CN112163288B

CN112163288B - Fluid mechanical blade optimal design method based on large vortex simulation

Info

Publication number: CN112163288B
Application number: CN202010831303.3A
Authority: CN
Inventors: 丁亮; 李卉; 周翔; 郑楠帧
Original assignee: 704th Research Institute of CSIC
Current assignee: 704th Research Institute of CSIC
Priority date: 2020-08-18
Filing date: 2020-08-18
Publication date: 2023-09-19
Anticipated expiration: 2040-08-18
Also published as: CN112163288A

Abstract

The invention relates to a fluid machinery blade optimal design method based on large vortex simulation, firstly, a fluid machinery blade optimal design method for establishing an implicit relation optimal proposition of a performance objective function and design variables is adopted, the matched blades are optimally designed by controlling the optimal performance objective in combination with the characteristics of a fluid machinery, then, the initial blades are parameterized by Bezier curves to determine the design variables, a sample database of the blades is formed in a constraint space by using a uniform design method, and the objective function corresponding to each sample blade is calculated by using a fluid numerical simulation method; starting from the parallel neural network, establishing an implicit relation or calculation model of the design variables and the objective function; finally, under the condition of meeting constraint conditions, carrying out iterative computation by using a global optimizing method-niche genetic algorithm, and finally obtaining the fluid mechanical blade with optimal performance. The aerodynamic resistance of the blade profile after optimization is reduced by 3.5% compared with that of the blade profile of the initial blade, and the lift-drag ratio is increased by 8.6%.

Description

Fluid mechanical blade optimal design method based on large vortex simulation

Technical Field

The invention relates to an optimal design method of a fluid mechanical blade, in particular to an optimal design method of a fluid mechanical blade based on large vortex simulation.

Background

The fluid machinery plays a very important role in national economy, especially heavy industry systems, and is a key product and an 'energy consumption consumer' of ethylene fertilizer, petroleum machinery, coal chemical industry, metallurgical industry, military systems and the like. As a core component-blade of the working performance of the fluid machinery, the structure is optimized, and the improvement of the performance efficiency has important practical significance for energy conservation, cost reduction and benefit increase.

From the product design flow, for any one design parameter or user requirement, a plurality of feasible schemes can be generally made, then the best scheme is selected preferentially, and the comparison fluid mechanical blade design can be reduced to an optimization proposition under the constraint condition. The existing fluid mechanical blade optimization design method basically belongs to flow field analysis and experience, is finally verified by experiments, and mostly utilizes hypothesis to simplify a fluid control equation or solve a flow equation in Reynolds time to obtain an explicit function relation of an objective function and a design variable, and obtains an optimization solution by a steepest descent method or a conjugate gradient method under the condition of considering constraint. The existing defects are that: 1. the functional relationship is not sufficient to reflect the fluid mechanical mechanism due to the introduction of excessive assumptions and simplifications. 2. Traditional optimization algorithms tend to fall into locally optimal solutions. 3. The effort and cost of verification experiments is enormous.

The invention patent with the Chinese application number of 200510030772.0 proposes a blade profile design method of turbine compression fluid machinery, which processes a flow control equation of an N-S partial differential equation into a flow system described by a normal differential equation set through reasonable simplification, establishes a function equation through an optimal control theory, and obtains an optimal solution by utilizing a direct optimizing method. This approach reduces to some extent the assumptions needed to obtain the objective function equation, but does not address the deficiencies and problems described above.

Disclosure of Invention

The invention aims to provide an optimal design method for a fluid mechanical blade, in particular to an optimal design method for a fluid mechanical blade based on a fine flow simulation algorithm-large vortex simulation, aiming at the defects of the prior art.

In order to achieve the above purpose, the invention adopts the following technical scheme: the optimal design method of the fluid mechanical blade based on the large vortex simulation comprises the steps of firstly, adopting an optimal design method of the fluid mechanical blade for establishing an implicit relation of a performance objective function and design variables, optimally designing matched blades by controlling an optimal performance objective in combination with the characteristics of the fluid mechanical, then parameterizing initial blades by Bezier curves to determine design variables, forming a sample database of the blades by using a uniform design method in a constraint space, and calculating an objective function corresponding to each sample blade by using a fluid numerical simulation method; starting from the parallel neural network, establishing an implicit relation or calculation model of the design variables and the objective function; finally, under the condition of meeting constraint conditions, carrying out iterative computation by using a global optimizing method-niche genetic algorithm, and finally obtaining the fluid mechanical blade with optimal performance.

Further, the fluid mechanical blade optimal design method based on the large vortex simulation comprises the following specific steps:

(1) Determining a design target according to the design requirement of the fluid product, and giving a performance target function: flow, pressure ratio, efficiency and constraint conditions, and establishing optimization life of blade design;

(2) Selecting an initial blade, and carrying out parameterization treatment on the blade by adopting a Bezier curve, so as to replace a plurality of blade type value points by using a smaller number of control parameter points, namely design variables, thereby achieving the aim of completing optimal design by using acceptable calculated quantity;

(3) Generating a blade sample database under the constraint condition by adopting a uniform design method, carrying out blade modeling, calculating grid generation and turbulent flow model selection procedures, and then carrying out flow field analysis by utilizing a fluid numerical calculation method to obtain an objective function value of a sample blade;

(4) Based on a nonlinear information processing technology, establishing an implicit relation or approximate model of design variables and performance objective functions in a sample database by means of a parallel neural network tool, and classifying a blade design problem into a corresponding optimal control form by combining constraint conditions and the approximate model;

(5) Adopting a niche genetic algorithm to carry out global optimization on the optimization proposition of the fluid mechanical blade design in a constraint space, and solving to obtain an optimized blade with the optimal objective function;

(6) Adopting a fine fluid numerical simulation method-large vortex simulation to replace experimental verification of the optimal blade so as to save time and cost, and evaluating aerodynamic performance and objective function of the optimal blade;

(7) The final step of the optimization process is convergence judgment, if the convergence condition cannot be met, the blade is put into a sample database, and a new round of optimization process is started from the step (4); and if the convergence condition is met, the optimization process is completed, and the optimal blade of the fluid machinery is designed.

The invention has the beneficial effects that:

by adopting the technical scheme, the aerodynamic resistance of the optimized blade profile is reduced by 3.5% compared with that of the original blade profile, and the lift-drag ratio is increased by 8.6%.

Drawings

FIG. 1 is a view of an initial blade and optimized blade profile;

FIG. 2 is an initial blade and optimized blade pressure coefficient.

Detailed Description

The following takes the blade optimization design of a certain axial flow fan as an example, further describes the fluid mechanical blade optimization design method based on large vortex simulation and the specific implementation mode, and performs the optimization design according to the following steps:

1. according to the product characteristics of the axial flow fan, an optimization proposition of blade design is established. Objective function: blade lift-drag ratio of direct relation energy efficiency; constraint conditions: the front and rear edge positions and the geometric inlet and outlet angles of the blade are unchanged, and the blade profile area of the optimized blade is not less than 80% of the blade profile of the original blade; optimizing the local minimum thickness of the blade profile to be not less than 85% of the local minimum thickness of the initial blade profile; the optimized blade profile drag coefficient is no greater than the initial blade profile.

2. The initial blade (shown in figure 1) is selected, the 6-order Bezier curve is adopted to fit the blade profile well, and the design variable is obtained through parameterization. And generating 50 blade sample data under the constraint condition by adopting a uniform design method to perform fluid numerical calculation, so that the corresponding blade lift-drag ratio can be obtained. Wherein, blade modeling is completed by cad and cae software, corresponding boundary conditions are given, and calculation grid generation is generated (structured grids are adopted for types, encryption processing is carried out on the near wall surface, and the grid number is determined by utilizing grid independence detection); the flow calculation adopts Fluent software, a second-order finite volume method discrete fluid control equation is adopted, a SIMPLE pressure correction method is applied to improve the convergence of flow field calculation, a three-layer fully hidden differential format is used in time, and a Reynolds stress model (wall surface is processed into a standard wall surface function) is selected by a turbulence model.

3. By means of the parallel neural network, an approximate model of parameter control points (design variables) and lift-drag ratios (objective functions) in a sample database can be established, and the constraint conditions are combined to form an optimization proposition of the blade design. And then, carrying out global optimization on the optimization proposition in a constraint space by adopting a niche genetic algorithm, and basically keeping the lift-drag ratio of the blade unchanged when the evolution iterates to 40 generations, wherein the optimization process converges, thus completing the optimization design of the blade. The obtained optimized Blade profile is shown in fig. 1, wherein Blade 1 and Blade 2 are two Blade profiles of the middle step number in optimization iteration.

4. And a fine fluid numerical simulation method, namely large vortex simulation (a mixed dynamic model is adopted by a sub-lattice model), is adopted to replace an experiment to verify the performance parameters of the optimal blade. The resulting pairs of initial blade and optimized blade pressure coefficients are shown in FIG. 2, where the data shows that the aerodynamic performance of the optimized blade profile is significantly better than the initial blade profile. Further aerodynamic force integral calculation results show that the aerodynamic resistance of the optimized blade profile is reduced by 3.5% compared with that of the initial blade profile, and the lift-drag ratio is increased by 8.6%.

Claims

1. A fluid mechanical blade optimization design method based on large vortex simulation is characterized by comprising the following steps of: firstly, an optimal design method of a fluid mechanical blade for establishing an implicit relation optimal proposition of a performance objective function and a design variable is adopted, the characteristics of the fluid machinery are combined, a matched blade is optimally designed by controlling an optimal performance objective, then, the initial blade is parameterized by a Bezier curve, the design variable is determined, a sample database of the blade is formed in a constraint space by using a uniform design method, and an objective function corresponding to each sample blade is calculated by using a fluid numerical simulation method; starting from the parallel neural network, establishing an implicit relation or calculation model of the design variables and the objective function; finally, under the condition of meeting constraint conditions, carrying out iterative computation by using a global optimizing method-niche genetic algorithm to finally obtain the fluid mechanical blade with optimal performance; the method specifically comprises the following steps:

(1) Determining a design target according to the design requirement of the fluid product, and giving a performance target function: flow, pressure ratio, efficiency and constraint conditions, and establishing an optimization proposition of blade design;