等透平机械的关键核心部件,对整机的工作性能起到决定性作用,在国家能源、国防等重要领域扮演着举足轻重的角色。随着科技进步,高性能透平机械对叶片的设计和加工提出新的挑战,要求更优气动表现的叶片形状及型面设计、更高的叶片加工精度和表面粗糙度等级。针对上述问题,本文着重研究了叶片形状及型面优化设计和叶片表面粗糙度对透平机械气动性能的影响。本文基于CFD模拟仿真技术,对透平机械叶片形状进行多目标气动优化设计研究。将网格变形技术引入到叶片形状的优化设计中,实现了对透平叶片流场网格的自动变形控制,通过更新网格模型的网格节点位置,产生新的叶片形状,规避了叶片几何参数化和网格重划分过程,减少设计参数,提高了优化设计效率。通过对响应面法、克里金法、径向基函数法三种近似模型及混合多目标梯度搜索算法、改进的非支配排序遗传算法两种多目标优化算法的分析对比,表明克里金近似模型拟合精度高,鲁棒性好,混合多目标梯度搜索算法计算效率高,结果精确,由此提出由最优拉丁超立方试验设计、克里金近似模型和混合多目标搜索算法构成的组合优化设计方法。通过上述优化设计方法构建了基于网格自由变形的叶片多目标气动自动优化设计平台,以等熵效率和总压比为目标函数,网格变形参数为设计变量对航空发动机叶片进行多目标气动优化设计,并对优化设计结果进行分析和总结,验证了叶片多目标气动优化设计平台的有效性,与传统叶片多目标气动优化设计方法相比具有更高的优化效率。开展了透平机械叶片型面的局部精细化气动优化设计研究。精细化设计使设计参数增多,优化计算量显著增加,一般的优化方法难以满足其设计要求,因此将计算量与设计变量数目近乎无关的伴随方法引入到叶片型面的精细化设计中。基于伴随方法及流场伴随方程,推导了目标函数对自由网格变形控制点的梯度;分析比较了最速下降法、共轭梯度法和拟牛顿法三种梯度优化算法的优劣性,确定采用Armijo-Goldstein线搜索算法的BFGS拟牛顿法为最佳梯度优化算法。由此构建了基于伴随方法、网格自由变形方法和梯度优化算法的叶片三维气动精细化优化设计平台。通过此优化设计平台,以出口熵增为目标函数,网格变形参数为设计变量,有效地实现了航空发动机叶片型面的精细化气动优化设计。结果表明,此优化设计平台一方面可以通过梯度信息深刻认识叶片型面变化对其性能的影响,另一方面计算量相比多目标全局搜索算法大幅降低,极大地提高了优化设计效率。此优化设计平台与叶片全局多目标气动优化设计平台形成互为补充的关系。开展了叶片表面粗糙度对其气动性能影响的研究。从叶片加工机理角度出II发,分析了叶片粗糙表面的微观形态。基于已有的空气动力学模型、表面粗糙度与空气动力学的关系,研究了航空发动机叶片表面粗糙度对其气动性能的影响,并在此基础上针对性地分析了叶根倒圆、叶顶、下端壁等难抛光区域的表面粗糙度对叶片气动性能的影响。结果表明,叶片表面粗糙度的增加使叶片损失增加,叶片性能降低,叶片吸力面的粗糙度比压力面的粗糙度对其性能的影响大,难抛光区域的表面粗糙度增加均使叶片损失增加,其中下端壁对其性能影响最大。当叶片表面粗糙度0.1μmaR￡时,其表面是水力光滑的,叶片可获得最理想的气动性能;当叶片表面粗糙度1.5μmaR>时,其表面粗糙度对性能的影响幅度开始变大。本文的研究工作,有助于发展和完善透平机械叶片三维气动优化设计体系,对叶片的加工、维护也具有指导意义。更多还原

【Abstract】 Blade was the key core component of the turbomachinery equipment, such as aircraft engine, steam turbine and so on, which played a decisive role in the performance of the whole machine. And it had an important influence on national energy, national defense and other important fields. Along with scientific and technological progress, the development of high performance turbomachinery put forward new challenges to the design and machining requirements for the blade. Such as higher performance of blade shape and surface profile design, higher machining precision and surface roughness level. Aiming at the above problems, this paper emphatically researched on the blade shape and surface profile optimization design, and the influence of surface roughness on the aerodynamic performance of turbomachinery.In this paper, based on the CFD simulation technology, multi-objective aerodynamic optimization design for the turbomachinery blade shape was studied. The mesh deformation technique was introduced into the optimization design of the blade shape, and the automatic deformation control of the fluid grid of blade was realized. The new blade shape was generated by updating the grid node position of the mesh model directly, which avoided the process of blade geometry parameters and the mesh re-dividing. The design parameters were reduced and the efficiency of the optimization design was improved. Response Surface Model, Kriging, Radial Based Function approximate models and Hybrid Multi-Gradient Exploration, modified Non-dominated Sorting Genetic Algorithm were compared and analyzed, the results show that the Kriging approximation model has high fitting accuracy and good robustness, Hybrid Multi-Gradient Exploration optimization algorithm has high efficiency and accurate result. Thus a combinatorial optimization design method was proposed, which formed by optimum Latin hypercube design method, Kriging approximation model and Hybrid Multi-Gradient Exploration. Through the above mentioned optimization design method, a new multi-objective aerodynamic optimization design platform was constructed, which was based on the mesh free-form deformation. And the equal entropy efficiency and the total pressure ratio were took as objective functions, the deformation of grids control parameters were took as design variations to implement the aerodynamic optimize design of aeroengine blade. Through the analysis and summary of the optimization results, the effectiveness of the blade aerodynamic optimization design platform was verified, and the optimization efficiency was higher compared with the traditional blade aerodynamic optimization design method.A study on the local detail fine aerodynamic optimization design of turbomachinery blade surface profile was carried out. The fine design made the design parameters increase, and the calculation quantity was also increased significantly. The general optimization method was difficult to fit the design requirements. Thus, the adjoint method, which is almost independent of the amount of computation and the number of design variables, was introduced into the fine design of the blade surface profile. Based on the adjoint method and the adjoint equation of the flow field, the gradient of the objective function to the control point of mesh free-form deformation was derived. The advantages and disadvantages of the three unconstrained gradient optimization algorithms were analyzed and compared, which were the steepest descent method, the conjugate gradient method and the quasi Newton method. And the BFGS quasi Newton method with Armijo-Goldstein line search algorithm was determined as the optimal gradient optimization algorithm. Thus, the blade three dimensional detail fine aerodynamic optimization design platform was constructed, which based on the adjoint method, the mesh free-form deformation method and the gradient optimization algorithm. Through the optimization design platform, took the outlet entropy as the objective function, mesh deformation parameters as design variables, aeroengine blade surface fine aerodynamic optimization design was achieved effectively. The results show that the influence of blade profile on its performance is deeply understood through gradient information, and compared to the global search algorithm, the computation quantity is greatly reduced, which greatly improves the efficiency of the optimization design. The optimization design platform and the blade global multi-objective aerodynamic optimization design platform form a complementary relationship.In the end, the influence of blade surface roughness on the aerodynamic performance of turbomachinery was studied. The micro morphology of rough surface was analyzed from the aspect of blade machining mechanism. Based on the existing aerodynamic model and the relationship between surface roughness and aerodynamics, the influence of blade surface roughness on the aerodynamic performance was studied, and the surface roughness of blade fillet, blade hub and blade tip, which were difficult to polished, was also researched. The results show that the increase of the surface roughness of aeroengine blade increases the blade loss and degrades the aerodynamic performance. The influence of the roughness of the suction surface of the blade on the performance is greater than that of the pressure surface. The surface roughness of the difficult polishing area also increases the blade loss, the blade hub has the greatest impact on its performance. When the blade surface roughness Ra is less than 0.1um, the surface is smooth and the blade can obtain the most ideal aerodynamic performance. When the blade surface roughness Ra is greater than 1.5um, the impact extent of surface roughness on the performance of the blade becomes larger.The research results obtained in this paper play an active role in the development and improvement of the 3D Aerodynamic Optimization Design System of the turbomachinery, also has guiding significance in the processing and maintenance of the turbomachinery blade.更多还原