A method for designing a variable cross-section winding of an electric machine and an electric machine designed using the method

Abstract

The application discloses a motor variable cross-section winding design method and a motor designed by the method, and relates to the field of motor winding design and manufacturing. The method comprises the following steps: obtaining stator slot inner cavity contour parameters, winding layer number and target operation condition parameters, and setting geometric constraint conditions; establishing a motor electromagnetic finite element model, solving motor iron loss and stator slot inner magnetic field space-time variation characteristics; further establishing a stator slot inner conductor loss analytical calculation model, calculating the average loss of each conductor in each candidate variable cross-section winding scheme; further establishing a motor thermal network model comprising a slot winding refined thermal network sub-model, solving conductor temperature distribution and winding maximum temperature; finally, taking the minimization of winding average loss and winding maximum temperature as the target, iteratively optimizing, outputting each layer conductor thickness and width parameters, and forming a variable cross-section winding design scheme. The application considers the loss calculation precision and optimization efficiency under high-frequency conditions, and is beneficial to improving motor efficiency and operation reliability.

Description

Technical Field

[0001] This invention relates to a method for designing variable cross-section motor windings and a motor designed using this method, belonging to the field of motor winding design and manufacturing technology. Background Technology

[0002] With the rapid development of new energy vehicles, high-dynamic industrial servo systems and robot integrated joints, motors are evolving towards high frequency, high speed and high power density. The problem of winding heating caused by high frequency and high current density operating conditions is becoming increasingly prominent.

[0003] In traditional motor stator design, conductors with uniform cross-sectional dimensions, such as standard round wire or flat wire with uniform cross-sectional dimensions, are typically used for winding. However, due to significant differences in leakage flux characteristics and heat dissipation conditions at different depths within the stator slots, uniform cross-sectional windings are prone to forming localized hot spots in the center or at the slot opening. To improve winding losses and heat distribution, the industry has recently proposed a design concept of "variable cross-sectional windings," which involves arranging conductors with different cross-sectional dimensions within the stator slot space to achieve targeted optimization.

[0004] However, existing design methods suffer from the following serious drawbacks in practical engineering: First, they neglect high-frequency AC losses. Existing schemes often only consider DC copper losses, failing to account for the severe AC eddy current losses caused by air gap harmonic penetration and slot leakage flux in the slot conductors under high-frequency operating conditions, leading to distorted calculations of winding losses under high-frequency conditions. Second, the optimization algorithms are inefficient. Variable cross-section design is a high-dimensional nonlinear multi-objective optimization problem. Existing local search strategies are prone to getting trapped in local optima, while directly using the finite element method for parameter sweeping is computationally too costly and cannot complete large-scale iterative optimization within a finite period.

[0005] In summary, there is an urgent need for a multi-objective optimization design method for variable cross-section windings that comprehensively considers both high-frequency eddy current effects and complex slot heat conduction, and has extremely high optimization efficiency, in order to break through the temperature rise bottleneck of high power density motors. Summary of the Invention

[0006] This invention aims to solve the technical problems of existing variable cross-section winding design methods, such as incomplete description of physical mechanism under high-frequency operating conditions, inability to adapt to complex geometric slot shapes, and low optimization efficiency. It proposes a variable cross-section winding design method for motors and a motor designed using this method.

[0007] The technical solution of the present invention:

[0008] A method for designing variable cross-section windings for electric motors includes the following steps:

[0009] S1. Obtain the stator slot cavity contour parameters, winding layer number and target operating condition parameters of the motor to be designed, and set the design variables and constraints of the variable cross-section winding. The design variables include the thickness and width parameters of the conductors of each winding layer, and the constraints include insulation spacing constraints, manufacturing process constraints and slot geometric boundary constraints.

[0010] S2. Establish an electromagnetic finite element model of the motor, solve the motor iron loss and the spatiotemporal variation characteristics of the magnetic field in the stator slot under the target operating conditions, and extract the magnetic field discrete sampling point set corresponding to each winding conductor region in the slot. The magnetic field discrete sampling point set includes the spatial coordinates of the sampling point and the magnetic field vector data at different times within one electric cycle.

[0011] S3. Based on the magnetic field discrete sampling point set obtained in step S2, establish an analytical calculation model for the loss of the stator slot winding conductor considering the proximity effect, and calculate the average AC loss of each winding conductor in each candidate variable cross-section winding scheme under the target operating condition.

[0012] S4. Establish a motor thermal network model that includes a refined thermal network sub-model of the slot winding. Apply the motor iron loss obtained in step S2 as a heat source to the corresponding thermal network node. Apply the average AC loss of each winding conductor obtained in step S3 as an independent heat source to the corresponding winding conductor node of the refined thermal network sub-model of the slot winding. Solve to obtain the temperature distribution of each winding conductor and the highest temperature of the winding.

[0013] S5. With minimizing the average winding loss and minimizing the maximum winding temperature as the dual optimization objectives, the thickness and width parameters of the conductors in each layer of winding are iteratively optimized to determine the size parameters of the target variable cross-section winding while satisfying all constraints.

[0014] S6. Output the target variable cross-section winding design scheme, which includes the conductor thickness and conductor width of each layer of windings.

[0015] Specifically, in step S1,

[0016] The stator slot inner cavity contour parameters include slot opening size, slot bottom size, slot depth size, and slot width size at different slot depth positions;

[0017] The insulation spacing constraints include the minimum insulation spacing constraints between adjacent layer winding conductors and the minimum insulation spacing constraints between winding conductors and stator slot walls;

[0018] The manufacturing process constraints include the constraint on the rate of change of thickness of adjacent layer winding conductors and the upper and lower limits of the manufacturable dimensions of the winding conductors.

[0019] The geometric boundary constraint within the slot requires that all winding conductor layers, after being insulated, must be completely arranged within the allowable space of the stator slot cavity.

[0020] Specifically, in step S2, the electromagnetic finite element model is solved only once to obtain the spatiotemporal variation characteristics of the motor iron loss and the magnetic field in the stator slot; in step S3, for all candidate variable cross-section winding schemes, the electromagnetic finite element model is not solved again, and the average AC loss of each winding conductor is calculated only through the analytical calculation model of winding conductor loss.

[0021] Specifically, step S3 includes:

[0022] S31. The discrete magnetic field sampling point set is reconstructed into a continuous magnetic field distribution function in the stator slot by an interpolation algorithm;

[0023] S32. Based on the reconstructed continuous magnetic field distribution function, the spatiotemporal distribution of magnetic vector potential in each winding conductor region is obtained by path integration;

[0024] S33. The distribution of induced current density in the winding conductor is obtained according to Faraday's law of electromagnetic induction and Ohm's law, and the spatiotemporal distribution of total current density is obtained by combining the source current constraint.

[0025] S34. Based on the total current density distribution and Joule's law, calculate the average loss of each winding conductor in one electrical cycle.

[0026] Specifically, in step S4, the refined thermal network sub-model of the slot winding treats each winding conductor in the slot as an independent heat source node and temperature node, and adopts an equivalent thermal resistance model for the insulation layer and air gap; for different candidate schemes, the thermal resistance value corresponding to each node is updated in a parameterized manner to achieve rapid temperature solution.

[0027] Specifically, in step S5, the termination condition for iterative optimization is reaching the preset maximum number of iterations, convergence of the optimization target, or stability of the non-dominated solution set. In each iteration, the geometric boundary of the winding conductor is first updated according to the candidate size parameters, and then the loss calculation in step S3 and the temperature calculation in step S4 are executed in sequence. The results are fed back to the optimization algorithm to generate the candidate parameters for the next round.

[0028] An electric motor includes a stator core with stator slots. A variable cross-section winding is provided in the stator slot. The variable cross-section winding is designed using the above-mentioned variable cross-section winding design method. The variable cross-section winding is arranged in non-uniform cross-section layers in the stator slot along the slot depth direction. The thickness and width of the conductors of each layer of winding are determined by electromagnetic-thermal coupling multi-objective optimization.

[0029] The beneficial effects of this invention are:

[0030] 1. Improved accuracy of heat distribution calculation under high-frequency operating conditions. This invention establishes an analytical calculation model for stator slot conductor losses considering proximity effects to accurately calculate the AC losses of variable cross-section windings under complex alternating magnetic fields, solving the design failure problem of high-frequency operating conditions caused by existing methods that only consider DC losses.

[0031] 2. Improve the efficiency of winding temperature analysis and enhance the ability to characterize the temperature distribution within the slot. This invention employs a motor thermal network model that includes a refined thermal network sub-model of the slot winding. Each conductor within the slot is treated as an independent heat source node, and the average loss of the corresponding conductor is applied. This allows for the rapid acquisition of the temperature distribution of each conductor and the highest temperature of the winding, balancing thermal analysis efficiency with the accuracy of temperature distribution calculation.

[0032] 3. Possesses good geometric topological universality. This invention defines the winding width as a variable that dynamically changes with the thickness and slot shape equation, which can adapt to complex slot structures such as trapezoidal slots and pear-shaped slots. It optimizes the heat conduction path while ensuring the slot fill factor, and realizes winding arrangement optimization under complex geometric constraints.

[0033] 4. Improve the efficiency of loss assessment during the optimization process. This invention uses an electromagnetic finite element model to extract the spatiotemporal variation characteristics of the magnetic field in the stator slot at the front end. In the iterative optimization process of candidate variable cross-section winding schemes, instead of repeatedly solving the finite element model for each candidate scheme, an analytical calculation model of conductor loss in the stator slot is used to quickly calculate the average loss of each conductor, thereby reducing the cost of a single assessment and improving the efficiency of optimization design. Attached Figure Description

[0034] Figure 1 This is a flowchart of the method of the present invention;

[0035] Figure 2 This is a schematic diagram of the cross-sectional distribution structure of the variable cross-section windings within the stator slots.

[0036] Among them, 1-stator slot; 2-winding conductor; 3-stator core. Detailed Implementation

[0037] The above content is merely a brief summary of the technical solution of this invention. To more thoroughly understand the technical means employed in this invention, it can be implemented according to the detailed description in the specification. Furthermore, to make the above and other objects, features, and advantages of this invention easier to understand, specific embodiments of this invention will be described below.

[0038] Example 1:

[0039] like Figure 1 As shown, this embodiment is a method for designing variable cross-section windings for a motor, including the following steps:

[0040] Step S1: Obtain the inner cavity contour parameters of the stator slot 1 of the motor to be designed, the number of winding layers, and the target operating condition parameters. Set the design variables and constraints of the variable cross-section winding. The design variables include at least the thickness parameters and width parameters of each winding conductor 2. The constraints include at least the insulation spacing constraints, manufacturing process constraints, and slot geometric boundary constraints.

[0041] First, obtain the internal contour parameters of stator slot 1, the number of winding layers, and the target operating condition parameters of the motor to be designed. The internal contour parameters of stator slot 1 include at least the slot opening size, slot bottom size, slot depth, and slot width at different slot depths; the target operating condition parameters include at least the operating frequency and current amplitude.

[0042] In this embodiment, the stator slot 1 winding is divided into multiple layers of winding conductors 2 along the slot depth direction, and the thickness and width parameters of each layer of winding conductors 2 are used as design variables. For the i-th layer of winding conductors 2, its thickness can be denoted as hi and its width as wi. This forms the set of dimensional parameters for the variable cross-section winding:

[0043] H={h1, h2,…,hn}, W={w1,w2,…,wn}

[0044] Where n is the number of 2 layers of winding conductor.

[0045] Simultaneously, constraints are defined. These constraints include at least the following:

[0046] (1) Insulation spacing constraints: The insulation spacing between adjacent layer winding conductors 2 shall not be less than the preset minimum insulation spacing; the insulation spacing between winding conductor 2 and stator slot 1 wall shall not be less than the preset minimum insulation spacing.

[0047] (2) Manufacturing process constraints: The thickness variation rate of adjacent layer winding conductor 2 does not exceed the preset threshold; the thickness and width of winding conductor 2 are within the upper and lower limits of the manufacturable size.

[0048] (3) Slot space constraints: All winding conductors 2, after the insulation layer is added, should be arranged within the allowable space inside the stator slot 1 and should not cross the slot wall outline boundary.

[0049] In subsequent iterative optimization processes, the corresponding solution is considered a valid candidate solution only when the set of candidate size parameters simultaneously satisfies the above constraints.

[0050] Step S2: Establish the electromagnetic finite element model of the motor, solve for the iron loss of the motor and the spatiotemporal variation characteristics of the magnetic field in the stator slot 1, and extract the discrete sampling point set of the magnetic field corresponding to each winding conductor 2 region in the slot.

[0051] After obtaining the motor's geometric parameters and operating condition parameters, an electromagnetic finite element model of the motor is established. This electromagnetic finite element model is used to solve for the motor's iron loss and the spatiotemporal variation characteristics of the magnetic field within stator slot 1 under the target operating condition. Specifically, the electromagnetic state of the motor over one or more electrical cycles under the target operating condition can be solved using finite element methods to obtain the magnetic flux density distribution data within stator slot 1 at each moment.

[0052] To facilitate subsequent analytical calculations of winding conductor 2 losses, a discrete magnetic field sampling point set S is extracted from the region where each winding conductor 2 is located within the slot. For a given candidate winding scheme, magnetic field vector data at multiple locations are recorded within each winding conductor 2 region, thus forming the discrete magnetic field sampling point set corresponding to that region of winding conductor 2. The sampling point set S includes at least the spatial coordinates of the sampling points and magnetic field vector data at different times within one electrical cycle. This discrete magnetic field data will serve as the input for the subsequent analytical calculation model of winding conductor 2 losses.

[0053] In this embodiment, the electromagnetic finite element model is mainly used to extract the spatiotemporal characteristics of the magnetic field and obtain the iron loss of the motor, rather than to repeatedly calculate the loss of all winding conductors 2 under each candidate scheme in each round of optimization.

[0054] Step S3: Based on the magnetic field discrete sampling point set obtained in step S2, establish an analytical calculation model for the loss of winding conductor 2 in stator slot 1 considering the proximity effect, and calculate the average loss of each winding conductor 2 in each candidate variable cross-section winding scheme under the target operating condition.

[0055] First, a model space coordinate system is established to describe the magnetic field distribution and current density distribution within the winding conductor 2 region. Preferably, the intersection of the slot centerline and the slot bottom boundary is taken as the origin, the direction along the slot depth and pointing towards the slot opening is taken as the positive y-axis, and the direction perpendicular to the y-axis and pointing to the right is taken as the positive x-axis.

[0056] The discrete magnetic field sampling point set S obtained in step S2 can be represented as follows:

[0057]

[0058] Where xi is the x-axis coordinate of the i-th sampling point, yi is the y-axis coordinate of the i-th sampling point, and B xi (t) represents the magnetic field component along the x-axis at the i-th sampling point at time t, B yi (t) represents the y-axis component of the magnetic field at the i-th sampling point at time t, and N represents the number of sampling points.

[0059] Since the finite element model outputs magnetic field data at discrete sampling points, while analytical calculations require a continuous magnetic field function, the discrete magnetic field sampling point set is first reconstructed into a continuous magnetic field distribution function using an interpolation algorithm. , :

[0060]

[0061] in, This is an interpolation operator that translates a discrete point set into a continuous computational domain. The interpolation method can be bilinear interpolation, spline interpolation, or other interpolation methods suitable for reconstructing discrete magnetic fields. Through this step, the magnetic field value at any location within the cross-section of winding conductor 2 can be obtained while maintaining the non-uniform spatial characteristics of the magnetic field obtained from finite element calculations.

[0062] After obtaining the continuous magnetic field distribution function, the magnetic vector potential distribution within the winding conductor 2 region is obtained through path integration. Preferably, the origin in the local coordinate system is selected as the potential zero point, and the magnetic field is integrated along an orthogonal path to obtain the temporal and spatial distribution A(x,y,t) of the magnetic vector potential within each layer of winding conductor 2 region:

[0063]

[0064] The purpose of this step is to map the magnetic field distribution to the magnetic vector potential distribution required for subsequent current density calculation, thereby facilitating the determination of the induced current density inside winding conductor 2 according to Faraday's law of electromagnetic induction.

[0065] After the magnetic vector potential distribution is determined, the current density distribution in winding conductor 2 is obtained according to Faraday's law of electromagnetic induction and Ohm's law. The time-space distribution J(x,y,t) of the current density in winding conductor 2 can be expressed as:

[0066]

[0067] in, Let i be the conductivity of winding conductor 2. s (t) represents the source current component of winding conductor 2, which is determined by the operating parameters. yes The surface average value on the cross-section of winding conductor 2, i.e.:

[0068]

[0069] Among them, S c Let be the cross-sectional area of ​​winding conductor 2.

[0070] Combining the differential form of Joule's law, the average loss P of conductor 2 in the i-th layer of windings avg,i The calculation formula is as follows:

[0071]

[0072] Among them, l c Let T be the axial length of winding conductor 2, and T be the electrical period.

[0073] By performing the above calculations on all winding conductors 2, the average loss distribution of each winding conductor 2 under the current candidate variable cross-section winding scheme can be obtained.

[0074] In this embodiment, as the thickness and width parameters of each layer of winding conductor 2 change, the geometric boundary of each winding conductor 2 region also changes. Therefore, in each round of optimization, the cross-sectional range of the winding conductor 2 under the corresponding candidate scheme will be re-determined, and the average loss of each winding conductor 2 will be calculated based on this.

[0075] Step S4: Establish a motor thermal network model. Use the motor iron loss obtained from solving the motor electromagnetic finite element model in step S2 as a heat source and apply it to the corresponding thermal network node. Use the average loss of each winding conductor 2 obtained in step S3 as a heat source and apply it to the corresponding thermal network node. Solve the motor thermal network model to obtain the temperature distribution of each winding conductor 2 and the highest temperature of the winding.

[0076] The motor thermal network model in this embodiment includes at least three nodes: environmental nodes, housing nodes, stator core nodes, slot winding nodes, end winding nodes, and other nodes related to motor heat conduction. Based on the principle of thermal balance, thermal balance equations between each node can be written to form an overall thermal network model. The motor iron loss obtained by finite element analysis in step S2 is used as a heat source and applied to the corresponding thermal network nodes.

[0077] To more accurately reflect the temperature gradient caused by the uneven distribution of heat sources within the slot, the motor thermal network model in this embodiment includes a refined thermal network sub-model for the slot windings. In this sub-model, each winding conductor 2 within the slot is treated as an independent heat source node, and the average loss of each winding conductor 2 obtained from the analytical calculation in step S3 is applied to the corresponding heat source node of the winding conductor 2. This allows for the direct calculation of the temperature value of each winding conductor 2, rather than just the average temperature of the entire winding within the slot.

[0078] When establishing a refined thermal network sub-model for the slot winding, the winding conductors 2, interlayer insulation, and air gaps can be simplified. Preferably, the winding conductor 2 portion adopts a thermal resistance network model that simultaneously considers radial, circumferential, and axial heat transfer, while the insulation and air gap portions adopt corresponding equivalent thermal resistance models. For different winding conductor 2 thicknesses, widths, and insulation layer thicknesses, the corresponding thermal resistance values ​​can be updated parametrically, thereby enabling rapid temperature analysis of different candidate variable cross-section winding schemes.

[0079] After applying the motor's iron loss and the average loss of each winding conductor 2, the motor's thermal network model is solved to obtain the temperature distribution of each winding conductor 2 and the highest winding temperature. The highest winding temperature is preferably defined as the temperature value of the winding conductor 2 with the highest temperature among all winding conductor 2 nodes within the slot. Therefore, both the loss distribution and temperature distribution can be used simultaneously in the subsequent optimization process.

[0080] Step S5: Using the average winding loss and maximum winding temperature obtained in steps S3 and S4 as optimization targets, iteratively optimize the design variables to determine the target variable cross-section winding size parameters under the conditions of satisfying the insulation spacing constraints, manufacturing process constraints and stator slot 1 geometric boundary constraints.

[0081] Preferably, the optimization problem is solved by minimizing the average winding loss and minimizing the maximum winding temperature. During the optimization process, the set of candidate size parameters must simultaneously satisfy the following conditions:

[0082] (1) The insulation spacing constraint must be met;

[0083] (2) Meet manufacturing process constraints;

[0084] (3) Arrange the conductors 2 of each layer of windings within the allowable space of the stator slot 1.

[0085] Specifically, in each iteration, the geometric boundary of winding conductor 2 is updated based on the thickness and width parameters of the current candidate winding conductor 2. Then, the average loss of each winding conductor 2 is calculated according to step S3, and the temperature distribution and maximum temperature of each winding conductor 2 are calculated according to step S4. The obtained optimization target value is fed back to the optimization algorithm, which generates the next set of candidate size parameters. This iteration is repeated until a preset termination condition is reached, such as the maximum number of iterations, the target convergence condition, or the stability condition of the non-dominated solution set.

[0086] Step S6: After optimization, output the target variable cross-section winding design scheme. The design scheme includes at least the thickness of each layer of winding conductor 2 and the width of each layer of winding conductor 2.

[0087] Example 2:

[0088] This embodiment provides a motor designed using the method described in Embodiment 1. For example... Figure 2 As shown, the motor includes a stator core 3, a stator slot 1 is provided on the stator core 3, and a variable cross-section winding is installed inside the stator slot 1. The winding is composed of multiple layers of winding conductors 2 arranged in layers along the depth direction of the stator slot 1. The winding conductors 2 are arranged in layers with non-uniform cross-sections in the stator slot 1.

[0089] The thickness and width dimensions of the conductors 2 in each layer of the winding vary. These dimensions are not predetermined but are optimized and determined using the variable cross-section winding design method described in Example 1. First, the spatiotemporal variation characteristics of the magnetic field within the stator slot 1 are obtained using an electromagnetic finite element model. Second, the average loss of each layer of winding conductors 2 is calculated using a conductor loss analytical calculation model. Then, the temperature distribution of each conductor is calculated using a thermal network model. Finally, the thickness and width parameters of each layer of winding conductors 2 are determined through an optimization process. The resulting variable cross-section winding system achieves superior overall performance in terms of loss and temperature rise while meeting insulation and manufacturing requirements.

Claims

1. A method for designing variable cross-section windings for an electric motor, characterized in that, Includes the following steps: S1. Obtain the inner cavity contour parameters, winding layers and target operating condition parameters of the stator slot (1) of the motor to be designed, and set the design variables and constraints of the variable cross section winding. The design variables include the thickness parameters and width parameters of each layer of winding conductor (2). The constraints include insulation spacing constraints, manufacturing process constraints and slot geometric boundary constraints. S2. Establish an electromagnetic finite element model of the motor, solve the motor iron loss and the spatiotemporal variation characteristics of the magnetic field in the stator slot (1) under the target operating conditions, and extract the magnetic field discrete sampling point set corresponding to the region of each winding conductor (2) in the slot. The magnetic field discrete sampling point set includes the spatial coordinates of the sampling point and the magnetic field vector data at different times within one electric cycle. S3. Based on the magnetic field discrete sampling point set obtained in step S2, establish an analytical calculation model for the loss of the winding conductor (2) in the stator slot (1) considering the proximity effect, and calculate the average AC loss of each winding conductor (2) in each candidate variable cross section winding scheme under the target operating condition. S4. Establish a motor thermal network model including a refined thermal network sub-model of slot windings. Apply the motor iron loss obtained in step S2 as a heat source to the corresponding thermal network node. Apply the average AC loss of each winding conductor (2) obtained in step S3 as an independent heat source to the corresponding winding conductor (2) node of the refined thermal network sub-model of slot windings. Solve to obtain the temperature distribution of each winding conductor (2) and the highest temperature of the winding. S5. With minimizing the average winding loss and minimizing the maximum winding temperature as the dual optimization objectives, the thickness and width parameters of the conductors (2) of each layer of winding are iteratively optimized, and the size parameters of the target variable cross-section winding are determined under the condition of satisfying all constraints. S6. Output the target variable cross-section winding design scheme, which includes the thickness of each layer of winding conductor (2) and the width of each layer of winding conductor (2).

2. The method for designing variable cross-section windings for a motor according to claim 1, characterized in that, In step S1, The inner cavity contour parameters of the stator slot (1) include the slot opening size, slot bottom size, slot depth size, and slot width size at different slot depth positions; The insulation spacing constraints include the minimum insulation spacing constraints between adjacent layer winding conductors (2) and the minimum insulation spacing constraints between the winding conductors (2) and the stator slot (1) wall; The manufacturing process constraints include the thickness variation rate constraint of adjacent layer winding conductors (2) and the upper and lower limits of the manufacturable dimensions of the winding conductors (2); The geometric boundary constraint within the slot is that all winding conductors (2) layers, after being insulated, must be completely arranged within the allowable space of the stator slot (1) cavity.

3. The method for designing variable cross-section windings for motors according to claim 1, characterized in that, In step S2, the electromagnetic finite element model is solved only once to obtain the spatiotemporal variation characteristics of the motor iron loss and the magnetic field in the stator slot (1); in step S3, for all candidate variable cross-section winding schemes, the electromagnetic finite element model is not solved again, and the average AC loss of each winding conductor (2) is calculated only through the analytical calculation model of the winding conductor (2) loss.

4. The method for designing variable cross-section windings for a motor according to claim 1, characterized in that, Step S3 specifically includes: S31. The discrete magnetic field sampling point set is reconstructed into a continuous magnetic field distribution function in the stator slot (1) by interpolation algorithm; S32. Based on the reconstructed continuous magnetic field distribution function, the spatiotemporal distribution of magnetic vector potential in the region of each winding conductor (2) is obtained by path integration; S33. The induced current density distribution in the winding conductor (2) is obtained according to Faraday's law of electromagnetic induction and Ohm's law, and the spatiotemporal distribution of the total current density is obtained by combining the source current constraint. S34. Based on the total current density distribution and Joule's law, the average loss of each winding conductor (2) in one electrical cycle is calculated.

5. The method for designing variable cross-section windings for a motor according to claim 1, characterized in that, In step S4, the refined thermal network sub-model of the slot winding treats each winding conductor (2) in the slot as an independent heat source node and temperature node, and adopts an equivalent thermal resistance model for the insulation layer and air gap. For different candidate schemes, the thermal resistance value corresponding to each node is updated by parameterization to achieve rapid temperature solution.

6. The method for designing variable cross-section windings for a motor according to claim 1, characterized in that, In step S5, the termination condition for iterative optimization is to reach the preset maximum number of iterations, converge the optimization target, or stabilize the non-dominated solution set. In each iteration, the geometric boundary of the winding conductor (2) is updated according to the candidate size parameters, and then the loss calculation in step S3 and the temperature calculation in step S4 are executed in sequence. The results are fed back to the optimization algorithm to generate the candidate parameters for the next round.

7. An electric motor, characterized in that, The motor includes a stator core (3), and stator slots (1) are opened on the stator core (3). A variable cross-section winding is provided in the stator slot (1). The variable cross-section winding is designed using the motor variable cross-section winding design method according to any one of claims 1 to 6. The variable cross-section winding is arranged in non-uniform cross-section layers in the stator slot (1) along the slot depth direction. The thickness and width of the conductors (2) of each layer of winding are determined by electromagnetic-thermal coupling multi-objective optimization.

Images