WO2013091222A1 - 径流式液力透平优化设计方法 - Google Patents

径流式液力透平优化设计方法 Download PDF

Info

Publication number
WO2013091222A1
WO2013091222A1 PCT/CN2011/084460 CN2011084460W WO2013091222A1 WO 2013091222 A1 WO2013091222 A1 WO 2013091222A1 CN 2011084460 W CN2011084460 W CN 2011084460W WO 2013091222 A1 WO2013091222 A1 WO 2013091222A1
Authority
WO
WIPO (PCT)
Prior art keywords
optimization
design
algorithm
cfd
flow
Prior art date
Application number
PCT/CN2011/084460
Other languages
English (en)
French (fr)
Inventor
孙金菊
宋鹏
王科
Original Assignee
西安交通大学
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 西安交通大学 filed Critical 西安交通大学
Priority to PCT/CN2011/084460 priority Critical patent/WO2013091222A1/zh
Publication of WO2013091222A1 publication Critical patent/WO2013091222A1/zh

Links

Classifications

    • FMECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F03MACHINES OR ENGINES FOR LIQUIDS; WIND, SPRING, OR WEIGHT MOTORS; PRODUCING MECHANICAL POWER OR A REACTIVE PROPULSIVE THRUST, NOT OTHERWISE PROVIDED FOR
    • F03BMACHINES OR ENGINES FOR LIQUIDS
    • F03B3/00Machines or engines of reaction type; Parts or details peculiar thereto
    • F03B3/10Machines or engines of reaction type; Parts or details peculiar thereto characterised by having means for functioning alternatively as pumps or turbines
    • F03B3/103Machines or engines of reaction type; Parts or details peculiar thereto characterised by having means for functioning alternatively as pumps or turbines the same wheel acting as turbine wheel and as pump wheel
    • FMECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F05INDEXING SCHEMES RELATING TO ENGINES OR PUMPS IN VARIOUS SUBCLASSES OF CLASSES F01-F04
    • F05BINDEXING SCHEME RELATING TO WIND, SPRING, WEIGHT, INERTIA OR LIKE MOTORS, TO MACHINES OR ENGINES FOR LIQUIDS COVERED BY SUBCLASSES F03B, F03D AND F03G
    • F05B2240/00Components
    • F05B2240/20Rotors
    • F05B2240/24Rotors for turbines
    • F05B2240/242Rotors for turbines of reaction type
    • FMECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F05INDEXING SCHEMES RELATING TO ENGINES OR PUMPS IN VARIOUS SUBCLASSES OF CLASSES F01-F04
    • F05BINDEXING SCHEME RELATING TO WIND, SPRING, WEIGHT, INERTIA OR LIKE MOTORS, TO MACHINES OR ENGINES FOR LIQUIDS COVERED BY SUBCLASSES F03B, F03D AND F03G
    • F05B2260/00Function
    • F05B2260/84Modelling or simulation
    • YGENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02EREDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
    • Y02E10/00Energy generation through renewable energy sources
    • Y02E10/20Hydro energy

Definitions

  • the invention belongs to the residual pressure recovery technology in the fields of petrochemical, chemical fertilizer, chemical industry, etc., and relates to a design method of a hydraulic turbine, in particular to a design optimization method of a radial flow hydraulic machine. Background technique
  • Hydraulic turbines entered the market in the late 1970s, with reverse pumps (REVERSE PUMP), Francis pump (FRANCIS), Kaplan (KAPLAX) and Pelton (PELTON).
  • the reverse pump type hydraulic turbine realizes the pressure recovery of the liquid medium by the reverse operation of the centrifugal pump.
  • the structure is simple, the cost is low, and the design is mature, so it is widely used in petrochemical plants.
  • the flow requirements are strict. When the flow rate is too large, the recovery efficiency will be greatly reduced, and even there will be no recovery. Due to the above drawbacks, such hydraulic turbines tend to be replaced by radial turbines.
  • the pressure drop required by the petrochemical plant process is often quite large, and there is a large amount of recyclable liquid pressure energy.
  • the single-stage turbine is not suitable for high-pressure differential hydraulic recovery, only the addition liquid
  • the level of the force turbine can increase the scale of energy recovery.
  • the radial blade turbine hydraulic turbine is more efficient than the reversible pump hydraulic turbine, the number of turbines used in the same differential pressure is often less than that of the reverse pump turbine. The reduction will effectively improve the reliability of the rotor.
  • the diameter of the radial vane turbine is relatively small, the structure is compact and the cost is low.
  • variable geometry nozzles can be used because the nozzles on the annular runner are bolted and easy to replace, so that the hydraulic turbine has the best working efficiency and the design flow range is also better. width.
  • the design theory of radial hydraulic turbine flow components there is limited information on the design theory of radial hydraulic turbine flow components.
  • the invention discloses a radial hydraulic turbine optimization design method to provide support for the design of the hydraulic turbine flow passage component. Summary of the invention
  • the object of the present invention is to provide a complete machine optimization design method for a radial flow turbine flow passage component, including a one-dimensional optimization design, an initial shape design of a flow passage component, and an optimized design method (a platform).
  • the invention provides a radial flow hydraulic turbine optimization design method, comprising a one-dimensional thermal optimization design method, a three-dimensional modeling method of a flow component and a whole machine optimization method; wherein the overall machine optimization method comprises the following steps: nozzle guide vane and Impeller blade parameterization step, co-evolutionary genetic algorithm, adaptive approximation model, and CFD (Computational Fluid Dynamics) automatic calling algorithm.
  • the radial flow hydraulic turbine optimization design method is as follows:
  • the initial shape of the flow-through component is designed, including nozzle vane shape, impeller meter and volute design; (3) Perform CFD analysis on the flow field of the through-flow components such as the volute, nozzle, impeller, etc. in the whole machine environment, obtain the performance data of the initial design, and analyze the flow loss;
  • the algorithm has the ability to deal with multi-variable highly nonlinear problems; because the original problem is reasonably segmented on the basis of considering the correlation of variables, considering the correlation between sub-problems, at the maximum On the basis of retaining the characteristics of the original problem, the amount of calculation required for optimization is reduced;
  • (8M is used in the optimization process, using the dynamic sampling strategy to update the approximate model, automatically calling fewer CFD calculations, continuously improving the accuracy of the model in the potential optimal region, and improving the optimization efficiency;
  • the one-way thermal optimization method includes the following steps:
  • the one-dimensional thermal design uses the genetic algorithm to automatically complete the optimization process through the computer, while the non-traditional relies on continuous trial and error;
  • the whole machine optimization method includes the following steps:
  • the modules of the optimization method are not simply sequenced, but are organically integrated.
  • CFD needs to be called to complete the establishment of the initial sample database, and the optimization algorithm is called to complete the approximation.
  • the fitting work of the model; the fitted approximation model is used as the objective function to guide the optimization; in the process of optimization, the CFD calculation is called according to the needs of the algorithm calculation, and the appropriate sample points are added.
  • the accuracy of the approximate model in the potential optimal region is updated.
  • both the optimization algorithm and the CFD are called, and the approximate model is updated at the same time.
  • the purpose of the one-dimensional thermal optimization design is to satisfy various constraints within the range of parameter values, match the relevant design parameters of each component, determine the diameter of the impeller inlet and outlet, the speed triangle, and the shape of the blade wheel, to lay the foundation for the design of the three-dimensional flow channel. .
  • the wheel cycle efficiency is selected as the objective function, the corresponding dimensionless parameter is the design variable, various constraints are given, and the genetic algorithm is introduced to form a one-dimensional thermal optimization design method.
  • the impeller wheel cycle efficiency ⁇ can be expressed as a function of the following dimensionless parameters: In the above formula
  • D y Di wheel diameter ratio (average diameter of impeller outlet / impeller inlet diameter)
  • the range of values for the above dimensionless parameters is largely determined based on empirical data combined with actual design requirements. Optimization needs to ensure a wide parameter search space as much as possible, but there is no too unreasonable parameter combination; depending on the application, the variable value range may be different, and it needs to be adjusted according to experience. For example, the following is a more appropriate range of variable values:
  • the Mach number of the nozzle outlet is constrained by the reaction degree ⁇ : Proper selection of a large ⁇ can give full play to the advantage of using inertial force to work. However, if the ⁇ is too large, the impeller load will be too large, and the inside of the turbine will be The increase in external air leakage loss and the friction loss of the wheel also causes an insignificant increase, and the axial force that the rotor is subjected to increases accordingly. ⁇ should not be too small, otherwise c may be too large and exceed the local speed of sound.
  • the relative leaf height is within a certain range and meets the minimum leaf height requirement: 0.03 ⁇ ⁇ 0.15, ⁇ 3mm
  • Hydraulic turbines need to operate at fixed or variable flow rates in different processes.
  • a fixed-angle nozzle configuration can be used, and the vanes are pneumatically shaped to reduce flow losses, as shown in Figure 1(a).
  • an adjustable angle nozzle configuration is required to improve the performance of the hydraulic turbine. Since the rotating mechanism is to be mounted on the nozzle, the elongated pneumatic blade is no longer suitable, and it is necessary to find a suitable blade shape.
  • Fig. 1(b) shows a guide vane type suitable for the adjustable structure.
  • the three-dimensional shape of the impeller includes a meridian profile, a work wheel and an exit inducer design.
  • the meridian lines of the wheel and the wheel cover are all elliptical.
  • the working wheel blades take into account the strength requirements and are generally radial straight blades.
  • the hydraulic turbine outlet inducer should meet several requirements: It has good hydraulic properties; the formed blade has better strength characteristics; it is easy to process and inspect.
  • the non-expandable parabolic paraboloid of the cylindrical base has advantages in terms of hydraulic performance, strength, processing technology, etc., and can meet several requirements for the induction of vane type. Therefore, the present invention adopts a straight-line paraboloid with a non-expandable cylindrical surface to face the wind deflector for blade modeling. As shown in Figure 3, through the impeller axis The plane is called the meridional plane, the plane perpendicular to the axis of rotation of the impeller is called the radial plane, and the end surface perpendicular to the axis of rotation of the impeller is the radial end surface.
  • Fig. 4 is the program flow chart.
  • the volute consists of three parts: the inlet tube, the coil tube and the ring accelerator.
  • the ternary flow in the volute is simplified to a constant flow of binary adiabatic, and it is assumed that the mass of the liquid of a certain section of the scroll is concentrated on the core of the section. Therefore, by determining the variation of the fluid parameters along the corrugated section core connection line, it is possible to understand the fluid flow in the scroll and design the volute accordingly.
  • the flow rate of the working fluid passing through any section F of the scroll tube is in the following relationship with the cross-sectional azimuth angle ⁇ and the number of scroll tubes Z v :
  • is the height of the volute, is the radius of the section circle, is the width of the nozzle blade, is the number of speed, is the azimuth angle, 0 ⁇ is the liquid exit angle, and the azimuth angle 0 in each of the above calculation formulas appears in an implicit form.
  • the parameters of a given section can be solved by computer programming to design the volute. 3.
  • Parametric expression of the nozzle blades is carried out using two B-spline curves (which can also be described by Bezier curves or NURBS).
  • control points are selected to control the upper and lower curves.
  • the blade tail and the leading edge points (points 1 and 8) are fixed; to reduce the control variables used, the shape of the nozzle vanes is changed only by changing the ordinates of the remaining control points (a total of 6 variables).
  • the nozzle mounting angle is also used as an optimization variable and is determined by performance optimization, as shown in Figure 503). For adjustable nozzles, a total of 7 variables are used to parameterize the nozzle vanes.
  • the elliptical curve is used for the leaf wheel noon line
  • the radial straight blades are used for the working wheel
  • the induction wheel is described by the cylindrical-parabolic equation.
  • the parameterization of the impeller blades is based on the initial geometry of the blades.
  • the five modeling parameters c ⁇ c ⁇ A ⁇ and ⁇ 3 ⁇ 4 are selected as the optimized variables for the impeller blades.
  • the optimized design of the hydraulic turbine needs to be carried out in the whole machine environment.
  • the CFD analysis is carried out in the whole machine environment, and the parameter optimization method and the advanced co-evolutionary genetic algorithm (CCGA) are combined to establish a complete design method of the hydraulic turbine.
  • CCGA advanced co-evolutionary genetic algorithm
  • the objective function is a combination of the two, expressed as
  • Pr represents the expansion ratio of the liquid expander
  • Pr° is the expansion ratio of the initial design
  • cnc 2 is the empirical coefficient
  • the optimization platform consists of four main modules: parametric generation of nozzle vanes and impeller blades, co-evolutionary genetic algorithm (CCGA), adaptive approximation model technique, and CFD Automatic call calculations.
  • CCGA co-evolutionary genetic algorithm
  • adaptive approximation model technique adaptive approximation model technique
  • CFD Automatic call calculations CFD Automatic call calculations.
  • these four modules are organically integrated. The figure shows the flow chart of the optimization platform. A detailed description of each module is given below.
  • Parametric module parameterize the blade of the nozzle by using the spline curve, change the profile of the nozzle blade by changing the coordinates of the control point; if it is an adjustable nozzle, it can also describe the change of the installation angle of the nozzle blade; -
  • the parabolic method expresses the impeller blades, and the impeller design can be modified by adjusting the geometric parameters. The above method is implemented by the program and is conveniently outputted in the CFD software data interface format to automatically generate nozzle and impeller blade data during the optimization process.
  • Approximate model module Construct a certain number of group sample designs, and perform CFD analysis on the samples in the whole environment to obtain the objective function values to establish the approximate model for initialization. In the subsequent optimization calculation, this approximation model is used to replace the time-consuming CFD calculation to complete the objective function evaluation and accelerate the optimization process. In the optimization process, new points are added according to the algorithm needs to improve the potential model. The most advantageous prediction accuracy.
  • Variable grouping algorithm module The key problem in the co-evolution algorithm is reasonable variable grouping.
  • correlation data between different variables is obtained by statistically distributing distribution data of individual populations of genetic algorithms.
  • the optimization variables are divided into multiple groups (ie, variable space segmentation) to optimize the calculation using the co-evolution genetic algorithm.
  • Co-evolutionary Genetic Algorithm (CCGA) module Co-evolutionary genetic algorithm divides a complex multivariate optimization problem into multiple relatively independent sub-problems, and uses genetic algorithm to solve each sub-problem one by one. Reasonable segmentation of the problem based on the correlation between variables can effectively reduce the amount of computation required for optimization. Since each genetic algorithm individual has only a part of the optimization variables, when it is necessary to calculate a candidate individual objective function value, it needs to be combined with variables from other populations to obtain the objective function of the candidate individual. DRAWINGS
  • Figure 1 (a) and (b) show the fixed mounting angle and the adjustable mounting angle nozzle vane pattern.
  • Figure 2 Schematic diagram of the impeller meridian plane.
  • Figure 3 is a schematic view of an impeller blade.
  • the origin 0 is selected on the impeller rotation axis O.
  • the length Y of the arc on the 0-angle and the radial plane is positive when it is the same as the direction of rotation of the impeller, and vice versa, and is calculated from the mid-surface of a certain working wheel blade.
  • O. The axis of the center paraboloid. It is assumed that the intersection of the central paraboloid and the outer end face is a straight line and passes through this axis.
  • O and O The distance between the two shafts is R a and the angle of the top of the blade is reduced!
  • is a positive value of the axial projection of the angle between the intersection of the central parabola and the outer end face and the radial line at the average radius, and the angle ⁇ of the blade top wrap angle is also positive.
  • Fig. 5 is a schematic diagram of parameterization and adjustable installation angle of the nozzle guide vane.
  • Figure 6 shows a schematic diagram of how to implement automatic CFD call during the optimization process of the hydraulic turbine.
  • Figure 7 shows the flow chart of the hydraulic turbine optimization design program platform. detailed description
  • the initial shape of the flow-through component is designed, including the nozzle guide vane shape, the impeller meter, and the volute design.
  • a fixed-mount angle nozzle configuration is recommended, and the vanes are pneumatically shaped to reduce flow losses, as shown in Figure 1 (a).
  • Figure 1 (a) For hydraulic turbines with variable flow operation, in order to improve the hydraulic turbine's changing conditions The performance requires the use of an adjustable angle nozzle structure, which in turn requires the use of a vane type suitable for adjustable structural requirements, Figure l(b).
  • the three-dimensional shape of the impeller includes a meridian profile, a work wheel and an exit inducer design. As shown in Fig.
  • the meridian line of the hydraulic turbine impeller wheel and wheel cover adopts an elliptic curve
  • the working wheel blades are radial straight blades
  • the exit induction wheel adopts a cylindrical base surface non-expandable straight grain parabolic surface guide.
  • the wind wheel performs blade shape.
  • Figures 3(a), (), and (c) show the meridional projection of the impeller, the projection, and the development of the cylindrical section of the induction wheel. According to the above scheme, the numerical simulation and computer programming make it easy to realize three-dimensional forming of the impeller.
  • Figure 4 shows the flow chart of the impeller design procedure.
  • Equation (4) gives the relationship of the geometric parameters of the volute.
  • the parameterization of the geometry is performed and the optimized design variables are extracted.
  • the suction and pressure surfaces of the nozzle vanes are fitted by two spline curves, and the coordinate values of the control points are obtained, which are used as the initial design expression, and some coordinate values are reasonably selected as the parameters of the nozzle airfoil optimization.
  • Variable For adjustable nozzles, 1 variable can be added to indicate the nozzle mounting angle change.
  • the impeller is parameterized and five variables ⁇ , ⁇ are selected. , ⁇ and as an optimization variable for the impeller blades, as shown in Figures 3(a), (b) and (c).
  • and as an optimization variable for the impeller blades, as shown in Figures 3(a), (b) and (c).
  • 12 variables were used to parameterize the nozzle and impeller of the liquid expander. Six of these variables are used to control nozzle blade changes, one variable is used to control nozzle mounting angle changes, and five variables are used to optimize the impeller.
  • Equation (5) gives its mathematical expression.
  • a CFD-based optimization platform is established, which is suitable for the optimization of hydraulic turbine level.
  • the automatic call of CFD is completed through a series of own programs and batch scripts to complete CFD analysis including candidate blade profile generation, grid.
  • the process of partitioning, CFD setup and solver auto-solving, and numerical result acquisition is convenient for integration with the optimization program.
  • the process is shown in Figure 6.
  • Figure 7 is a flow chart of the optimized design platform, which mainly includes four modules: nozzle vane and impeller blade parameterization, co-evolutionary genetic algorithm, adaptive approximation model technology, and CFD automatic call.
  • the optimization method effectively reduces the amount of calculation and accelerates the optimization convergence by the following measures: Using the approximate model to predict the objective function instead of calling the time-consuming CFD calculation each time; Using the approximate model algorithm with automatic update capability, continuously improve the model Improve the accuracy of the optimal optimal area and improve the efficiency of optimization.
  • the introduction of co-evolutionary algorithms makes the algorithm capable of dealing with multivariable highly nonlinear problems.
  • variable grouping algorithm detecting the (tight or loose) quantitative relationship between variables, segmenting a complex multivariate optimization problem into multiple relatively independent sub-problems (dividing the optimization variables into multiple groups) and using the genetic algorithm Solve each sub-problem one by one; based on the correlation between variables, reasonably segment the problem, consider the correlation between sub-problems, and effectively reduce the optimization problem based on the maximum retention of the original problem characteristics. The amount of calculation required.
  • the adaptive genetic operator is used in the algorithm to improve the global search ability of the co-evolutionary algorithm and accelerate convergence.
  • the modules of the optimization platform are not simply sequenced, but are organically integrated.

Landscapes

  • Engineering & Computer Science (AREA)
  • Chemical & Material Sciences (AREA)
  • Combustion & Propulsion (AREA)
  • Mechanical Engineering (AREA)
  • General Engineering & Computer Science (AREA)
  • Turbine Rotor Nozzle Sealing (AREA)

Abstract

一种径流式液力透平优化设计方法包含一元热力优化设计方法,通流部件三维造型方法以及整机优化方法;其中,整机优化方法包括如下步骤:喷嘴导叶和叶轮叶片参数化步骤、协同进化遗传算法、自适应的近似模型以及CFD自动调用算法。这种设计方法有效减少了计算量。

Description

说明书 径流式液力透平优化设计方法 技术领域
本发明属于石油化工、 化肥、 化学工业等领域余压回收技术, 涉及 一种液力透平的设计方法, 特别是一种径流式液力机械整机设计优化方 法。 背景技术
在高能耗的石油化工装置流程中,需要将高压液体节流至低压以满 足工艺流程的需要。 通常这一降压过程是通过节流闽来完成的, 节流闽 节流过程是极其不可逆的剧烈降压过程, 高压液体能量将被白白浪费 掉, 且会导致液体介质的汽化、 结构的汽蚀破坏。 石油化工企业也越来 越重视节能降耗以降低产品成本和提高企业自身的竞争能力, 节能技术 和设备的应用及推广就显得愈加重要。在石油化工装置流程通过液力透 平回收高压流体的能量用以驱动其他转动机械如泵等, 可以取得显著的 节能降耗效益。
液力透平于 70 年代末期进入市场, 有反转泵( REVERSE PUMP) 、 法兰西斯泵型( FRANCIS ) 、 卡普兰型(KAPLAX) 和佩尔顿型 ( PELTON)。 反转泵式液力透平通过离心泵的反转运行实现液体介质的 压力回收, 其结构简单, 成本较低,设计成熟, 故在石化装置中有较广泛 的应用。 但对流量要求比较严格, 流量过大过小时, 回收效率将大幅降 低, 甚至出现无回收情况。 由于上述缺陷, 此类液力透平有逐歩被径向 透平取代的趋势。
石油化工装置工艺流程所求的压降往往相当大, 蕴藏着大量可回收 的液体压力能, 单级透平不适用于高压差的液力回收利用, 只有增加液 力透平的级数才能提高能量回收规模。但随着级数的增加势必会增加转 子的轴长, 从而降低其可靠性。 而径向叶片涡轮液力透平与可逆泵式液 力透平相比具有更高的效率, 同样压差下所用透平的级数往往要比反转 泵式透平要少, 级数的减少将有效地提高转子的可靠性。 同时, 由于径 向叶片涡轮的直径相对较小, 因此结构紧凑、 造价低。 径向叶片涡轮设 计的更大优点是可以使用变几何喷嘴, 因为环形流道上的喷嘴为螺栓连 接, 容易更换, 这样, 液力透平就有最佳的工作效率, 同时设计流量的 范围也较宽。 但就国内外文献看, 径向液力透平通流部件的设计理论方 面资料有限。
本发明公开了径向液力透平优化设计方法, 为液力透平通流部件的 设计提供支撑。 发明内容
本发明的目的是提供径流式液力透平通流部件的整机优化设计方 法, 包括一元优化设计、 通流部件的初始形状设计、 优化设计方法 (平 台)。
本发明提供了一种径流式液力透平优化设计方法, 包含一元热力优 化设计方法, 通流部件三维造型方法以及整机优化方法; 其中, 整机优 化方法包括如下歩骤: 喷嘴导叶和叶轮叶片参数化歩骤、 协同进化遗传 算法、 自适应的近似模型以及 CFD (计算流体力学) 自动调用算法。
所述径流式液力透平优化设计方法, 按照如下歩骤:
(1)给定流量、进出口工质参数,将一元优化模型与优化算法相结合, 考虑各种约束条件, 通过编程求解优化问题, 确定叶轮进出口直径、 速 度三角形, 为通流部件三维设计打下基础;
(2)基于一元优化设计结果, 设计通流部件初始形状, 包括喷嘴导 叶叶型、 叶轮计以及蜗壳设计; (3)在整机环境下, 对已设计的蜗壳、 喷嘴、 叶轮等通流部件的流场 进行 CFD分析, 获得初始设计的性能数据, 同时分析流动损失;
(4) 基于通流部件初始形状设计, 进行几何形状的参数化并提取优 化设计变量; 依靠 CFD 数值研究, 通过变量灵敏度分析确定优化变量 的范围; 优化目标是在整机环境下, 同时提高整机的效率以及膨胀。
(5)利用试验方法构造实验样本, 并用 CFD对实验样本进行分析, 获得其目标函数值; 利用实验样本数据建立近似模型, 替代耗时的 CFD 计算, 作为目标函数指导优化进行;
(6)利用遗传算法计算若干歩, 完成对整个设计空间的初歩探索; 分 析遗传算法种群数据, 计算变量之间紧密或松散的量化关系; 通过变量 分组算法, 将复杂的多变量优化问题分解为多个相对独立又互相影响的 子问题;
(7)使用协同进化算法优化,该算法具有处理多变量高度非线性问题 的能力; 由于是在考虑变量相关性的基础上对原问题进行合理分割, 考 虑了子问题间的相关性, 在最大限度保留原问题特性的基础上, 减少了 寻优所需的计算量;
(8M吏用在优化的过程中, 使用动态采样策略更新近似模型, 自动 调用较少次数 CFD计算, 不断提高模型在潜在最优区域的精度, 提高 寻优效率;
(9)当计算量达到预定最大值时, 优化算法结束, 获得最优化设计。 所述一元热力优化方法包括如下歩骤:
(1)一元热力设计使用遗传算法, 通过计算机自动完成优化过程, 而 非传统的依靠不断试凑得出;
(2)在优化的过程中, 需要工质在通流部件不同截面的物性数据, 均 通过调用物性子程序完成, 不但保证了计算精度, 而且, 由于物性子程 序可计算不同工质, 优化程序具有设计不同介质膨胀机的能力。
所述整机优化方法包括如下歩骤:
为同时考虑膨胀机各个部件对整机流场及性能的影响,优化各个 部件工作在最优化状态, 引入了协同进化算法, 使得优化方法具有处理 多变量高度非线性问题的能力; 此种算法将一个复杂的多变量优化问题 分割为多个相对独立的子问题, 并利用遗传算法对每一个子问题进行逐 歩求解; 在考虑变量之间相关性的基础上对问题进行合理分割, 在最大 限度保留原问题特性的基础上, 有效减少寻优所需的计算量;
(2)由于整机优化使用了较多变量,为保证优化在有限计算资源下完 成, 利用利用克里金近似模型预测目标函数, 而不是每次都调用耗时的
CFD计算;采用了有自动更新能力的近似模型算法,在仅调用少量 CFD 计算次数基础上,不断提高模型在潜在最优区域的精度,提高寻优效率;
(3) CFD 的自动调用通过一系列自有程序和批处理脚本, 完成 CFD 分析包括候选叶片型线生成、 网格划分、 CFD设置和求解器自动求解、 以及数值结果获取的全部过程, 可直接与优化程序集成;
(4)该优化方法的各个模块不是简单按照顺序依次进行, 而是被有机 地整合在一起, 譬如, 在建立初始近似模型时, 需要调用 CFD完成 初始样本数据库的建立, 并调用优化算法完成近似模型的拟合工作; 拟合后的近似模型被作为目标函数, 用来指导优化的进行; 在寻优 的过程中, 又会根据算法计算的需要, 调用 CFD计算, 增加适当的 样本点, 来更新近似模型在潜在最优区域的精度, 这时候优化算法 和 CFD均被调用, 近似模型同时完成一次更新。 1. 一元热力优化设计方法
一元热力优化设计的目的是在参数取值范围内, 满足各项约束条 件, 匹配各部件的相关设计参数, 确定叶轮进出口直径、 速度三角形、 叶轮子午面形状, 为三维流道的设计打下基础。
本发明中所建立的一元优化设计模型中, 选取轮周效率为目标函 数,相应的无量纲参数为设计变量, 给出各项约束条件, 引入遗传算法, 形成了一元热力优化设计方法。 叶轮轮周效率^可表示为下列无量纲 参数的函数:
Figure imgf000006_0001
上式中
—导向装置 (包括蜗壳和喷嘴) 的速度系数
—叶轮的速度系数
Ω—反动度
—叶轮进口绝对液流角
一叶轮出口相对液流角
"/一级的速比 (叶轮进口周速 /等熵膨胀速度)
DyDi一轮径比 (叶轮出口平均直径 /叶轮进口直径) 上述无量纲参数的取值范围, 很大程度上基于经验数据并结合实际 设计要求来确定。 优化需要尽可能地保证较宽的参数搜索空间, 但又不 会出现过于不合理的参数组合; 根据不同的应用情况, 变量取值范围或 许不同, 需要根据经验做适当调整。 例如, 下列是一组较为合适的变量 取值范围:
0.3< Ω < 0.6
0.85 < <^ < 0.97
Figure imgf000007_0001
0.5≤^/¾≤0.8
10 ≤^≤30
20 <β2<50
0.2≤D2/D1≤0.6
在一元优化设计中, 除了要满足优化变量的取值范围外, 还需要同 时满足其它约束条件, 以获得较好的流动特性及整机性能。 主要包括:
(1) 喷嘴出口马赫数对反动度 Ω的约束: 适当地选用较大的 Ω可以充 分发挥利用惯性力做功的优点, 但若 Ω过大, 将会造成叶轮负载 过大, 透平的内、 外漏气损失以及轮盘摩擦损失等增加, 还会造 成 无意义地增加, 转子承受的轴向力也会相应地增加。 Ω也不 能太小, 否则 c 可能过大而超过当地音速。
(2) 叶轮加速因子的限制作用:叶轮内无减速流动: w2 > Wl
(3) 入口冲角: ί = β、 β、, 对于高速的径向-轴流式透平, 叶轮流道中 的全部能量损失中, 入口冲击损失要占很大比例。 这项损失的大 小主要取决于入口冲角。 建议叶轮进口保持一定负冲角或较小正 冲角: -10 <i<5 。
(4) 许用速比 的约束: 根据叶轮入口能量损失系数 和冲角 之间关 系, 从水力的观点来看, 应尽量采用负冲角或者较小的正冲角。 然而, 从结构设计的观点来看, 有的时候需要有一定的正冲角, 因为这有利于降低透平 的数值。
(5) 相对叶高在一定范围 内 , 且满足最低叶高要求: 0.03 < < 0.15, < 3mm
(6) 设计中根据实际需要还要考虑别的约束条件 (如最小轴颈要求 等)。 传统设计方法利用循环迭代方法试凑出上述各参数; 但由于设计变 量较多、 变量取值范围广、 约束条件众多, 将产生大量的候选解, 很难 通过试凑的方法获得满足约束条件的最佳设计参数。因此,在本发明中, 将上述一元设计模型和遗传算法相结合, 用优化的方法来确定满足各种 设计约束条件下的最优设计参数。通过一元优化优化设计可以确定叶轮 进出口速度三角形以及子午面形状, 为三维流道的设计优化打下了基 础。
2. 通流部件初始形状设计
2.1喷嘴导叶
在不同工艺流程中, 液力透平需要在固定流量或变流量下运行。 对 于固定流量运行的液力透平, 可以采用固定安装角式喷嘴结构, 导叶采 用气动叶型以减少流动损失, 如图 1(a)所示。 对于变流量运行的液力透 平,为了改善液力透平的变工况性能,需要使用可调安装角式喷嘴结构。 由于要在喷嘴上安装旋转机构, 瘦长的气动型叶片不再适用, 需要寻找 合适的叶型, 图 1(b)给出了一种适合可调结构要求的导叶叶型。
2.2叶轮
叶轮的三维造型包括子午面型线、 工作轮和出口诱导轮设计。
如图 2所示, 轮盘、 轮盖的子午型线均采用椭圆曲线。
工作轮叶片考虑强度要求, 一般为径向直叶片。
液力透平出口诱导轮应满足几方面的要求: 具有良好的水力学特性; 成型后的叶片具有较好的强度特性; 便于加工、 检验等。 圆柱基面的非 可展直纹抛物面在水力性能、 强度方面、 加工工艺等方面都具有优势, 可以满足对诱导轮叶型的几方面要求。 因此, 本发明采用圆柱基面非可 展的直纹抛物面对导风轮进行叶片造型。 如图 3所示, 通过叶轮旋转轴 的平面称为子午面, 垂至于叶轮旋转轴的平面称为径向面, 垂直于叶轮 旋转轴的端面为径向端面。
根据叶轮特点, 可开发出叶轮设计程序, 图 4是程序流程图。
2.3蜗壳造型设计
蜗壳由进液管, 蜗管和环形加速器三部分组成。 在本发明中, 蜗壳 中的三元流动被简化为二元绝热的定常流动, 同时假设涡管某截面液体 的质量全部集中在该截面的型心上。 因此, 确定了流体的参数沿蜗管截 面型心连接线的变化规律, 就可以了解蜗管中流体流动情况, 并据此设 计蜗壳。
根据二元流动假设,通过蜗管任一截面 F的工质的流量 ^与截面方 位角 Θ和蜗管数 Zv呈下列关系:
G, = '^ = ί cu pdF = (^-^-2^llNplclutgalT ( 2 ) 蜗壳截面的形状, 根据不同情况, 有圆形, 椭圆形以及非对称的梨 型等。 其中, 圆形蜗壳的截面形状 Ge可表示为:
Figure imgf000009_0001
Figure imgf000009_0002
其中, Δ为蜗壳高, 为截面圆半径, 为喷嘴叶片宽度, 为速度 数, 为方位角, 0^为液体出口角, 上述各计算式中方位角 0均以隐式形式出现, 若求某一指定截面的 何参数, 可借助于计算机编程求解, 对蜗壳进行造型设计。 3. 参数化
3.1喷嘴叶片的参数化
利用两条 B样条曲线(也可采用 Bezier曲线或 NURBS等曲线描述 方法), 对喷嘴的叶片进行参数化表达。
如图 5 所示, 选用 8个控制点用来控制上下两条曲线。 其中叶片 尾翼以及前缘点 (点 1和点 8) 固定不动; 为减少使用的控制变量, 仅 通过改变其余各控制点的纵坐标(共 6个变量)来改变喷嘴叶片的形状。 喷嘴安装角亦作为一个优化变量,通过性能优化来确定,如图 503)所示。 对于可调喷嘴, 共用 7个变量对喷嘴叶片进行参数化表述。
3.2叶轮叶片参数化
如上所述, 叶轮子午型线使用了椭圆曲线, 工作轮使用了径向直叶 片,诱导轮使用了圆柱 -抛物面方程描述。叶轮叶片的参数化正是根据叶 片的初始几何形状特点展开的。
如图 4(a)、 (b)、 (c)所示, 五个造型参数 c^ c^A ^以及 <¾, 被选择 作为叶轮叶片的优化变量。
其中
a,一诱导轮外端面与径向面夹角
«3—诱导轮内端面与径向面夹角
。一诱导轮外端面平均半径处的叶片角度
<¾一工作轮叶片压力面和速度面的径向夹角
<¾一工作轮压力面和速度面的轴向夹角
综上所述, 在液力透平整级的优化中, 共采用了 12个变量, 对液 体膨胀机的喷嘴及叶轮进行参数化表达。其中 6个变量用于控制喷嘴叶 片变化, 1 个变量用于控制喷嘴安装角变化, 5个变量用于优化叶轮。 由于优化变量较多, 需要对液力透平的整机优化设计方法进行研究。 4. 整机优化设计方法
由于膨胀机整机各个部件及其相互作用(尤其是喷嘴和叶轮之间的 相互作用)对液力透平整机流场及性能有重要的影响。 鉴于此, 液力透 平的优化设计需要在整机环境下进行。 在本发明中, 在整机环境中进行 CFD 分析, 结合参数化方法和先进的协同进化遗传算法 (CCGA), 建 立了液力透平的整机优化设计方法。
4.1 目标函数
在液力透平优化中, 为了使液力透平效率提高的同时, 压降也能达 到要求, 目标函数由二者组合而成, 表示为
Figure imgf000011_0001
其中 Pr代表了液体膨胀机的膨胀比, Pr°是初始设计的膨胀比; 代 表了液体膨胀机的等熵效率, c n c2是经验系数。
4.2 CFD的自动调用
在优化计算的过程中, 目标函数 (包括效率, 膨胀比) 需要使用 CFD进行评估以更新近似模型,于是包括候选叶片形线生成、网格划分、 CFD调用计算, 以及 CFD结果的获取, 均通过自有程序以及批处理调 用 CFD流场计算软件完成。 图 6显示了一个候选设计从叶片生成到完 成目标函数值计算的完整过程:
4.3优化平台
由于优化变量个数较多、 优化问题本身高度非线性、 以及 CFD整 机数值模拟相当费时, 传统的遗传算法和近似模型技术无法保证寻优过 程的完成。 为在有限的计算资源下完成对高度非线性多变量问题的优 化, 针对基于 CFD 的多变量优化问题, 建立了一个较为通用的优化设 计平台。 该优化平台包含四个主要模块: 喷嘴导叶和叶轮叶片的参数化 生成、协同进化遗传算法(CCGA)、 自适应的近似模型技术、 以及 CFD 的自动调用计算。 在优化平台中, 这四个模块中有机地整合在一起。 图 给出了优化平台的流程图, 下面给出了对各模块较详细的描述。
( 1 ) 参数化模块: 使用样条曲线对喷嘴的叶片进行参数化, 通过改变 控制点坐标来改变喷嘴叶片的型线; 若为可调喷嘴, 也可描述喷嘴 叶片的安装角变化; 利用圆柱-抛物面方法表述叶轮叶片, 可通过调 整几何参数修改叶轮设计。 以上方法通过程序实现并方便以 CFD软 件数据接口格式输出, 用以在优化过程中自动生成喷嘴及叶轮叶片 数据。
( 2 ) 近似模型模块: 构造一定数量的组样本设计, 并在整级环境下对 样本进行 CFD分析, 获得其目标函数值, 用以建立初始化的近似模 型。 在后续的优化计算中, 此近似模型被用来替代耗时的 CFD计算 完成目标函数评估, 加速寻优的进程; 并在优化过程中根据算法需 要不断增加新的点, 以提高近似模型在潜在最优点的预测精度。
( 3 ) 变量分组算法模块: 协同进化算法中的关键问题是合理的变量分 组。 在本发明中, 通过统计遗传算法种群个体的分布数据, 获得不 同变量间的相关性数据。 基于变量之间 (紧密或松散) 的量化关系, 将优化变量分为多组(即进行变量空间分割), 以便使用协同进化遗 传算法进行优化计算。
( 4 ) 协同进化遗传算法 (CCGA)模块: 协同进化遗传算法将一个复杂 的多变量优化问题分割为多个相对独立的子问题, 并利用遗传算法 对每一个子问题进行逐歩求解, 它在考虑变量之间相关性的基础上 对问题进行合理分割, 能有效地减少寻优所需的计算量。 由于每个 遗传算法个体只拥有一部分优化变量, 当需要计算一个候选个体目 标函数值的时候, 它就需要与来自其它种群的变量进行合并, 以获 得候选个体的目标函数。 附图说明
图 1 ( a) 和 (b)分别给出了固定安装角和可调安装角喷嘴导叶叶型。 图 2叶轮子午面示意图。
图 3为叶轮叶片示意图。 原点 0选在叶轮旋转轴 O上。 0角及径向 面上的圆弧长度 Y取与叶轮旋转方向相同时为正值, 反之, 为负值, 且 都从某一工作轮叶片中面起算。 O。为中心抛物面的轴。假定中心抛物面 与外端面的交线为直线并通过这根轴。 O和 O。两根轴之间的距离为 Ra, 并令叶片顶部包角 减小时的!^为正值 为中心抛物面与外端面的相 交直线和平均半径处的径向线间夹角的轴向投影, 同样取使叶片顶部包 角 减小时的 </角为正值。
图 4叶轮设计程序流程图。
图 5喷嘴导叶叶型参数化及可调安装角示意图。
图 6给出了在液力透平整机优化过程中如何实现 CFD 自动调用的 示意图。
图 7给出了液力透平整机优化设计程序平台流程图。 具体实施方式
为了方便本发明的实施, 以下结合附图及方程等对本发明作进一歩 详细的描述。
首先, 将方程 (1 ) 所描述的一元优化模型与优化算法相结合, 考 虑各种约束条件, 通过编程求解优化问题, 确定叶轮进出口直径、 速度 三角形等, 是通流部件三维设计的基础。
第二, 基于一元优化设计结果, 设计通流部件初始形状, 包括喷嘴 导叶叶型、 叶轮计、 以及蜗壳设计。 对于固定流量运行的液力透平, 推 荐采用固定安装角式喷嘴结构, 导叶采用气动叶型以减少流动损失, 如 图 1 (a)所示。 对于变流量运行的液力透平, 为了改善液力透平的变工况 性能, 需要使用可调安装角式喷嘴结构, 相应地需要采用适合可调结构 要求的导叶叶型, 图 l(b)。 叶轮的三维造型包括子午面型线、 工作轮和 出口诱导轮设计。 如图 2所示, 液力透平叶轮轮盘、 轮盖的子午型线均 采用椭圆曲线, 工作轮叶片为径向直叶片, 出口诱导轮采用圆柱基面非 可展的直纹抛物面对导风轮进行叶片造型。 图 3(a)、 ( ), (c)分别给出了 叶轮的子午面投影, 投影, 以及诱导轮圆柱截面展开图。 按照上述 方案, 利用数值方法和计算机编程, 很容易实现叶轮三维成型, 图 4 给 出了叶轮设计程序流程图。蜗壳的设计基于蜗壳内二元绝热定常流动的 假设, 以及涡管某截面液体的质量全部集中在该截面型心上的假设。 方 程 (4) 给出了蜗壳几何参数的关系式, 通过编程求解该隐式方程可以 得到各位置角度上对应的几何参数。
第三, 基于通流部件初始形状设计, 进行几何形状的参数化并提取 优化设计变量。 如图 5所示, 利用两条样条曲线拟合喷嘴叶片的吸力面 及压力面, 获取控制点的坐标值, 将其作为初始设计表达, 并合理筛选 部分坐标值作为喷嘴翼型优化的参变量; 对于可调喷嘴, 可增加 1个变 量用于表示喷嘴安装角变化。 基于叶轮的初始几何形状特点 (即叶轮子 午型线为椭圆曲线, 工作轮为径向直叶片, 诱导轮则为圆柱 -抛物面), 对叶轮的进行参数化并选择五个变量 ^,^ 。,^以及 作为叶轮叶片 的优化变量, 如图 3(a)、(b)和 (c)所示。综上,在液力透平整级的优化中, 共采用了 12个变量, 对液体膨胀机的喷嘴及叶轮进行参数化表达。 其 中 6个变量用于控制喷嘴叶片变化, 1个变量用于控制喷嘴安装角变化, 5个变量用于优化叶轮。
第四, 液力透平的整机优化方法。 目标函数由液力透平总效率和压 比组合而成, 方程 (5)给出了其数学表述。 建立了一个基于 CFD的优化 平台, 适用于液力透平整级的优化, 其中 CFD 的自动调用通过一系列 自有程序和批处理脚本, 完成 CFD分析包括候选叶片型线生成、 网格 划分、 CFD设置和求解器自动求解、 以及数值结果获取的全部过程, 方 便与优化程序集成, 过程如图 6所示。 图 7是优化设计平台流程图, 主 要包括四个模块: 喷嘴导叶和叶轮叶片参数化、 协同进化遗传算法、 自 适应近似模型技术、 以及 CFD 自动调用。 在优化方法通过下列措施有 效减少计算量, 并加速寻优收敛: 利用近似模型预测目标函数, 而不是 每次都调用耗时的 CFD计算; 采用了有自动更新能力的近似模型算法, 不断提高模型在潜在最优区域的精度, 提高寻优效率。 协同进化算法的 引入, 使得算法具有处理多变量高度非线性问题的能力。 通过变量分组 算法, 侦测变量之间 (紧密或松散) 的量化关系, 将一个复杂的多变量 优化问题分割为多个相对独立的子问题 (将优化变量分为多组), 并利 用遗传算法对每一个子问题进行逐歩求解; 在考虑变量之间相关性的基 础上对问题进行合理分割, 考虑子问题间的相关性, 在最大限度保留原 问题特性的基础上, 有效减少寻优所需的计算量。 算法中同时使用自适 应遗传算子, 改善了协同进化算法的全局搜索能力并加速收敛。 该优化 平台的各个模块不是简单按照顺序依次进行, 而是被有机地整合在一 起, 譬如, 在建立初始近似模型时, 需要调用 CFD完成初始样本数据 库的建立, 并调用优化算法完成近似模型的拟合工作; 拟合后的近似模 型被作为目标函数, 用来指导优化的进行; 在寻优的过程中, 又会根据 算法计算的需要, 调用 CFD计算, 增加适当的样本点, 来更新近似模 型在潜在最优区域的精度, 这时候优化算法和 CFD均被调用, 近似模 型同时完成一次更新。 以上内容是结合具体的优选实施方式对本发明所作的进一歩详细 说明, 不能认定本发明的具体实施方式仅限于此, 对于本发明所属技术 领域的普通技术人员来说, 在不脱离本发明构思的前提下, 还可以做出 若干简单的推演或替换, 都应当视为属于本发明由所提交的权利要求书 确定专利保护范围。

Claims

权利要求书
1、 一种径流式液力透平优化设计方法, 其特征在于, 包含一元热力优 化设计方法, 通流部件三维造型方法以及整机优化方法; 其中, 整机优 化方法包括如下歩骤: 喷嘴导叶和叶轮叶片参数化歩骤、 协同进化遗传 算法、 自适应的近似模型以及 CFD自动调用算法。
2、 如权利要求 1所述径流式液力透平优化设计方法, 其特征在于, 按照如下歩骤:
(1)给定流量、进出口工质参数,将一元优化模型与优化算法相结合, 考虑各种约束条件, 通过编程求解优化问题, 确定叶轮进出口直径、 速 度三角形, 为通流部件三维设计打下基础;
(2)基于一元优化设计结果, 设计通流部件初始形状, 包括喷嘴导 叶叶型、 叶轮计以及蜗壳设计;
(3)在整机环境下, 对已设计的蜗壳、 喷嘴、 叶轮等通流部件的流场 进行 CFD分析, 获得初始设计的性能数据, 同时分析流动损失;
(4) 基于通流部件初始形状设计, 进行几何形状的参数化并提取优 化设计变量; 依靠 CFD 数值研究, 通过变量灵敏度分析确定优化变量 的范围; 优化目标是在整机环境下, 同时提高整机的效率以及膨胀。
(5)利用试验方法构造实验样本, 并用 CFD对实验样本进行分析, 获得其目标函数值; 利用实验样本数据建立近似模型, 替代耗时的 CFD 计算, 作为目标函数指导优化进行;
(6)利用遗传算法计算若干歩, 完成对整个设计空间的初歩探索; 分 析遗传算法种群数据, 计算变量之间紧密或松散的量化关系; 通过变量 分组算法, 将复杂的多变量优化问题分解为多个相对独立又互相影响的 子问题; (7)使用协同进化算法优化,该算法具有处理多变量高度非线性问题 的能力; 由于是在考虑变量相关性的基础上对原问题进行合理分割, 考 虑了子问题间的相关性, 在最大限度保留原问题特性的基础上, 减少了 寻优所需的计算量;
(8M吏用在优化的过程中, 使用动态采样策略更新近似模型, 自动 调用较少次数 CFD计算, 不断提高模型在潜在最优区域的精度, 提高 寻优效率;
(9)当计算量达到预定最大值时, 优化算法结束, 获得最优化设计。
3、 如权利要求 2所述径流式液力透平优化设计方法, 其特征在于: 所述一元热力优化方法, 包括如下歩骤:
(1)一元热力设计使用遗传算法, 通过计算机自动完成优化过程, 而 非传统的依靠不断试凑得出;
(2)在优化的过程中, 需要工质在通流部件不同截面的物性数据, 均 通过调用物性子程序完成, 不但保证了计算精度, 而且, 由于物性子程 序可计算不同工质, 优化程序具有设计不同介质膨胀机的能力。
4、 如权利要求 2所述径流式液力透平优化设计方法, 其特征在于: 所述整机优化方法, 包括如下歩骤:
(1)为同时考虑膨胀机各个部件对整机流场及性能的影响,优化各个 部件工作在最优化状态, 引入了协同进化算法, 使得优化方法具有处理 多变量高度非线性问题的能力; 此种算法将一个复杂的多变量优化问题 分割为多个相对独立的子问题, 并利用遗传算法对每一个子问题进行逐 歩求解; 在考虑变量之间相关性的基础上对问题进行合理分割, 在最大 限度保留原问题特性的基础上, 有效减少寻优所需的计算量;
(2)由于整机优化使用了较多变量,为保证优化在有限计算资源下完 成, 利用利用克里金近似模型预测目标函数, 而不是每次都调用耗时的 CFD计算;采用了有自动更新能力的近似模型算法,在仅调用少量 CFD 计算次数基础上,不断提高模型在潜在最优区域的精度,提高寻优效率;
(3) CFD 的自动调用通过一系列自有程序和批处理脚本, 完成 CFD 分析包括候选叶片型线生成、 网格划分、 CFD设置和求解器自动求解、 以及数值结果获取的全部过程, 可直接与优化程序集成;
(4)该优化方法的各个模块不是简单按照顺序依次进行,而是被有机 地整合在一起, 在建立初始近似模型时, 需要调用 CFD 完成初始样本 数据库的建立, 并调用优化算法完成近似模型的拟合工作; 拟合后的近 似模型被作为目标函数, 用来指导优化的进行; 在寻优的过程中, 又会 根据算法计算的需要, 调用 CFD计算, 增加适当的样本点, 来更新近 似模型在潜在最优区域的精度, 这时候优化算法和 CFD 均被调用, 近 似模型同时完成一次更新。
PCT/CN2011/084460 2011-12-22 2011-12-22 径流式液力透平优化设计方法 WO2013091222A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
PCT/CN2011/084460 WO2013091222A1 (zh) 2011-12-22 2011-12-22 径流式液力透平优化设计方法

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2011/084460 WO2013091222A1 (zh) 2011-12-22 2011-12-22 径流式液力透平优化设计方法

Publications (1)

Publication Number Publication Date
WO2013091222A1 true WO2013091222A1 (zh) 2013-06-27

Family

ID=48667671

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2011/084460 WO2013091222A1 (zh) 2011-12-22 2011-12-22 径流式液力透平优化设计方法

Country Status (1)

Country Link
WO (1) WO2013091222A1 (zh)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1724847A (zh) * 2004-07-22 2006-01-25 中国科学院工程热物理研究所 径流式涡轮径向元素叶轮等强度叶片造型及其方法
CN1828024A (zh) * 2005-03-04 2006-09-06 徐大懋 提高能量转换效率的叶轮机械叶片设计方法
CN1847664A (zh) * 2006-04-07 2006-10-18 刘昌喆 径流叶栅压气机
CN101915130A (zh) * 2010-06-25 2010-12-15 北京理工大学 可变几何涡轮增压器喷嘴环三维叶片及其设计方法

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1724847A (zh) * 2004-07-22 2006-01-25 中国科学院工程热物理研究所 径流式涡轮径向元素叶轮等强度叶片造型及其方法
CN1828024A (zh) * 2005-03-04 2006-09-06 徐大懋 提高能量转换效率的叶轮机械叶片设计方法
CN1847664A (zh) * 2006-04-07 2006-10-18 刘昌喆 径流叶栅压气机
CN101915130A (zh) * 2010-06-25 2010-12-15 北京理工大学 可变几何涡轮增压器喷嘴环三维叶片及其设计方法

Similar Documents

Publication Publication Date Title
CN102608914B (zh) 径流式液力透平优化设计方法
Kim et al. Multi-objective optimization of a centrifugal compressor impeller through evolutionary algorithms
Drtina et al. Hydraulic turbines—basic principles and state-of-the-art computational fluid dynamics applications
Kim et al. Effect of blade thickness on the hydraulic performance of a Francis hydro turbine model
Deng et al. Whole flow field performance prediction by impeller parameters of centrifugal pumps using support vector regression
CN115017843A (zh) 一种离心压缩机气动性能优化设计方法
Nejad et al. Parametric study and performance improvement of regenerative flow pump considering the modification in blade and casing geometry
Avdieieva et al. Development of the typical design of the high-pressure stage of a steam turbine
Bai et al. Optimal design and performance improvement of an electric submersible pump impeller based on Taguchi approach
Wang et al. Energy efficiency optimization of water pump based on heuristic algorithm and computational fluid dynamics
Han et al. Multiple parameters and multiple conditions optimization based on two steps strategy for an axial flow pump
Kim et al. Design optimization of mixed-flow pump in a fixed meridional shape
Kong et al. An optimization on the stacking line of low-pressure axial-flow fan using the surrogate-assistant optimization method
Maftouni et al. Effects of implementing a Diffuser around the Wind Turbine
Li et al. Investigations on ORC radial inflow turbine three-dimensional geometry design and off-design performance prediction
Sun et al. Effect of impeller and diffuser matching optimization on broadening operating range of storage pump
WO2013091222A1 (zh) 径流式液力透平优化设计方法
Sun et al. Optimization design of IGV profile in centrifugal compressor
Obrovský et al. Development of high specific speed Francis turbine for low head HPP
Rajesh et al. Design of impeller Blade by varying blades and type of blades using Analytical
Han et al. A novel method for work capacity calculation of centrifugal compressor impellers in energy storage systems
Popov et al. Optimization algorithm for axial multistage compressor workflow
CN104235055B (zh) 一种大口径弯管渣浆循环泵的水力模型设计方法
CN106971019A (zh) 一种高比转速轴流泵导叶水力设计方法
Lei et al. Three-Dimensional Multi-Objective Design Optimization of a 6.5-Stage Axial Flow Compressor Blades With Lean and Twist

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 11877976

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 11877976

Country of ref document: EP

Kind code of ref document: A1

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205N DATED 12/03/2015)

122 Ep: pct application non-entry in european phase

Ref document number: 11877976

Country of ref document: EP

Kind code of ref document: A1