Method and apparatus for determining noise and electromagnetic vibration levels

Abstract

The invention relates to a method and a device for determining noise and electromagnetic vibration levels in a machine equipped with an electric motor. To this end, the computer device: acquires (31, 32) a set of load conditions and characteristic modes of the machine; determines (33) a significant subset; acquires (34) a set of relevant nodes and calculates (35) a set of frequency response functions limited to these relevant nodes; acquires (36, 37) a set of operating points and magnetic states of the machine; determines (38) a set of operating loads of the machine; then determines (39) the noise and electromagnetic vibration levels based on the frequency response functions and the operating loads.

Description

Technical Field

[0001] This invention relates to an electrical system for emitting electromagnetically generated acoustic noise in virtual prototyping.

[0002] The present invention relates more particularly to a method for determining the electromagnetic origin noise and vibration levels of a machine containing an electric motor.

[0003] The present invention also relates to apparatus for carrying out this method.

[0004] The present invention also relates to designing a machine having an electric motor and implementing this method.

[0005] Therefore, the present invention will find many advantageous applications in the design of electrical machines in fields such as transportation (automobiles, railways, maritime), industry, energy, medical and household applications. Background Technology

[0006] All mechanical systems generate noise. Often, resonance between structural modes and excitation forces is the primary cause of noise and vibration problems, or greatly contributes to them. These resonances can originate from specific components of the system, or from the entire system or structure.

[0007] These noise and vibration phenomena occur in operating machines, especially during the rotational motion of rotating machines. High noise or vibration levels can lead to a loss of acoustic or manual comfort, and even health problems with prolonged exposure to noise. They are also a source of fatigue failure and damage to the physical structure of the system.

[0008] In the simplest case, the origin of resonant vibrations can be identified empirically (e.g., problems with clamping mechanical parts) or by adequately adjusting the system structure to limit noise (e.g., by adding an acoustic enclosure). However, due to the complexity of the excitation forces involved, in the case of electromagnetically originating noise, this origin quickly becomes much more difficult to identify.

[0009] The applicant observed that solutions proposed to date for predicting the magnetic vibrations of a given machine are based on numerical simulations coupled with different finite element software, which are used for general purposes in low-frequency electromagnetics, structural dynamics, and linear acoustics. These solutions thus provide detailed calculations of the magnetic force at variable velocities in the first step, calculations of vibrations via modal extension in the second step, and assessment of acoustic radiation around the machine in the third step. These solutions enable accurate results regarding the level of vibration produced by a given machine, but require very high computation time, approximately 15 hours per machine. Therefore, this approach is incompatible with the design phase of electrical machines, as time-consuming simulations are required to estimate their electromagnetic noise with each iteration or design change. The applicant also proposes that other complementary methods have been developed, based on approximate calculations of the magnetic force or synthesis of electromagnetic vibrations, and capable of utilizing the linearity of the structure to accelerate computation time. Therefore, these computation times can be reduced to approximately 1.5 hours in the simplest cases.

[0010] However, the applicant points out that these calculations remain cumbersome, consuming significant amounts of memory during computation (on RAM) and even more during output computation (on hard disk), becoming even heavier under complex load conditions, such as machines with skewed rotors. Furthermore, these solutions only provide a final estimate of the total noise generated by the machine. Therefore, it is difficult to identify how to reduce noise, and improving the machine design requires multiple simulations to find a solution that minimizes noise.

[0011] The applicant argues that the current solution has a very long computation time, is incompatible with iterative design methods, consumes a large amount of memory during computation (on RAM) and when calculating output (on hard disk), and the results are susceptible to numerical noise due to grid projection, digital interpolation and spectral spread, and does not provide information or distinction regarding the origin of magnetic noise and vibration.

[0012] Therefore, the solutions proposed by the applicant for determining electromagnetic noise and vibration levels are unsatisfactory and particularly unsuitable for industrial design and prototyping needs. Summary of the Invention

[0013] The present invention aims to improve the above situation.

[0014] The present invention aims to remedy the aforementioned drawbacks, particularly by providing a method and apparatus for determining electromagnetically originating noise, which enables such determination to be integrated into the design phase, thereby providing a simplifying and assistive tool for the design of machines with electric motors without slowing them down. According to a first aspect, the subject matter of the invention relates to a method for determining the electromagnetically originating noise and vibration levels of an electric machine with a mechanical structure, said machine comprising a rotor and a stator separated by an air gap, the method being implemented by at least one processor, the method comprising the following steps: - Obtain first data representing a set of machine-related load cases, each load case corresponding to a magnetic load component decomposed according to a mathematical basis, preferably a Fourier series; - Obtain second data representing the set of intrinsic modes of the mechanical structure; - Determine a subset of load conditions and significant intrinsic modes from the set of load conditions and the set of intrinsic modes; - Obtain third data representing the set of relevant nodes associated with the mechanical structure; - Calculate the set of frequency response functions by limiting modal expansion to the relevant set of nodes, load conditions, and subsets of significant intrinsic modes; - Obtain fourth data representing the first set of operating points of the machine, each operating point being associated with the machine's torque and speed; - Obtain the fifth data representing the set of magnetic states of the machine, where each magnetic state is associated with an operating point; - Determine the set of working loads from the set of magnetic states; and - Determine information representing the level of electromagnetic noise and vibration originating from the set of working loads and the set of frequency response functions.

[0015] This is understood to mean that load conditions correspond to the components of the force applied to the electric machine, particularly at different points on the rotor and stator. The magnetic loads of an electric machine are defined, for example, according to the machine's design parameters, i.e., they follow the structure of the machine's magnetic circuit. Load conditions are associated with any fixed amplitude or operating level, such as unit load, that is, associated with an amplitude of 1 N to simplify calculations.

[0016] Each load case includes, for example, all of the following information: - Load type: force or torque; - Direction of application, for example, for force: radial, circumferential, axial; - Applying structure: rotor or stator; and - Wave number: r = 0, 1, 2, etc.

[0017] According to a specific embodiment, the set of load cases is limited to the lowest wavenumber present in the machine, for example, only excitations with wavenumber r = 0. Those skilled in the art will understand that the lowest wavenumber is associated with the most important excitation, and therefore this limitation can avoid unnecessary calculations without affecting the accuracy of the results.

[0018] Further understanding reveals that the set of eigenmodes corresponds to representing the resonant behavior of the machine in a manner suitable for vibration analysis. Each eigenmode is characterized, for example, by modal deformation, natural frequency, and modal damping. Those skilled in the art will understand that the response to any excitation will correspond to a unique linear combination of eigenmodes. Determining the load conditions and significant subsets of eigenmodes corresponds to identifying the combinations of load conditions and eigenmodes that lead to significant vibrations. In other words, this determination enables the identification of the mode whose shape is closest for each load condition. This design thus allows for the reduction of elements in the set of load conditions and the set of eigenmodes that significantly contribute to electromagnetically generated noise, thereby avoiding redundant calculations.

[0019] It is also understood that the relevant set of nodes corresponds to all nodes generated relative to the meshing of the machine structure, reduced to nodes relevant to the computation. These nodes may be relevant for the visualization of modes of interest, for computation, or for verifying design criteria (e.g., vibration criteria). Relevant nodes are reduced, for example, to the tip of a motor tooth where most of the magnetic force is applied, a machine anchor point to which vibration design criteria are applied, and the structural envelope point responsible for acoustic radiation by its normal vibration. The modal deformations of the eigenmodes of the aforementioned determined subset are reduced to this relevant set of nodes; that is, during the computation of the set of frequency response functions, modal deformations are computed only for the relevant set of nodes.

[0020] Those skilled in the art will understand that the set of frequency response functions makes it possible to obtain, for example, the amplitude or sound pressure amplitude of vibration under load conditions on the intrinsic modes at each relevant node.

[0021] The modal expansion principle is known from existing technologies and represents a significant source of computation time, especially at high frequencies where modal density is higher. By limiting it to relevant modes, relevant excitations, and relevant nodes, it is possible to greatly limit the computation time and memory load associated with noise determination without affecting the accuracy of the results.

[0022] Therefore, the first set of steps enables the characterization of the structure and vibration behavior of electric machines.

[0023] Secondly, or, for example, in parallel with the first set of steps, the second set of steps enables the characterization of the machine's working force level, that is, the complex amplitude (amplitude and phase) of each magnetic load condition applied to the machine at the machine's operating point.

[0024] Therefore, it can be understood that the set of operating points, i.e., the combination of torque and speed of the electric motor, corresponds to the various speeds at which the machine can operate. The set of magnetic states thus corresponds to the magnetic flux and / or force generated in the machine when the latter is operating at a given operating point.

[0025] Each magnetic state includes, for example, the magnetic dipole of each stator tooth, including radial force, circumferential force, and torque. Therefore, the set of magnetic states makes it possible to determine the set of operating levels and evaluate the strength of a given load at the operating point, for example, in N.

[0026] In other words, the set of magnetic states corresponds to a set of excitation forces associated with the operating point.

[0027] Therefore, the set of working loads can be determined from the set of magnetic states, for example, by integrating the Maxwell tensor along the tooth pitch over the magnetic flux distribution in the middle of the air gap. A more accurate method for calculating the working load is to apply a method based on the virtual works of the magnetic field distribution on the magnetic grid, and then integrate the force dipole along the tooth pitch. Regardless of the method chosen for calculating the magnetic dipole, the dipole is decomposed into a Fourier series to determine the amplitude in N and the phase in radians for each load condition (e.g., a radial force applied to the stator with wavenumber r = 0) and each operating point. Thus, the determination of information representing the electromagnetic noise and vibration level corresponds to a mapping of the frequency response function, that is, the vibration produced by a unit load, to the working load, i.e., the intensity of the load during machine operation.

[0028] Therefore, information representing the level of electromagnetic noise and vibration includes noise and vibration levels as a function of the operating point (i.e., the combination of torque and speed), making it possible to identify the operating point that generates noise. Based on this information, the control of the electric machine can be adjusted to avoid maximum vibration and minimize noise and failure risks.

[0029] The applicant therefore argues that, due to this invention, by calculating only a reduced number of nodes, intrinsic modes, and load conditions, the determination of the magnetic origin noise and vibration level of electric machines is simplified, the calculation time is shortened, and electromagnetic vibration analysis can be effectively integrated into the design process of electric machines.

[0030] In an advantageous embodiment of the invention, the method further includes displaying information belonging to a set of information via a human-machine interface, the set of information including: - Information representing noise and vibration levels; - Information representing the load conditions and a subset of significant intrinsic modes; and - Information representing the set of frequency response functions.

[0031] It is understood here that the display of representative information corresponds to the presentation of noise and vibration levels determined during the method, preferably visual or graphical, or other relevant elements that can assist the design process of electric machines. Of course, such a display depends on the representative information determined, particularly its level of detail. According to the variant embodiments described below, the method of the present invention enables, for example, the estimation of the overall noise or vibration level generated by the machine in a global manner or based on determined points or surfaces, or the estimation of a specific noise or vibration level generated by the considered modes and / or load conditions in a specific manner, based on the operating level determined during the execution of the method. According to another variant, the method of the present invention enables the estimation of coefficients representing the correspondence between load conditions and intrinsic modes of a subset of load conditions and intrinsic modes, for example, through projection as described below. The display then corresponds, for example, to the display of a set of projections between load conditions and intrinsic modes from which a subset of load conditions and intrinsic modes are determined. It is further understood that the method can be configured to display a variety of determined quantities or functions during execution, particularly frequency response functions associated with multiple load conditions.

[0032] In a particular embodiment, the method further includes receiving information representing a set of magnetic loads associated with a machine, wherein obtaining the first data corresponds to determining a set of load conditions associated with the machine from the set of magnetic loads.

[0033] This is understood to mean that the method determines the set of load cases by applying the principle of mathematical basis decomposition, which is applied to the set of magnetic loads. The decomposition corresponds to, for example, Fourier series decomposition or any other mathematical basis known to those skilled in the art.

[0034] The set of magnetic loads is defined, for example, by a set of analytical equations on the structure, characterizing the frequency and wavenumber of the stator and rotor loads depending on the machine type and its fault condition.

[0035] According to another example, information representing a set of magnetic loads includes: - Information representing the number of stator slots in the machine; - Information representing the number of machine pole pairs; and - Information representing the number of phases in the machine.

[0036] It will be understood that this type of information corresponds to the discrete design parameters of the electric machine, from which analytical equations supporting the load spectrum can be established.

[0037] According to the alternative variant, the acquisition of the first data item corresponds to the direct receipt of the load case set, for example, determined by analysis outside of method execution.

[0038] In an additional embodiment, the method further includes receiving information representing a mechanical structure, wherein obtaining the second data corresponds to determining a set of intrinsic modes from the structure.

[0039] Those skilled in the art will understand here that all intrinsic modes can be determined from the structure, for example, by modeling the stator as an equivalent cylinder or by more precisely using a beam model, where the stator teeth and yoke are modeled as beam elements. Information representing the machine structure corresponds, for example, a model of the machine or a set of parameters representing the machine, enabling it to be modeled.

[0040] According to other variant embodiments, the intrinsic mode set is received by the method and determined outside of its execution by methods known to those skilled in the art.

[0041] The intrinsic mode set is determined, for example, by mechanical finite element calculations, i.e., by solving the eigenvalue problem of the machine stiffness and mass matrix, or by experimental modal analysis.

[0042] In an additional embodiment, the method includes the following steps: - Determine the set of projections between the load condition and the intrinsic modes, where each projection in the set is associated with both the load condition and the intrinsic mode; and - Each projection in the projection set is compared with a threshold, and the load condition and subset of significant intrinsic modes are determined based on the comparison results.

[0043] In other words, when the projection of the load condition onto the intrinsic mode exceeds a threshold, each significant combination of the load condition and the intrinsic mode is identified. The load condition and the intrinsic mode are represented, for example, as representative vectors, with each projection corresponding to a scalar product between the two vectors.

[0044] Therefore, it can be understood here that the projection exceeding the threshold corresponds to a combination of modes whose load condition shape is similar to that of a specific mode, that is, the excitation and the defined structural mode enter spatial resonance.

[0045] Each projection in the projection set is compared with a threshold, for example, to form a subset of significant projections, and to extend to subsets of load cases and significant intrinsic modes.

[0046] According to another variant, the calculation of the projection set makes it possible to identify the mode whose shape is closest to that force for each load case, such as the load case with wave number r = 0, i.e., the mode that projects the highest value to the load case.

[0047] In yet another embodiment, the method further includes determining a set of resonant velocities based on load conditions and a subset of significant intrinsic modes, each resonant velocity being associated with an intrinsic mode of the subset, and a first set of operating points being associated with the set of resonant velocities.

[0048] It is understood here that determining the load conditions and a significant subset of eigenmodes enables the determination of one or more velocities at which the combination of load conditions and eigenmodes leads to resonance. Each eigenmode is associated, for example, with a specific frequency or angular frequency, from which the resonant velocity can be determined, for example, via the analytical equations described above.

[0049] The first set of operating points, which correlates operating points with resonant velocities (i.e., operating points corresponding to velocities and resonant velocities), ensures that calculations are performed relative to the velocities that generate the most noise and vibration, thereby ensuring accurate calculations. This first set of operating points may be expanded, for example, to include each identified resonant velocity to improve calculation accuracy and / or limited to operating points associated with the identified resonant velocities to minimize computation time.

[0050] In another embodiment that can be combined with the aforementioned mode, the method further includes determining a set of cross-projection coefficients between load conditions, each cross-projection coefficient being associated with a combination of two load conditions in the set of load conditions, and representative information on electromagnetic origin noise and vibration levels being determined based on the set of cross-projection coefficients.

[0051] The applicant argues that the use of cross-projection coefficients allows for the acquisition of multiple independent coefficients, independent of the operating load, that participate in the expression of the envelope mean square vibration velocity representing the overall noise level. This limits the calculations to determining the electromagnetic source noise and vibration level information, thus simplifying the matching of the response function with the operating load. Therefore, determining the set of cross-projection coefficients reduces the total computational load under variable speed conditions and facilitates handling a large number of operating points. For example, the cross-projection coefficients, similar to the aforementioned projection set, correspond to the scalar product between the vectors representing the load conditions. For instance, the cross-projection coefficients can advantageously be determined only for the aforementioned load conditions and load conditions within the subset of characteristic modes, or more generally, only for load conditions that take into account the determination of electromagnetic source noise and vibration levels.

[0052] In another embodiment, the method further includes acquiring sixth data representing a set of radiation factors, each radiation factor being associated with a characteristic mode in the characteristic mode set, and the electromagnetic source noise and vibration level representation information being determined based on the sixth data.

[0053] The applicant argues that using a set of radiation coefficients associated with the intrinsic modes allows for more refined calculations, particularly power and sound pressure levels, i.e., representative information on noise and vibration levels. Therefore, the accuracy of the results is improved, especially in the low-frequency range. According to yet another example, radiation coefficients are used only in the excitation frequency range below a given threshold.

[0054] As mentioned above, the set of radiation coefficients is obtained, for example, only for eigenmodes that belong to the loading condition and a subset of significant eigenmodes.

[0055] Preferably, the method further includes receiving information representing the mechanical structure, wherein obtaining the sixth data item corresponds to determining a set of radiation coefficients from the structure.

[0056] The applicant proposed that the set of radiation coefficients could be determined analytically during the execution of the method, for example, by modeling the stator as an equivalent cylinder.

[0057] Those skilled in the art will also understand that the set of radiation coefficients can be calculated from the modal deformation of each intrinsic mode by acoustic finite element method.

[0058] The set of radiation coefficients can be received directly in parallel with the set of modes, or determined based on the same information representation structure as the intrinsic modes, for example, in parallel with the determination of intrinsic modes, or at a later time, once the intrinsic modes have been reduced to a subset.

[0059] In one embodiment, the relevant node set includes a first subset of nodes that apply magnetic force to the rotor and stator, a second subset of nodes that generate acoustic noise radiation, and a third subset of isolation interface nodes.

[0060] The applicant proposes that these three subsets correspond to important nodes in the calculation, and the relevant node set is determined, for example, by identifying and combining these three subsets. It is further understood that the relevant node set can be determined in other ways, based on the requirements and knowledge of those skilled in the art, to consider nodes whose behavior is important for the vibration calculation.

[0061] In one particular embodiment, all frequency response functions include a set of response functions for each load case and optionally a subset of load cases and for each eigenmode of significant eigenmodes, with representative information on machine electromagnetic origin noise and vibration levels determined separately for each load case and optionally for each eigenmode.

[0062] The applicant proposes here that the set of frequency response functions in the modal extension of the method corresponds to, for example, a linear combination of the responses for each load condition and each intrinsic mode, e.g., a finite sum. This design thus makes it possible to compute the response for each load condition or the response for each intrinsic mode separately in order to specifically determine and identify the noise and vibration levels generated by each load condition and optionally each intrinsic mode, and then determine the total noise level generated by their combination.

[0063] This provides a more refined analysis of the noise origins. Users can then identify the load conditions and / or intrinsic modes that generate vibrations at each operating point and adjust the design of the electric machinery accordingly to specifically attenuate them. Thus, users understand which type of force excites which type of mode and can act on excitation, such as by adjusting the design of the machine's magnetic circuitry, and on the structure, such as by shifting specific natural frequencies, to reduce noise and vibration.

[0064] In an additional embodiment, the set of frequency response functions includes a set of audio frequency response functions.

[0065] In other words, the frequency response function corresponds to the sound pressure response under a given load (e.g., a unit load). The frequency response function is then calculated in the acoustic domain.

[0066] In another embodiment, the set of frequency response functions includes a set of vibration frequency response functions.

[0067] In other words, the frequency response function corresponds to the vibration response under a given load.

[0068] Of course, it is further understood that the representative information of the machine's electromagnetic noise and vibration levels is determined based on the determined frequency response function. In the context of the acoustic frequency response function, the representative information corresponds, for example, to the sound pressure field under working excitation. The acoustic frequency response function is determined, for example, on an acoustic grid, again limited to the set of nodes that a person skilled in the art considers relevant (e.g., points around the machine used to calculate sound power according to current standards). Conversely, in the context of the vibration frequency response function, the representative information corresponds, for example, to the mean square vibration associated with at least a portion of the machine.

[0069] It is also possible to determine a set of frequency response functions that includes both the set of acoustic frequency response functions and the set of vibration frequency response functions. This design corresponds, for example, to a combined mechano-acoustic calculation, or, by combining with the above variants, to an acoustic calculation separate from the radiation calculation, where the radiation and vibration frequency response functions are combined to obtain the set of acoustic frequency response functions.

[0070] In an additional embodiment, the method further includes receiving information representing a second set of working points for the machine, wherein obtaining the first set of working points includes processing the second set of working points.

[0071] Here it is understood that the second set of operating points corresponds to a set of points provided for performing the method, and the generation of the first set of operating points enables the completion and / or reduction of the number and properties of operating points, for example, to correspond the operating points to the resonant velocities determined above or to reduce the number of operating points and related calculations. The determination of additional operating points is achieved, for example, by interpolation or extrapolation from the second set of operating points.

[0072] In one embodiment that can be combined with the foregoing embodiments, the magnetic state set is obtained from a first set of operating points.

[0073] Here it is understood that the magnetic state corresponds to the magnetic field in the machine associated with the operating point. The machine's magnetic state can be calculated in various ways, such as through magnetic finite element methods, magnetic permeability or magnetomotive force methods, magnetoresistive networks, or equivalent circuits. The magnetic state then allows the associated operating load to be determined based on operating points such as Maxwell's tensor or virtual work methods.

[0074] In a particular embodiment, the method includes receiving information representing the magnetic state, such as directly receiving fifth data, or receiving information on the magnetic flux distribution in the air gap, from which the method determines the magnetic dipoles on the stator teeth using the Maxwell tensor method.

[0075] Within certain operating ranges, the amplitude of the magnetic load can be a known constant or varying value. Therefore, the determination of the set of magnetic states and the operating point can include extrapolation, a method that determines the machine's magnetic state at a single operating point within the range and extrapolates the amplitude and frequency of the excitation harmonics to the operating range, or interpolation between two known operating points. Of course, the operating point used for interpolation is chosen to minimize the error committed, i.e., the inaccuracy during interpolation. It is further understood that using interpolation and / or extrapolation within a defined range also allows for minimizing the computational load, and the identification of such ranges ensures that the accuracy of the results is not affected. For example, the amplitude of a known magnetic load can remain constant within the operating range of an asynchronous motor with constant flux, an open-circuit magnet machine, or a constant current angle.

[0076] According to another variant embodiment, the first set of operating points and the set of magnetic states associated with the operating points are directly imported, and the magnetic states are determined, for example, by magnetic finite element method outside of method execution.

[0077] In yet another embodiment, the method further includes receiving information on target selection belonging to a target set, which includes: Root mean square vibration of the machine surface; The root mean square vibration of at least one node of the machine; and The representative information of the acoustic power radiated by part or all of the machine, the electromagnetic noise origin of the machine, and the vibration level is also determined based on this selection.

[0078] It is understood here that various types of information, each representing electromagnetic noise and vibration levels, can be obtained through the method of execution according to the invention, and their use and usefulness may vary depending on the application. In particular, a user may seek results for a confined part of a machine, derived from the vibrational behavior of the entire machine, such as a given surface or point. This design therefore enables precise determination of which noise and vibration level to be determined during execution, particularly displaying this information according to the variations described above.

[0079] According to a second aspect, the present invention relates to an apparatus for determining the electromagnetic origin noise and vibration levels of an electric machine, the apparatus including a memory associated with a processor configured to implement the steps of the method of the first aspect of the invention.

[0080] According to a third aspect, the present invention relates to a computer program comprising instructions for performing the method of the first aspect of the invention, particularly when such instructions are executed by at least one processor.

[0081] According to a fourth aspect, the present invention relates to a computer-readable storage medium having a computer program recorded thereon, the computer program including instructions for carrying out the method steps of the first aspect of the invention. In one aspect, the recording medium can be any entity or device capable of storing a program. For example, the medium can include a storage device, such as a ROM memory, a CD-ROM, or a ROM memory of the microelectronic circuit type, or a magnetic recording device or a hard disk.

[0082] Furthermore, the recording medium can also be a transmissible medium, such as an electrical or optical signal, which can be transmitted via electrical or optical cables, via conventional radio, or via the air, or via a self-guided laser beam or other means. The computer program according to the invention can be downloaded specifically from a network of the Internet type.

[0083] Alternatively, the recording medium may be an integrated circuit in which a computer program is embedded, which is adapted to perform or be used to perform the method.

[0084] According to a fifth aspect, the present invention relates to the use of the method of the first aspect of the invention in the design of electric machines. It is understood here that the design of electric machines involves producing multiple prototypes or models of the machine, in which it is advantageous to determine the magnetic behavior (e.g., torque, losses) and vibroacoustic behavior as early as possible in order to identify and correct any defects, particularly by adjusting the magnetic excitation and / or physical structure, to minimize investment and manufacturing costs.

[0085] Therefore, the method according to the invention enables rapid execution, thereby enabling the initiation of sensitivity and optimization studies, i.e., the method is integrated into the design phase of the electric machine, and carried out on an industrial scale.

[0086] Therefore, through the aforementioned different functional and structural technical features, the applicant proposes a method and apparatus for determining the electromagnetic origin noise and vibration levels of electric machines, which enables a significant reduction in computation time and can be integrated into the industrial design and / or prototyping methods of electric machines. Attached Figure Description

[0087] See the following reference appendix Figures 1 to 6 Other features and advantages of the invention will become apparent from the description of specific, non-limiting exemplary embodiments thereof, wherein: Figure 1 The structure and mechanical grid of an electric machine according to a specific non-limiting example embodiment of the present invention are shown; Figure 2This illustration schematically shows a configuration for determining according to a specific non-limiting example embodiment of the present invention. Figure 1 The electromagnetic sources of the machine, noise levels, and vibration levels of the equipment; Figure 3 A specific non-limiting example embodiment of the present invention is shown for determining Figure 1 A flowchart of the steps involved in methods for determining the electromagnetic origins of noise and vibration levels in a machine. Figure 4 It shows the results based on Figure 3 The method determined Figure 1 The first figure shows the electromagnetic origins of the machine, noise levels, and vibration levels. Figure 5 It shows the results based on Figure 3 The method determined Figure 1 The second figure shows the electromagnetic noise and vibration levels of the machine. Figure 6 It shows the results based on Figure 3 The method determined Figure 1 The third figure shows the electromagnetic origin of the machine, noise, and vibration levels. Detailed Implementation

[0088] Now refer to Figures 1 to 6 This describes methods and apparatus for determining the electromagnetic origin noise and vibration levels of electric machinery. Throughout the following description, the same elements are identified by the same reference numerals.

[0089] As described in the introduction, current solutions for determining electromagnetic noise and vibration levels correspond to general-purpose software, which is not suitable for this specific task. Their computation time is quite long and unsuitable for industrial needs.

[0090] One of the objectives of this invention is to provide a method for determining electromagnetic noise and vibration levels that is adaptable to integration into the design of electric machinery, particularly in sensitivity and optimization studies.

[0091] This is made possible in the example described below, where the method according to the invention is used in the design phase for electric machines.

[0092] It is understood here that this example is non-limiting, and the method according to the invention can be integrated into various different stages involving vibration analysis of electric machinery.

[0093] according to Figure 1For example, machine 1 includes an electric motor comprising a rotor 11 and a stator 12 separated by an air gap 13. Machine 1 corresponds, for example, to an electric machine model in the design phase, or a physical prototype whose vibration behavior is to be studied in detail, for example, to more accurately identify its vibration sources. Machine 1 corresponds, for example, to a synchronous machine, an asynchronous machine, or a variable reluctance machine. Of course, according to other examples, machine 1 may include multiple rotors 11 and / or multiple stators 12.

[0094] A rotating electric machine typically consists of two main parts: a rotor 11 and a stator 12, connected by bearings. An air gap 13 defines the space between the rotor 11 and the stator 12, in which electromechanical conversion takes place. The electric machine may contain multiple rotors and multiple stators; this does not affect the method if only the number of magnetic loads is considered. The stator 12 is a stationary component made of ferromagnetic material and contains slots called stator slots 12a, 12b, and 12c, which are filled with windings connected to a power source capable of supplying power to the windings of the stator 12 to generate so-called stator currents. These stator currents generate a rotating magnetic field in the air gap 13 of the machine 1.

[0095] The rotor 11 itself is a slender component made of ferromagnetic material, capable of rotating about its own longitudinal axis, for example, about shaft 15.

[0096] according to Figure 1 For example, machine 1 corresponds to a synchronous machine with magnet 14, and stator slots 12a, 12b, 12c include a first set of slots 12a associated with a first phase, a second set of slots 12b associated with a second phase, and a third set of slots 12c associated with a third phase.

[0097] In another example, machine 1 corresponds to an asynchronous machine. In such an example, rotor 11 contains so-called rotor slots filled with coils through which current flows, called rotor current, which is induced by the magnetic field fluctuations generated by stator 12.

[0098] During the operation of machine 1, i.e., during the operation of the electric motor, rotor 11 begins to rotate in an attempt to follow the magnetic field generated by stator 12 according to Lenz-Faraday's law. In asynchronous machines, the mechanical rotational speed of rotor 11 is slightly lower than the rotational speed of the stator field, while in synchronous machines, rotor 11 rotates synchronously with the stator magnetic field.

[0099] The electromagnetic force that generates torque in the air gap 13 is also the cause of vibration and acoustic noise: the generated force deforms the structure of the rotor 11 and stator 12 and propagates more widely to the entire physical structure of the machine 1. The generated vibration can propagate to the surrounding air and radiate at audible frequencies, representing a source of discomfort, and can also cause mechanical fatigue, which may damage applications that must ensure the long-term operation of the machine, such as hydroelectric or wind turbines.

[0100] These electromagnetic forces mainly originate from the interaction between different magnetic flux harmonic groups (Maxwell forces); the magnetic flux harmonics themselves originate from the interaction between different so-called magnetic permeability and magnetomotive force harmonics.

[0101] Therefore, identifying and addressing electromagnetic vibration sources, or more simply, determining whether a given machine is likely to vibrate, represents a complex problem, especially in the preliminary design phase. While combinations of software solutions exist capable of estimating the electromagnetic noise of a given machine, these solutions are complex and extremely time-consuming.

[0102] To solve this problem, machine 1 is associated with a computer device configured to implement a method for determining its electromagnetically originating noise and vibration levels, for example... Figure 3 The method. For example... Figure 2 As shown, such computer devices are advantageously integrated, for example, into an electronic device 2, such as a computer (hereinafter referred to as a "calculator"). Computer 2 is configured, for example, to transmit and receive data within a communication network. The components of computer 2, individually or in combination, can be integrated into a single integrated circuit, multiple integrated circuits, and / or discrete components. Computer 2 can be manufactured as electronic circuit or software (or computer) modules, or even a combination of electronic circuits and software modules.

[0103] Computer 2 includes one or more processors configured to execute instructions for performing method steps and / or for executing software embedded in computer 2. The processor may include integrated memory, input / output interfaces, and various circuits known to those skilled in the art. Computer 2 also includes at least one memory, such as volatile and / or non-volatile memory, and / or includes a storage device that may include volatile and / or non-volatile memory, such as EEPROM, ROM, PROM, RAM, DRAM, SRAM, flash memory, magnetic disk, or optical disk.

[0104] Computer code for embedded software that is to be loaded and executed by a processor is stored, for example, in the memory of computer 2.

[0105] According to a variant embodiment, computer 2 is configured to implement the method for determining the level of electromagnetically originating noise and vibration as part of a broader method, such as a method for designing electric machinery, wherein in this example, quantities calculated during the method of the invention, such as acoustic power or per-load condition and optionally per eigenmode noise and vibration levels, are calculated during each iteration of the design method. According to one example, the method of the invention may also be iteratively performed for multiple machines 1 corresponding to different designs to provide a comparison between the different designs. The method for designing electric machinery includes, for example, creating a model of machine 1, with information associated with the model recorded in the memory 20 of computer 2.

[0106] In the first step 31 of the method for determining noise and vibration levels, computer 2 obtains first data representing a set of load conditions associated with machine 1. Each load condition here corresponds to a magnetic load component decomposed according to a mathematical basis, for example, in a Fourier series. Preferably, each load condition corresponds to a unit load and is characterized by the load type, application direction, application structure, and wave number r. For example, a magnetic load condition corresponds to a radial force on a rotor at a spatial frequency r = 0.

[0107] The first data item is received, for example, by a beacon unit 22 of computer 2, which communicates with the human-machine interface 220 of computer 2. Computer 2 forms, for example, a communication network, such as a multiplexed communication network, in which data is transmitted via wireless or wired links. Therefore, computer 2 establishes communication between the human-machine interface 220 (used as a data acquisition peripheral device) and the beacon unit 22 to allow data exchange.

[0108] According to another example consistent with the above variant, the first data is obtained from the memory 20 of computer 2. In both cases, the first data corresponds to the input data, and the set of load conditions has been determined outside the process execution, for example, through analysis.

[0109] According to a variant embodiment, computer 2 receives information representing a set of magnetic loads associated with machine 1, for example via beacon unit 22 or memory 20. The computer's processor 21 then determines a set of load conditions associated with machine 1 from the set of magnetic loads.

[0110] Information representing the set of magnetic loads corresponds to, for example, a set of analytical equations that allow characterization of the load's frequency and wavenumber, for example, based on the type of machine 1 and / or the parameters of these analytical equations. Such analytical equations are also established based on the discrete design parameters of machine 1.

[0111] According to an exemplary embodiment, the open-circuit permanent magnet machine is associated with the following analytical equation:

[0112] Where r is the wave number of the load, f is the frequency of the load, and k is the frequency of the load. s and h r The two relative integers f are the magnetic flux Fourier series decompositions. e It is the electrical frequency Z, which is proportional to the operating speed. s This represents the number of slots in stator 12, and p is the number of pole pairs in the machine. Modulus [Z] s / 2] This relates to the wavenumber of the force seen at the stator teeth due to spatial sampling phenomena.

[0113] For example, in the context of machine 1: Formula 3 ([Math 3]) p = 4 and [Math 4] Z s = 48 Machine 1 includes wavenumber excitation: [Math 5] r = 0 = 48 - 6 * 8 and related frequencies: [Math 6] f = -12 f e Similarly [Math 7] h r = 6 This type of excitation occurs in both the stator 12 and the rotor 11 in the radial and circumferential directions.

[0114] For other types of machines 1, especially synchronous machines, asynchronous machines, and variable reluctance machines, different analysis formulas are used. These analysis formulas can also be based on the number of phases q of machine 1. s Establishment. The analysis formula can also be modified to account for defective cases, such as adding eccentricity for new load conditions.

[0115] The higher the wave number of the excitation, the less vibration occurs. Therefore, these analytical formulas also allow for limiting the scope of the study by both wave number and maximum frequency, thereby shortening the computation time. Prior knowledge of the frequency also enables optimization of the time sampling in electromagnetic simulations.

[0116] Therefore, according to one example, the information representing the magnetic load set includes machine type information and multiple design parameters representing machine 1, specifically parameter Z. s p and q s .

[0117] In the second step 32, computer 2 obtains second data representing the set of intrinsic modes of the mechanical structure of machine 1.

[0118] As described in the alternative embodiments above, according to one example, the second data is directly received from the reference unit 22 or the memory 20 of the computer 2. In this case, the second data is determined, for example, by mechanical finite element analysis or by experimental modal analysis using a vibrating hammer measurement on the physical machine 1.

[0119] According to another example, computer 2 receives information representing a mechanical structure, such as a model of machine 1 or a set of parameters representing its structure, such as rotor 11 and / or stator 12, from which processor 21 determines a set of eigenmodes. Processor 21 determines the eigenmodes, for example, based on parameters representing stator 12, by modeling it as an equivalent cylinder or more precisely by modeling the teeth and yoke of stator 12 as beam elements.

[0120] In these exemplary embodiments, the intrinsic mode set of the mechanical structure corresponds to a series of modes characterized by modal deformation, natural frequency, and modal damping, respectively.

[0121] In the third step 33, the computer 2, such as processor 21, determines a subset of load conditions and significant eigenmodes. In other words, processor 21 determines which load conditions and which eigenmodes are most likely to cause vibration. Since vibration is generated by resonance between load conditions and eigenmodes, it is understood here that subsets of load conditions and eigenmodes correspond to pairs of load conditions whose shapes are close to or opposite to the eigenmodes.

[0122] Depending on the variant, processor 21 determines the set of projections between the load conditions and the eigenmodes. Each load condition and each eigenmode is represented, for example, in vector form, and processor 21 computes a scalar product between the two vectors corresponding to the projections for each combination of load conditions and eigenmodes.

[0123] For example, the determination of load conditions and subsets of salient intrinsic modes stems from comparing the relevant projection of each combination with a threshold, such as a threshold recorded in memory 20 of computer 2 or a threshold corresponding to the input data received by beacon unit 22. Thresholds are determined, for example, to establish a level of computational precision; higher thresholds can limit the number of load conditions and salient intrinsic modes, thereby limiting the computation performed.

[0124] Optionally, computer 2 also determines the resonant velocity from the load conditions and a subset of significant eigenmodes. The resonant velocity corresponds, for example, to the aforementioned electrical frequency f. e Since each eigenmode is characterized by a natural frequency f0, and machine 1 is advantageously characterized by its frequency f as its electrical frequency f0. e The function is characterized by analytical equations, and therefore the resonant velocity, i.e., the electrical frequency f related to the combination of eigenmodes and load conditions, can be determined, for example, via the aforementioned analytical equations. e .

[0125] Based on the example above, processor 21, for instance, calculates the excitation projection at wavenumber r = 0 (corresponding to the most important excitation) to identify the mode whose shape is closest to that force. For example, the projection between the mode and the unit force at radial r = 0 allows for the identification of specific eigenmodes, referred to as stator breathing modes. Based on the analytical equations above, we then obtain:

[0126] Furthermore, the resonant velocity, i.e., the electrical frequency f, is related to the combination of intrinsic modes and load conditions. e It is determined by the natural frequency f0 of the breathing mode of stator 12.

[0127] In step 34, computer 2 obtains third data representing a set of relevant nodes associated with the mechanical structure. From this set of relevant nodes, computer 2 uses modal expansion to calculate a set of frequency response functions in step 35. Modal expansion, as is known to those skilled in the art, is specifically limited to relevant nodes, load conditions, and a subset of significant intrinsic modes, thereby reducing computation time and memory consumption—major sources of computational time.

[0128] For example, the general modal expansion formula is given as follows:

[0129] Where x is the complex harmonic displacement restricted to the set of nodes of interest I, ω m The natural frequency (rad / s) of mode m, the modal damping of mode m, ξ m is the modal shape of mode m in the modal subset J, F is the magnetic excitation force (N), and co is the excitation frequency (rad / s).

[0130] In other words, the computer determines a set of nodes for machine 1, and the modal extension computation will be limited to these nodes. The modal extension is also limited to the aforementioned determined subset in order to specifically study responses that can produce resonance.

[0131] According to a favorable variant, the relevant node set includes a first subset of nodes that apply magnetic forces to the rotor 11 and stator 12, a second subset of nodes that generate acoustic noise radiation, and a third subset of isolation interface nodes. In other words, the subsets correspond to different criteria for identifying relevant nodes to ensure correct calculation while limiting the total number of nodes. According to another example, the relevant nodes correspond to the nodes located at the tooth tip and machine stationary point in the structure of machine 1.

[0132] This is understood to mean that each frequency response function corresponds to the behavior of a node as a function of the motor frequency, generated by the interaction of load conditions and intrinsic modes. The response can correspond to the sound pressure response (in the context of the sound frequency response function) or the vibration response (i.e., the deformation of the node, in the context of the vibration frequency response function).

[0133] For example, in a finite element mechanical model, the outer envelope S of machine 1 is discretized into N. elem The index e is contained in the interval [1, N]. elem The element mesh, with basic surface dS e Normal vector ne And point C e Centered on the element, the normal deformation U n Modal extension is given by the following formula:

[0134] Where N mode It is the total number of modes considered, ψ m (f) is the modal projection factor of mode m, defined as:

[0135] Where F k (f) Load conditions associated with the projection, H m It is a magnification factor, such that:

[0136] The applicant specifically pointed out that directly calculating the frequency response function through modal extension based on the above-mentioned general modal extension formula is beneficial for shortening the calculation time when dealing with a large number of loads.

[0137] According to specific variant embodiments, a frequency response function is determined separately for each load condition and / or each intrinsic mode. Those skilled in the art will understand here that the response of a node to the set of load conditions and the set of intrinsic modes corresponds to a linear combination of the individual responses of each pair of load conditions and intrinsic modes. Therefore, the frequency response function can be calculated individually for each load condition relative to all intrinsic modes, or calculated separately for each load condition and intrinsic mode combination. In particular, the frequency response function can be determined only for a unique subset of load conditions and intrinsic mode combinations determined above, whose pre-selection corresponds to the combination that produces vibration.

[0138] Therefore, to illustrate, for example, the deformation U of the node n It can be calculated separately for each independent excitation, in particular using the following formula:

[0139] Among them WFRF i The frequency response function corresponding to load condition i, where i is contained in the interval [1, N] lc ], N lc For the number of load cases, F i Let be the complex amplitude under load condition i.

[0140] Similarly, the frequency response function can be separated according to its intrinsic modes using the following formula:

[0141] In particular, the frequency response function (also known as FRF) makes it possible to determine the response proportional to a given amplitude load (e.g., a unit load).

[0142] In step 36, computer 2 obtains fourth data representing the first set of operating points of machine 1. Each operating point corresponds to the torque and speed of machine 1, i.e., the speed at which the engine can operate.

[0143] As described in the above variant, the fourth data is received, for example, directly from the homing unit 22 or the memory 20, or determined by the processor 21. The computer 2 receives, for example, a second set of operating points, from which the processor 21 determines the first set of operating points.

[0144] This means that determining the first set of working points corresponds to establishing a sufficient number of working points to determine the behavior of machine 1. The number and nature of the working points naturally depend on operating standards known to those skilled in the art. For example, a first set containing 200 points can be constructed.

[0145] According to a favorable variant, the first set of operating points is associated with the resonant velocity determined above to ensure that calculations are performed at the velocity indicating resonance, based on the determined load conditions and the subset of significant eigenmodes. It is understood here that, in this case, the acquisition of the fourth data item occurs after the determination of the load conditions and the subset of significant eigenmodes. In other cases, the acquisition of the fourth data item can be performed at a different time, for example, in parallel or upstream of the aforementioned operations.

[0146] Based on the above example, the first set of operating points includes, for example, the electrical frequency f generated by the breathing mode of stator 12 and the velocity. e One or more corresponding points are associated with one or more torque levels. In step 37, computer 2 obtains fifth data representing the set of magnetic states of machine 1. Specifically, computer 2 obtains the distribution and amplitude of magnetic forces in machine 1.

[0147] Furthermore, the acquisition of the fifth data varies depending on the design and the type of input data. According to several examples, the computer determines the electromagnetic state of the machine using magnetic finite element methods, magnetomotive force / permeability methods, or magnetoresistive network methods, which can be coupled with equivalent circuits. It is understood here that those skilled in the art can choose one or another of these methods according to different criteria, particularly based on various trade-offs between computation time and accuracy. According to yet another variation, computer 2 directly receives the fifth data, particularly in parallel with the first set of operating points.

[0148] According to one example, each magnetic state includes a magnetic dipole for each stator tooth, and each dipole includes radial force, circumferential force, and torque.

[0149] Furthermore, each magnetic state in the magnetic state set is associated with an operating point in the first set of operating points. Of course, obtaining the magnetic state set also depends on obtaining the first set of operating points.

[0150] In particular, when processor 21 determines a set of magnetic states associated with a large number of operating points, computation can be reduced by interpolating or extrapolating forces over a specific operating range (e.g., a fixed torque or speed, or even the same torque / speed ratio). Depending on the operating range under consideration, the load amplitude can be constant, in which case processor 21 determines the magnetic states associated with the operating point and extrapolates them to the operating range. For example, such extrapolation can be applied to an asynchronous machine with constant magnetic flux, an open-circuit magnet machine, or a machine 1 with a constant current angle.

[0151] Depending on other operating ranges, processor 21 can interpolate the force between two operating points, selecting the operating point to minimize the error committed. For example, the change in the current vector between two consecutive operating points is finite.

[0152] In step 38, computer 2 then determines the set of working loads from the set of magnetic states.

[0153] In particular, as with load cases, the working load corresponds to the decomposition of the magnetic force in order to determine the magnitude (e.g., in N) of each load case for a given operating point.

[0154] Therefore, computer 2 derives the set of working loads from the fifth data, for example, using the Maxwell tensor method. According to another variant, the fifth data includes the magnetic field distribution on the magnetic grid of machine 1; computer 2 applies the virtual work method and integrates the force dipole tooth-by-tooth to derive the set of working loads.

[0155] Therefore, the workload enables the determination of the amplitude of the aforementioned determined load condition, such as the complex amplitude Fi(f). Finally, in step 39, the computer determines information representing the electromagnetic noise and vibration level of machine 1 based on the workload set and the frequency response function set.

[0156] This means that the determination corresponds to a mapping of the frequency response function, that is, the vibration or pressure response under a given load, mapped to the actual intensity of the load on machine 1 at each operating point.

[0157] It is further understood that information can be determined in more or less detail according to the steps of the method. In particular, for example, information can be determined separately for each load case and / or each intrinsic mode, and the total noise level corresponds to a linear combination of the determined set of information.

[0158] For example, representative information about electromagnetic noise and vibration levels corresponds to the root mean square velocity v determined by the so-called electromagnetic vibration synthesis formula.2 rms :

[0159] Where C i,j The cross-projection coefficient between the load cases corresponding to indices i and j is given by the following formula:

[0160] Where φ m , l The modal shape projections corresponding to modes m and l are, for example:

[0161] In particular, it can be seen that the cross-projection coefficient C i,j With modal shape projection φ m , l Similarly, it is independent of the operating load. According to an advantageous variant embodiment, computer 2 thus determines a set of cross-projection coefficients between load conditions and calculates information representing noise and vibration levels, such as root-mean-square velocity, based on this set of cross-projection coefficients. In other words, the set of cross-projection coefficients is recorded in the computer's memory 20 to simplify calculations during the matching of the frequency response function with the operating load.

[0162] Furthermore, according to the following identity:

[0163] Alternatively, only half of the cross-projection coefficients can be determined. According to another example, electromagnetic noise and vibration levels are determined in the acoustic domain, similar to the frequency response function. Computer 2, for example, meshes machine 1 according to the ISO 3744 standard. For any point M in the ISO 3744 mesh, the sound pressure p is given by the following formula:

[0164] Among them WFRF ap is the frequency response function in the acoustic domain, and OLC is the working load.

[0165] In the same example, the acoustic power W of machine 1 is given by the following formula:

[0166] Where W0 equals 10 -12 W, P0 equals 20 μPa, S ISO3744 N represents the total area of ​​the grid according to the ISO 3744 standard. ISO3744 This represents the total number of nodes according to the ISO 3744 standard.

[0167] It is further understood that the representative information sought regarding the electromagnetic noise and vibration levels, and therefore the calculation methods used, may vary depending on the objectives of those skilled in the art. Therefore, according to a particular variant embodiment, computer 2 also receives information representing the selection of targets belonging to the target set, for example, from beacon unit 22 communicating with human-machine interface 220.

[0168] This set of objectives includes, for example: Root mean square vibration of the machine surface; The root mean square vibration of at least one node of the machine; and The sound power radiated by part or all of the machine.

[0169] Then, representative information about the machine's electromagnetic noise and vibration levels is determined based on the choices made. Of course, it's understood that other elements may depend on this choice, such as calculating the frequency response function in the acoustic or vibration domain.

[0170] According to yet another variant, computer 2 obtains a sixth set of data representing the set of radiation coefficients, each radiation coefficient associated with an eigenmode in the set of eigenmodes. Computer 2, for example, directly receives the set of eigenmodes and the associated set of radiation coefficients, the radiation coefficients being determined outside of the method execution, for example, by acoustic finite element analysis from modal deformation of each mode. In another example, the set of radiation coefficients is determined, for example, based on mechanical structural information of machine 1 for each eigenmode under load conditions and a significant subset of eigenmodes, and then the radiation coefficients are determined by analysis, for example, by modeling stator 12 as an equivalent cylinder.

[0171] For example, the radiation coefficient c of the eigenmode m of a cylinder of finite length m Given by the following formula:

[0172] Where k is the wave number, R c L is the outer radius of the cylinder. st k is the length of the cylinder z k is the longitudinal wave number. r denoted as radial wavenumber.

[0173] In this variant, representative information about the machine's electromagnetic origin noise and vibration levels is also determined based on a set of radiation coefficients associated with the intrinsic modes. The applicant argues that using radiation coefficients particularly ensures more accurate results in low-frequency calculations.

[0174] With the radiation coefficient applied, the formula for calculating the vibration frequency response can be artificially modified as follows:

[0175] This allows the electromagnetic vibration synthesis formula to be retained, and the influence of the modal radiation factor to be considered at the acoustic power level.

[0176] Therefore, step 39 enables the acquisition of a set of noise levels as a function of the operating point of machine 1, at least allowing the identification of the torque and speed combinations that generate noise.

[0177] Optionally, the computer then continues to display information representing the noise and vibration levels via a human-machine interface that communicates with the computer. The computer 2 communicates with the human-machine interface 220, for example, via a beacon unit 22 or another dedicated unit, to allow the display of graphical content representing the determined noise and vibration levels.

[0178] Graphic content, for example, corresponds to Figures 4 to 6 One or more diagrams are shown.

[0179] therefore, Figure 4 First Figure 4 This corresponds to a spectrum diagram showing the evolution of sound power 43 on a curve defined by frequency 41 (Hz) and rotational speed 42 (rpm) of machine 1. The noisiest excitations, such as excitations H24 and H48, corresponding to the ratio between frequency 41 and rotational speed 42, can then be deduced.

[0180] Figure 5 The second Figure 5 The contribution 52 (%) of the overall noise level of machine 1 under different load conditions at a rotational speed 51 (rpm) is shown. This second Figure 5 For example, it enables the determination of which forces are the source of noise at a given speed, such as the circumferential force with a wave number r = 0 on the stator at low speeds and the radial force with a wave number r = 0 on the stator at high speeds.

[0181] Figure 6 The third Figure 6 Another spectrum diagram is shown, illustrating the acoustic power level 63 at a given velocity as a function of the intrinsic mode order 61 and the load condition 62. This third... Figure 6 Then it is possible to separate the contribution of the load condition from the contribution of the electrical noise mode.

[0182] According to other designs, the contributions of different radiating surfaces of machine 1 can also be separated, for example, to distinguish the radiation of the flange and cylinder head of machine 1.

[0183] Therefore, this diversity and the precision delivered to the user enable the understanding of which type of force excites which type of mode, in order to reduce vibration by selectively applying excitation and structure to machine 1, for example, by designing different magnetic circuits, machine controls, or by moving specific natural frequencies of machine 1. It will be understood that this contribution allows for parallel work in the electrical and mechanical domains and assists design teams in both areas.

[0184] According to one variant, computer 2 displays another representative piece of information, such as a supplement to or alternative to information representing noise and vibration levels. For example, the computer displays information corresponding to... Figure 7 and 8 The graphic content shown.

[0185] Figure 7 The fourth Figure 7 Therefore, the projection set between load condition 72 and intrinsic mode 71 is shown. Fourth Figure 7 Each point thus shows the value of the scalar product 73 of the load condition 72 and the specific mode 71, in kg. 0.5 .ms 2 .fourth Figure 7 The points, for example, correspond only to combinations of load conditions and subsets of intrinsic modes, i.e., the projections used that are greater than a threshold. This fourth... Figure 7 Therefore, it is possible to easily visualize the loading conditions and subsets of significant intrinsic modes, as well as the most important combinations within that subset. For example, the threshold used can be adjusted based on this fourth figure to better calibrate the level of detail in the calculations. Advantageously, the fourth... Figure 7 The associated natural frequencies 74 of each intrinsic mode 71 are also shown.

[0186] Figure 8 The fifth Figure 8 The evolution of a defined set of frequency response functions 82, for example in m / N, is shown as a function of frequency 81 (for example in Hz). This fifth Figure 8 Therefore, it is possible to visualize the vibration behavior of machine 1 under a given load (e.g., unit load) without considering the operating point of machine 1.

[0187] Therefore, it will be understood that the present invention provides a method and apparatus for determining the electromagnetic origin noise and vibration levels of electric machinery. This method is particularly capable of accelerating and greatly simplifying calculations performed in this field, especially compared to solutions used in the prior art. This method thus allows electromagnetic vibration analysis to be integrated into the design phase of electric machinery, allowing results to be obtained within the time constraints of this field without sacrificing accuracy.

[0188] It will be understood that this determination method can be applied to a variety of electric machines and integrated into a wider range of electric machine design and / or modeling methods.

[0189] It should be noted that this detailed description relates to specific exemplary embodiments of the invention, but this description is in no way limiting in nature to the subject matter of the invention; rather, its aim is to eliminate any possible inaccuracies or misinterpretations of the following claims.

[0190] Of course, the present invention is not limited to the above-described exemplary embodiments, but extends to methods for determining representative information of electromagnetic origins, including secondary steps, without departing from the scope of the invention. The same applies to systems and / or devices configured to implement this method.

[0191] It should also be noted that the reference marks placed between parentheses in the following claims are not restrictive in any way; their sole purpose is to enhance the comprehensibility and understanding of the following claims and the scope of protection sought.

Claims

1. A method for determining the electromagnetic origin noise and vibration levels of an electric machine (1) having a mechanical structure, the machine (1) comprising a rotor (11) and a stator (12) separated by an air gap, the method being implemented by at least one processor, the method comprising the steps of: - Obtain (31) first data representing a set of load cases associated with the machine (1), each load case corresponding to a magnetic load component decomposed according to a mathematical basis, preferably a Fourier series; - Obtain (32) second data representing the set of intrinsic modes of the mechanical structure; - Determine (33) a subset of load conditions and significant intrinsic modes from the set of load conditions and the set of intrinsic modes; - Obtain (34) third data representing the set of relevant nodes associated with the mechanical structure; - Calculate the set of frequency response functions (35) by limiting the modal extension to the relevant set of nodes and the load conditions and significant intrinsic mode subsets; - Obtain (36) fourth data representing the first set of operating points of the machine (1), each operating point being associated with the torque and speed of the machine (1); - Obtain (37) fifth data representing a set of magnetic states of the machine (1), each magnetic state in the set of magnetic states being associated with an operating point; - Determine the set of working loads (38) from the set of magnetic states; as well as - Determine (39) information representing the electromagnetic origin noise and vibration level of the machine (1) from the set of working loads and the set of frequency response functions.

2. The method according to claim 1 further includes displaying information belonging to an information set via a human-machine interface (220), the information set including: - Information representing the noise and vibration levels; - Information representing the load conditions and a subset of significant intrinsic modes; as well as - Information representing the set of frequency response functions.

3. The method according to claim 1 or 2, further comprising receiving information representing a set of magnetic loads associated with the machine (1), wherein obtaining (31) the first data corresponds to determining a set of load conditions associated with the machine (1) from the set of magnetic loads.

4. The method according to any one of claims 1 to 3, further comprising receiving information representing the mechanical structure, wherein the acquisition (32) of the second data corresponds to determining the intrinsic mode set from the structure.

5. The method according to any one of claims 1 to 4, further comprising the following step: - Determine a set of projections between the load condition and the intrinsic mode, wherein each projection in the set is associated with a load condition and an intrinsic mode; as well as - Each projection of the projection set is compared with a threshold, and the subset of the load condition and significant intrinsic modes is determined (33) as a function of the comparison result.

6. The method according to any one of claims 1 to 5, further comprising determining a set of resonant velocities from the load condition and a subset of significant intrinsic modes, each resonant velocity being associated with an intrinsic mode of the subset, and the first set of operating points being associated with the set of resonant velocities.

7. The method according to any one of claims 1 to 6, further comprising determining a set of cross-projection coefficients between the load conditions, each cross-projection coefficient being associated with a combination of two load conditions in the set of load conditions, wherein information representing the electromagnetic origin noise and vibration level is determined (39) as a function of the set of cross-projection coefficients.

8. The method according to any one of claims 1 to 7, further comprising obtaining sixth data, the sixth data representing a set of radiation factors, each radiation factor being associated with an eigenmode of the set of eigenmodes, and information representing the electromagnetic origin noise and vibration level being determined (39) as a function of the sixth data.

9. The method of claim 8, further comprising receiving information representing the mechanical structure, wherein obtaining the sixth data corresponds to determining the set of radiation factors from the structure.

10. The method according to any one of claims 1 to 9, characterized in that, The relevant node set includes a first node subset on which magnetic loads are applied to the rotor (11) and the stator (12), a second node subset that generates acoustic noise radiation, and a third isolation interface node subset.

11. The method according to any one of claims 1 to 10, characterized in that, The set of frequency response functions includes a set of response functions for each load condition and optionally for each intrinsic mode of the load condition and a subset of significant intrinsic modes, and information representing the electromagnetic origin noise and vibration level of the machine (1) is determined (39) for each load condition and optionally for each intrinsic mode.

12. The method according to any one of claims 1 to 11, characterized in that, The set of frequency response functions includes the set of audio frequency response functions.

13. The method according to any one of claims 1 to 12, characterized in that, The set of frequency response functions includes a set of vibration frequency response functions.

14. The method according to any one of claims 1 to 13, further comprising receiving information representing a second set of working points of the machine (1), wherein obtaining (36) the first set of working points includes processing the second set of working points.

15. The method according to any one of claims 1 to 14, characterized in that, The set of magnetic states is obtained from the first set of operating points (37).

16. The method according to any one of claims 1 to 15, further comprising receiving information representing targets selected from a set of targets, the set of targets including: - Root mean square vibration of the surface of the machine (1); - Root mean square vibration of at least one node of the machine (1); as well as - The acoustic power radiated by part or all of the machine (1), representing information about the electromagnetic origin noise and vibration level of the machine (1), is determined (39) as a function of the selection.

17. A computer program comprising instructions for performing the method according to any one of the preceding claims when the instructions are executed by a processor.

18. A computer-readable storage medium having a computer program recorded thereon, the computer program including instructions for carrying out the steps of the method according to any one of claims 1 to 16.

19. A device (2) for determining the level of electromagnetic origin noise and vibration of an electric machine, the device (2) comprising a memory (21) associated with at least one processor (20) configured to implement the method according to any one of claims 1 to 16.

20. Use of a method for designing an electric machine, the method being the method according to any one of claims 1 to 16.

Images