A brake mechanical aging detection method based on an SPMSM driving motor

Abstract

The application discloses a kind of brake mechanical aging detection methods based on SPMSM driving motor, comprising the following steps: when surface-mounted permanent magnet synchronous motor SPMSM zero speed, high-frequency voltage of sweep frequency is injected to the q axis of SPMSM, the frequency when the angular velocity sweep frequency component fluctuation amplitude of encoder feedback is maximum, the optimal excitation frequency is corrected by the frequency deviation compensation model modified in advance;In brake gap elimination stage, make SPMSM run at rated speed, inject the high-frequency voltage of q axis with the frequency of to the q axis of SPMSM, inject the high-frequency voltage of d axis with different frequency to the d axis of SPMSM;The d-axis voltage and current are bandpass filtered, the d-axis inductance is identified, and the fundamental amplitude is obtained by Fourier transform on the time sequence of;According to the mechanical aging degree index of brake, the aging degree is output according to preset grading standard.

Description

Technical Field

[0001] This invention relates to the field of vehicle brake-by-wire technology, specifically to a method for detecting the mechanical aging of brakes based on an SPMSM drive motor. Background Technology

[0002] Electro-mechanical braking (EMB) systems are the core actuators of next-generation brake-by-wire systems. They generate braking force directly via electrical signals through a series structure of a drive motor, planetary gearbox, ball screw, and brake caliper. EMB systems offer advantages such as fast response, high control precision, and ease of integration, making them a crucial development direction for braking systems in new energy vehicles and intelligent driving vehicles. Among these components, the surface-mounted permanent magnet synchronous motor (PMSM) serves as the core power component. Its rotor shaft is rigidly connected to the sun gear shaft of the planetary gearbox, allowing the motor rotor shaft to directly sense vibration changes in the mechanical structure. However, during long-term service, EMBs are susceptible to progressive aging and degradation due to factors such as vehicle vibration, braking impact, high and low temperature cycling, dust corrosion, and component wear. Typical manifestations include gear tooth wear, increased bearing clearance, sun gear shaft bending deformation, decreased assembly coaxiality, and ball screw wear. These defects can trigger gearbox vibration, which is transmitted back to the motor rotor shaft through the rigid connection, resulting in delayed braking response, fluctuating braking force, brake judder, and increased noise. In severe cases, this can even lead to brake failure, posing a significant threat to driving safety. Therefore, achieving early, accurate, and online aging detection of brake mechanical structures has significant engineering value for improving system reliability and reducing operation and maintenance costs.

[0003] Currently, for aging detection of brake mechanical structures, the main engineering approach is to use external vibration sensors. This approach involves installing accelerometers on the outer surfaces of key components such as gearboxes and ball screws to collect vibration signals generated during mechanical operation. These signals are then combined with methods such as spectrum analysis, time-domain statistics, or envelope demodulation to extract frequency components or energy indicators related to aging characteristics, thereby determining the health status of the mechanical structure. This type of vibration signal monitoring technology has a relatively mature theoretical foundation and widespread application in the field of rotating machinery fault diagnosis. It can reflect the operating status of mechanical components to a certain extent and provide a reference for maintenance decisions.

[0004] However, while external vibration sensor solutions are theoretically feasible, they exhibit several bottlenecks under automotive operating conditions, failing to meet the actual needs of EMB (Electronic Manufacturing Machine) lifecycle health monitoring. Firstly, sensors require additional mounting on the surfaces of components such as gearboxes and ball screws, making installation inherently cumbersome. Furthermore, the limited internal space of an EMB prevents complete coverage of all critical components. Moreover, the vibration signals collected by the sensors are easily affected by external factors such as vehicle vibration, road surface excitation, and in-vehicle electromagnetic interference, resulting in low detection accuracy and difficulty in identifying early, minor aging defects. Additionally, the sensors themselves are wear-prone components, aging and failing over time, requiring regular replacement and calibration, which undoubtedly increases detection costs and maintenance workload. These solutions can only collect vibration signals from the surface of mechanical structures, failing to deeply reflect the aging conditions of internal components such as sun gear shaft eccentricity and gear wear, resulting in highly limited detection results. Furthermore, these solutions typically rely on a single characteristic to determine aging conditions; once the motor itself experiences faults such as demagnetization or short circuits between winding turns, its vibration or electrical characteristics can easily be confused with the signs of mechanical aging, making effective differentiation impossible. Summary of the Invention

[0005] This invention proposes a brake mechanical aging detection method based on SPMSM drive motor, which solves the problem of how to achieve online detection of the aging state of the brake mechanical structure without additional vibration sensors or changes to the brake mechanical structure.

[0006] To address the aforementioned technical problems, this invention provides a method for detecting the mechanical aging of a brake based on an SPMSM drive motor, wherein the rotor shaft of the surface-mounted permanent magnet synchronous motor (SPMSM) is rigidly connected coaxially to the sun gear shaft of the planetary gearbox of the brake. The method includes the following steps: Step S1: When the SPMSM is at zero speed, inject a high-frequency sweep voltage into the q-axis of the SPMSM. Take the frequency at which the amplitude of the angular velocity sweep frequency component fluctuation is the largest, and correct the frequency to the optimal excitation frequency using a pre-calibrated frequency offset compensation model. ; Step S2: During the gap elimination phase of the brake, the SPMSM is operated at a constant speed at its rated speed, and a frequency of [frequency value missing] is injected into the q-axis of the SPMSM. q-axis high-frequency voltage, injection frequency into the d-axis and Different d-axis high-frequency voltages; Step S3: Perform bandpass filtering on the d-axis voltage and current to identify the d-axis inductance. ,right Perform a Fourier transform on the time series over at least one rotor mechanical cycle to obtain the fundamental amplitude at the rotor mechanical frequency. ; Step S4: According to and Calculate the mechanical aging degree index of the brake and output the aging degree of the brake according to the preset grading standard.

[0007] Preferably, the amplitude of the sweep frequency high-frequency voltage in step S1 is 5% to 10% of the DC bus voltage of the SPMSM, and the frequency of the sweep frequency high-frequency voltage increases in the range of 500Hz to 1500Hz.

[0008] Preferably, the expression for the sweep frequency voltage in step S1 is: ; In the formula, This represents the instantaneous value of the high-frequency sweep voltage; The amplitude of the sweep frequency high-frequency voltage; The sweep frequency is the angular frequency that increases over time. For time; The frequency corresponding to the sweep frequency is higher than the fundamental frequency of the SPMSM, higher than the current loop bandwidth of the SPMSM, and lower than the switching frequency of the inverter driving the SPMSM.

[0009] Preferably, the frequency offset compensation model in step S1 is pre-calibrated in the following manner: the resonant frequency of the brake at zero speed and the resonant frequency of the brake at rated speed are collected respectively, the difference between the resonant frequency at rated speed and the resonant frequency at zero speed is used as the frequency offset compensation amount of the frequency offset compensation model, and the frequency offset compensation amount is written into the controller of the brake.

[0010] Preferably, the frequency of the d-axis high-frequency voltage in step S2 is... The absolute value of the difference is not less than 200Hz and not greater than 500Hz.

[0011] Preferably, the d-axis inductor in step S3 The time series is composed of d-axis inductance sampling segments, which are obtained in the following way: from the first... k The d-axis high-frequency voltage and d-axis high-frequency current within the injection cycle are used to calculate the d-axis inductance in the [number]th injection cycle. k Value per injection cycle In the formula k The sequence number of the injection cycle and k For positive integers, the values ​​obtained from multiple consecutive injection cycles within the same gap elimination phase are considered. By serial number k Arranged from smallest to largest.

[0012] Preferably, the Calculate using the following formula: ; In the formula, The amplitude of the high-frequency voltage along the d-axis; ω is the angular frequency of the high-frequency voltage along the d-axis; For the first k The amplitude of the d-axis high-frequency current obtained after bandpass filtering within each injection cycle.

[0013] Preferably, multiple d-axis inductance sampling segments are obtained in each of the multiple gap elimination stages of the brake, wherein the d-axis inductance The time series is obtained by fusing multiple d-axis inductance sampling segments in the following manner: each d-axis inductance sampling segment is resampled at equal mechanical angle intervals based on the synchronously acquired rotor single-turn absolute mechanical angle, and then connected in the order of the absolute mechanical angle scale of the beginning and end of each d-axis inductance sampling segment. The connected data is then cut into whole mechanical cycle lengths based on the rotor mechanical phase and averaged point by point. The rotor single-turn absolute mechanical angle is synchronously acquired by the encoder, which is a single-turn absolute rotary encoder.

[0014] Preferably, the expression for the mechanical aging degree index is: ; ; ; In the formula, As an indicator of the degree of mechanical aging; The characteristic value of the eccentricity of the brake's mechanical structure; This refers to the stiffness degradation of the brake's mechanical structure. The fundamental amplitude reference value is the one corresponding to the mechanical structure of the brake that has not aged. The optimal excitation frequency reference value is the one corresponding to the mechanical structure of the brake that has not aged. and The proportionality coefficient is determined by the experiment; and The normalized weighting coefficients are determined by experiments and ; This is the stiffness retention coefficient of the brake's mechanical structure; The d-axis inductor fundamental amplitude value relative to the reference value The relative change.

[0015] Preferably, the preset grading standard in step S4 is: Output has no aging when it is no greater than 3%. When the percentage is greater than 3% but not greater than 5%, the output shows slight aging. When the percentage is greater than 5% but not greater than 10%, the output indicates moderate aging. When the output is greater than 10%, the output is severely aged.

[0016] The advantages of this invention include at least the following: 1. Because the sweeping high-frequency voltage injected into the q-axis locks the excitation frequency at the resonance point of the brake mechanical structure, the mechanical structure enters a resonance state at this frequency, and the eccentric vibration of the sun gear shaft receives continuous energy replenishment, which offsets the energy loss caused by mechanical damping, thus making the originally weak eccentric characteristics amplified in a specific direction. 2. At the same time, the fundamental amplitude of the d-axis inductor, which characterizes the degree of eccentricity, and the optimal excitation frequency, which characterizes the structural stiffness, were extracted. These two values ​​correspond to the amplitude dimension and frequency dimension of mechanical structure aging, respectively, and the degree of aging and fault type can be accurately determined. 3. Since the high-frequency signals injected into the q-axis and d-axis are set to different frequencies, and the frequency interval between them is greater than the transition band width of the filter, the electromagnetic coupling component from the q-axis in the acquired d-axis voltage can be completely eliminated by the frequency division bandpass filter. The identification result of the d-axis inductance no longer drifts with the amplitude and phase of the q-axis injected signal. The influence of vehicle electromagnetic interference and road vibration interference on the detection result is significantly suppressed, and the detection result is stable and reliable. 4. Since the optimal excitation frequency is determined only by the equivalent stiffness and equivalent mass of the mechanical structure, and is independent of the magnetic properties of the motor permanent magnet and the electrical parameters of the winding, when the motor body demagnetizes or the winding fails, the amplitude of the d-axis inductance fundamental wave will change while the optimal excitation frequency remains stable; when the mechanical structure ages, both will change simultaneously. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the electromechanical braking system according to an embodiment of the present invention; Figure 2 This is a control block diagram of the closed-loop control and high-frequency injection of the drive motor in an embodiment of the present invention; Figure 3 This is a flowchart of the brake mechanical aging detection method according to an embodiment of the present invention; Figure 4 This is a schematic diagram of the motion trajectory of the motor rotor with mixed eccentricity in an embodiment of the present invention; Figure 5 The waveform diagram of the q-axis sweep frequency high-frequency injection signal and the angular velocity fluctuation of the motor at zero speed in an embodiment of the present invention; Figure 6 This is a comparison waveform diagram of the d-axis high-frequency current component and the d-axis inductance fluctuation component before and after q-axis injection in an embodiment of the present invention. Figure 7 This is a comparison waveform diagram of d-axis inductance fluctuation under different aging degrees in an embodiment of the present invention.

[0018] In the diagram: 1-Rotary encoder; 2-Drive motor; 3-Planetary gearbox; 31-Sun gear shaft; 4-Ball screw; 5-Pressure sensor; 6-Piston; 7-Brake pad; 8-Brake disc; 9-Brake caliper. Detailed Implementation

[0019] The technical solutions in this embodiment 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 protection scope of the present invention.

[0020] Example 1 This invention provides a method for mechanical aging testing of brakes based on SPMSM drive motors, applicable to both the final calibration and testing stages of brake assembly. The testing system implementing this method is as follows: Figure 1 and Figure 2 As shown, the system includes a surface-mounted permanent magnet synchronous motor (SPMSM), a brake mechanical structure, a drive motor control module, a high-frequency voltage generator, an angular velocity fluctuation detection module, a frequency offset compensation model, a signal acquisition module composed of a current sensor and a rotary encoder, a frequency division filtering module, a parameter identification module, a spectrum analysis module, and a dual feature extraction and aging assessment module. The drive motor 2, i.e., the SPMSM, has its rotor shaft rigidly connected coaxially to the sun gear shaft 31 of the planetary gearbox 3. The output of the SPMSM, after being reduced in speed and increased in torque by the planetary gearbox 3, drives the ball screw 4 for linear feed. This feed is then transmitted via piston 6 to push the brake pads 7 within the brake caliper 9 closer to or further away from the brake disc 8, and the braking force is detected by pressure sensor 5. A single-turn absolute rotary encoder 1 is installed on the SPMSM's body. The rotary encoder outputs the single-turn absolute mechanical position of the rotor. The real-time electrical angle of the rotor is calculated by combining the single-turn absolute mechanical position with the number of motor pole pairs, and the real-time angular velocity of the rotor is simultaneously fed back. The complete detection process of this embodiment is as follows: Figure 3 As shown, it includes steps S1 to S4.

[0021] This embodiment embeds all detection actions into the existing working sequence of the brake, without adding independent detection conditions. Specifically, the frequency sweep process in step S1 is executed when the brake has not received a braking command from the upper level. At this time, the brake pads remain at their rated initial position, the ball screw does not perform any feed movement, and the brake does not undertake any braking task; therefore, the frequency sweep process does not occupy the effective braking time of the brake. Steps S2 to S4 are executed during the gap elimination phase. During the gap elimination phase, the brake pads and brake disc have not yet made contact, the system does not generate braking force, the load is stable, and the mechanical structure is in an unloaded excitation state, which is an ideal safe window for carrying out aging detection. During this phase, the drive motor adopts dual-loop control with an outer speed loop and an inner current loop, with the control setpoint set to the rated speed for uniform rotation, providing a stable working condition basis for high-frequency injection and feature extraction.

[0022] Step S1: When the surface-mounted permanent magnet synchronous motor SPMSM is at zero speed, inject a high-frequency sweep voltage into the q-axis of the SPMSM. Take the frequency at which the amplitude of the angular velocity sweep frequency component fluctuation is the largest, as fed back by the encoder, and correct the frequency to the optimal excitation frequency using a pre-calibrated frequency offset compensation model. .

[0023] The purpose of this step is to find the zero-speed resonant frequency that causes the brake's mechanical structure to enter a resonant state without interrupting the brake's operating time, and then convert it into the resonant frequency under rated speed conditions, i.e., the optimal excitation frequency. .

[0024] In this embodiment of the invention, the frequency sweep process is completed at zero motor speed, while the formal test is completed at rated speed. The dynamic characteristics of the mechanical structure are not exactly the same under the two operating conditions. The rotational speed will affect the equivalent stiffness of the system through gyroscopic effect, bearing oil film stiffness, etc., thus causing the resonant frequency of the same mechanical structure to deviate at the two rotational speeds.

[0025] The core essence of brake mechanical aging is the continuous decrease in the structural equivalent stiffness. Regardless of the specific form of aging—whether it manifests as tooth surface wear, increased clearance, fatigue of elastic components, or loosening of connections—its impact on the structural dynamics is consistent: the static natural frequency decreases synchronously, and the operating resonant frequency at rated speed also decreases synchronously. Therefore, rotational speed only introduces a fixed operating condition bias to the resonant frequency and does not alter the monotonic mapping relationship between stiffness degradation and frequency decrease. When judging aging based on the relative change in frequency, the systematic error introduced by rotational speed can be offset.

[0026] Based on this, this embodiment further employs an offline fixed compensation correction method to eliminate the fixed bias. During the brake's offline calibration phase, the resonant frequencies of the same brake at zero speed and at rated speed are collected separately. Using the frequency deviation between the two as a sample, a frequency offset compensation model is established. The resonant frequency at zero speed is collected as in step S1; the resonant frequency at rated speed is collected on the offline calibration bench. The SPMSM is made to run at a constant speed at rated speed and a high-frequency sweep voltage of the same frequency band is injected into the q-axis. The angular velocity fed back by the encoder is extracted with the same frequency as the sweep frequency, and the frequency at which the fluctuation amplitude of the same frequency component is the largest is taken. The frequency offset compensation model is adopted as a fixed bias form, that is, the difference between the resonant frequency at rated speed and the resonant frequency at zero speed is used as the frequency offset compensation amount. After calibration, the compensation amount is written into the bottom parameter area of ​​the brake controller, which can realize one-click calibration under fixed operating conditions. In each subsequent test, the controller directly calls the model to correct the resonant frequency extracted at zero speed in real time, restoring a value close to the natural frequency of the real structure, and eliminating the dynamic deviation caused by the gyro effect and oil film stiffness.

[0027] The given speed of the drive motor is set to zero, and a sweeping high-frequency voltage is injected into the q-axis of the motor. Its expression is: ; In the formula, This represents the instantaneous value of the high-frequency sweep voltage; The amplitude of the sweep frequency high-frequency voltage; The sweep frequency is the angular frequency that increases over time. For time.

[0028] Amplitude of sweep frequency high-frequency voltage The value should be 5% to 10% of the DC bus voltage. The lower limit of this range ensures the injected signal has sufficient excitation energy, while the upper limit ensures the injection process does not excessively encroach on the voltage modulation margin, thus avoiding interference with the normal operation of the motor. Sweep frequency. The corresponding frequency increases at a set rate within the range of 500Hz to 1500Hz. The selection of this frequency band simultaneously satisfies three constraints: the sweep frequency is higher than the fundamental frequency of the SPMSM, higher than the current loop bandwidth of the SPMSM, and lower than the switching frequency of the inverter driving the SPMSM. Specifically, being higher than the fundamental frequency ensures that the high-frequency components are separable from the fundamental components in the frequency domain; being higher than the current loop bandwidth ensures that the current loop does not suppress the injected signal, allowing the high-frequency voltage to directly act on the motor windings; and being lower than the inverter switching frequency ensures that the injected signal can be accurately reproduced by the pulse width modulation stage without aliasing.

[0029] If a braking command is received from the upper level during the frequency sweep, the frequency sweep is immediately stopped and normal braking control is switched to. The data of this round is invalidated, and the frequency sweep is restarted from the starting frequency when the condition of not receiving a braking command is met again.

[0030] In the synchronous rotating dq coordinate system, neglecting core saturation, eddy current losses, and hysteresis losses, the voltage equation for the surface-mounted permanent magnet synchronous motor is: ; ; In the formula, and These are the stator voltages along the d-axis and q-axis, respectively. and These are the stator currents along the d-axis and q-axis, respectively. Stator resistance; and These are the stator inductances along the d-axis and q-axis, respectively. Electric angular velocity; The excitation flux generated by the permanent magnet.

[0031] Since the fundamental component and the high-frequency component do not overlap in the frequency domain, the voltage equation for the q-axis high-frequency component can be extracted separately: ; In the formula, and These are the high-frequency current components along the q-axis and d-axis, respectively. It represents the electric angular velocity.

[0032] The above equation is the general form of the q-axis high-frequency voltage equation, which is simplified as follows under the two operating conditions of this invention: In the zero-speed frequency sweep condition of this step, the given speed of the motor is zero, and the electrical angular velocity is... The value is zero, and at this point, no high-frequency voltage has been injected into the d-axis. Zero, cross-coupling term Since it is always zero, the above equation degenerates into: .

[0033] Under the rated speed detection conditions in steps S2 to S4, the electric angular velocity Non-zero and simultaneously injected with high-frequency voltage along the d-axis, cross-coupling term It cannot be omitted; however, the oscillation frequency of this term is determined by... The determination is equal to the angular frequency of the high-frequency injection along the d-axis. angular frequency of high-frequency injection along the q-axis They are different from each other, therefore the center of the transit band is located at After bandpass filtering, this term is completely filtered out and does not participate in the solution of the q-axis high-frequency current.

[0034] As can be seen from the above, the q-axis high-frequency current component can be obtained from the following formula under both operating conditions: ; Phase difference for: ; At this time, the total voltage along the q-axis Total current along the q-axis They are represented as follows: ; ; In the formula, and These are the q-axis voltage and q-axis current output by the original inner current loop to maintain the control setpoint.

[0035] The electromagnetic torque equation for a surface-mounted permanent magnet synchronous motor is: ; In the formula, The electromagnetic torque of the motor; coefficient The amplitude invariance coefficient for the transformation from a three-phase coordinate system to a two-phase coordinate system; This represents the number of pole pairs of the motor.

[0036] For surface-mounted permanent magnet synchronous motors and Since they are approximately equal and the changes caused by eccentricity are in the same direction, the reluctance torque term is approximately zero. The following formula retains this term to maintain the generality of the expression.

[0037] Substituting the q-axis current containing high-frequency components into the above equation, we can obtain the electromagnetic torque after injecting the high-frequency signal as follows: ; In the formula, This represents the amplitude of the q-axis high-frequency current. ; This represents the hysteresis phase angle of the q-axis high-frequency current relative to the q-axis high-frequency voltage. The electromagnetic torque required to drive the motor to operate normally; It is a high-frequency pulsating torque. The high-frequency pulsating torque drives the rotor shaft to generate high-frequency vibration, which is then transmitted to the sun gear shaft of the gearbox via a coaxial rigid connection, thus exciting the resonance of the entire mechanical structure of the brake.

[0038] During the frequency sweep, the absolute mechanical position of a single revolution fed back by the encoder is acquired in real time at a sampling rate synchronized with the inverter switching frequency. The angular velocity is calculated differentially and a bandpass extraction is performed at the same frequency as the sweep. The envelope of the bandpass output is taken as the angular velocity fluctuation amplitude. At the resonance point, the angular displacement amplitude of the rotor torsional vibration is greater than the angle corresponding to the single position resolution of the encoder, and the fluctuation amplitude can be effectively distinguished at this frequency. As the sweep frequency increases, when the excitation frequency approaches the resonance point of the mechanical structure, the rotor vibration amplitude excited by the same high-frequency pulsating torque increases significantly, and the angular velocity fluctuation amplitude reaches a peak. When the angular velocity fluctuation amplitude reaches its maximum, the current frequency is updated and locked as the zero-speed optimal excitation frequency. This frequency corresponds to the resonance point of the brake's mechanical structure at zero speed, directly reflecting the stiffness and natural frequency changes of components such as gearboxes, bearings, and ball screws.

[0039] Subsequently, based on the pre-calibrated frequency offset compensation model, The resonant frequency converted to the rated speed, i.e., the optimal excitation frequency. It is used in subsequent assessments of the degree of aging. Figure 5 The measured waveforms of the q-axis sweep frequency injection signal and the motor angular velocity fluctuation at zero speed in this embodiment are presented. For example... Figure 5 As shown, the injection frequency can excite the resonant response of the mechanical structure as it increases within a set range. Selecting the frequency corresponding to the resonance point yields the optimal excitation frequency at zero velocity. Actual measurements revealed the optimal excitation frequency obtained after frequency offset compensation under oscillation conditions of mechanical structures with varying degrees of aging. It has a significant offset.

[0040] Step S2: During the brake clearance elimination phase, the SPMSM is operated at a constant speed at its rated speed, and a frequency of [frequency value missing] is injected into the q-axis of the SPMSM. q-axis high-frequency voltage, injection frequency into the d-axis and Different d-axis high-frequency voltages.

[0041] The control brake enters the gap elimination stage, causing the SPMSM to run at a constant speed at the rated speed, after which two high-frequency injections are executed simultaneously.

[0042] The first path is q-axis excitation injection. The injection frequency into the q-axis is the optimal excitation frequency determined in step S1. The q-axis high-frequency voltage has an amplitude that is still 5% to 10% of the DC bus voltage. At this time, the excitation frequency is consistent with the natural frequency of the eccentric vibration of the sun gear shaft, and the mechanical structure enters a resonant state. This amplification effect does not indiscriminately amplify all vibrations, but rather focuses on the periodic vibration caused by the eccentricity of the sun gear shaft, continuously replenishing it with energy to offset mechanical damping losses, so that the eccentric vibration characteristics are stably maintained, providing a strong characteristic signal for high-precision detection.

[0043] The second path is d-axis identification injection. The injection frequency to the d-axis is... The expressions for different d-axis high-frequency voltages are as follows: ; In the formula, This represents the instantaneous value of the high-frequency voltage along the d-axis. The amplitude of the high-frequency voltage along the d-axis; Let be the angular frequency of the d-axis high-frequency voltage. The frequency of the d-axis high-frequency voltage is related to... The absolute value of the difference is not less than 200Hz and not greater than 500Hz. The lower limit of this frequency interval ensures that the two signals have sufficient spacing in the frequency domain so that the passbands of the bandpass filters do not overlap; the upper limit ensures that the d-axis injection frequency does not deviate too far from the range where high-frequency approximation is valid. In this embodiment, the frequency interval is set to 300Hz, while ensuring that the d-axis injection frequency avoids the mechanical natural frequency and the eccentric frequency to prevent signal aliasing.

[0044] From the aforementioned voltage equation, it can be seen that the d-axis voltage Includes This cross-interference term. After a high-frequency signal is injected along the q-axis... The high-frequency components contained therein will affect the d-axis voltage through this cross-interference term. This causes interference components with the same frequency as the high-frequency signal on the q-axis to appear in the acquired d-axis voltage, which will produce significant errors if the d-axis inductance is directly identified using this voltage.

[0045] This invention employs high-frequency signals of different frequencies injected into the q-axis and d-axis, respectively. A bandpass filter is designed based on this frequency difference, with its passband center frequency set at the d-axis injection frequency and its passband width smaller than the frequency interval between the two axes. By performing bandpass filtering on the acquired d-axis voltage and current, the response components corresponding only to the high-frequency d-axis injection signal can be separated, excluding those from the q-axis. Interference, thus laying the foundation for accurate identification of d-axis inductance.

[0046] During long-term service, the mechanical structure of a brake is affected by factors such as braking load impact, vibration, and wear, resulting in a series of aging defects such as gear wear, bearing wear, sun gear shaft deformation, and increased assembly clearance. These defects disrupt the rotational symmetry of the brake's mechanical structure, especially the planetary gearbox, causing the rotational center of the sun gear shaft to no longer coincide with its geometric center during operation, forming an initial eccentricity. This eccentricity is transmitted to the motor rotor shaft end via a rigid coaxial connection in a manner with no phase difference and no time delay. On the one hand, this causes the motor load torque to change periodically with the rotor angle; on the other hand, because the sun gear shaft is rigidly connected to the rotor shaft and the radial offset is within the bearing clearance range, the radial eccentricity of the sun gear shaft is transmitted approximately equally as the radial offset of the rotor core relative to the stator, causing the motor air gap to no longer be uniformly distributed along the circumference.

[0047] Under normal operating conditions without active excitation, the aforementioned initial eccentricity can only generate weak periodic disturbances during motor operation. The signal amplitude is low, the signal-to-noise ratio is poor, and it is easily masked by operating noise, making it difficult to accurately capture and failing to meet the requirements of high-precision detection. However, by actively injecting oscillations through the q-axis, the sun gear shaft can be forced to vibrate periodically around the eccentric center, forming a stable periodic eccentric torque and achieving directional amplification of the eccentric vibration characteristics.

[0048] The mechanical oscillations induced by the high-frequency injection of the q-axis at the motor rotor end generate a periodic radial excitation force on the sun gear shaft of the gearbox via a rigid connection. This excitation force only causes high-frequency radial vibration on the motor shaft and does not actively lead to motor eccentricity. For an unaged brake mechanical structure, its rotational symmetry is good, and the radial excitation force, after transmission, acts uniformly in the circumferential direction of the sun gear shaft. The sun gear shaft also only undergoes high-frequency radial vibration without significant eccentric vibration. However, for an aged brake mechanical structure, the sun gear shaft has initial eccentricity, and defects such as uneven gear meshing clearance and increased bearing clearance will cause uneven distribution of the radial excitation force in the circumferential direction of the sun gear shaft, forming a periodic eccentric torque. This eccentric torque amplifies the original aging pre-fault characteristics, making them appear at a higher magnitude in actual testing, thus distinguishing them from noise interference.

[0049] The aging of the brake's mechanical structure will not present as absolute static eccentricity, nor as absolute dynamic eccentricity; eccentricity faults will always exist in the form of mixed eccentricity. Figure 4 The motion trajectory of the motor rotor under mixed eccentricity is shown in the figure. The length of the air gap. , , These are static eccentricity, dynamic eccentricity, and mixed eccentricity, as labeled in the figure. , , For the corresponding eccentricity distance, The initial eccentricity angle, The minimum air gap position angle, For the stator side spatial mechanical angle, This represents the rotor angle.

[0050] Based on graphical calculations, the mixed eccentricity Mechanical angle at the minimum air gap with rotor eccentricity The expression is as follows: ; .

[0051] It should be noted that in the above formula... The instantaneous mixed eccentricity varies with the rotor angle. Periodic variation is a time-varying variable describing the rotor's geometric position, used to characterize the spatial and temporal distribution of eccentricity; while step S4 is used to calculate the aging degree index. The eccentricity characteristic value is the relative change of the d-axis inductance fundamental amplitude relative to a reference value, expressed as a proportionality coefficient. The converted value is a scalar representing the severity of eccentricity in a single detection. The relationship between the two is: Instantaneous Mixed Eccentricity The amplitude of the fluctuation within a rotor mechanical cycle determines the periodic variation of the air gap non-uniformity, which in turn determines the fundamental amplitude of the d-axis inductance fluctuation. Therefore, the eccentricity eigenvalue With instantaneous mixing eccentricity The fluctuation amplitude within a mechanical cycle is monotonically positively correlated. The larger the value, the more severe the eccentricity of the brake's mechanical structure.

[0052] For an ideal, unbiased surface-mounted permanent magnet synchronous motor, the air gap is uniformly distributed along the circumference, and the air gap length is a constant value. The ideal air gap function is: .

[0053] When the aforementioned mixed eccentricity is transmitted to the motor rotor shaft without delay or phase difference via a coaxial rigid connection, the motor air gap exhibits a cosine-shaped non-uniform distribution: ; In the formula, Eccentricity is defined as the ratio of the distance between the stator center and the rotor center to the air gap length, and its value ranges from [value missing]. , The larger the value, the more severe the eccentricity. This is the minimum air gap position angle, used to determine the eccentricity direction; This represents the motor's rotation angle. After eccentricity, the air gap reaches its minimum point along the circumference. With the maximum air gap point The air gap non-uniformity varies with It increases linearly.

[0054] Therefore, when the motor is in an eccentric fault state, the air gap is no longer evenly distributed along the circumference. Due to this change in space, the eccentric fault will cause the three-phase inductance values ​​of the motor to change. The inductance value increases on the side with a smaller air gap, and the inductance value decreases on the side with a larger air gap.

[0055] The severity of system eccentricity is directly related to the air gap size. Among the available motor parameters, the motor inductance is strongly correlated with the air gap. The expression for the d-axis inductance of a surface-mounted permanent magnet synchronous motor under normal conditions is: ; In the formula, This refers to the size of the air gap in the motor. The thickness of the permanent magnet material; The vacuum permeability; This refers to the number of turns in series per phase of the stator winding; Calculate the area for the armature.

[0056] Based on the above relationship, this embodiment locates the aging characteristic value of the brake's mechanical structure onto the d-axis inductance of the drive motor connected to the same rotating shaft. Since eccentric vibration can be detected on the motor side, It will change periodically with the frequency of eccentric vibration, therefore the identified It is a quantity that changes periodically with time, and its frequency of change coincides with the eccentric vibration frequency, that is, with the mechanical rotational frequency of the rotor. Fourier decomposition can yield the DC component. That is, the average inductance and the fundamental component That is, the amplitude of inductance fluctuation, where Amplitude and instantaneous mixing eccentricity The fluctuation amplitude within a mechanical cycle is strictly positively correlated, which is the core characteristic of eccentric faults.

[0057] Step S3: Perform bandpass filtering on the d-axis voltage and current to identify the d-axis inductance. ,right Perform a Fourier transform on the time series over at least one rotor mechanical cycle to obtain the rotor mechanical frequency. .

[0058] After eliminating q-axis interference, based on the filtered d-axis high-frequency voltage With d-axis high-frequency current Impedance identification is used to identify the d-axis inductance. At higher frequency signals, the stator resistance... The resulting voltage drop is much smaller than that generated by the inductor, and its effect can be ignored. Furthermore, crosstalk has been eliminated by bandpass filtering. Therefore, the relationship between the d-axis high-frequency voltage and the d-axis high-frequency current can be simplified as follows: .

[0059] Will Substituting the expression into the above formula and integrating over time, we obtain the identification formula for the d-axis inductance. This embodiment addresses the issue of the d-axis high-frequency voltage at the [missing information - likely a specific voltage level]. Within each injection cycle, the d-axis inductance is calculated from the d-axis high-frequency voltage and d-axis high-frequency current within that injection cycle. Value per injection cycle The formula for its calculation is: ; In the formula, The amplitude of the high-frequency voltage along the d-axis; ω is the angular frequency of the high-frequency voltage along the d-axis; For the first The time within each injection cycle and take The moment when it equals π / 2; For the first The d-axis high-frequency current obtained after bandpass filtering within each injection cycle; It is the sequence number of the injection cycle and is a positive integer.

[0060] In practical implementation, take within each injection cycle The sampling time is defined as the moment when the high-frequency current along the d-axis reaches its amplitude. The value is 1, and the sampled This refers to the amplitude of the d-axis high-frequency current during the injection cycle. The above formula simplifies to This yields a uniquely determined d-axis inductance value within that injection cycle. Using a fixed-phase sampling method ensures that the extracted values ​​across different injection cycles are consistent. Having the same computational benchmark avoids issues caused by sampling phase drift. Introduce spurious fluctuations into the sequence.

[0061] The results obtained in multiple consecutive injection cycles By serial number Arranged from smallest to largest, they form a d-axis inductance sampling segment. In this embodiment, the d-axis inductance sampling segment collected during a single gap elimination phase is taken as the d-axis inductance. The time series. It should be noted that the eccentricity caused by mechanical aging... The high-frequency current of the d-axis varies periodically with the rotor's mechanical cycle. It is no longer a pure sine wave signal, but a time-varying signal. The modulated signal, therefore, is extracted periodically. Able to restore The complete profile changes over time. The d-axis high-frequency injection frequency is much higher than the mechanical eccentric vibration frequency, therefore a sufficient number of [instruments / items] can be obtained within one cycle of the eccentric vibration. The sampling points satisfy the requirements of the sampling theorem.

[0062] Regarding the above The time series was subjected to a Fast Fourier Transform to decompose it into DC components. With fundamental component The amplitude of the fundamental component is taken as the first characteristic quantity characterizing the degree of eccentricity. Figure 6 The comparison waveforms of the d-axis high-frequency current component and the d-axis inductance fluctuation component extracted before and after q-axis injection in this embodiment are given. For example... Figure 6 As shown, when no high-frequency signal is injected into the q-axis, the envelope of the high-frequency current on the d-axis is flat, and the d-axis inductance fluctuation is weak; when a frequency of [missing information] is injected into the q-axis... After the high-frequency voltage, the d-axis high-frequency current exhibits obvious periodic envelope fluctuations, and the fluctuation amplitude of the d-axis inductance increases significantly. This indicates that the eccentric fluctuation characteristic amplitude at the natural frequency during the aging process of the brake is amplified by the q-axis injection, making the characteristic value more effectively distinguishable from noise.

[0063] Step S4: According to and Calculate the mechanical aging index of the brake and output the aging degree of the brake according to the preset grading standard.

[0064] After the above steps, this embodiment obtains two characteristic quantities. The first characteristic quantity is the d-axis inductance fundamental amplitude. The first characteristic directly reflects the degree of eccentricity of the sun gear shaft and is positively correlated with the severity of aging. The second characteristic is the optimal excitation frequency. It reflects the stiffness and natural frequency shift of the mechanical structure, and characterizes aging types such as wear, increased clearance, and deformation.

[0065] Will and As input parameters for the dual-feature evaluation model, durability experiments were used to establish... to eccentricity eigenvalue A library of monotonic positive correlation mappings, and The system utilizes a monotonic positive correlation mapping library for stiffness degradation and eliminates the possibility of motor-specific faults through dual-feature amplitude-frequency curves, diagnosing the characteristic frequencies of mechanical structure aging. Finally, it combines multi-parameter data with actual acquired data to establish... and Aging level indicators The mapping library and hierarchical standard are expressed as follows: ; ; ; In the formula, As an indicator of the degree of mechanical aging; The eccentricity characteristic value of the brake's mechanical structure is obtained by converting the relative change of the d-axis inductance fundamental wave amplitude value relative to the reference value using a proportional coefficient. This is the stiffness retention coefficient of the brake's mechanical structure, which is the square of the ratio of the current optimal excitation frequency to the reference optimal excitation frequency. This refers to the stiffness degradation of the brake's mechanical structure. The fundamental amplitude reference value is the one corresponding to the mechanical structure of the brake that has not aged. The optimal excitation frequency reference value is the one corresponding to the mechanical structure of the brake that has not aged. and The proportionality coefficient is determined by the experiment; and Let be the normalized weighting coefficients calibrated by experiments, and satisfy . .

[0066] The calculated aging index The aging level is output according to the preset grading standard, which is as follows: If the output is no greater than 3%, it indicates that the performance is normal. A value greater than 3% but not greater than 5% indicates slight aging, meaning regular monitoring is required. A value greater than 5% but not greater than 10% indicates moderate aging, meaning timely maintenance is required. If the output is greater than 10%, it indicates severe aging and requires immediate replacement.

[0067] because and Corresponding to different physical sources, when the motor body demagnetizes or there is a short circuit between winding turns, the air gap magnetic field and the winding inductance change. The values ​​change accordingly, but the equivalent stiffness and equivalent mass of the mechanical structure remain unchanged. Maintain stability; when a mechanical structure ages, both eccentricity and equivalent stiffness change simultaneously. and Simultaneous changes occur. Therefore, the dual-feature joint assessment can effectively distinguish between mechanical structural aging and motor body demagnetization and winding faults, solving the confusion problem existing in traditional single-feature diagnostic schemes.

[0068] In this embodiment, the SPMSM has a rated speed of 3000 rpm, 5 pole pairs, a DC bus voltage of 48V, a stator resistance of 0.35Ω, a d-axis inductance reference value of 1.2mH, an inverter switching frequency of 10kHz, and a current loop bandwidth of 300Hz. In step S1, the amplitude of the sweep frequency high-frequency voltage is taken as 8% of the DC bus voltage, i.e., 3.84V. The sweep frequency increases from 500Hz to 1500Hz at a rate of 50Hz per second. The encoder is a single-turn absolute rotary encoder with a single-turn absolute position resolution of 4096 divisions per revolution. The optimal excitation frequency at zero speed is obtained through frequency sweeping. The optimal excitation frequency was 860Hz, which was obtained after correction by the frequency offset compensation model. The frequency is 892Hz. In step S2, the frequency of the q-axis high-frequency voltage is set to 892Hz, and the amplitude is set to 3.84V; the frequency of the d-axis high-frequency voltage is set to 1192Hz, which is the same as... The frequency interval is 300Hz, and the amplitude is taken as 6% of the DC bus voltage, i.e., 2.88V. In step S3, the passband center frequency of the bandpass filter is set to 1192Hz, and the passband width is set to 120Hz. The sequence was resampled at equal mechanical angle intervals to obtain 240 points, covering 10 complete rotor mechanical cycles. In step S4, the scaling factor... The calibration value is 0.42. The calibration is 0.65, and the weighting coefficient is... The calibration value is 0.6. Calibrated to 0.4, reference value It is 0.018mH, the reference value. The frequency is 940Hz. Taking a certain test in this embodiment as an example, the fundamental amplitude of the d-axis inductor is obtained through step S3. If it is 0.0192mH, then It is 0.067. It is 0.028; by For 892Hz and For 940Hz It is 0.900. It is 0.065; substituting it, we get... The value is 0.043, which is 4.3%, falling within the range of greater than 3% but not greater than 5%, indicating slight aging and suggesting the need for regular monitoring.

[0069] In this embodiment, in order to enable signal acquisition, Identification and feature extraction can be completed within a single gap elimination cycle, and the detection time can be controlled within the hundreds of milliseconds by setting the size of the braking gap. After completing the offline analysis of the fast Fourier transform and outputting the aging degree detection results, the rated braking gap is restored, bringing the brake back to the stage of consistency of engineering operation indicators.

[0070] Example 2 This embodiment provides a method for detecting the mechanical aging of brakes during the vehicle's engineering operation phase. The difference from Embodiment 1 is that this embodiment does not change the rated braking clearance of the brake, thus completely preserving the normal braking performance of the brake. Instead, it uses a multiple sampling accumulation method to improve data accuracy. Steps S1, S2, and S4 of this embodiment are the same as in Embodiment 1, the difference being the d-axis inductance in step S3. The time series is composed of multiple d-axis inductance sampling segments fused together, rather than being directly composed of a single d-axis inductance sampling segment.

[0071] In actual vehicle operation, the duration of each gap elimination phase is relatively short, and the length of a single d-axis inductance sampling segment may not be sufficient to complete a full Fast Fourier Transform. Therefore, in this embodiment, multiple d-axis inductance sampling segments are obtained in each of the multiple gap elimination phases of the brake, and then these d-axis inductance sampling segments are merged into a long sequence.

[0072] Because the brake undergoes clamping and releasing processes between each sampling, the rotor phases corresponding to the beginning and end of each d-axis inductor sampling segment are not continuous. If they are directly spliced ​​in chronological order, phase jumps will be introduced at the splicing points, resulting in spectral leakage and harmonic aliasing, thus distorting the results of the Fast Fourier Transform. To avoid the above problems, this embodiment adopts an integrated method of mechanical angle domain regularization and phase synchronization fusion for overall data fusion and sequence processing, specifically including the following three steps.

[0073] The first step is equal-interval resampling in the mechanical angle domain. Data is acquired synchronously throughout the process. The data and the single-turn absolute mechanical position signal output by the rotary encoder are used to resample each d-axis inductance sampling segment at equal mechanical angle intervals using linear interpolation based on the corresponding rotor mechanical angle. The number of resampling points is set to no more than the number of points within the same mechanical cycle. The original point count transforms the discrete, unequally spaced time-domain signal into a periodic signal uniformly distributed across the mechanical angle domain. After this processing step, the sequence's independent variable is converted from time to rotor mechanical angle, and the sampling points are evenly distributed along the mechanical angle axis, eliminating the problem of uneven sampling angles caused by minute fluctuations in motor speed.

[0074] The second step involves connecting the absolute mechanical angles across segments. The absolute mechanical angle scales at the beginning and end of each d-axis inductor sampling segment are matched one by one, and the seamless connection across segments is achieved according to the angular progression of mechanical rotation, thereby eliminating phase loss and signal jumps caused by sampling intervals. After this step, the various d-axis inductor sampling segments are connected end-to-end on the mechanical angle axis, continuous and uninterrupted, equivalent to a long sequence continuously acquired in the mechanical angle domain.

[0075] The third step is mechanical cycle synchronization averaging. Combining the mechanical cycle synchronization averaging mechanism, the concatenated data is uniformly cut to the standard length corresponding to a complete mechanical cycle, i.e., a full rotor rotation, and then averaged point-by-point with the rotor mechanical phase as the reference. This step suppresses random sampling noise, electromagnetic spurious interference, and noise components introduced by segmented phase deviations without disrupting the mechanical vibration modulation characteristics and stiffness-related resonant components corresponding to the q-axis excitation.

[0076] It should be noted that in this embodiment, the number of pole pairs of the SPMSM is greater than 1, and each mechanical cycle of the rotor corresponds to multiple electrical cycles. The eccentric vibration of the sun gear shaft caused by the aging of the brake's mechanical structure, and the d-axis inductance fluctuation modulated by this eccentric vibration, have a fundamental period that is consistent with the mechanical rotation period of the rotor, but not with the electrical period. If resampling, connection, and synchronous averaging are performed based on the electrical angle domain and the electrical period, multiple electrical cycles in different mechanical phases within the same mechanical cycle will be superimposed. The inductance fluctuation component caused by eccentricity cancels each other out during the superposition process, thus weakening the aging characteristics. Therefore, this embodiment uniformly completes resampling, absolute angle connection, and synchronous averaging of the entire mechanical cycle within the mechanical angle domain. The rotary encoder used in this embodiment can directly output the absolute mechanical position of the rotor in a single revolution. Even if clamping, braking, and release processes occur between each gap elimination stage, the absolute mechanical phase relationship between each d-axis inductance sampling segment can still be recovered based on the absolute mechanical position of the single revolution, providing a unique and definite angle reference for cross-segment connection.

[0077] After the above three steps, a mechanical periodic average sequence with continuous mechanical angles, uninterrupted mechanical phases, and complete periodicity is finally formed, which is then used as the d-axis inductance. The time series data avoids the distortion problem caused by splicing multiple data segments, ensuring the accuracy and stability of extracting the resonant frequency of the brake's mechanical structure under rated speed conditions. A fast Fourier transform is then performed on the average mechanical cycle sequence to obtain the DC component. With fundamental component The subsequent aging assessment process is the same as in Example 1.

[0078] In this embodiment, the SPMSM has a rated speed of 3000 rpm, 5 pole pairs, and a DC bus voltage of 48V. In step S1, the amplitude of the sweep frequency voltage is 5% of the DC bus voltage, i.e., 2.4V, and the sweep frequency increases from 500Hz to 1500Hz at a rate of 40Hz per second. After sweep frequency and frequency offset compensation correction, the optimal excitation frequency is obtained. The frequency is 885Hz. In step S2, the frequency of the d-axis high-frequency voltage is set to 1085Hz, which is the same as... The frequency interval is 200Hz, and the amplitude is 5% of the DC bus voltage. In step S3, 16 points are resampled at equal mechanical angle intervals for each rotor mechanical cycle. A single d-axis inductor sampling segment covers one mechanical cycle, accumulating 16 d-axis inductor sampling segments. After mechanical angle domain normalization and absolute mechanical angle concatenation, a long fusion sequence with a length of 256 points covering 16 consecutive mechanical cycles is obtained. This sequence is then trimmed to the length of the entire mechanical cycle and averaged point by point to obtain a single mechanical cycle average sequence with a length of 16 points. The calibration coefficients and reference values ​​in step S4 are the same as in Example 1.

[0079] Example 3 This embodiment provides a method for testing the mechanical aging of brakes under different motor parameters and frequency values, illustrating the applicability of the technical solution to brakes of different specifications. The execution logic of steps S1 to S4 in this embodiment is exactly the same as that in Embodiment 1, the only difference being the specific values ​​of the motor parameters and injection parameters.

[0080] In this embodiment, the SPMSM has a rated speed of 4500 rpm, 3 pole pairs, a DC bus voltage of 24V, a stator resistance of 0.28Ω, a d-axis inductance reference value of 0.85mH, an inverter switching frequency of 8kHz, and a current loop bandwidth of 250Hz. The d-axis high-frequency injection frequency of 665Hz avoids the mechanical natural frequency of 1120Hz and the rotor mechanical rotation frequency of 75Hz and its lower harmonics. In step S1, the amplitude of the sweep frequency high-frequency voltage is taken as 10% of the DC bus voltage, i.e., 2.4V, and the sweep frequency increases from 600Hz to 1400Hz at a rate of 60Hz per second. The zero-speed optimal excitation frequency is obtained through frequency sweeping. The optimal excitation frequency was 1120Hz, which was obtained after correction by the frequency offset compensation model. The frequency is 1165Hz. In step S2, the frequency of the q-axis high-frequency voltage is set to 1165Hz; the frequency of the d-axis high-frequency voltage is set to 665Hz, which is the same as... The frequency interval is 500Hz, and the amplitude is 7% of the DC bus voltage. In step S3, the passband center frequency of the bandpass filter is set to 665Hz, and the passband width is set to 150Hz. In step S4, the scaling factor... The calibration value is 0.38. The calibration is 0.71, and the weighting coefficient is... The calibration value is 0.55. Calibrated at 0.45, reference value It is 0.013mH, the reference value. It is 1240Hz.

[0081] As the number of service cycles increases, monotonically increasing and Monotonic decrease, aging degree index It then increases monotonically. Figure 7 The comparison waveforms of d-axis inductance fluctuations under four conditions—no aging, slight aging, moderate aging, and severe aging—are presented in this embodiment. Figure 7 As shown, under the same coordinate system, the greater the degree of aging, the more significant the increase in the amplitude of the high-frequency inductor fundamental wave obtained from the real-time waveform. This phenomenon is consistent with the aforementioned mechanism analysis and also provides a basis for the establishment of a mapping library from dual features to aging degree.

[0082] To verify the feasibility and effectiveness of the technical solution in this embodiment, experiments were conducted on the above embodiment, and the results are as follows.

[0083] Regarding the effects of active excitation, such as Figure 6 As shown, when no high-frequency signal is injected into the q-axis, The fluctuations are weak and easily drowned out by noise; after implementing q-axis sweep frequency excitation... The fluctuation amplitude increased by 3 to 5 times, the aging characteristics were clearly distinguishable, and the ability to identify early aging was greatly improved.

[0084] In terms of diagnostic accuracy, The amplitude reflects the degree of aging. Reflecting changes in structural stiffness, the combination of these two methods can achieve accurate determination of aging degree and fault type, with higher accuracy than using only one method. The single-feature approach improves performance by more than 20%.

[0085] Dual-axis heterogeneous frequency injection and frequency division filtering can effectively suppress vehicle-mounted electromagnetic interference and vibration interference. Under actual vehicle operation conditions, the dispersion of aging degree indicators obtained from repeated tests remains at a low level, and the test results are stable and reliable. The entire testing process is embedded in the gap elimination stage, requiring no additional sensors and not changing the mechanical structure of the brake, making it suitable for mass industrial applications.

[0086] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described; only preferred embodiments of the present invention are illustrated. The descriptions are relatively specific and detailed, but they should not be construed as limiting the scope of the present invention. As long as the combination of these technical features does not contradict each other, it should be considered within the scope of this specification.

[0087] It should be noted that those skilled in the art can make various modifications and improvements without departing from the inventive concept, and these all fall within the scope of protection of this invention. Therefore, the scope of protection of this invention should be determined by the appended claims.

Claims

1. A method for detecting the mechanical aging of a brake based on an SPMSM drive motor, wherein the rotor shaft of the surface-mounted permanent magnet synchronous motor (SPMSM) is rigidly connected coaxially to the sun gear shaft of the planetary gearbox of the brake, characterized in that... Includes the following steps: Step S1: When the SPMSM is at zero speed, inject a high-frequency sweep voltage into the q-axis of the SPMSM. Take the frequency at which the amplitude of the angular velocity sweep frequency component fluctuation is the largest, and correct the frequency to the optimal excitation frequency using a pre-calibrated frequency offset compensation model. ; Step S2: During the gap elimination phase of the brake, the SPMSM is operated at a constant speed at its rated speed, and a frequency of [frequency value missing] is injected into the q-axis of the SPMSM. q-axis high-frequency voltage, injection frequency into the d-axis and Different d-axis high-frequency voltages; Step S3: Perform bandpass filtering on the d-axis voltage and current to identify the d-axis inductance. ,right Perform a Fourier transform on the time series over at least one rotor mechanical cycle to obtain the fundamental amplitude at the rotor mechanical frequency. ; Step S4: According to and Calculate the mechanical aging degree index of the brake and output the aging degree of the brake according to the preset grading standard.

2. The method for detecting mechanical aging of brakes based on SPMSM drive motors according to claim 1, characterized in that: In step S1, the amplitude of the sweep frequency high-frequency voltage is 5% to 10% of the DC bus voltage of the SPMSM, and the frequency of the sweep frequency high-frequency voltage increases in the range of 500Hz to 1500Hz.

3. The method for detecting mechanical aging of brakes based on SPMSM drive motors according to claim 2, characterized in that: The expression for the sweep frequency voltage mentioned in step S1 is: ； In the formula, This represents the instantaneous value of the high-frequency sweep voltage; The amplitude of the sweep frequency high-frequency voltage; The sweep frequency is the angular frequency that increases over time. For time; The frequency corresponding to the sweep frequency is higher than the fundamental frequency of the SPMSM, higher than the current loop bandwidth of the SPMSM, and lower than the switching frequency of the inverter driving the SPMSM.

4. The method for detecting mechanical aging of brakes based on SPMSM drive motors according to claim 1, characterized in that: The frequency offset compensation model in step S1 is pre-calibrated as follows: the resonant frequency of the brake at zero speed and the resonant frequency of the brake at rated speed are collected respectively, and the difference between the resonant frequency at rated speed and the resonant frequency at zero speed is used as the frequency offset compensation amount of the frequency offset compensation model, and the frequency offset compensation amount is written into the controller of the brake.

5. The method for detecting mechanical aging of brakes based on SPMSM drive motors according to claim 1, characterized in that: The frequency of the d-axis high-frequency voltage in step S2 is... The absolute value of the difference is not less than 200Hz and not greater than 500Hz.

6. The method for detecting mechanical aging of brakes based on SPMSM drive motors according to claim 1, characterized in that: The time series described in step S3 consists of d-axis inductance sampling segments, which are obtained in the following way: from the first k The d-axis high-frequency voltage and d-axis high-frequency current within the injection cycle are used to calculate the d-axis inductance in the [number]th injection cycle. k Value per injection cycle In the formula k The sequence number of the injection cycle and k For positive integers, the values ​​obtained from multiple consecutive injection cycles within the same gap elimination phase are considered. By serial number k Arranged from smallest to largest.

7. The method for detecting mechanical aging of brakes based on SPMSM drive motors according to claim 6, characterized in that: The Calculate using the following formula: ； In the formula, The amplitude of the high-frequency voltage along the d-axis; ω is the angular frequency of the high-frequency voltage along the d-axis; For the first k The amplitude of the d-axis high-frequency current obtained after bandpass filtering within each injection cycle.

8. The method for detecting mechanical aging of brakes based on SPMSM drive motors according to claim 6, characterized in that: Multiple d-axis inductance sampling segments are obtained during the multiple gap elimination stages of the brake. The time series is obtained by fusing the multiple d-axis inductance sampling segments in the following manner: each d-axis inductance sampling segment is resampled at equal mechanical angle intervals based on the synchronously acquired rotor single-turn absolute mechanical angle. The segments are then connected in the order of the absolute mechanical angle scale of the beginning and end of each d-axis inductance sampling segment. The connected data is then cut into whole mechanical cycle lengths based on the rotor mechanical phase and then averaged point by point. The rotor single-turn absolute mechanical angle is synchronously acquired by the encoder, which is a single-turn absolute rotary encoder.

9. The method for detecting mechanical aging of brakes based on SPMSM drive motors according to claim 1, characterized in that: The expression for the mechanical aging degree index is: ； ； ； In the formula, As an indicator of the degree of mechanical aging; The characteristic value of the eccentricity of the brake's mechanical structure; This refers to the stiffness degradation of the brake's mechanical structure. The fundamental amplitude reference value is the one corresponding to the mechanical structure of the brake that has not aged. The optimal excitation frequency reference value is the one corresponding to the mechanical structure of the brake that has not aged. and The proportionality coefficient is determined by the experiment; and The normalized weighting coefficients are determined by experiments and ; The stiffness retention coefficient of the brake's mechanical structure; The d-axis inductor fundamental amplitude value relative to the reference value The relative change.

10. The method for detecting mechanical aging of brakes based on SPMSM drive motors according to claim 9, characterized in that: The preset grading standard mentioned in step S4 is: Output has no aging when it is no greater than 3%. When the percentage is greater than 3% but not greater than 5%, the output shows slight aging. When the percentage is greater than 5% but not greater than 10%, the output indicates moderate aging. When the output is greater than 10%, the output is severely aged.

Images