Design method of motor flat copper wire stator winding

Abstract

This invention relates to the field of motor design and manufacturing technology, specifically a design method for flat copper wire stator windings of motors. The method includes: constructing a basic data source layer to integrate stator slot space data, electromagnetic boundary data, and conductor physical parameters; constructing a topology generation layer to plan the winding spatial distribution and end connection paths, wherein a transposition connection module plans the inter-layer connection relationship of conductors within the slot, and an end topology module plans the spatial layout of end welding points; constructing a dynamic adjustment layer to establish a three-dimensional spatial matrix based on the stator core end face, setting a stepped misalignment mechanism including stepped misalignment amount and design tolerance, and adjusting the axial height distribution of end welding points; and generating and outputting a three-dimensional design model or manufacturing parameters of the stator winding based on the adjustment results, achieving a unified approach to electromagnetic optimization, end space avoidance, and welding accessibility.

Description

Technical Field

[0001] This invention relates to the field of motor design and manufacturing technology, specifically to a design method for flat copper wire stator windings of motors. Background Technology

[0002] In the stator winding design of drive motors and high power density motors for new energy vehicles, flat copper wire windings have gradually become a common solution for stator structure design due to their high slot fill factor and good conductivity.

[0003] Currently, the design of flat copper wire stator windings generally supports conventional modeling based on the arrangement of conductors in the slot and the end connection relationship. If it is necessary to take into account the end welding layout, high-frequency electromagnetic distribution and manufacturing equipment adaptation under the high-layer winding, it is usually necessary to complete the design by performing electromagnetic design, structural design and process design separately to obtain the corresponding winding scheme.

[0004] However, when designing high-layer flat copper wire stator windings using the above method, data needs to be repeatedly converted between different design stages. This makes it difficult to unify and coordinate the interlayer connection relationship within the slot, the spatial path at the end, and the distribution of welding height. Furthermore, it is impossible to dynamically adjust the winding topology based on leakage flux distribution, conductor bending constraints, and welding process constraints, resulting in insufficient control of high-frequency AC losses, high risk of end interference, and poor welding accessibility. At the same time, this approach lacks effective overall planning for the spatial distribution of three-dimensional end weld points, and it is prone to frequent path backtracking and low design convergence efficiency as the number of layers increases. Moreover, it relies heavily on designers to perform a large amount of manual verification and local correction, making it difficult to adapt to the rapid development needs of high slot fill factor and multi-layer flat copper wire stator windings, resulting in low design stability and engineering adaptability. Summary of the Invention

[0005] The purpose of this invention is to provide a design method for flat copper wire stator windings of electric motors, solving the following technical problems:

[0006] This design avoids the physical interference problems caused by the concentration of end welding points on the same plane, such as local welding point overlap, bending area collision, or insufficient insulation distance. It also avoids the problems of branch induced potential imbalance and high-frequency AC loss caused by each parallel branch staying in different leakage magnetic intensity areas for a long time. Furthermore, it makes it easier to unify the slot electromagnetic optimization, end space avoidance, and welding accessibility at the manufacturing end into the same design process, realizing the transformation of the winding design scheme from two-dimensional connection logic to three-dimensional manufacturable topology, and effectively adapting to the three-dimensional dynamic zoom laser welding equipment on the production line.

[0007] The objective of this invention can be achieved through the following technical solutions:

[0008] The design method for electric motor flat copper wire stator windings includes:

[0009] A basic data source layer is constructed to integrate stator slot spatial data, electromagnetic boundary data, and conductor physical parameters;

[0010] A topology generation layer is constructed, which plans the winding spatial distribution and end connection path based on the integrated stator slot space data, electromagnetic boundary data and conductor physical parameters. The topology generation layer includes a transposition connection module and an end topology module. The transposition connection module plans the inter-layer connection relationship of conductors in the slot, and the end topology module plans the spatial layout of end welding points.

[0011] A dynamic adjustment layer is constructed, and a three-dimensional spatial matrix is ​​established with the stator core end face as the reference surface. A stepped misalignment mechanism based on the three-dimensional spatial matrix is ​​set. The stepped misalignment mechanism includes the preset stepped misalignment amount and design tolerance. The axial height distribution of the end welding points is adjusted according to the stepped misalignment mechanism.

[0012] Based on the adjusted axial height distribution of the end welding points and the planned winding space distribution, a three-dimensional design model or manufacturing parameters of the stator winding are generated and output.

[0013] Optionally, before inputting the stator slot space data, electromagnetic boundary data, and conductor physical parameters into the basic data source layer, the following steps are also included:

[0014] The stator slot spatial data, electromagnetic boundary data, and conductor physical parameters are preprocessed to obtain the preprocessed stator slot spatial data, electromagnetic boundary data, and conductor physical parameters. The preprocessing includes extracting leakage flux density feature values, extracting preset bending limit radii, and data dimensionality reduction mapping. Data dimensionality reduction mapping is to perform spatial coordinate system transformation on the physical stator dimensions in the stator slot spatial data to eliminate geometric redundancy.

[0015] Optionally, a transposition connection module is used to plan the interlayer connection relationship of conductors in the slot based on a preset asymmetric transposition connection logic combined with the electromagnetic field distribution characteristics in the electromagnetic boundary data.

[0016] The transposition connection module is specifically used for: initializing a preset connection matrix based on the number of layers and phases of the stator slot, refining the preset connection matrix to each stator slot through asymmetric transposition connection logic, extracting the electromagnetic high-frequency leakage magnetic field region and weak leakage magnetic field region of each stator slot from the electromagnetic boundary data, extracting the conductor dwell length from the conductor physical parameters, and planning the interlayer span of each stator slot in combination with the extracted data. The connection matrix includes a cross-layer connection matrix.

[0017] The interlayer connection relationship is forcibly defined by the cross-layer connection matrix, and the electromagnetic equivalent spiral transposition of the same phase winding on the stator circumference is quantified by the short pitch configuration with a span smaller than the pole pitch and the long pitch configuration with a span larger than the pole pitch.

[0018] Spatial location data is processed by electromagnetic field distribution characteristics, and the results of conductor current sharing planning in different regions are refined based on the relationship between the leakage flux density and interlayer span in the electromagnetic high-frequency leakage magnetic field region and the weak leakage magnetic field region.

[0019] Optionally, an end topology module is used to divide each end of the stator winding into odd-numbered slot end welding points and even-numbered slot end welding points, and decompose the spatial distribution of the end connection path to obtain the three-dimensional coordinates of each end welding point.

[0020] End-topology modules are specifically used for:

[0021] Based on the winding layer number, stator circumferential dimensions, end straight section extension length, and anti-interference principles derived from stator slot space data and conductor physical parameters, the spatial layout of the end welding points is evaluated. The evaluation process includes:

[0022] The global welding height is predicted using a pre-defined stepped end topology model. The axial offset and circumferential physical spacing of the end connection path are extracted. The axial offset and circumferential physical spacing are combined to evaluate the non-interference feasibility and final axial height of various stepped staggered layouts.

[0023] Optionally, the stepped misalignment mechanism includes a first height plane, a second height plane, and a transition connection plane. The first height plane is the axial height of the welding points at the ends of the odd-numbered slots, the second height plane is the axial height of the welding points at the ends of the even-numbered slots, and the transition connection plane is the wire bending area connecting the first height plane and the second height plane.

[0024] Optionally, the axial height distribution of the end welding points can be adjusted according to the stepped misalignment mechanism, including:

[0025] Select the target misalignment height according to the design requirements, and based on the target misalignment height, use the topology generation layer to re-optimize and arrange the end connection paths that meet the target misalignment height.

[0026] The system acquires preset welding process constraints and uses these constraints as dynamic inputs to dynamically adjust the height distribution. The welding process constraints include three-dimensional dynamic zoom laser welding parameters. When mechanical interference at the end is detected or the welding height exceeds the preset welding height threshold, the system switches to the corresponding step misalignment mechanism in real time. The system also uses the topology generation layer to replan the future end connection path and end welding point distribution through real-time calculation and simulation.

[0027] Based on the preset asymmetric transposition connection logic and the preset stepped end topology model data synchronization mechanism, the planning results are automatically updated.

[0028] Optionally, when the end weld point is at the first height plane, the axial height distribution of the end weld point is adjusted, including:

[0029] The axial height of the welding points at the ends of the odd-numbered slots is set to remain at the current level. The relationship between the weld penetration and the axial height is set according to the preset tolerance range. The coordinates of the end welding points are output based on the space data of the stator slots.

[0030] When the end weld point is at the second height plane, adjust the axial height distribution of the end weld point, including:

[0031] The plan for axial offset is set, and the welding points at the ends of even-numbered slots are staggered. The extreme value of circumferential spacing without physical interference is calculated based on the spacing between adjacent slots and the width of the conductor. The welding points at the ends of even-numbered slots are staggered according to the extreme value of circumferential spacing. By combining the extension length of the straight section at the end with the preset stepped end topology model, the height difference of the end welding points and the proportion of the end welding points in the overall end space are predicted. The preset fluctuation threshold is set. The preset fluctuation threshold is the maximum allowable relative change ratio of the height difference or the proportion of the space under different design layers.

[0032] Among them, in the prediction results, the rate of change of the height difference of the end welding point and the proportion of the end welding point in the overall end space with the increase of the number of layers is less than the preset fluctuation threshold; if the rate of change is greater than or equal to the preset fluctuation threshold, the extreme value of the circumferential spacing and the axial offset are recalculated and reassigned.

[0033] When the end weld point is on the transition connection plane, adjust the axial height distribution of the end weld point, including:

[0034] Assuming a spatial interference conflict occurs, a three-dimensional spatial matrix is ​​used to simulate the future avoidance path, assess the impact of the spatial interference conflict on the end insulation creepage distance and the overall welding height, and adjust the axial height distribution of the end welding points based on the assessment results. The spatial interference conflict includes physical overlap.

[0035] Optionally, the method further includes a closed-loop verification feedback step, which includes:

[0036] The three-dimensional point cloud data of the end wires is collected by a visual positioning device, and the actual spatial coordinates of each end welding point are extracted.

[0037] Verify whether the stair misalignment meets the design tolerance based on actual spatial coordinates;

[0038] A pulse width modulation waveform is injected using a high-frequency impedance testing device, and transient voltage and current data at the terminals are collected.

[0039] Fast Fourier Transform is used to process transient voltage and transient current data to extract the fundamental current component and higher harmonic current components.

[0040] The actual AC resistance is calculated based on the fundamental current component and the higher harmonic current component, and the actual AC resistance is fed back to the basic data source layer for parameter correction.

[0041] Optionally, the actual AC resistance can be fed back to the basic data source layer for parameter correction, including:

[0042] Set a first AC resistance threshold and a second AC resistance threshold, wherein the first AC resistance threshold is greater than the second AC resistance threshold;

[0043] If the actual AC resistance is greater than the first AC resistance threshold, it is determined that there is a local circulating current in the interlayer connection relationship of the conductors in the slot, and the transposition connection module is triggered to re-plan the interlayer span.

[0044] If the actual AC resistance is less than the second AC resistance threshold, it is determined that the interlayer connection relationship of the conductors in the slot has reached electromagnetic equilibrium, and the final winding design scheme is output.

[0045] If the actual AC resistance is greater than or equal to the second AC resistance threshold and less than or equal to the first AC resistance threshold, the design is determined to be in a critical state, and the sampling frequency of the high-frequency impedance test equipment is increased to reacquire transient voltage and transient current data.

[0046] The beneficial effects of this invention are:

[0047] 1. This invention integrates spatial, electromagnetic and physical parameters by constructing a basic data source layer, and combines topology generation and dynamic adjustment mechanisms to achieve unified coordination of slot connection, end path and welding height, avoiding frequent path backtracking caused by multi-stage data conversion, and improving the convergence efficiency of high-layer winding design.

[0048] 2. This invention combines asymmetric transposition connection logic with electromagnetic field distribution characteristics. By defining a cross-layer connection matrix and quantifying the long and short pitch configuration, it dynamically plans the conductor current sharing in the high and low leakage magnetic regions, effectively suppressing local circulating current between parallel branches and solving the problem of insufficient AC loss control under high-frequency operating conditions.

[0049] 3. This invention divides the end welding points into odd and even slots and assigns them to planes of different heights. Combined with process constraints such as dynamic zoom laser welding parameters, it performs interference-free evaluation and dynamic adjustment of height distribution, thereby mitigating the interference risk caused by end congestion in high-layer windings and realizing effective overall planning of three-dimensional end welding points.

[0050] 4. This invention introduces a closed-loop feedback step that includes visual positioning geometric verification and high-frequency impedance testing. It uses a preset AC resistance dual threshold mechanism to automatically determine and correct the actual electromagnetic performance, eliminating the reliance on a large amount of manual verification and improving the stability and production line adaptability of high slot full rate stator winding design. Attached Figure Description

[0051] The invention will now be further described with reference to the accompanying drawings.

[0052] Figure 1 This is a flowchart illustrating the design method of the motor flat copper wire stator winding provided in the embodiments of this application. Detailed Implementation

[0053] 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, and 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.

[0054] Please see Figure 1 The design method for flat copper wire stator windings of motors includes: constructing a basic data source layer and integrating stator slot space data, electromagnetic boundary data and conductor physical parameters;

[0055] A topology generation layer is constructed, which plans the winding spatial distribution and end connection path based on the integrated stator slot space data, electromagnetic boundary data and conductor physical parameters. The topology generation layer includes a transposition connection module and an end topology module. The transposition connection module plans the inter-layer connection relationship of conductors in the slot, and the end topology module plans the spatial layout of end welding points.

[0056] A dynamic adjustment layer is constructed, and a three-dimensional spatial matrix is ​​established with the stator core end face as the reference surface. A stepped misalignment mechanism based on the three-dimensional spatial matrix is ​​set, which includes a preset stepped misalignment amount. and design tolerances And adjust the axial height distribution of the end welding points according to the stepped misalignment mechanism;

[0057] Based on the adjusted axial height distribution of the end welding points and the planned winding space distribution, a three-dimensional design model or manufacturing parameters of the stator winding are generated and output.

[0058] This embodiment provides a design mechanism for a flat copper wire stator winding for a motor. Specifically, the application scenario is set as follows: an 800V new energy vehicle drive motor project needs to design a 10-layer flat copper wire stator winding. The goal is to control the AC resistance increment under 1000Hz operating conditions within a predetermined range while maintaining a slot fill factor of not less than 60%, and to ensure that the absolute height of the end welding area does not exceed 3.5mm, so as to adapt to the three-dimensional dynamic zoom laser welding equipment on the production line.

[0059] Specifically, the entire design process can be divided into a basic data source layer, a topology generation layer, a dynamic adjustment layer, and a model output layer. The basic data source layer receives three types of input data: first, stator slot space data, such as slot width, slot depth, tooth shoulder size, core stack thickness, number of slots, number of pole pairs, and core end face coordinate system; second, electromagnetic boundary data, such as leakage flux density distribution, phase current amplitude, frequency, and phase near each slot under rated operating conditions and high-frequency operating conditions; and third, conductor physical parameters, such as the width, thickness, insulation layer thickness, minimum bending radius, and allowable weld penetration of flat copper wire. The basic data source layer maps the above data to the same coordinate system, for example, establishing a design coordinate system with the stator core center axis as the Z-axis, the circumferential direction as the θ-axis, and the radial direction as the R-axis, thereby providing a unified reference for subsequent spatial planning.

[0060] To avoid ambiguity caused by the same notation throughout the text, this embodiment adopts the following consistent convention: This only indicates the amount of staircase misalignment. The numbers 0.42T, 0.35T, etc., following the leakage flux density values ​​indicate only the design tolerance; the letter T represents the magnetic flux density unit Tesla and is not a design tolerance symbol.

[0061] Meanwhile, plane A and plane B used in this embodiment for ease of deduction are only exemplary pronouns for the first height plane and the second height plane, wherein plane A corresponds to the first height plane where the welding points at the ends of the odd-numbered slots are located, and plane B corresponds to the second height plane where the welding points at the ends of the even-numbered slots are located.

[0062] In the topology generation layer, the winding space distribution is planned first, and then the end connection path is planned. For ease of explanation, it is assumed that only 4 stator slots of a certain phase winding are selected as local analysis objects, which are denoted as slot 1, slot 2, slot 3 and slot 4 respectively. Each slot contains 10 layers of conductors, and the layer number is 1 to 10 from the side closest to the slot opening to the bottom of the slot.

[0063] The transposition connection module does not simply connect layers 1 to 2, 2 to 3 in sequence, but instead outputs an inter-layer connection matrix; this connection matrix can be understood as a 10×10 relational table, where the first layer connects to the second layer, the second layer connects to the third layer, and so on. Line number When the column value is 1, it indicates that the first column is the 1st column. The wires of the layer are connected to the first layer at the end of the slot. Layer, in which Corresponding to the current layer number of the outgoing wire, Corresponding target connection layer number;

[0064] For example, a local matrix can be set as follows: layer 1 connects to layer 3, layer 2 connects to layer 5, layer 3 connects to layer 4, and layer 8 connects to layer 10; after forming an asymmetric cross-layer relationship, the same current branch will alternately pass through the high leakage magnetic field region and the low leakage magnetic field region in different slots, thereby achieving the purpose of balanced dwell time;

[0065] The end topology module further expands each end welding point from a two-dimensional circular arrangement to a three-dimensional circular + axial arrangement; its processing can be understood as follows: first, the initial end point list is obtained according to the exit sequence of each wire in the slot, and then candidate connection paths are generated according to the wire width, insulation layer thickness and minimum bending radius.

[0066] For example, if slot 1 and slot 2 are adjacent, and the first layer of wire in slot 1 needs to be connected to the third layer in slot 2, the end topology module calculates an initial bending path from the exit coordinate P1 of slot 1 to the candidate coordinate P2 of the solder joint in slot 2.

[0067] If the minimum distance between this path and the candidate path of the second layer conductor of slot 1 is less than the insulation creepage distance requirement, its axial height is reallocated to form another candidate path P1-P3-P2, where P1 is the coordinate of the starting slot exit point, P2 is the coordinate of the target weld point candidate, and P3 is the transition bend point between the two; after iterative calculation, the spatial layout of the end weld points can be formed.

[0068] The dynamic adjustment layer establishes a three-dimensional spatial matrix with the stator core end face as the reference surface; the three-dimensional spatial matrix is ​​composed of an index table organized by slot number × layer number × height plane; for example, when encoding the above 4 slots, 10 layers per slot, and two welding height planes A and B, a 4×10×2 occupancy matrix can be obtained; if a cell in the matrix has a value of 1, it means that the corresponding slot layer has been occupied in that height plane;

[0069] If the value is 0, it indicates that the arrangement is possible; the system adjusts according to the preset step misalignment amount. and design tolerances The height of the welding points for the odd and even slots is allocated; for example, if the axial height of plane A is 1.2mm and the axial height of plane B is 3.2mm, then the corresponding stepped misalignment is... It is 2.0mm;

[0070] When design tolerance When the setting is ±0.15mm, the subsequent output of the solder joint target height can be judged as qualified if it falls into the corresponding range; through this stepped staggered mechanism, the solder joints that were originally crowded on the same plane can be split into two discrete planes, thereby alleviating the physical interference under high-layer structure;

[0071] In the model output layer, the system outputs a three-dimensional design model or manufacturing parameters of the stator winding based on the adjusted axial height distribution of the weld points and the planned winding spatial distribution; the three-dimensional design model may include the spatial polyline of each conductor segment, the arc transition radius, the weld point coordinates and the weld point orientation;

[0072] Manufacturing parameters may include the bending length of each wire, the groove angle, the axial offset, the laser welding focal plane switching sequence, and the station execution cycle time;

[0073] As a supplementary explanation, if there are missing items in the basic data, such as a batch of wires only giving the width but not the minimum bending radius, the system will first call the default process library for the corresponding material grade; if the parameter does not exist in the process library either, the wire will be marked as pending confirmation and will not enter the automatic path planning, but will instead output a manual confirmation prompt.

[0074] If the insulation spacing and welding height constraints cannot be satisfied simultaneously through parameter adjustment during topology generation, the system first locks the transposition relationship within the slot unchanged and only performs local rearrangement of the end paths; if it still cannot be satisfied, it reverts to the previous round of interlayer connection matrix and searches for feasible connection relationships again.

[0075] If the number of searches exceeds the preset limit, the system outputs a conflict report, indicating that there is no solution that satisfies the constraints under the current combination of slot fill rate and layer number;

[0076] For example, in the above-mentioned 800V drive motor design scenario, the stator has a 72-slot, 8-pole, 10-layer flat copper wire structure. After the system imports the slot width, slot depth, and wire size, it identifies that the first 1 to 2 layers near the slot are high leakage magnetic fields, while the 7th to 10th layers are weak leakage magnetic fields. Therefore, it automatically generates a set of asymmetric transposition connection relationships and arranges the odd-numbered slot solder joints on the 1.2mm plane and the even-numbered slot solder joints on the 3.2mm plane. It outputs a model file containing the three-dimensional coordinates of all solder joints and a laser welding processing parameter table to guide subsequent manufacturing.

[0077] The purpose of this step is to unify in-slot electromagnetic optimization, end space avoidance, and welding accessibility at the manufacturing end into the same design process, thereby realizing the transformation of the winding design scheme from two-dimensional connection logic to three-dimensional manufacturable topology.

[0078] In this embodiment of the invention, before inputting the stator slot space data, electromagnetic boundary data, and conductor physical parameters into the basic data source layer, the method further includes: preprocessing the stator slot space data, electromagnetic boundary data, and conductor physical parameters to obtain preprocessed stator slot space data, electromagnetic boundary data, and conductor physical parameters. The preprocessing includes extracting leakage flux density feature values, extracting preset bending limit radii, and data dimensionality reduction mapping. The data dimensionality reduction mapping involves transforming the physical stator dimensions in the stator slot space data into a spatial coordinate system to eliminate geometric redundancy.

[0079] This embodiment provides a parameter purification mechanism for preprocessing. Specifically, in the design process of the above-mentioned 10-layer flat copper wire drive motor, if the original simulation model and process log are directly input into the subsequent modules, it is easy to have inconsistent data scales, redundant geometric elements, and unclear electromagnetic key areas, which will lead to a decrease in the efficiency of subsequent topology planning and even meaningless local detour paths. Therefore, preprocessing is performed before formally integrating the basic data.

[0080] Specifically, the extraction of leakage flux density characteristic values ​​can be completed based on the electromagnetic simulation results of the stator slot cross section; assuming that a slot is divided into 5 sampling zones in the radial direction, from the slot opening to the slot bottom, they are denoted as zone 1 to zone 5, and the corresponding average leakage flux densities are 0.42T, 0.35T, 0.21T, 0.10T, and 0.08T, respectively.

[0081] The system does not need to retain the complete field map. Instead, it extracts feature vectors that can characterize the leakage magnetic gradient of the slot, such as [0.42, 0.35, 0.21, 0.10, 0.08]. Furthermore, it calibrates bands 1 and 2 as high-frequency sensitive regions and bands 4 and 5 as weakly sensitive regions. In this way, the subsequent transposition and connection module does not need to repeatedly read the complete grid data to know which layers are more suitable for arranging high-frequency dwell paths and which layers are more suitable as compensation paths.

[0082] The extraction of the preset bending limit radius is oriented towards the conductor forming constraint; for a certain specification of flat copper wire, if the bare copper thickness is 1.2mm, the width is 4.0mm, and the total insulation layer thickness is 0.15mm, the process database can give its minimum bending radius under stable production conditions as 3.5mm;

[0083] The system writes the radius into the physical parameters of the conductor and records its corresponding risk range simultaneously. For example, when the bending radius is between 3.0mm and 3.5mm in the design, it is marked as a critical manufacturable zone; when it is less than 3.0mm, it is judged as an unmanufacturable zone. The reason for this is that being able to connect geometrically is not the same as being able to form a shape in the process. Pre-extraction can reduce repeated backtracking in the future.

[0084] Data dimensionality reduction mapping is used to eliminate stator geometric redundancy. In the actual 3D computer-aided design model, the 72 slots of the stator are repeatedly distributed along the circumference. If all surfaces and boundary elements of each slot are completely preserved, it will bring unnecessary burden to the calculation. Therefore, the stator slot spatial data can be mapped from a complete 3D solid to a standard slot template + slot number rotation angle.

[0085] For example, first, using slot 1 as a standard slot template, its two-dimensional cross-sectional profile and end face reference position are recorded, and then the slot is... This is represented as groove 1 rotating around the central axis. Thus, the system only needs to save one template and 72 rotation parameters, instead of saving 72 duplicate geometries; for end planning, the global coordinates can be recovered using the slot number and rotation angle.

[0086] As a fallback, if the resolution of the electromagnetic simulation results is insufficient, resulting in the difference in leakage flux density between adjacent sampling bands being less than the noise threshold, for example, all falling between 0.10T and 0.12T, then the high leakage flux region and the weak leakage flux region will no longer be forcibly divided. Instead, the slot will be recorded as a uniform region and processed according to the default symmetrical dwell strategy.

[0087] If the bending limit radius of the corresponding wire specification is not found in the process library, it can be estimated by interpolation based on wires of the same material and similar thickness, and the estimation result is accompanied by a confidence level; when the confidence level is lower than the preset value, the system outputs a process mark that needs to be tested and verified; if it is found that the local groove shape is not completely consistent when restoring geometry after data dimensionality reduction, such as the presence of cooling grooves or local reinforcing ribs, the original geometry of these special grooves is retained separately and they are not involved in the template processing.

[0088] For example, in this drive motor design scenario, the original electromagnetic model contains a complete 72-slot 3D mesh, and the amount of data exceeds the preset processing threshold. After system preprocessing, only one standard slot template, 72 rotation parameters, and 5 leakage magnetic characteristic values ​​for each slot are retained. At the same time, the wire process library extracts the minimum bending radius of 3.5mm and gives the critical manufacturing zone so that subsequent modules can complete accelerated planning without losing key physical information.

[0089] The purpose of this step is to retain the key physical quantities and process boundaries that determine the design results before proceeding to the 3D topology solution, and to eliminate duplicate or redundant data, thereby forming a more stable basis for subsequent planning.

[0090] In this embodiment of the invention, the transposition connection module is used to plan the interlayer connection relationship of the conductors in the slot based on the preset asymmetric transposition connection logic and the electromagnetic field distribution characteristics in the electromagnetic boundary data.

[0091] The transposition connection module is specifically used for: initializing a preset connection matrix based on the number of layers and phases of the stator slot, refining the preset connection matrix to each stator slot through asymmetric transposition connection logic, extracting the electromagnetic high-frequency leakage magnetic field region and weak leakage magnetic field region of each stator slot from the electromagnetic boundary data, extracting the conductor dwell length from the conductor physical parameters, and planning the interlayer span of each stator slot in combination with the extracted data. The connection matrix includes a cross-layer connection matrix.

[0092] The interlayer connection relationship is forcibly defined by the cross-layer connection matrix, and the electromagnetic equivalent spiral transposition of the same phase winding on the stator circumference is quantified by the short pitch configuration with a span smaller than the pole pitch and the long pitch configuration with a span larger than the pole pitch.

[0093] Spatial location data is processed by electromagnetic field distribution characteristics, and the results of conductor current sharing planning in different regions are refined based on the relationship between the leakage flux density and interlayer span in the electromagnetic high-frequency leakage magnetic field region and the weak leakage magnetic field region.

[0094] This embodiment provides a planning mechanism for asymmetric transposition connection; specifically, although the basic three-dimensional spatial avoidance can disperse the solder joints, under high-frequency operating conditions, if each parallel branch stays in a region with different leakage magnetic flux for a long time, the branch induced potential imbalance and local circulating current may still occur.

[0095] In other words, simply solving the end interference cannot automatically yield a low AC loss structure; therefore, it is necessary to further introduce interlayer transposition logic coupled with electromagnetic distribution.

[0096] Specifically, the system first establishes a cross-layer connection matrix for each slot. Taking a single slot with 10 layers of conductors as an example, if sequential connection is used, the connection path is layer 1 to layer 2, layer 2 to layer 3, and layer 3 to layer 4. The advantage of this method is that the path is simple, but the disadvantage is that the same branch may continuously pass through adjacent high leakage magnetic layers, resulting in high-frequency loss accumulation.

[0097] In this embodiment, it is replaced with an asymmetrical connection; for example, for a slot group of a certain phase, layer 1→layer 4, layer 2→layer 5, layer 3→layer 6, layer 4→layer 2, layer 5→layer 8 can be set; in this way, the conductors in the high leakage magnetic field region will not only circulate locally with the adjacent layers, but will be forced to alternately pair with the medium and low leakage magnetic field regions.

[0098] Furthermore, the system extracts high-frequency and weak magnetic leakage regions based on the electromagnetic boundary data of each slot. If, in the radial 10 layers, the average sampling magnetic leakage values ​​for layers 1 to 3 are 0.40T, 0.34T, and 0.29T respectively, and for layers 8 to 10 are 0.11T, 0.09T, and 0.08T respectively, then layers 1 to 3 can be marked as high-frequency magnetic leakage regions, and layers 8 to 10 can be marked as weak magnetic leakage regions. The dwell length of the conductor is determined by the length of the slot segment continuously occupied by the branch in a certain layer.

[0099] For example, if the cumulative dwell length of a branch in the high leakage magnetic field region is 18mm and the cumulative dwell length in the weak leakage magnetic field region is only 6mm, its high frequency loss assessment value exceeds the preset safety threshold; the system will adjust the interlayer span accordingly, so that the branch will enter the weak leakage magnetic field region in the subsequent slot, thereby gradually approaching the balance.

[0100] The short pitch and long pitch here can be used to quantify electromagnetic equivalent helical transposition; for example, if the slot span corresponding to the motor pole pitch is 9 slots, then the short pitch can be set to 8 slots and the long pitch can be set to 10 slots; the same phase winding uses 8 slots for bridging in some slots and 10 slots for bridging in other slots, and the target layer of bridging is not symmetrical, which is equivalent to making the conductor form an equivalent helical traversal path on the stator circumference;

[0101] Although the physical conductor was not actually wound into a spiral, judging from the leakage magnetic environment it experienced, the exposure conditions of each branch in the circumferential and radial directions were balanced.

[0102] To illustrate the current sharing planning results more intuitively, a micro-example can be constructed. Assume that a phase has two parallel branches A and B. In the initial design, the dwell length of branch A in the high leakage magnetic field region is 20mm and the dwell length in the weak leakage magnetic field region is 5mm; while that of branch B is 9mm and 16mm respectively.

[0103] The system calculated that the difference in dwell length between the two exceeded a preset difference threshold. Therefore, the connection of branch A in slot 3 from layer 1 to layer 4 was changed to layer 1 to layer 7, and the connection of branch B in slot 4 from layer 8 to layer 9 was changed to layer 8 to layer 3. After the adjustment, branch A became a high leakage magnetic field area of ​​12mm and a weak leakage magnetic field area of ​​13mm, while branch B became 11mm and 14mm. As a result, the exposure of the two branches to high-frequency leakage magnetic field is closer, and the risk of local circulating current is reduced.

[0104] As a fallback, if certain slots cannot perform the specified cross-layer connection due to geometric limitations, such as the connection from layer 1 to layer 7 crossing an unavoidable area, the system will first keep the short pitch or long pitch unchanged and only replace the alternative layer with the target layer number within the same leakage flux level.

[0105] If the alternative layer is also not feasible, the circumferential span is reversed and another slot combination is searched again; if all alternatives fail, the slot group is marked as a structurally sensitive area and the avoidance priority is increased in the subsequent end topology module; on the other hand, if it is found after calculation that the difference in the dwell time of each branch in the high leakage magnetic field area and the weak leakage magnetic field area is lower than the preset threshold, such as lower than 2mm, the transposition complexity will not be increased to avoid unnecessary increase in the end path length;

[0106] For example, in this 800V drive motor design scenario, the system identified that the first three layers near the slot opening had high leakage flux at 1000Hz. Therefore, the conventional connection between adjacent layers was changed to a mixed connection across three and four layers, and some slots used short-pitch 8 slots and some slots used long-pitch 10 slots. Ultimately, the cumulative dwell length difference between the two parallel branches in the same phase in the high leakage flux region was reduced from the initial 11mm to less than 2mm.

[0107] The purpose of this mechanism is to transform the interlayer connection within the slot from a simple geometric connectivity problem into an electromagnetic equalization problem, thereby suppressing high-frequency AC losses and local circulating currents.

[0108] In this embodiment of the invention, the end topology module is used to divide each end of the stator winding into odd-numbered slot end welding points and even-numbered slot end welding points, and to decompose the spatial distribution of the end connection path to obtain the three-dimensional coordinates of each end welding point.

[0109] The end topology module is specifically used to: evaluate the spatial layout of end welding points by combining the number of winding layers, stator circumferential dimensions, end straight section extension length, and anti-interference principles derived from stator slot space data and conductor physical parameters. The evaluation process includes: using a preset stepped end topology model to predict the global welding height, extracting the axial offset and circumferential physical spacing of the end connection path, and combining the axial offset and circumferential physical spacing to evaluate the non-interference feasibility and final axial height of various stepped misalignment layouts.

[0110] This embodiment provides a three-dimensional topology decomposition mechanism for end welding points. Specifically, after introducing asymmetric transposition, although the electromagnetic balance in the slot is improved, the end connection path will be more complex than the traditional design. If all welding points are still forcibly arranged in the same plane, the circumferential spacing of the welding points will decrease rapidly after the number of layers increases, eventually resulting in local welding point overlap, collision in the bending area, or insufficient insulation distance. Therefore, it is necessary to evolve the end path from planar arrangement to a three-dimensional topology with odd and even slot layered arrangement.

[0111] Specifically, the end topology module first groups the welding points according to the parity of the slot number; assuming that in a 72-slot stator, slots 1, 3, 5, etc. correspond to odd-numbered slot groups, and slots 2, 4, 6, etc. correspond to even-numbered slot groups; each group forms an independent welding point sequence at the end, and then the three-dimensional coordinates of each welding point are calculated;

[0112] When determining the coordinates, the center point of the slot is not directly taken. Instead, the path is decomposed by combining the layer number of the conductor, the length of the straight section at the end, and the bending direction. For example, the two-dimensional circumferential position of a weld point is determined by the slot number, the radial position is determined by the layer number and the conductor width, and the axial position is determined by the plane to which it belongs and the local clearance amount.

[0113] Therefore, each weld point can be expressed as a set of X, Y, Z coordinates, where X represents the circumferential projection coordinate component of the weld point in the stator end face coordinate system, Y represents another orthogonal coordinate component in the end face paired with X, and Z represents the axial height coordinate component relative to the stator core end face; in specific implementation, the corresponding XYZ representation can also be obtained by conversion from the aforementioned R-θ-Z design coordinate system.

[0114] Global welding height prediction can be accomplished using a pre-defined stepped end topology model; this model does not directly output a unique answer, but provides multiple candidate layouts; for example, scheme A sets odd-numbered slots in the 1.0mm plane and even-numbered slots in the 3.0mm plane;

[0115] Option B places odd-numbered slots on a 1.3mm plane and even-numbered slots on a 3.3mm plane; Option C adds a local +0.2mm compensation for individual sensitive weld points on top of the above. The system will check whether the circumferential physical spacing, bending radius, and total welding height are satisfied for each option. If the circumferential spacing is insufficient, it can be alleviated by increasing the axial offset. If the total height exceeds the limit after increasing the axial offset, the option is eliminated.

[0116] A simplified example can be used to illustrate the evaluation process; assume that the circumferential projection distance between two adjacent solder joints on the same plane is only 1.1 mm, while the minimum physical distance required by the process is 1.6 mm;

[0117] If one of the weld points is raised to the second plane and an axial offset of 2.0mm is introduced, the actual minimum three-dimensional distance between the two is approximately 2.28mm according to spatial distance calculation, which is greater than 1.6mm. Therefore, it is determined that there is no interference and it is feasible.

[0118] If another solder joint is simultaneously surrounded by the other two wires, resulting in a minimum local radius of only 3.1mm after bending, which is less than the minimum forming requirement of 3.5mm, then although the solution meets the spacing requirements, it is still considered infeasible and the path needs to be reallocated.

[0119] As a fallback, if multiple solder joints still cluster in a local area after odd and even slot grouping, for example, if the three solder joints fall in the same circumferential sector due to the concentrated span of a certain phase winding in this area, the system can further introduce intra-group sorting adjustment, changing the solder joints originally arranged in layer number order to be sorted according to the target connection slot distance.

[0120] If the conditions are still not met, the odd and even slot double-group structure will be retained, and only the transition bend point in the area will be increased so that it avoids the weld in the axial direction first; if the end straight section extension length is insufficient to complete any manufacturable bend, the system will prompt to increase the straight section length or reduce the slot fill factor, instead of outputting an unmanufacturable topology.

[0121] For example, in this drive motor design scenario, there are a large number of 72 slot end weld points in the 10-layer structure. After the system groups the slots into odd and even slots, it first places the odd slot weld points on the first height plane, and then places the even slot weld points on the second height plane, and calculates the three-dimensional distance from each weld point to the adjacent weld points and the adjacent bending segment.

[0122] After global prediction, it was found that there was only local congestion near slots 17 and 18. Therefore, a transition point was added to this area and the length of the end straight section was finely adjusted to finally obtain a global end coordinate scheme that meets the requirements of no interference.

[0123] The purpose of this mechanism is to decompose complex end connection paths into a calculable and verifiable three-dimensional solder point layout problem, thereby enabling end space reuse under high-layer winding conditions.

[0124] In this embodiment of the invention, the stepped misalignment mechanism includes a first height plane, a second height plane, and a transition connection plane. The first height plane is the axial height of the welding points at the ends of the odd-numbered slots, the second height plane is the axial height of the welding points at the ends of the even-numbered slots, and the transition connection plane is the wire bending area connecting the first height plane and the second height plane.

[0125] Adjusting the axial height distribution of the end welding points according to the stepped misalignment mechanism includes: selecting the target misalignment height according to the design requirements, and using the topology generation layer to re-optimize and arrange the end connection paths that meet the target misalignment height based on the target misalignment height;

[0126] The system acquires preset welding process constraints and uses these constraints as dynamic inputs to dynamically adjust the height distribution. The welding process constraints include three-dimensional dynamic zoom laser welding parameters. When mechanical interference at the end is detected or the welding height exceeds the preset welding height threshold, the system switches to the corresponding step misalignment mechanism in real time. The system also uses the topology generation layer to replan the future end connection path and end welding point distribution through real-time calculation and simulation.

[0127] Based on the preset asymmetric transposition connection logic and the preset stepped end topology model data synchronization mechanism, the planning results are automatically updated;

[0128] This embodiment provides a dual-plane stepped misalignment and its dynamic switching mechanism; specifically, if only the odd and even slots are grouped without a clear height plane and switching rules, the end layout may still lose consistency in different design rounds. Especially when a three-dimensional dynamic zoom laser welding equipment is introduced at the manufacturing end, the weld point height not only affects spatial interference, but also the focusing cycle and welding stability.

[0129] Therefore, in this embodiment, the end height structure is explicitly defined as three parts: a first height plane, a second height plane, and a transition connection plane.

[0130] Specifically, the first height plane can be understood as the main bearing plane of the odd-numbered slot weld joints, the second height plane is the main bearing plane of the even-numbered slot weld joints, and the transition connection plane is not an independent welding plane, but a bending transition area located between the two.

[0131] For example, if the first height plane is set to be 1.1 mm from the end face of the iron core and the second height plane is set to be 3.1 mm, then the 2.0 mm height difference between the two is the target misalignment height; after the wire is led out from the slot, if the target solder joint is located on the second height plane, it needs to be lifted on the transition connection plane before reaching the target solder joint.

[0132] The dynamic programming process can be linked with welding process constraints; the system first selects the target misalignment height based on design requirements, for example, 2.0mm in the conventional scheme, and switches to 2.3mm if the estimated weld density is higher;

[0133] The three-dimensional dynamic zoom laser welding parameters are used as input conditions, including the optimal focusing height of focal plane A and focal plane B, focal plane switching time, allowable weld penetration range and maximum welding height threshold.

[0134] If mechanical interference still occurs in a certain area of ​​the current topology generation result at a misalignment height of 2.0mm, the system does not need to be completely rebuilt. Instead, it will switch to the 2.3mm misalignment mechanism first, and keep the verified and feasible in-slot transposition logic unchanged, only replanning the end path.

[0135] The data synchronization mechanism here is used to ensure consistency between the in-slot connections and the end paths; because after the end solder joint height changes, the bending length of some cross-layer connections will change, thus affecting the conductor dwell length and balance relationship; therefore, when the end is replanned, the system synchronously updates the conductor path length statistics in the transposition connection module.

[0136] For example, if the length of a branch increases by 1.8mm due to the lifting of the solder joint, the total dwell length of that branch needs to be re-accounted for; if the update causes two parallel branches to be exposed to imbalance in the high leakage magnetic field area, exceeding the threshold, the system will reallocate the cross-layer connection of another slot to restore overall balance.

[0137] A set of microscopic data can be used to illustrate dynamic switching; assuming that in the initial scheme, the weld points of slot 11 and slot 12 fall on the 1.1mm and 3.1mm planes respectively, after collision detection, it was found that the minimum spatial distance is only 1.4mm, which is lower than the required 1.6mm; at the same time, the total welding height is 3.3mm, which has not exceeded the limit;

[0138] The system then prioritizes maintaining the total height and adjusts the curvature of the transition path. If it still cannot reach 1.6mm, it switches to a 2.3mm offset scheme, raising the second plane to 3.4mm. If the total welding height reaches 3.6mm at this point, exceeding the 3.5mm threshold, it further triggers replanning, shortening the straight section at the local end or changing the circumferential sequence of the target weld points. Only after both spacing and height constraints are met is the scheme for that area confirmed.

[0139] As a fallback measure, if the focal plane switching time of the 3D dynamic zoom laser equipment exceeds the cycle time requirement, for example, the design allows a switching time of 8ms, but the actual switching time of the equipment is 15ms, the system can limit the number of plane switching and sort the welding points on the same plane for continuous processing to reduce the back-and-forth switching between the A / B planes.

[0140] If a batch of production uses a regular fixed focal depth device instead of a dual focal plane device, the system can freeze the second plane as an emergency plane for only a small number of sensitive solder joints, and return the remaining solder joints to the main plane as much as possible to take into account manufacturing compatibility; if all switching strategies result in the total height exceeding the limit, a design alarm will be output requiring adjustment of the number of wire layers or slot fill factor.

[0141] For example, in this drive motor design scenario, the production line is equipped with a three-dimensional dynamic zoom laser welding head that can quickly switch between two discrete focal planes; the system adopts a 1.1mm and 3.1mm dual-plane scheme, and then detects local interference near slot 17, so it automatically switches to the 1.1mm and 3.4mm scheme and rearranges the order of adjacent welding points; since the equipment can complete the focal plane switching in milliseconds, the new dual-plane matrix still meets the production cycle time;

[0142] The purpose of this mechanism is to synergistically link end height design with manufacturing equipment capabilities, thereby achieving a dynamic balance between weld point space avoidance, welding height control, and production cycle time.

[0143] In this embodiment of the invention, when the end welding point is at the first height plane, the axial height distribution of the end welding point is adjusted, including: setting the axial height of the odd-numbered slot end welding points to remain at the current level, setting the relationship between the welding penetration and the axial height according to a preset tolerance range, mapping based on the stator slot space data, and realizing the output of the end welding point coordinates;

[0144] When the end welding point is at the second height plane, the axial height distribution of the end welding point is adjusted, including: setting the plan for the axial offset and staggering the end welding points of even-numbered slots; calculating the extreme value of the circumferential spacing without physical interference based on the spacing between adjacent slots and the width of the conductor; staggering the end welding points of even-numbered slots according to the extreme value of the circumferential spacing; predicting the height difference of the end welding point and the proportion of the end welding point in the overall end space by combining the preset stepped end topology model with the extension length of the end straight section; setting a preset fluctuation threshold, which is the maximum allowable relative change ratio of the height difference or space proportion under different design layers;

[0145] Among them, in the prediction results, the rate of change of the height difference of the end welding point and the proportion of the end welding point in the overall end space with the increase of the number of layers is less than the preset fluctuation threshold; if the rate of change is greater than or equal to the preset fluctuation threshold, the extreme value of the circumferential spacing and the axial offset are recalculated and reassigned.

[0146] When the end weld point is in the transition connection plane, the axial height distribution of the end weld point is adjusted, including: assuming that a spatial interference conflict occurs, simulating the future avoidance path using a three-dimensional spatial matrix, evaluating the impact of the spatial interference conflict on the end insulation creepage distance and the overall weld height, and adjusting the axial height distribution of the end weld point based on the evaluation results. The spatial interference conflict includes physical overlap.

[0147] This embodiment provides a sub-plane refinement adjustment mechanism. Specifically, although the double-plane stepped misalignment can significantly alleviate end congestion, in actual design, the first height plane, the second height plane, and the transition connection plane play different roles. If the three adopt a unified adjustment rule, it is easy to cause excessive disturbance of the main plane, excessive lifting of the secondary plane, or abrupt bending in the transition zone. Therefore, this embodiment sets different calculation logics for the three regions respectively.

[0148] Specifically, when the current weld point is located on the first height plane, the odd-numbered slot weld points should, in principle, remain at the current level and not easily fluctuate up or down. The reason is that the first plane usually serves as the main welding plane and undertakes a high proportion of welding tasks. Maintaining its height stability is more conducive to manufacturing cycle control.

[0149] The system mainly adjusts the matching relationship between the weld penetration depth and the axial height within this plane; for example, if the design height of the first plane is 1.1mm and the allowable weld penetration depth range is 0.8mm to 1.2mm, then the relative position of the geometric center of the weld point and the laser focus can be controlled within the preset tolerance range.

[0150] If the estimated penetration depth of a solder joint exceeds the preset upper limit due to an increase in local copper content, the solder joint is not raised immediately. Instead, its welding posture or local path is fine-tuned first. Only when the posture adjustment is insufficient to meet the process requirements will a request be made to the next higher dynamic switching mechanism for local plane migration.

[0151] When the current solder joint is located in the second height plane, even-numbered slot solder joints are allowed to introduce axial offset more aggressively; here it is necessary to calculate the extreme value of the circumferential spacing without physical interference; for example, if the arc length corresponding to the angular distance between adjacent slots is 4.8mm, the width of a single conductor including insulation is 1.6mm, and considering the minimum insulation creepage distance of 0.4mm, then theoretically the minimum circumferential center distance between two solder joints in the same plane should not be less than 2.0mm;

[0152] If the actual calculation is only 1.5mm, the system will keep one of the weld points in the current circumferential position, and the other weld point will be axially lifted and slightly offset circumferentially to form a three-dimensional misalignment; then, the height difference of the weld point and its proportion in the overall end space will be predicted by combining the length of the straight section protruding from the end.

[0153] In this embodiment, the term "proportion" refers to the ratio of the local envelope volume formed by the second height plane weld point and its corresponding transition bend to the total end envelope volume. If only a cross-sectional level rapid evaluation is performed, the ratio of the second height plane projection area to the total end projection area within the same circumferential sector can be used as an equivalent substitute.

[0154] Accordingly, the so-called height difference refers to the difference between the target axial height of the weld point and the axial height of its exit reference point, rather than the height difference between any two weld points. In the design, it is hoped that as the number of layers increases from 8 to 10 or even 12, the rate of change of this ratio will remain stable, rather than a nonlinear abrupt change that exceeds the preset limit.

[0155] If the rate of change exceeds the preset fluctuation threshold, such as exceeding 10%, it indicates that the current misalignment strategy does not have the stability of layer expansion and the layout needs to be rebuilt.

[0156] To avoid comparison distortion caused by changes in the proportion of different design rounds, the system adopts the same end envelope definition method within the same project. For example, the overall end envelope can be defined by the axial range between the outer side of the iron core end face and the highest weld point, the radial range between the outermost conductor envelope radius and the innermost conductor envelope radius, and the circumferential coverage range of the corresponding phase group end.

[0157] The space ratio of the second height planar solder joints is always calculated based on this fixed envelope; thus, when comparing 8-layer, 10-layer, and 12-layer schemes, the rate of change reflects the scalability of the layout strategy itself, rather than the change in statistical caliber.

[0158] When the current path passes through the transition connection plane, the focus is on handling spatial interference conflicts; these conflicts include not only physical overlap, but also compression of creepage distance and increase in total welding height; the system uses a three-dimensional spatial matrix to simulate future avoidance paths;

[0159] For example, if conductor A is expected to intersect conductor B near coordinate interval Z=2.0mm when transitioning from the first plane to the second plane, the system generates two alternative paths: Path 1 uses an earlier rise and a gentler transition; Path 2 uses a later rise and a more abrupt transition. The minimum insulation distance and the final total welding height are evaluated for each of the two paths.

[0160] If the minimum distance of path 1 is 1.7mm and the height is 3.4mm, and the minimum distance of path 2 is 1.5mm and the height is 3.2mm, while the safety requirement is at least 1.6mm, then only path 1 is acceptable.

[0161] As a fallback measure, if a few weld points still exceed the penetration tolerance under the first plane stability control, these weld points can be marked as special weld points, and their energy parameters can be configured separately in the welding program, instead of immediately changing their spatial height.

[0162] If the number of solder joints on the second plane continues to increase, causing its end space ratio to grow too fast, the system will first check the repositioning connection logic to determine whether there is a possibility of dispersing the end cluster by changing the slot span.

[0163] If, in the transition connection plane, the insulation distance and total height requirements cannot be met simultaneously no matter how the plan is redesigned, the local inter-layer span should be reduced, and if necessary, the branch should be reverted to a more conservative connection relationship.

[0164] For example, in this drive motor design scenario, the first plane bears about 55% of the solder joints, and the second plane bears about 45% of the solder joints; the system keeps the main plane unchanged at 1.1mm for the odd-numbered slot solder joints, and only finely adjusts the penetration depth in the welding parameters; the even-numbered slot solder joints are staggered according to the calculated minimum circumferential center distance, so that they are concentrated in the range of 3.1mm to 3.4mm.

[0165] For several conductors that are raised from the first plane to the second plane, two raising paths are simulated in the transition zone, and a scheme that takes into account both insulation distance and total height is selected. After calculation, after the number of layers increases from 8 to 10, the rate of change of the space ratio of the solder joints on the second plane is controlled within the preset threshold.

[0166] The purpose of this mechanism is to apply differentiated constraints to different spatial levels, thereby achieving a refined layout of end solder joints that is stable, scalable, and manufacturable.

[0167] In this embodiment of the invention, the method further includes a closed-loop verification feedback step, which includes: acquiring three-dimensional point cloud data of the end conductors through a visual positioning device and extracting the actual spatial coordinates of each end welding point; verifying whether the step misalignment meets the design tolerance based on the actual spatial coordinates; injecting a pulse width modulation waveform through a high-frequency impedance testing device and acquiring transient voltage data and transient current data of the end conductors.

[0168] The transient voltage and transient current data are processed using Fast Fourier Transform to extract the fundamental current component and the higher harmonic current component. The actual AC resistance is calculated based on the fundamental current component and the higher harmonic current component, and the actual AC resistance is fed back to the basic data source layer for parameter correction.

[0169] This embodiment provides a design-manufacturing-testing closed-loop verification mechanism. Specifically, the aforementioned scheme can output a three-dimensional design model, but without physical verification, it is still impossible to confirm whether the actual bending springback, welding heat-affected zone, and high-frequency impedance performance are consistent with the design. Especially in a 10-layer flat copper wire stator, even if the height deviation of several solder joints is only 0.2mm, it may lead to local interference or laser focusing deviation. Therefore, this embodiment introduces visual positioning and high-frequency impedance testing as closed-loop feedback methods.

[0170] Specifically, after the prototype is manufactured, a visual positioning device is used to collect three-dimensional point cloud data of the end wire. This device can be composed of an industrial camera and a structured light module, and its output is a dense set of points on the surface of the end wire. After the system segments the point cloud, it extracts the geometric center coordinates of each weld point and the weld boundary contour.

[0171] For example, if the target coordinates of an odd-numbered slot weld point in the design are 28.4, 15.2, 1.1, and the measured center coordinates are 28.5, 15.1, 1.18, then its axial deviation is 0.08mm, which is within the tolerance of ±0.15mm and can be judged as qualified; if the designed height of an even-numbered slot weld point is 3.1mm, and the measured height is 3.34mm, then the deviation is 0.24mm, which exceeds the tolerance and needs to be included in the rework or parameter correction.

[0172] To make the verification rules for the stepped misalignment clearer, the system does not arbitrarily select two weld points for height comparison. Instead, it pairs adjacent odd and even slots within the same end region. For example, the odd-numbered weld points of slot 17 and the even-numbered weld points of slot 18 are grouped together. For each group, the actual stepped misalignment is calculated based on its measured axial coordinates.

[0173]

[0174] in, This represents the measured axial coordinates of the even-numbered slot weld joints in this pairing group. This represents the measured axial coordinates of the odd-numbered slot weld points in the pairing group; the actual stepped misalignment amount obtained from this is... misalignment with the design specifications Compare, if satisfied

[0175]

[0176] Then it can be determined that the group of weld points meets the stepped misalignment requirement; if multiple consecutive groups exceed the design tolerance... This indicates that it is not only a single-point processing deviation, but also that there may be an overall drift of the end topology, which requires the data to be sent back to the dynamic adjustment layer to recalibrate the plane allocation.

[0177] After passing the geometric verification, a high-frequency impedance test is then performed; the test equipment injects a preset pulse width modulation waveform into the stator winding and collects transient voltage and transient current data;

[0178] For ease of explanation, assume that a current sequence is obtained by sampling at 1000Hz, and the fundamental current component is decomposed by Fast Fourier Transform. and several higher harmonic current components ,in This represents the amplitude of the fundamental current, while These correspond to the amplitudes of the 3rd, 5th, and 7th harmonic currents, respectively.

[0179] The system estimates the actual AC resistance accordingly. There is no need to go into complicated electromagnetic derivation here. It is only necessary to explain that when the proportion of higher harmonic components exceeds the preset harmonic proportion threshold, it indicates that the proximity effect is aggravated and the local circulating current is increased. The actual AC resistance will be higher than the design expectation.

[0180] The system compares the AC resistance value with the predicted value in the design model and then feeds it back to the basic data source layer to correct the leakage flux distribution calibration or conductor path parameters.

[0181] To avoid the misinterpretation that calculating the actual AC resistance based on the fundamental and higher harmonic current components is simply a direct calculation of the resistance value based on the frequency spectrum, this embodiment employs a two-stage processing approach. The first stage utilizes Fast Fourier Transform to extract... Equal components are formed to create harmonic structure features, which are used to identify the sources of increased AC loss;

[0182] The second stage involves determining the average active power from transient voltage and transient current within the same test window. And deduct the fixed losses obtained from the no-load calibration of the fixture, leads, and test circuit. Combined with the effective value of the current within that window Calculate the actual AC resistance:

[0183]

[0184] in, The corresponding average active power within the test window. Corresponding to the fixed loss obtained through no-load calibration, The corresponding effective value of the current within the test window, and The actual AC resistance obtained using the unified power conversion method;

[0185] In other words, the fundamental current component and higher harmonic current components are used to construct the pair The attribution criteria and version comparison standards, and It still outputs according to the same power conversion algorithm, thereby ensuring comparability between different similar machines and different batches;

[0186] To make the two-stage processing more feasible, the system first performs time-domain synchronous sampling within each test window, and then performs a fast Fourier transform to obtain the component proportions of the fundamental frequency and higher harmonics. Its expression is:

[0187]

[0188] This indicator This represents the combined proportion of higher harmonic currents relative to the fundamental current; when When the frequency increase is significant but the geometric deviation of the weld point is still within the tolerance, it is primarily judged to be an increase in high-frequency additional loss caused by in-slot transposition.

[0189] when When the height of the second plane solder joints increases as a whole, the primary judgment is that the increased additional loss is caused by the lengthening of the end path or the steepness of the transition zone. Thus, the system retains the verification path based on the fundamental current component and the higher harmonic current component, while avoiding the simple confusion between the harmonic ratio and the AC resistance value.

[0190] An example of local analysis is: the design predicted the AC resistance of a certain prototype to be 12.0mΩ, while the actual measured resistance at 1000Hz was 12.9mΩ; after analysis, visual data showed that the actual height of the two second plane solder joints near slot 17 was slightly higher, resulting in a steeper transition bend and increasing the length of the relevant branch ends;

[0191] Meanwhile, the results of the fast Fourier transform showed that the fifth harmonic current component was higher than that of similar prototypes; the system then fed back the actual spatial coordinates of the area and the measured AC resistance, corrected the record of the wire dwell length of the slot group, and redistributed the local cross-layer connection in the next round of design.

[0192] As a supplementary explanation, if there is occlusion in the point cloud acquisition, making the outline of some weld points incomplete, the system can call the prior positions in the design model to perform point filling and fitting on the point cloud; if the fitting residual is still higher than the threshold, the weld point will be marked for manual re-inspection.

[0193] If the sampling noise is too high during the high-frequency impedance test, causing the harmonic components to be unstable after the fast Fourier transform, the sampling window can be increased or the average of multiple test results can be calculated before calculating the AC resistance. If all the geometric dimensions are qualified but the AC resistance is still abnormally high, the priority should be to check whether the transposition in the slot is really carried out according to the design, rather than looking for the cause only from the welding height.

[0194] For example, during the prototype development phase of the drive motor project, the system first acquires the three-dimensional point cloud of the entire end region using a structured light device to verify whether the actual step misalignment between odd and even slots is maintained at around 2.0 mm.

[0195] Then, a 1000Hz pulse width modulation waveform is injected into the prototype using a high-frequency impedance testing device to obtain transient voltage and current, and the harmonic components are extracted by fast Fourier transform. If the AC resistance of a certain batch of prototypes is found to be too high, the system will automatically associate it with the corresponding design version and start the next round of parameter correction.

[0196] The purpose of this mechanism is to advance the design goal from theoretical feasibility to experimental closed-loop verifiability, thereby achieving the linkage correction of geometric accuracy and electromagnetic performance.

[0197] In this embodiment of the invention, the actual AC resistance is fed back to the basic data source layer for parameter correction, including: setting a first AC resistance threshold and a second AC resistance threshold, wherein the first AC resistance threshold is greater than the second AC resistance threshold; if the actual AC resistance is greater than the first AC resistance threshold, it is determined that there is a local circulating current in the interlayer connection relationship of the conductor in the slot, and the transposition connection module is triggered to replan the interlayer span.

[0198] If the actual AC resistance is less than the second AC resistance threshold, it is determined that the interlayer connection relationship of the conductors in the slot has reached electromagnetic equilibrium, and the final winding design scheme is output; if the actual AC resistance is greater than or equal to the second AC resistance threshold and less than or equal to the first AC resistance threshold, it is determined that the design is in a critical state, and the sampling frequency of the high-frequency impedance test equipment is increased to reacquire transient voltage and transient current data.

[0199] This embodiment provides a feedback decision mechanism based on AC resistance threshold. Specifically, simply sending the measured AC resistance back to the basic data layer is not enough, because without clear judgment rules, the system cannot determine whether the reconnection should be redone immediately or only additional testing is needed. To avoid repeated design oscillations, this embodiment introduces a dual threshold mechanism to process the feedback results in a graded manner.

[0200] Specifically, the system presets a first AC resistance threshold and a second AC resistance threshold, with the former being higher than the latter; for example, for a certain specification prototype, the design target AC resistance is 12.0mΩ, the first threshold can be set to 13.0mΩ and the second threshold can be set to 12.3mΩ.

[0201] If the measured value is greater than 13.0mΩ, it means that the difference between the measured value and the target value exceeds the set judgment range. It is usually no longer attributed to test fluctuations, but more likely to be due to local circulation or uneven residence caused by interlayer connection relationship. At this time, the system directly triggers the repositioning connection module to replan the interlayer span. During replanning, the bridging target and circumferential pitch of the relevant layers in the high leakage magnetic field area are adjusted first, rather than changing the welding process parameters that have been verified to be relatively stable.

[0202] To avoid threshold misjudgment caused by differences in temperature rise, fixture wear compensation, or sampling window variations, the first and second AC resistance thresholds in this embodiment both correspond to a unified comparison benchmark, i.e., the converted actual AC resistance. The comparisons were conducted under the same reference temperature, the same pulse width modulation spectrum window, and the same test fixture conditions.

[0203] In other words, the system does not directly compare the original voltage and current peaks or uncorrected resistance values ​​with the dual thresholds; if the sample test temperature deviates from the reference temperature, it is first converted to a unified reference before entering the threshold decision; this can avoid the same topology scheme being misjudged as needing to be redesigned simply because of different test environments.

[0204] If the measured value is less than 12.3mΩ, the AC loss is determined to meet the preset performance requirements. The system can determine that the design has reached the electromagnetic equilibrium state and solidify the current winding scheme as the final design version. The final version here does not exclude subsequent process fine-tuning, but means that there is no need to continue to reconstruct at the topology logic level.

[0205] If the measured value is between 12.3mΩ and 13.0mΩ, it is considered to be in a critical state. At this time, direct redesign may be too aggressive, while direct release may leave performance fluctuation risks. Therefore, the system chooses to increase the sampling frequency of the high-frequency impedance test equipment and re-acquire transient waveforms.

[0206] For example, the initial sampling frequency is 20kHz, which can be increased to 50kHz and the sampling time can be extended; if the AC resistance drops below 12.25mΩ after retesting, it enters the pass state; if it is stable above 13.0mΩ after retesting, it is upgraded to the replanning state; if it is still in the critical region, it is determined whether the test noise or the local structural problem is dominant by combining the geometric deviation data.

[0207] To further suppress the jitter in the judgment of the critical zone, the system does not immediately draw a conclusion based on the result of a single test window after entering the critical state. Instead, it acquires at least three sets of independent test windows and performs a re-judgment based on the corresponding point cloud data.

[0208] If the median or batch mean of the three sets of results still falls within the critical region, and the point cloud shows that the local height deviation is concentrated in the same slot group, then the slot group is identified as the structurally sensitive area first, and the end path is locally corrected.

[0209] If the three sets of results are significantly discrete and the point cloud deviations do not show obvious clustering, then the sampling window should be extended or the test wiring should be checked first to avoid mistaking the test noise as a transposition imbalance. In this way, the dual threshold mechanism not only provides the upper and lower limit boundaries, but also provides the stabilization processing rules for the intermediate gray area.

[0210] A continuous prototype iteration example can be used to illustrate this; the measured AC resistance of the first prototype was 13.4mΩ, which exceeded the first threshold. Based on this, the system determined that there was obvious unevenness in the high leakage magnetic field area of ​​the slot. The layer 1→layer 4 was changed to layer 1→layer 7, and the short pitch and long pitch combination of the relevant slots were adjusted simultaneously.

[0211] The second prototype measured 12.7mΩ, falling into the critical region. Instead of immediately modifying the structure, the system increased the sampling frequency for retesting. The retest result was 12.85mΩ, and the corresponding point cloud data showed a large deviation in the height of the solder joints in slot 17. Therefore, the system determined that this deviation had an amplifying effect on the impedance and first corrected the local path at the end. The third prototype measured 12.18mΩ, which was below the second threshold. Based on this, the system output the final winding design scheme.

[0212] As a supplementary explanation, if the AC resistance distribution of multiple samples of the same design version is significantly dispersed due to individual differences in the samples, the system will use batch statistics rather than individual sample values ​​for judgment, for example, using the mean plus the standard deviation to form a conservative criterion.

[0213] If the harmonic characteristics cannot be stably extracted even after increasing the sampling frequency, the results of DC resistance, solder joint height deviation and local temperature rise test can be temporarily combined for auxiliary judgment to avoid distortion of a single indicator; if the system stays in the critical region for several consecutive rounds, the system can trigger a conservative strategy of freezing the mechanical structure and only adjusting the test window to prevent frequent design fluctuations.

[0214] For example, in this drive motor design scenario, the system sets 13.0mΩ as the upper threshold and 12.3mΩ as the lower threshold. If the measured value of the prototype is higher than the upper threshold, the system automatically returns to the transposition connection module to replan the interlayer span. If it is lower than the lower threshold, the current design version is output as the final stator winding manufacturing parameters. If it is in the middle range, the test sampling frequency is increased first and the point cloud deviation is combined for secondary confirmation before deciding whether to iterate.

[0215] The purpose of this mechanism is to establish clear decision boundaries for closed-loop correction, thereby stabilizing and engineering the design convergence process.

[0216] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.

Claims

1. A design method for a flat copper wire stator winding of an electric motor, characterized in that, include: A basic data source layer is constructed to integrate stator slot spatial data, electromagnetic boundary data, and conductor physical parameters; A topology generation layer is constructed, which plans the winding spatial distribution and end connection path based on the integrated stator slot space data, electromagnetic boundary data and conductor physical parameters. The topology generation layer includes a transposition connection module and an end topology module. The transposition connection module plans the inter-layer connection relationship of conductors in the slot, and the end topology module plans the spatial layout of end welding points. A dynamic adjustment layer is constructed, and a three-dimensional spatial matrix is ​​established with the stator core end face as the reference surface. A stepped misalignment mechanism based on the three-dimensional spatial matrix is ​​set. The stepped misalignment mechanism includes the preset stepped misalignment amount and design tolerance. The axial height distribution of the end welding points is adjusted according to the stepped misalignment mechanism. Based on the adjusted axial height distribution of the end welding points and the planned winding space distribution, a three-dimensional design model or manufacturing parameters of the stator winding are generated and output. The transposition connection module is used to plan the interlayer connection relationship of conductors in the slot based on the preset asymmetric transposition connection logic and the electromagnetic field distribution characteristics in the electromagnetic boundary data. The transposition connection module is specifically used for: initializing a preset connection matrix based on the number of layers and phases of the stator slot, refining the preset connection matrix to each stator slot through asymmetric transposition connection logic, extracting the electromagnetic high-frequency leakage magnetic field region and weak leakage magnetic field region of each stator slot from the electromagnetic boundary data, extracting the conductor dwell length from the conductor physical parameters, and planning the interlayer span of each stator slot in combination with the extracted data. The connection matrix includes a cross-layer connection matrix. The interlayer connection relationship is forcibly defined by the cross-layer connection matrix, and the electromagnetic equivalent spiral transposition of the same phase winding on the stator circumference is quantified by the short pitch configuration with a span smaller than the pole pitch and the long pitch configuration with a span larger than the pole pitch. Spatial location data is processed by electromagnetic field distribution characteristics, and the results of conductor current sharing planning in different regions are refined based on the relationship between the leakage flux density and interlayer span in the electromagnetic high-frequency leakage magnetic field region and the weak leakage magnetic field region. The end topology module is used to divide each end of the stator winding into odd-numbered slot end welding points and even-numbered slot end welding points, and to decompose the spatial distribution of the end connection path to obtain the three-dimensional coordinates of each end welding point. The end topology module is specifically used to: evaluate the spatial layout of end welding points by combining the number of winding layers, stator circumference dimensions, end straight section extension length, and anti-interference principles derived from stator slot space data and conductor physical parameters. The evaluation process includes: The global welding height is predicted using a pre-defined stepped end topology model. The axial offset and circumferential physical spacing of the end connection path are extracted. The axial offset and circumferential physical spacing are combined to evaluate the non-interference feasibility and final axial height of various stepped staggered layouts.

2. The design method for the flat copper wire stator winding of a motor according to claim 1, characterized in that, Before inputting the stator slot space data, electromagnetic boundary data, and conductor physical parameters into the basic data source layer, the following steps are also included: The stator slot spatial data, electromagnetic boundary data, and conductor physical parameters are preprocessed to obtain the preprocessed stator slot spatial data, electromagnetic boundary data, and conductor physical parameters. The preprocessing includes extracting leakage flux density feature values, extracting preset bending limit radii, and data dimensionality reduction mapping. Data dimensionality reduction mapping is to perform spatial coordinate system transformation on the physical stator dimensions in the stator slot spatial data to eliminate geometric redundancy.

3. The design method for the flat copper wire stator winding of a motor according to claim 1, characterized in that, The stepped misalignment mechanism includes a first height plane, a second height plane, and a transition connection plane. The first height plane is the axial height of the welding points at the ends of the odd-numbered slots, the second height plane is the axial height of the welding points at the ends of the even-numbered slots, and the transition connection plane is the wire bending area connecting the first height plane and the second height plane.

4. The design method for the flat copper wire stator winding of a motor according to claim 3, characterized in that, Adjusting the axial height distribution of the end welding points based on the stepped misalignment mechanism includes: Select the target misalignment height according to the design requirements, and based on the target misalignment height, use the topology generation layer to re-optimize and arrange the end connection paths that meet the target misalignment height. The system acquires preset welding process constraints and uses these constraints as dynamic inputs to dynamically adjust the height distribution. The welding process constraints include three-dimensional dynamic zoom laser welding parameters. When mechanical interference at the end is detected or the welding height exceeds the preset welding height threshold, the system switches to the corresponding step misalignment mechanism in real time. The system also uses the topology generation layer to replan the future end connection path and end welding point distribution through real-time calculation and simulation. Based on the preset asymmetric transposition connection logic and the preset stepped end topology model data synchronization mechanism, the planning results are automatically updated.

5. The design method for the flat copper wire stator winding of a motor according to claim 4, characterized in that, When the end weld point is at the first height plane, adjust the axial height distribution of the end weld point, including: The axial height of the welding points at the ends of the odd-numbered slots is set to remain at the current level. The relationship between the weld penetration and the axial height is set according to the preset tolerance range. The coordinates of the end welding points are output based on the space data of the stator slots. When the end weld point is at the second height plane, adjust the axial height distribution of the end weld point, including: The plan for axial offset is set, and the welding points at the ends of even-numbered slots are staggered. The extreme value of circumferential spacing without physical interference is calculated based on the spacing between adjacent slots and the width of the conductor. The welding points at the ends of even-numbered slots are staggered according to the extreme value of circumferential spacing. By combining the extension length of the straight section at the end with the preset stepped end topology model, the height difference of the end welding points and the proportion of the end welding points in the overall end space are predicted. The preset fluctuation threshold is set. The preset fluctuation threshold is the maximum allowable relative change ratio of the height difference or the proportion of the space under different design layers. Among them, in the prediction results, the rate of change of the height difference of the end welding point and the proportion of the end welding point in the overall end space with the increase of the number of layers is less than the preset fluctuation threshold; if the rate of change is greater than or equal to the preset fluctuation threshold, the extreme value of the circumferential spacing and the axial offset are recalculated and reassigned. When the end weld point is on the transition connection plane, adjust the axial height distribution of the end weld point, including: Assuming a spatial interference conflict occurs, a three-dimensional spatial matrix is ​​used to simulate the future avoidance path, assess the impact of the spatial interference conflict on the end insulation creepage distance and the overall welding height, and adjust the axial height distribution of the end welding points based on the assessment results. The spatial interference conflict includes physical overlap.

6. The design method for the flat copper wire stator winding of a motor according to claim 1, characterized in that, The method also includes a closed-loop verification feedback step, which includes: The three-dimensional point cloud data of the end wires is collected by a visual positioning device, and the actual spatial coordinates of each end welding point are extracted. Verify whether the stair misalignment meets the design tolerance based on actual spatial coordinates; A pulse width modulation waveform is injected using a high-frequency impedance testing device, and transient voltage and current data at the terminals are collected. Fast Fourier Transform is used to process transient voltage and transient current data to extract the fundamental current component and higher harmonic current components. The actual AC resistance is calculated based on the fundamental current component and the higher harmonic current component, and the actual AC resistance is fed back to the basic data source layer for parameter correction.

7. The design method for the flat copper wire stator winding of a motor according to claim 6, characterized in that, The actual AC resistance is fed back to the basic data source layer for parameter correction, including: Set a first AC resistance threshold and a second AC resistance threshold, wherein the first AC resistance threshold is greater than the second AC resistance threshold; If the actual AC resistance is greater than the first AC resistance threshold, it is determined that there is a local circulating current in the interlayer connection relationship of the conductors in the slot, and the transposition connection module is triggered to re-plan the interlayer span. If the actual AC resistance is less than the second AC resistance threshold, it is determined that the interlayer connection relationship of the conductors in the slot has reached electromagnetic equilibrium, and the final winding design scheme is output. If the actual AC resistance is greater than or equal to the second AC resistance threshold and less than or equal to the first AC resistance threshold, the design is determined to be in a critical state, and the sampling frequency of the high-frequency impedance test equipment is increased to reacquire transient voltage and transient current data.

Images