A direct current charging-oriented charging pile voltage stability control method

Abstract

The application discloses a charging pile voltage stability control method for direct current charging and relates to the technical field of grid-connected voltage control. The method comprises the following steps: starting continuous sampling of coupling point voltage and injected current and establishing a circular buffer to generate an initial detection time slot index, performing conflict detection and resolution, and triggering data interception of a collection window; generating a discrete angular frequency point set, obtaining a passive margin sample value sequence, and generating a passive margin function; generating a passive margin target, writing a consistent iteration initial value, performing wave quantity coding interaction and consistent iteration, ruling on a unique master pile, generating a master pile identifier and shaping task division; generating a guaranteed admittance increment, superimposing a unique master pile to generate a supplementary admittance increment, generating a compensation current, performing frequency domain criterion checking and triggering closed-loop updating, and improving controllability and scalability of parallel operation; and keeping the coupling point voltage stable under working condition disturbance and realizing self-adaptive maintenance.

Description

Technical Field

[0001] This invention relates to the field of grid-connected voltage control technology, and in particular to a method for stabilizing the voltage of charging piles for DC charging. Background Technology

[0002] DC fast charging piles are connected to the distribution network coupling point through rectification and power conversion stages. In scenarios with multiple piles connected in parallel and rapid changes in vehicle-end load, voltage stability and power quality control at the coupling point become crucial for grid-connected operation. In existing engineering practices, a closed-loop control strategy combining an outer voltage loop and an inner current loop is typically adopted. This strategy combines droop characteristics, feedforward compensation, and virtual impedance to shape the dynamic equivalent admittance. By sharing the operating status through communication, power distribution and stable operation among multiple piles can be achieved.

[0003] However, conventional practices still face challenges in parameter tuning and multi-pile coordination. On the one hand, control parameters are mostly set based on typical grid-side impedance. When coupling conditions and the number of parallel connections change, the key resonant frequency band is prone to drift, making it difficult to maintain passive margin in the long term. On the other hand, multi-pile coordination often relies on centralized coordination or preset priorities, making it difficult to achieve distributed consistency in the division of labor for detection scheduling, main pile decision-making, and admittance shaping, thus limiting online adaptive updates. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a voltage stability control method for charging piles oriented towards DC charging, which solves the problems in the prior art where it is difficult to adaptively maintain passive margin in key resonant frequency bands and where it is difficult to keep the collaborative admittance shaping and online update mechanism consistent when multiple charging piles are running in parallel.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: This invention provides a voltage stability control method for charging piles oriented towards DC charging, comprising: initiating continuous sampling of coupling point voltage and injected current and establishing a cyclic buffer; calculating the number of detection time slots; generating an initial detection time slot index; performing conflict detection and resolution; solidifying the final detection time slot index into a schedule; and triggering data interception of the acquisition window; receiving acquisition window data; generating a discrete angular frequency point set; obtaining a passive margin sampling value sequence; determining the dominant peak and key resonant frequency band boundary; generating a passive margin function; generating a passive margin target based on the passive margin sampling value sequence and key resonant frequency band boundary; writing the initial value of the consistency iteration; performing wave quantity coding interaction and consistency iteration; determining the unique master pile; generating a master pile identifier and shaping task allocation; based on the master pile identifier and shaping task allocation, combined with the passive margin function, generating a minimum admittance increment; superimposing the unique master pile to generate a supplementary admittance increment; generating a compensation current; performing frequency domain criterion verification; and triggering closed-loop update.

[0007] As a preferred embodiment of the voltage stability control method for charging piles oriented towards DC charging described in this invention, the generation of the initial detection time slot index includes: starting coupling point voltage sampling and injection current sampling, obtaining the sampling period, generating the number of sampling points in the acquisition window and setting the capacity of the circular buffer, initializing the circular buffer write pointer, and forming a circular buffer; broadcasting the charging pile number and receiving the neighboring charging pile numbers, generating a neighboring set, calculating the number of detection time slots, and generating the initial detection time slot index.

[0008] As a preferred embodiment of the voltage stability control method for charging piles oriented towards DC charging described in this invention, the trigger acquisition window data interception includes: based on the initial detection time slot index, performing conflict detection and resolution, using the maximum number concession update until there is no conflict, generating the final detection time slot index, generating trigger conditions, and when the trigger conditions are met, the charging pile intercepts acquisition window data from the cyclic buffer to obtain the coupling point voltage acquisition window data sequence and the injection current acquisition window data sequence.

[0009] As a preferred embodiment of the voltage stability control method for charging piles oriented towards DC charging described in this invention, the step of obtaining the passive margin sampling value sequence includes: freezing the acquisition window boundary based on the coupling point voltage acquisition window data sequence and the injection current acquisition window data sequence, generating a discrete angular frequency point set, performing a discrete Fourier transform to obtain the frequency domain voltage and frequency domain current; and calculating the port impedance and port admittance point by point based on the discrete angular frequency point set, the frequency domain voltage, and the frequency domain current to obtain the port impedance sequence and the port admittance sequence, thereby generating the passive margin sampling value sequence.

[0010] As a preferred embodiment of the voltage stability control method for charging piles oriented towards DC charging described in this invention, the generation of the passive margin function includes: generating a resonance identification threshold based on the port impedance sequence and the passive margin sampling value sequence, screening candidate peak points and determining the dominant peak, generating the key resonance frequency band boundary, and generating the passive margin function according to the piecewise constant value rule.

[0011] As a preferred embodiment of the voltage stability control method for charging piles oriented towards DC charging described in this invention, the generation of the passive margin target includes: traversing the set of discrete angular frequency points, obtaining a set of non-critical frequency point indices based on the lower boundary angular frequency and the upper boundary angular frequency of the critical resonant frequency band; reading the set of non-critical frequency point indices, combining it with the passive margin sampling value sequence, extracting the non-critical margin sequence, and obtaining the passive margin target.

[0012] As a preferred embodiment of the voltage stability control method for charging piles oriented towards DC charging described in this invention, the step of writing the initial value of the consistency iteration includes: performing range integration and bandwidth normalization of the key resonant frequency band according to the passive margin target, obtaining the margin gap index, writing the initial value of the consistency iteration, and initializing the buffer amount.

[0013] As a preferred embodiment of the voltage stability control method for DC charging piles described in this invention, the generation of the main pile identifier and the division of shaping tasks include: generating wave quantity parameters based on the resonance identification threshold and combined with the low-frequency port impedance amplitude; performing wave quantity encoding interaction and consistency iteration based on the initial value of consistency iteration; updating the consistency iteration value and updating the cache amount using the maximum consistency rule; generating the maximum value of the consistency iteration value; reading the maximum value of the consistency iteration value and collecting the charging pile number; performing the minimum number decision; obtaining the unique main pile number; writing the main pile identifier and distinguishing the main pile identifier status; and generating the division of shaping tasks.

[0014] As a preferred embodiment of the voltage stability control method for charging piles oriented towards DC charging described in this invention, the generation of guaranteed admittance increment includes: generating a shaping execution state based on the main pile identifier and the shaping task division, and generating a guaranteed admittance increment within the key resonant frequency band by combining the passive margin function.

[0015] As a preferred embodiment of the voltage stability control method for charging piles oriented towards DC charging described in this invention, the generation of compensation current includes: generating supplementary admittance increment based on the minimum admittance increment and combined with the passive margin target, and synthesizing the total admittance increment; performing frequency domain to time domain transformation according to the total admittance increment, obtaining the time domain kernel function, generating compensation current, superimposing the compensation current onto the current command, performing frequency domain criterion verification, and triggering closed-loop update.

[0016] The beneficial effects of this invention are as follows: By using wave quantity coding interaction and consistency iteration to determine the unique master pile, the shaping task division is generated, so that the compensation responsibility of each charging pile for the key resonant frequency band is clearly divided and coordinated, effectively avoiding mutual interference between detection and injection, and improving the controllability and scalability of parallel operation; According to the passive margin function, a minimum admittance increment is generated, and the admittance increment is superimposed by the unique master pile to form a compensation current. Combined with frequency domain criterion verification and closed-loop update strategy, the coupling point voltage is kept stable under operating condition disturbance, and adaptive maintenance is achieved. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 A flowchart of a voltage stabilization control method for charging piles oriented towards DC charging.

[0019] Figure 2 This is a flowchart captured from the capture window.

[0020] Figure 3 The flowchart for generating the passive margin function.

[0021] Figure 4 The flowchart for the closed-loop update of the sole main pile adjudication.

[0022] Figure 5 This is a comparison diagram of the voltage at the coupling point.

[0023] Figure 6 Stability diagram of main pile number. Detailed Implementation

[0024] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0025] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0026] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0027] Reference Figures 1-6 This is one embodiment of the present invention, which provides a voltage stabilization control method for charging piles oriented towards DC charging, including the following steps: S1. Start continuous sampling of coupling point voltage and injection current and establish a cyclic buffer, calculate the number of detection time slots, generate the initial detection time slot index, perform conflict detection and resolution, solidify the final detection time slot index into a schedule, and trigger the acquisition window data interception.

[0028] Start sampling of coupling point voltage and injection current, obtain sampling period, generate sampling window number of sampling points and set cyclic buffer capacity, initialize cyclic buffer write pointer to form cyclic buffer; broadcast charging pile number and receive neighboring charging pile number, generate neighborhood set, calculate detection time slot number, generate initial detection time slot index.

[0029] Furthermore, the coupling point voltage sampling and injection current sampling are initiated, and the sampling circuit continuously outputs the sampling sequence; a circular buffer is established in the memory, the circular buffer capacity is set, and the latest sampling point is written in a fixed step manner.

[0030] Read the sampling period and generate the number of sampling points in the acquisition window and the capacity of the circular buffer, and determine the constraint relationship between the number of sampling points in the acquisition window and the capacity of the circular buffer.

[0031] Specifically, the number of sampling points in the acquisition window is the ratio of the acquisition window duration to the sampling period. Specifically, the sampling circuit configuration is read, the sampling period is obtained, and the sampling period is written into the control parameter area. The acquisition window duration is set by the controller to be an integer multiple of the sampling period. The capacity of the loop buffer is the number of sampling points that the loop buffer can hold, and it must cover at least twice the number of sampling points in the acquisition window.

[0032] Furthermore, the write pointer of the circular buffer is initialized to zero, and a timer interrupt is started. The timer interrupt period is consistent with the sampling period. Each time the timer interrupt arrives, the target charging pile collects the coupling point voltage sampling value and the injection current sampling value. The injection current sampling value is taken from the output current sampling channel of the power conversion control loop. The coupling point voltage sampling value and the injection current sampling value are written to the current position of the circular buffer with the same timestamp. The timer interrupt sampling and writing operations are continuously executed, and the circular buffer is continuously updated over time.

[0033] Furthermore, the target charging pile establishes a rule for generating the test current component. The test current component is used to identify the controllable excitation during the frequency domain period. The test current component is formed by superimposing multiple sinusoidal components, with a total of 16 sinusoidal components. The frequency of the sinusoidal component is taken as an integer multiple of the frequency corresponding to the reciprocal of the acquisition window duration, with the integer multiple ranging from 1 to 16. The initial phase of the sinusoidal component is 0. The peak amplitude of each sinusoidal component is 0.125% of the upper limit of the rated output current of the charging pile.

[0034] It should be noted that the number of sinusoidal components of the test current is 16. The 16 sinusoidal components cover the integer multiples of the fundamental frequency from 1 to 16 times the frequency, ensuring that the effective impedance amplitude forms a sufficiently dense sampling point in the low-frequency to mid-frequency range, so as to realize the identification of the resonant frequency range. Under the constraints of the acquisition window length, the cyclic buffer capacity and the operation cycle, the 16 sinusoidal components take into account both frequency resolution and computational load, ensuring the real-time performance of voltage stability control and limiting voltage fluctuations.

[0035] In this embodiment, the frequency of the sinusoidal component of the test current is set as an integer multiple of the fundamental frequency, with the integer multiple range set from 1 to 16, determined according to the frequency resolution. This ensures that the discrete Fourier transform landing point covers the sensitive range of low-frequency to mid-frequency resonance, while controlling the number of sinusoidal components to limit the computational load and buffer capacity, thus meeting the requirements of real-time detection and control cycle.

[0036] The peak amplitude of each sinusoidal component of the test current is taken as 0.125% of the upper limit of the rated output current. This is determined based on the minimum resolvable current amplitude of the sampling channel, the upper limit of the allowable disturbance of the charging output current, and the effective impedance amplitude to calculate the signal-to-noise ratio requirement. 0.125% ensures that the frequency domain current amplitude is more than ten times higher than the resolution and keeps the output voltage fluctuation within the allowable range of voltage stability control.

[0037] Furthermore, updating the circular cache write pointer is represented as follows: ; in, This indicates the write pointer corresponding to the current sampling point's write position. This represents the write pointer after the previous sampling point was written to the circular buffer. Indicates the circular cache size. This represents the remainder operator.

[0038] Furthermore, the target charging pile number is periodically broadcast through the communication interface, and the charging pile numbers broadcast by other charging piles are received within the same discovery period. The different numbers received are deduplicated and written into the neighborhood set. The target charging pile number is removed, the neighborhood set is obtained, and the neighborhood set is frozen until the end of the scheduling period. The discovery period is the neighborhood discovery time window in the scheduling period, and the scheduling period is the time window between the generation of two neighborhood sets.

[0039] Furthermore, based on the number of elements in the neighborhood set, the number of probe slots is calculated. Specifically, the number of elements in the neighborhood set is incremented by 1 to obtain the number of probe slots.

[0040] Furthermore, the target charging pile number is read and mapped to the detection time slot range to generate an initial detection time slot index.

[0041] The initial probe slot index is represented as: ; in, Indicates the initial probe slot index. Indicates the charging station number. This indicates the number of detection time slots.

[0042] Based on the initial probe time slot index, conflict detection and resolution are performed. The maximum number is used to yield and update until there is no conflict, generating the final probe time slot index and generating trigger conditions. When the trigger conditions are met, the charging pile extracts the acquisition window data from the cyclic buffer and obtains the coupling point voltage acquisition window data sequence and the injection current acquisition window data sequence.

[0043] Furthermore, in each scheduling iteration, the target charging pile sends its current probe slot index to each charging pile in its neighborhood set and receives the probe slot indexes from each charging pile in the neighborhood set. When it is detected that at least one charging pile in the neighborhood set has the same probe slot index as the target charging pile, a conflict is determined to have occurred. For the set of charging piles that have a conflict, the conflict detection and resolution adopts the maximum number concession update, allowing only the charging pile with the largest number in the set of charging piles that have a conflict to update its probe slot index by incrementing by 1 and taking the remainder of the number of probe slots, while the other charging piles keep their probe slot index unchanged. The exchange and detection are repeated until the target charging pile no longer detects a probe slot index that is the same as any neighboring member after one exchange, then the final probe slot index is generated and put into the scheduling for solidification.

[0044] Furthermore, the final detection time slot index is written into the scheduling status, and the current time slot count equals the final detection time slot index as the trigger condition. When triggered, the data acquisition window is entered for data interception. When not triggered, normal charging control and continuous sampling are maintained, the probe current component remains at zero value, and detection interception is not performed.

[0045] When the scheduling triggering condition is met, the target charging pile enters the detection time slot execution stage. Within the acquisition window duration, the current command of the charging pile is superimposed with the test current component. Within the acquisition window duration, continuous sampling is maintained and written to the loop buffer. When the acquisition window duration ends, the most recent continuous sampling data with a length equal to the number of sampling points in the acquisition window is extracted from the loop buffer as the acquisition window data.

[0046] Specifically, the calculation of the starting read pointer of the acquisition window is represented as follows: ; in, This indicates the starting read pointer for the data in the circular buffer within the acquisition window. Indicates the number of sample points in the acquisition window, subscript Indicates the sampling point number.

[0047] Furthermore, starting from the initial read pointer, several sampling points of the acquisition window are continuously read to form the coupling point voltage acquisition window data sequence and the injection current acquisition window data sequence, respectively. The injection current acquisition window data sequence contains the output current response component corresponding to the probe current component.

[0048] Furthermore, after the target charging pile current command is superimposed with the probe current component, the existing current limiting logic of the charging pile continues to perform amplitude limiting to prevent the output current from exceeding the hardware allowable range.

[0049] S2. Receive the acquisition window data, generate a set of discrete angular frequency points, obtain the passive margin sampling value sequence, determine the dominant peak and the boundary of the key resonant frequency band, and generate the passive margin function.

[0050] Based on the data sequence of the coupling point voltage acquisition window and the data sequence of the injection current acquisition window, the acquisition window boundary is frozen, a set of discrete angular frequency points is generated, and a discrete Fourier transform is performed to obtain the frequency domain voltage and frequency domain current. Based on the set of discrete angular frequency points, frequency domain voltage and frequency domain current, the port impedance and port admittance are calculated point by point, the port impedance sequence and port admittance sequence are obtained, and a passive margin sampling value sequence is generated.

[0051] Furthermore, the coupling point voltage acquisition window data sequence and the injection current acquisition window data sequence are read, and the acquisition window duration and the number of acquisition window samples are read. The coupling point voltage acquisition window data sequence, the injection current acquisition window data sequence, the acquisition window duration, and the number of acquisition window samples are written into the frequency domain identification working area. The start sample and end sample of the acquisition window are marked in the frequency domain identification working area, and the acquisition window boundary is frozen in the frequency domain identification working area.

[0052] Furthermore, a discrete angular frequency point set is generated based on the acquisition window duration. This set covers the frequency point indices from zero to the non-negative frequency range corresponding to the sampled Nyquist angular frequency. Specifically, the acquisition window duration and the number of acquisition window samples are read to generate a frequency point index list. The frequency point index list increments from zero, and the ending index is the integer part of half the number of acquisition window samples. Discrete angular frequency points are calculated for each item in the frequency point index list, and the discrete angular frequency point set is written into the frequency domain identification working area.

[0053] Furthermore, when the number of sampling points in the acquisition window is even, the end number of the frequency point number list corresponds to the sampling Nyquist angular frequency; when the number of sampling points in the acquisition window is odd, the end number of the frequency point number list corresponds to the maximum non-negative angular frequency, which is less than the sampling Nyquist angular frequency.

[0054] Furthermore, the frequency point number list is extended by one from the end number to the negative frequency mirror point number corresponding to the number of sampling points minus one. The negative frequency mirror point number is not written into the discrete angular frequency point set.

[0055] A Discrete Fourier Transform (DFT) is performed on the voltage acquisition window data sequence at the coupling point to obtain the frequency domain voltage. A DFT is also performed on the injection current acquisition window data sequence to obtain the frequency domain current. The DFT is performed point-by-point by point according to the number of samples in the acquisition window. The frequency domain voltage and frequency domain current are extracted according to the index range of the frequency point number list to form the frequency domain voltage sequence and frequency domain current sequence. The discrete angular frequency point set, the frequency domain voltage sequence, and the frequency domain current sequence are written into the frequency domain identification working area.

[0056] It should be noted that the process involves frequency-by-frequency calculation of port impedance, frequency-by-frequency calculation of port admittance, generation of passive margin sampling value sequence, generation of impedance amplitude sequence, generation of resonance identification threshold, screening of candidate peak points, determination of dominant peak, search for key resonant frequency band boundaries, generation of passive margin function, and acquisition of non-key frequency point number set. The entire frequency point number list is traversed, and the traversal range does not include negative frequency mirror frequency point numbers.

[0057] Furthermore, the discrete angular frequency point set is traversed to calculate the frequency domain current amplitude sequence; the minimum effective frequency domain current amplitude threshold is taken as the maximum value between the minimum resolvable current amplitude of the injected current sampling channel and the peak amplitude of one-tenth of the probe current component; based on the frequency domain current amplitude sequence and the minimum effective frequency domain current amplitude threshold, a set of effective frequency point numbers and a set of invalid frequency point numbers are generated and written into the frequency domain identification working area.

[0058] It should be noted that the minimum effective frequency domain current amplitude threshold is one-tenth of the peak amplitude of the trial current component. The frequency domain amplitude estimation error increases with the increase of noise and leakage effects. When the amplitude is less than one-tenth of the peak amplitude, the effective impedance amplitude calculation is more sensitive to noise, and the stability of the resonance amplitude mutation identification decreases. Using one-tenth of the peak amplitude ensures that the signal-to-noise ratio of the participating decision frequency points meets the stable identification requirements, and is constrained together with the minimum resolvable current amplitude of the sampling channel to avoid invalid frequency points from entering the resonance frequency point identification.

[0059] Furthermore, the port impedance and port admittance are calculated point by point according to the discrete angular frequency point set. When the discrete angular frequency point number belongs to the set of valid frequency point numbers, the port impedance is taken as the complex ratio of the frequency domain voltage to the frequency domain current, and the port admittance is taken as the reciprocal of the complex number of the port impedance. The real part of the port admittance is taken as the passive margin sampling value. When the discrete angular frequency point number belongs to the set of invalid frequency point numbers, the calculation of port impedance and port admittance is skipped and an invalid mark is written. The writing of the passive margin sampling value is skipped, and the passive margin sampling value sequence is obtained.

[0060] The port impedance sequence, port admittance sequence, and passive margin sample value sequence are written into the frequency domain identification working area. When the frequency domain identification is completed, the frequency domain identification working area contains the discrete angular frequency point set, frequency domain voltage, frequency domain current, port impedance sequence, port admittance sequence, passive margin sample value sequence, effective frequency point number set, and invalid frequency point number set.

[0061] Based on the port impedance sequence and the passive margin sampling value sequence, a resonance identification threshold is generated, candidate peaks are screened and the dominant peak is determined, the key resonant frequency band boundary is generated, and the passive margin function is generated according to the piecewise constant value rule.

[0062] Furthermore, the impedance amplitude sequence of the port impedance sequence is extracted, and the impedance amplitude is extracted item by item according to the effective frequency point sequence to form an effective impedance amplitude sequence. The median of the effective impedance amplitude sequence is calculated to obtain the resonance identification threshold, which is used for candidate peak point screening.

[0063] It should be noted that the median is used for the resonance identification threshold because parallel coupling and transient disturbances are prone to generating a small number of spikes. The median is not sensitive to spikes, thus improving the repeatability of the critical resonant frequency band boundary.

[0064] Specifically, the resonance identification threshold is expressed as: ; in, Indicates the resonance identification threshold. Indicates port impedance. Represents discrete angular frequency points. Represents the imaginary unit. Represents the median operator. Indicates the index of the discrete angular frequency point.

[0065] Furthermore, based on the impedance amplitude sequence, candidate peaks are screened. Specifically, the candidate peak selection rules include: the candidate peak number set does not contain elements of the invalid frequency number set; the candidate peak selection uses the port impedance amplitudes of two adjacent discrete angular frequency points within the valid frequency number set as adjacent comparison objects; the impedance amplitude is not less than the resonance identification threshold; the impedance amplitude is greater than the impedance amplitudes of two adjacent discrete angular frequency points; and the peak with the largest impedance amplitude is selected as the dominant peak from the candidate peaks.

[0066] Specifically, the dominant peak number is represented as: ; in, Indicates the dominant peak number. The set of candidate peak indexes. This represents the maximal ordinal operator.

[0067] Furthermore, the nearest local valley is searched point by point from the dominant peak toward the low-frequency direction, and the nearest local valley is searched point by point from the dominant peak toward the high-frequency direction. The local valley satisfies that the impedance amplitude is less than the impedance amplitude of two adjacent discrete angular frequency points. The angular frequency corresponding to the nearest local valley at the low frequency is recorded as the lower boundary angular frequency of the key resonant frequency band, and the angular frequency corresponding to the nearest local valley at the high frequency is recorded as the upper boundary angular frequency of the key resonant frequency band. The lower boundary angular frequency and the upper boundary angular frequency of the key resonant frequency band are written into the frequency domain identification working area.

[0068] It should be noted that the determination of the dominant peak and the search for the boundary of the key resonant frequency band traverse the set of valid frequency point indices, while the search process for the lower and upper bound angular frequencies of the key resonant frequency band skips the set of invalid frequency point indices.

[0069] Furthermore, the passive margin sampled value sequence is continuously converted into a passive margin function according to the set of discrete angular frequency points corresponding to the set of valid frequency point indices. The passive margin function adopts a piecewise constant value rule. Specifically, when the frequency falls between two adjacent discrete angular frequency points within the set of valid frequency point indices, the passive margin function takes the passive margin sampled value corresponding to the lower discrete angular frequency point. The piecewise constant value rule skips the invalid mark of the passive margin sampled value corresponding to the set of invalid frequency point indices.

[0070] Specifically, the piecewise constant rule is expressed as follows: ; in, Represents the passive margin function. Represents the independent variable of angular frequency. Represents discrete angular frequency points Passive margin sample value at the location, Represents discrete angular frequency points The next discrete angular frequency point with a higher adjacent angular frequency.

[0071] It should be noted that the range of the piecewise constant passive margin function lies between the minimum and maximum values ​​of the passive margin sampled value sequence, and its dimension is Siemens.

[0072] S3. Based on the passive margin sampling value sequence and the key resonant frequency band boundary, generate the passive margin target, write the initial value of the consistency iteration, perform wave quantity coding interaction and consistency iteration, determine the unique master stake, generate the master stake identifier and the shaping task division.

[0073] Traverse the set of discrete angular frequency points, and obtain the set of non-critical frequency point indices based on the lower and upper bound angular frequencies of the critical resonant frequency band; read the set of non-critical frequency point indices, and extract the non-critical margin sequence by combining it with the passive margin sampling value sequence to obtain the passive margin target.

[0074] Furthermore, the passive margin sampled value sequence, the lower boundary angular frequency of the key resonant band, the upper boundary angular frequency of the key resonant band, the set of discrete angular frequency points, and the piecewise constant rule of the passive margin function are read from the frequency domain identification working area.

[0075] Traverse the set of discrete angular frequency points, write the index of discrete angular frequency points that fall between the lower and upper bound angular frequencies of the critical resonant frequency band into the critical frequency point index set, and write the index of discrete angular frequency points that fall outside the range of the lower and upper bound angular frequencies of the critical resonant frequency band into the non-critical frequency point index set; write the critical frequency point index set and the non-critical frequency point index set into the collaborative computing work area.

[0076] Furthermore, the set of non-critical frequency point numbers is read, the passive margin sample value sequence is read from the frequency domain identification working area, the passive margin sample value is extracted item by item according to the set of non-critical frequency point numbers to form a non-critical margin sequence, and it is determined whether the non-critical margin sequence is empty. If the non-critical margin sequence is empty, the passive margin sample value sequence corresponding to the set of valid frequency point numbers is taken as the non-critical margin sequence.

[0077] Perform median calculation on non-critical margin sequences to obtain median margin, apply non-negative constraints to the median margin to generate a passive margin target, and write the passive margin target into the collaborative computing workspace.

[0078] Specifically, the passive margin objective is expressed as: ; in, Indicates a passive margin target. Indicates the lower boundary angular frequency of the key resonant frequency band. Indicates the upper limit angular frequency of the key resonant frequency band, subscript Indicates the lower bound, subscript Indicates the upper bound.

[0079] Based on the passive margin target, perform range integration and bandwidth normalization of the key resonant frequency band to obtain the margin gap index, write the initial value of the consensus iteration and initialize the buffer size; based on the resonance identification threshold and combined with the low-frequency port impedance amplitude, generate wave quantity parameters, combine with the initial value of the consensus iteration, perform wave quantity encoding interaction and consensus iteration, update the consensus iteration value and the buffer size using the maximum consensus rule, and generate the maximum value of the consensus iteration value.

[0080] Furthermore, the integration range is defined based on the lower and upper angular frequencies of the key resonant frequency band; the passive margin sampled value sequence is extended into a passive margin function according to the piecewise constant rule of the passive margin function; the passive margin target is subtracted from the positive part of the passive margin function within the key resonant frequency band range, and normalization is performed according to the bandwidth of the key resonant frequency band to obtain the margin gap index.

[0081] Specifically, the margin gap index is expressed as: ; in, Indicates the margin gap index, This represents the positive part operator.

[0082] It should be noted that when the margin gap index is not less than 0 and the passive margin function is not less than the passive margin target within the critical resonant frequency band, the margin gap index is equal to zero.

[0083] Furthermore, the margin gap index is written into the initial value of the consistency iteration, and the cache size is initialized to zero; the initial value of the consistency iteration and the cache size are written into the collaborative computing work area.

[0084] Furthermore, the low-frequency port impedance amplitude and resonance identification threshold in the frequency domain identification working area are read; when the resonance identification threshold is greater than 0 and the low-frequency port impedance amplitude is greater than 0, the wave quantity parameter is the ratio of the low-frequency port impedance amplitude to the resonance identification threshold; when the resonance identification threshold is equal to zero or the low-frequency port impedance amplitude is equal to zero, the wave quantity parameter is 1.

[0085] Specifically, the wave quantity parameter is expressed as: ; in, Represents wave quantity parameters, This represents the port impedance magnitude at the lowest discrete angular frequency point. This represents the angular frequency corresponding to the lowest discrete angular frequency point.

[0086] Furthermore, during each round of consensus iteration, a wave quantity is sent to each charging pile in the neighborhood set, and a wave quantity sent by each charging pile in the neighborhood set is received; the received wave quantity is decoded into a neighborhood consensus value, and the consensus iteration value is updated with the neighborhood consensus value; after the update is completed, the next round of iteration begins.

[0087] Specifically, wave quantity encoding, decoding, and buffer size updates are represented as follows: ; ; ; in, Indicates the first The amount of wave volume transmitted in the round. Indicates the first Round-consistent iteration value, Indicates the first Round cache size, Indicates the decoding consistency value. This indicates the cache size for the next round.

[0088] It should be noted that the consistency iteration value update adopts the maximum consistency rule, which compares the consistency iteration value of the current round with all decoded consistency values ​​in the neighborhood set one by one, and selects the maximum value as the consistency iteration value of the next round. The consistency iteration is terminated when the consistency iteration value has not changed after a round of interaction and all decoded consistency values ​​in the neighborhood set are not greater than the consistency iteration value.

[0089] Read the maximum value of the consistency iteration and collect the charging pile number, execute the minimum number decision, obtain the unique main pile number, write the main pile identifier and distinguish the main pile identifier status, and generate the shaping task allocation.

[0090] Furthermore, after the consensus iteration is completed, each charging pile has the same maximum consensus iteration value. Each charging pile in the neighborhood set sends its charging pile number and consensus iteration value simultaneously. The smallest charging pile number is selected as the unique master pile number from the set of charging piles whose consensus iteration value is equal to the maximum value. The unique master pile number is written into the master pile identifier. The master pile identifier is true when it is the only master pile, and false when it is a non-master pile.

[0091] The task assignment for shaping is generated. Specifically, when the main pile identifier is true, the task of calculating and injecting the supplementary admittance increment is assigned, and the task of calculating and injecting the minimum admittance increment is assigned; when the main pile identifier is false, only the task of calculating and injecting the minimum admittance increment is assigned; the task assignment for shaping is written into the collaborative computing work area.

[0092] S4. Based on the main pile identification and shaping task division, combined with the passive margin function, generate the minimum admittance increment, generate the supplementary admittance increment by superimposing the unique main pile, generate the compensation current, perform frequency domain criterion verification and trigger closed-loop update.

[0093] Based on the main pile identification and the division of shaping tasks, the shaping execution status is generated, and combined with the passive margin function, the minimum admittance increment is generated in the key resonant frequency band.

[0094] Furthermore, the lower boundary angular frequency and upper boundary angular frequency of the key resonant frequency band, as well as the passive margin function, are read in the frequency domain identification working area; the passive margin target, main pile identifier, and shaping task division are read in the collaborative calculation working area; when the main pile identifier is true, the supplementary admittance increment calculation is enabled; when the main pile identifier is false, the supplementary admittance increment calculation is disabled, and only the minimum admittance increment calculation is enabled.

[0095] Furthermore, the angular frequency points within the range from the lower boundary angular frequency of the key resonant frequency band to the upper boundary angular frequency of the key resonant frequency band are traversed, and the passive margin function value is extracted according to the piecewise constant passive margin function rule; the positive part operation is performed on the passive margin function value to generate the minimum admittance increment, so that the negative part of the margin function is raised to zero, and the minimum admittance increment is written into the shaping working area.

[0096] Specifically, the guaranteed admittance increment is expressed as: ; in, This indicates the minimum allowable increment.

[0097] Based on the minimum admittance increment and combined with the passive margin target, a supplementary admittance increment is generated, and the total admittance increment is synthesized. According to the total admittance increment, a frequency domain to time domain transformation is performed to obtain the time domain kernel function, a compensation current is generated, the compensation current is superimposed on the current command, a frequency domain criterion check is performed, and a closed-loop update is triggered.

[0098] Furthermore, when the main pile identifier is true, the passive margin target and the minimum admittance increment are read, the passive margin function value is read, and the positive part operation is performed on the margin gap in the key resonant frequency band to generate the supplementary admittance increment, so that the passive margin function reaches the passive margin target after the minimum. When the main pile identifier is false, the supplementary admittance increment is set to zero and the calculation is skipped.

[0099] Specifically, the supplementary admittance increment is expressed as: ; in, This indicates that the admittance increment is being supplemented.

[0100] Furthermore, the total admittance increment within the key resonant frequency band is synthesized, where the total admittance increment is equal to the sum of the minimum admittance increment and the supplementary admittance increment; the total admittance increment is transformed from the frequency domain to the time domain within the key resonant frequency band to generate a time-domain kernel function, and a finite-time convolution is performed on the coupling point voltage to generate a compensation current.

[0101] Specifically, the time-domain kernel function and the compensation current are expressed as follows: ; ; in, Represents the time-domain kernel function. Representing the independent variable of time, Represents a complex exponential term. Indicates the compensation current, subscript Indicating compensation, Indicates the duration of the data collection window. Represents the integral variable. This represents the voltage at the coupling point.

[0102] It should be noted that the real-valued time-domain kernel function is formed by taking the real part of the complex exponential integral result. The angular frequency integral is performed by piecewise accumulation using a set of discrete angular frequency points and follows the piecewise constant passive margin function rule. The piecewise accumulation of frequencies is completed within the frequency index range corresponding to the set of discrete angular frequency points. The frequency indexes outside the frequency index range corresponding to the set of discrete angular frequency points are used to construct the conjugate symmetric mirror frequency indexes. The discrete sequence of the total admittance increment satisfies the conjugate symmetry constraint. The discrete sequence of the time-domain kernel function obtained by the inverse transform is a real-valued sequence. The time convolution is performed by piecewise accumulation using the voltage sequence provided by the circular buffer and the convolution length is fixed to the acquisition window duration, consistent with the acquisition window.

[0103] Furthermore, the length of the discrete sequence of the time-domain kernel function is equal to the number of sampling points in the acquisition window. The index of the discrete sequence of the time-domain kernel function ranges from zero to the number of sampling points in the acquisition window minus one. The time independent variable corresponding to the index of the discrete sequence of the time-domain kernel function ranges from zero to the acquisition window duration. When the time independent variable is less than zero, the time-domain kernel function takes a zero value. When the time independent variable is greater than the acquisition window duration, the time-domain kernel function takes a zero value. The compensation current generation process traverses the discrete sequence of the time-domain kernel function from zero to the number of sampling points in the acquisition window minus one. The voltage sequence index of the cyclic buffer is decremented according to the cyclic buffer read pointer. The segmented accumulation result is used as the compensation current output.

[0104] Furthermore, the compensation current is superimposed on the charging pile current command and output to the power conversion control loop, which executes according to the updated current command. When the updated current command triggers the existing current limiting logic of the charging pile, the existing current limiting logic of the charging pile is used to perform the limiting to avoid exceeding the hardware allowable range.

[0105] Furthermore, after triggering the next detection time slot and completing frequency domain identification, the port admittance within the key resonant frequency band is read, the minimum real part of the port admittance is calculated, the calculation of the minimum real part of the port admittance traverses the intersection of the key frequency point index set and the effective frequency point index set, the calculation of the minimum real part of the port admittance skips the invalid port admittance mark corresponding to the invalid frequency point index set, and performs an inequality judgment between the minimum real part of the port admittance and the passive margin target.

[0106] Specifically, the inequality determination is expressed as: ; in, Indicates port admittance. This represents the operator for taking the real part.

[0107] When the inequality determination is true, the total admittance increment and compensation current generation process continue to run and enter the next scheduling cycle; when the inequality determination is false, the neighborhood set and detection time slot index are regenerated and the frequency domain identification, unique main pile decision and the generation of shaping execution state are re-executed and the compensation current is generated to realize closed-loop adaptive update for power distribution coupling changes.

[0108] In this embodiment, the beneficial effects of the proposed method are verified through simulation experiments. Specifically, the detection environment is constructed using numerical simulation to create a parallel charging pile-coupled point scenario for DC charging. The simulation objects include the coupling point voltage measurement channel, the current output channel of each charging pile, the equivalent network parameters related to the key resonant frequency band, and the nearest-neighbor bidirectional digital message communication interaction link. The number of parallel charging piles is set to 8, and the communication topology is a nearest-neighbor sparse interconnection topology constructed based on the nearest-neighbor bidirectional digital message communication interaction link. Normal disturbances such as delay jitter and message loss are set in the digital message communication interaction link to cover the uncertainties commonly encountered in engineering operation. The coupling network parameters are set to be inconsistent and weakly damped, making the key resonant frequency band more easily excited by operating condition disturbances, which facilitates the observation of the improvement effects of insufficient passive margin and admittance shaping. The operating condition script includes load step, load characteristic switching, slow drift of coupling network parameters, and charging pile offline and re-online events.

[0109] Figure 5 The time-domain comparison results of the coupling point voltage within a 60-second simulation window are shown, including two voltage curves of the method in this embodiment and the control method, with voltage reference values ​​and stability constraint bands superimposed. The control method adopts parallel control commonly used in current engineering. Specifically, each charging pile performs voltage-current dual closed-loop control locally, and a fixed virtual impedance or fixed droop characteristic is superimposed on the current command or voltage command to achieve parallel current sharing. In the centralized master-slave collaborative control scenario, the master pile uniformly issues the coupling point voltage reference value and droop-related parameters, and other charging piles operate according to the issued parameters. Figure 5 The closed-loop update trigger time is marked with a vertical line, and the charging pile offline and charging pile online events are marked with red and green vertical lines, respectively.

[0110] like Figure 5 As shown, after load step and load characteristic switching, the coupling point voltage of the control method shows a larger drop and oscillation, and the fluctuation is further aggravated during the 40-second offline phase, with a slower recovery process. The method of this embodiment maintains a voltage reference value closer to the voltage under the same disturbance, the voltage fluctuation amplitude is significantly reduced, and it can still maintain near the stable constraint band after the offline event occurs. After the closed-loop update is triggered, it shows a faster fluctuation decay and recovery trend. Based on the passive margin function, a minimum admittance increment is generated, and the admittance increment is supplemented by the unique main pile to form a compensation current. With the frequency domain criterion verification and closed-loop update strategy, the coupling point voltage stability constraint can be maintained continuously under operating condition disturbance and topology change.

[0111] Figure 6 The diagram illustrates a comparison of how the main pile-related numbers change over time. In this embodiment, the curve represents the unique main pile number used in the consistent iterative adjudication, while the curve in the comparison method represents the dominant number proxy quantity when no adjudication mechanism is implemented. Figure 6As shown, the method in this embodiment maintains a single stable number for most of the time period, indicating that the method can still converge to a unique master pile under conditions such as sparse interconnection and communication jitter through wavelet coding interaction and consistency iteration. Near key events (such as closed-loop update triggering, neighborhood changes caused by charging pile offline and online), the number can still quickly return to a stable state, reflecting the uniqueness and robustness of the master pile decision. The number of the control method changes frequently over time, reflecting the instability of the dominant behavior of each charging pile when there is no unified decision, which can easily lead to unclear shaping responsibility boundaries, increase the probability of mutual interference between detection and injection, and reduce the controllability of parallel operation. Figure 6 From the perspective of the collaborative mechanism, the division of responsibilities for the sole main pile adjudication and reshaping tasks ensures a clear division and consistent coordination of responsibilities for compensation in the key resonant frequency band.

[0112] In summary, this invention generates a division of shaping tasks by using wave quantity coding interaction and consistency iterative adjudication to determine the unique master pile. This clearly defines and coordinates the compensation responsibilities of each charging pile for key resonant frequency bands, effectively avoiding mutual interference between detection and injection, and improving the controllability and scalability of parallel operation. Based on the passive margin function, a minimum admittance increment is generated, which is then superimposed by the unique master pile to form a compensation current. Combined with frequency domain criterion verification and closed-loop update strategy, the coupling point voltage remains stable under operating condition disturbances, achieving adaptive maintenance.

[0113] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A voltage stabilization control method for charging piles oriented towards DC charging, characterized in that, include: Start continuous sampling of coupling point voltage and injection current and establish a cyclic buffer, calculate the number of detection time slots, generate an initial detection time slot index, perform conflict detection and resolution, solidify the final detection time slot index into a schedule, and trigger data capture of the acquisition window; Receive data from the acquisition window, generate a set of discrete angular frequency points, obtain a passive margin sampling value sequence, determine the dominant peak and the boundary of the key resonant frequency band, and generate a passive margin function; Based on the passive margin sampling value sequence and the key resonant frequency band boundary, a passive margin target is generated, the initial value of the consistency iteration is written, wave quantity coding interaction and consistency iteration are performed, a unique master stake is determined, and a master stake identifier and shaping task division are generated. Based on the main pile identification and shaping task division, combined with the passive margin function, a minimum admittance increment is generated, a supplementary admittance increment is generated by superimposing the unique main pile, a compensation current is generated, frequency domain criterion verification is performed and closed-loop update is triggered.

2. The voltage stabilization control method for charging piles oriented towards DC charging as described in claim 1, characterized in that, The generation of the initial probe time slot index includes: Start sampling of coupling point voltage and injection current, obtain the sampling period, generate the number of sampling points in the acquisition window and set the capacity of the circular buffer, initialize the circular buffer write pointer, and form a circular buffer; Broadcast the charging pile number and receive the neighboring charging pile numbers, generate a neighbor set, calculate the number of detection time slots, and generate an initial detection time slot index.

3. The voltage stabilization control method for charging piles oriented towards DC charging as described in claim 1, characterized in that, The trigger acquisition window data interception includes: based on the initial detection time slot index, performing conflict detection and resolution, using the maximum number concession update until there is no conflict, generating the final detection time slot index, generating trigger conditions, and when the trigger conditions are met, the charging pile intercepts acquisition window data from the cyclic buffer to obtain the coupling point voltage acquisition window data sequence and the injection current acquisition window data sequence.

4. The voltage stabilization control method for charging piles oriented towards DC charging as described in claim 3, characterized in that, The process of obtaining the passive margin sample value sequence includes: Based on the data sequence of the coupling point voltage acquisition window and the data sequence of the injection current acquisition window, the acquisition window boundary is frozen, a set of discrete angular frequency points is generated, and a discrete Fourier transform is performed to obtain the frequency domain voltage and frequency domain current. Based on the discrete angular frequency point set, frequency domain voltage, and frequency domain current, the port impedance and port admittance are calculated point by point to obtain the port impedance sequence and port admittance sequence, and a passive margin sampling value sequence is generated.

5. The voltage stabilization control method for charging piles oriented towards DC charging as described in claim 1, characterized in that, The generation of the passive margin function includes: Based on the port impedance sequence and the passive margin sampling value sequence, a resonance identification threshold is generated, candidate peaks are screened and the dominant peak is determined, the key resonant frequency band boundary is generated, and the passive margin function is generated according to the piecewise constant value rule.

6. The voltage stabilization control method for charging piles oriented towards DC charging as described in claim 1 or 5, characterized in that, The objective for generating passive margin includes: Traverse the set of discrete angular frequency points, and obtain the set of non-critical frequency point indices based on the lower boundary angular frequency and the upper boundary angular frequency of the critical resonant frequency band. Read the set of non-critical frequency points, combine it with the passive margin sampling value sequence, extract the non-critical margin sequence, and obtain the passive margin target.

7. The voltage stabilization control method for charging piles oriented towards DC charging as described in claim 6, characterized in that, The initial values ​​for the write consistency iteration include: Based on the passive margin target, perform range integration and bandwidth normalization of the key resonant frequency band to obtain the margin gap exponent, write the initial value of the consistency iteration and initialize the cache size.

8. The voltage stabilization control method for charging piles oriented towards DC charging as described in claim 5, characterized in that, The division of labor for generating the main pile identifier and shaping the task includes: Based on the resonance identification threshold and combined with the low-frequency port impedance amplitude, wave quantity parameters are generated. Combined with the initial value of the consensus iteration, wave quantity encoding interaction and consensus iteration are performed. The consensus iteration value is updated and the buffer amount is updated using the maximum consensus rule to generate the maximum value of the consensus iteration value. Read the maximum value of the consistency iteration and collect the charging pile number, execute the minimum number decision, obtain the unique main pile number, write the main pile identifier and distinguish the main pile identifier status, and generate the shaping task allocation.

9. The voltage stabilization control method for charging piles oriented towards DC charging as described in claim 8, characterized in that, The generation of the guaranteed admittance increment includes: Based on the main pile identification and the division of shaping tasks, the shaping execution status is generated, and combined with the passive margin function, the minimum admittance increment is generated in the key resonant frequency band.

10. The voltage stabilization control method for charging piles oriented towards DC charging as described in claim 9, characterized in that, The generated compensation current includes: Based on the minimum admittance increment, combined with the passive margin target, a supplementary admittance increment is generated, and the total admittance increment is synthesized. Based on the total admittance increment, perform frequency domain to time domain transformation, obtain the time domain kernel function, generate compensation current, superimpose the compensation current onto the current command, perform frequency domain criterion verification, and trigger closed-loop update.

Images