A casing structure optimization design method for suppressing electromagnetic noise of forced air-cooled induction motors

Abstract

The present invention discloses a casing structure optimization design method for suppressing electromagnetic noise of a forced air-cooled induction motor, comprising the following steps: Step 1: establishing a casing physical model with different heat dissipation rib structures, performing modal and heat dissipation simulations in finite element simulation software, and selecting a structure that helps to strengthen structural rigidity while ensuring heat dissipation performance and production process; Step 2: performing multi-parameter and multi-objective optimization of the casing structure of the improved scheme based on support vector regression (SVR) and the second-generation non-dominated sorting genetic algorithm (NAGA‑Ⅱ); Step 3: performing multi-physical field coupling simulation verification after structural optimization. In order to explore a noise reduction design method for a forced air-cooled induction motor and to make up for the low efficiency defect of direct optimization, the present invention takes the casing as the optimization object and proposes a multi-parameter and multi-objective indirect optimization design method that can comprehensively consider the heat dissipation performance of the motor and reduce noise, with little impact on other motor properties and low optimization cost.

Description

Technical Field

[0001] The present invention relates to the technical field of motor optimization, and in particular to a casing structure optimization design method for suppressing electromagnetic noise of a forced air-cooled induction motor. Background Art

[0002] Forced-air-cooled induction motors are essential power equipment for both industrial manufacturing and daily life. They boast low cost, mature technology, simple structure, and reliable operation, making them widely used in electric traction equipment. In traditional motor optimization technology, in addition to high efficiency, low consumption, and intelligent design goals, noise is also an essential indicator for evaluating a motor's overall performance. Electromagnetic noise has a significant impact and offers significant optimization benefits, making it a key target for analysis and suppression.

[0003] Existing optimization methods for suppressing electromagnetic noise in induction motors can be broadly categorized into direct and indirect optimization. Direct optimization involves establishing an analytical or finite element model of the motor, using electromagnetic or structural design parameters as optimization variables, calculating the noise as the optimization target, and then using a heuristic optimization algorithm to directly optimize the noise results. However, calculating motor noise involves multi-physics coupling, and analytical models generally have low accuracy. Direct optimization using multi-physics coupling finite element models is time-consuming, making indirect optimization more popular in commercial applications.

[0004] In terms of motor structural design, optimization of structural parameters primarily focuses on the stator and rotor. There are relatively few multi-parameter, multi-objective optimization methods for casing design. In particular, there is a lack of research on optimization design methods that comprehensively consider the heat dissipation performance and noise reduction of the motor casing. The casing is a significant source of electromagnetic noise radiation in forced-air-cooled motors and also accelerates heat dissipation. Optimizing the casing design can help these motors further reduce noise while ensuring other performance characteristics. Analysis of the motor vibration analytical model reveals that changing the casing's stiffness or mass can affect the amount of vibration on the casing surface, and thus the radiated noise. Therefore, for forced-air-cooled motors with relatively weak casing structures, increasing the overall casing stiffness and adjusting the heat dissipation rib structure can help reduce noise.

[0005] Therefore, in order to improve the noise reduction design of forced air-cooled motors and avoid the low efficiency problem caused by direct optimization, the present invention takes the casing as the optimization object and proposes a multi-parameter and multi-objective indirect optimization method that can comprehensively consider the motor heat dissipation performance and noise reduction. Summary of the Invention

[0006] The purpose of the present invention is to solve the problems raised in the above background technology and provide a casing structure optimization design method for suppressing the electromagnetic noise of a forced air-cooled induction motor.

[0007] To solve the above technical problems, the technical solution of the present invention is: a casing structure optimization design method for suppressing electromagnetic noise of a forced air-cooled induction motor, comprising the following steps:

[0008] Step 1: Create physical models of the chassis with different heat dissipation rib structures, perform modal and heat dissipation simulations in finite element simulation software, and select a structure that helps strengthen structural rigidity while ensuring heat dissipation performance and production process. Step 1 includes: 1.1 Create physical models of various heat dissipation rib structure schemes; 1.2 Perform modal simulations on the chassis models of various schemes; 1.3 Select the heat dissipation rib structure scheme that best improves chassis rigidity; 1.4 Select the heat dissipation rib structure scheme that minimizes heat dissipation.

[0009] Step 2: Based on support vector regression (SVR) and the second-generation non-dominated sorting genetic algorithm (NAGA-II), the casing structure of the improved scheme is optimized for multiple parameters and multiple objectives. Step 2 includes: 2.1 Establishing an optimization mathematical model for the casing structure; 2.2 Establishing a parameterized physical model; 2.3 Calculating the overall sensitivity change within the value range of each parameter variable; 2.4 Establishing a parameter sample set and performing modal simulation calculations; 2.5 Establishing an optimization agent model based on SVR; 2.6 Using NAGA-II to perform parameter optimization.

[0010] Step 3: Perform multi-physics coupling simulation verification after structural optimization; Step 3 includes: 3.1 establishing a multi-physics coupling simulation model of the motor; 3.2 calculating the electromagnetic noise results of the motor before and after optimization.

[0011] In the above-mentioned casing structure optimization design method for suppressing electromagnetic noise of a forced air-cooled induction motor, step 1.1 is: based on the basic structural characteristics of the original casing, a physical model of the casing with different heat dissipation rib structures is established; taking a vertical induction motor for a centrifugal pump as an example, nine heat dissipation rib structure schemes are established for comparison, among which Scheme I has a casing including a set of equidistantly arranged annular heat dissipation ribs, Scheme II is the original casing, including four sets of 12 axial heat dissipation ribs, Scheme III has the heat dissipation ribs tilted at a certain angle. A large tilt angle will affect the gas flow in some air ducts, and this scheme sets the tilt angle to 15°. Scheme IV has a casing with wavy heat dissipation ribs, the number of which is the same as that of Scheme I, and Scheme V has a casing with spiral heat dissipation ribs, and this scheme sets the spiral angle to 15°. Schemes VI-IX are combined structures of annular heat dissipation ribs and other types of heat dissipation ribs; in order to control variables, the effects of different heat dissipation rib structures on stiffness are studied, and the theoretical masses of each model are controlled to be close to each other.

[0012] In the above-mentioned casing structure optimization design method for suppressing electromagnetic noise of a forced air-cooled induction motor, step 1.2 is: using finite element simulation software to perform modal simulation calculations on each casing physical model to solve the first six free modal natural frequencies of each scheme; because the mass of each model has been controlled to remain basically consistent, the calculated natural frequencies can better reflect the overall radial stiffness of the casing.

[0013] In the above-mentioned casing structure optimization design method for suppressing electromagnetic noise of a forced air-cooled induction motor, step 1.3 is: comprehensively analyze the modal calculation results and casing processing technology of various heat dissipation rib schemes, and select a structural scheme that is easy to produce and process, has a suitable appearance, and has a higher low-order natural frequency; compared with the original casing scheme, the low-order natural frequencies of various other schemes are all improved, among which schemes VI-IX with added circumferential ribs have a more obvious optimization effect, and are selected as alternative schemes.

[0014] In the above-mentioned casing structure optimization design method for suppressing electromagnetic noise of a forced air-cooled induction motor, step 1.4 is: establishing an equivalent simplified finite element model of the casing of each alternative scheme, and performing fluid heat dissipation simulation on the outer surface of the casing. To simplify the simulation process, only the stator core and the casing are considered in the model, and an ideal working state is set. The heat loss is estimated based on the efficiency of the motor, and the stator core is set as the only heat source. Only the heat exchange between the casing surface and the air is considered. After the results are solved, the heat dissipation performance of each scheme can be quickly compared and analyzed to select the scheme that is most conducive to heat dissipation or has the least impact on heat dissipation. In this test, the heating power of the stator core is set to 178482W / m3, and the axial temperature distribution of the outer surface of the casing of each scheme is solved. In addition, because the casing in the test is an aluminum alloy die-casting, the final improvement scheme is selected by comprehensively considering the layout of the heat dissipation ribs, the direction of demolding, the motor heat dissipation method and the simulation results.

[0015] In the above-mentioned casing structure optimization design method for suppressing electromagnetic noise of forced air-cooled induction motor, step 2.1 is: determine the structural parameter variables and optimization targets to be optimized, wherein the selection of the structural parameter variation range needs to take into account the production and processing technology, and the specific values ​​can be selected with the help of engineering experience; the optimization target can be determined as the maximum first-order natural frequency and the minimum mass change, because when the mass change is not large, the first-order natural frequency can better reflect the change of radial stiffness, and controlling the mass is also conducive to the lightweight design of the motor; in this test method, a total of q number of heat dissipation ribs are selected w , number of circumferential reinforcement q r , housing length l, housing thickness t, heat dissipation rib height h w 、Thickness of heat dissipation rib t w 、Slope of heat dissipation rib i w , circumferential reinforcement height h r, hoop reinforcement thickness t r , hoop reinforcement inclination i r Ten structural parameters such as and so on are calculated, and the initial values ​​and value ranges of each structural parameter are summarized; the first-order natural frequency f n＝1 The optimization goal is to maximize and minimize the mass change Δm. The optimization model of the casing can be described as:

[0016]

[0017] stp i.min ≤p i ≤p i,max ,p i =[q w ,q r ,l,t,h w ,t w ,i w ,h r ,t r ,i r ]

[0018] In the above-mentioned casing structure optimization design method for suppressing electromagnetic noise of a forced air-cooled induction motor, step 2.2 is: based on the determined structural parameters, parameter modeling is performed on the physical model of the casing to facilitate subsequent variable parameter simulation calculations in finite element simulation software; step 2.3 is: evenly select a points within the value range of each parameter, and perform modal solution on the casing model in the finite element software, while keeping other parameters unchanged, to obtain the response relationship of the natural frequency and mass with the change of each parameter; the sensitivity value is represented by the coefficient obtained by least squares fitting the response result. Here, in order to avoid the influence caused by the difference in the dimension and value range of each parameter, the basic definition of sensitivity is multiplied by the value interval of each parameter. The definition of sensitivity is described as:

[0019]

[0020] Among them, s ωn 、s m are the nth order natural frequency and mass sensitivity, c ωn 、c m is the coefficient of least square fitting of the response results, that is, the approximate slope of the response curve, Δp i is the value interval of the i-th parameter; in this test method, a=6 is taken, and the sensitivity value between each parameter variable and the low-order natural frequency and mass is selected. By comparing the sensitivity, the hoop reinforcement inclination i with less influence is eliminated. r , and optimize the remaining 9 structural parameters.

[0021] In the above-mentioned casing structure optimization design method for suppressing electromagnetic noise of a forced air-cooled induction motor, step 2.4 is: selecting structural parameters that have a more significant impact on the optimization target, setting the variation range of each structural parameter in the finite element simulation software, using the Latin hypercube sampling method NAGA-Ⅱ to select b groups of parameter variable samples, and performing modal simulation calculations on the casing structure under each parameter combination; wherein the value of b is adjusted according to the computational cost. In theory, the larger b is, the higher the accuracy of the subsequent proxy model establishment. In this test method, b is set to 400.

[0022] In the above-mentioned casing structure optimization design method for suppressing electromagnetic noise of forced air-cooled induction motor, step 2.5 is: randomly divide the calculated group b samples into x training sets and y validation sets, use the training sets to train the SVR model, use the grid search method to adjust the model parameters, build the corresponding SVR proxy model, and then use the validation set to test the model to determine the coefficient R 2 , root mean square proportional error RMSPE, mean absolute percentage error MAPE measure the goodness of fit of the model; in this test method, take x = 360, y = 40, and get the prediction results and evaluation indicators of the validation set. The R 2 The values ​​are greater than 0.99, and the RMSPE and MAPE are less than 1%, indicating a good fit.

[0023] In the above-mentioned casing structure optimization design method for suppressing electromagnetic noise of forced air-cooled induction motors, step 2.6 is: relying on the established SVR agent model, parameter optimization is performed on the two optimization objectives, and after several iterations, the Pareto frontiers of the two optimization objectives are searched in the variable space, and the optimal solution is selected; the optimization results in this test are obtained. Considering that the lightweight goal and the improvement of structural rigidity are relatively contradictory, this shows a similar law in most structural optimization problems. For the optimization goal of the casing in this example, improving the radial stiffness of the casing has a higher priority than controlling the weight. It is only necessary to ensure that the change in mass is as small as possible. The sample as the optimal solution is selected, and the optimized first-order natural frequency and mass change are obtained. The first-order natural frequency is increased by 84.2%, and the mass remains basically unchanged, which proves that the radial stiffness of the casing has been significantly improved.

[0024] In the above-mentioned casing structure optimization design method for suppressing the electromagnetic noise of a forced air-cooled induction motor, step 3.1 is: establishing an electromagnetic, structural, and acoustic multi-physics field coupling simulation model of the original motor and the optimized motor based on finite element simulation analysis software; in this test method, the basic framework of the coupling simulation model is established; step 3.2 is: in this test method, the A-weighted sound pressure level spectrum results of the casing at a position 0.5m away from the outer surface of the casing before and after optimization are obtained, and the total sound pressure level drops from 81dBA to 72dBA, indicating that after the casing uses the optimization method proposed in the present invention, the electromagnetic noise of the motor is significantly reduced.

[0025] In the above-mentioned casing structure optimization design method for suppressing electromagnetic noise of a forced air-cooled induction motor, the transmission wheel is provided with a transmission tooth portion, and the transmission wheel is meshed with the motor shaft of the motor through the transmission tooth portion for transmission.

[0026] The beneficial effects of the present invention are:

[0027] In order to explore the noise reduction design method of forced air-cooled induction motors and make up for the low efficiency defect of direct optimization, the present invention takes the casing as the optimization object and proposes a multi-parameter and multi-objective indirect optimization design method that can comprehensively consider the motor heat dissipation performance and noise reduction. This method only improves the casing based on the analysis of the vibration and noise characteristics of the prototype motor, has little impact on other motor performance, and has a low optimization cost. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 It is the optimization flow chart of the present invention;

[0029] Figure 2 It is a schematic diagram of various heat dissipation rib structure scheme models in the present invention;

[0030] Figure 3 This is a comparison of modal simulation results of various heat dissipation rib structure solutions in the present invention;

[0031] Figure 4 This is the simulation result of the axial temperature distribution on the outer surface of the casing for different heat dissipation rib solutions in the present invention;

[0032] Figure 5 This is a schematic diagram of optimized structural parameters of the casing in the present invention;

[0033] Figure 6 is the sensitivity calculation result between the optimization parameters and the target in the present invention;

[0034] Figure 7 This is a comparison chart of the prediction results of the SVR proxy model in the present invention;

[0035] Figure 8 This is the result of the casing structure optimization in the present invention;

[0036] Figure 9 It is the basic framework of the multi-physics field coupling simulation model in the present invention;

[0037] Figure 10 This is a comparison chart of electromagnetic noise simulation results before and after optimization in the present invention;

[0038] Figure 11 are the initial values ​​and value ranges of the structural parameters in the present invention;

[0039] Figure 12 It is an evaluation index of the goodness of fit of the surrogate model in the present invention;

[0040] Figure 13 This is a comparison of the results before and after the optimization of the casing structure of this Daming machine. DETAILED DESCRIPTION

[0041] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the present invention will be further explained below with reference to the accompanying drawings in the embodiments of the present invention.

[0042] See also Figures 1 to 13 The present invention provides a casing structure optimization design method for suppressing electromagnetic noise of a forced air-cooled induction motor, comprising the following steps:

[0043] Step 1: Create a physical model of the chassis with different heat dissipation rib structures, perform modal and heat dissipation simulations in finite element simulation software, and select a structure that helps to strengthen structural rigidity while ensuring heat dissipation performance and production process. Step 1 specifically includes:

[0044] 1.1 Establish a variety of physical model schemes with different heat dissipation rib structures; Based on the basic structural characteristics of the original casing, establish a casing physical model with different heat dissipation rib structures; Take a vertical induction motor for a centrifugal pump as an example, establish Figure 2 The nine heat dissipation rib structure schemes shown are used for comparison. Among them, the housing of Scheme I includes a set of equidistantly arranged annular heat dissipation ribs. Scheme II is the original housing, including four groups of 12 axial heat dissipation ribs. The housing of Scheme III tilts the heat dissipation ribs at a certain angle. Too large an angle will affect the air flow in some air ducts. In this example, the tilt angle is set to 15°. The housing of Scheme IV uses wavy heat dissipation ribs, the number of which is the same as that of Scheme I. The housing of Scheme V uses spiral heat dissipation ribs. In this example, the spiral angle is set to 15°. Schemes VI-IX are a combination of annular heat dissipation ribs and other types of heat dissipation ribs. In order to control variables and study the effect of different heat dissipation rib structures on stiffness, the theoretical mass of each model is controlled to be close to each other.

[0045] 1.2 Perform modal simulation on the casing models of various schemes; use finite element simulation software to perform modal simulation calculation on the physical models of each casing, and solve the first six free modal natural frequencies of each scheme; because the quality of each model has been controlled to be basically consistent, the calculated natural frequencies can better reflect the overall radial stiffness of the casing. The results of this example are as follows Figure 3 ;

[0046] 1.3 Select the heat dissipation rib structure that best improves the casing stiffness. Step 1.3 involves comprehensively analyzing the modal calculation results and casing processing technology of various heat dissipation rib schemes to select a structural scheme that is easy to manufacture, has a suitable appearance, and has a high low-order natural frequency. Compared to the original casing scheme, the low-order natural frequencies of the other schemes were all improved. Among them, schemes VI-IX, which added circumferential ribs, had the most significant optimization effect and were selected as alternative schemes.

[0047] 1.4 Select the heat dissipation rib structure scheme with the least impact on heat dissipation; establish an equivalent simplified finite element model of the casing of each alternative scheme, and perform fluid heat dissipation simulation on the outer surface of the casing. To simplify the simulation process, only the stator core and the casing are considered in the model, and the ideal working state is set. The heat loss is estimated based on the efficiency of the motor, and the stator core is set as the only heat source. Only the heat exchange between the casing surface and the air is considered. After solving the results, the heat dissipation performance of each scheme can be quickly compared and analyzed to select the scheme that is most conducive to heat dissipation or has the least impact on heat dissipation. In this test, the heating power of the stator core is set to 178482W / m3, and the axial temperature distribution comparison of the outer surface of the casing of each scheme is solved. Figure 4 ,In addition, because the casing in the test is an aluminum alloy die-casting, ,the layout of the heat dissipation ribs, the direction of demoulding, the motor heat dissipation method and ,the simulation results are comprehensively considered, and Scheme VIII is selected as the final ,improvement scheme.

[0048] Step 2: Based on support vector regression (SVR) and the second-generation non-dominated sorting genetic algorithm (NAGA-Ⅱ), the casing structure of the improved scheme is optimized with multiple parameters and multiple objectives. Step 2 specifically includes:

[0049] 2.1 Establish an optimization mathematical model for the casing structure; determine the structural parameter variables and optimization targets that need to be optimized. The selection of the structural parameter variation range needs to take into account the production and processing technology, and the specific values ​​can be selected with the help of engineering experience; the optimization target can be determined as the maximum first-order natural frequency and the minimum mass change, because when the mass change is not large, the first-order natural frequency can better reflect the change in radial stiffness, and controlling the mass is also conducive to the lightweight design of the motor. In this test method, a total of q heat dissipation ribs are selected. w , number of circumferential reinforcement q r , housing length l, housing thickness t, heat dissipation rib height h w , heat dissipation rib thickness tw 、Slope of heat dissipation rib i w , circumferential reinforcement height h r , hoop reinforcement thickness t r , hoop reinforcement inclination i r Ten structural parameters, such as Figure 5 The initial values ​​and value ranges of various structural parameters are summarized as follows: Figure 11 . With the first-order natural frequency f n＝1 The optimization goal is to maximize and minimize the mass change Δm. The optimization model of the casing can be described as:

[0050]

[0051] stp i，min ≤p i ≤p i,max , p i =[q w ,q r ,l,t,h w ,i w ,h r ,t r ,i r ]

[0052] 2.2 Establish a parametric physical model; Based on the determined structural parameters, perform parametric modeling on the physical model of the casing to facilitate subsequent variable parameter simulation calculations in finite element simulation software;

[0053] 2.3 Calculate the overall sensitivity change within the range of each parameter variable; uniformly select a points within the value interval of each parameter and perform a modal solution on the housing model in the finite element software, while keeping other parameters unchanged, to obtain the response relationship between the natural frequency and mass as each parameter changes. The sensitivity value is represented by the coefficient obtained by least squares fitting the response results. Here, in order to avoid the influence of the differences in the dimensions and value ranges of each parameter, the basic definition of sensitivity is multiplied by the value interval of each parameter. The definition of sensitivity is described as:

[0054]

[0055] Among them, s ωn 、s m are the nth order natural frequency and mass sensitivity, c ωn 、c m is the coefficient of least square fitting of the response results, that is, the approximate slope of the response curve, Δp i is the interval between the values ​​of the i-th parameter. In this test method, a=6, and the sensitivity between each parameter variable and the low-order natural frequency and mass is as follows: Figure 6By comparing the sensitivity, the hoop reinforcement slope i with less influence is eliminated. r , optimize the remaining 9 structural parameters;

[0056] 2.4 Establish a parameter sample set and perform modal simulation calculations; select the structural parameters that have a significant impact on the optimization target, set the variation range of each structural parameter in the finite element simulation software, use the Latin hypercube sampling method NAGA-Ⅱ to select b groups of parameter variable samples, and perform modal simulation calculations on the casing structure under each parameter combination. Among them, the value of b is adjusted according to the computational cost. In theory, the larger b is, the higher the accuracy of the subsequent proxy model. In this test method, b = 400;

[0057] 2.5 Establish an optimized proxy model based on SVR; randomly divide the calculated group b samples into x training sets and y validation sets, use the training set to train the SVR model, use the grid search method to adjust the model parameters, and then use the validation set to test the model after building the corresponding SVR proxy model to determine the coefficient R 2 , root mean square proportional error RMSPE, mean absolute percentage error MAPE measure the goodness of fit of the model. In this test method, take x = 360, y = 40, the prediction results of the validation set are as follows Figure 7 , evaluation indicators such as Figure 12 , R in this test method 2 greater than 0.99, and RMSPE and MAPE are less than 1%, indicating goodness of fit;

[0058] 2.6 Use NAGA-Ⅱ to optimize parameters; Based on the established SVR agent model, optimize the parameters of the two optimization objectives. After several iterations, search the Pareto frontier of the two optimization objectives in the variable space and select the optimal solution. In this example, the optimization results are as follows: Figure 8 , considering that the lightweight goal and the improvement of structural rigidity are relatively contradictory, which shows a similar law in most structural optimization problems, for the optimization goal of the casing in this example, improving the radial rigidity of the casing takes priority over weight control. It is only necessary to ensure that the change in mass is as small as possible, so we choose Figure 8 The sample points marked by the black box are taken as the optimal solution. The first-order natural frequency and mass changes after optimization are shown as follows: Figure 13 , the first-order natural frequency increased by 84.2%, and the mass remained basically unchanged, proving that the radial stiffness of the casing has been significantly improved.

[0059] Step 3: Perform multi-physics coupling simulation verification after structural optimization; Step 3 specifically includes:

[0060] 3.1 Establish a multi-physics coupling simulation model of the motor; Based on the finite element simulation analysis software, establish the electromagnetic, structural, and acoustic multi-physics coupling simulation model of the original motor and the optimized motor. In this test method, the basic framework of the coupling simulation model is established;

[0061] 3.2 Calculate the electromagnetic noise results of the motor before and after optimization; In this test method, the A-weighted sound pressure level spectrum results of the casing at a position 0.5m away from the outer surface of the casing before and after optimization are as follows: Figure 10 ,The total sound pressure level dropped from 81dBA to 72dBA, indicating that the electromagnetic noise of the motor was significantly reduced after the ,case was optimized using the optimization method proposed in this ,invention.

[0062] The above is a detailed introduction to a casing structure optimization design method for suppressing electromagnetic noise of a forced air-cooled induction motor provided by an embodiment of the present invention. Specific examples are used in this article to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the technical solutions disclosed by the present invention. At the same time, for general technical users in this field, according to the ideas of the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as a limitation of the present invention.

Claims

1. A casing structure optimization design method for suppressing electromagnetic noise of a forced air-cooled induction motor, characterized by: The following steps are involved: Step 1: Create physical models of the chassis with different heat dissipation rib structures, perform modal and heat dissipation simulations in finite element simulation software, and select a structure that helps strengthen structural rigidity while ensuring heat dissipation performance and production process. Step 1 includes: 1.1 Create physical models of various heat dissipation rib structure schemes; 1.2 Perform modal simulations on the chassis models of various schemes; 1.3 Select the heat dissipation rib structure scheme that best improves chassis rigidity; 1.4 Select the heat dissipation rib structure scheme that minimizes heat dissipation. Step 2: Based on support vector regression (SVR) and the second-generation non-dominated sorting genetic algorithm (NAGA-II), the casing structure of the improved scheme is optimized for multiple parameters and multiple objectives. Step 2 includes: 2.1 Establishing an optimization mathematical model for the casing structure; 2.2 Establishing a parameterized physical model; 2.3 Calculating the overall sensitivity change within the value range of each parameter variable; 2.4 Establishing a parameter sample set and performing modal simulation calculations; 2.5 Establishing an optimization agent model based on SVR; 2.6 Using NAGA-II to perform parameter optimization. Step 3: Perform multi-physics coupling simulation verification after structural optimization; Step 3 includes: 3.1 establishing a multi-physics coupling simulation model of the motor; 3.2 calculating the electromagnetic noise results of the motor before and after optimization.

2. The casing structure optimization design method for suppressing electromagnetic noise of a forced air-cooled induction motor according to claim 1, characterized in that: The step 1.1 is as follows: based on the basic structural characteristics of the original casing, a casing physical model with different heat dissipation rib structures is established; nine heat dissipation rib structure schemes are established for a vertical induction motor used for a centrifugal pump for comparison, wherein the casing of Scheme I includes a group of equidistantly arranged annular heat dissipation ribs, Scheme II is the original casing, including four groups of 12 axial heat dissipation ribs, the casing of Scheme III tilts the heat dissipation ribs at a certain angle. A large tilt angle will affect the gas flow in some air ducts, and this scheme sets the tilt angle to 15°, the casing of Scheme IV uses wavy heat dissipation ribs, the number of which is the same as that of Scheme I, the casing of Scheme V uses spiral heat dissipation ribs, and this scheme sets the spiral angle to 15°, and Schemes VI-IX are combined structures of annular heat dissipation ribs and other types of heat dissipation ribs; in order to control variables, the influence of different heat dissipation rib structures on stiffness is studied, and the theoretical mass of each model is controlled to be close to each other.

3. The casing structure optimization design method for suppressing electromagnetic noise of a forced air-cooled induction motor according to claim 1, characterized in that: The step 1.2 is: using finite element simulation software to perform modal simulation calculations on the physical models of each casing to solve the first six free modal natural frequencies of each scheme; because the mass of each model has been controlled to be basically consistent, the calculated natural frequencies can better reflect the overall radial stiffness of the casing.

4. The casing structure optimization design method for suppressing electromagnetic noise of a forced air-cooled induction motor according to claim 1, characterized in that: Step 1.3 is: comprehensively analyze the modal calculation results and casing processing technology of each heat dissipation rib scheme, and select a structural scheme that is easy to produce and process, has a suitable appearance, and has a higher low-order natural frequency; compared with the original casing scheme, the low-order natural frequencies of various other schemes are all improved, among which schemes VI-IX with added circumferential ribs have a more obvious optimization effect and are selected as alternative schemes.

5. The casing structure optimization design method for suppressing electromagnetic noise of a forced air-cooled induction motor according to claim 1, characterized in that: The step 1.4 is: establishing an equivalent simplified finite element model of the casing of each alternative scheme, and performing fluid heat dissipation simulation on the outer surface of the casing. To simplify the simulation process, only the stator core and the casing are considered in the model, and an ideal working state is set. The heat loss is estimated based on the efficiency of the motor, and the stator core is set as the only heat source. Only the heat exchange between the casing surface and the air is considered. After the results are solved, the heat dissipation performance of each scheme can be quickly compared and analyzed to select the scheme that is most conducive to heat dissipation or has the least impact on heat dissipation. In this test, the heating power of the stator core is set to 178482W / m3, and the axial temperature distribution of the outer surface of the casing of each scheme is solved. In addition, because the casing in the test is an aluminum alloy die-casting, the layout of the heat dissipation ribs, the direction of demolding, the motor heat dissipation method and the simulation results are comprehensively considered to select the final improvement scheme.

6. The casing structure optimization design method for suppressing electromagnetic noise of a forced air-cooled induction motor according to claim 1, characterized in that: The step 2.1 is: determine the structural parameter variables and optimization targets that need to be optimized, wherein the selection of the structural parameter variation range needs to take into account the production and processing technology, and the specific values ​​can be selected with the help of engineering experience; the optimization target can be determined as the maximum first-order natural frequency and the minimum mass change, because when the mass change is not large, the first-order natural frequency can better reflect the change in radial stiffness, and controlling the mass is also conducive to the lightweight design of the motor; in this test method, a total of q heat dissipation ribs are selected w , number of circumferential reinforcement q r , housing length l, housing thickness t, heat dissipation rib height h w 、Thickness of heat dissipation rib t w 、Slope of heat dissipation rib i w , circumferential reinforcement height h r , hoop reinforcement thickness t r , hoop reinforcement inclination i r Ten structural parameters, and summarize the initial values ​​and value ranges of each structural parameter; take the first-order natural frequency f n＝1 The optimization goal is to maximize and minimize the mass change Δm. The optimization model of the casing can be described as: s.t.p i.min ≤p i ≤p i,max ,p i =[q w ,q r ,l,t,h w ,t w ,i w ,h r ,t r ,i r ]。 7. The casing structure optimization design method for suppressing electromagnetic noise of a forced air-cooled induction motor according to claim 1, characterized in that: The step 2.2 is: based on the determined structural parameters, parametric modeling is performed on the physical model of the casing to facilitate subsequent variable parameter simulation calculations in the finite element simulation software; the step 2.3 is: evenly select a points within the value range of each parameter, and perform modal solution on the casing model in the finite element software, while keeping other parameters unchanged, to obtain the response relationship between the natural frequency and mass as the parameters change; the sensitivity value is represented by the coefficient obtained by least squares fitting the response result. Here, in order to avoid the influence caused by the difference in the dimension and value range of each parameter, the basic definition of sensitivity is multiplied by the value interval of each parameter. The definition of sensitivity is described as: Among them, s ωn 、s m are the nth order natural frequency and mass sensitivity, c ωn 、c m is the coefficient of least square fitting of the response results, that is, the approximate slope of the response curve, Δp i is the value interval of the i-th parameter; in this test method, a=6 is taken, and the sensitivity value between each parameter variable and the low-order natural frequency and mass is selected. By comparing the sensitivity, the hoop reinforcement inclination i with less influence is eliminated. r , and optimize the remaining 9 structural parameters.

8. The casing structure optimization design method for suppressing electromagnetic noise of a forced air-cooled induction motor according to claim 1, characterized in that: Step 2.4 is as follows: selecting structural parameters that have a significant impact on the optimization target, setting the variation range of each structural parameter in the finite element simulation software, using the Latin hypercube sampling method NAGA-Ⅱ to select b groups of parameter variable samples, and performing modal simulation calculations on the casing structure under each parameter combination; wherein the value of b is adjusted according to the computational cost. In theory, the larger b is, the higher the accuracy of the subsequent proxy model establishment. In this test method, b is set to 400.

9. The casing structure optimization design method for suppressing electromagnetic noise of a forced air-cooled induction motor according to claim 1, characterized in that: The step 2.5 is as follows: randomly divide the calculated group b samples into x training sets and y validation sets, use the training sets to train the SVR model, use the grid search method to adjust the model parameters, build the corresponding SVR proxy model, and then use the validation set to test the model to determine the coefficient R 2 , root mean square proportional error RMSPE, mean absolute percentage error MAPE measure the goodness of fit of the model; in this test method, take x = 360, y = 40, and get the prediction results and evaluation indicators of the validation set. The R 2 The values ​​are greater than 0.99, and the RMSPE and MAPE are less than 1%, indicating a good fit.

10. The casing structure optimization design method for suppressing electromagnetic noise of a forced air-cooled induction motor according to claim 1, characterized in that: The step 2.6 is as follows: relying on the established SVR agent model, the parameters of the two optimization objectives are optimized, and after several iterations, the Pareto frontiers of the two optimization objectives are searched in the variable space, and the optimal solution is selected; the optimization results in this test are obtained. Considering that the lightweight goal and the improvement of structural rigidity are relatively contradictory, for the optimization goal of the casing, the priority of improving the radial stiffness of the casing is higher than the control of weight. It is only necessary to ensure that the change in mass is as small as possible. The sample as the optimal solution is selected, and the optimized first-order natural frequency and mass change are obtained. The first-order natural frequency is increased by 84.2%, and the mass remains basically unchanged, which proves that the radial stiffness of the casing has been significantly improved.

11. The casing structure optimization design method for suppressing electromagnetic noise of a forced air-cooled induction motor according to claim 1, characterized in that: Step 3.1 is: establishing electromagnetic, structural, and acoustic multi-physics field coupling simulation models of the original motor and the optimized motor based on finite element simulation analysis software; Step 3.2 is: in this test method, obtaining the A-weighted sound pressure level spectrum results of the casing at a position 0.5m away from the outer surface of the casing before and after optimization, and the total sound pressure level dropped from 81dBA to 72dBA.

Images