A data processing method and system of a voltage detection sensor

Abstract

The application relates to the technical field of voltage detection, in particular to a data processing method and system of a voltage detection sensor, which comprises the following steps: acquiring an initial voltage sequence; performing data processing on the initial voltage sequence to obtain a first voltage sequence; performing a wavelet transform algorithm on the first voltage sequence to construct a voltage feature vector set of a subband; performing a principal component analysis algorithm based on the voltage feature vector set to identify a principal component sequence; judging whether the fluctuation amplitude of the principal component sequence exceeds a first threshold value; if yes, increasing the sampling rate and the quantization bit number of an analog-to-digital converter, and reacquiring the initial voltage sequence. The application adopts error correction and time consistency processing on initial data, and significantly improves the adaptability and robustness of voltage data in a severe fluctuation scene through a feature-driven sampling rate adjustment mechanism.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of voltage detection, and in particular to a data processing method and system of a voltage detection sensor. BACKGROUND

[0002] Voltage detection technology is crucial in power systems, industrial automation and intelligent devices, and its precision and real-time performance directly affect the stability and safety of device operation. Voltage detection technology refers to the whole process of continuous or intermittent sampling, filtering, conversion and analysis of voltage signals of a target system by sensors and data processing modules, aiming to achieve high-precision perception, real-time feedback and intelligent decision support of system voltage state.

[0003] However, existing voltage detection methods often perform poorly in dealing with dynamic changes. Existing systems rely on fixed-precision collection modes and are difficult to adapt to rapid fluctuations in voltage signals, resulting in inaccurate feature extraction. SUMMARY

[0004] The present application aims to at least solve one of the technical problems existing in the prior art or related art.

[0005] To this end, the present application provides a data processing method of a voltage detection sensor, which uses error correction and time consistency processing on initial data, and significantly improves the adaptability and robustness of voltage data in a severe fluctuation scenario through a feature-driven sampling rate adjustment mechanism. Compared with traditional static sampling methods, it has intelligent sampling rate control capability, lower error sensitivity and stronger system self-adjustment capability.

[0006] The present application also provides a data processing system of a voltage detection sensor, an electronic device and a storage medium comprising the above-mentioned data processing method of a voltage detection sensor.

[0007] According to the data processing method of a voltage detection sensor of the first aspect of the present application, the method comprises:

[0008] An initial voltage sequence is obtained, which includes an initial A-phase voltage sequence, an initial B-phase voltage sequence and an initial C-phase voltage sequence. The initial voltage sequence is obtained by collecting analog voltage signals of A-phase voltage, B-phase voltage and C-phase voltage of a three-phase motor through a voltage sensor, digitizing the analog voltage signals through an analog-to-digital converter at a basic sampling rate and a basic quantization bit number, and obtaining discrete voltage values with timestamps.

[0009] The initial voltage sequence is processed to obtain a first voltage sequence. The data processing at least includes error correction and time alignment.

[0010] A wavelet transform algorithm is performed on the first voltage sequence to decompose the first voltage sequence into sub-bands of different frequencies, and a voltage feature vector set of the sub-bands is constructed. The voltage feature vector set is used to analyze the operating state of the three-phase motor.

[0011] Based on the voltage feature vector set, a principal component analysis algorithm is executed to identify the principal component sequence; the principal component sequence is the most representative feature sequence extracted from the voltage feature vector set.

[0012] Determine whether the fluctuation amplitude of the principal component sequence exceeds a first threshold;

[0013] If so, increase the sampling rate and quantization bits of the analog-to-digital converter to reacquire the initial voltage sequence;

[0014] If not, then output the voltage feature vector set.

[0015] Optionally, the step of processing the initial voltage sequence to obtain the first voltage sequence includes:

[0016] The initial voltage sequence is corrected for errors using a multi-channel joint Kalman filter algorithm to generate a second voltage sequence.

[0017] The second voltage sequence is time-aligned using linear interpolation to generate the first voltage sequence.

[0018] Optionally, the step of using a multi-channel joint Kalman filter algorithm to correct the error in the initial voltage sequence and generate the second voltage sequence includes:

[0019] Using the initial A-phase voltage sequence, initial B-phase voltage sequence, and initial C-phase voltage sequence as state vectors, the state equation and observation equation of the multi-channel joint Kalman filter algorithm are established.

[0020] Based on the state equation and the observation equation, state prediction, covariance prediction, Kalman gain calculation, state update and covariance update are performed sequentially, and the prediction results are fused with the current state value to obtain the estimated three-phase voltage values.

[0021] Apply a three-phase summation constraint to normalize and compensate the estimated three-phase voltage values ​​to generate a second voltage sequence.

[0022] Optionally, the step of applying a three-phase summation constraint, normalizing and compensating the estimated three-phase voltage values, and generating a second voltage sequence includes:

[0023] Based on the three-phase summation constraint, the correction error value is calculated as follows:

[0024] ε k =U A +UB +U C -U target

[0025] Where, ε k To correct for error values; U A This is the estimated value of phase A voltage; U B This is the estimated value of phase B voltage; U C This is the estimated value of phase C voltage; U target This is the theoretical sum of the three-phase voltages;

[0026] The correction error is evenly distributed among the three-phase voltage estimates to generate a second voltage sequence:

[0027]

[0028] Among them, U i 输出 For the second voltage sequence; U i This is the estimated value of the three-phase voltage; ε k This is to correct for error values.

[0029] Optionally, the step of time-aligning the second voltage sequence using linear interpolation to generate the first voltage sequence includes:

[0030] Obtain a reference time series; the reference time series is generated based on the sampling rate of the analog-to-digital converter, corresponding to the timestamps of each sampling point in the initial voltage series;

[0031] Calculate the time difference between each sampling point in the second voltage sequence and the reference time sequence;

[0032] Determine whether the time difference exceeds the second threshold;

[0033] If so, then linear interpolation is used to obtain the corresponding voltage value using the timestamp of the reference time series, and the sampling points in the second voltage series are updated;

[0034] If not, then retain the original sampling points in the second voltage sequence.

[0035] Optionally, the step of performing principal component analysis algorithm to identify principal component sequences based on the voltage feature vector set includes:

[0036] The voltage feature vector sets obtained at multiple sampling times are arranged in chronological order to construct the original feature matrix;

[0037] Perform mean-reduction processing on each feature dimension of the original feature matrix to obtain a centered feature matrix;

[0038] Based on the centered feature matrix, the covariance matrix between features is calculated;

[0039] Eigenvalue decomposition is performed on the covariance matrix to obtain a set of eigenvectors and their corresponding eigenvalues ​​arranged in descending order of eigenvalues. The eigenvector corresponding to the principal component with the largest contribution rate is selected as the first principal component vector.

[0040] The centered feature matrix is ​​projected onto the first principal component vector to obtain the principal component sequence.

[0041] Optionally, the step of determining whether the fluctuation amplitude of the principal component sequence exceeds the first threshold includes:

[0042] Set the sliding window length and sliding step size;

[0043] Extract the subsequence corresponding to each sliding window in the principal component sequence;

[0044] Perform linear fitting on the data points of the subsequence to obtain the slope of the fitted line;

[0045] If the absolute value of the slope of n consecutive subsequences is greater than the first threshold, then it is determined that the fluctuation range of the principal component sequence exceeds the first threshold; otherwise, it is determined that the fluctuation range of the principal component sequence does not exceed the first threshold.

[0046] A data processing system for a voltage detection sensor according to a second aspect embodiment of this application includes:

[0047] The acquisition module is used to acquire an initial voltage sequence, which includes an initial A-phase voltage sequence, an initial B-phase voltage sequence, and an initial C-phase voltage sequence. The initial voltage sequence is obtained by acquiring analog voltage signals of the A-phase voltage, B-phase voltage, and C-phase voltage of a three-phase motor through a voltage sensor, and then digitizing them using an analog-to-digital converter at a basic sampling rate and a basic number of quantization bits to obtain discrete voltage values ​​with timestamps.

[0048] The processing module is used to perform data processing on the initial voltage sequence to obtain a first voltage sequence; the data processing includes at least: error correction and time alignment;

[0049] The extraction module is used to perform a wavelet transform algorithm on the first voltage sequence, decompose the first voltage sequence into sub-bands of different frequencies, construct a voltage feature vector set of the sub-bands, and use the voltage feature vector set to analyze the operating state of the three-phase motor.

[0050] The dimensionality reduction module is used to perform principal component analysis algorithm to identify principal component sequences based on the voltage feature vector set; the principal component sequence is the most representative feature sequence extracted from the voltage feature vector set.

[0051] The judgment module is used to determine whether the fluctuation amplitude of the principal component sequence exceeds a first threshold.

[0052] If so, increase the sampling rate and quantization bits of the analog-to-digital converter to reacquire the initial voltage sequence;

[0053] If not, then output the voltage feature vector set.

[0054] An electronic device according to a third aspect of this application includes:

[0055] The processor and memory, the memory being used to store computer programs, and the processor being used to call and run the computer programs stored in the memory to execute the method of the first aspect.

[0056] A computer-readable storage medium according to a fourth aspect of this application is used to store a computer program that causes a computer to perform the method of the first aspect.

[0057] One of the above technical solutions has at least the following advantages or beneficial effects:

[0058] A data processing method for a voltage detection sensor according to an embodiment of this application includes: acquiring an initial voltage sequence, the initial voltage sequence including an initial A-phase voltage sequence, an initial B-phase voltage sequence, and an initial C-phase voltage sequence; the initial voltage sequence is an analog voltage signal of the A-phase voltage, B-phase voltage, and C-phase voltage of a three-phase motor acquired by a voltage sensor, which is then digitized by an analog-to-digital converter at a basic sampling rate and a basic number of quantization bits to obtain discrete voltage values ​​with timestamps; and performing data processing on the initial voltage sequence to obtain a first voltage sequence; the data processing includes at least: error correction and time alignment. A wavelet transform algorithm is applied to the first voltage sequence to decompose it into sub-bands of different frequencies, constructing a voltage feature vector set for each sub-band. This voltage feature vector set is used to analyze the operating state of the three-phase motor. Based on the voltage feature vector set, a principal component analysis algorithm is performed to identify the principal component sequence. The principal component sequence is the most representative feature sequence extracted from the voltage feature vector set. It is determined whether the fluctuation amplitude of the principal component sequence exceeds a first threshold. If so, the sampling rate and quantization bits of the analog-to-digital converter are increased, and the initial voltage sequence is reacquired. If not, the voltage feature vector set is output. This application employs error correction and time consistency processing on the initial data, and significantly improves the adaptability and robustness of voltage data in scenarios with severe fluctuations through a feature-driven sampling rate adjustment mechanism. Compared with traditional static sampling methods, it has intelligent sampling rate control capabilities, lower error sensitivity, and stronger system self-adjustment capabilities.

[0059] The data processing system, electronic device, and storage medium for a voltage detection sensor provided in this application embodiment are equipped with the aforementioned data processing method for the voltage detection sensor. Since the data processing method for the voltage detection sensor has the aforementioned technical effects, the data processing system, electronic device, and storage medium for a voltage detection sensor equipped with the data processing method for the voltage detection sensor should also have corresponding technical effects. Attached Figure Description

[0060] Figure 1 A schematic flowchart of a data processing method for a voltage detection sensor provided in an embodiment of this application is shown;

[0061] Figure 2 A schematic diagram of the error correction process using a multi-channel joint Kalman filter algorithm provided in an embodiment of this application is shown.

[0062] Figure 3 A schematic diagram of the process for time alignment using linear interpolation provided in an embodiment of this application is shown;

[0063] Figure 4 This document illustrates a flowchart of the principal component analysis algorithm for identifying principal component sequences provided in an embodiment of this application.

[0064] Figure 5 This document illustrates a flowchart of an embodiment of the present application for determining whether the fluctuation amplitude of a principal component sequence exceeds a first threshold.

[0065] Figure 6 A schematic diagram of a data processing system for a voltage detection sensor provided in an embodiment of this application is shown;

[0066] Figure 7 A schematic diagram of an electronic device provided in an embodiment of this application is shown. Detailed Implementation

[0067] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0068] As mentioned above, three-phase motors are widely used in industrial automation, energy systems, and manufacturing equipment, and their operating status directly affects the stability and production efficiency of the equipment. Since three-phase motors typically operate under varying loads and speeds, their supply voltage fluctuates, experiences transient disturbances, or becomes unbalanced. Therefore, real-time and accurate monitoring of three-phase motor voltage changes is crucial for achieving equipment status assessment, fault early warning, and intelligent control. Voltage detection technology is not only used to determine power supply stability but also assists in identifying problems such as load changes, electrical faults, and wiring abnormalities. For example, Chinese patent application number 202510000331.3 discloses an improved empirical wavelet transform-based method and system for diagnosing permanent magnet synchronous motor faults. This method uses wavelet filter banks to extract each modal component from each power spectrum interval; it constructs a fault feature vector of the fault signal using the energy torque of each modal component within each power spectrum interval, thus solving the mode aliasing problem caused by unclear signal spectrum boundaries during motor diagnosis.

[0069] However, existing voltage detection methods have the following drawbacks:

[0070] 1. Existing systems rely on a fixed-precision acquisition mode, meaning the sampling rate and quantization bit count of the analog-to-digital converter are fixed, such as using a 10kHz sampling rate and 10-bit quantization bit count. This leads to insufficient feature extraction accuracy when the voltage signal fluctuates rapidly. On the other hand, if high-precision acquisition is consistently used, such as a 20kHz sampling rate and 12-bit quantization bit count, the amount of data acquired will be too large when the voltage signal changes slowly, resulting in slow processing.

[0071] 2. In a practical three-phase motor voltage detection system, phases A, B, and C are sampled by independent voltage sensors. Although these sensors are consistent in nominal parameters, differences in sensor sensitivity and hardware zero-point offset are inevitable due to factors such as manufacturing process errors, aging effects, temperature drift, and inconsistencies in transmission links. At the same time, there may be time delays between different acquisition systems, leading to data distortion and affecting subsequent wavelet analysis.

[0072] Therefore, there is an urgent need for a three-phase voltage detection data processing method with error correction capabilities, a time consistency guarantee mechanism, and intelligent control of sampling accuracy. This method can dynamically model the sensitivity mismatch and zero-point drift caused by manufacturing differences, environmental interference, and aging effects in multi-channel voltage sensors during actual operation, and correct them through joint filtering. Simultaneously, based on real-time analysis of the fluctuation trend of the signal principal components, it intelligently judges the intensity of voltage state changes and adaptively adjusts the sampling frequency and quantization bit depth of the analog-to-digital converter to improve the signal recovery capability under severe disturbances or abnormal conditions.

[0073] To address at least one of the technical problems existing in the prior art or related technologies, this application provides a data processing method and system for a voltage detection sensor. The method includes: acquiring an initial voltage sequence, wherein the initial voltage sequence includes an initial A-phase voltage sequence, an initial B-phase voltage sequence, and an initial C-phase voltage sequence; the initial voltage sequence is an analog voltage signal acquired by a voltage sensor from the A-phase, B-phase, and C-phase voltages of a three-phase motor, which is then digitized using an analog-to-digital converter at a basic sampling rate and a basic quantization bit count to obtain discrete voltage values ​​with timestamps; and performing data processing on the initial voltage sequence to obtain a first voltage sequence; the data processing at least... The method includes: error correction and time alignment; performing wavelet transform on the first voltage sequence to decompose it into sub-bands of different frequencies, constructing a voltage feature vector set for each sub-band, which is used to analyze the operating state of the three-phase motor; based on the voltage feature vector set, performing principal component analysis to identify the principal component sequence; the principal component sequence is the most representative feature sequence extracted from the voltage feature vector set; determining whether the fluctuation amplitude of the principal component sequence exceeds a first threshold; if so, increasing the sampling rate and quantization bits of the analog-to-digital converter to reacquire the initial voltage sequence; if not, outputting the voltage feature vector set. This application employs error correction and time consistency processing on the initial data, and also significantly improves the adaptability and robustness of voltage data in scenarios with severe fluctuations through a feature-driven sampling rate adjustment mechanism. Compared with traditional static sampling methods, it has intelligent sampling rate control capabilities, lower error sensitivity, and stronger system self-adjustment capabilities.

[0074] The following describes, with reference to the accompanying drawings, a data processing method and system for a voltage detection sensor according to some embodiments provided in this application.

[0075] See Figures 1 to 7 The first aspect of this application provides a data processing method for a voltage detection sensor, comprising:

[0076] S100: Obtain the initial voltage sequence, which includes: initial phase A voltage sequence, initial phase B voltage sequence and initial phase C voltage sequence; the initial voltage sequence is obtained by acquiring the analog voltage signals of phase A voltage, phase B voltage and phase C voltage of a three-phase motor through a voltage sensor, and then digitizing them through an analog-to-digital converter at the basic sampling rate and basic quantization bit number to obtain discrete voltage values ​​with timestamps.

[0077] Three independent voltage sensors synchronously acquire the voltage signals of phases A, B, and C of a three-phase motor. These three analog voltage signals are then fed into analog-to-digital converters (ADCs). Synchronous digital sampling is performed according to a set base sampling rate and base quantization depth (e.g., 10kHz base sampling rate and 10-bit base quantization depth), with each sampling point marked with a timestamp. The output initial voltage sequence contains three sets of timestamped discrete voltage values: the initial phase A voltage sequence, the initial phase B voltage sequence, and the initial phase C voltage sequence. These initial voltage sequences serve as the input basis for subsequent error correction, feature extraction, and trend recognition algorithms.

[0078] S200: Perform data processing on the initial voltage sequence to obtain a first voltage sequence; the data processing includes at least: error correction and time alignment.

[0079] In one illustrative embodiment, the step of processing the initial voltage sequence to obtain a first voltage sequence includes:

[0080] S210: The initial voltage sequence is corrected for errors using a multi-channel joint Kalman filter algorithm to generate a second voltage sequence.

[0081] Considering the systematic errors such as sensitivity differences, zero-point offset, and environmental interference in the three-channel voltage sensors in practical applications, a multi-channel joint Kalman filter algorithm is used to dynamically correct the initial voltage sequence, which can significantly reduce the error accumulation caused by sensor inconsistency.

[0082] In one illustrative embodiment, such as Figure 2 As shown, the step of using a multi-channel joint Kalman filter algorithm to correct the error of the initial voltage sequence and generate the second voltage sequence includes:

[0083] S211: Using the initial A-phase voltage sequence, initial B-phase voltage sequence, and initial C-phase voltage sequence as state vectors, establish the state equation and observation equation of the multi-channel joint Kalman filter algorithm.

[0084] The initial phase A voltage sequence, initial phase B voltage sequence, and initial phase C voltage sequence are collectively constructed into a three-dimensional state vector:

[0085]

[0086] Where, x k This represents the state vector of the k-th sampling point.

[0087] Equations of state:

[0088] x k =A·xk-1

[0089] Where, x k x represents the state vector of the k-th sampling point; k-1 A represents the state vector of the previous sampling point; A is the state transition matrix, usually set as the identity matrix I, indicating that the voltage state changes smoothly in a short time, that is, the current state is directly equal to or linearly continues the previous state, which belongs to the static state model.

[0090] Observation equation:

[0091] z k =H·x k

[0092] Among them, z k H is the observation vector composed of actual measured values; H is the observation matrix, usually set as the identity matrix I, which represents the sensor's direct observation state itself, without projection, filtering or transformation.

[0093] S212: Based on the state equation and the observation equation, perform state prediction, covariance prediction, Kalman gain calculation, state update and covariance update in sequence, and fuse the current state value to correct the prediction results to obtain the estimated three-phase voltage values.

[0094] Based on the above model, the following filtering process is performed for each time point k:

[0095] State prediction, used to predict the current voltage:

[0096] x^ k∣k-1 =A·x^ k-1

[0097] Where, x^ k∣k-1 Let x be the predicted state value at the current time k, i.e., the estimated three-phase voltage value; A is the state transition matrix, usually taken as the identity matrix I, representing the stable evolution of voltage over a short period of time; x^ k-1 This is the state estimate from the previous moment.

[0098] Covariance prediction is used to estimate the error range.

[0099] P k∣k-1 =A·P k-1 ·A T +Q

[0100] Among them, P k∣k-1 The prediction error covariance matrix reflects the confidence level of the predicted values; P k-1A represents the uncertainty of the state estimate at the previous moment; A is the state transition matrix, usually taken as the identity matrix I, which represents the steady evolution of voltage over a short period of time; Q is the process noise covariance matrix, which reflects the uncertainty brought about by the natural evolution of the state over time, and can be set using empirical methods.

[0101] Kalman gain calculation, used to harmonize prediction and observation, involves selecting a fusion weight that minimizes the covariance matrix of the estimation error, specifically:

[0102] K k =P k∣k-1 ·H T ·(H·P k∣k-1 ·H T +R) -1

[0103] This formula reflects the trade-off between prediction uncertainty and observation reliability, where K k K represents the Kalman gain, indicating the weight of the current observation's influence on the state correction. k The larger the value, the higher the reliability of the current observation and the larger the correction range; conversely, the smaller the value, the more it relies on historical predictions. H is the observation matrix, usually the identity matrix I, representing the direct observation of the three-phase voltage by the sensor. R is the observation noise covariance matrix, reflecting the uncertainty of the sensor observation, which can be set using empirical methods.

[0104] State updates are used to merge observations for correction.

[0105] x^ k =x^ k∣k-1 +K k ·(z k -H·x^ k∣k-1 )

[0106] Where, x^ k Z represents the optimal state estimate after fusion of observations; H is the observation matrix; z k This represents the actual three-phase voltage observation vector collected at the current moment; (z k -H·x^ k∣k-1 ) represents the observation residual, indicating the deviation between the predicted and actual values.

[0107] Covariance update, used for corrected error estimation:

[0108] P k =(IK k ·H)·P k∣k-1

[0109] Among them, P k K represents the updated state covariance, reflecting the confidence interval of the corrected state estimate. k H is the Kalman gain; H is the observation matrix.

[0110] Kalman filters obtain dynamic estimates of three-phase voltages with minimal mean square error by dynamically balancing predictions and actual observations and fusing historical states with new information.

[0111] S213: Apply a three-phase summation constraint, normalize and compensate the estimated three-phase voltage values, and generate a second voltage sequence.

[0112] After obtaining the estimated values ​​of the three-phase voltage at the current moment through Kalman filtering, physical consistency constraints are further applied to the estimation results to improve the physical rationality and numerical symmetry of the voltage measurement results.

[0113] Under ideal operating conditions, the symmetrical power supply system of a three-phase motor satisfies the following relationship:

[0114] U A (k)+U B (k)+U C (k)≈U target

[0115] Among them, U A (k), U B (k), U C (k) represents the voltage of phase A, phase B, and phase C, respectively; U target Theoretically, this is the sum of the three-phase voltages. For example, in a three-phase balanced system (Y-connection), a 380V line voltage (Y-connection) is widely used in three-phase motors and power distribution systems. Simultaneously, 220V single-phase power can be supplied through the neutral line. Therefore, theoretically, the sum of the three-phase voltages, U, is... target ≈660V.

[0116] Since Kalman filtering may still retain some residual bias during the processing, the three-phase estimates need to be summed and normalized to enhance the symmetry of the estimates.

[0117] In one illustrative embodiment, the step of applying a three-phase summation constraint, normalizing the three-phase voltage estimates, and generating a second voltage sequence includes:

[0118] Based on the three-phase summation constraint, the correction error value is calculated as follows:

[0119] ε k =U A +U B +U C -U target

[0120] Where, ε k To correct for error values; U A This is the estimated value of phase A voltage; U BThis is the estimated value of phase B voltage; U C This is the estimated value of phase C voltage; U target This is the theoretical sum of the three-phase voltages;

[0121] The correction error is evenly distributed among the three-phase voltage estimates to generate a second voltage sequence:

[0122]

[0123] Among them, U i 输出 For the second voltage sequence; U i This is the estimated value of the three-phase voltage; ε k This is to correct for error values.

[0124] For example, taking the three-phase voltage detection of a three-phase motor system under standard operating conditions as an example, the sampling time points are t1, t2, and t3, and the following initial voltage sequence, i.e., the observed values, are collected:

[0125] Initial phase A voltage sequence: t1 is 219.3V, t2 is 218.7V, t3 is 219.5V;

[0126] Initial phase B voltage sequence: t1 is 220.1V, t2 is 220.3V, t3 is 220.0V;

[0127] Initial C-phase voltage sequence: t1 is 221.4V, t2 is 221.0V, and t3 is 220.3V;

[0128] The initial state value of the Kalman filter is set to the observation value at time t1, the state transition matrix A = I, and the observation matrix H = I. Based on the sensor noise experimental data, the covariance is initialized as: P0 = 0.5·I, process noise Q = 0.1·I, and observation noise R = 0.2·I.

[0129] Time t1:

[0130] State vector:

[0131]

[0132] State prediction:

[0133]

[0134] Covariance prediction:

[0135] P 1|0 =P0+Q=0.5+0.1=0.6·I

[0136] Kalman gain calculation:

[0137]

[0138] Status Update:

[0139]

[0140] Since the initial predictions were consistent with the observations, there were no changes in the first iteration.

[0141] Covariance update:

[0142] P1=(1-K1)·P 1|0 = (1-0.75)×0.6 = 0.15·I

[0143] Total Target: U target =660.0V

[0144] Calculate the correction error value:

[0145] ε1=219.3+220.1+221.4-660.0=0.8

[0146] Generate the second voltage sequence at time t1:

[0147]

[0148] Time t2:

[0149] Estimated value:

[0150]

[0151] State prediction:

[0152]

[0153] Covariance prediction:

[0154] P 2|1 =P1+Q=0.15+0.1=0.25·I

[0155] Kalman gain calculation:

[0156]

[0157] Status Update:

[0158]

[0159] Covariance update:

[0160] P2=(1-K2)·P 2|1 = (1-0.556)×0.25 = 0.111·I

[0161] Total Target: U target =660.0V

[0162] Calculate the correction error value:

[0163] ε2=218.966+220.211+221.178-660.0=0.355

[0164] Generate the second voltage sequence at time t2:

[0165]

[0166] Time t3:

[0167] Estimated value:

[0168]

[0169] State prediction:

[0170]

[0171] Covariance prediction:

[0172] P 3|2 =P2+Q=0.111+0.1=0.211·I

[0173] Kalman gain calculation:

[0174]

[0175] Status Update:

[0176]

[0177] Covariance update:

[0178] P3=(1-K3)·P 3|2 = (1-0.513)×0.211≈0.103

[0179] Total Target: U target =660.0V

[0180] Calculate the correction error value:

[0181] ε3=219.240+220.103+220.727-660.0=0.07

[0182] Generate the second voltage sequence at time t3:

[0183]

[0184] The final second voltage sequence is:

[0185] Phase A voltage sequence: t1 is 219.003V, t2 is 218.848V, t3 is 219.217V;

[0186] Phase B voltage sequence: t1 is 219.833V, t2 is 220.093V, and t3 is 220.080V;

[0187] C-phase voltage sequence: t1 is 221.133V, t2 is 221.060V, and t3 is 220.704V.

[0188] In summary, by employing a multi-channel joint Kalman filter model and recursively fusing the state prediction from the previous moment with the current observation residual, interferences such as sensor observation noise, sensitivity differences, and zero-point drift are effectively suppressed. The output three-phase voltage estimates exhibit a more stable and continuous numerical structure than the original observations at multiple moments.

[0189] S220: The second voltage sequence is time-aligned using linear interpolation to generate the first voltage sequence.

[0190] Since the analog-to-digital converter may experience slight clock skew or interruption interference, causing the sampling points to be not fully synchronized, a time alignment process is further performed on the second voltage sequence. This process ensures that all voltage sequences have valid voltage values ​​at a uniform time point, resulting in temporal structural consistency for subsequent wavelet decomposition and feature extraction, and providing high-quality input for subsequent signal feature extraction and analysis.

[0191] In one illustrative embodiment, such as Figure 3 As shown, the step of using linear interpolation to time-align the second voltage sequence and generate the first voltage sequence includes:

[0192] S221: Obtain a reference time series; the reference time series is generated based on the sampling rate of the analog-to-digital converter, with a timestamp corresponding to each sampling point in the initial voltage series.

[0193] The system is based on the sampling rate f of the analog-to-digital converter. s Generate a baseline time series:

[0194]

[0195] Where t0 is the starting sampling timestamp; Δt is the fixed sampling interval; and N is the total number of sampling points.

[0196] The reference time series represents the standard sampling time points that each phase voltage should correspond to under ideal conditions.

[0197] S222: Calculate the time difference between each sampling point in the second voltage sequence and the reference time sequence.

[0198] The timestamp t of each actual sampling point in the second voltage sequence k ′ and the corresponding time stamp t of the reference time series k Time difference δ k calculate:

[0199] δ k =|t′ k -t k |

[0200] S223: Determine whether the time difference exceeds the second threshold;

[0201] If so, then linear interpolation is used to obtain the corresponding voltage value using the timestamp of the reference time series, and the sampling points in the second voltage series are updated;

[0202] If not, then retain the original sampling points in the second voltage sequence.

[0203] Set the maximum allowed offset, i.e., the second threshold δ. max The value is set to 15%-25% of the sampling period. For example, for a 10kHz sample, Δt = 0.1ms, and the second threshold is set to 20% of the sampling period, i.e., δ max =0.02ms.

[0204] If the time difference exceeds the second threshold, then the following condition is met: δ k >δ max If the offset of the sampling point is not negligible, then compensation is required.

[0205] For sampling points that are determined to have exceeded the offset limit, the system performs linear interpolation based on their adjacent time points and corresponding voltage values.

[0206] With t k ′ <t k For example, adjacent time points (t) k ′, t k+1 ′), voltage value (U) k U k+1 The updated value is:

[0207]

[0208] Soon k The voltage at time ′ is reconstructed with respect to the reference time series t. k Consistent interpolation values.

[0209] For sampling points whose time difference does not exceed the second threshold, the original value is retained; for those that exceed the threshold, interpolation is used to update, ultimately forming a first voltage sequence with a unified timestamp structure. This can significantly improve the synchronization and structural regularity of the three-phase voltage sequence in the time dimension, avoid multi-phase feature alignment deviations caused by time misalignment, and is particularly suitable for wavelet analysis, trend modeling, and frequency-sensitive feature extraction algorithms, providing accurate, continuous, and time-consistent input data for subsequent system processing.

[0210] S300: Perform wavelet transform algorithm on the first voltage sequence to decompose the first voltage sequence into sub-bands of different frequencies, and construct voltage feature vector set of the sub-bands. The voltage feature vector set is used to analyze the operating state of the three-phase motor.

[0211] Wavelet transform is a time-frequency analysis method that decomposes a signal into different frequency ranges (called sub-bands). It is suitable for processing non-stationary signals. Unlike Fourier transform, wavelet transform can simultaneously preserve the temporal locality and frequency information of a signal, making it particularly suitable for detecting abrupt changes, perturbations, and trend changes in voltage signals.

[0212] In this embodiment, a Discrete Wavelet Transform (DWT) algorithm can be first applied to the first voltage sequence. A suitable wavelet basis (e.g., Daubechies db4) and a decomposition level (e.g., 5 levels) are selected to decompose the first voltage sequence into multiple sub-band signals at different frequency scales, each including multiple detail components (high-frequency components) and an approximation component (low-frequency trend). This multi-scale decomposition can simultaneously capture high-frequency disturbances, transient impacts, and low-frequency stable changes in the first voltage sequence. Then, for each sub-band signal, representative statistical and frequency features are further extracted to construct a sub-band voltage feature vector. For example, feature dimensions such as root mean square value, energy, kurtosis, skewness, and dominant frequency are extracted from each sub-band level to form a structured, fixed-length feature vector. The feature vectors of all sub-bands are then concatenated to form a complete voltage feature vector set. For example, after 5-level wavelet decomposition, each sampling window can form a vector containing 30 dimensions of features to comprehensively describe the voltage behavior in the time-frequency domain.

[0213] The voltage feature vector set extracted through wavelet transform can effectively cover multiple scenarios, including steady-state identification, long-term trend tracking, and transient anomaly detection, enabling the voltage detection system to have more comprehensive state perception and dynamic response capabilities. For example, by accurately characterizing the steady-state level of voltage through amplitude features, energy distribution, and statistics (such as mean and variance) in the low-frequency subband, it helps determine whether the equipment is in typical states such as standard power supply, off-center load, no-load, or overload, and is widely applicable to the operation monitoring of various equipment such as motors, power supplies, and inverters. By detecting sudden increases in peak value, kurtosis, or abnormal energy fluctuations in the high-frequency subband, the system can quickly respond to fault signals such as spikes, arcs, instantaneous voltage loss, and harmonic impacts within milliseconds, achieving early detection, source location, and automatic reporting of electrical disturbances, significantly improving the system's fault recovery speed and protection accuracy.

[0214] S400: Based on the voltage feature vector set, execute the principal component analysis algorithm to identify the principal component sequence; the principal component sequence is the most representative feature sequence extracted from the voltage feature vector set.

[0215] After completing the wavelet transform and feature extraction of the three-phase voltage signal, the system has obtained a voltage feature vector set for subsequent analysis. However, this voltage feature vector set is extracted based on the default base sampling rate and quantization bit count of the analog-to-digital converter. Therefore, there may be insufficient accuracy issues. For example, if the base sampling rate is set too low (e.g., below 10kHz), when the signal contains high-frequency disturbances (e.g., spikes, arcs, harmonic interference), it may not be able to completely capture short-time high-frequency components, leading to feature distortion or desensitization. The base quantization bit count (e.g., 10 bits) may introduce quantization errors in high dynamic range scenarios, resulting in a decrease in the accuracy of key statistical indicators (e.g., root mean square, voltage energy, kurtosis) in the feature vector, affecting the subsequent model's discrimination ability.

[0216] Therefore, it is necessary to further analyze the changing trend of the current voltage feature vector set to determine whether the current data is sufficient to reflect the true fluctuation characteristics of the voltage operating state. To this end, principal component analysis (PCA) is introduced to extract principal component sequences, and the fluctuation amplitude is used as a criterion to trigger the adaptive adjustment mechanism of the sampling parameters.

[0217] Principal Component Analysis (PCA) is a commonly used technique for dimensionality reduction and feature extraction. Its core idea is to project the original high-dimensional data into a low-dimensional space through linear transformation, ensuring that the projected data maintains the maximum variance in the low-dimensional space, thus preserving the most important information from the original data.

[0218] In one illustrative embodiment, such as Figure 4 As shown, the step of performing principal component analysis algorithm to identify principal component sequences based on the voltage feature vector set includes:

[0219] S410: Arrange the voltage feature vector sets obtained at multiple sampling times in chronological order to construct the original feature matrix.

[0220] The extracted voltage feature vectors from multiple consecutive sampling times, such as M sampling points, are arranged in chronological order to form the original feature matrix X:

[0221]

[0222] Where M is the number of sampling points; N is the number of feature dimensions extracted for each sampling point; each row x i This represents the voltage characteristic vector at a certain moment.

[0223] S420: Perform mean removal processing on each feature dimension of the original feature matrix to obtain a centered feature matrix.

[0224] To eliminate scale differences between features, a mean-removing operation is first performed on each column (feature dimension) of the matrix to obtain a centered matrix X′:

[0225] x′ ij =x ij -μ j ,in

[0226] S430: Based on the centered feature matrix, the covariance matrix between features is calculated.

[0227] Based on the centered feature matrix X′, calculate the covariance matrix C among the features:

[0228]

[0229] The covariance matrix C describes the degree of linear correlation between the feature dimensions.

[0230] S440: Perform eigenvalue decomposition on the covariance matrix to obtain a set of eigenvectors and their corresponding eigenvalues ​​arranged in descending order of eigenvalues. Select the eigenvector corresponding to the principal component with the largest contribution rate as the first principal component vector.

[0231] Perform eigenvalue decomposition on the covariance matrix C:

[0232] C = V·Λ·V -1 Where V = [v1, v2, ..., v N ]

[0233] Where Λ is the eigenvalue diagonal matrix; V is the corresponding eigenvector.

[0234] Arrange the eigenvalues ​​in descending order, with larger values ​​indicating a greater contribution to the overall change in that direction. Select the eigenvector v1 corresponding to the principal eigenvalue with the largest contribution rate as the first principal component vector.

[0235] S450: Project the centered feature matrix onto the first principal component vector to obtain the principal component sequence.

[0236] Project each row of the centered feature matrix X′, i.e., the feature vector at each time step, onto the first principal component vector v1:

[0237] p i =x i ′·v i i = 1, 2, ..., M

[0238] The principal component sequence P is obtained:

[0239] P = {p1, p2, ..., p} M}

[0240] Wherein, P is the principal component sequence; x i ′ represents the row vector in the centered feature matrix X′; v i This is a column vector in the first principal component vector v1. The principal component sequence P is one-dimensional time-series data, representing the core changing trend among the multi-dimensional features, used for subsequent fluctuation amplitude judgment. Retaining the most representative direction of change among the features improves the clarity of the signal's changing trend, providing a reliable dynamic basis for whether to subsequently increase the sampling rate and quantization bit count.

[0241] S500: Determine whether the fluctuation amplitude of the principal component sequence exceeds the first threshold;

[0242] If so, increase the sampling rate and quantization bits of the analog-to-digital converter to reacquire the initial voltage sequence;

[0243] If not, then output the voltage feature vector set.

[0244] After extracting the principal component sequence, it is necessary to determine whether there are significant fluctuations in the principal component sequence, so as to determine whether the current sampling parameters meet the accuracy requirements of voltage state analysis.

[0245] If the fluctuation amplitude of the principal component sequence exceeds the first threshold, it indicates that the current sampling parameters do not meet the accuracy requirements. Therefore, the sampling rate of the analog-to-digital converter is increased from the basic value (e.g., 10kHz) to a high-resolution value (e.g., 20kHz); simultaneously, the quantization bit depth is increased from the basic bit depth (e.g., 10bit) to a high precision (e.g., 12bit); and step S100 is re-executed to acquire a new round of initial voltage sequence to obtain higher quality voltage data. The first threshold can be obtained by statistically analyzing the historical principal component slope distribution, selecting the upper limit of principal component fluctuation under normal conditions as the judgment criterion.

[0246] If the fluctuation amplitude of the principal component sequence does not exceed the first threshold, the current voltage state is considered to be stable, and the basic sampling parameters can be used to directly output the current voltage feature vector set for subsequent analysis or further feature processing.

[0247] In one illustrative embodiment, such as Figure 5 As shown, the step of determining whether the fluctuation amplitude of the principal component sequence exceeds the first threshold includes:

[0248] S510: Set the sliding window length and sliding step size;

[0249] To achieve dynamic analysis of the volatility of the principal component sequences, a sliding window strategy is adopted to evaluate the principal component sequences segment by segment. The sliding window length W represents the number of principal component data points used in each fitting; the sliding step size S represents the interval at which the window moves within the principal component sequences.

[0250] S520: Extract the subsequence corresponding to each sliding window in the principal component sequence;

[0251] For the principal component sequence P, the system extracts several sliding window subsequences with a set step size:

[0252] P i ={p i ,p i+1 ,...p i+W-1}, i = 1, 1 + S, 1 + 2 S, ...

[0253] Among them, P i is the subsequence; W is the sliding window length; S is the sliding step size.

[0254] S530: Perform linear fitting on the data points of the subsequence to obtain the slope of the fitted line;

[0255] For each subsequence P i Perform first-order least squares linear fitting:

[0256] p(t) = k i ·t+b

[0257] Where k i is the fitting slope of the current subsequence, reflecting the rate of change of the principal components during this period; b is the intercept of the fitting line with the vertical axis; t is the relative time position of each point in the window, used to establish the abscissa variable of the fitting function.

[0258] S540: If the absolute value of the slope of n consecutive subsequences is greater than the first threshold, then it is determined that the fluctuation amplitude of the principal component sequence exceeds the first threshold; otherwise, it is determined that the fluctuation amplitude of the principal component sequence does not exceed the first threshold.

[0259] The system for each k i Perform the following judgment: If |k i If the first threshold is reached, the subsequence is recorded as a significant fluctuation segment; the system records the number N of consecutive significant fluctuation segments. cont If n is ≥ n, for example, n is 3, and three consecutive significant fluctuation segments are recorded, then the fluctuation amplitude of the principal component sequence is determined to exceed the first threshold. Otherwise, the current principal component fluctuation is considered to be within an acceptable range, that is, the fluctuation amplitude of the principal component sequence is determined not to exceed the first threshold.

[0260] Sliding trend analysis of principal component sequences can effectively identify structural changes in three-phase voltage characteristics over time. Compared with traditional single-point feature judgments such as mean and standard deviation, slope fitting is better able to characterize trend-type disturbances, such as gradual increases, gradual decreases, and sudden drops, effectively improving the early identification capability of fluctuation states.

[0261] like Figure 6 As shown, a second aspect of this application provides a data processing system for a voltage detection sensor, comprising:

[0262] The acquisition module is used to acquire an initial voltage sequence, which includes an initial A-phase voltage sequence, an initial B-phase voltage sequence, and an initial C-phase voltage sequence. The initial voltage sequence is obtained by acquiring analog voltage signals of the A-phase voltage, B-phase voltage, and C-phase voltage of a three-phase motor through a voltage sensor, and then digitizing them using an analog-to-digital converter at a basic sampling rate and a basic number of quantization bits to obtain discrete voltage values ​​with timestamps.

[0263] The processing module is used to perform data processing on the initial voltage sequence to obtain a first voltage sequence; the data processing includes at least: error correction and time alignment;

[0264] The extraction module is used to perform a wavelet transform algorithm on the first voltage sequence, decompose the first voltage sequence into sub-bands of different frequencies, construct a voltage feature vector set of the sub-bands, and use the voltage feature vector set to analyze the operating state of the three-phase motor.

[0265] The dimensionality reduction module is used to perform principal component analysis algorithm to identify principal component sequences based on the voltage feature vector set; the principal component sequence is the most representative feature sequence extracted from the voltage feature vector set.

[0266] The judgment module is used to determine whether the fluctuation amplitude of the principal component sequence exceeds a first threshold.

[0267] If so, increase the sampling rate and quantization bits of the analog-to-digital converter to reacquire the initial voltage sequence;

[0268] If not, then output the voltage feature vector set.

[0269] like Figure 7 As shown, a third aspect of this application provides an electronic device, including:

[0270] A memory and a processor are provided, wherein the memory stores a computer program and transfers the program code to the processor. In other words, the processor can retrieve and run the computer program from the memory to implement the methods described in the embodiments of this application.

[0271] For example, the processor can be used to execute the above-described method embodiments according to instructions in the computer program.

[0272] In some embodiments of this application, the processor may include, but is not limited to:

[0273] General-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.

[0274] In some embodiments of this application, the memory includes, but is not limited to:

[0275] Volatile memory and / or non-volatile memory. Non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. Volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of RAM are available, such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDR SDRAM), Enhanced Synchronous DRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), and Direct Rambus RAM (DR RAM).

[0276] In some embodiments of this application, the computer program may be divided into one or more modules, which are stored in the memory and executed by the processor to perform the method provided in this application. The one or more modules may be a series of computer program instruction segments capable of performing a specific function, and these instruction segments describe the execution process of the computer program in the controller.

[0277] The electronic device may also include:

[0278] A transceiver that can be connected to the processor or memory.

[0279] The processor can control the transceiver to communicate with other devices; specifically, it can send data to or receive data from other devices. The transceiver may include a transmitter and a receiver. It may further include antennas, which can be one or more.

[0280] It should be understood that the various components in the electronic device are connected through a bus system, which includes a data bus, a power bus, a control bus, and a status signal bus.

[0281] A fourth aspect of this application also provides a computer storage medium storing a computer program thereon, which, when executed by a computer, enables the computer to perform the methods of the above-described method embodiments. Alternatively, one embodiment of this application also provides a computer program product containing instructions that, when executed by a computer, cause the computer to perform the methods of the above-described method embodiments.

[0282] When implemented using software, it can be implemented entirely or partially as a computer program product. This computer program product includes one or more computer instructions. When these computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another via wired (e.g., coaxial cable, fiber optic, Digital Subscriber Line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium accessible to a computer or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., Digital Video Disc (DVD)), or a semiconductor medium (e.g., Solid State Disk (SSD)).

[0283] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments claimed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0284] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or modules may be electrical, mechanical, or other forms.

[0285] It should be noted that the terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented, for example, in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0286] In the description of this specification, the terms "one embodiment," "some embodiments," "embodiment," "exemplary model," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0287] Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make modifications, alterations, substitutions and variations to the above embodiments within the scope of this application.

Claims

1. A data processing method for a voltage detection sensor, characterized in that, include: An initial voltage sequence is obtained, which includes an initial A-phase voltage sequence, an initial B-phase voltage sequence, and an initial C-phase voltage sequence. The initial voltage sequence is obtained by acquiring analog voltage signals of the A-phase voltage, B-phase voltage, and C-phase voltage of a three-phase motor through a voltage sensor, and then digitizing them using an analog-to-digital converter at a basic sampling rate and a basic number of quantization bits to obtain discrete voltage values ​​with timestamps. The initial voltage sequence is processed to obtain a first voltage sequence; the data processing includes at least: error correction and time alignment. A wavelet transform algorithm is performed on the first voltage sequence to decompose the first voltage sequence into sub-bands of different frequencies, and a voltage feature vector set of the sub-bands is constructed. The voltage feature vector set is used to analyze the operating state of the three-phase motor. Based on the voltage feature vector set, a principal component analysis algorithm is executed to identify the principal component sequence; the principal component sequence is the most representative feature sequence extracted from the voltage feature vector set. Determine whether the fluctuation amplitude of the principal component sequence exceeds a first threshold; If so, increase the sampling rate and quantization bits of the analog-to-digital converter to reacquire the initial voltage sequence; If not, then output the voltage feature vector set.

2. The data processing method for a voltage detection sensor as described in claim 1, characterized in that, The step of processing the initial voltage sequence to obtain the first voltage sequence includes: The initial voltage sequence is corrected for errors using a multi-channel joint Kalman filter algorithm to generate a second voltage sequence. The second voltage sequence is time-aligned using linear interpolation to generate the first voltage sequence.

3. The data processing method for a voltage detection sensor as described in claim 2, characterized in that, The step of using a multi-channel joint Kalman filter algorithm to correct the error in the initial voltage sequence and generate the second voltage sequence includes: Using the initial A-phase voltage sequence, initial B-phase voltage sequence, and initial C-phase voltage sequence as state vectors, the state equation and observation equation of the multi-channel joint Kalman filter algorithm are established. Based on the state equation and the observation equation, state prediction, covariance prediction, Kalman gain calculation, state update and covariance update are performed sequentially, and the prediction results are fused with the current state value to obtain the estimated three-phase voltage values. Apply a three-phase summation constraint to normalize and compensate the estimated three-phase voltage values ​​to generate a second voltage sequence.

4. The data processing method for a voltage detection sensor as described in claim 3, characterized in that, The step of applying a three-phase summation constraint, normalizing and compensating the estimated three-phase voltage values, and generating a second voltage sequence includes: Based on the three-phase summation constraint, the correction error value is calculated as follows: ε k =U A +U B +U C -IN target Where, ε k To correct for error values; U A This is the estimated value of phase A voltage; U B This is the estimated value of phase B voltage; U C This is the estimated value of phase C voltage; U target This is the theoretical sum of the three-phase voltages; The correction error value is evenly distributed among the three-phase voltage estimates to generate a second voltage sequence: Among them, U i 输出 For the second voltage sequence; U i This is the estimated value of the three-phase voltage; ε k This is to correct for error values.

5. The data processing method for a voltage detection sensor as described in claim 2, characterized in that, The step of time-aligning the second voltage sequence using linear interpolation to generate the first voltage sequence includes: Obtain a reference time series; the reference time series is generated based on the sampling rate of the analog-to-digital converter, corresponding to the timestamps of each sampling point in the initial voltage series; Calculate the time difference between each sampling point in the second voltage sequence and the reference time sequence; Determine whether the time difference exceeds the second threshold; If so, then linear interpolation is used to obtain the corresponding voltage value using the timestamp of the reference time series, and the sampling points in the second voltage series are updated; If not, then retain the original sampling points in the second voltage sequence.

6. The data processing method for a voltage detection sensor as described in claim 1, characterized in that, The step of performing principal component analysis algorithm to identify principal component sequences based on the voltage feature vector set includes: The voltage feature vector sets obtained at multiple sampling times are arranged in chronological order to construct the original feature matrix; Perform mean-reduction processing on each feature dimension of the original feature matrix to obtain a centered feature matrix; Based on the centered feature matrix, the covariance matrix between features is calculated; Eigenvalue decomposition is performed on the covariance matrix to obtain a set of eigenvectors and their corresponding eigenvalues ​​arranged in descending order of eigenvalues. The eigenvector corresponding to the principal component with the largest contribution rate is selected as the first principal component vector. The centered feature matrix is ​​projected onto the first principal component vector to obtain the principal component sequence.

7. The data processing method for a voltage detection sensor as described in claim 1, characterized in that, The step of determining whether the fluctuation amplitude of the principal component sequence exceeds the first threshold includes: Set the sliding window length and sliding step size; Extract the subsequence corresponding to each sliding window in the principal component sequence; Perform linear fitting on the data points of the subsequence to obtain the slope of the fitted line; If the absolute value of the slope of n consecutive subsequences is greater than the first threshold, then it is determined that the fluctuation range of the principal component sequence exceeds the first threshold; otherwise, it is determined that the fluctuation range of the principal component sequence does not exceed the first threshold.

8. A data processing system for a voltage detection sensor, characterized in that, include: The acquisition module is used to acquire an initial voltage sequence, which includes an initial A-phase voltage sequence, an initial B-phase voltage sequence, and an initial C-phase voltage sequence. The initial voltage sequence is obtained by acquiring analog voltage signals of the A-phase voltage, B-phase voltage, and C-phase voltage of a three-phase motor through a voltage sensor, and then digitizing them using an analog-to-digital converter at a basic sampling rate and a basic quantization bit count to obtain discrete voltage values ​​with timestamps. The processing module is used to perform data processing on the initial voltage sequence to obtain a first voltage sequence; the data processing includes at least: error correction and time alignment; The extraction module is used to perform a wavelet transform algorithm on the first voltage sequence, decompose the first voltage sequence into sub-bands of different frequencies, construct a voltage feature vector set of the sub-bands, and use the voltage feature vector set to analyze the operating state of the three-phase motor. The dimensionality reduction module is used to perform principal component analysis algorithm to identify principal component sequences based on the voltage feature vector set; the principal component sequence is the most representative feature sequence extracted from the voltage feature vector set. The judgment module is used to determine whether the fluctuation amplitude of the principal component sequence exceeds a first threshold. If so, increase the sampling rate and quantization bits of the analog-to-digital converter to reacquire the initial voltage sequence; If not, then output the voltage feature vector set.

9. An electronic device, characterized in that, include: A processor and a memory, the memory being used to store a computer program, the processor being used to invoke and run the computer program stored in the memory to perform the method of any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, Used to store a computer program that causes a computer to perform the method as described in any one of claims 1-7.

Images