A distribution box power distribution system with subcircuit dynamic equalization regulation

Abstract

The application discloses a kind of distribution box power distribution systems of subcircuit dynamic equal load regulation, comprising: harmonic characteristic analysis module, for determining the harmonic pollution level label and neutral line current nonlinear amplification coefficient corresponding to each subcircuit;Active filter compensation module, for determining active filter compensation reference phase and output compensation current;Equal load optimization distribution module, for obtaining the harmonic pollution factor of each loop, determine phase correction load, and generate equal load regulation scheme;Switching verification execution module, for adjusting each loop access phase based on equal load regulation scheme, and judge whether neutral line is overloaded;Overload protection control module, for executing loop access phase locking and equal load regulation scheme updating according to neutral line overload judgment result, and implementing neutral line overload protection control.The application can realize the dynamic optimization regulation of load state inside distribution box, reduce three-phase imbalance and neutral line overload risk, and improve power distribution system operation stability.

Description

Technical Field

[0001] This invention relates to the field of power system automation and intelligent operation and maintenance technology, and in particular to a distribution box power distribution system with dynamic load sharing regulation of individual circuits. Background Technology

[0002] With the widespread application of modern power electronics technology, the penetration rate of nonlinear loads in low-voltage distribution networks, such as variable frequency drives, switching power supplies, LED lighting drivers, charging piles, and uninterruptible power supplies, continues to rise. These loads generate a large amount of non-sinusoidal harmonic currents during operation, injecting them into the distribution system and causing severe distortion of the current waveforms in each circuit within the distribution box. In low-voltage distribution systems using a three-phase four-wire power supply, harmonic currents, especially the third harmonic and its odd multiples of the zero-sequence harmonics, cannot cancel each other out between phase lines due to their phase consistency. This results in an arithmetic superposition on the neutral line, causing the effective value of the neutral line current to far exceed the effective values ​​of the phase line currents, leading to a series of operational safety problems such as neutral line overload, localized transformer overheating, and malfunctioning protection devices. At the same time, the spatiotemporal distribution characteristics of user-side loads are becoming increasingly complex. Different circuits within the same distribution box connect to diverse types of equipment, overlapping operating periods, and significant differences in load rates, resulting in inherent imbalances in the three-phase loads. Traditional load balancing in distribution boxes primarily relies on manual calculation of the effective values ​​of each phase current for static phase sequence adjustment, or on automatic phase-switching devices based on the effective values ​​of the current. The former has a slow response speed and long adjustment cycle, making it difficult to track load changes in a timely manner. The latter, while achieving dynamic adjustment, only uses the effective values ​​of each phase current as the balancing target, without considering the actual impact of the amplitude and phase of harmonic currents in each sub-circuit on the neutral current. Since the phase relationship of harmonic currents between different circuits directly affects the degree of harmonic superposition on the neutral line, load balancing based solely on the effective values ​​of the current may lead to the neglect of harmonic vector superposition, resulting in a theoretical balance but an actual increase in the neutral current. Furthermore, existing distribution box protection mechanisms typically design neutral line overload protection and harmonic mitigation as independent systems. Overload protection relies on the thermal-magnetic tripping characteristics of circuit breakers, resulting in a delayed response and an inability to distinguish the differences in thermal effects generated by fundamental current and harmonic current. Harmonic mitigation is achieved through the installation of passive or active filters for independent compensation, lacking coordination with load balancing. When a distribution box faces both three-phase imbalance and harmonic pollution simultaneously, coupling disturbances may occur between the two adjustment methods, causing the load-sharing adjustment scheme and the harmonic compensation scheme to affect each other, making it difficult to achieve the overall optimal power distribution operation. Therefore, it is urgent to propose a dynamic load-sharing adjustment method for distribution boxes that can comprehensively consider the harmonic pollution characteristics of individual circuits, the harmonic vector coupling effect, and the nonlinear amplification characteristics of the neutral current, so as to achieve synergistic optimization of load balancing adjustment and harmonic control, and improve the safety of distribution box operation. Summary of the Invention

[0003] To address the problems existing in the prior art, this invention provides a power distribution system with dynamic load balancing adjustment for each circuit, mainly comprising: The harmonic characteristic analysis module is used to extract the current waveforms of each sub-circuit and the neutral line current waveform based on the historical operation logs of the distribution box's weak current network, and to determine the harmonic pollution level label and the nonlinear amplification factor of the neutral line current corresponding to each sub-circuit. The active filter compensation module is used to collect the neutral line current waveform, grid voltage waveform and harmonic phasor data of each circuit according to the detection results of the harmonic pollution level of each circuit, determine the active filter compensation reference phase and output the compensation current, and decompose the current waveform of each circuit into the fundamental component and the harmonic component of the preset order in real time. The load sharing optimization and allocation module is used to obtain the harmonic pollution factor of each circuit based on the effective value of the fundamental component and the effective value of the harmonic component, determine the phase correction load in combination with the load state and the harmonic vector coupling relationship, and generate a load sharing adjustment scheme. The switching verification execution module is used to adjust the access phase of each circuit based on the load sharing regulation scheme, perform transient disturbance shielding and collect the neutral line current waveform, combine the fundamental current vector and nonlinear amplification coefficient to obtain the predicted effective value of the neutral line current, and determine whether the neutral line is overloaded. The overload protection control module is used to perform circuit access phase locking and load sharing adjustment scheme updates based on the neutral line overload judgment result, adjust the circuit phase configuration in combination with overload status feedback information, and implement neutral line overload protection control.

[0004] Furthermore, the harmonic characteristic analysis module is used to extract the current waveforms of each sub-circuit and the neutral line current waveform based on the historical operation logs of the distribution box's weak current network, and to determine the harmonic pollution level label and neutral line current nonlinear amplification factor corresponding to each sub-circuit, including: By analyzing the historical operation logs of the distribution box's weak current network, current waveform data for each circuit, neutral line current waveform data, and voltage harmonic distortion rate data were obtained. Based on the acquired current waveform data, Fast Fourier Transform was used to extract the fundamental current amplitude, harmonic current amplitude, and harmonic phase angle for each circuit. The harmonic content, total harmonic distortion rate, and harmonic phase angle were then used to construct harmonic feature vectors. K-means clustering was employed to perform cluster analysis on the harmonic feature vectors. Based on the total harmonic distortion rate and harmonic content corresponding to each cluster center, the clustering results were sorted according to pollution level, resulting in low-pollution, medium-pollution, and high-pollution level labels for each circuit. Based on the harmonic feature vectors and the corresponding neutral line current... Based on the flow data and the effective value of the three-phase fundamental unbalanced current, the nonlinear amplification factor of the neutral current is calculated, and the harmonic feature vector and the corresponding nonlinear amplification factor of the neutral current are stored in the distribution box operation monitoring database. Through the distribution box operation monitoring database, historical data of the harmonic feature vector and the nonlinear amplification factor of the neutral current are obtained, and the gradient boosting decision tree algorithm is used to train the model to construct a harmonic-neutral current nonlinear amplification prediction model. Based on the real-time current waveforms of each sub-circuit of the three-phase four-wire distribution box, the real-time harmonic feature vector is extracted using fast Fourier transform, and the nonlinear amplification factor of the neutral current corresponding to each sub-circuit is determined using the harmonic-neutral current nonlinear amplification prediction model.

[0005] Furthermore, the active filter compensation module is used to collect the neutral current waveform, grid voltage waveform, and harmonic phasor data of each circuit based on the harmonic pollution level detection results of each sub-circuit, determine the active filter compensation reference phase and output the compensation current, and decompose the current waveform of each circuit into the fundamental component and harmonic components of a preset order in real time, including: Based on the harmonic pollution level of each circuit, it is determined whether there are any circuits with harmonic pollution levels exceeding the preset threshold. If so, the neutral current waveform, grid voltage waveform, and current waveforms of each circuit are acquired in real time. The grid voltage phase is obtained using a synchronous reference coordinate system phase-locked loop (PLL). The amplitude and phase of the preset-order zero-sequence harmonic component in the neutral current are extracted using a second-order generalized integrator. Harmonic phasor data of each circuit are obtained through a current transformer array. The phase curl exponent is calculated to determine the risk of PLL lockout failure and to determine the active filter compensation reference phase. Based on the determined compensation reference phase, the active filter connected in parallel on the outgoing side of the distribution box is activated. Based on the harmonic component amplitude, phase, and reference phase, a harmonic compensation reference current is synthesized. The harmonic compensation reference current is compared with the actual output current. The deviation is used as an error signal. The proportional resonance control algorithm is used to calculate the modulation wave signal of the active filter. The drive pulse generated by SPWM controls the switching device to output a compensation current that is opposite in phase to the neutral line harmonic current. The amplitude of the harmonic compensation reference current is adjusted by closed-loop feedback. When the harmonic component of the preset number in the compensated neutral line current is lower than the preset threshold for a preset number of consecutive power frequency cycles, the effective value of the compensated neutral line current is calculated. The effective value of the neutral line current is compared with the upper limit of the rated neutral line current of the circuit breaker. If it is less than the rated upper limit, the neutral line overload state is determined to be released. The instantaneous current waveform data of each sub-circuit is obtained through the current transformer array. The current waveform of each circuit is decomposed into the fundamental component and the harmonic component of the preset number in real time.

[0006] It also includes acquiring harmonic phasor data for each circuit through a current transformer array, calculating the phase curl index, determining the risk of phase-locked loop (PLL) loss of lock-in, and determining the active filter compensation reference phase, specifically including: The instantaneous current waveforms of each sub-circuit are acquired using a current transformer array. Fast Fourier Transform (FFT) is then used to extract the amplitude and phase of the same preset-order harmonic components in each sub-circuit, yielding harmonic phasor data for each circuit fed into the neutral line. Based on the phase angles in the harmonic phasor data of each circuit, a harmonic phase angle distribution sequence is constructed. The phase angles are normalized using the grid voltage phase as a reference to obtain the normalized phase angles for each circuit. Finally, the phase curl exponent formula is applied based on the normalized phase angle sequence. Calculate the phase curl index PWI of the current harmonic phase distribution, where, Let be the normalized phase angle of the k-th loop. The average normalized phase angle of all loops is N, where N is the total number of loops. If the phase curl exponent is less than or equal to the preset exponent threshold, the grid voltage phase output in real time by the synchronous reference coordinate system phase-locked loop is used as the reference phase. If the phase curl exponent is greater than the preset exponent threshold, it is determined that there is a risk of phase-locked loop loss in the current harmonic phase distribution. The current phase-locked loop output phase is temporarily stored by updating the output flag bit of the synchronous reference coordinate system phase-locked loop, and the temporarily stored phase is used as the reference phase of the active filter harmonic compensation reference current.

[0007] Furthermore, the load sharing optimization allocation module is used to obtain the harmonic pollution factor of each circuit based on the effective values ​​of the fundamental component and harmonic components, determine the phase correction load by combining the load state and harmonic vector coupling relationship, and generate a load sharing adjustment scheme, including: Based on the RMS values ​​of the fundamental and harmonic components, the total harmonic RMS value is calculated. Then, based on the total harmonic RMS value and the fundamental RMS value, the harmonic pollution factor for each circuit is obtained. Finally, based on the harmonic pollution factor and instantaneous active power data, the equivalent load factor is calculated using the formula... Calculate the equivalent load rate of each circuit. ,in, Let be the fundamental active power of the i-th circuit. Let be the harmonic active power of the i-th circuit. Let be the harmonic pollution factor of the i-th circuit; obtain the current switching status and maximum current carrying capacity parameters of each circuit through the intelligent circuit breaker communication interface, and map the equivalent load rate to a preset range using normalization processing to obtain the normalized load rate; calculate the harmonic vector coupling factor between circuits in the distribution box based on the amplitude and phase data of the preset harmonic components of each circuit, and determine the phase correction load; based on the phase correction load and the maximum current carrying capacity of each circuit, use a constrained genetic algorithm with the objective function of minimizing the variance of each phase correction load and minimizing the effective value of the predicted neutral current, and with the upper limit of the number of switching operations of each circuit, phase sequence consistency, and circuit breaker capacity constraints as constraints, construct a load sharing regulation optimization model, and generate a load sharing regulation scheme.

[0008] It also includes calculating the harmonic vector coupling factor between circuits in the distribution box based on the amplitude and phase data of the preset harmonic components of each circuit, and determining the phase correction load, specifically including: Based on the amplitude and phase data of the preset harmonic components of each circuit, the harmonic vector coupling factor formula is used. Calculate the harmonic vector coupling factor between any two circuits i and j within the distribution box. The inter-loop harmonic vector coupling factor matrix is ​​obtained, where, and Let be the current amplitudes of loops i and j under the h-th harmonic, respectively. and These represent the corresponding harmonic current phases; based on the normalized load factor and harmonic vector coupling factor matrix, the phase correction load formula is used. Determine the phase correction load ,in, Let be the normalized load factor of loop i.

[0009] Furthermore, the switching verification execution module is used to adjust the access phase of each circuit based on the load sharing adjustment scheme, perform transient disturbance shielding and acquire the neutral line current waveform, combine the fundamental current vector and the nonlinear amplification factor to obtain the predicted effective value of the neutral line current, and determine whether the neutral line is overloaded, including: Based on the load sharing regulation scheme, the intelligent circuit breaker sends switching commands to each circuit to determine the access phase of each circuit. Simultaneously with issuing the switching command, a transient disturbance shielding mechanism is activated, with the shielding window duration set to a preset number of power frequency cycles. During the shielding period, neutral current sampling data is temporarily stored, but overload comparison judgment is not performed. Using a digital signal processor, the current waveforms of each phase are acquired in real time after the shielding window ends. The effective value and peak value of the neutral current are calculated using the sliding window method, with the sliding window length being a preset number of power frequency cycles. Based on the nonlinear amplification factor, combined with the current fundamental current vector sum of each phase and the calculated effective value of the fundamental unbalanced current, the predicted effective value of the neutral current is obtained. The effective value of the neutral current is the product of the effective value of the fundamental unbalanced current and the nonlinear amplification factor. The predicted effective value of the neutral current is compared with the upper limit of the rated neutral current of the circuit breaker to determine whether the neutral line is overloaded.

[0010] Furthermore, the overload protection control module is used to perform loop access phase locking and load sharing adjustment scheme updates based on the neutral line overload judgment result, adjust the loop phase configuration in conjunction with overload state feedback information, and implement neutral line overload protection control, including: If the neutral line overload judgment result is no overload, the intelligent circuit breaker locks the current access phase of each circuit and generates a load sharing regulation scheme execution completion signal. If the neutral line overload judgment result is overload, the current load sharing regulation scheme is discarded, and the overload judgment result and overload phase sequence identifier are fed back to the load sharing regulation optimization model as new constraints. The load sharing regulation optimization model solution and neutral line overload verification are re-executed until the neutral line overload judgment result is no overload, and the optimized load sharing regulation scheme is determined. If the overload cannot be relieved after reaching the preset maximum number of optimizations, a neutral line overload alarm message is output. The system generates overload protection control commands, puts the distribution box in neutral line overload protection state, locks the current phase of each circuit connection and stops executing new load sharing adjustment operations. At the same time, it controls the active filter to enter continuous compensation mode to suppress neutral line harmonics in real time, and uploads the current neutral line current, load rate of each circuit, harmonic characteristic vector, overload phase sequence identifier and alarm time to the distribution box operation monitoring platform. When the neutral line current is detected to be lower than the upper limit of the rated neutral line current of the circuit breaker for a preset number of power frequency cycles, the neutral line overload protection state is released and the load sharing adjustment optimization model is restarted.

[0011] The technical solutions provided by the embodiments of the present invention may include the following beneficial effects: This invention provides a power distribution system with dynamic load sharing regulation for each circuit. By comprehensively analyzing historical operation logs and real-time operation data of the distribution box's weak current network, this invention establishes a correlation between the harmonic pollution characteristics of each circuit, the nonlinear variation characteristics of the neutral current, and the circuit load status, achieving dynamic perception of the distribution box's operating status. By collecting the current waveforms of each circuit, the neutral current waveform, and the grid voltage waveform, and extracting the fundamental component, preset-order harmonic components, and harmonic phasor information, the degree of harmonic pollution in different circuits can be accurately identified. Furthermore, by combining the harmonic vector coupling relationship, the superposition effect of harmonics between circuits can be determined, optimizing the traditional load sharing regulation method based solely on load current. This invention introduces a phase-corrected load evaluation method under the influence of harmonic coupling, combining the load level of each circuit and the harmonic interaction relationship between circuits to generate a dynamic load sharing regulation scheme. It also uses intelligent circuit breakers to automatically adjust the circuit connection phase, transforming the three-phase load distribution of the distribution box from static configuration to dynamic optimization regulation. This invention achieves real-time safety verification during load sharing adjustment through transient disturbance shielding, neutral current prediction, and overload status feedback mechanisms, avoiding misjudgments caused by transient processes during circuit switching or sudden load changes. Furthermore, this invention combines the neutral overload judgment results to perform closed-loop correction of the load sharing adjustment scheme, continuously optimizing the circuit-phase distribution relationship while meeting operational constraints. When an unresolvable neutral overload state is detected, abnormal state handling is achieved through protection control, active filter compensation, and operational data uploading. The distribution box power distribution system with dynamic load sharing adjustment for each circuit provided by this invention can simultaneously consider circuit load changes and harmonic coupling effects, achieving dynamic optimization and adjustment of the load state inside the distribution box, reducing the risk of three-phase imbalance and neutral overload, and improving the operational stability of the power distribution system. Attached Figure Description

[0012] Figure 1 A flowchart of a power distribution system with dynamic load sharing regulation of distribution boxes for each circuit according to the present invention; Figure 2 This is a schematic diagram of a power distribution box system with dynamic load sharing adjustment for branch circuits according to the present invention. Detailed Implementation

[0013] To make the objectives, technical strategies, and advantages of this invention clearer, the invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0014] like Figures 1-2 This embodiment describes a power distribution system with dynamic load balancing adjustment for each circuit, which may specifically include: Step S101, Harmonic Feature Analysis Module, is used to extract the current waveforms of each sub-circuit and the neutral line current waveform based on the historical operation log of the distribution box weak current network, and determine the harmonic pollution level label and the nonlinear amplification coefficient of the neutral line current corresponding to each sub-circuit.

[0015] By analyzing the historical operation logs of the distribution box's weak current network, current waveform data for each circuit, neutral current waveform data, and voltage harmonic distortion rate data were obtained. Based on the acquired current waveform data, a Fast Fourier Transform was used to extract the fundamental current amplitude, harmonic current amplitude, and harmonic phase angle for each circuit. The harmonic content, total harmonic distortion rate, and harmonic phase angle were then used to construct a harmonic feature vector. A K-means clustering algorithm was employed to cluster the harmonic feature vectors. Based on the total harmonic distortion rate and harmonic content corresponding to each cluster center, the clustering results were sorted according to pollution level, resulting in low-pollution, medium-pollution, and high-pollution level labels for each circuit. Based on the harmonic feature vectors and the corresponding neutral current data, combined with the effective value of the three-phase fundamental unbalanced current, the neutral current nonlinear amplification factor was calculated. The harmonic feature vectors and the corresponding neutral current nonlinear amplification factor were then stored in the distribution box operation monitoring database. Historical data on harmonic feature vectors and neutral current nonlinear amplification coefficients were obtained from the distribution box operation monitoring database. A gradient boosting decision tree algorithm was used to train the model, constructing a harmonic-neutral current nonlinear amplification prediction model. Based on the real-time acquired current waveforms of each circuit in the three-phase four-wire distribution box, a fast Fourier transform was used to extract real-time harmonic feature vectors. The harmonic-neutral current nonlinear amplification prediction model was then used to determine the neutral current nonlinear amplification coefficient for each circuit.

[0016] For example, a three-phase four-wire intelligent distribution box has 12 sub-circuits, with 4 circuits connected to phase A, 4 circuits to phase B, and 4 circuits to phase C. The rated neutral current limit of the circuit breaker is 160A. The historical operation logs of the distribution box for the past 30 days were retrieved, yielding 86,400 sets of historical sampling data. Each set of data includes the current waveforms of the 12 sub-circuits, the neutral current waveform, and the voltage harmonic distortion rate data at the corresponding time. Both current and voltage were sampled at a frequency of 10kHz, and one set of analysis results was saved every 100ms. Taking a set of historical data from the third sub-circuit as an example, after performing a Fast Fourier Transform analysis on its current waveform, the effective value of the fundamental current was found to be 48.2A, the amplitude of the 3rd harmonic current was 7.1A, the amplitude of the 5th harmonic current was 3.6A, and the amplitude of the 7th harmonic current was 1.9A, with corresponding harmonic phase angles of 35.6°, 118.4°, and 201.8°, respectively. The calculated harmonic content rates of the 3rd, 5th, and 7th harmonics were 14.73%, 7.47%, and 3.94%, respectively, and the total harmonic distortion rate was 16.98%. These data constituted the harmonic feature vector of this circuit. After extracting the harmonic feature vectors of the remaining 11 circuits using the same method, all historical harmonic feature vectors were input into a K-means clustering algorithm with a cluster size of 3. The average total harmonic distortion rates corresponding to the three cluster centers were 4.82%, 10.64%, and 18.35%, respectively. After sorting by pollution level from smallest to largest, they were labeled as low pollution, medium pollution, and high pollution levels. The total harmonic distortion rate of the third sub-circuit is 16.98%, which is closest to the third cluster center, therefore it is marked as a high pollution level. Simultaneously, the effective value of the neutral current at the corresponding moment in this set of historical data is 41.3A. Based on the three-phase fundamental current vector, the effective value of the fundamental unbalanced current is calculated to be 24.6A. According to the principle that the nonlinear amplification factor of the neutral current is equal to the ratio of the effective value of the neutral current to the effective value of the fundamental unbalanced current, the nonlinear amplification factor at this moment is calculated to be 41.3 / 24.6 = 1.68. The harmonic feature vector of the third sub-circuit and the corresponding nonlinear amplification factor of 1.68 are stored as a set of training samples in the distribution box operation monitoring database. After the database has accumulated 86,400 sets of historical samples, a gradient boosting decision tree algorithm is used for training, with the harmonic feature vector as input and the nonlinear amplification factor as output, to construct a harmonic-neutral current nonlinear amplification prediction model.When the distribution box enters the real-time operation phase, the current waveform of the third sub-circuit is collected at the current moment. After fast Fourier transform, the effective value of the fundamental current is 50.1A, and the amplitudes of the 3rd, 5th, and 7th harmonic currents are 7.5A, 3.9A, and 2.1A, respectively, with corresponding harmonic phase angles of 37.2°, 120.1°, and 203.4°. The total harmonic distortion rate is calculated to be 17.35%. The real-time harmonic feature vector is constructed from the above data and input into the trained prediction model, resulting in a nonlinear amplification factor of 1.72 for the current third sub-circuit.

[0017] Step S102, the active filter compensation module is used to collect the neutral line current waveform, grid voltage waveform and harmonic phasor data of each circuit according to the detection results of the harmonic pollution level of each circuit, determine the active filter compensation reference phase and output the compensation current, and decompose the current waveform of each circuit into the fundamental component and the harmonic component of the preset order in real time.

[0018] Based on the harmonic pollution level of each circuit, it is determined whether any circuit has a harmonic pollution level higher than a preset threshold. If so, the neutral current waveform, grid voltage waveform, and current waveforms of each circuit are acquired in real time. The grid voltage phase is obtained using a synchronous reference coordinate system phase-locked loop (PLL). The amplitude and phase of the preset-order zero-sequence harmonic component in the neutral current are extracted using a second-order generalized integrator. Harmonic phasor data for each circuit are obtained through a current transformer array. The phase curl exponent is calculated to assess the risk of PLL lockout failure and determine the active filter compensation reference phase. Based on the determined compensation reference phase, the active filter connected in parallel on the outgoing side of the distribution box is activated. A harmonic compensation reference current is synthesized based on the harmonic component amplitude, phase, and reference phase. The deviation between the harmonic compensation reference current and the actual output current is used as the error signal. A proportional resonant control algorithm is used to calculate the modulation wave signal of the active filter. SPWM is used to generate drive pulses to control the switching devices to output a compensation current with a phase opposite to the neutral harmonic current. The amplitude of the harmonic compensation reference current is adjusted through closed-loop feedback. When the harmonic component of the preset order in the compensated neutral current is lower than the preset threshold for a preset number of consecutive power frequency cycles, the effective value of the compensated neutral current is calculated. The effective value of the neutral current is compared with the upper limit of the rated neutral current of the circuit breaker. If it is less than the upper limit, the neutral overload condition is determined to be relieved. The instantaneous current waveform data of each sub-circuit is obtained through the current transformer array, and the current waveform of each circuit is decomposed in real time into the fundamental component and the harmonic component of the preset order.

[0019] For example, after the distribution box enters real-time operation, the total harmonic distortion rate of the third sub-circuit is 17.35%, which belongs to the high pollution level and exceeds the preset medium pollution level threshold, thus initiating the harmonic compensation process. At this time, the neutral line current waveform, the three-phase voltage waveform of the power grid, and the current waveforms of the 12 sub-circuits are collected in real time. The effective value of the neutral line current is measured to be 41.8A. The three-phase voltage is phase-locked using a synchronous reference coordinate system phase-locked loop, and the current power grid voltage phase is obtained as 26.4°. Here, the power grid voltage phase refers to the instantaneous phase angle obtained after establishing a synchronous rotating coordinate system with the power grid fundamental voltage as a reference. Its function is to provide a unified phase reference for all harmonic currents in the subsequent harmonic compensation process, avoiding phase deviation of the compensation current due to slight fluctuations in the power grid frequency. A second-order generalized integrator is used to separate harmonics in the neutral line current. Since the zero-sequence third harmonic is usually the main influencing component of the neutral line current in a three-phase four-wire distribution system, the preset analysis number is set to the third harmonic. After separation processing, the amplitude of the third harmonic current of the neutral line was found to be 18.6A, with a phase of 114.7°. Simultaneously, current waveforms of 12 circuits were acquired using current transformer arrays installed on each sub-circuit, and the phasor of the third harmonic current of each circuit was extracted using Fast Fourier Transform. The current phase curl index was calculated to be 2.60. Since subsequent calculations use radians, this value is converted to 0.045 rad. The preset phase curl exponent threshold is 0.05 rad. Because the calculated PWI is 0.045 rad, which is less than the preset threshold, it is determined that the harmonic phase distribution of each circuit is relatively concentrated, and there is no risk of PLL lockout. Therefore, the 26.4° real-time output of the synchronous reference coordinate system PLL is used as the active filter compensation reference phase. Subsequently, the parallel active filter installed on the outgoing side of the distribution box is started. Based on the neutral line third harmonic current amplitude of 18.6A, phase of 114.7°, and the compensation reference phase of 26.4°, a harmonic compensation reference current is generated. This reference current amplitude is consistent with the neutral line target harmonic current amplitude, but its phase is set opposite to the target harmonic current to cancel the zero-sequence harmonic component in the neutral line. The deviation between the reference compensation current and the actual output current of the active filter is used as the error signal. The modulation wave is calculated using a proportional resonance control algorithm, and then SPWM is used to control the inverter switching devices to output the compensation current. In the initial compensation phase, the actual output compensation current was 17.8A, with a 0.8A error compared to the target value of 18.6A. After closed-loop adjustment, the compensation current increased to 18.5A in the 5th power frequency cycle, and the neutral line 3rd harmonic current decreased from 18.6A to 5.2A; it decreased to 2.8A in the 10th power frequency cycle. A harmonic compensation stability threshold of 3.0A was set, requiring the current to remain below this threshold for 5 consecutive power frequency cycles. The detected values ​​for the 11th to 15th cycles were 2.8A, 2.7A, 2.6A, 2.8A, and 2.7A, respectively, all meeting the requirement, thus determining that the compensation was stable. The effective value of the neutral line current was recalculated, decreasing from 41.8A before compensation to 29.4A, which is less than the circuit breaker's rated upper limit of 160A, thus determining that the neutral line overload condition was resolved. After compensation, the 3rd, 5th, and 7th harmonic currents of the 3rd sub-circuit decreased to 2.1A, 1.4A, and 0.8A, respectively, and the total harmonic distortion rate decreased from 17.35% to 5.84%.

[0020] Among them, harmonic phasor data of each circuit are obtained through current transformer array, phase curl index is calculated, phase-locked loop loss risk is judged, and active filter compensation reference phase is determined.

[0021] The instantaneous current waveforms of each sub-circuit are acquired using a current transformer array. Fast Fourier Transform (FFT) is then used to extract the amplitude and phase of the same preset-order harmonic components in each sub-circuit, yielding harmonic phasor data for each circuit fed into the neutral line. Based on the phase angles in the harmonic phasor data of each circuit, a harmonic phase angle distribution sequence is constructed. The phase angles are normalized using the grid voltage phase as a reference to obtain the normalized phase angles for each circuit. Based on the normalized phase angle sequence, the phase curl exponent formula is used... Calculate the phase curl index PWI of the current harmonic phase distribution, where, Let be the normalized phase angle of the k-th loop. The average normalized phase angle of all loops is given by N, where N is the total number of loops. If the phase curl exponent is less than or equal to a preset exponent threshold, the grid voltage phase output in real time by the synchronous reference coordinate system phase-locked loop is used as the reference phase. If the phase curl exponent is greater than the preset exponent threshold, it is determined that there is a risk of phase-locked loop loss in the current harmonic phase distribution. The current phase-locked loop output phase is temporarily stored by updating the output flag bit of the synchronous reference coordinate system phase-locked loop, and the temporarily stored phase is used as the reference phase of the active filter harmonic compensation reference current.

[0022] For example, current waveforms of 12 circuits are acquired by an array of current transformers installed on each sub-circuit, and the third harmonic current phasor of each circuit is extracted using Fast Fourier Transform. The third harmonic current amplitude of the third sub-circuit is 7.5A and the phase is 37.2°; the third harmonic current amplitude of the fourth sub-circuit is 6.8A and the phase is 43.5°; the third harmonic current amplitude of the seventh sub-circuit is 5.9A and the phase is 39.8°; and the third harmonic current amplitudes of the remaining circuits are 2.6A, 1.9A, 2.3A, 1.7A, 2.1A, 1.5A, 1.3A, 1.8A, and 2.0A, respectively, with corresponding phases of 34.8°, 40.6°, 41.7°, 36.5°, 38.9°, 42.3°, 35.4°, 39.2°, and 40.1°. Using the grid voltage phase of 26.4° as a unified reference, the third harmonic phase of each circuit was normalized, resulting in normalized phase angle sequences of 10.8°, 17.1°, 8.4°, 14.2°, 13.4°, 15.3°, 10.1°, 12.5°, 15.9°, 9.0°, 12.8°, and 13.7°. Further calculation yielded an average normalized phase angle of 12 circuits of 12.77°. Based on the phase curl exponent formula... The phase curl index (PWI) of the current harmonic phase distribution is calculated to be 2.60. Since subsequent calculations use radians, it is converted to PWI = 0.045 rad, where... Let be the normalized phase angle of the k-th loop. The average normalized phase angle of all loops is given by N, where N is the total number of loops. The preset phase curl exponent threshold is 0.05 rad. Since the calculated PWI is 0.045 rad, which is less than the preset threshold, it is determined that the harmonic phase distribution of each loop is relatively concentrated, and there is no risk of PLL lockout. The 26.4° real-time output of the synchronous reference coordinate system PLL is used as the active filter compensation reference phase. At another operating moment, due to the simultaneous operation of multiple high-power nonlinear loads, the harmonic current phase of each loop changes rapidly. According to the phase curl exponent calculation formula, PWI = 0.381 rad, which is greater than the preset threshold of 0.05 rad. Therefore, it is determined that the harmonic phase distribution of each loop is highly discrete, and the PLL output phase may be affected by sudden load changes, resulting in large fluctuations and a risk of PLL lockout. To avoid the PLL output phase change causing a shift in the active filter compensation reference current direction, a phase protection mechanism is triggered. The current PLL output is frozen by the output update flag of the synchronous reference coordinate system PLL, and the PLL output phase in the stable period before the lockout risk occurs is read as a temporary phase.

[0023] Step S103, the load sharing optimization allocation module is used to obtain the harmonic pollution factor of each circuit based on the effective value of the fundamental component and the effective value of the harmonic component, determine the phase correction load in combination with the load state and the harmonic vector coupling relationship, and generate a load sharing adjustment scheme.

[0024] Based on the RMS values ​​of the fundamental and harmonic components, the total harmonic RMS value is calculated. Then, based on the total harmonic RMS value and the fundamental RMS value, the harmonic pollution factor for each circuit is obtained. Based on the harmonic pollution factor and instantaneous active power data, the equivalent load factor is calculated using the formula... Calculate the equivalent load rate of each circuit. ,in, Let be the fundamental active power of the i-th circuit. Let be the harmonic active power of the i-th circuit. Let be the harmonic pollution factor of the i-th circuit. Through the intelligent circuit breaker communication interface, the current switching status and maximum current carrying capacity parameters of each circuit are obtained. Normalization is applied to map the equivalent load rate to a preset range, yielding the normalized load rate. Based on the amplitude and phase data of the preset harmonic components of each circuit, the harmonic vector coupling factor between circuits in the distribution box is calculated, and the phase-corrected load is determined. Based on the phase-corrected load and the maximum current carrying capacity of each circuit, a constrained genetic algorithm is used. The objective function is to minimize the variance of each phase-corrected load and the effective value of the predicted neutral current. Constraints include the upper limit of the number of switching operations for each circuit, phase sequence consistency, and circuit breaker capacity constraints. A load-sharing adjustment optimization model is constructed, generating a load-sharing adjustment scheme and obtaining the optimal circuit-phase distribution mapping relationship.

[0025] For example, the preset analysis objects are the 3rd, 5th, and 7th harmonics. Therefore, the total harmonic effective value is calculated by taking the square root of the sum of the squares of the effective values ​​of each harmonic current, resulting in a total harmonic effective value of approximately 2.65A. Here, the total harmonic effective value represents the magnitude of the combined harmonic current after the combined effect of all harmonics, reflecting the overall harmonic pollution level of the circuit. The ratio of the total harmonic effective value to the fundamental current effective value is used as the harmonic pollution factor. Therefore, the harmonic pollution factor of the 3rd sub-circuit is 2.65 ÷ 50.1 = 0.053, or 5.3%, indicating that the harmonic current in this circuit accounts for approximately 5.3% of the fundamental current. Based on the real-time collected voltage and current data, and the fundamental active power of the 3rd sub-circuit (10.42kW) and the harmonic active power (0.38kW), the equivalent load factor calculation formula is used. The calculated equivalent load factor of the third sub-circuit is 0.998, indicating that although there is some harmonic current, the actual effective load capacity of this circuit is almost entirely determined by the fundamental active power because the proportion of harmonic power is low. Let be the fundamental active power of the i-th circuit. Let be the harmonic active power of the i-th circuit. Let be the harmonic pollution factor of the i-th circuit. After calculating the remaining 11 circuits using the same method, the equivalent load factors of the 12 circuits are 0.812, 0.845, 0.998, 0.763, 0.794, 0.838, 0.871, 0.802, 0.856, 0.781, 0.829, and 0.847, respectively. Subsequently, the current status of each circuit is read through the intelligent circuit breaker communication interface. All 12 circuits are in the closed state, and the maximum current carrying capacity of each circuit is 63A. After processing using the maximum-minimum normalization method, the normalized load factors of the 12 circuits are 0.209, 0.349, 1.000, 0, 0.132, 0.319, 0.460, 0.166, 0.396, 0.077, 0.281, and 0.357, respectively. The third harmonic current amplitude of the existing sub-circuit 3 is 2.1A, with a phase of 37.2°; the third harmonic current amplitude of the fourth sub-circuit is 1.8A, with a phase of 43.5°; the third harmonic current amplitude of the seventh sub-circuit is 1.9A, with a phase of 39.8°; and the third harmonic current amplitude of the tenth sub-circuit is 1.2A, with a phase of 58.6°. The calculated harmonic vector coupling factor is 0.852 between the third and fourth sub-circuits, 0.904 between the third and seventh sub-circuits, and 0.532 between the third and tenth sub-circuits. The harmonic vector coupling factors between the remaining circuits are all below 0.35. The above calculation results show that the third and seventh sub-circuits have strong harmonic coupling due to their larger harmonic amplitudes and smaller phase differences. Currently, phase A is connected to the 3rd, 4th, and 10th sub-circuits, with normalized load rates of 1.000, 0, and 0.396, respectively, and coupling factors among them of 0.852, 0.532, and 0.487, respectively. Based on the phase-corrected load formula, the phase-corrected load for phase A is calculated to be 3.267. Phase B is connected to the 7th sub-circuit and two other sub-circuits with low coupling, resulting in a calculated phase-corrected load of 2.814 for phase B and 2.936 for phase C. The overall corrected load for phase A is higher than that for phases B and C, indicating that the current load distribution method not only suffers from the traditional load unevenness problem but is also affected by harmonic superposition effects. Subsequently, using the phase-corrected loads of each phase, the maximum allowable current carrying capacity of 63A for each circuit, and the harmonic coupling relationship between circuits as input parameters, a constrained genetic algorithm load-sharing optimization model is constructed. The genetic algorithm uses the minimum variance of the optimized three-phase phase-corrected load and the minimum effective value of the predicted neutral current as its objective functions. Constraints include a maximum of one switching operation per circuit, maintaining the original phase sequence direction, and ensuring that the load current of each phase does not exceed the circuit breaker's rated capacity of 63A. After iterative solving, the algorithm obtains a load-sharing adjustment scheme, adjusting the 3rd branch circuit originally connected to phase A to phase C, adjusting the 7th branch circuit originally connected to phase C to phase B, and maintaining the original phase-line connections of the remaining 10 branch circuits.After adjustment, the three-phase load correction values ​​were changed from 3.267, 2.814, and 2.936 before optimization to 3.08, 3.04, and 3.05, resulting in a more balanced three-phase load distribution. This verifies that the load-sharing optimization method based on harmonic vector coupling constraints can effectively reduce the additional harmonic imbalance caused by loop combination. Furthermore, combining the effective value of the neutral line current after active filter compensation (29.4A) and the nonlinear amplification factor of 1.72, the predicted effective value of the neutral line current after optimization is 22.1A, meeting the constraint requirements.

[0026] Specifically, based on the amplitude and phase data of the preset harmonic components of each circuit, the harmonic vector coupling factor between circuits in the distribution box is calculated, and the phase correction load is determined.

[0027] Based on the amplitude and phase data of the preset harmonic components of each circuit, the harmonic vector coupling factor formula is used. Calculate the harmonic vector coupling factor between any two circuits i and j within the distribution box. The inter-loop harmonic vector coupling factor matrix is ​​obtained, where, and Let be the current amplitudes of loops i and j under the h-th harmonic, respectively. and These represent the corresponding harmonic current phases. Based on the normalized load factor and harmonic vector coupling factor matrix, the phase-corrected load formula is used. Determine the phase correction load ,in, Let be the normalized load factor of loop i.

[0028] For example, after calculating the normalized load rate for each circuit, the normalized load rates for the 3rd, 4th, 7th, and 10th circuits are 1.000, 0, 0.460, and 0.396, respectively, while the normalized load rates for the remaining circuits remain unchanged from the calculation results in the previous embodiment. The harmonic vector coupling factor between any two circuits is calculated using the extracted amplitude and phase data of the preset third harmonic current for each circuit. Here, the harmonic vector coupling factor considers not only the magnitude of the harmonic currents in the two circuits but also the phase difference between them. When the harmonic phases of the two circuits are closer, the cosine value is closer to 1, indicating that the two harmonic currents are more likely to superimpose on the neutral line. When the phase difference is large, the cosine value decreases, and the two harmonic currents have a certain canceling effect. Therefore, the harmonic vector coupling factor can more realistically reflect the degree of influence of different circuits connected to the same phase line on the neutral line harmonic current. Taking the 3rd and 7th sub-circuits as examples, the amplitude of the 3rd harmonic current in the 3rd sub-circuit is 2.1A, and the phase is 37.2°. The amplitude of the 3rd harmonic current in the 7th sub-circuit is 1.9A, and the phase is 39.8°. The phase difference between the two is 2.6°. According to the harmonic vector coupling factor formula... Calculate the harmonic vector coupling factor between any two circuits i and j within the distribution box. The harmonic vector coupling factor between the 3rd and 7th sub-circuits is found to be 0.904, where... and Let be the current amplitudes of loops i and j under the h-th harmonic, respectively. and These represent the corresponding harmonic current phases. Similarly, the harmonic vector coupling factor between the 3rd and 4th sub-circuits is calculated to be 0.852, and the harmonic vector coupling factor between the 3rd and 10th sub-circuits is 0.532. After calculating the harmonic vector coupling factors between all pairs of the 12 circuits using the same method, a 12×12 harmonic vector coupling factor matrix can be constructed, where all elements on the main diagonal of the matrix are 1, and the remaining elements represent the degree of harmonic coupling between any two circuits. Based on the above harmonic vector coupling factor matrix and the normalized load rate of each circuit, the phase correction load formula is used. Calculate the phase-corrected load for each phase. Here, the phase-corrected load is a comprehensive load index that incorporates the harmonic coupling effect on top of the traditional load rate. It considers both the load magnitude of the circuit itself and the degree of harmonic superposition between circuits. Let be the normalized load factor of loop i. Currently, phase A is connected to sub-loops 3, 4, and 10. Therefore, the phase-corrected load of phase A consists of three parts: first, the sum of the normalized load factors of the three loops, i.e., 1.000 + 0 + 0.396 = 1.396; second, the harmonic vector coupling factors between the loops within phase A are added, including 0.852 between sub-loops 3 and 4, 0.532 between sub-loops 3 and 10, and 0.487 calculated between sub-loops 4 and 10. Therefore, the final phase-corrected load of phase A is 1.396 + 0.852 + 0.532 + 0.487 = 3.267. If phase B is currently only connected to sub-loop 7 and two other loops with low coupling, the calculated phase-corrected load of phase B is 2.814, and the calculated phase-corrected load of phase C is 2.936. It can be seen that although the normalized load factor of phase A is not significantly different from that of the other two phases, the large harmonic vector coupling factor of 0.852 between the third and fourth sub-circuits makes the overall load of phase A after correction significantly higher than that of the other two phases. This indicates that phase A currently has a strong harmonic superposition effect, which is not conducive to reducing the neutral line current.

[0029] Step S104, the switching verification execution module is used to adjust the access phase of each circuit based on the load sharing adjustment scheme, perform transient disturbance shielding and collect the neutral line current waveform, combine the fundamental current vector and nonlinear amplification coefficient to obtain the predicted effective value of the neutral line current, and determine whether the neutral line is overloaded.

[0030] Based on the load-sharing regulation scheme, a switching command is sent to each circuit via an intelligent circuit breaker to determine the access phase of each circuit. Simultaneously with issuing the switching command, a transient disturbance shielding mechanism is activated, with the shielding window duration set to a preset number of power frequency cycles. During the shielding period, neutral current sampling data is temporarily stored, but overload comparison is not performed. Using a digital signal processor, the current waveforms of each phase are acquired in real time after the shielding window ends. The effective value and peak value of the neutral current are calculated using a sliding window method, with the sliding window length being a preset number of power frequency cycles. Based on the nonlinear amplification factor, combined with the current fundamental current vector sum of each phase and the calculated effective value of the fundamental unbalanced current, the predicted effective value of the neutral current is obtained. The effective value of the neutral current is the product of the effective value of the fundamental unbalanced current and the nonlinear amplification factor. The predicted effective value of the neutral current is compared with the upper limit of the circuit breaker's rated neutral current to determine whether the neutral line is overloaded.

[0031] For example, based on the load sharing regulation scheme, a switching control command is sent to the corresponding circuit via the intelligent circuit breaker communication interface to switch the 3rd sub-circuit from phase A to phase C, and the 7th sub-circuit from phase C to phase B, while the remaining 10 sub-circuits maintain their original phase line connections. Since the circuit breaker contact operation, line impedance changes, and load redistribution during circuit switching can cause transient current fluctuations in a short time, a transient disturbance shielding mechanism is activated simultaneously with the sending of the switching command. Taking a 50Hz power frequency grid as an example, the shielding window length is set to 5 power frequency cycles, i.e., 100ms. During this shielding time, the digital signal processor continues to collect neutral line current sampling data and temporarily stores the corresponding data, but does not perform neutral line overload judgment to avoid the instantaneous current peak generated during the switching transient process causing misjudgment of the protection logic. After the transient shielding window ends, the system re-enters stable detection mode. The digital signal processor acquires the three-phase current waveforms (A, B, and C) in real time, and calculates the effective value of the neutral current using a sliding window method. The sliding window length is set to 10 power frequency cycles (200ms). The current effective value of the neutral current is obtained by continuously updating the sampled data within the window. At this time, based on the aforementioned optimized loop-phase distribution results, the fundamental current vectors of each phase are recalculated, yielding an optimized effective value of 12.85A for the fundamental unbalanced current, representing the zero-sequence unbalanced component that still exists after the three-phase fundamental current vectors are superimposed. Combining this with the neutral current nonlinear amplification factor of 1.72 obtained after active filter compensation in the aforementioned embodiment, the formula for predicting the effective value of the neutral current is used. The optimized predicted effective value of the neutral current was calculated to be 22.1A, where, This indicates the predicted effective value of the neutral current. This represents the nonlinear amplification factor of the neutral line current. This represents the effective value of the fundamental unbalanced current. Subsequently, this predicted value is compared with the upper limit of the rated neutral current set by the intelligent circuit breaker, 160A. Since 22.1A is significantly lower than 160A, it indicates that after circuit redistribution, the three-phase load imbalance has been effectively reduced, and the neutral current has not reached the overload threshold. Therefore, the current neutral line operation of the distribution box is deemed safe. Simultaneously, the sliding window acquisition results are retained, and the predicted neutral current value is periodically updated to continuously monitor the risk of neutral line overload under subsequent load changes.

[0032] Step S105, the overload protection control module is used to perform circuit access phase locking and load sharing adjustment scheme update based on the neutral line overload judgment result, adjust the circuit phase configuration in combination with overload status feedback information, and implement neutral line overload protection control.

[0033] If the neutral line overload judgment result is no overload, the intelligent circuit breaker locks the current access phase of each circuit and generates a load sharing adjustment scheme execution completion signal. If the neutral line overload judgment result is overload, the current load sharing adjustment scheme is discarded, and the overload judgment result and overload phase sequence identifier are fed back to the load sharing adjustment optimization model as new constraints. The load sharing adjustment optimization model solution and neutral line overload verification are re-executed until the neutral line overload judgment result is no overload, and the optimized load sharing adjustment scheme is determined. If the overload cannot be relieved after reaching the preset maximum number of optimizations, a neutral line overload alarm message is output, an overload protection control command is generated, the distribution box is placed in the neutral line overload protection state, the current access phase of each circuit is locked and the execution of new load sharing adjustment operations is stopped. At the same time, the active filter is controlled to enter the continuous compensation mode to suppress neutral line harmonics in real time, and the current neutral line current, load rate of each circuit, harmonic characteristic vector, overload phase sequence identifier and alarm time are uploaded to the distribution box operation monitoring platform. When the neutral line current is detected to be lower than the upper limit of the rated neutral line current of the circuit breaker for a preset number of consecutive power frequency cycles, the neutral line overload protection state is released and the load sharing adjustment optimization model is restarted.

[0034] For example, the current predicted effective value of the neutral line current calculated according to the sliding window is 22.1A, while the upper limit of the rated neutral line current of the intelligent circuit breaker is set to 160A. Since 22.1A is less than 160A, it is determined that the neutral line is not currently overloaded. The intelligent circuit breaker locks the access phases of the current 12 circuits, keeping the optimized phase line connection relationship of the 3rd sub-circuit connected to phase C, the 7th sub-circuit connected to phase B, and the remaining circuits unchanged, and generates a load sharing adjustment scheme execution completion signal, thus completing the three-phase load sharing optimization process of the distribution box. In another operating state, if a sudden increase in load causes an abnormal rise in the neutral current, an overload test is performed again. For example, if a high-power nonlinear load is added at a certain moment, and the effective value of the neutral current calculated through the sliding window rises to 168.5A, while the upper limit of the rated neutral current of the circuit breaker is still 160A, then the current load sharing regulation scheme is determined to be unable to meet the neutral current constraint requirements, and the result is an overload. The current load sharing regulation scheme is immediately eliminated, and the overload information and overload phase sequence identifier obtained in this test are fed back to the load sharing regulation optimization model as new constraints. Based on the real-time acquisition results, phase B is determined to be the main overload phase sequence, and a "phase B load over-limit" constraint identifier is generated. The genetic algorithm optimization process is then re-executed. In the new optimization process, in addition to maintaining the maximum number of switching times of each circuit not exceeding 1, the load current of each phase not exceeding 63A, and the phase sequence consistency constraint, a phase B load limit constraint is further added to prevent the optimization model from generating a circuit-phase distribution scheme that would cause excessive phase B load. After re-optimization, a new load sharing regulation scheme is obtained, and the neutral overload verification is performed again. If, after the second optimization, the three-phase phase correction load is readjusted from 3.08, 3.04, and 3.05 before optimization to 3.01, 3.06, and 3.02, and the effective value of the fundamental unbalanced current decreases to 13.20A, combined with the aforementioned nonlinear amplification factor of 1.72, the new predicted effective value of the neutral line current is calculated to be 22.70A. Since 22.70A is lower than 160A, the neutral line overload risk is determined to be eliminated, confirming the effectiveness of the re-optimized load sharing adjustment scheme, and locking in the new load sharing adjustment scheme. If the neutral line overload risk still cannot be eliminated during continuous optimization, such as after reaching the preset maximum number of optimizations (5 times), if the neutral line current remains above 160A, assuming the effective value of the neutral line current is detected as 164.3A during the fifth verification, then the load sharing optimization calculation is stopped, a neutral line overload alarm message is output, and an overload protection control command is generated to switch the distribution box to the neutral line overload protection state. In the protected state, the intelligent circuit breaker locks the current phase of each circuit, stops new load sharing adjustment operations, and controls the active filter to enter the continuous compensation mode to continuously suppress the neutral line harmonic current based on the real-time detected harmonic current characteristics.Simultaneously, the current neutral current value of 164.3A, load rate data for each circuit, harmonic characteristic vector, overload phase sequence identifier, and alarm time information are uploaded to the distribution box operation monitoring platform for operational status recording and fault tracking. When the neutral current is continuously detected to be below 160A for 10 consecutive power frequency cycles, and if it stably drops to 35.6A and remains above 200ms, the overload risk is determined to be eliminated, the neutral overload protection state is automatically exited, and the load sharing adjustment optimization model is restarted. The load sharing adjustment scheme is recalculated based on the current load changes, thereby ensuring three-phase load balance and neutral current safety during the long-term operation of the distribution box.

[0035] The above description is merely a preferred embodiment of this application and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in this application is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the concept of this application. For example, technical solutions formed by substituting the above features with (but not limited to) technical features with similar functions disclosed in this application.

Claims

1. A power distribution system with dynamic load balancing regulation for each circuit, characterized in that, The system includes: The harmonic characteristic analysis module is used to extract the current waveforms of each sub-circuit and the neutral line current waveform based on the historical operation logs of the distribution box's weak current network, and to determine the harmonic pollution level label and the nonlinear amplification factor of the neutral line current corresponding to each sub-circuit. The active filter compensation module is used to collect the neutral line current waveform, grid voltage waveform and harmonic phasor data of each circuit according to the detection results of the harmonic pollution level of each circuit, determine the active filter compensation reference phase and output the compensation current, and decompose the current waveform of each circuit into the fundamental component and the harmonic component of the preset order in real time. The load sharing optimization allocation module is used to obtain the harmonic pollution factor of each circuit based on the effective value of the fundamental component and the effective value of the harmonic component, determine the phase correction load in combination with the load state and the harmonic vector coupling relationship, and generate a load sharing adjustment scheme. The switching verification execution module is used to adjust the access phase of each circuit based on the load sharing regulation scheme, perform transient disturbance shielding and collect the neutral line current waveform, combine the fundamental current vector and nonlinear amplification coefficient to obtain the predicted effective value of the neutral line current, and determine whether the neutral line is overloaded. The overload protection control module is used to perform circuit access phase locking and load sharing adjustment scheme updates based on the neutral line overload judgment result, adjust the circuit phase configuration in combination with overload status feedback information, and implement neutral line overload protection control.

2. The system according to claim 1, wherein, The harmonic characteristic analysis module is used to extract the current waveforms of each sub-circuit and the neutral line current waveform based on the historical operation logs of the distribution box's weak current network, and to determine the harmonic pollution level label and neutral line current nonlinear amplification factor corresponding to each sub-circuit, including: By analyzing the historical operation logs of the distribution box's weak current network, current waveform data for each circuit, neutral line current waveform data, and voltage harmonic distortion rate data were obtained. Based on the acquired current waveform data, Fast Fourier Transform was used to extract the fundamental current amplitude, harmonic current amplitude, and harmonic phase angle for each circuit. The harmonic content, total harmonic distortion rate, and harmonic phase angle were then used to construct harmonic feature vectors. K-means clustering was employed to perform cluster analysis on the harmonic feature vectors. Based on the total harmonic distortion rate and harmonic content corresponding to each cluster center, the clustering results were sorted according to pollution level, resulting in low-pollution, medium-pollution, and high-pollution level labels for each circuit. Based on the harmonic feature vectors and the corresponding neutral line current... Based on the flow data and the effective value of the three-phase fundamental unbalanced current, the nonlinear amplification factor of the neutral current is calculated, and the harmonic feature vector and the corresponding nonlinear amplification factor of the neutral current are stored in the distribution box operation monitoring database. Through the distribution box operation monitoring database, historical data of the harmonic feature vector and the nonlinear amplification factor of the neutral current are obtained, and the gradient boosting decision tree algorithm is used to train the model to construct a harmonic-neutral current nonlinear amplification prediction model. Based on the real-time current waveforms of each sub-circuit of the three-phase four-wire distribution box, the real-time harmonic feature vector is extracted using fast Fourier transform, and the nonlinear amplification factor of the neutral current corresponding to each sub-circuit is determined using the harmonic-neutral current nonlinear amplification prediction model.

3. The system according to claim 1, wherein, The active filter compensation module is used to collect neutral current waveforms, grid voltage waveforms, and harmonic phasor data of each circuit based on the harmonic pollution level detection results of each sub-circuit, determine the active filter compensation reference phase, and output the compensation current. It decomposes the current waveform of each circuit into a fundamental component and harmonic components of a preset order in real time, including: Based on the harmonic pollution level of each circuit, it is determined whether there are any circuits with harmonic pollution levels exceeding the preset threshold. If so, the neutral current waveform, grid voltage waveform, and current waveforms of each circuit are acquired in real time. The grid voltage phase is obtained using a synchronous reference coordinate system phase-locked loop (PLL). The amplitude and phase of the preset-order zero-sequence harmonic component in the neutral current are extracted using a second-order generalized integrator. Harmonic phasor data of each circuit are obtained through a current transformer array. The phase curl exponent is calculated to determine the risk of PLL lockout failure and to determine the active filter compensation reference phase. Based on the determined compensation reference phase, the active filter connected in parallel on the outgoing side of the distribution box is activated. Based on the harmonic component amplitude, phase, and reference phase, a harmonic compensation reference current is synthesized. The harmonic compensation reference current is compared with the actual output current. The deviation is used as an error signal. The proportional resonance control algorithm is used to calculate the modulation wave signal of the active filter. The drive pulse generated by SPWM controls the switching device to output a compensation current that is opposite in phase to the neutral line harmonic current. The amplitude of the harmonic compensation reference current is adjusted by closed-loop feedback. When the harmonic component of the preset number in the compensated neutral line current is lower than the preset threshold for a preset number of consecutive power frequency cycles, the effective value of the compensated neutral line current is calculated. The effective value of the neutral line current is compared with the upper limit of the rated neutral line current of the circuit breaker. If it is less than the rated upper limit, the neutral line overload state is determined to be released. The instantaneous current waveform data of each sub-circuit is obtained through the current transformer array. The current waveform of each circuit is decomposed into the fundamental component and the harmonic component of the preset number in real time.

4. The system according to claim 3, wherein, The process of acquiring harmonic phasor data for each circuit through a current transformer array, calculating the phase curl index, determining the risk of phase-locked loop (PLL) loss of lock-in, and determining the active filter compensation reference phase includes: The instantaneous current waveforms of each sub-circuit are acquired using a current transformer array. Fast Fourier Transform (FFT) is then used to extract the amplitude and phase of the same preset-order harmonic components in each sub-circuit, yielding harmonic phasor data for each circuit fed into the neutral line. Based on the phase angles in the harmonic phasor data of each circuit, a harmonic phase angle distribution sequence is constructed. The phase angles are normalized using the grid voltage phase as a reference to obtain the normalized phase angles for each circuit. Finally, the phase curl exponent formula is applied based on the normalized phase angle sequence. Calculate the phase curl index PWI of the current harmonic phase distribution, where, Let be the normalized phase angle of the k-th loop. The average normalized phase angle of all loops is N, where N is the total number of loops. If the phase curl exponent is less than or equal to the preset exponent threshold, the grid voltage phase output in real time by the synchronous reference coordinate system phase-locked loop is used as the reference phase. If the phase curl exponent is greater than the preset exponent threshold, it is determined that there is a risk of phase-locked loop loss in the current harmonic phase distribution. The current phase-locked loop output phase is temporarily stored by updating the output flag bit of the synchronous reference coordinate system phase-locked loop, and the temporarily stored phase is used as the reference phase of the active filter harmonic compensation reference current.

5. The system according to claim 1, wherein, The load sharing optimization allocation module is used to obtain the harmonic pollution factor of each circuit based on the effective values ​​of the fundamental and harmonic components, determine the phase correction load by combining the load state and harmonic vector coupling relationship, and generate a load sharing adjustment scheme, including: Based on the RMS values ​​of the fundamental and harmonic components, the total harmonic RMS value is calculated. Then, based on the total harmonic RMS value and the fundamental RMS value, the harmonic pollution factor for each circuit is obtained. Finally, based on the harmonic pollution factor and instantaneous active power data, the equivalent load factor is calculated using the formula... Calculate the equivalent load rate of each circuit. ,in, Let be the fundamental active power of the i-th circuit. Let be the harmonic active power of the i-th circuit. Let be the harmonic pollution factor of the i-th circuit; obtain the current switching status and maximum current carrying capacity parameters of each circuit through the intelligent circuit breaker communication interface, and map the equivalent load rate to a preset range using normalization processing to obtain the normalized load rate; calculate the harmonic vector coupling factor between circuits in the distribution box based on the amplitude and phase data of the preset harmonic components of each circuit, and determine the phase correction load; based on the phase correction load and the maximum current carrying capacity of each circuit, use a constrained genetic algorithm with the objective function of minimizing the variance of each phase correction load and minimizing the effective value of the predicted neutral current, and with the upper limit of the number of switching operations of each circuit, phase sequence consistency, and circuit breaker capacity constraints as constraints, construct a load sharing regulation optimization model, and generate a load sharing regulation scheme.

6. The system according to claim 5, wherein, The calculation of the harmonic vector coupling factor between circuits in the distribution box, based on the amplitude and phase data of the preset harmonic components of each circuit, and the determination of the phase correction load, includes: Based on the amplitude and phase data of the preset harmonic components of each circuit, the harmonic vector coupling factor formula is used. Calculate the harmonic vector coupling factor between any two circuits i and j within the distribution box. The inter-loop harmonic vector coupling factor matrix is ​​obtained, where, and Let be the current amplitudes of loops i and j under the h-th harmonic, respectively. and These represent the corresponding harmonic current phases; based on the normalized load factor and harmonic vector coupling factor matrix, the phase correction load formula is used. Determine the phase correction load ,in, Let be the normalized load factor of loop i.

7. The system according to claim 1, wherein, The switching verification execution module is used to adjust the access phase of each circuit based on the load sharing adjustment scheme, perform transient disturbance shielding and acquire the neutral line current waveform, combine the fundamental current vector and the nonlinear amplification factor to obtain the predicted effective value of the neutral line current, and determine whether the neutral line is overloaded, including: Based on the load sharing regulation scheme, the intelligent circuit breaker sends switching commands to each circuit to determine the access phase of each circuit. Simultaneously with issuing the switching command, a transient disturbance shielding mechanism is activated, with the shielding window duration set to a preset number of power frequency cycles. During the shielding period, neutral current sampling data is temporarily stored, but overload comparison judgment is not performed. Using a digital signal processor, the current waveforms of each phase are acquired in real time after the shielding window ends. The effective value and peak value of the neutral current are calculated using the sliding window method, with the sliding window length being a preset number of power frequency cycles. Based on the nonlinear amplification factor, combined with the current fundamental current vector sum of each phase and the calculated effective value of the fundamental unbalanced current, the predicted effective value of the neutral current is obtained. The effective value of the neutral current is the product of the effective value of the fundamental unbalanced current and the nonlinear amplification factor. The predicted effective value of the neutral current is compared with the upper limit of the rated neutral current of the circuit breaker to determine whether the neutral line is overloaded.

8. The system according to claim 1, wherein, The overload protection control module is used to perform circuit access phase locking and load sharing adjustment scheme updates based on the neutral line overload judgment result, adjust the circuit phase configuration in conjunction with overload status feedback information, and implement neutral line overload protection control, including: If the neutral line overload judgment result is no overload, the intelligent circuit breaker locks the current access phase of each circuit and generates a load sharing regulation scheme execution completion signal. If the neutral line overload judgment result is overload, the current load sharing regulation scheme is discarded, and the overload judgment result and overload phase sequence identifier are fed back to the load sharing regulation optimization model as new constraints. The load sharing regulation optimization model solution and neutral line overload verification are re-executed until the neutral line overload judgment result is no overload, and the optimized load sharing regulation scheme is determined. If the overload cannot be relieved after reaching the preset maximum number of optimizations, a neutral line overload alarm message is output. The system generates overload protection control commands, puts the distribution box in neutral line overload protection state, locks the current phase of each circuit connection and stops executing new load sharing adjustment operations. At the same time, it controls the active filter to enter continuous compensation mode to suppress neutral line harmonics in real time, and uploads the current neutral line current, load rate of each circuit, harmonic characteristic vector, overload phase sequence identifier and alarm time to the distribution box operation monitoring platform. When the neutral line current is detected to be lower than the upper limit of the rated neutral line current of the circuit breaker for a preset number of power frequency cycles, the neutral line overload protection state is released and the load sharing adjustment optimization model is restarted.

Images