阅读说明：本技术 一种径流式透平机械一维设计评估及优化方法 (One-dimensional design evaluation and optimization method for radial-flow type turbine machinery ) 是由 谢永慧 李金星 施东波 张荻 于 2021-08-27 设计创作，主要内容包括：本发明一种径流式透平机械一维设计评估及优化方法,包括：建立备选损失模型数据库,生成初始损失模型配置,根据任一初始损失模型进行一维计算获得验证透平机械几何参数；根据验证透平机械几何参数进行建模,随后针对验证透平机械进行变工况CFD计算,后处理获得透平机械真实性能数据,总结验证透平机械几何参数、变工况边界条件集及透平机械真实性能数据获得验证数据集；以初始损失模型配置作为初始化种群,根据验证数据集中的透平机械几何参数及变工况边界条件集进行一维计算,采用遗传算法对损失模型配置进行优化,获得最佳一维损失模型配置；根据获得的最佳一维损失模型配置进行透平机械设计。本发明能够快速、准确获得最佳损失模型配置。(The invention discloses a one-dimensional design evaluation and optimization method of a radial flow type turbine machine, which comprises the following steps: establishing an alternative loss model database, generating initial loss model configuration, and performing one-dimensional calculation according to any initial loss model to obtain geometric parameters of the verification turbomachine; modeling according to geometric parameters of the verification turbine machine, then carrying out variable working condition CFD calculation on the verification turbine machine, carrying out post-processing to obtain real performance data of the turbine machine, summarizing the geometric parameters of the verification turbine machine, the variable working condition boundary condition set and the real performance data of the turbine machine to obtain a verification data set; taking initial loss model configuration as an initialization population, performing one-dimensional calculation according to geometrical parameters of the turbine machinery in the verification data set and the variable working condition boundary condition set, and optimizing the loss model configuration by adopting a genetic algorithm to obtain the optimal one-dimensional loss model configuration; and designing the turbomachinery according to the obtained optimal one-dimensional loss model configuration. The method can quickly and accurately obtain the optimal loss model configuration.)

1. A one-dimensional design evaluation and optimization method for a radial-flow turbomachine is characterized by comprising the following steps:

step 1: establishing an alternative loss model database, generating initial loss model configuration, and performing one-dimensional calculation according to any initial loss model to obtain geometric parameters of the verification turbine machinery aiming at the current application scene and application working media;

step 2: modeling according to geometric parameters of the verification turbine machine, then carrying out variable working condition CFD calculation on the verification turbine machine, carrying out post-processing to obtain real performance data of the turbine machine, summarizing the geometric parameters of the verification turbine machine, the variable working condition boundary condition set and the real performance data of the turbine machine to obtain a verification data set;

and step 3: taking initial loss model configuration as an initialization population, performing one-dimensional calculation according to geometrical parameters of the turbine machinery in the verification data set and the variable working condition boundary condition set, and optimizing the loss model configuration by adopting a genetic algorithm to obtain the optimal one-dimensional loss model configuration;

and 4, step 4: and designing the turbomachinery according to the obtained optimal one-dimensional loss model configuration.

2. The method for evaluating and optimizing the one-dimensional design of a radial flow turbomachine according to claim 1, wherein step 1 specifically comprises:

firstly, summarizing the existing various loss model formulas, establishing a loss model database, wherein the loss model of the turbo machine in one-dimensional design comprises the following types: incident loss, channel loss, wake loss, tip clearance loss, lost afterspeed, and blast loss;

the incident loss, the channel loss, the wake loss, the tip clearance loss, the afterspeed loss and the blowing loss are numbered from 1 to 6, and a one-dimensional design variable matrix x forming a loss model is (x is equal to₁,x₂,…,x₆)，x_1-6Values are integers, and different values represent different loss model formulas; randomly generating N sets of initial variables x_nWherein N is 1, … N, and N is the total number of initial variables;

according to the current application scene and the application working medium, M groups of different design parameters are given, any loss model combination is selected for one-dimensional thermodynamic design, and the geometric parameters z of M verification turbomachines are obtained_mWherein M is 1, … M, M is the total number of verification turbomachines, and M is 3-5.

3. The method of claim 2, wherein the loss model formula used in step 1 is as follows:

the common loss model formula of the incident loss is as follows:

in the formula, W_inIs the relative velocity of the inlet, beta_inIs the actual inlet relative flow angle, beta_in,optFor optimum inlet relative flow angle, Z_rIs the number of leaves, C_θ,inTangential component of absolute velocity of the inlet, M_inIs the inlet Mach number, alpha_inIs the inlet absolute air flow angle, U_inIs the inlet linear velocity, kappa is the weight coefficient, K₁The coefficient can be adjusted for the incident loss;

the common loss model formula for channel loss is as follows:

in the formula, W_inIs the inlet relative velocity, W_outIs the relative velocity of the outlet, beta_inIs the actual inlet relative flow angle, beta_in,optFor optimum inlet relative draft angle, C_inIs the inlet absolute velocity, r_inIs the inlet radius, r_CIs the mean radius of curvature, Z_rNumber of leaves, K₂Adjustable for passage lossSection coefficient;

the common loss model formula for wake loss is as follows:

wherein γ is the adiabatic coefficient, M_out,relIs the actual Mach number of the outlet, P_0,out,relIs the outlet relative pressure, p_outIs the exit density;

the common loss model formula for tip clearance loss is as follows:

in the formula of U_inTo the inlet linear velocity, Z_rIs the number of blades,. epsilon_xIs axial tip clearance, epsilon_rIs the radial tip clearance, C_xAs axial absolute velocity, C_rFor radial absolute velocity, Δ H_isFor isentropic enthalpy drop, t is the thickness of the blade, b_inIs the height of the inlet leaf r_out,shIs the radius of the outlet tip, r_out,hbIs the exit root radius, C_θ,inIs the tangential component of the absolute velocity of the inlet, C_out,mMeridional component of absolute velocity of the outlet, b_outIs the height of the outlet blade, K_x、K_r、K_xrAll the adjustable coefficients of the blade tip clearance loss are adjustable coefficients;

the common loss model formula for the residual velocity loss is as follows:

in the formula, C_outAs the absolute velocity of the outlet, C_dThe coefficient can be adjusted for the remaining speed loss;

the blast loss common loss model formula is as follows:

in the formula (I), the compound is shown in the specification,is average density, U_inIs the inlet linear velocity, r_inIs the inlet radius, r_outIs the average radius of the outlet, m is the mass flow, C_in,mIs the meridional component of the absolute velocity of the inlet, D_inIs the diameter of the inlet, A_NTo the leakage area, W_inFor relative speed, N is the wheel diameter ratio and Re isReynolds number.

4. The method according to claim 3, wherein the step 2 comprises:

for M verification turbomachines, the inlet temperature T is selected_inInlet pressure P_inInlet airflow angle alpha₁Mass flow rate ofAnd a rotational speed omega_RChanging working condition parameters; determining the variable condition parameter variation range, then Sampling all parameters in an empirical design space by adopting a Latin Hypercube Sampling mode to obtain a variable condition calculation boundary condition set b_m,l,aWherein M is 1, … M, M is the total number of verification turbomachines, L is 1, … L, L is the total number of variable operating points sampled by the Latin Hypercube Sampling mode, a is 1, … a, and a is the total number of the input variables;

subsequently, the geometrical parameter z of the turbomachine is verified for M_mCalling three-dimensional modeling software to generate a three-dimensional impeller model, introducing the obtained geometric model into meshing software to perform fluid domain meshing, and then performing variable working condition CFD (computational fluid dynamics) calculation, wherein a turbulence model in the CFD calculation is selected as an SST (Steady-shear) k-omega turbulence model, and boundary conditions are selected from a boundary condition set b obtained by sampling_m,l,a；

Preprocessing the variable working condition CFD results of M verification turbomachines to obtain normalized turbomachines real performance data y_m,l,hTrue performance data y of turbomachinery_m,l,hAnd variable working condition boundary condition set b_m,l,aThe verification data sets D, z and H are obtained by arranging, wherein M is 1, … M, M is total number of verification turbomachines, L is 1, … L, L is total number of variable operating point obtained by Sampling in a Latin Hypercube Sampling mode, H is 1, … H and H are turbomachines performance parameters, and the verification data sets D, z and z are obtained by arranging_m,b_m,l,a,y_m,l,h}；

The parameters of the turbomachinery performance include total enthalpy loss, turbomachinery power, turbomachinery efficiency and the like.

5. The method of claim 4, wherein the inlet temperature T is selected from the group consisting of temperature, and mass of the turbine_inInlet pressure P_inInlet airflow angle alpha₁Mass flow rate ofAnd speed of rotation omega of the turbomachine_RThe value range is +/-10% to +/-40% of the design value.

6. The method of claim 4, wherein the step 3 comprises:

for any individual x in genetic algorithm_iAccording to verification of geometrical parameters z of the turbomachine_mAnd variable condition boundary condition set b_m,l,aPerforming one-dimensional design to obtain normalized one-dimensional calculation performance data of the turbomachineThe mapping relationship is as follows:

wherein, f is a one-dimensional calculation process of the turbine machinery;

one-dimensional calculation of performance parameters using turbomachineryData y of actual performance of turbine_m,l,hThe error of the genetic algorithm is used as an optimization parameter of the genetic algorithm, then the genetic algorithm population number is selected to be 50-200, the cross probability is 0.6-0.8, the mutation probability is 0.01-0.1, the termination algebra is 50-100, the genetic algorithm is optimized, and the optimal one-dimensional loss model configuration x is obtained_o。

7. According to claimThe method for evaluating and optimizing the one-dimensional design of a radial flow turbomachine of claim 6, wherein in step 3, the one-dimensional calculation performance parameters of the turbomachine are evaluated by using an O indexData y of actual performance of turbine_m,l,hThe O index is a combination of the average relative error and the root mean square error, and the calculation formula is as follows:

wherein, O₁To average relative error, O₂Is root mean square error, Ψ₁、Ψ₂Is an adjustable factor.

8. The method for evaluating and optimizing the one-dimensional design of a radial flow turbomachine according to claim 1, wherein in step 4, the turbomachine design is optimized by a genetic algorithm, a wolf algorithm or a particle swarm optimization.

Technical Field

The invention belongs to the field of energy power, and particularly relates to a one-dimensional design evaluation and optimization method for a radial-flow type turbine machine.

Background

The performance of turbomachinery, which is a core component of power cycle, directly affects the power and efficiency of the cycle system. The radial-flow turbo machinery is widely applied to medium and small power cycles such as waste heat utilization, organic working media and the like due to the characteristics of small size, light weight, simple structure, easy maintenance, low price and the like.

The one-dimensional design is the first step of the design of the turbine machine, and the main geometric dimension of the turbine machine, the prediction of the thermal parameters and the performance parameters of the turbine machine can be quickly obtained through the one-dimensional design. The level of the one-dimensional design directly affects the finally obtained mechanical performance and the design period of the turbine. Loss estimation is the most important loop in one-dimensional design, as it directly affects the flow parameter variation along the way. In the long-term technological development process, institutions and researchers in various countries accumulate a lot of empirical data and formulas. However, with the continuous development of turbomachines, new application scenarios and new application working mediums add new tests to the design of turbomachines, and the applicability of traditional experience data and formulas is difficult to guarantee due to the novel turbomachines structure, the physical properties of special working mediums and the particularity of heat exchange rules. Although organizations and researchers in various countries summarize a great number of loss models, the selection of the loss model in the one-dimensional design at present depends on the experience of designers, and is very blind. In addition, the prediction accuracy of the one-dimensional design on the performance parameters of the turbomachinery under the variable working condition is often poor, and the variable working condition performance of the turbomachinery is often obtained by CFD (computational fluid dynamics) solution based on a physical model, so that the research and development cost and the time consumption of a new product are greatly increased.

Therefore, a method for evaluating and optimizing the one-dimensional design of the radial-flow turbomachine is urgently needed to be provided, an optimal loss model can be selected according to an application scene and an application working medium, so that the one-dimensional design can rapidly and efficiently design the turbomachine with excellent performance, and meanwhile, the mechanical performance and variable working condition parameters of the turbomachine are accurately predicted.

Disclosure of Invention

The invention aims to provide a one-dimensional design evaluation and optimization method for a radial-flow type turbine machine, so as to solve the existing technical problems. The method summarizes the existing various loss models to establish a loss model database, constructs a verification data set of the turbine machinery according to an application scene and an application working medium, optimizes the loss model configuration in one-dimensional design by adopting a genetic algorithm, can quickly and accurately obtain the optimal loss model configuration, and improves the accuracy of the one-dimensional design of the turbine machinery and the prediction precision of variable working conditions.

The invention is realized by adopting the following technical scheme:

a one-dimensional design evaluation and optimization method for a radial flow type turbine machine comprises the following steps:

step 1: establishing an alternative loss model database, generating initial loss model configuration, and performing one-dimensional calculation according to any initial loss model to obtain geometric parameters of the verification turbine machinery aiming at the current application scene and application working media;

step 2: modeling according to geometric parameters of the verification turbine machine, then carrying out variable working condition CFD calculation on the verification turbine machine, carrying out post-processing to obtain real performance data of the turbine machine, summarizing the geometric parameters of the verification turbine machine, the variable working condition boundary condition set and the real performance data of the turbine machine to obtain a verification data set;

and step 3: taking initial loss model configuration as an initialization population, performing one-dimensional calculation according to geometrical parameters of the turbine machinery in the verification data set and the variable working condition boundary condition set, and optimizing the loss model configuration by adopting a genetic algorithm to obtain the optimal one-dimensional loss model configuration;

and 4, step 4: and designing the turbomachinery according to the obtained optimal one-dimensional loss model configuration.

The further improvement of the invention is that the step 1 specifically comprises:

firstly, summarizing the existing various loss model formulas, establishing a loss model database, wherein the loss model of the turbo machine in one-dimensional design comprises the following types: incident loss, channel loss, wake loss, tip clearance loss, lost afterspeed, and blast loss;

the incident loss, the channel loss, the wake loss, the tip clearance loss, the afterspeed loss and the blowing loss are numbered from 1 to 6, and a one-dimensional design variable matrix x forming a loss model is (x is equal to₁,x₂,…,x₆)，x_1-6Values are integers, and different values represent different loss model formulas; randomly generating N sets of initial variables x_nWherein N is 1, … N, and N is the total number of initial variables;

according to the current application scene and the application working medium, M groups of different design parameters are given, any loss model combination is selected for one-dimensional thermodynamic design, and the geometric parameters z of M verification turbomachines are obtained_mWherein M is 1, … M, M is the total number of verification turbomachines, and M is 3-5.

A further improvement of the present invention is that the loss model formula commonly used in step 1 is as follows:

the common loss model formula of the incident loss is as follows:

in the formula, W_inIs the relative velocity of the inlet, beta_inIs the actual inlet relative flow angle, beta_in,optFor optimum inlet relative flow angle, Z_rIs the number of leaves, C_θ,inTangential component of absolute velocity of the inlet, M_inIs the inlet Mach number, alpha_inIs the inlet absolute air flow angle, U_inIs the inlet linear velocity, kappa is the weight coefficient, K₁The coefficient can be adjusted for the incident loss;

the common loss model formula for channel loss is as follows:

in the formula, W_inIs the inlet relative velocity, W_outIs the relative velocity of the outlet, beta_inIs the actual inlet relative flow angle, beta_in,optFor optimum inlet relative draft angle, C_inIs the inlet absolute velocity, r_inIs the inlet radius, r_CIs the mean radius of curvature, Z_rNumber of leaves, K₂The coefficients can be adjusted for channel losses;

the common loss model formula for wake loss is as follows:

wherein γ is the adiabatic coefficient，M_out,relIs the actual Mach number of the outlet, P_0,out,relIs the outlet relative pressure, p_outIs the exit density;

the common loss model formula for tip clearance loss is as follows:

in the formula of U_inTo the inlet linear velocity, Z_rIs the number of blades,. epsilon_xIs axial tip clearance, epsilon_rIs the radial tip clearance, C_xAs axial absolute velocity, C_rFor radial absolute velocity, Δ H_isFor isentropic enthalpy drop, t is the thickness of the blade, b_inIs the height of the inlet leaf r_out,shIs the radius of the outlet tip, r_out,hbIs the exit root radius, C_θ,inIs the tangential component of the absolute velocity of the inlet, C_out,mMeridional component of absolute velocity of the outlet, b_outIs the height of the outlet blade, K_x、K_r、K_xrAll the adjustable coefficients of the blade tip clearance loss are adjustable coefficients;

the common loss model formula for the residual velocity loss is as follows:

in the formula, C_outAs the absolute velocity of the outlet, C_dThe coefficient can be adjusted for the remaining speed loss;

the blast loss common loss model formula is as follows:

in the formula (I), the compound is shown in the specification,is average density, U_inIs the inlet linear velocity, r_inIs the inlet radius, r_outIs the average radius of the outlet, m is the mass flow, C_in,mIs the meridional component of the absolute velocity of the inlet, D_inIs the diameter of the inlet, A_NTo the leakage area, W_inFor relative velocity, N is the wheel diameter ratio and Re is the Reynolds number.

The further improvement of the invention is that the step 2 specifically comprises:

for M verification turbomachines, the inlet temperature T is selected_inInlet pressure P_inInlet airflow angle alpha₁Mass flow rate ofAnd a rotational speed omega_RChanging working condition parameters; determining the variable condition parameter variation range, then Sampling all parameters in an empirical design space by adopting a Latin Hypercube Sampling mode to obtain a variable condition calculation boundary condition set b_m,l,aWherein M is 1, … M, M is the total number of verification turbomachines, L is 1, … L, L is Latin HypeThe total number of the variable operating points obtained by Sampling in an rc tube Sampling mode, wherein a is 1, … a, and a is the total number of the input variables;

subsequently, the geometrical parameter z of the turbomachine is verified for M_mCalling three-dimensional modeling software to generate a three-dimensional impeller model, introducing the obtained geometric model into meshing software to perform fluid domain meshing, and then performing variable working condition CFD (computational fluid dynamics) calculation, wherein a turbulence model in the CFD calculation is selected as an SST (Steady-shear) k-omega turbulence model, and boundary conditions are selected from a boundary condition set b obtained by sampling_m,l,a；

Preprocessing the variable working condition CFD results of M verification turbomachines to obtain normalized turbomachines real performance data y_m,l,hTrue performance data y of turbomachinery_m,l,hAnd variable working condition boundary condition set b_m,l,aThe verification data sets D, z and H are obtained by arranging, wherein M is 1, … M, M is total number of verification turbomachines, L is 1, … L, L is total number of variable operating point obtained by Sampling in a Latin Hypercube Sampling mode, H is 1, … H and H are turbomachines performance parameters, and the verification data sets D, z and z are obtained by arranging_m,b_m,l,a,y_m,l,h}；

The parameters of the turbomachinery performance include total enthalpy loss, turbomachinery power, turbomachinery efficiency and the like.

A further development of the invention is that the inlet temperature T is_inInlet pressure P_inInlet airflow angle alpha₁Mass flow rate ofAnd speed of rotation omega of the turbomachine_RThe value range is +/-10% to +/-40% of the design value.

The invention has the further improvement that the step 3 specifically comprises the following steps:

for any individual x in genetic algorithm_iAccording to verification of geometrical parameters z of the turbomachine_mAnd variable condition boundary condition set b_m,l,aPerforming one-dimensional design to obtain normalized one-dimensional calculation performance data of the turbomachineThe mapping relationship is as follows:

wherein, f is a one-dimensional calculation process of the turbine machinery;

one-dimensional calculation of performance parameters using turbomachineryData y of actual performance of turbine_m,l,hThe error of the genetic algorithm is used as an optimization parameter of the genetic algorithm, then the genetic algorithm population number is selected to be 50-200, the cross probability is 0.6-0.8, the mutation probability is 0.01-0.1, the termination algebra is 50-100, the genetic algorithm is optimized, and the optimal one-dimensional loss model configuration x is obtained_o。

The invention is further improved in that in step 3, O index is adopted to judge one-dimensional calculation performance parameters of the turbomachineryData y of actual performance of turbine_m,l,hThe O index is a combination of the average relative error and the root mean square error, and the calculation formula is as follows:

wherein, O₁To average relative error, O₂Is root mean square error, Ψ₁、Ψ₂Is an adjustable factor.

The invention is further improved in that in step 4, the turbo-machine design adopts a genetic algorithm, a wolf algorithm or a particle swarm optimization method.

Compared with the prior art, the invention has the following beneficial technical effects:

the invention provides a method for evaluating and optimizing one-dimensional design of radial-flow turbomachinery, which is characterized in that a loss model database is established by summarizing various existing loss models, then a turbomachinery verification data set is established according to an application scene and an application working medium, and the loss model configuration in one-dimensional design is optimized by adopting a genetic algorithm, so that the optimal loss model configuration can be quickly and conveniently obtained, and the accuracy of one-dimensional design is improved. In the optimization process, the influence of variable working conditions on the one-dimensional design precision is considered, the obtained optimal loss model configuration can accurately predict the pneumatic performance of the variable working conditions of the turbine machinery, and the performance parameters of the turbine unit can be conveniently mastered in the actual operation of the turbine machinery. In addition, the optimal loss model configuration is combined with the corresponding optimization algorithm, so that the accuracy of one-dimensional design is ensured, and the research and development period of the turbomachinery is further shortened. The method disclosed by the invention is suitable for various application scenes and application working media, has strong universality, is easy to realize and has wide application prospect.

Drawings

FIG. 1 is a schematic flow diagram of a method for radial flow turbomachinery design evaluation and optimization in one dimension in accordance with the present invention;

FIG. 2 is a schematic illustration of a CFD calculation for a supercritical carbon dioxide turbomachine utilizing the present invention;

FIG. 3 is a process for optimizing a supercritical carbon dioxide turbomachinery loss model configuration utilizing the present invention;

fig. 4 is a parameter distribution of a supercritical carbon dioxide turbomachinery flow field using the present invention, in which fig. 4(a) is a pressure distribution and fig. 4(b) is a temperature distribution.

Detailed Description

Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be embodied in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art. It should be noted that the embodiments and features of the embodiments may be combined with each other without conflict. The present invention will be described in detail below with reference to the embodiments with reference to the attached drawings.

Referring to fig. 1, the present invention provides a method for evaluating and optimizing a one-dimensional design of a radial turbine, comprising the following steps:

step 1: establishing an alternative loss model database, and generating an initial loss model configuration x_nAiming at the current application scene and the application working medium, one-dimensional calculation is carried out according to any initial loss model to obtain the geometric parameter z of the verification turbine machine_m。

Firstly, summarizing the existing loss model formulas and establishing a loss model database. Loss models for turbomachinery in one-dimensional designs generally include the following categories: incident losses, channeling losses, wake losses, tip clearance losses, lost velocities and windage losses.

The common loss model formula of the incident loss is as follows:

in the formula, W_inIs the relative velocity of the inlet, beta_inIs the actual inlet relative flow angle, beta_in,optFor optimum relative gas inletFlow angle, Z_rIs the number of leaves, C_θ,inTangential component of absolute velocity of the inlet, M_inIs the inlet Mach number, alpha_inIs the inlet absolute air flow angle, U_inIs the inlet linear velocity, kappa is the weight coefficient, K₁The coefficients can be adjusted for the incident losses.

The common loss model formula for channel loss is as follows:

in the formula, W_inIs the inlet relative velocity, W_outIs the relative velocity of the outlet, beta_inIs the actual inlet relative flow angle, beta_in,optFor optimum inlet relative draft angle, C_inIs the inlet absolute velocity, r_inIs the inlet radius, r_CIs the mean radius of curvature, Z_rNumber of leaves, K₂The coefficients can be adjusted for channel losses.

The common loss model formula for wake loss is as follows:

wherein γ is the adiabatic coefficient, M_out,relIs the actual Mach number of the outlet, P_0,out,relIs the outlet relative pressure, p_outIs the outlet density.

The common loss model formula for tip clearance loss is as follows:

in the formula of U_inTo the inlet linear velocity, Z_rIs the number of blades,. epsilon_xIs axial tip clearance, epsilon_rIs the radial tip clearance, C_xAs axial absolute velocity, C_rFor radial absolute velocity, Δ H_isFor isentropic enthalpy drop, t is the thickness of the blade, b_inIs the height of the inlet leaf r_out,shIs the radius of the outlet tip, r_out,hbIs the exit root radius, C_θ,inIs the tangential component of the absolute velocity of the inlet, C_out,mMeridional component of absolute velocity of the outlet, b_outIs the height of the outlet blade, K_x、K_r、K_xrAll are adjustable coefficients of tip clearance loss.

The common loss model formula for the residual velocity loss is as follows:

in the formula, C_outAs the absolute velocity of the outlet, C_dThe coefficients can be adjusted for the remaining speed loss.

The blast loss common loss model formula is as follows:

in the formula (I), the compound is shown in the specification,is average density, U_inIs the inlet linear velocity, r_inIs the inlet radius, r_outIs the average radius of the outlet, m is the mass flow, C_in,mIs the meridional component of the absolute velocity of the inlet, D_inIs the diameter of the inlet, A_NTo the leakage area, W_inFor relative velocity, N is the wheel diameter ratio and Re is the Reynolds number.

The incident loss, the channel loss, the wake loss, the tip clearance loss, the afterspeed loss and the blowing loss are numbered from 1 to 6, and a one-dimensional design variable matrix x forming a loss model is (x is equal to₁,x₂,…,x₆)，x_1-6Values are integers, and different values represent different loss model formulas; randomly generating N sets of initial loss model configurations x_nWhere N is 1, … N, N being the total number of initial variables.

According to the current application scene and the application working medium, M groups of different design parameters are given, any loss model combination is selected for one-dimensional thermodynamic design, and the geometric parameters z of M verification turbomachines are obtained_mWherein M is 1, … M, M is the total number of verification turbomachines.

Wherein M is 3-5;

step 2: according to verification of geometrical parameters z of the turbomachine_mModeling, and subsequently varying conditions for verification of turbomachineryCFD calculation, post-processing to obtain turbine machine real performance data, summarizing and verifying turbine machine geometric parameter z_mVariable working condition boundary condition set b_m,l,aAnd turbine machine true performance data y_m,l,hObtaining a validation dataset D ═ z_m,b_m,l,a,y_m,l,h}。

For M verification turbomachines, the inlet temperature T is selected_inInlet pressure P_inInlet airflow angle alpha₁Mass flow rate ofAnd a rotational speed omega_RAn equal variation working condition parameter; determining the variable condition parameter variation range, then Sampling all parameters in an empirical design space by adopting a Latin Hypercube Sampling mode to obtain a variable condition calculation boundary condition set b_m,l,aWherein M is 1, … M, M is the total number of verification turbomachines, L is 1, … L, L is the total number of variable operating points sampled by the Latin Hypercube Sampling mode, a is 1, … a, and a is the total number of the input variables;

wherein the inlet temperature T_inInlet pressure P_inInlet airflow angle alpha₁Mass flow rate ofAnd speed of rotation omega of the turbomachine_RThe value range is +/-10% to +/-40% of the design value;

subsequently, the geometrical parameter z of the turbomachine is verified for M_mAnd calling three-dimensional modeling software to generate a three-dimensional impeller model, introducing the obtained geometric model into meshing software to perform fluid domain meshing, and then performing variable working condition CFD calculation. Selecting a turbulence model in CFD calculation as an SST (shear Stress transport) k-omega turbulence model, and selecting a boundary condition set b obtained by sampling according to the boundary conditions_m,l,a。

Preprocessing the variable working condition CFD results of M verification turbomachines to obtain normalized turbomachines real performance data y_m,l,hTrue performance data y of turbomachinery_m,l,hAnd variable working condition boundary condition set b_m,l,aOne-to-one correspondenceWherein M is 1, … M, M is the total number of verification turbomachines, L is 1, … L, L is the total number of variable operating points sampled by a Latin Hypercube Sampling mode, H is 1, … H, and H is a turbomachines performance parameter. Sorting to obtain a verification data set D ═ { z ═ z_m,b_m,l,a,y_m,l,h}。

The parameters of the turbomachinery performance include total enthalpy loss, turbomachinery power, turbomachinery efficiency and the like.

And step 3: configuring x with initial loss model_nAs an initialization population, based on the geometrical parameters z of the turbomachinery in the verification data set D_mAnd variable working condition boundary condition set b_m,l,aOne-dimensional calculation is carried out, the loss model configuration is optimized by adopting a genetic algorithm, and the optimal one-dimensional loss model configuration x is obtained_o。

For any individual x in genetic algorithm_iAccording to verification of geometrical parameters z of the turbomachine_mAnd variable condition boundary condition set b_m,l,aPerforming one-dimensional design to obtain normalized one-dimensional calculation performance data of the turbomachineThe mapping relationship is as follows:

wherein, f is a one-dimensional calculation process of the turbine machinery.

One-dimensional calculation of performance data using turbomachineryData y of actual performance of turbine_m,l,hThe O index is used as an optimization parameter of the genetic algorithm, the O index is the combination of the average relative error and the root mean square error, and the calculation formula is as follows:

wherein, O₁To average relative error, O₂Is root mean square error, Ψ₁、Ψ₂Is an adjustable factor.

Selecting genetic algorithm population number of 50-200, cross probability of 0.6-0.8, mutation probability of 0.01-0.1, and termination algebra of 50-100, and optimizing genetic algorithm to obtain optimal one-dimensional loss model configuration x_o。

And 4, step 4: configuring x according to the obtained optimal one-dimensional loss model_oThe turbo machine design can adopt optimization methods such as genetic algorithm, Hui wolf algorithm, particle swarm algorithm and the like.

Example 1

Referring to fig. 1, the supercritical carbon dioxide centripetal turbine one-dimensional design method is evaluated and optimized by using the radial flow type turbomachine one-dimensional design evaluation and optimization method of the invention, and the specific steps are as follows:

step 1, numbering incident loss, channel loss, wake loss, blade tip clearance loss, residual speed loss and blast loss from 1 to 6 to form a one-dimensional design variable matrix x of a loss model (x ═ x)₁,x₂,…,x₆)，x_1-6Values are integers, and different values represent different loss model formulas; randomly generating 50 sets of initial variables x_nWherein n is 1, … 50.

Selecting an application working medium as supercritical carbon dioxide, selecting an application scene as a centripetal power generation turbine, giving 3 groups of different design parameters, and selecting a loss model configuration x₁One-dimensional design is carried out to obtain 3 groups of verification turbomachinery design parameters and geometric parameters z₃As shown in table 1.

TABLE 1 supercritical carbon dioxide centripetal turbine design parameters and geometric parameters

Step 2, for each verification turbomachine, selecting an inlet temperature T_inInlet pressure P_inInlet airflow angle alpha₁Mass flow rate ofAnd a rotational speed omega_RAs a variable working condition parameter, the variable working condition parameter is changed within the range of +/-20% of the design value, 300 points are sampled by adopting a Latin Hypercube Sampling mode, and a boundary condition set b is obtained_3,300,a. Performing variable-condition CFD calculation on all working condition points, wherein the calculation process is as shown in FIG. 2, preprocessing the variable-condition CFD results of 3 verification turbomachines, deriving total enthalpy loss, turbomachines power and turbomachines efficiency, and obtaining real performance data y of the normalized turbomachines_3,300,h。

Step 3, with x_nAs an initial population, aiming at any individual x in genetic algorithm_iAccording to verification of geometrical parameters z of the turbomachine₃And boundary condition set b of turbine mechanical variable working condition_3,300,aOne-dimensional design is carried out to obtain normalized one-dimensional calculation performance parameters of the turbine machineryPerforming genetic algorithm optimization by using the O index as an optimization parameter of the genetic algorithm, selecting the population number of 50, the cross probability of 0.8, the mutation probability of 0.05 and the termination algebra of 50, and optimizing to obtain the optimal loss model configuration x as shown in FIG. 3_oAs shown in table 2.

TABLE 2 optimal loss model configuration

Step 4, configuring x by the optimal loss model_oAnd (3) carrying out turbomachinery design, and optimizing by adopting a genetic algorithm according to design parameters to obtain an optimal design scheme, wherein the CFD calculation result is shown in figure 4. It can be seen that the optimal loss model configuration x obtained by the present invention_oThe method can accurately predict the performance of the turbomachinery, quickly and conveniently obtain the turbomachinery with excellent performance, improve the accuracy of one-dimensional design and shorten the research and development period of the turbomachinery.

Although the invention has been described in detail hereinabove with respect to a general description and specific embodiments thereof, it will be apparent to those skilled in the art that modifications or improvements may be made thereto based on the invention. Accordingly, such modifications and improvements are intended to be within the scope of the invention as claimed.