Method for optimizing impeller of high-cavitation-resistance low-temperature deep-well pump

Abstract

The invention provides a high-cavitation-resistance low-temperature deep-well pump impeller optimization method. The method comprises the steps that S1, reverse parametric modeling is conducted on an impeller of a low-temperature deep-well pump; s2, the impeller subjected to reverse parametric modeling and other flow passage components are combined, numerical simulation calculation is conducted in CFX software, and the cavitation performance value of the low-temperature deep-well pump model is obtained; s3, comparing the cavitation performance value of the low-temperature deep-well pump model with the cavitation performance value in practical application, and combining the difference value after comparison with the standard cavitation performance value to obtain a target cavitation performance value of the low-temperature deep-well pump model; s4, performing test design on the input parameters of the low-temperature deep-well pump impeller and the cavitation performance of the low-temperature deep-well pump through an Isight optimization system, constructing a kriging agent model by taking a plurality of design points obtained by the test design as training points, and performing iterative solution on the constructed kriging agent model by utilizing a genetic algorithm to obtain impeller parameters meeting a target cavitation performance value, so as to obtain the low-temperature deep-well pump impeller. And cavitation performance optimization of the low-temperature deep-well pump is completed.

Description

Technical Field

[0001] The present invention relates to the technical field of low-temperature deep well pump impeller optimization, and specifically relates to an optimization method for a high cavitation-resistant low-temperature deep well pump impeller. Background Technique

[0002] Due to the special operating conditions of low-temperature deep well pumps, compared with other industries, their cavitation problems are more prominent. At present, relatively few optimization studies on existing low-temperature deep well pumps are based on intelligent optimization algorithms. Therefore, it is very meaningful to optimize the design of existing pumps to improve their cavitation resistance.

[0003] Currently, the methods for improving the cavitation performance of low-temperature deep well pumps mainly include empirical formula methods and intelligent optimization algorithm methods. For designers without rich experience, the empirical formula method requires a large amount of computational cost and time cost to find a relatively good design scheme. The intelligent optimization algorithm method can construct a response surface based on a surrogate model and quickly find the optimal model through an intelligent optimization algorithm, optimizing a design scheme with better performance with higher efficiency and faster time, and can effectively improve the cavitation performance of existing low-temperature deep well pumps.

[0004] Therefore, how to optimize a high cavitation performance design scheme for a low-temperature deep well pump within a limited time and meet the higher and more stringent cavitation performance requirements in actual applications has become a problem to be solved. Summary of the Invention

[0005] Aiming at the serious cavitation phenomenon of existing low-temperature deep well pumps under higher and more stringent cavitation requirements, the purpose of the present invention is to provide an optimization method for a high cavitation-resistant low-temperature deep well pump impeller. By combining the reverse parameterization of the existing product impeller model, neural network model and genetic algorithm, it can quickly and effectively optimize the cavitation performance of the low-temperature deep well pump, so that when operating under higher and more stringent cavitation conditions, serious cavitation phenomenon will not occur and smooth flow can be achieved.

[0006] To achieve the above purpose, the present invention adopts the following technical solutions:

[0007] An optimization method for a high cavitation-resistant low-temperature deep well pump impeller includes the following steps:

[0008] S1. Conduct reverse parameterization modeling on the impeller of the existing low-temperature deep well pump;

[0009] S2. Combine the impeller after reverse parameterization modeling with other flow components and perform numerical simulation calculations in CFX software to obtain the cavitation performance value of the low-temperature deep well pump model;

[0010] S3. Compare the cavitation performance values of the low-temperature deep-well pump model with those under actual application, and then combine the difference after comparison with the standard cavitation performance value to obtain the target cavitation performance value of the low-temperature deep-well pump model;

[0011] S4. Through the Isight optimization system, conduct experimental design on the input parameters of the low-temperature deep-well pump impeller and the cavitation performance of the low-temperature deep-well pump, use multiple design points obtained from the experimental design as training points to construct a kriging surrogate model, and then use the genetic algorithm to iteratively solve the constructed kriging surrogate model to obtain the impeller parameters that meet the target cavitation performance value, thus completing the cavitation performance optimization of the low-temperature deep-well pump.

[0012] Preferably, step S4 includes:

[0013] S41. Apply the Isight optimization system, use the impeller inlet diameter, blade inlet edge position, and blade inlet attack angle of the low-temperature deep-well pump as input parameters, and use the cavitation performance value of the low-temperature deep-well pump model as the output parameter;

[0014] S42. Select appropriate parameter variable ranges for the input parameters;

[0015] S43. Within the appropriate parameter variable ranges of the input parameters, generate multiple design points based on the optimal space-filling experimental design method for experimental design;

[0016] S44. Conduct numerical simulation calculations on multiple design points in the CFX software to obtain multiple cavitation performance values of the low-temperature deep-well pump model, and use multiple design points and the corresponding multiple cavitation performance values to construct a kriging surrogate model;

[0017] S45. Use the constructed kriging surrogate model to determine the response surface, and reconstruct the response surface by generating multiple new design points through experimental design to obtain the response surface model of the required net positive suction head;

[0018] S46. According to the response surface model of the required net positive suction head, select the genetic algorithm for iterative solution to obtain the impeller parameters that meet the target cavitation performance value.

[0019] Preferably, the cavitation performance value of the low-temperature deep-well pump model includes the required net positive suction head.

[0020] Preferably, step S2 includes:

[0021] S21. Import the impeller model after reverse parametric modeling into the Turbogird software for mesh generation, and combine the completed impeller mesh model with the mesh models of other flow components;

[0022] S22. Numerically simulate the gas-liquid two-phase flow of the combined low-temperature deep-well pump grid model in CFX software to obtain the required net positive suction head (NPSH) of the low-temperature deep-well pump model.

[0023] Preferably, step S3 includes:

[0024] S31. Compare the required NPSH obtained in step S22 with the required NPSH of the low-temperature deep-well pump under actual application to obtain the difference.

[0025] S32. Combine the difference with the standard required NPSH to obtain the target required NPSH of the low-temperature deep-well pump model.

[0026] Preferably, step S42 includes:

[0027] Control the value range of each input parameter within ±15% of the initial value of each input parameter.

[0028] Preferably, step S43 includes: within the appropriate parameter variable range of the input parameters, generate 100 design points based on the optimal space-filling experimental design method for experimental design.

[0029] Preferably, step S46 includes: set the configuration parameters of the genetic algorithm, define the individual as a binary encoding length of 6, the initial population size of 100, the maximum number of genetic generations of 100 times, the required NPSH represents the fitness, the generation gap is 0.8, the crossover probability is 0.7, and the mutation probability is 0.01.

[0030] Compared with the prior art, the present invention has the following beneficial effects:

[0031] The present invention combines the kriging surrogate model and the genetic algorithm, which can quickly and effectively obtain a hydraulic model that meets higher and more stringent cavitation performance requirements. Compared with the traditional optimization method, the present invention enables non-experienced designers to significantly improve the cavitation performance of the existing product, the low-temperature deep-well pump, within an effective time. Description of the Drawings

[0032] Figure 1 is the optimization flow chart of the low-temperature deep-well pump impeller optimization platform;

[0033] Figure 2 is the response surface representation diagram of the impeller inlet diameter parameter and the required NPSH;

[0034] Figure 3 is the response surface representation diagram of the position parameter of the blade inlet edge and the required NPSH;

[0035] Figure 4 is the response surface representation diagram of the blade inlet incidence angle parameter and the required NPSH;

[0036] Figure 5 It is a flowchart of a genetic algorithm. Specific implementation manners

[0037] In order to make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings.

[0038] Refer to Figure 1 , an embodiment of the present application provides an optimization method for the impeller of a high cavitation resistance low-temperature deep well pump, including the following steps:

[0039] S1. Perform reverse parametric modeling on the impeller of the existing product low-temperature deep well pump.

[0040] This step S1 specifically includes: in the CFturbo software, perform reverse parametric modeling on the three-dimensional model of the impeller of the low-temperature deep well pump by capturing key geometric positions.

[0041] S2. Combine the impeller after reverse parametric modeling with other flow-through components and perform numerical simulation calculations in the CFX software to obtain the cavitation performance value of the low-temperature deep well pump model.

[0042] Step S2 specifically includes:

[0043] Import the impeller model after reverse parametric modeling into the Turbogird software for mesh generation, combine the mesh model of the impeller after completion with the mesh models of other flow-through components; finally, perform gas-liquid two-phase flow numerical simulation calculations on the combined mesh model of the low-temperature deep well pump in the CFX software to obtain the cavitation performance value of the low-temperature deep well pump model.

[0044] In the embodiment of the present application, the cavitation performance value of the low-temperature deep well pump model includes the required net positive suction head.

[0045] The CFX software is a computational fluid dynamics (CFD) software tool developed by ANSYS, Inc.

[0046] S3. Compare the cavitation performance value obtained from the numerical simulation of the low-temperature deep well pump with the cavitation performance value under actual application, and then combine the difference after comparison with the standard cavitation performance value to obtain the target cavitation performance value of the low-temperature deep well pump model.

[0047] The standard cavitation performance value is a cavitation performance value with higher and stricter pre-calibrated cavitation requirements.

[0048] S4. Use the Isight optimization system to conduct experimental design on the input parameters of the impeller of the cryogenic deep well pump and the cavitation performance of the cryogenic deep well pump. Use multiple design points obtained from the experimental design as training points to construct a Kriging surrogate model, and then use the genetic algorithm to iteratively solve the constructed Kriging surrogate model to obtain the impeller parameters that meet the target cavitation performance value, thus completing the optimization of the cavitation performance of the cryogenic deep well pump.

[0049] Step S4 includes:

[0050] S41. Apply the Isight optimization system, define the impeller inlet diameter, the position of the blade inlet edge, and the blade inlet incidence angle as input parameters, and define the cavitation performance value of the cryogenic deep well pump model as the output parameter;

[0051] S42. Optimize by integrating these two parameters; to ensure meeting the requirements of the standard cavitation performance value, select an appropriate parameter variable range for the input parameters;

[0052] S43. Within the appropriate parameter variable range of the input parameters, based on the optimal space filling experimental design method, generate a large number of design points for experimental design;

[0053] S44. After numerically simulating and calculating the large number of generated design points using the CFX software, the optimization system can use these large number of design points as training points. Since the variables and the input parameters are highly non - linear, these training points can be used to construct a Kriging surrogate model;

[0054] S45. After completing the calculation of all design points, use the Kriging surrogate model to construct a response surface; when the accuracy of the response surface model meets the requirements, each input parameter has different effects on the required net positive suction head. Considering all input parameters comprehensively, determine the response surface model of the required net positive suction head;

[0055] S46. According to the response surface model of the required net positive suction head, select the genetic algorithm for solution. After a large number of iterative calculations, the optimal solution point can be found to obtain a cryogenic deep well pump model with better cavitation performance.

[0056] The above - mentioned method of the present invention combines the Kriging surrogate model and the genetic algorithm, which can quickly and effectively obtain a hydraulic model that meets higher and more stringent cavitation performance requirements. Compared with traditional optimization methods, the above - mentioned method of the present invention enables designers without rich experience to significantly improve the cavitation performance of the existing product, the cryogenic deep well pump, within an effective time.

[0057] In this embodiment, a numerical simulation is carried out on a certain low-temperature deep-well pump of an existing product to optimize the cavitation performance. Since the cavitation performance of the existing product low-temperature deep-well pump in actual application does not meet the specified requirement of the required net positive suction head (NPSH) of 1.2 m for actual application, it is necessary to optimize the parameters of the impeller of the existing product low-temperature deep-well pump.

[0058] First, the three-dimensional model of the impeller of the existing product low-temperature deep-well pump is reversely parametrically modeled in CFturbo by capturing key geometric positions; then, the impeller model data obtained from the reverse parametric modeling is imported into the Turbogird software for mesh generation. After the mesh generation of the impeller is completed, the impeller mesh model is combined with the mesh models of other flow-through components. Finally, the combined low-temperature deep-well pump mesh model is numerically simulated for gas-liquid two-phase flow in the CFX software, and the required net positive suction head (NPSH) of the low-temperature deep-well pump model in the numerical simulation is obtained as 1.6 m.

[0059] The required net positive suction head (NPSH) of 1.6 m obtained from the numerical simulation of the low-temperature deep-well pump model is compared with the required net positive suction head (NPSH) of 1.8 m under actual application. After comparison, the difference in the required net positive suction head (NPSH) of -0.2 m is combined with the standard required net positive suction head (NPSH) value of 1.2 m (where the standard required net positive suction head (NPSH) value is the required net positive suction head (NPSH) value for higher and more stringent cavitation requirements), and the target required net positive suction head (NPSH) of the low-temperature deep-well pump model after optimization is obtained as 1.0 m.

[0060] Through the input parameters constructed by the impeller reverse parametric model of the CFturbo software in the above numerical simulation process, and the cavitation performance parameters calculated by the CFX software to construct the output parameters, the Isight optimization system is applied to integrate the input parameters and output parameters, and the impeller inlet diameter, the position of the blade inlet edge, and the blade inlet attack angle are defined as the input parameters, and the required net positive suction head (NPSH) of the cavitation performance of the low-temperature deep-well pump is defined as the output parameter.

[0061] To ensure a proper matching degree between the overall structure of the low-temperature deep-well pump and the optimized impeller, the shape of the impeller inlet position is not changed significantly. To ensure meeting the specified required net positive suction head (NPSH) requirement, the variation range of each input parameter value is within ±15% of the initial value. Table 1 shows each input parameter and the value range of the input parameter.

[0062] Input parameter name [Abbreviation] Initial value Minimum value Maximum value Impeller inlet diameter (mm) [x1] 77 73 81 Blade inlet incidence angle (°) [θ1] 3 1 8 Blade inlet edge position [y1] 3 1 15

[0063] Table 1

[0064] After determining the parameter variable range of the input parameters, within this parameter variable range, based on the optimal space-filling experimental design method, 100 design points are generated. Table 2 shows the design point parameters for experimental design.

[0065]

[0066]

[0067]

[0068] Table 2

[0069] After generating 100 design points, numerical simulation calculations are performed on all the design points using the CFX software; during this numerical simulation process, the Isight optimization system uses these 100 design points as training points. Since the input parameters of the impeller and the required NPSH have a highly non-linear relationship, these training points can be used to construct a kriging surrogate model.

[0070] After completing the numerical simulation calculations of all the design points in the CFX software within the Isight optimization system, a response surface is constructed using the kriging surrogate model. To improve the accuracy of the response surface, 100 new design points can be generated through experimental design and numerical simulation calculations of these new design points are performed using the CFX software; these new design points and their simulation results are added to the kriging surrogate model to update and refine the kriging surrogate model, thereby reconstructing the response surface; this process of reconstructing the response surface can be iterated, and each iteration may improve the accuracy and reliability of the response surface. After each reconstruction of the response surface, it is necessary to verify the accuracy of the response surface so that each error value of the response surface is within 5%; as shown in Table 3, after multiple reconstructions of the response surface, the specified requirements for the accuracy of the response surface can be met.

[0071] Error type Required NPSH Mean Reciprocal Rank MRR 0.167253 Relative Mean Absolute Error RMAE 5.7318 Root Mean Square Error RMSE 0.074829 Relative Root Mean Square Error RRMSE 0.105327

[0072] Table 3

[0073] When the accuracy of the response surface model meets the specified requirements, the response surface relationship between the variable of the input parameter and the required NPSH can be input, such as Figure 2 , Figure 3 and Figure 4 shown.

[0074] By analyzing the response surface relationship between the variable of the input parameter and the required NPSH, the impeller inlet diameter, the position of the blade inlet edge, and the blade inlet incidence angle have different degrees of influence on the required NPSH. Considering the impeller inlet diameter, the position of the blade inlet edge, and the blade inlet incidence angle comprehensively, making them all at a reasonable value, the cavitation performance will be improved at the reasonable value. The response surface of the impeller inlet diameter, the position of the blade inlet edge, and the blade inlet incidence angle with respect to the required NPSH can determine the three-dimensional required NPSH response surface.

[0075] According to the obtained required NPSH response surface model, the genetic algorithm is selected. The flow model of the genetic algorithm is as Figure 5As shown in the figure, combined with the optimized mathematical model, the configuration parameters of the genetic algorithm are set as shown in Table 4.

[0076] Individual binary coding length 6 Initial population size 100 Maximum number of generations 100 Generation gap of required NPSH fitness 0.8 Crossover probability 0.7 Mutation probability 0.01

[0077] Table 4

[0078] After the settings of the genetic algorithm are completed, the optimization problem of its required net positive suction head can be described in the standard form of a mathematical expression, as shown in Table 5

[0079] Input parameter name [Abbreviation] Mathematical expression Impeller inlet diameter (mm) [x1] 73≤x1≤81 Blade inlet incidence angle (°) [θ1] 1 ≤ θ1 ≤ 8 Blade inlet edge position [y1] 1≤y1≤15 Required NPSH (m) [z] Minf = z(x1, θ1, y1)

[0080] Table 5

[0081] After the settings of the genetic algorithm and the mathematical expression of the optimization problem of the required net positive suction head are completed, through a large number of iterative calculations using numerical simulation and simulation methods, the optimal solution point is found in this large number of iterative calculations. Based on the geometric parameters of the solution point, the optimized impeller model is found, and it is combined with other flow-through components to obtain a low-temperature deep-well pump model with better cavitation performance. Table 6 shows the performance comparison of the optimized input parameter variables and the net positive suction head.

[0082] Parameter variable Before optimization After optimization Impeller inlet diameter (mm) 77 80 Blade inlet incidence angle (°) 3 5 Blade inlet edge position 3 9 Required NPSH (m) 1.6 1.0

[0083] Table 6

[0084] Obviously, the above embodiments are the preferred embodiments of the present invention to illustrate the technical solutions of the present invention, rather than to limit it, and certainly not to limit the patent scope of the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. An optimization method for the impeller of a high cavitation-resistant low-temperature deep well pump, characterized in that It includes the following steps: S1. Conduct reverse parametric modeling on the impeller of the existing low-temperature deep-well pump; S2. Combine the impeller after reverse parametric modeling with other flow-through components, and perform numerical simulation calculations in the CFX software to obtain the cavitation performance value of the low-temperature deep-well pump model; S3. Compare the cavitation performance value of the low-temperature deep-well pump model with the cavitation performance value under actual application, and then combine the difference after comparison with the standard cavitation performance value to obtain the target cavitation performance value of the low-temperature deep-well pump model; S4. Through the Isight optimization system, conduct experimental design on the input parameters of the low-temperature deep-well pump impeller and the cavitation performance of the low-temperature deep-well pump, use multiple design points obtained from the experimental design as training points to construct a kriging surrogate model, and then use the genetic algorithm to iteratively solve the constructed kriging surrogate model to obtain the impeller parameters that meet the target cavitation performance value, and complete the cavitation performance optimization of the low-temperature deep-well pump.

2. The optimization method of the impeller of the high anti-cavitation low-temperature deep well pump according to claim 1, characterized in that Step S4 includes: S41. Apply the Isight optimization system, use the impeller inlet diameter, blade inlet edge position, and blade inlet attack angle of the low-temperature deep-well pump as input parameters, and use the cavitation performance value of the low-temperature deep-well pump model as the output parameter; S42. Select appropriate parameter variable ranges for the input parameters; S43. Within the appropriate parameter variable ranges of the input parameters, generate multiple design points based on the optimal space-filling experimental design method for experimental design; S44. Perform numerical simulation calculations on multiple design points in the CFX software to obtain multiple cavitation performance values of the low-temperature deep-well pump model, and use multiple design points and the corresponding multiple cavitation performance values to construct a kriging surrogate model; S45. Use the constructed kriging surrogate model to determine the response surface, and generate multiple new design points through experimental design to reconstruct the response surface to obtain the response surface model of the required net positive suction head; S46. According to the response surface model of the required net positive suction head, select the genetic algorithm for iterative solution to obtain the impeller parameters that meet the target cavitation performance value.

3. The optimization method of the impeller of the high anti-cavitation low-temperature deep well pump according to claim 1, characterized in that, The cavitation performance value of the low-temperature deep-well pump model includes the required net positive suction head.

4. An optimization method for the impeller of a high cavitation-resistant low-temperature deep well pump according to claim 3, characterized in that, Step S2 includes: S21. Import the impeller model after reverse parametric modeling into the Turbogird software for mesh generation, and combine the completed impeller mesh model with the mesh models of other flow-through components; S22. Perform gas-liquid two-phase flow numerical simulation calculations on the combined low-temperature deep-well pump mesh model in the CFX software to obtain the required net positive suction head of the low-temperature deep-well pump model.

5. The optimization method of the impeller of a high cavitation-resistant low-temperature deep well pump according to claim 4, characterized in that, Step S3 includes: S31. Compare the required net positive suction head obtained through Step S22 with the required net positive suction head of the low-temperature deep-well pump under actual application to obtain the difference; S32. Combine the difference with the standard required net positive suction head to obtain the target required net positive suction head of the low-temperature deep-well pump model.

6. The optimization method of the impeller of a high anti-cavitation low-temperature deep well pump according to claim 2, characterized in that, Step S42 includes: Control the value range of each input parameter within ±15% of the initial value of each input parameter.

7. An optimization method for the impeller of a high cavitation-resistant low-temperature deep well pump according to claim 2, characterized in that, Step S43 includes: Within the appropriate parameter variable ranges of the input parameters, generate 100 design points based on the optimal space-filling experimental design method for experimental design.

8. The optimization method of the impeller of a high anti-cavitation low-temperature deep well pump according to claim 1, characterized in that, Step S46 includes: setting the configuration parameters of the genetic algorithm, defining the individual as a binary encoding with a length of 6, the initial population size as 100, the maximum number of genetic generations as 100 times, the required net positive suction head representing the fitness, the generation gap as 0.8, the crossover probability as 0.7, and the mutation probability as 0.01.

Images