Intelligent charging pile power quality monitoring method and system

Abstract

This invention discloses an intelligent charging pile power quality monitoring method and system, belonging to the field of power quality analysis technology. The method includes: charging pile monitoring data acquisition, optimal wavelet basis dynamic optimization, adaptive harmonic parameter extraction, power quality index calculation, and charging pile power quality monitoring. This scheme performs wavelet packet decomposition based on each set of wavelet basis parameters, calculates the energy density of the leaf node subband corresponding to each target harmonic, obtains a comprehensive score for the wavelet basis parameters, selects the optimal wavelet basis parameters, and achieves cycle-by-cycle adaptive optimization of the wavelet basis parameters as the charging power changes. Based on the energy density, it marks the terminating subband and candidate subdivision nodes, constructs an optimal decomposition tree, records the subband to which the target harmonic belongs, and reconstructs the discrete time-domain sampling sequence based on wavelet packet inverse transform to obtain the amplitude and initial phase angle of each target harmonic in voltage and current, thus improving the reliability of the charging pile power quality monitoring results.

Description

Technical Field

[0001] This invention belongs to the field of power quality analysis technology, specifically referring to an intelligent power quality monitoring method and system for charging piles. Background Technology

[0002] The power quality monitoring method for charging piles is a technical solution that uses artificial intelligence technology to dynamically evaluate the power quality of charging piles by collecting their operating signals in real time. This provides scientific data for decision-making by grid operators, charging facility managers, and power users, thereby effectively improving the power quality level of charging piles connected to the grid and ensuring the safe and stable operation of the power grid.

[0003] However, existing power quality monitoring methods for charging piles suffer from several problems. The harmonic spectrum changes dynamically with the charging power, making it difficult to match the actual time-frequency structure. This results in insufficient effectiveness and consistency of feature extraction, limiting the accuracy of power quality measurement and reducing the reliability of monitoring results. Furthermore, existing power quality monitoring methods for charging piles also suffer from the problem that the energy of each harmonic is non-uniformly distributed in the frequency domain and migrates with the charging state. Weak harmonics are easily submerged by noise or mixed with adjacent harmonics, leading to increased deviations in amplitude and phase extraction and insufficient accuracy in power quality monitoring. Summary of the Invention

[0004] To address the above issues and overcome the shortcomings of existing technologies, this invention provides an intelligent charging pile power quality monitoring method and system. Addressing the problem that existing charging pile power quality monitoring methods suffer from dynamic changes in harmonic spectra with charging power, making it difficult to match the actual time-frequency structure and resulting in insufficient effectiveness and consistency of feature extraction, thus limiting the accuracy of power quality measurement and reducing the reliability of monitoring results, this solution extracts voltage digital sequences from the charging pile monitoring dataset. Based on each set of wavelet basis parameters, wavelet packet decomposition is performed, calculating the energy density of the leaf node subband corresponding to each target harmonic. A comprehensive score for the wavelet basis parameters is obtained, and the optimal wavelet basis parameters are selected. This ensures that the selected parameters simultaneously meet the extraction requirements of the fundamental wave and all characteristic harmonics, achieving periodic adaptive optimization of the wavelet basis parameters, ensuring that the decomposition structure always closely matches the current time-frequency structure. This approach improves the accuracy and reliability of power quality measurement and monitoring results by optimizing the frequency distribution of harmonic energy. Addressing the issues in existing charging pile power quality monitoring methods, such as the non-uniform frequency distribution of harmonic energy and its migration with charging status, weak harmonics being easily submerged by noise or mixed with adjacent harmonics, leading to increased amplitude and phase extraction deviations and insufficient power quality monitoring accuracy, this solution performs wavelet packet decomposition based on optimal wavelet basis parameters. Terminating subbands and candidate subdivision nodes are marked according to energy density, constructing an optimal decomposition tree to obtain a list of frequency intervals for all terminating subbands. The subband to which the target harmonic belongs is recorded, and discrete time-domain sampling sequences are reconstructed based on inverse wavelet packet transform. This yields the amplitude and initial phase angle of each target harmonic in voltage and current, supporting accurate calculation of subsequent power quality indicators and improving the reliability of charging pile power quality monitoring results.

[0005] The technical solution adopted by this invention is as follows: This invention provides an intelligent charging pile power quality monitoring method, which includes the following steps:

[0006] Step S1: Data collection for charging pile monitoring;

[0007] Step S2: Dynamic optimization of the optimal wavelet basis;

[0008] Step S3: Adaptive harmonic parameter extraction;

[0009] Step S4: Calculation of power quality indicators;

[0010] Step S5: Power quality monitoring of charging piles.

[0011] Further, in step S1, the charging pile monitoring data acquisition is carried out at the output end of the charging pile by picking up the voltage and current analog signals through voltage Hall sensors and current Hall sensors respectively, and connecting the signals to a dual-channel synchronous analog-to-digital converter. All channels share the same trigger signal to start synchronous sampling, converting the analog signals into digital sequences to obtain voltage digital sequences and current digital sequences. The obtained digital sequences are organized and stored according to the power frequency cycle as units to construct the charging pile monitoring dataset.

[0012] Further, in step S2, the optimal wavelet basis dynamic optimization is to extract the voltage digital sequence of the current power frequency cycle from the charging pile monitoring dataset cycle by cycle, and send the current cycle data into 12 sets of complex Morlet wavelet bases in the candidate wavelet basis parameter library for evaluation; for each set of wavelet basis parameters, a fixed 5-level wavelet packet decomposition is performed on the current cycle data to locate the leaf node subband in which each target harmonic falls, and the energy density is calculated for each leaf node subband to obtain the comprehensive score of the set of wavelet basis parameters. The parameter corresponding to the highest comprehensive score is selected as the optimal wavelet basis parameter for the current cycle.

[0013] Furthermore, in step S3, the adaptive harmonic parameter extraction specifically includes the following steps:

[0014] Step S31: Constructing the optimal decomposition tree; Based on the optimal wavelet basis parameters, perform 5-level wavelet packet decomposition on the voltage digital sequence of the current power frequency cycle layer by layer; After each layer of decomposition is completed, calculate the energy density of the sub-bands of all leaf nodes and calculate their arithmetic mean; Mark the sub-bands with energy densities lower than the arithmetic mean as termination sub-bands; Mark the sub-bands with energy densities not lower than the arithmetic mean as candidate sub-sub-nodes, and obtain the candidate sub-sub-node set; If the number of decomposition layers reaches the maximum number of decomposition layers or the candidate sub-sub-node set is empty, stop the decomposition; otherwise, proceed to the next layer of decomposition and only further decompose the candidate sub-sub-nodes; After the decomposition is completed, the set of all termination sub-bands constitutes the optimal decomposition tree of the current voltage digital sequence;

[0015] Step S32: Target Harmonic Frequency Band Assignment; After the optimal decomposition tree is constructed, obtain a list of frequency intervals for all terminating sub-bands, and calculate the frequency f for each target harmonic. k Search the frequency range list for frequencies that satisfy the lower limit frequency ≤ frequency f. k The sub-band number of the upper limit frequency is recorded as the sub-band to which the target harmonic belongs;

[0016] Step S33: Dual-channel harmonic parameter reconstruction; Using the optimal decomposition tree structure constructed in step S31 as a standard template, read the voltage and current digital sequences of the current power frequency cycle from the charging pile monitoring dataset. Both channels use the same decomposition tree template for wavelet packet decomposition. After decomposition, for each digital sequence, based on the sub-band determined in step S32, extract the wavelet packet coefficient sequence of the corresponding sub-band from the decomposition result. Perform inverse wavelet packet transform on each extracted coefficient sequence to reconstruct the discrete time-domain sampling sequence of each target harmonic. Perform amplitude detection and zero-crossing phase detection on each reconstructed component to obtain the amplitude and initial phase angle of each target harmonic in the current cycle. Summarize the amplitude and initial phase angle of voltage and current on each target harmonic to form a complete set of harmonic parameters.

[0017] Furthermore, in step S4, the power quality index calculation is based on the amplitude and initial phase angle of voltage and current on each target harmonic, superimposing each target harmonic component in frequency order, reconstructing the denoised discrete time-domain waveform sequence, and calculating the various power quality indices of the current power frequency cycle.

[0018] Furthermore, in step S5, the power quality monitoring of the charging pile is based on the various power quality indicators of the current power frequency cycle, which are compared with the corresponding national standard limits item by item, and corresponding monitoring signals are issued according to the comparison results and the changing trend to complete the power quality monitoring of the charging pile.

[0019] The present invention provides an intelligent charging pile power quality monitoring system, including a charging pile monitoring data acquisition module, an optimal wavelet basis dynamic optimization module, an adaptive harmonic parameter extraction module, a power quality index calculation module, and a charging pile power quality monitoring module.

[0020] The charging pile monitoring data acquisition module collects analog voltage and current signals, converts them into digital sequences, obtains digital voltage sequences and digital current sequences, constructs a charging pile monitoring dataset, and sends the data to the optimal wavelet basis dynamic optimization module.

[0021] The optimal wavelet basis dynamic optimization module extracts the voltage digital sequence from the charging pile monitoring dataset, performs wavelet packet decomposition based on each set of wavelet basis parameters, calculates the energy density of the leaf node subband corresponding to each target harmonic, obtains the comprehensive score of the wavelet basis parameters, selects the optimal wavelet basis parameters, and sends the data to the adaptive harmonic parameter extraction module.

[0022] The adaptive harmonic parameter extraction module performs wavelet packet decomposition based on the optimal wavelet basis parameters, constructs the optimal decomposition tree based on energy density-labeled terminating subbands and candidate subdivision nodes, obtains a list of frequency intervals for all terminating subbands, records the subband to which the target harmonic belongs, reconstructs the discrete time-domain sampling sequence based on wavelet packet inverse transform, obtains the amplitude and initial phase angle of each target harmonic of voltage and current, and sends the data to the power quality index calculation module.

[0023] The power quality index calculation module reconstructs the denoised discrete time domain waveform sequence based on the amplitude and initial phase angle of voltage and current on each target harmonic, calculates various power quality indicators, and sends the data to the charging pile power quality monitoring module.

[0024] The charging pile power quality monitoring module sends out corresponding monitoring signals based on various power quality indicators to complete the power quality monitoring of the charging pile.

[0025] The beneficial effects achieved by the present invention using the above solution are as follows:

[0026] (1) In view of the problem that the harmonic spectrum in the existing power quality monitoring method of charging piles changes dynamically with the charging power, making it difficult to match the actual time-frequency structure, resulting in insufficient effectiveness and consistency of feature extraction, which limits the accuracy of power quality measurement and reduces the credibility of monitoring results, this scheme extracts the voltage digital sequence from the charging pile monitoring dataset, performs wavelet packet decomposition based on each set of wavelet basis parameters, comprehensively evaluates the adaptability of candidate parameters to the time-frequency characteristics of the current period, calculates the energy density of the leaf node sub-band corresponding to each target harmonic, quantifies the effectiveness of harmonic extraction into a calculable evaluation index, obtains the comprehensive score of wavelet basis parameters, selects the optimal wavelet basis parameters, and ensures that the selected parameters take into account the extraction requirements of the fundamental wave and each characteristic harmonic, realizes the period-by-period adaptive optimization of wavelet basis parameters, and makes the decomposition structure always closely follow the current spectrum distribution, thereby improving the accuracy of power quality measurement and the credibility of monitoring results.

[0027] (2) In view of the problem that the energy of each harmonic in the existing charging pile power quality monitoring method is not uniformly distributed in the frequency domain and migrates with the charging state, and the weak harmonics are easily submerged by noise or mixed with adjacent harmonics, resulting in increased deviation of amplitude and phase extraction and insufficient power quality monitoring accuracy, this scheme performs wavelet packet decomposition based on the optimal wavelet basis parameters, marks the terminating subband and candidate subdivision nodes according to energy density, constructs the optimal decomposition tree, concentrates the decomposition granularity on the frequency band where harmonic energy is concentrated, and improves the distinguishability of weak harmonic signals; obtains a list of frequency intervals of all terminating subbands, records the subband to which the target harmonic belongs, avoids the mixing of adjacent harmonics or noise components into the reconstruction results, and reduces the frequency band crosstalk error of amplitude and phase extraction; obtains discrete time domain sampling sequence based on wavelet packet inverse transform reconstruction, obtains the amplitude and initial phase angle of each target harmonic of voltage and current, supports the accurate calculation of subsequent power quality indicators, and improves the reliability of charging pile power quality monitoring results. Attached Figure Description

[0028] Figure 1 A flowchart illustrating an intelligent charging pile power quality monitoring method provided by the present invention;

[0029] Figure 2 A schematic diagram of an intelligent charging pile power quality monitoring system provided by the present invention;

[0030] Figure 3 This is a flowchart illustrating step S3.

[0031] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention and do not constitute a limitation thereof. Detailed Implementation

[0032] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0033] In the description of this invention, it should be understood that the terms "upper", "lower", "front", "rear", "left", "right", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0034] Example 1, see Figure 1 The present invention provides an intelligent power quality monitoring method for charging piles, which includes the following steps:

[0035] Step S1: Charging pile monitoring data acquisition; Collect analog voltage and current signals, convert them into digital sequences, obtain digital voltage sequences and digital current sequences, and construct a charging pile monitoring dataset;

[0036] Step S2: Dynamic optimization of optimal wavelet basis; extract the voltage digital sequence from the charging pile monitoring dataset, perform wavelet packet decomposition based on each set of wavelet basis parameters, calculate the energy density of the leaf node subband corresponding to each target harmonic, obtain the comprehensive score of the wavelet basis parameters, and select the optimal wavelet basis parameters;

[0037] Step S3: Adaptive harmonic parameter extraction; perform wavelet packet decomposition based on the optimal wavelet basis parameters, mark the terminating sub-band and candidate subdivision nodes according to energy density, construct the optimal decomposition tree, obtain the frequency range list of all terminating sub-bands, record the sub-band to which the target harmonic belongs, reconstruct the discrete time domain sampling sequence based on wavelet packet inverse transform, and obtain the amplitude and initial phase angle of each target harmonic of voltage and current.

[0038] Step S4: Calculation of power quality indicators; Based on the amplitude and initial phase angle of voltage and current on each target harmonic, reconstruct the denoised discrete time-domain waveform sequence and calculate various power quality indicators;

[0039] Step S5: Power quality monitoring of charging piles; Based on various power quality indicators, corresponding monitoring signals are issued to complete the power quality monitoring of charging piles.

[0040] Example 2, see Figure 1This embodiment is based on the above embodiment. In step S1, the charging pile monitoring data acquisition is carried out at the output end of the charging pile by picking up the voltage and current analog signals through voltage Hall sensors and current Hall sensors respectively. The signals are uniformly connected to a dual-channel synchronous analog-to-digital converter (ADC). All channels share the same trigger signal to start synchronous sampling, converting the analog signals into digital sequences to obtain voltage digital sequences and current digital sequences. The sampling rate is set to 12800Hz, that is, 256 sampling points are collected in each power frequency cycle (20ms). The obtained digital sequences are organized and stored in units according to the power frequency cycle to construct the charging pile monitoring dataset.

[0041] Example 3, see Figure 1 This embodiment is based on the above embodiment. In step S2, the optimal wavelet basis dynamic optimization is aimed at the fact that during the operation of the charging pile, the harmonic spectrum changes dynamically with the charging power, and the fixed-parameter wavelet basis is difficult to adapt to all operating conditions. From the charging pile monitoring dataset constructed in step S1, the voltage digital sequence of the current power frequency cycle is extracted cycle by cycle, and the current cycle data is sent to 12 sets of complex Morlet wavelet bases in the candidate wavelet basis parameter library for evaluation. The candidate wavelet basis parameter library is constructed using two-dimensional uniform grid sampling, where the discrete values ​​of the center frequency cover the interval [0.6, 1.8], with a total of 3 sampling points, and the discrete values ​​of the bandwidth cover the interval [0.5, 2.0], with a total of 4 sampling points. The two are cross-combined to form all 12 wavelet basis parameters. For each set of wavelet basis parameters, a fixed 5-level wavelet packet decomposition is performed on the current cycle data. After decomposition, each node in the 5th level is called a leaf node. For a sub-band with a fixed frequency width, the leaf nodes of the 5th layer are numbered in ascending order of frequency to locate the leaf node sub-band into which each target harmonic falls. The target harmonics include the fundamental (50Hz), 5th (250Hz), 7th (350Hz), 11th (550Hz), and 13th (650Hz) harmonics. For each leaf node sub-band, the ratio of the sum of squares of the wavelet packet coefficients within the sub-band to the frequency bandwidth of that sub-band is calculated as the energy density of that sub-band. The energy densities of all leaf node sub-bands are then summed to obtain the comprehensive score of the wavelet basis parameters. After traversing all 12 candidate wavelet bases, the parameters corresponding to the highest comprehensive score are selected as the optimal wavelet basis parameters for the current period. This strategy directly utilizes the degree of matching of the wavelet base to the time-frequency structure of the signal to complete the base selection. At the same time, using the charging pile harmonics as the target guide ensures the effectiveness of the base selection for subsequent harmonic extraction tasks. The formula used is as follows:

[0042] ;

[0043] ;

[0044] In the formula, ρ θIt is a comprehensive score when using the θ-th group of wavelet basis parameters. Its physical meaning is the sum of the energy densities of all target harmonics in the sub-bands. This value reflects the degree to which the target harmonic energy is concentrated in each leaf node sub-band by the wavelet basis. The larger the value, the higher the matching degree between the wavelet basis and the current signal harmonic structure. is the energy density of the subband containing the k-th target harmonic when using the θ-th set of wavelet basis parameters, representing the signal energy concentrated within a unit frequency width; θ is the parameter group index in the candidate wavelet basis parameter library. It is the complete set of candidate wavelet basis parameter libraries; It is the set of target harmonic order indices. Where 1 represents the 50Hz fundamental frequency, 5 represents 250Hz (5th harmonic), 7 represents 350Hz (7th harmonic), 11 represents 550Hz (11th harmonic), and 13 represents 650Hz (13th harmonic); k is the target harmonic order index. It is the frequency bandwidth of the leaf node subband where the k-th target harmonic falls after 5-level wavelet packet decomposition using the θ-th wavelet basis. f s J is the sampling rate, J is the number of decomposition levels, and n is the index of the wavelet packet coefficients within the subband of the leaf node. It is the coefficient of the nth wavelet packet within the subband containing the kth target harmonic after performing 5-level wavelet packet decomposition using the θth wavelet basis; θ opt These are the optimal wavelet basis parameters selected for the current power frequency cycle.

[0045] By performing the above operations, this solution addresses the problem in existing charging pile power quality monitoring methods where the harmonic spectrum dynamically changes with charging power, making it difficult to match the actual time-frequency structure. This results in insufficient effectiveness and consistency of feature extraction, limiting the accuracy of power quality measurement and reducing the reliability of monitoring results. The solution extracts voltage digital sequences from the charging pile monitoring dataset, performs wavelet packet decomposition based on each set of wavelet basis parameters, and comprehensively evaluates the adaptability of candidate parameters to the current period's time-frequency characteristics. It calculates the energy density of the leaf node sub-band corresponding to each target harmonic, quantifying the effectiveness of harmonic extraction into a calculable evaluation index. A comprehensive score for the wavelet basis parameters is obtained, and the optimal wavelet basis parameters are selected. This ensures that the selected parameters simultaneously meet the extraction requirements of the fundamental wave and all characteristic harmonics, achieving period-by-period adaptive optimization of the wavelet basis parameters. This ensures that the decomposition structure always closely follows the current spectral distribution, improving the accuracy of power quality measurement and the reliability of monitoring results.

[0046] Example 4, see Figure 1 and Figure 3 This embodiment is based on the above embodiment. In step S3, the adaptive harmonic parameter extraction specifically includes the following steps:

[0047] Step S31: Constructing the optimal decomposition tree; The harmonic energy of each charging pile exhibits a non-uniform distribution in the frequency domain and migrates with changes in charging state, requiring a decomposition structure that adaptively adjusts to the energy distribution; Based on the optimal wavelet basis parameters selected in step S2, the voltage digital sequence of the current power frequency cycle is decomposed into 5 layers of wavelet packets, starting from layer 1; After each layer of decomposition is completed, the energy density of the sub-bands of all leaf nodes is calculated, and their arithmetic mean is calculated; Sub-bands with energy densities lower than the arithmetic mean are marked as termination sub-bands and will not participate in subsequent decompositions; Sub-bands with energy densities not lower than the arithmetic mean are marked as termination sub-bands. The sub-bands with the arithmetic mean are marked as candidate subdivision nodes, allowing further decomposition to obtain a set of candidate subdivision nodes. Decomposition stops if the maximum number of decomposition levels is reached or the set of candidate subdivision nodes is empty; otherwise, the next level of decomposition is entered, and only the candidate subdivision nodes are further decomposed. After decomposition, the set of all terminating sub-bands constitutes the optimal decomposition tree for the current voltage digital sequence. This decomposition tree adaptively changes with the energy distribution of the voltage digital sequence, prioritizing computational resources to the frequency bands richest in harmonic information while avoiding over-decomposition of noise-dominated frequency bands. The formula used is as follows:

[0048] ;

[0049] in, It is the energy density of the subband of the m-th leaf node in the j-th layer; j is the wavelet packet decomposition layer index, and m is the leaf node subband index; It is the frequency bandwidth of the sub-band of the m-th leaf node in the j-th layer; It is the coefficient of the nth wavelet packet within the subband of the mth leaf node in the jth layer;

[0050] Step S32: Target harmonic frequency band assignment; After dynamic decomposition, the positions and widths of each terminating sub-band on the frequency axis are inconsistent. It is necessary to accurately locate the sub-band into which each target harmonic falls in order to extract the corresponding component; After the optimal decomposition tree is constructed, obtain a list of frequency intervals for all terminating sub-bands. For each target harmonic, calculate its frequency f. k =k×50Hz, see the frequency range list. Search for frequencies that satisfy the lower limit frequency ≤ frequency f k The sub-band number of the upper limit frequency is recorded as the sub-band to which the target harmonic belongs. k It can accurately determine the position of each target harmonic in the dynamic decomposition structure, effectively avoiding crosstalk between adjacent harmonics or noise components; among which, and These are the lower and upper frequency limits of the q-th subband, respectively. q is the node number of the q-th sub-band, q is the index of the terminating sub-band, and Q is the total number of terminating sub-bands;

[0051] Step S33: Dual-channel harmonic parameter reconstruction; Power quality monitoring requires complete harmonic parameters for voltage and current, and must ensure the physical consistency of the phase relationship between each harmonic of voltage and current; Using the optimal decomposition tree structure constructed in step S31 as a standard template, read the voltage and current digital sequences of the current power frequency cycle from the charging pile monitoring dataset. Both channels use the same decomposition tree template for wavelet packet decomposition, that is, the decomposition structure, number of decomposition layers, and leaf node partitioning method of the two channels are completely identical; After decomposition, for each digital sequence, according to the sub-band determined in step S32, extract the wavelet packet coefficient sequence of the corresponding sub-band from the decomposition results. Each extracted coefficient sequence is subjected to inverse wavelet packet transform to reconstruct the discrete-time sampling sequence of each target harmonic. During single-branch reconstruction, all wavelet packet coefficients outside the assigned sub-band are set to zero to ensure that the reconstructed component contains only the frequency components of that sub-band. Amplitude and zero-crossing phase detection are performed on each reconstructed component to obtain the amplitude and initial phase angle of each target harmonic in the current period. The amplitudes and initial phase angles of voltage and current on each target harmonic are summarized to form a complete set of harmonic parameters. The digital sequences share the same decomposition tree to ensure that the frequency bands of each harmonic of voltage and current strictly correspond, and their phase relationship is not affected by the time delay differences introduced by different decomposition structures. The formulas used are as follows:

[0052] ;

[0053] ;

[0054] ;

[0055] In the formula, It is the discrete time-domain sampling sequence after the reconstruction of the kth target harmonic of the p-th channel digital sequence; p is the channel index, p∈{1,2}, p=1 represents the voltage channel, p=2 represents the current channel; It is the digital sequence of channel p in its assigned subband. k The wavelet packet coefficient sequence within; It is a wavelet packet inverse transform reconstruction operator that performs an inverse wavelet packet transform on the wavelet packet coefficient sequence to restore it from the transform domain to the time domain waveform sequence. It is the root mean square value operator; t0 is the amplitude of the kth target harmonic in the p-th channel digital sequence; t0 is the zero-crossing reference time of the fundamental component, which is the time corresponding to the first zero-crossing point of the reconstructed fundamental waveform that changes from negative to positive within this period. It is the discrete Hilbert transform operator; construct its analytic signal. It is an operator for calculating the complex argument angle, taking the phase value of the analytic signal at the reference time t0 as the initial phase angle of the harmonic; It is the initial phase angle of the k-th target harmonic in the p-th channel digital sequence.

[0056] By performing the above operations, this solution addresses the problems in existing charging pile power quality monitoring methods, such as the non-uniform distribution of harmonic energy in the frequency domain, its migration with charging status, and the susceptibility of weak harmonics to noise or superposition with adjacent harmonics, leading to increased amplitude and phase extraction deviations and insufficient power quality monitoring accuracy. This solution performs wavelet packet decomposition based on optimal wavelet basis parameters, marks the terminating subbands and candidate subdivision nodes according to energy density, constructs an optimal decomposition tree, and concentrates the decomposition granularity on the frequency bands where harmonic energy is concentrated, improving the distinguishability of weak harmonic signals. It obtains a list of frequency intervals for all terminating subbands, records the subband to which the target harmonic belongs, avoids the mixing of adjacent harmonics or noise components into the reconstruction results, and reduces the frequency band crosstalk error in amplitude and phase extraction. Based on the inverse wavelet packet transform, it reconstructs a discrete time-domain sampling sequence, obtaining the amplitude and initial phase angle of each target harmonic in voltage and current, supporting the accurate calculation of subsequent power quality indicators and improving the reliability of charging pile power quality monitoring results.

[0057] Example 5, see Figure 1 This embodiment is based on the above embodiment. In step S4, the power quality index calculation is for charging piles as nonlinear loads. Their output waveforms contain broadband noise in addition to characteristic harmonics. Directly calculating the power quality index will introduce significant errors. Therefore, it needs to be calculated after reconstructing a clean waveform to ensure the accuracy of the evaluation results. Based on the amplitude and initial phase angle of the voltage and current on each target harmonic obtained in step S3, the target harmonic components are superimposed in frequency order to reconstruct a denoised discrete-time waveform sequence. During reconstruction, only the frequency band components of each target harmonic are included; the remaining unselected terminating sub-bands are not included in the reconstruction, thus achieving a denoising effect. Based on the reconstructed discrete-time waveform sequence, various power quality indices for the current power frequency cycle are calculated. The power quality indices include the total harmonic distortion rate of voltage, the total harmonic distortion rate of current, and the power factor. During reconstruction, pure noise sub-bands are excluded to achieve waveform denoising, and all comprehensive indices required for power quality evaluation are calculated simultaneously to ensure that the indices have a unified source, clear physical meaning, and are not affected by noise. The formulas used are as follows:

[0058] ;

[0059] ;

[0060] ;

[0061] ;

[0062] In the formula, It is the instantaneous value of the reconstructed discrete-time waveform sequence of the p-th channel at the r-th sampling point; r is the sampling point index. N is the number of sampling points per cycle, and f0 is the fundamental frequency; It is the effective value of the k-th target harmonic in the p-th channel; and These are the effective values ​​of the kth target harmonics of voltage and current, respectively; It is the total effective value of the p-th channel; and These are the total effective value of voltage and the total effective value of current, respectively; THD p It is the total harmonic distortion rate of the p-th channel; It is the effective value of the fundamental frequency of the p-th channel. , PF is the amplitude of the fundamental frequency of the digital sequence in channel p; PF is the power factor. and These are the initial phase angles of the kth target harmonic of the voltage and current digital sequences, respectively.

[0063] Example 6, see Figure 1 This embodiment is based on the above embodiment. In step S5, the power quality monitoring of the charging pile is based on the power quality indicators of the current power frequency cycle calculated in step S4. These indicators are compared with the corresponding national standard limits item by item, and corresponding monitoring signals are issued according to the comparison results and changing trends to complete the power quality monitoring of the charging pile. If any power quality indicator value in the current cycle exceeds its corresponding national standard limit (the national standard limit for total harmonic distortion of voltage and total harmonic distortion of current is 5%, and the national standard limit for power factor is 0.90), an alarm signal is immediately issued to remind the maintenance personnel that the current power quality indicator has exceeded the standard and needs to be investigated and dealt with immediately. If none of the power quality indicators in the current cycle exceed their corresponding national standard limits, then each power quality indicator is processed separately. The trend prediction method involves performing a first-order linear fit on historical data from the most recent 60 periods of the indicator, calculating the fitting slope and intercept, and extrapolating to predict the indicator value one second from now. If the predicted value of the total harmonic distortion (THD) of voltage or current exceeds 4.5%, or the predicted value of the power factor is below 0.99, an early warning signal is issued, prompting maintenance personnel to pay attention to the changing trend of the power quality indicator and prepare accordingly. If all power quality indicators in the current period do not exceed their corresponding national standard limits, and the predicted values ​​of THD and THD do not exceed 4.5%, and the predicted value of the power factor is not lower than 0.99, a normal signal is issued, indicating that the current power quality is within a controllable range and the equipment is operating well. The formula used is as follows:

[0064] ;

[0065] ;

[0066] ;

[0067] In the formula, This is the predicted value of the s-th power quality indicator, where s is the power quality indicator index, and a... s and b s These are the fitting slope and fitting intercept of the s-th power quality index, respectively, and t cur It is the current cycle number. It is the number of leading cycles predicted. W is the fitting window length, W=60, t c It is the cycle number of the c-th historical cycle, where c is the historical cycle index. It is the measured value of the s-th power quality indicator in the c-th cycle.

[0068] Example 7, see Figure 2 Based on the above embodiments, this embodiment provides an intelligent charging pile power quality monitoring system, including a charging pile monitoring data acquisition module, an optimal wavelet basis dynamic optimization module, an adaptive harmonic parameter extraction module, a power quality index calculation module, and a charging pile power quality monitoring module.

[0069] The charging pile monitoring data acquisition module collects analog voltage and current signals, converts them into digital sequences, obtains digital voltage sequences and digital current sequences, constructs a charging pile monitoring dataset, and sends the data to the optimal wavelet basis dynamic optimization module.

[0070] The optimal wavelet basis dynamic optimization module extracts the voltage digital sequence from the charging pile monitoring dataset, performs wavelet packet decomposition based on each set of wavelet basis parameters, calculates the energy density of the leaf node subband corresponding to each target harmonic, obtains the comprehensive score of the wavelet basis parameters, selects the optimal wavelet basis parameters, and sends the data to the adaptive harmonic parameter extraction module.

[0071] The adaptive harmonic parameter extraction module performs wavelet packet decomposition based on the optimal wavelet basis parameters, constructs the optimal decomposition tree based on energy density-labeled terminating subbands and candidate subdivision nodes, obtains a list of frequency intervals for all terminating subbands, records the subband to which the target harmonic belongs, reconstructs the discrete time-domain sampling sequence based on wavelet packet inverse transform, obtains the amplitude and initial phase angle of each target harmonic of voltage and current, and sends the data to the power quality index calculation module.

[0072] The power quality index calculation module reconstructs the denoised discrete time domain waveform sequence based on the amplitude and initial phase angle of voltage and current on each target harmonic, calculates various power quality indicators, and sends the data to the charging pile power quality monitoring module.

[0073] The charging pile power quality monitoring module sends out corresponding monitoring signals based on various power quality indicators to complete the power quality monitoring of the charging pile.

[0074] It should be noted that, in this document, the terms “comprising,” “including,” or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0075] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention.

[0076] The present invention and its embodiments have been described above. This description is not restrictive, and the accompanying drawings are only one embodiment of the present invention; the actual structure is not limited thereto. In conclusion, if those skilled in the art are inspired by this description and design similar structures and embodiments without departing from the spirit of the invention, such designs should fall within the protection scope of the present invention.

Claims

1. An intelligent charging pile power quality monitoring method, characterized in that: The method includes the following steps: Step S1: Charging pile monitoring data acquisition; Collect voltage and current analog signals, convert them into digital sequences, obtain voltage digital sequences and current digital sequences, and construct a charging pile monitoring dataset; Step S2: Dynamic optimization of optimal wavelet basis; extract the voltage digital sequence from the charging pile monitoring dataset, perform wavelet packet decomposition based on each set of wavelet basis parameters, calculate the energy density of the leaf node subband corresponding to each target harmonic, obtain the comprehensive score of the wavelet basis parameters, and select the optimal wavelet basis parameters; Step S3: Adaptive harmonic parameter extraction; perform wavelet packet decomposition based on the optimal wavelet basis parameters, mark the terminating sub-band and candidate subdivision nodes according to energy density, construct the optimal decomposition tree, obtain the frequency range list of all terminating sub-bands, record the sub-band to which the target harmonic belongs, reconstruct the discrete time domain sampling sequence based on wavelet packet inverse transform, and obtain the amplitude and initial phase angle of each target harmonic of voltage and current. Step S4: Calculation of power quality indicators; Based on the amplitude and initial phase angle of voltage and current on each target harmonic, reconstruct the denoised discrete time-domain waveform sequence and calculate various power quality indicators; Step S5: Power quality monitoring of charging piles; Based on various power quality indicators, corresponding monitoring signals are issued to complete the power quality monitoring of charging piles. 2.The intelligent charging pile electric energy quality monitoring method according to claim 1, characterized in that: In step S2, the optimal wavelet basis dynamic optimization involves extracting the voltage digital sequence of the current power frequency cycle from the charging pile monitoring dataset cycle by cycle, and sending the current cycle data into 12 sets of complex Morlet wavelet bases in the candidate wavelet basis parameter library for evaluation. For each set of wavelet basis parameters, a fixed 5-level wavelet packet decomposition is performed on the current cycle data to locate the leaf node subband into which each target harmonic falls. For each leaf node subband, the energy density is calculated to obtain the comprehensive score of the set of wavelet basis parameters. The parameter corresponding to the highest comprehensive score is selected as the optimal wavelet basis parameter for the current cycle. 3.The intelligent charging pile power quality monitoring method of claim 2, characterized in that: In step S3, the adaptive harmonic parameter extraction specifically includes the following steps: Step S31: Constructing the optimal decomposition tree; Based on the optimal wavelet basis parameters, perform 5-level wavelet packet decomposition on the voltage digital sequence of the current power frequency cycle layer by layer; After each layer of decomposition is completed, calculate the energy density of the sub-bands of all leaf nodes and calculate their arithmetic mean; Mark the sub-bands with energy densities lower than the arithmetic mean as termination sub-bands; Mark the sub-bands with energy densities not lower than the arithmetic mean as candidate sub-sub-nodes, and obtain the candidate sub-sub-node set; If the number of decomposition layers reaches the maximum number of decomposition layers or the candidate sub-sub-node set is empty, stop the decomposition; otherwise, proceed to the next layer of decomposition and only further decompose the candidate sub-sub-nodes; After the decomposition is completed, the set of all termination sub-bands constitutes the optimal decomposition tree of the current voltage digital sequence; Step S32: Target harmonic frequency band assignment; Step S33: Dual-channel harmonic parameter reconstruction. 4.The intelligent charging pile power quality monitoring method of claim 3, characterized in that: In step S32, the target harmonic frequency band assignment is determined after the optimal decomposition tree is constructed, by obtaining a list of frequency intervals for all terminating sub-bands, and for each target harmonic, calculating its frequency f. k Search the frequency range list for frequencies that satisfy the lower limit frequency ≤ frequency f. k The sub-band number of the upper limit frequency is recorded as the sub-band to which the target harmonic belongs. 5.The intelligent charging pile power quality monitoring method of claim 4, characterized in that: In step S33, the dual-channel harmonic parameter reconstruction uses the optimal decomposition tree structure constructed in step S31 as a standard template. The voltage and current digital sequences for the current power frequency cycle are read from the charging pile monitoring dataset. Both channels use the same decomposition tree template for wavelet packet decomposition. After decomposition, for each digital sequence, based on the assigned sub-band determined in step S32, the wavelet packet coefficient sequence of the corresponding sub-band is extracted from the decomposition result. Inverse wavelet packet transform is performed on each extracted coefficient sequence to reconstruct the discrete-time sampling sequence of each target harmonic. Amplitude detection and zero-crossing phase detection are performed on each reconstructed component to obtain the amplitude and initial phase angle of each target harmonic in the current cycle. The amplitudes and initial phase angles of voltage and current on each target harmonic are summarized to form a complete set of harmonic parameters. 6.The intelligent charging pile power quality monitoring method of claim 5, characterized in that: In step S4, the power quality index is calculated by superimposing the target harmonic components in frequency order based on the amplitude and initial phase angle of voltage and current on each target harmonic, reconstructing the denoised discrete time-domain waveform sequence, and calculating the power quality index of the current power frequency cycle.

7. The intelligent charging pile power quality monitoring method according to claim 6, characterized in that: In step S1, the charging pile monitoring data acquisition is carried out at the output end of the charging pile by picking up the voltage and current analog signals through voltage Hall sensors and current Hall sensors respectively, and connecting the signals to a dual-channel synchronous analog-to-digital converter. All channels share the same trigger signal to start synchronous sampling, converting the analog signals into digital sequences to obtain voltage digital sequences and current digital sequences. The obtained digital sequences are organized and stored according to the power frequency cycle as units to construct the charging pile monitoring dataset.

8. The intelligent charging pile power quality monitoring method according to claim 7, characterized in that: In step S5, the power quality monitoring of the charging pile is carried out by comparing each power quality indicator of the current power frequency cycle with the corresponding national standard limit, and issuing corresponding monitoring signals according to the comparison results and the changing trend to complete the power quality monitoring of the charging pile.

9. An intelligent charging pile power quality monitoring system, used to implement the intelligent charging pile power quality monitoring method as described in any one of claims 1-8, characterized in that: It includes a charging pile monitoring data acquisition module, an optimal wavelet basis dynamic optimization module, an adaptive harmonic parameter extraction module, a power quality index calculation module, and a charging pile power quality monitoring module; The charging pile monitoring data acquisition module collects analog voltage and current signals, converts them into digital sequences, obtains digital voltage sequences and digital current sequences, constructs a charging pile monitoring dataset, and sends the data to the optimal wavelet basis dynamic optimization module. The optimal wavelet basis dynamic optimization module extracts the voltage digital sequence from the charging pile monitoring dataset, performs wavelet packet decomposition based on each set of wavelet basis parameters, calculates the energy density of the leaf node subband corresponding to each target harmonic, obtains the comprehensive score of the wavelet basis parameters, selects the optimal wavelet basis parameters, and sends the data to the adaptive harmonic parameter extraction module. The adaptive harmonic parameter extraction module performs wavelet packet decomposition based on the optimal wavelet basis parameters, constructs the optimal decomposition tree based on energy density-labeled terminating subbands and candidate subdivision nodes, obtains a list of frequency intervals for all terminating subbands, records the subband to which the target harmonic belongs, reconstructs the discrete time-domain sampling sequence based on wavelet packet inverse transform, obtains the amplitude and initial phase angle of each target harmonic of voltage and current, and sends the data to the power quality index calculation module. The power quality index calculation module reconstructs the denoised discrete time domain waveform sequence based on the amplitude and initial phase angle of voltage and current on each target harmonic, calculates various power quality indicators, and sends the data to the charging pile power quality monitoring module. The charging pile power quality monitoring module sends out corresponding monitoring signals based on various power quality indicators to complete the power quality monitoring of the charging pile.

Images