Sensorless FOC ripple feature rotor positioning method and system

Abstract

The application discloses a non-inductive FOC ripple feature rotor positioning method and system, and the method comprises the following steps: in the motor zero-speed starting stage, a multi-directional detection voltage vector with gradually increasing amplitude is injected into the stator winding, and the multi-directional detection voltage vector contains at least three non-collinear spatial directions; the three-phase current transient response under the action of each voltage vector is synchronously collected, and the current ripple characteristic component matched with the rotor salient pole frequency is extracted through band-pass filtering; based on the spatial distribution anisotropy of the current ripple characteristic component, a position sensitive factor set corresponding to the voltage direction is constructed, and each position sensitive factor in the position sensitive factor set represents the distribution weight of the ripple energy under the excitation of the voltage direction. Through the multi-dimensional collaborative detection mechanism and the dynamic feature compensation architecture, the application realizes robust control in all working conditions.

Description

Sensorless FOC Ripple Characteristic Rotor Positioning Method and System Technical Field

[0001] This application relates to the field of motor control technology, and in particular to a sensorless FOC ripple characteristic rotor positioning method and system. Background Technology

[0002] In brushless motor drive solutions, detecting the initial rotor position during the sensorless startup phase is a key challenge.

[0003] Traditional high-frequency signal injection methods rely on the inherent salient pole effect of motors. However, in motors with low salient pole ratios or under deep magnetic saturation conditions, the current response signal is extremely weak, making it difficult to extract effective features using conventional unidirectional excitation, resulting in severely insufficient zero-speed positioning accuracy. More seriously, motors face complex dynamic conditions in actual operation: drastic changes in winding temperature cause resistance parameter drift, sudden load changes lead to distortion of core saturation characteristics, and demagnetization of permanent magnets weakens the magnetic field strength. These factors collectively cause amplitude attenuation, frequency drift, and phase distortion in the current ripple characteristics. Existing ripple detection technologies have three limitations:

[0004] First, fixed-direction voltage injection cannot capture the spatial anisotropy of magnetic saturation, and weak ripple characteristics are easily overwhelmed by inverter nonlinear noise. Second, static filter parameters are difficult to adapt to hot-state salient pole frequency shifts, leading to a sharp decrease in feature extraction reliability. Third, the position calculation model does not consider the strong coupling relationship between flux decay and saturation effects, resulting in systematic deviations in angle calculations under demagnetization conditions. These shortcomings lead to the risk of positioning failure under extreme conditions using traditional methods. Summary of the Invention

[0005] To address the aforementioned problems, embodiments of the present invention provide a sensorless FOC ripple characteristic rotor positioning method, the method comprising:

[0006] Acquire visible light reflection images and near-infrared scattering spectrum data of the PCB surface;

[0007] Spatial filtering is performed on the visible light reflection image to extract pixels in the interface region. At the same time, the coating thickness gradient distribution is calculated based on the Mie scattering features in the scattering spectrum data, and the edges are fused to generate an edge gradient distribution map.

[0008] Based on the gradient magnitude change rate in the edge gradient distribution map, an adaptive dynamic threshold algorithm is used to segment the effective coverage area of ​​the coating layer, and the curvature extreme point density and boundary fractal dimension parameters at the segmentation boundary are extracted.

[0009] The curvature extremum point density and boundary fractal dimension parameters are input into a pre-trained coating curing morphology prediction model, and the predicted value of the coating interface bonding strength and the coordinates of potential peeling risk areas are output by coupling thermodynamic simulation data.

[0010] The surface of the coating layer is scanned by an array of dielectric coupling sensors, and the impedance phase response of each scanning point at different excitation frequencies is measured. The micro porosity distribution map is calculated based on the slope of the phase offset as a function of frequency.

[0011] The interface is combined with the strength prediction value, the coordinates of the potential peeling risk area and the micro porosity distribution map for spatiotemporal registration. A comprehensive evaluation matrix of coating quality is generated by a multi-source data decision fusion algorithm, and a defect location report and quality level classification results are output.

[0012] Furthermore, spatial filtering methods include:

[0013] A non-uniform illumination compensation algorithm is used to perform illumination equalization processing on visible light reflection images. The algorithm calculates pixel compensation coefficients by establishing a local illumination distribution model, wherein the compensation coefficients are dynamically adjusted according to the surface roughness characteristics in the near-infrared band scattering spectrum data.

[0014] Furthermore, methods for generating edge gradient distribution maps include:

[0015] The gradient phase information of the visible light reflection image is superimposed with the Mie scattering phase delay of the near-infrared scattering spectrum, and the interference of the substrate material texture on the coating boundary recognition is eliminated by phase gradient consistency verification.

[0016] Furthermore, the adaptive dynamic threshold algorithm includes:

[0017] A probability density function is established based on the statistical distribution of the gradient magnitude change rate. Variational mode decomposition technology is used to separate the threshold response characteristics of the coating body and the boundary transition zone, and the segmentation threshold interval is dynamically determined.

[0018] Furthermore, the training methods for the coating curing morphology prediction model include:

[0019] A finite element simulation model coupling thermal-humidity-mechanical multiphysics is constructed. By introducing the time-varying viscoelastic constitutive equation of the coating material, a training set of interface stress distribution under different curing conditions is generated. The training set includes the mapping relationship between temperature gradient, humidity penetration depth and interface peeling stress.

[0020] Furthermore, the methods for analyzing impedance phase response include:

[0021] An equivalent circuit model of the micropores in the coating layer is established. The phase shift at different frequencies is fitted by the Cole-Cole distribution function, and the pore connectivity index and equivalent dielectric relaxation time parameters are extracted.

[0022] Furthermore, methods for achieving spatiotemporal registration include:

[0023] A three-dimensional registration coordinate system is established based on the spatial topology of the coating surface. The mechanical confidence of the interface bonding strength prediction value and the electrical sensitivity of the micro porosity distribution are weighted and fused to generate a quality evaluation weight matrix with physical interpretability.

[0024] Furthermore, the sensorless FOC ripple characteristic rotor positioning method also includes:

[0025] A multi-frequency eddy current sensing unit is integrated into an array-type dielectric coupling sensor, and the coating dielectric phase response and substrate eddy current skin effect signal are simultaneously acquired through frequency division multiplexing technology.

[0026] A frequency-domain blind source separation model is constructed based on the inverse square relationship between the skin depth and frequency of eddy current signals, and the cross-coupled noise component caused by defects in the metal substrate is decoupled from the dielectric phase response.

[0027] By utilizing the frequency domain characteristics of high-frequency eddy currents, which are sensitive to the pore closure of the coating surface, and low-frequency eddy currents, which characterize the continuity of the substrate metal, a micro-porosity distribution map with substrate artifacts removed is reconstructed.

[0028] A sensorless FOC ripple characteristic rotor positioning method, the system includes:

[0029] The dual-modal acquisition module acquires visible light reflection images and near-infrared scattering spectrum data of the PCB surface.

[0030] The gradient map generation module performs spatial filtering on the visible light reflection image to extract pixels in the interface region, and calculates the coating thickness gradient distribution based on the Mie scattering features in the scattering spectrum data, and fuses them to generate an edge gradient distribution map.

[0031] The dynamic threshold segmentation module uses an adaptive dynamic threshold algorithm to segment the effective coverage area of ​​the coating layer based on the gradient magnitude change rate in the edge gradient distribution map, and extracts the curvature extreme point density and boundary fractal dimension parameters at the segmentation boundary.

[0032] The interface strength prediction module inputs the curvature extreme point density and boundary fractal dimension parameters into a pre-trained coating curing morphology prediction model, and outputs the predicted value of the coating interface bonding strength and the coordinates of potential peeling risk areas by coupling thermodynamic simulation data.

[0033] The dielectric porosity analysis module scans the surface of the coating layer using an array of dielectric coupling sensors, measures the impedance phase response of each scanning point at different excitation frequencies, and calculates the micro porosity distribution map based on the slope of the phase shift as a function of frequency.

[0034] The results output module combines the interface with the predicted strength value, the coordinates of the potential peeling risk area, and the micro porosity distribution map for spatiotemporal registration. It then generates a comprehensive evaluation matrix of coating quality through a multi-source data decision fusion algorithm and outputs a defect location report and quality grade classification results.

[0035] The technical effects and advantages of the sensorless FOC ripple characteristic rotor positioning method provided by this invention are as follows:

[0036] This invention addresses two core technical challenges in sensorless start-up: insufficient zero-speed positioning accuracy and poor adaptability to operating conditions. Through a multi-dimensional collaborative detection mechanism and a dynamic feature compensation architecture, it achieves robust control across all operating conditions. The invention employs a multi-directional rotating voltage excitation strategy to deeply excite the spatial anisotropic response of magnetic saturation, enabling the extraction of robust rotor orientation features even under weak salient pole and deep saturation conditions. It constructs a ripple spectrum adaptive tracking mechanism and a flux linkage-saturation coupling compensation model to effectively suppress feature distortion caused by parameter drift and demagnetization effects, ensuring the stability of angle calculations under extreme conditions. Pulse excitation and ripple analysis interact in a closed-loop manner in the time-frequency domain, dynamically avoiding single-mode detection blind spots and significantly improving the system's fault tolerance in complex electromagnetic environments. Attached Figure Description

[0037] Figure 1 is a flowchart of the non-sensory FOC ripple characteristic rotor positioning method according to Embodiment 1 of this application;

[0038] Figure 2 is a flowchart of the sensorless FOC ripple characteristic rotor positioning method according to Embodiment 2 of this application;

[0039] Figure 3 is a schematic diagram of the connection of the sensorless FOC ripple characteristic rotor positioning system according to Embodiment 3 of this application. Detailed Implementation

[0040] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0041] Example 1: As shown in Figure 1, this embodiment of the invention provides a rotor positioning method for non-inductive FOC ripple characteristics, the method including:

[0042] Step S1: During the zero-speed start-up phase of the motor, a multi-directional detection voltage vector with gradually increasing amplitude is injected into the stator winding. The multi-directional detection voltage vector contains at least three non-collinear spatial directions.

[0043] Step S2: Synchronously acquire the transient response of the three-phase current under the action of each voltage vector, and extract the current ripple characteristic component that matches the rotor salient pole frequency through bandpass filtering;

[0044] Step S3: Based on the spatial anisotropy of the characteristic components of current ripple, construct a set of position sensitive factors corresponding to the voltage direction. Each position sensitive factor in the set represents the distribution weight of ripple energy under excitation in that voltage direction.

[0045] Step S4: Input the set of position-sensitive factors into the pre-trained ripple-position mapping model and output the initial position angle of the rotor. The ripple-position mapping model establishes a nonlinear relationship between ripple characteristics and position through the magnetic saturation effect of the motor.

[0046] Step S5: Switch to FOC closed-loop operation based on the rotor's initial position angle;

[0047] In step S2, the extraction of current ripple characteristic components is independent of the back electromotive force signal, and the ripple-position mapping model in step S4 completes the position calculation when the motor is stationary.

[0048] In the current ripple feature component extraction process in step S2, frequency domain feature separation is required for the acquired three-phase current transient response. Since the current ripple frequency excited by the rotor salient pole effect is directly related to the motor structural parameters, this method uses a dynamically adjustable bandpass filter for feature extraction. The passband center frequency of the bandpass filter needs to be matched with the fundamental frequency of the rotor salient pole harmonic in real time. This frequency is determined by the number of pole pairs and salient pole characteristics of the motor. It is a fixed value in the static state but needs to be identified autonomously by the algorithm. In specific implementation, the control system dynamically calibrates the center frequency according to the spectral characteristics of the current injected voltage vector to ensure that the ripple energy passes through to the maximum extent (for example, when the estimated salient pole fundamental frequency is 250Hz, the passband center frequency is automatically locked to 250Hz).

[0049] Meanwhile, the filter bandwidth setting is related to the anisotropy intensity of the motor's magnetic circuit; according to electromagnetic principles, the difference between the direct and quadrature axis inductance (L... d -L q The larger the current ripple energy, the more concentrated it is on the frequency spectrum. Therefore, the filter is designed with a bandwidth of -3dB that is proportional to the difference between the direct and quadrature axis inductances. For built-in permanent magnet motors with high saliency (such as L...), d / L q=0.5), a narrower bandwidth is used to suppress noise; while for surface-mount motors with low saliency (Ld≈Lq), the bandwidth is appropriately widened to capture dispersed ripple energy. This adaptive bandwidth mechanism ensures the signal-to-noise ratio of feature extraction while being compatible with the saliency characteristics of different types of motors.

[0050] The spatial distribution characteristics of the current ripple feature component after this processing are further transformed into a position-sensitive factor in step S3. The entire process is based entirely on the current response and does not rely on the back electromotive force signal, which meets the core requirements of static positioning.

[0051] When constructing the set of position-sensitive factors in step S3, it is necessary to quantify the spatial anisotropy characteristics of the current ripple feature components. For the current ripple signal acquired in each voltage direction, two key features are first extracted: its kurtosis coefficient and the root mean square value of the envelope. The extraction methods include:

[0052] The kurtosis coefficient reflects the sharpness of the amplitude distribution of the ripple signal. Directions with significant salient pole effect (such as near the direct axis) will generate sharper current pulses due to magnetic saturation, resulting in higher kurtosis values.

[0053] The root mean square value of the envelope characterizes the overall intensity of ripple energy and is related to the projection component of the voltage vector in the rotor salient pole direction.

[0054] Multiplying the two yields the original sensitivity factor, which incorporates both waveform morphology and energy intensity information. For example, when the injected voltage direction is aligned with the rotor direct axis (d-axis), the current ripple exhibits high-frequency narrow pulses and concentrated energy, resulting in a significantly increased product. However, when the voltage direction is close to the quadrature axis (q-axis), the ripple distribution is smooth and the energy is dispersed, leading to a sharp decrease in the product.

[0055] To unify the dimensions and enhance directional contrast, the original sensitivity factor is normalized using an exponential function. Map to the [0,1] interval:

[0056] ;

[0057] In the formula, For the first The original sensitivity factor of direction , The minimum and maximum values ​​in all directions. As an adjustable gain coefficient, this nonlinear mapping amplifies the difference between the high-sensitivity direction and the low-sensitivity direction, ultimately forming a set of position-sensitive factors that characterize the weight of ripple energy distribution in each voltage direction, providing highly discriminative input features for the ripple-position mapping model in step S4.

[0058] Before constructing the ripple-position mapping model in step S4, offline training is required for the motor pre-calibration stage. The training methods include:

[0059] Data acquisition across all locations, including the following methods:

[0060] Control the motor rotor to rotate mechanically to N evenly distributed positions (e.g., N=512, corresponding to 0.7° resolution);

[0061] At each position θ g Repeat steps S1 to S3 to obtain the location sensitivity vector F of that point. g (Including features such as kurtosis coefficient and ripple energy);

[0062] Forming the training dataset {θ g ,F g g = 1, 2, ..., N;

[0063] Modeling methods for Support Vector Regression (SVR) include:

[0064] The position angle θ is used as the regression target, and the multidimensional factor vector F is used as the input feature.

[0065] Nonlinear mappings are constructed using radial basis function (RBF) kernels, including:

[0066] ;

[0067] in, For the number of support vectors, , and For training parameters, The subscript indicates the first... indivual.

[0068] Simultaneously, hyperparameters are optimized through grid search cross-validation.

[0069] The RBF kernel can characterize the fluctuation of factors caused by magnetic saturation (such as the third harmonic of the position factor caused by cogging torque), the ϵ-insensitive band ignores small measurement noise, and the support vector retains only key samples to meet the real-time position estimation requirements. The trained SVR model is burned into the controller ROM in the form of a coefficient matrix. During the online operation phase, the rotor position estimate θ^ is output in real time through the position sensitive factor input in step S4.

[0070] In the multi-directional voltage vector injection stage of step S1, a precise voltage excitation is generated using a space vector pulse width modulation (SVPWM) sequence. Specifically, the control system plans at least three non-collinear spatial voltage directions (e.g., 0°, 60°, and 120° directions) in the α-β coordinate system, and decomposes each target voltage vector into the timing sequence of the three-phase full-bridge circuit using the SVPWM algorithm; taking the 60° direction voltage injection as an example:

[0071] Calculate the duration of the basic voltage vector in that direction (e.g., vector U1 accounts for 70% and U2 accounts for 30%).

[0072] Generate the PWM drive waveform for the corresponding power transistor, where a dead time (e.g., 3μs) is forcibly inserted when switching between adjacent vectors.

[0073] During the dead time, the drive signals of the upper and lower bridge arms are shut off to completely block the current path and eliminate the risk of short circuit in the bridge arms.

[0074] The voltage vector amplitude gradually increases by linearly increasing the modulation depth. For example, the initial injection is 5% of the rated voltage (to avoid inrush current), and then gradually increases to 20% of the rated voltage (to enhance the ripple signal-to-noise ratio). Several PWM cycles (e.g., 5 cycles) are used in each direction to ensure stable current response, and the current decays naturally through dead time when the direction is switched. This injection method ensures the spatial pointing accuracy of the voltage vector while also taking into account the safety of the power circuit.

[0075] Voltage vector in step S1 During the injection process, the amplitude gradual increase strategy uses an exponentially rising curve to achieve smooth excitation, and its mathematical expression includes:

[0076] ;

[0077] In the formula, The target peak voltage (e.g., 20% of the rated voltage). The key time constant, For the current time, The current is constant; this design overcomes the current step problem caused by linear rise, i.e., the initial slope of the exponential curve is large ( When = 0, d / d = / It can quickly establish a magnetic field, and the slope gradually decreases in the later stages. As d approaches infinity, / d →0) suppresses the risk of overshoot.

[0078] time constant Only then will an adaptive mechanism be used, including:

[0079] During motor operation, the winding resistance R increases significantly with increasing temperature (for example, R increases by 50% at 150℃ compared to the cold state). If the current rises too slowly in the hot state, the ripple feature extraction will be delayed.

[0080] The control system acquires the online identified winding resistance value in real time (data shared via the FOC current loop parameter tuning module), and then... ∝1 / R is dynamically adjusted; for example, when the cold resistance R = 1Ω at 25℃, set... =50ms, automatically reduced to when the hot state R100℃=1.5Ω. ≈33ms, which keeps the transient response speed of the current stable.

[0081] This design ensures that the current ripple energy accumulation rate is consistent under different operating conditions, avoiding both cold-state overcurrent impact and hot-state feature extraction delay, thus laying the foundation for reliable acquisition in step S2.

[0082] In the current ripple feature extraction in step S2, the acquired three-phase current transient response is first subjected to Park transformation to transform it from the stationary coordinate system (abc) to the rotor synchronous rotating coordinate system (dq). This transformation eliminates the DC bias of the fundamental current through angle offset compensation, so that the ripple alternating component that was originally submerged by the fundamental current is significantly separated in the q-axis current. For example, when the injected voltage direction coincides with the d-axis, the q-axis current exhibits a high-frequency ripple with an amplitude of about 0.5A, while the d-axis current ripple is almost zero.

[0083] For the separated ripple components, their spectral kurtosis is further calculated to quantify the frequency domain energy concentration characteristics, including:

[0084] The ripple signal is analyzed in the time-frequency two-dimensional plane using short-time Fourier transform (STFT), and the window function length is set to cover the main harmonic periods (e.g., a 100μs window length corresponds to a 10kHz analysis bandwidth).

[0085] Calculate the kurtosis coefficient at each frequency point fk The calculation methods include:

[0086] ;

[0087] In the formula, For time spectrum, The spectral kurtosis value represents the average over the time dimension. When the spectral kurtosis value is greater than 3 (e.g., SK=4.2 at 5kHz along the d-axis), it indicates that there is pulsed energy accumulation caused by magnetic saturation in this frequency band. The spectral kurtosis along the q-axis is close to 2 (wideband white noise characteristic), reflecting a smooth energy distribution without salient pole effect.

[0088] This method overcomes the information loss problem of traditional bandpass filtering, and the extracted spectral kurtosis feature matrix will serve as the key input for constructing the position-sensitive factor in step S3.

[0089] By monitoring the confidence index of location-sensitive factors in real time, online calibration of ripple-location mapping model parameters is triggered when the confidence level falls below a threshold.

[0090] Real-time acquisition of the generated location-sensitive factor vector, including methods such as:

[0091] The input source is the position-sensitive factor vector F. g = (m is the feature dimension), here The symbol for vector transpose;

[0092] Time-domain stability analysis was performed, and the analysis methods included:

[0093] Establish a sliding time window and calculate the coefficient of variation sequence for each feature component. :

[0094] ;

[0095] In the formula, here For the width of the sliding window, For time offset index ( =1,2,3,... ), The index of the eigenvalue, i.e., the first... Each feature value.

[0096] Take the weighted norm of the coefficient of variation for each dimension ,Right now In the formula, The feature weight matrix; th1 This is the warning threshold (low-risk critical value). th2 This is for calibrating the threshold (high-risk critical value).

[0097] Feature weight matrix Based on the feature sensitivity configuration (e.g., the ripple energy feature weight is higher than the kurtosis coefficient), as shown in the table below:

[0098]

[0099] Example 2: As shown in Figure 2, this example is a further improvement on Example 1. The difference is that, in actual operation, Example 1 showed that its fixed amplitude and limited direction of the probe voltage vector led to distortion of the ripple characteristics under high temperature and demagnetization conditions, resulting in insufficient spatial resolution of the position sensitivity factor. Therefore, the sensorless FOC ripple characteristic rotor positioning method further includes:

[0100] The amplitude of the detection voltage is adjusted in real time based on the winding temperature and the rate of change of the q-axis inductance, and an orthogonal detection direction is added to the initial three directions; when an abnormal increase in temperature is detected or the rate of change of inductance exceeds the critical value, the amplitude of the base voltage is dynamically increased according to the temperature change; an additional compensation voltage is superimposed based on the rate of change of inductance; and two sets of orthogonal directions are added to the initial at least three detection directions.

[0101] For example:

[0102] A permanent magnet synchronous motor (PMSM) used to drive a compressor stops after running at full load for an extended period. At this time:

[0103] The winding temperature is abnormally high: reaching 110°C (far higher than the 25°C at room temperature calibration).

[0104] Based on a preset temperature-voltage mapping table or function, the basic detection voltage amplitude is increased from 0.15V calibrated at room temperature to 0.25V (to compensate for the ripple attenuation and resistance increase effect caused by high temperature).

[0105] Based on the detected inductance change rate of -18%, the additional compensation voltage is calculated to be approximately 0.0045V, resulting in a final injected amplitude of 0.2545V.

[0106] A temperature decay function is used to suppress the ripple amplitude decay effect caused by high temperature; a demagnetization compensation matrix is ​​applied to eliminate phase distortion caused by flux weakening; and the corrected position sensitivity factor is input into the ripple-position mapping model to ensure stable rotor angle calculation accuracy under all operating conditions.

[0107] Example 3: As shown in Figure 3, based on the same inventive concept as the sensorless FOC ripple characteristic rotor positioning method in the previous examples, this application provides a sensorless FOC ripple characteristic rotor positioning system. The system and method embodiments in this application are based on the same inventive concept. The system includes:

[0108] The gradient injection module injects a multi-directional detection voltage vector with gradually increasing amplitude into the stator winding during the zero-speed start-up phase of the motor. The multi-directional detection voltage vector contains at least three non-collinear spatial directions.

[0109] The ripple extraction module synchronously acquires the transient response of the three-phase current under the action of each voltage vector, and extracts the current ripple characteristic components that match the rotor salient pole frequency through bandpass filtering.

[0110] The sensitive configuration module constructs a set of position sensitive factors corresponding to the voltage direction based on the spatial anisotropy of the characteristic components of the current ripple. Each position sensitive factor in the set of position sensitive factors represents the distribution weight of the ripple energy under the excitation of the voltage direction.

[0111] The angle mapping module inputs the set of position-sensitive factors into the pre-trained ripple-position mapping model and outputs the initial position angle of the rotor. The ripple-position mapping model establishes a nonlinear relationship between ripple characteristics and position through the magnetic saturation effect of the motor.

[0112] The closed-loop switching module switches to FOC closed-loop operation based on the rotor's initial position angle.

[0113] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

[0114] The above description is merely a preferred embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present application, based on the technical solution and concept of the present application, should be covered within the scope of protection of the present application.

Claims

1. A sensorless FOC ripple characteristic rotor positioning method, characterized in that, The methods include: During the zero-speed start-up phase of the motor, a multi-directional detection voltage vector with gradually increasing amplitude is injected into the stator winding. The multi-directional detection voltage vector contains at least three non-collinear spatial directions. The transient response of the three-phase current under the action of each voltage vector is collected synchronously, and the current ripple characteristic component matching the rotor salient pole frequency is extracted by bandpass filtering; the passband center frequency of the bandpass filter dynamically tracks the fundamental frequency of the rotor salient pole harmonics, and its -3dB bandwidth is proportional to the difference between the direct and quadrature axis inductance of the motor. Based on the spatial anisotropy of the characteristic components of current ripple, a set of position-sensitive factors corresponding to the voltage direction is constructed. Each position-sensitive factor in the set represents the distribution weight of ripple energy under excitation in that voltage direction. The method for constructing position-sensitive factors includes: calculating the product of the kurtosis coefficient and the root mean square value of the envelope of the ripple energy under excitation in each voltage direction, and normalizing it to a weight value in the [0,1] interval using an exponential function; inputting the set of position-sensitive factors into a pre-trained ripple-position mapping model, outputting the rotor initial position angle, and establishing a nonlinear relationship between ripple characteristics and position through the motor magnetic saturation effect; during the motor pre-calibration stage, obtaining the set of position-sensitive factors for the entire position interval, and constructing a nonlinear mapping function between the position angle and the multidimensional factor vector using a support vector regression mechanism; and switching to FOC closed-loop operation according to the rotor initial position angle.

2. The sensorless FOC ripple characteristic rotor positioning method according to claim 1, characterized in that, The injection of multi-directional probe voltage vectors employs a space vector pulse width modulation sequence, and dead time is inserted between adjacent vectors to suppress the risk of bridge arm shoot-through.

3. The sensorless FOC ripple characteristic rotor positioning method according to claim 2, characterized in that, The voltage vector amplitude gradually increases using an exponential rising curve, and its time constant is adaptively adjusted according to the hot winding resistance value of the motor.

4. The sensorless FOC ripple characteristic rotor positioning method according to claim 1, characterized in that, The method for extracting the characteristic components of current ripple includes: performing Parker transform on the three-phase current, separating the alternating ripple components in a rotating coordinate system, and calculating the spectral kurtosis of its short-time Fourier transform.

5. The sensorless FOC ripple characteristic rotor positioning method according to claim 1, characterized in that, During motor operation, the confidence index of the position-sensitive factor is monitored in real time. When the confidence level is lower than the threshold, the ripple-position mapping model parameters are calibrated online.

6. The sensorless FOC ripple characteristic rotor positioning method according to claim 1, characterized in that, The method also includes: adjusting the amplitude of the detection voltage in real time according to the winding temperature and the rate of change of q-axis inductance, and adding an orthogonal detection direction on the basis of the initial three directions; correcting the distortion of the position sensitivity factor through the temperature decay function and the demagnetization compensation matrix, and outputting it to the ripple-position mapping model.

7. A sensorless FOC ripple characteristic rotor positioning system, characterized in that, The system includes: The system includes a gradient injection module, which injects a multi-directional detection voltage vector with gradually increasing amplitude into the stator winding during the zero-speed start-up phase of the motor. The multi-directional detection voltage vector contains at least three non-collinear spatial directions. A ripple extraction module synchronously acquires the transient response of the three-phase current under the action of each voltage vector and extracts the current ripple characteristic components that match the rotor salient pole frequency through bandpass filtering. The passband center frequency of the bandpass filter dynamically tracks the fundamental frequency of the rotor salient pole harmonics, and its -3dB bandwidth is proportional to the difference between the direct and quadrature axis inductance of the motor. A sensitive configuration module constructs a set of position sensitive factors corresponding to the voltage direction based on the spatial anisotropy of the current ripple characteristic components. Each position sensitive factor in the set represents the distribution weight of ripple energy under excitation in that voltage direction. The method for constructing the position sensitivity factor includes: calculating the product of the kurtosis coefficient of the ripple energy under excitation in each voltage direction and the root mean square value of the envelope, and normalizing it to a weight value in the [0,1] interval using an exponential function; An angle mapping module inputs a set of position-sensitive factors into a pre-trained ripple-position mapping model and outputs the initial rotor position angle. The ripple-position mapping model establishes a nonlinear relationship between ripple features and position through the motor's magnetic saturation effect. During the motor pre-calibration stage, a set of position-sensitive factors across the entire position range is obtained, and a nonlinear mapping function between the position angle and multidimensional factor vectors is constructed using a support vector regression mechanism. A closed-loop switching module switches to FOC closed-loop operation based on the initial rotor position angle.

Images