Impeller optimization design method based on centrifugal pump pressure pulsation line spectrum suppression

Abstract

This invention belongs to the field of centrifugal pump technology, and discloses an impeller optimization design method based on the suppression of pressure pulsation line spectrum in centrifugal pumps. This method mainly obtains the pressure pulsation characteristics and internal velocity field of the impeller flow channel through numerical simulation, and obtains the various modes P of this pressure field through modal energy decomposition. mode and velocity field modes V mode And the time coefficients corresponding to each mode, and then obtain P through fast Fourier transform. mode Time coefficient line spectrum, select P mode The mode with the highest peak value in the time coefficient line spectrum is denoted as P. mode m Using P mode m Time coefficients and V of each order mode Cross-correlation analysis was performed on the time coefficients of V, which showed a high correlation. mode The modalities are recorded and denoted as V. mode x Regarding the V selected above mode x The impeller parameters are optimized to suppress pressure pulsation in the flow field structure. This invention can effectively suppress pressure pulsation within a centrifugal pump, thereby reducing vibration and noise.

Description

An Impeller Optimization Design Method Based on Suppression of Pressure Pulsation Line Spectrum in Centrifugal Pumps Technical Field

[0001] This invention relates to the fields of fluid machinery and centrifugal pump technology, and specifically to an impeller optimization design method based on the suppression of centrifugal pump pressure pulsation line spectrum. Background Technology

[0002] In engineering tests, unstable pressure pulsations during the operation of some special pumps can severely affect their safe and stable operation, causing vibration and noise, and in severe cases, damaging the pump equipment and leading to safety accidents. Therefore, effectively suppressing pressure pulsations and vibration characteristics is of practical significance for engineering tests in my country.

[0003] Pumps generate both static and dynamic pressure components, known as pressure pulsations, which are superimposed on the static pressure component like an AC signal. The pressure pulsations within a centrifugal pump are extremely complex; for the same pump, different operating conditions result in different types of pressure pulsations. Furthermore, pressure pulsations are a contributing factor to vibration and noise. Therefore, suppressing pressure pulsations within the pump can reduce vibration and noise.

[0004] For example, existing technologies CN110909422A discloses a method for predicting and optimizing the high-efficiency operating range of a centrifugal pump impeller; CN111159941A discloses a method for transient numerical simulation of the flow field inside an automotive hydraulic torque converter; CN111400941A discloses a numerical prediction method for internal backflow and cavitation of backflow vortices in a vane pump; and CN102141064A discloses a method for constructing a turbulence model using spatial filtering. Zhao Weiguo et al. studied "numerical simulation and experiment of suppressing cavitation by arranging obstacles on the surface of centrifugal pump blades." However, the design methods in existing technologies still suffer from problems such as cumbersome or limited optimization methods for pressure pulsation of the centrifugal pump impeller and poor pressure pulsation suppression effects. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of existing technologies and provide an impeller optimization design method based on the suppression of pressure pulsation line spectrum in centrifugal pumps. This method mainly obtains the pressure pulsation characteristics and internal velocity field of the impeller flow channel through numerical simulation, and obtains the various modes P of the pressure field through modal energy decomposition. mode and velocity field modes V mode And the time coefficients corresponding to each mode, and then obtain P through Fast Fourier Transform (FFT). mode Time coefficient line spectrum, select P mode The mode with the highest peak value in the time coefficient line spectrum is denoted as P. mode m Using P modem Time coefficients and V of each order mode Cross-correlation analysis was performed on the time coefficients of V, and for those V that showed a high correlation (correlation coefficient of ±0.80 to ±1.00), the results were analyzed. mode The modalities are recorded and denoted as V. mode x , where x represents the order of the mode; for the V selected above mode x The flow field structure is optimized by designing the impeller parameters to suppress pressure pulsation.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0007] An impeller optimization design method based on the suppression of pressure pulsation line spectrum in centrifugal pumps includes the following steps:

[0008] Step S1: Obtain velocity field data using numerical simulation methods, set dynamic monitoring points on the blades, obtain pressure field data of the impeller flow channel, and simultaneously plot the pressure pulsation spectrum.

[0009] Step S2: Use MATLAB software to perform modal energy decomposition on the velocity field data and the pressure field data obtained from the monitoring points, and obtain the modes P of the pressure field through modal energy decomposition. mode and velocity field modes V mode And the time coefficients corresponding to each mode;

[0010] Step S3: Use MATLAB software to analyze P mode and V mode Perform a Fast Fourier Transform (FFT) on the time coefficients and plot the time coefficient line spectrum for each mode.

[0011] Step S4: For P mode By comparing the time coefficient line spectrum plots, the mode with the largest peak value is found and denoted as P. mode m ;

[0012] Step S5: Use P mode m Time coefficients and V of each order mode Cross-correlation analysis was performed on the time coefficients, and all orders V showed high correlation (correlation coefficients ranging from ±0.80 to ±1.00). mode All are recorded and denoted as V according to their order. mode x Where x represents the order of the mode (e.g., a first-order mode is denoted as V). mode 1. The second-order mode is denoted as V. mode 2) Simultaneously record the above V mode xThe corresponding energy percentage;

[0013] Step S6: Parameterize the impeller meridional plane using the Bezier curve method, set parameter control points, move them within a specified range, and use numerical simulation to obtain impeller flow channel velocity field data under different parameters. Perform modal decomposition to obtain the V values ​​recorded in step S5. mode x The mode corresponding to the order;

[0014] Step S7: For the modes obtained in step S6 above, determine the parameter control points that can reduce the energy proportion of this mode to below 20%, and record them as the optimization parameters of the impeller;

[0015] Step S8: Without affecting the pump head and efficiency, optimize the impeller design by adjusting the position of the control points of the parameters recorded in step S7 above, with the goal of suppressing the pressure pulsation intensity and controlling the pressure pulsation spectrum.

[0016] Furthermore, in step S3 above, the specific method for performing a Fast Fourier Transform (FFT) on the time coefficient using MATLAB is as follows:

[0017]

[0018] The time-domain function f(t) is expressed as an integral of the frequency-domain function F(ω).

[0019] Furthermore, in step S5 above, P mode m and V mode Cross-correlation analysis was performed on the time coefficients, and the specific method is as follows:

[0020] (1) Write a function in MATLAB to calculate the cross-correlation coefficient. The cross-correlation function is defined as:

[0021]

[0022] (2) The obtained P mode m and V mode Cross-correlation analysis was performed on the time coefficient line spectrum to obtain the correlation coefficients, where the correlation coefficients represent the degree of correlation as shown in the table below:

[0023] Table 1. Correlation Coefficients and Correlation Strength

[0024]

[0025]

[0026] (3) Select velocity field modes with correlation coefficients of ±0.80 to ±1.00.

[0027] Furthermore, in step S6 above, the specific method for setting the optimization parameter control points is as follows:

[0028] Parameter control points p1, p2, p3, p4, p5, p6, p7, p8, and p9 are set on the impeller meridional plane. Among them, p7 and p8 are control points for the width of the closed impeller outlet, and p6 and p9 are control points for the diameter of the impeller inlet. p1 moves on the horizontal line where p9 is located. p1, p8, and p9 generate the rear cover plate profile of the impeller meridional plane through Bezier curve fitting. p2 moves on the horizontal line where p6 is located. p2, p6, and p7 generate the front cover plate profile of the impeller meridional plane through Bezier curve fitting. The horizontal and vertical coordinates of p3 both move between p4 and p5. p4 moves on the front cover plate profile, and p5 moves on the rear cover plate profile. p3, p4, and p5 generate the water flow section generation line through Bezier curve fitting, as shown in Figure 8.

[0029] Furthermore, in step S7 above, the specific method for determining the parameters that need to be optimized is as follows:

[0030] (1) Using the controlled variable method, the parameter control points are changed within a specified range. p7 and p8 move horizontally, with a range of 10% of the initial impeller outlet width; p9 and p6 move vertically, with a range of 10% of the initial impeller inlet diameter; the angle between line segments p1 and p8 and the vertical line is between 10° and -10°, thus determining the movement range of p1; the angle between line segments p2 and p7 and the vertical line is between 10° and -10°, thus determining the movement range of p2; P3, P Both P4 and P5 are dimensionless parameter control points, with values ​​ranging from [0,1]. The range of P4 represents the movement range on the front cover line of the impeller in the axial projection diagram, and the range of P5 represents the movement range on the rear cover line in the axial projection diagram. When P4 equals 0, P4 coincides with P6; when P4 equals 1, P4 coincides with P7. When both the horizontal and vertical coordinates of P3 are 1, parameter control point P3 coincides with P4; when the horizontal and vertical coordinates of P3 are 0, control point P3 coincides with P5. The specific movement range is shown in Table 2.

[0031] Table 2 Optimization parameters and parameter ranges

[0032]

[0033]

[0034] Where D1 is the initial impeller inlet diameter, D2 is the initial impeller outlet diameter, b2 is the impeller outlet width, z1 and z2 are the x-coordinates of the original parameter control points corresponding to p1 and p2, and r6, r7, r8, and r9 are the y-coordinates of the original parameter control points corresponding to p6, p7, p8, and p9.

[0035] (2) Use numerical simulation to obtain the velocity field after the control point position changes, and perform mode decomposition to obtain the modes described in step S6 above. Record the energy proportion of the mode under different parameters, and select the parameter control point that can reduce the energy proportion of the mode to below 20% as the optimization parameter.

[0036] Further optimization methods in step S8 above include:

[0037] (1) The SVR-HDMR centrifugal pump impeller optimization design method based on vector regression is adopted. Based on the decision boundary of the impeller design variable control point selected in step S7 above, SVR-HDMR training samples are generated, SVR-HDMR model is constructed, and numerical simulation is used to set dynamic monitoring points on the blades to obtain the pressure pulsation amplitude corresponding to the training samples.

[0038] The above HDMR theory is specifically as follows:

[0039] The input variable is a d-dimensional vector x = (x1, x2, x3, ... xn) d If the system output is f(x), then the relationship between the input variables and the output is:

[0040]

[0041] Where f0 is the zeroth-order function term, f i (x i ) is a first-order function term, representing x i The individual action of f(x); f i,j (x i ,x j Let ) be a second-order function term, representing two variables x. i and x j The coupling has a combined effect on f(x); similarly, f i，j，...，r (x i x j , ...x r ) represents the combined effect of r input variables on f(x);

[0042] The Cut-HDMR solution method is introduced based on the original HDMR, that is, each component function in HDMR is represented by the cut point x0 as:

[0043] f0 = f(c)

[0044] f i (x i )=f i (x i c i )-f0

[0045] f i，j (x i x j )=f i，j (x i x j c i，j )-f i (x i )-f j (x j )-f0

[0046] In the first-order function term, (x i c i )express In the vector, all elements except the i-th element are related to the center point. Same elements This represents the sampled value of the i-th element at the center point along the i-th coordinate axis; similarly, the other terms have the same meaning.

[0047] From the above, the expression for the SVR-HDMR model can be expressed as:

[0048]

[0049] Furthermore, the rules for determining the sample size are shown in Table 3:

[0050] Table 3. Rules for determining sample size

[0051]

[0052] Where n represents the number of variables;

[0053] Furthermore, the specific construction process of the centrifugal pump SVR-HDMR surrogate model is shown in Figure 2;

[0054] (2) Construct the SVR-HDMR function of variables and pressure pulsation peak based on the SVR-HDMR training sample data;

[0055] (3) After machine learning of the training samples, the simulated annealing algorithm is used to optimize the SVR-HDMR function and obtain the best design parameters;

[0056] (4) Verify whether the optimization target has been achieved through numerical simulation. If the pressure pulsation reduction after optimization is greater than 6% of the original data, the optimization target has been achieved. Otherwise, update the sample points, reconstruct the SVR-HDMR function, and repeat steps (3) and (4) until the optimization target is achieved.

[0057] The present invention provides an impeller optimization design method based on the suppression of pressure pulsation line spectrum in centrifugal pumps, which can effectively suppress pressure pulsation in centrifugal pumps, thereby reducing vibration and noise. Attached Figure Description

[0058] Figure 1 is a schematic diagram of the steps of the optimization design method for suppressing pressure pulsation line spectrum of centrifugal pumps according to the present invention.

[0059] Figure 2 is a flowchart of the construction process of the centrifugal pump SVR-HDMR model of the present invention;

[0060] Figure 3 is a schematic diagram showing the location of the pressure monitoring point on the suction surface of the impeller blades in an embodiment of the present invention.

[0061] Figure 4 is a pressure pulsation spectrum diagram of the original impeller in an embodiment of the present invention;

[0062] Figure 5 shows P in an embodiment of the present invention. mode Time coefficient line spectrum of 1;

[0063] Figure 6 shows P in an embodiment of the present invention. mode 1 and V mode 1. Correlation coefficient curve of normalized time coefficients;

[0064] Figure 7 shows the original impeller flow channel V in an embodiment of the present invention. mode 1% of the energy;

[0065] Figure 8 is a schematic diagram of the control points for the optimized parameters of the impeller meridional plane in an embodiment of the present invention;

[0066] Figure 9 shows the impeller flow channel V of P1 when z = 20 in an embodiment of the present invention. mode 1% of the energy;

[0067] Figure 10 is a schematic diagram of the optimized impeller meridional plane in an embodiment of the present invention;

[0068] Figure 11 is a comparison diagram of pressure pulsation between the initial impeller and the optimized impeller in an embodiment of the present invention. Detailed Implementation

[0069] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0070] The present invention will now be described in further detail with reference to the accompanying drawings.

[0071] The optimization design method of this invention mainly obtains the pressure pulsation characteristics and internal velocity field of the impeller flow channel through numerical simulation, and obtains the various modes P of the pressure field through modal energy decomposition. mode and velocity field modes V mode And the time coefficients corresponding to each mode, and then obtain P through Fast Fourier Transform (FFT). mode Time coefficient line spectrum, select P mode The mode with the highest peak value in the time coefficient line spectrum is denoted as P. mode m Using P mode m Time coefficients and V of each order mode Cross-correlation analysis was performed on the time coefficients of V, and for those V that showed a high correlation (correlation coefficient of ±0.80 to ±1.00), the results were analyzed. mode The modalities are recorded and denoted as V. mode x , where x represents the order of the mode; for the V selected above mode x The flow field structure is optimized by designing the impeller parameters to suppress pressure pulsation.

[0072] In this embodiment, a low specific speed single-stage single-suction centrifugal pump was selected. The test object was a closed impeller model with 5 blades. The pump's specific speed (ns) was 40.1, the design speed was 3000 rpm, the design head was 35 m, and the design flow rate was 10 m³ / h. 3 / h.

[0073] Figure 1 is a schematic flowchart illustrating the steps of the impeller optimization design method based on the suppression of pressure pulsation line spectrum in centrifugal pumps according to the present invention. The main steps are as follows:

[0074] Step S1: Use numerical simulation to set dynamic monitoring points on the blades. The specific locations are shown in Figure 3. Ten monitoring points on the suction surface are evenly distributed on the blades, and the monitoring points on the pressure surface are symmetrically distributed with the suction surface. Then, acquire the pressure pulsation data of the impeller blades and the velocity field data of the impeller flow channel, and plot the pressure pulsation spectrum diagram, as shown in Figure 4.

[0075] Step S2: Use MATLAB software to analyze P mode and V mode Perform a Fast Fourier Transform (FFT) on the time coefficients and plot the time coefficient line spectrum for each mode.

[0076] Step S3: Use MATLAB software to process low-order P mode and V mode Perform a Fast Fourier Transform (FFT) on the time coefficients and plot the time coefficient line spectrum for each mode.

[0077] Step S4: For P mode By comparing the time coefficient line spectrum plots, the mode with the largest peak value is found. In this example, it is P. mode1 The time coefficient line spectrum is shown in Figure 5.

[0078] Step S5: Use P mode m Time coefficients and V of each order mode Cross-correlation analysis was performed on the time coefficients, and modes exhibiting high correlation (correlation coefficients ranging from ±0.80 to ±1.00) were recorded and denoted as V. mode x , where x represents the order of the mode, and the energy percentage of that mode is recorded. After calculation, V in this embodiment... mode 1 and P mode 1. High correlation is achieved, P mode 1 and V mode The normalized correlation coefficient curve of the cross-correlation analysis of the time coefficient is shown in Figure 6. The energy proportion of the first-order mode of the velocity field is 24.5%, as shown in Figure 7.

[0079] Step S6: The optimization design method aimed at reducing pressure pulsation without affecting other pump performance characteristics is as follows:

[0080] (1) In this embodiment, the front and rear cover plates and the inlet are fitted using Bezier curves in the axial projection diagram of the centrifugal pump impeller, and they are parameterized, as shown in Figure 8. By changing 9 parameter control points, the velocity field after the control point position changes is obtained using numerical simulation, and the first-order mode is obtained by modal decomposition. At the same time, the energy ratio of the mode under different parameters is obtained, and the parameter control points that can reduce the energy ratio to below 20% are recorded. In this embodiment, the original impeller control point P1 has an energy ratio of 9.5% when z = 20, which is lower than 20%, as shown in Figure 9. The same method is used to calculate the energy ratio of the other parameter control points. Through experiments, it can be found that changing control points P1, P2, P3, P4, and P5 can all produce coordinate points that can reduce the energy ratio of the first-order mode of the velocity field to below 20%. Therefore, these are used as design parameters that need to be optimized.

[0081] Among them, P3, P4, and P5 are dimensionless control points with a value range of [0,1]. The range of P4 represents the movement range on the front cover line of the impeller in the axial projection diagram, and the range of P5 represents the movement range on the rear cover line in the axial projection diagram. When P4 equals 0, P4 coincides with P6; when P4 equals 1, P4 coincides with P7. When both the horizontal and vertical coordinates of P3 are 1, control point P3 coincides with P4; when the horizontal and vertical coordinates of P3 are 0, control point P3 coincides with P5. The optimization parameters and parameter ranges are shown in Table 4.

[0082] Table 4 Optimization parameters and parameter ranges

[0083]

[0084] Table 5 Optimization Parameter Boundaries

[0085]

[0086]

[0087] (2) Generate SVR-HDMR training samples based on the decision boundary of the impeller design variable control point, construct the SVR-HDMR model, and perform numerical simulation calculations. Set dynamic monitoring points to obtain the corresponding pressure pulsation peak value A of the training samples as shown in Table 6 (to avoid the appearance of singular data in the training samples, the units of x1, x2 and x3 are converted to m when constructing the SVR-HDMR model).

[0088] Table 6 Training Samples for Variable x1 in SVR-HDMR Model

[0089] Serial Number x1 x2 x3 x4 x5 x6 A0 0.014 0.011 0.15 0.60 0.30 0.5144343.241 0.016 0.011 0.15 0.60 0.30 0.5142252.852 0.018 0.011 0.15 0.60 0.30 0.5134962.613 0.020 0.011 0.15 0.60 0.30 0.5125464.50 surface

[0090] (3) Fit the SVR-HDMR function of variable x1 and pressure pulsation amplitude based on the training sample data. The other variables are similar to x1, so they will not be listed again.

[0091] (4) After machine learning of the training samples, the fitting function of the pressure pulsation amplitude is optimized by the simulated annealing algorithm to obtain the best design parameters. In this embodiment, the impeller meridional surface after optimization is shown in Figure 10. It can be seen that compared with the original impeller, the curvature of the front cover plate is larger and the curvature of the rear cover plate is smaller, and the flow channel is therefore wider.

[0092] (5) Numerical simulation was performed on the optimized impeller and the pressure pulsation amplitude was compared with that of the original impeller. As shown in Figure 11, it can be clearly seen that the pressure pulsation amplitude has decreased.

[0093] The present invention provides an impeller optimization design method based on the suppression of pressure pulsation line spectrum in centrifugal pumps, which can effectively suppress pressure pulsation in centrifugal pumps, thereby reducing vibration and noise.

[0094] The above embodiments are illustrative of the present invention and not intended to limit the invention. It is understood that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the invention. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. An impeller optimization design method based on the suppression of pressure pulsation line spectrum in centrifugal pumps, comprising the following steps: Step S1: Obtain velocity field data using numerical simulation and set dynamic monitoring points on the blades to acquire pressure field data in the impeller flow channel, while simultaneously plotting the pressure pulsation spectrum. Step S2: Use MATLAB software to perform modal energy decomposition on the velocity field data and the pressure field data acquired from the monitoring points, obtaining the modes P of each order of the pressure field through modal energy decomposition. mode and velocity field modes V mode and the time coefficients corresponding to each mode; Step S3: Use MATLAB software to analyze P mode and V mode Perform a Fast Fourier Transform on the time coefficients and plot the time coefficient line spectra for each mode; Step S4: Perform a Fast Fourier Transform on P mode By comparing the time coefficient line spectrum plots, the mode with the largest peak value is found and denoted as P. modem Step S5: Use P modem Time coefficients and V of each order mode Cross-correlation analysis was performed on the time coefficients of all orders V that showed high correlation. mode All are recorded and denoted as V according to their order. modex Where x represents the order of the mode, and the above V is recorded. modex The corresponding energy percentage; Step S6: Parameterize the impeller meridional plane using the Bezier curve method, set parameter control points, move them within a specified range, and use numerical simulation to obtain impeller flow channel velocity field data under different parameters, perform modal decomposition, and obtain the V recorded in step S5. modex Step S7: For the modes obtained in Step S6 above, determine the parameter control points that can reduce the energy proportion of the mode to below 20%, and record them as the optimization parameters of the impeller; Step S8: Without affecting the pump head and efficiency, optimize the impeller design by adjusting the position of the parameter control points recorded in Step S7 above, with the goal of suppressing the pressure pulsation intensity and controlling the pressure pulsation spectrum.

2. The impeller optimization design method based on the suppression of centrifugal pump pressure pulsation line spectrum as described in claim 1, characterized in that, In step S6 above, the specific method for setting the optimization parameter control points is as follows: set parameter control points p1, p2, p3, p4, p5, p6, p7, p8, and p9 on the impeller meridional plane, where p7 and p8 are the control points for the width of the closed impeller outlet, and p6 and p9 are the control points for the impeller inlet diameter. p1 moves on the horizontal line where p9 is located. p1, p8, and p9 generate the impeller rear cover plate profile on the impeller meridional plane by fitting Bezier curves. p2 moves on the horizontal line where p6 is located. p2, p6, and p7 generate the impeller front cover plate profile on the impeller meridional plane by fitting Bezier curves. The horizontal and vertical coordinates of p3 both move between p4 and p5. p4 moves on the front cover plate profile, and p5 moves on the rear cover plate profile. p3, p4, and p5 generate the water flow section generation line by fitting Bezier curves.

3. The impeller optimization design method based on the suppression of centrifugal pump pressure pulsation line spectrum as described in claim 2, characterized in that, In step S7 above, the specific method for determining the parameters to be optimized is as follows: (1) The control variable method is used to change the parameter control points within the specified position range. p7 and p8 move in the horizontal direction, and the movement range is 10% of the initial impeller outlet width; p9 and p6 move in the vertical direction, and the movement range is 10% of the initial impeller inlet diameter; the angle between line segment p1p8 and the vertical line is between 10° and -10°, thus determining the movement range of p1; the angle between line segment p2p7 and the vertical line is between 10° and -10°, thus determining the movement range of p2; P3, P4 and P5 are all dimensionless parameter control points, and the value range is [0,1]; the range of p4 represents the movement range on the front cover line in the axial projection diagram, and the range of p5 represents the movement range on the rear cover line in the axial projection diagram; when P4 equals When the coordinates are 0, P4 and P6 coincide; when P4 equals 1, P4 and P7 coincide. When both the x and y coordinates of P3 are 1, control point P3 coincides with P4; when the x and y coordinates of P3 are 0, control point P3 coincides with P5. The specific position movement range of each parameter control point is as follows: Control point coordinate position movement range: P1z[z1-D2*tan10°, z1+D2*tan10°] P2z[z2-(D2-D1)*tan10°, z2+(D2-D1)*tan10°] P6r[r6-0.1*D1, r6+0.1*D1] P7z[r7-0.1*b2, r7+0.1*b2] P8z[r8-0.1*b2, r8+0.1*b2] P9r[r9-0.1*D1, r9+0.1*D1] Where D1 is the initial impeller inlet diameter, D2 is the initial impeller outlet diameter, b2 is the impeller outlet width, z1 and z2 are the abscissas of the original parameter control points corresponding to p1 and p2, and r6, r7, r8, and r9 are the ordinates of the original parameter control points corresponding to p6, p7, p8, and p9; (2) Use numerical simulation to obtain the velocity field after the control point position changes, and perform modal decomposition to obtain the modes described in step S6 above, and record the energy proportion of the mode under different parameters, and select the parameter control point that can reduce the energy proportion of the mode to below 20% as the optimization parameter.

4. The impeller optimization design method based on the suppression of centrifugal pump pressure pulsation line spectrum as described in claim 3, characterized in that, In step S8 above, the specific method for optimizing the impeller meridional plane based on the recorded parameter control points includes: (1) using the high-dimensional model SVR-HDMR centrifugal pump impeller optimization design method based on vector regression, generating SVR-HDMR training samples based on the decision boundary of the impeller design variable control points selected in step S7 above, constructing the SVR-HDMR model, and using numerical simulation to set dynamic monitoring points on the blades to obtain the pressure pulsation amplitude corresponding to the training samples; wherein the expression of the SVR-HDMR model is expressed as: (2) Construct the SVR-HDMR function of variables and pressure pulsation peak based on the SVR-HDMR training sample data; (3) After machine learning of the training samples, use the simulated annealing algorithm to optimize the SVR-HDMR function and obtain the best design parameters; (4) Verify whether the optimization target has been achieved through numerical simulation. If the pressure pulsation reduction after optimization is greater than 6% of the original data, the optimization target has been achieved. Otherwise, update the sample points, reconstruct the SVR-HDMR function, and repeat steps (3) and (4) until the optimization target is achieved.

Images