Multi-zone load balancing and overload processing method and device and computer equipment

Abstract

The application relates to a multi-zone load balancing and overload processing method, device and computer equipment. The method comprises the following steps: collecting load operation data of multi-zone distribution transformers and voltage data of a direct-current bus, determining a power redundancy gap coefficient based on the load operation data of each zone distribution transformer, determining a voltage retention rate based on the voltage data, generating a coupling instability risk index based on the power redundancy gap coefficient and the voltage retention rate, determining a target zone requiring power support and a receiving zone capable of receiving power from the multiple zones according to the load state and the coupling instability risk index of each zone, and generating a power coordination instruction, and adjusting the output power of a converter connected between the direct-current bus and each zone through a fuzzy control strategy based on the power coordination instruction. The method can effectively suppress voltage fluctuation and system oscillation caused by power mutation, improve the success rate of overload power transfer and the overall operation safety under the premise of ensuring system dynamic stability.

Description

Technical Field

[0001] This application relates to the field of intelligent control technology for power distribution networks, and in particular to a method, device, and computer equipment for multi-region load balancing and heavy overload handling. Background Technology

[0002] Flexible DC interconnection across multiple distribution substations is a crucial technological direction for achieving load balancing and managing heavy overloads in the distribution network sector. With the increasing digitalization of distribution networks, industrial data acquisition technology has been widely applied to monitor the operational status of distribution substations. Operating parameters such as the load rate of each substation's transformers and the DC bus voltage can be acquired in real time through sensor networks and aggregated to the control center via communication modules. Currently, existing solutions connect multiple distribution substations on the DC side using DC interconnection devices based on the acquired data. Power is allocated according to the load rate of each substation or the rated capacity of the transformers. When a substation experiences a heavy overload due to a sudden load surge or fluctuations in distributed power output, a converter is needed to transfer the excess power to other substations with the capacity to accommodate it, thus achieving power sharing and load balancing.

[0003] However, existing power coordination strategies mostly rely on single steady-state operating indicators such as the static remaining capacity, real-time load rate, or voltage deviation of each candidate receiving area to determine the priority and amount of power allocation. Although such methods can achieve basic power redistribution under normal operating conditions, in complex scenarios where multiple areas participate in power interaction simultaneously, system inertia decreases, or network topology frequently switches, steady-state indicators alone are insufficient to accurately characterize the actual stable bearing boundary of each area during transient processes. This can easily lead to power allocation commands exceeding local voltage stability limits, thereby inducing voltage drops, harmonic amplification, or even cascading instability. As a result, the expected power support is difficult to implement effectively and instead exacerbates the risk of system operation. Summary of the Invention

[0004] Therefore, it is necessary to provide a method, apparatus, and computer equipment for multi-zone load balancing and heavy overload handling that can effectively achieve load balancing and heavy overload handling, in response to the above-mentioned technical problems.

[0005] Firstly, this application provides a method for multi-station load balancing and heavy overload handling, including:

[0006] Collect load operation data and DC bus voltage data from multiple distribution transformers in the area;

[0007] The power redundancy gap coefficient is determined based on the load operation data of each distribution transformer area, and the voltage holding rate is determined based on the voltage data. The power redundancy gap coefficient is used to characterize the relative ability of the distribution transformer area to carry or release additional power without triggering overload protection.

[0008] Based on the power redundancy gap coefficient and the voltage holding rate, a coupling instability risk index is generated;

[0009] Based on the load status of each distribution area and the coupling instability risk index, the target distribution area requiring power support and the receiving distribution area that can accept power are determined from multiple distribution areas, and a power coordination command is generated.

[0010] Based on the power coordination command, the output power of the converter connected between the DC bus and each distribution area is adjusted through a fuzzy control strategy.

[0011] Secondly, this application also provides a multi-zone load balancing and overload handling device, comprising:

[0012] The data acquisition module is used to collect load operation data of multiple distribution transformers and voltage data of the DC bus.

[0013] The parameter determination module is used to determine the power redundancy gap coefficient based on the load operation data of each distribution transformer area, and to determine the voltage holding rate based on the voltage data. The power redundancy gap coefficient is used to characterize the relative ability of the distribution transformer area to carry or release additional power without triggering overload protection.

[0014] An index determination module is used to generate a coupling instability risk index based on the power redundancy gap coefficient and the voltage holding rate;

[0015] The instruction generation module is used to determine the target substation that needs power support and the receiving substation that can accept power from multiple substations based on the load status of each substation and the coupling instability risk index, and to generate power coordination instructions.

[0016] The adjustment module is used to adjust the output power of the converter connected between the DC bus and each distribution area through a fuzzy control strategy based on the power coordination command.

[0017] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in any of the above embodiments of the multi-area load balancing and heavy overload processing method.

[0018] The aforementioned multi-region load balancing and heavy overload handling methods, devices, and computer equipment overcome the limitations of existing technologies that rely solely on a single steady-state index for power allocation by introducing a power redundancy gap coefficient and voltage holding rate to construct a coupled instability risk index. This comprehensively characterizes the dynamic coupling relationship between the active power carrying capacity and voltage support capacity of each region during power transfer, thereby accurately identifying the true stability boundary of each candidate receiving region under transient interactions. Based on this, the target and receiving regions are determined according to this risk index, and power coordination commands are generated. This ensures that power is preferentially allocated to regions with both sufficient power margin and strong voltage holding capacity, reducing the risk of instability induced by excessive power injection into regions with weak voltage stability. Furthermore, by combining fuzzy control strategies to adjust the converter output power, adaptive smooth control of the power transfer process can be achieved under complex multi-region operating conditions, effectively suppressing voltage fluctuations and system oscillations caused by power mutations. This improves the success rate of heavy overload power transfer and overall operational safety while ensuring system dynamic stability. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a diagram illustrating the application environment of a multi-station load balancing and heavy overload handling method in one embodiment.

[0021] Figure 2 This is a flowchart illustrating a multi-zone load balancing and heavy overload handling method in one embodiment;

[0022] Figure 3 This is a flowchart illustrating a multi-zone load balancing and heavy overload handling method in another embodiment;

[0023] Figure 4 This is a flowchart illustrating the steps for determining the coupling instability risk index in one embodiment;

[0024] Figure 5 This is a detailed flowchart illustrating the multi-station load balancing and heavy overload handling method in another embodiment;

[0025] Figure 6 This is a structural block diagram of a multi-station load balancing and heavy overload handling device in one embodiment.

[0026] Figure 7 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0027] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0028] It should be noted that the terms "first," "second," etc., used in this application can be used to describe various elements, but these elements are not limited by these terms. These terms are only used to distinguish the first element from the second element. The terms "comprising" and "having," and any variations thereof, used in this application, are intended to cover non-exclusive inclusion. The term "multiple" used in this application refers to two or more. The term "and / or" used in this application refers to one of the embodiments, or any combination of multiple embodiments.

[0029] The multi-zone load balancing and heavy overload handling method provided in this application embodiment can be applied to, for example... Figure 1 In the application environment shown, the power terminal 102 communicates with the server 104 via a network. The data storage system can store the data that the server 104 needs to process. The data storage system can be integrated onto the server 104, or it can be located in the cloud or on another network server.

[0030] Specifically, server 104 can first collect load operation data of distribution transformers in each distribution area and voltage data of DC bus in real time through communication network; then, server 104 calculates the power redundancy gap coefficient of each distribution area based on the collected load operation data to characterize its relative ability to carry or release additional power, and determines the voltage holding rate based on voltage data, and integrates the two to generate a coupling instability risk index; next, server 104 integrates the real-time load status of each distribution area with the coupling instability risk index to identify the target distribution area that needs power support and the receiving distribution area that can safely accept power from multiple distribution areas, and generates a power coordination command accordingly; the command is sent to power terminal 102 through communication network or directly transmitted to converter control unit, and the fuzzy control strategy dynamically adjusts the output power of the converter connected between DC bus and each distribution area according to the command content.

[0031] Among them, server 104 can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides cloud computing services.

[0032] In one exemplary embodiment, such as Figure 2 As shown, a method for load balancing and heavy overload handling across multiple transformer areas is provided, which is then applied to... Figure 1Taking server 104 as an example, the explanation includes steps 100 to 500. Wherein:

[0033] Step 100: Collect load operation data of multiple distribution transformers and voltage data of DC bus.

[0034] Multi-zone distribution transformers refer to multiple distribution transformers within the same distribution network area, connected together via DC interconnection devices to share their respective low-voltage DC ports. Each distribution transformer independently supplies power to a load area, and the load characteristics of each zone differ. Asynchronous load fluctuations at a given time are the fundamental reason why multi-zone distribution transformers need to be interconnected and mutually supportive. Load operation data refers to the electrical parameters collected from each distribution transformer during operation, including real-time load power and load factor. Real-time load power is the active power calculated by collecting current and voltage values ​​in real time through instrument transformers, measured in kilowatts (kW). It dynamically changes with the connection and disconnection of the zone's load and is a direct basis for determining whether a zone has entered a state of heavy overload. The load factor is the percentage of the current real-time load power to the rated capacity of the distribution transformer in the zone. The DC bus refers to the common conductive busbar connecting the DC sides of each converter in the DC interconnection device. It serves as a common channel for power exchange between distribution zones. Power transmission between any two distribution zones is indirectly completed through bidirectional power flow between their respective converters and the DC bus. The DC bus voltage data refers to the real-time inter-pole voltage value between the positive and negative poles on the DC bus. Since the DC sides of all converters are connected in parallel to the same DC bus, this voltage is the common voltage faced by all converters in each distribution area, reflecting the voltage level of the entire DC interconnection network.

[0035] In practical applications, the server can configure current transformers and voltage transformers at the high-voltage side incoming terminals of each distribution transformer in the distribution area to collect current and voltage values ​​in real time and calculate real-time load power and load factor. Simultaneously, a DC voltage sensor is installed on the DC bus to collect inter-pole voltage values ​​in real time. All data collection is performed at a preset sampling frequency. High-voltage side data collection refers to installing the transformers at the high-voltage side incoming terminals of the distribution transformers, rather than the low-voltage side. This is because the power measured on the high-voltage side includes the transformer's own losses, providing a more accurate reflection of the distribution area's load occupancy from the upstream power grid. The server stores the collected load factor data in chronological order as load factor time-series data, providing a complete time-series basis for subsequent data processing.

[0036] The sampling frequency is determined based on the load change rate of the transformer substation and the system's real-time response requirements. For industrial transformer substations with rapid load changes, the sampling frequency can be set to tens to hundreds of times per second; for substations primarily supplied by residential loads, the sampling frequency can be appropriately reduced. The server synchronously collects load operation data and DC bus voltage data from all transformer substations within each sampling cycle, storing the collected load rate data in chronological order as load rate time-series data. For example, the load rate sampling cycle is set to be less than or equal to half the converter control cycle (i.e., the time interval between the converter control module performing a complete voltage and current calculation and drive signal update). The data interaction cycle of the communication module is synchronized with the control cycle, ensuring that the server can complete the entire calculation process based on the latest batch of data within each control cycle, with an end-to-end delay not exceeding one complete control cycle. The communication module adopts the industrial Ethernet protocol, and exchanges load rate information (the percentage value of the current real-time load power of the transformer in each distribution area to the rated capacity), DC bus voltage information (the real-time voltage value collected by the voltage sensor on the common DC bus) and control commands (the target power command issued by the power coordination control module to each converter and the specific drive signal parameters generated by the converter control module) between the local controllers of each distribution area according to the preset communication cycle, so as to ensure that the load rate data and DC bus voltage data used by each distribution area during power coordination are aligned in time.

[0037] In other embodiments, DC bus voltage data acquisition can employ a redundant sensor configuration. Multiple DC voltage sensors are installed at different locations on the DC bus. The server performs median filtering or weighted averaging on the voltage values ​​collected by each sensor before using the final DC bus voltage data, thereby improving the reliability and anti-interference capability of voltage acquisition. Furthermore, when acquiring load operation data, in addition to real-time load power and load rate, the server can also simultaneously acquire auxiliary electrical parameters such as the power factor, three-phase unbalance, and harmonic distortion rate of the distribution transformers, providing richer information on the operating status of the distribution transformers for subsequent risk assessment.

[0038] Step 200: Determine the power redundancy gap coefficient based on the load operation data of each distribution transformer area, and determine the voltage holding rate based on the voltage data. The power redundancy gap coefficient is used to characterize the relative ability of a distribution transformer area to carry or release additional power without triggering overload protection.

[0039] The power redundancy gap factor is a dimensionless index characterizing the relative ability of a distribution transformer area to carry or release additional power without triggering overload protection. This factor is determined based on the load operation data of each distribution transformer area. The calculation logic is to obtain the ratio of the available power margin of the transformer area (i.e., the upper limit of the active power that the transformer area can carry) to the rated capacity of the transformer area. A factor close to 1 indicates that the transformer area currently has sufficient power redundancy space, while a factor close to 0 indicates that the transformer area is close to full load and has almost no margin to accommodate additional power. Rated capacity is the maximum output power of a distribution transformer as indicated on its nameplate under rated operating conditions, measured in kilovolt-amperes (kVA). Each distribution transformer in a distribution area has its own fixed rated capacity value based on its model and design parameters.

[0040] Voltage hold-up rate is a dimensionless index used to characterize the ability of a transformer substation to maintain DC bus voltage at its connection node. It is determined based on collected voltage data, specifically by comparing the real-time DC bus voltage with the system's preset safety voltage. A voltage hold-up rate greater than 1 indicates that the current voltage is above the safety voltage level, while a rate less than 1 indicates that the current voltage is below the safety voltage level. The lower the value, the weaker the voltage hold-up capability of the node. The system's preset safety voltage is a voltage setpoint determined based on the rated operating voltage range of each converter in the DC interconnection device and the nominal voltage of the DC port on the low-voltage side of the transformer substation. When the DC bus voltage is maintained near this setpoint, each converter is in its optimal operating range, resulting in the highest power transmission efficiency.

[0041] In practice, after receiving load operation data and DC bus voltage data from each distribution area, the server performs parallel calculations of two indicators: the power redundancy gap coefficient and the voltage holding rate. Specifically, for the power redundancy gap coefficient, the server calculates a dimensionless value for each distribution area based on its load operation data, reflecting the relative capacity of that area to receive or release additional power without triggering overload protection. The basic logic of this coefficient is to normalize and compare the current unused power receiving capacity of the distribution area with its total rated capacity. Rated capacity is the maximum output power marked on the nameplate of the distribution transformer under rated operating conditions, measured in kilovolt-amperes (kVA). Each distribution transformer in a distribution area has its own fixed rated capacity value based on its model and design parameters, which is entered into the server database during the system initialization phase. The value of the power redundancy gap coefficient is between 0 and 1: a value close to 1 indicates that almost all of the rated capacity of the distribution area can be used to receive additional power, while a value close to 0 indicates that the distribution area is close to full load and can hardly receive any more additional power. This coefficient answers the core question from a capacity perspective: "How much space does the distribution area have to receive power?"

[0042] For the voltage hold-up rate, the server calculates a dimensionless value reflecting the voltage hold-up capability of the node based on the collected DC bus voltage data. The basic logic of this indicator is to compare the current actual DC bus voltage level with the system's preset safety voltage. The system's preset safety voltage is a voltage setpoint determined by comprehensively considering the rated operating voltage range of each converter in the DC interconnect device and the nominal voltage of the DC port on the low-voltage side of the distribution transformer. The unit is volts. When the DC bus voltage is maintained near this setpoint, each converter is in its optimal operating range, with the highest power transmission efficiency and the lowest voltage stress. A voltage hold-up rate greater than 1 indicates that the current voltage level is sufficient and the distribution transformer has a good voltage support margin when receiving additional power. A value less than 1, and the lower the value, the weaker the voltage hold-up capability of the node, and the higher the risk of voltage drop when receiving additional power. This indicator answers the core question from the perspective of voltage support: whether received power will lead to voltage instability.

[0043] In other embodiments, the power redundancy gap factor can be determined by converting the rated capacity to active power units using the rated power factor before calculating the difference between the rated capacity and the real-time load power, thus unifying the two dimensions in the calculation to active power units. This conversion process can more accurately reflect the actual available active power margin of the distribution transformer area when there is a deviation between the rated power factor and the current operating power factor. In other embodiments, the voltage holding rate can be determined by introducing a moving average filtering process. The server first performs a moving average filtering on multiple consecutive sampled values ​​of the DC bus voltage, using the filtered voltage value as the input. The window length of the moving average filtering is determined based on the ripple frequency of the DC bus voltage and the sampling frequency, typically between 3 and 10 sampling periods, which can effectively suppress the interference of high-frequency voltage ripple and sensor measurement noise on the voltage holding rate.

[0044] Step 300: Based on the power redundancy gap coefficient and voltage retention rate, generate the coupling instability risk index.

[0045] The coupling instability risk index is a quantitative indicator generated jointly based on the power redundancy gap coefficient and voltage hold-up rate. It is used to characterize the degree of risk of a transformer area triggering a voltage-power positive feedback transmission chain when receiving power. The larger the index value, the more severe the mismatch between power acceptance capability and voltage support capability, and the higher the risk of triggering voltage drop positive feedback; the smaller the value, the better the matching between the two, and the higher the safety of received power.

[0046] In practice, after obtaining the power redundancy gap coefficient and voltage maintenance rate for each transformer area, the server generates a coupling instability risk index based on the joint information of these two indicators. The server analyzes the relationship between the power redundancy gap coefficient and the voltage maintenance rate to quantify the matching degree between power acceptance capacity and voltage support capacity. When the power redundancy gap coefficient and voltage maintenance rate are well matched (e.g., both power acceptance capacity and voltage support capacity are sufficient), the generated coupling instability risk index is at a low level. When there is a significant mismatch between the power redundancy gap coefficient and the voltage maintenance rate (e.g., the transformer area can accept power in terms of capacity but lacks voltage support capacity), the generated coupling instability risk index increases significantly. The core feature of this risk index generation lies in its nonlinear response characteristics: when the voltage maintenance rate is only slightly insufficient relative to the power redundancy gap coefficient, the risk index increases at a relatively gentle rate; when the voltage maintenance rate is significantly insufficient relative to the power redundancy gap coefficient, the risk index increases at an accelerated rate. This nonlinear response characteristic is consistent with the actual physical mechanism of the power positive feedback transmission chain, that is, the triggering of the power positive feedback transmission chain is not linearly amplified with the degree of voltage drop, but rather shows an accelerated deterioration trend after passing a certain critical state.

[0047] In other embodiments, the risk index can be generated by further incorporating system parameters such as the transformer capacity level and the scale of the DC interconnection network topology, based on the analysis of the relationship between the power redundancy gap coefficient and the voltage maintenance rate. For interconnection systems composed of large-capacity transformer substations, since the rated power of the converters is relatively large and the voltage fluctuations caused by power transfer are relatively small, the sensitivity of the risk index to voltage changes can be appropriately reduced. For interconnection systems composed of small-capacity transformer substations, since the rated power of the converters is relatively small and the voltage fluctuations caused by power transfer are relatively large, the sensitivity of the risk index to voltage changes can be appropriately increased.

[0048] Step 400: Based on the load status and coupling instability risk index of each transformer area, determine the target transformer area that needs power support and the receiving transformer area that can accept power from multiple transformer areas, and generate a power coordination command.

[0049] Load status refers to the deviation of the actual load rate of each distribution area from its normal operating range under current operating conditions. It can be quantified by indicators such as load rate deviation and is a direct basis for determining whether a distribution area needs power support. Target distribution area refers to the distribution area identified based on load status that needs to obtain power support from other distribution areas; that is, the distribution area whose own load has exceeded the safe operating range and needs to transfer the excess power. Receiving distribution area refers to the distribution area identified based on the coupling instability risk index that has the conditions to safely receive additional power; that is, the distribution area has sufficient voltage support capability and will not trigger significant voltage drops or power positive feedback risks after receiving power. Power coordination command refers to the power control target value of each distribution area's converter generated by the server after completing the matching and power allocation between the target and receiving distribution areas. It is used to indicate the amount and direction of power that each converter should deliver or absorb.

[0050] In practice, after obtaining the coupling instability risk index of each distribution area, the server combines the load operation data of each area and simultaneously generates distribution area role classification and power coordination instructions according to two dimensions: load status and risk index. Specifically, firstly, the server determines whether the current load of each distribution area exceeds the safe operating range, and identifies the distribution areas that urgently need power support as target distribution areas. For each target distribution area, the server determines the total amount of power that needs to be transferred to other distribution areas based on the amount of power exceeding the safe operating range. When the impact source distribution area is marked and its load rate deviation also exceeds the heavy overload threshold, it is identified as the power support distribution area with the highest priority, ensuring that the power that needs to be transferred is allocated first, so as to alleviate the heavy overload pressure caused by the impact load on the source distribution area as quickly as possible. In terms of the coupling instability risk index dimension, the server identifies distribution areas suitable as power receivers based on the risk index of each distribution area, and determines these distribution areas whose risk indices meet the predetermined safety conditions as receiving distribution areas that can accept power. By using the risk index as a screening criterion instead of just the current load rate, the server can effectively exclude those areas that have a remaining margin in terms of capacity statistics but weak voltage support capabilities.

[0051] After role assignment, the server allocates the total power to be transferred from the target substation to each receiving substation and generates power coordination instructions accordingly. The server prioritizes safety when allocating power, allocating power first to the receiving substation with the highest safety rating. If the highest safety rating cannot absorb all the power to be transferred, power is allocated in descending order of safety. Substations with a coupling instability risk index below a preset low-risk threshold are identified as candidate power receiving substations. This means that in a two-dimensional coordination control coordinate system, all substations with a vertical coordinate value below the horizontal line corresponding to the preset low-risk threshold are included in the candidate power receiving substation set. These substations may have various states in terms of capacity, such as light load, medium load, or heavy load, but due to their strong voltage support capability and low risk of coupling instability in the received power, they all possess the basic qualifications for safe power reception. After completing the division of the power support area and the power receiving candidate area, the power coordination control module calculates the total amount of power that needs to be transferred in the power support area, and then allocates the power share to be received by each power receiving candidate area in order of its coupling instability risk index from low to high. Each power receiving share is converted into the target power command value of the strain current transformer, in kilowatts.

[0052] Step 500: Based on the power coordination command, adjust the output power of the converter connected between the DC bus and each distribution area through a fuzzy control strategy.

[0053] The common DC bus refers to the common conductive busbar connecting multiple converters in parallel on their DC sides within the converter control module; it is the physical carrier for power exchange between different transformer substations. A converter is a bidirectional power conversion device connected between the common DC bus and each transformer substation. Each converter can operate in rectification mode (drawing power from the DC bus and injecting it into the substation) or inverter mode (absorbing power from the substation and transmitting it to the DC bus). Power transfer between substations is achieved by adjusting the converter's output power. Fuzzy control strategy refers to a closed-loop control method that uses fuzzy inference to generate converter control quantities. Unlike traditional proportional-integral control, which relies on precise mathematical models and parameter tuning, fuzzy control strategy directly generates control outputs by fuzzifying, matching, and defuzzifying the input deviation and its rate of change, offering advantages such as strong adaptability to system parameter changes and nonlinear characteristics. Output power refers to the active power delivered by the converter to or absorbed from the corresponding substation, measured in kilowatts. The converter adjusts the magnitude and direction of its output power according to power coordination commands to ensure that the power flow between substations aligns with the coordination commands.

[0054] After receiving power coordination commands from each converter, the server adjusts the output power of each converter using a fuzzy control strategy. The basic principle of the fuzzy control strategy is to take the current operating deviation and the trend of deviation change of the system as input, and directly generate control output quantities through preset fuzzy inference rules, thereby driving the converters to change their output power according to the coordination commands.

[0055] During the execution of the fuzzy control strategy, the server first acquires the deviation information between the current DC bus voltage and the system's preset safety voltage, i.e., the magnitude and rate of change of the voltage deviation from the target value. Simultaneously, the server also needs to acquire the deviation information between the actual output current of the converter and the target current value. This deviation information serves as input to the fuzzy inference, reflecting the difference between the current system state and the target state, as well as the trend of this difference.

[0056] The server has a pre-set fuzzy inference rule base containing multiple "if-then" fuzzy inference rules. Each rule provides a corresponding control output based on different combinations of input values. When the input enters the server, it sequentially performs four steps: fuzzification, rule matching, fuzzy inference, and defuzzification. Specifically, first, the precise input value is converted into the membership value of each fuzzy subset; then, the fuzzified input is compared with the preconditions of each rule in the rule base to determine the activation level of each rule; subsequently, the conclusions of all activated rules are combined to generate a fuzzy output set; finally, this fuzzy output set is converted into a precise control output value. Ultimately, the server converts this control output into a drive signal for the converter. The drive signal contains the turn-on time and turn-on sequence information of the switching transistors, directly controlling the switching on and off of each power switching device in the converter, enabling the actual output power of the converter to track the power coordination command.

[0057] The fuzzy control strategy employed in this embodiment has the advantage over traditional proportional-integral control in that its control rules are derived from expert experience and system operating characteristics, rather than from a precise mathematical model of the system. In multi-station DC interconnected systems, due to the large differences and strong randomness in the load characteristics of each station, the system operating conditions are complex and variable, making it difficult to accurately describe with a unified mathematical model. The fuzzy control strategy, with its strong adaptability to nonlinear characteristics and parameter changes, can maintain stable control performance under various operating conditions. Furthermore, through comprehensive analysis of the input deviation and the rate of change of the deviation, the fuzzy control strategy can adaptively adjust the intensity of the control output according to the magnitude of the deviation: when the deviation is small and the system is close to steady state, the output adjustment is small and gradual to ensure control accuracy; when the deviation is large and the system requires a fast response, the output adjustment is large and rapid to accelerate the recovery speed.

[0058] In other embodiments, the fuzzy control strategy can employ different types of fuzzy inference models. For example, in addition to the commonly used Mamdani inference model, the Takagi-Sugeno inference model can also be used. In the latter model, the conclusion of each rule is a linear function of the input variables rather than a fuzzy set, which has certain advantages in terms of the smoothness of the input-output mapping and is suitable for applications requiring high control accuracy. In other embodiments, the frequency of control execution can be dynamically adjusted according to the system's safety status: when the safety status of all transformer areas is good, a normal control frequency is used to ensure steady-state control accuracy; when the system safety status requires a faster response, the control frequency is automatically increased to accelerate the response speed to changes.

[0059] In the aforementioned multi-region load balancing and heavy overload handling method, a coupled instability risk index is constructed by introducing a power redundancy gap coefficient and a voltage holding rate. This overcomes the limitation of existing technologies that rely solely on a single steady-state index for power allocation. It can comprehensively characterize the dynamic coupling relationship between the active power carrying capacity and voltage support capacity of a region during power transfer, thereby accurately identifying the true stability boundary of each candidate receiving region under transient interaction. Based on this, the target region and receiving region are determined according to this risk index, and a power coordination command is generated. This ensures that power is preferentially allocated to regions with both sufficient power margin and strong voltage holding capacity, reducing the excessive injection of power into regions with weak voltage stability that could induce instability. Furthermore, by combining a fuzzy control strategy to adjust the converter output power, adaptive smooth control of the power transfer process can be achieved under complex operating conditions in multiple regions. This effectively suppresses voltage fluctuations and system oscillations caused by power mutations, ultimately improving the success rate of heavy overload power transfer and overall operational safety while ensuring system dynamic stability.

[0060] In one exemplary embodiment, determining the power redundancy gap coefficient based on the load operation data of each transformer area includes:

[0061] For each transformer substation, the rated capacity and current real-time active power of the transformer substation are obtained. Based on the difference between the rated capacity and the current real-time active power, the available power margin of the transformer substation is determined. The ratio of the available power margin to the rated capacity is determined as the power redundancy gap coefficient.

[0062] Current real-time active power refers to the active power value calculated by real-time acquisition of current and voltage values ​​from current transformers and voltage transformers installed on the high-voltage side of the transformer substation. The unit is kilowatts. Unlike reactive power and apparent power, active power is the electrical power actually consumed or output by the transformer substation for performing work, and it changes dynamically with the connection and disconnection of the load. Available power margin refers to the additional active power that the transformer substation can carry under its current operating conditions. It is the difference between the rated capacity of the transformer substation and the current real-time active power, representing the maximum additional power that the substation can accept without triggering overload protection.

[0063] In practice, for each distribution area, the server first obtains the rated capacity of the transformer in that area. Rated capacity is the maximum output power of the distribution transformer as indicated on its nameplate under rated operating conditions, measured in kilovolt-amperes (kVA). Each distribution area's transformer has its own fixed rated capacity value based on its model and design parameters, which is entered into the server database during system initialization. Simultaneously, the server extracts the current real-time active power from the area's load operation data. The server calculates the difference between the rated capacity and the current real-time active power to obtain the available power margin for that area. This value represents the upper limit of additional active power that the area can bear under the current operating conditions. Subsequently, the server divides the available power margin by the rated capacity; the resulting ratio is the power redundancy gap coefficient. This coefficient, a dimensionless value between 0 and 1, directly reflects the proportion of the area's current available power capacity, facilitating horizontal comparison and unified sorting between distribution areas of different capacity levels. In other implementations, before calculating the difference, the server can first convert the rated capacity to active power units using the rated power factor before calculating the difference, unifying the two dimensions to active power units and further improving the physical consistency of the calculation.

[0064] In this embodiment, the difference between the rated capacity of the distribution transformer and the current real-time active power is used as the available power margin, and the ratio of the available power margin to the rated capacity is used as the power redundancy gap coefficient. This provides a simple calculation method with clear physical meaning for quantifying the power acceptance capacity of a distribution transformer area. This coefficient, with a dimensionless value from 0 to 1, directly reflects the proportion of the current available power acceptance space of the distribution transformer area, facilitating horizontal comparison and unified ranking between distribution transformer areas of different capacity levels.

[0065] like Figure 3 As shown, in an exemplary embodiment, step 300 includes:

[0066] Step 310: Determine the ratio of the power redundancy gap coefficient to the voltage holding rate, and determine the deviation of the ratio from the preset unit reference ratio as the original deviation. Perform nonlinear mapping on the original deviation to obtain the coupling instability risk index. Among them, when the voltage holding rate is insufficient relative to the power redundancy gap coefficient, the coupling instability risk index increases nonlinearly.

[0067] The preset unit reference ratio is the ratio corresponding to the ideal matching state of the power redundancy gap coefficient and the voltage maintenance rate, set to a value of 1. At this time, the power acceptance capacity and voltage support capacity of the transformer area are in a balanced matching state. The original deviation refers to the difference between the ratio of the power redundancy gap coefficient to the voltage maintenance rate and the preset unit reference ratio. When the voltage maintenance rate is insufficient relative to the power redundancy gap coefficient, the original deviation is negative, and the larger its absolute value, the more severe the mismatch. Nonlinear mapping refers to the process of converting the original deviation into a coupling instability risk index through an exponential function with the natural constant e as the base. This makes the risk index increase exponentially rather than nonlinearly when the voltage maintenance rate is insufficient relative to the power redundancy gap coefficient, which can more accurately characterize the accelerated deterioration characteristics of the power positive feedback transmission chain near the critical point than the linear mapping method.

[0068] In practice, after obtaining the power redundancy gap coefficient and voltage maintenance rate, the server first calculates the ratio of the two, and then determines the deviation of this ratio from the preset unit benchmark ratio as the original deviation. When the power redundancy gap coefficient equals the voltage maintenance rate, the original deviation is 0, indicating a balanced match between power acceptance capability and voltage support capability. When the voltage maintenance rate is insufficient relative to the power redundancy gap coefficient, the ratio is less than 1, and the original deviation is negative. The server performs an exponential function mapping on the original deviation with the natural constant e as the base to generate a coupling instability risk index. In an exemplary mapping relationship: when the original deviation is 0, the mapping output is 1, and the risk is at its lowest level; when the original deviation is -1, the mapping output is approximately 2.72, and the risk begins to rise significantly; when the original deviation is -2, the mapping output is approximately 7.39, and the risk has entered the high-level range. From the numerical trend, it can be seen that for every unit increase in the negative direction of the original deviation, the risk index increases by approximately 2.72 times compared to the previous level, exhibiting an exponential acceleration. The process of constructing this mapping relationship is as follows: System operation data triggered by continuous power injection is collected under different voltage sag rates. The power redundancy gap coefficient and voltage holding rate are used as input data, and the corresponding voltage sag degree is used as the output label. An exponential function is used for curve fitting to determine the optimal value of the exponent base. When the ratio of the power redundancy gap coefficient to the voltage holding rate is lower than a preset unit reference ratio (i.e., the voltage holding rate is relatively small compared to the power redundancy gap coefficient), the original deviation is negative, and after exponential function mapping, a value greater than 1 is output. When the ratio is higher than the preset unit reference ratio, the original deviation is positive, and after exponential function mapping, a value less than 1 and close to 0 is output. The physical meaning of the exponential function mapping is that when the voltage holding rate is insufficient relative to the power redundancy gap coefficient, the risk of triggering the power positive feedback transmission chain increases exponentially rather than nonlinearly. This nonlinear mapping can more sensitively capture the instability tendency in the early stages of weakened voltage support capability.

[0069] In this embodiment, the deviation between the ratio of the power redundancy gap coefficient to the voltage holding rate and the preset unit reference ratio is used as the original deviation and a nonlinear mapping with an exponential function as the core is performed. This allows the risk index to increase exponentially when the voltage holding rate is insufficient relative to the power redundancy gap coefficient. This enables the instability tendency to be sensitively detected in the early stage of the weakening of voltage support capability, providing a more accurate safety decision basis for power coordinated control.

[0070] like Figure 4 As shown, in an exemplary embodiment, after collecting load operation data of multiple distribution transformers and DC bus voltage data, the method further includes:

[0071] Step 320: Perform time-series differential calculation on the load operation data of each transformer area to obtain the load change rate.

[0072] Step 330: When the load change rate exceeds the preset mutation threshold, mark the corresponding transformer area as the impact source transformer area and record the impact event window characteristics.

[0073] Step 340: Determine the basic risk indicators based on the load operation data and voltage data.

[0074] Step 350: For the transformer areas marked as impact source areas, apply a preset positive offset to the basic risk index based on the impact event window characteristics to obtain the coupled instability risk index.

[0075] Load change rate refers to the rate at which the load changes over time, obtained through time-series differential calculation after collecting load operation data from multiple distribution transformer areas. Specifically, it is the difference between the load rate value at the current sampling moment and the load rate value at the previous sampling moment, divided by the time interval between the two sampling moments, expressed as a percentage per minute, reflecting the speed at which load is added or removed from the distribution area. The preset mutation threshold is a limit value used to determine whether a load change constitutes an abnormal mutation event. It is derived from statistical analysis of historical load fluctuation data for the distribution area. If the maximum load change rate of a distribution area under normal operating conditions does not exceed this threshold, then the load change rate exceeding this threshold is determined to be an abnormal mutation event caused by the addition of impulsive loads or the startup of large-capacity equipment. The impact source distribution area refers to a distribution area where the load change rate exceeds the preset mutation threshold; such distribution areas are experiencing instantaneous heavy overload caused by impulsive loads. The impact event window characteristic refers to the key characteristic quantities recorded within a continuous period when the load change rate exceeds the preset mutation threshold, including the peak time of the load rate and the peak time of the load change rate within that period. The basic risk index refers to the original value of the coupled instability risk index calculated through nonlinear mapping based solely on the power redundancy gap coefficient and voltage hold-up rate, without considering the marking of impulse source areas. The positive offset is a fixed numerical increment added to the basic risk index, which imposes stricter safety constraints on areas marked as impulse source areas during receiver area selection.

[0076] In practice, the server performs differential calculations on the load rate time-series data of each distribution area to obtain the load change rate. The differential calculation is performed once per sampling period, subtracting the load rate value at the current sampling moment from the load rate value at the previous sampling moment, and dividing the difference by the time interval between the two sampling moments to form a continuous load change rate sequence, expressed as a percentage per minute. The server compares the load change rate with a preset mutation threshold. In an exemplary parameter configuration, the preset mutation threshold is set to 1.2 times the rated change rate. When the load change rate of a distribution area exceeds the preset mutation threshold, the server marks that area as a source of the impact and records the characteristics of the impact event window. The server starts timing at the moment the load change rate first exceeds the threshold, continuously comparing the load change rate at each subsequent sampling moment with the threshold. When the load change rate falls back below the threshold and remains below it for a preset number of sampling periods, the period from the first time it exceeds the threshold to the time it stabilizes is defined as the impact event window. If the load change rate exceeds the threshold again within the window, the window is extended backward based on the last stable point, ensuring that the complete transient process of a single impact load is captured. The peak load rate moment refers to the sampling moment corresponding to the maximum load rate value among all load rate sampling data included in the entire impact event window, reflecting when the maximum load pressure on the transformer area occurs due to the impact load. The peak load rate change rate moment refers to the sampling moment corresponding to the maximum absolute value of the load rate change rate among all load rate change rate data included in the entire impact event window, reflecting when the most severe moment of impact load input or output occurs. The two peak moments may overlap or separate, depending on the characteristics of the impact load input. The identification information of the impact source transformer area is a status marker generated by the load mutation sensing unit for the transformer area after determining the impact event window is valid. It includes the transformer area number and the start time of the impact event window, and is sent to the voltage-power coupling instability risk assessment module in the form of a data frame via the communication module. After receiving the identification information, the voltage-power coupling instability risk assessment module adds an impact source flag to the output frame of the coupling instability risk index of the corresponding transformer area, so that the risk index of the transformer area in the current control cycle has a positive offset under the same calculation conditions, thus being more strictly screened in subsequent power coordination decisions.

[0077] Subsequently, the server determines the basic risk indicators based on load operation data and voltage data. For transformer areas marked as impact sources, the server applies a preset positive offset to the basic risk indicators to obtain the final coupling instability risk index for that transformer area. The positive offset is a fixed numerical increment added to the original calculation result of the coupling instability risk index, typically set to 5% to 15% of the maximum possible output value of the coupling instability risk index. The specific value is determined based on system safety operation requirements and statistical data of historical impact events in the transformer area. The purpose of adding the positive offset is to ensure that the ordinate value of the transformer area marked as an impact source is higher than the level without the offset under the same calculation conditions when the power coordination control module generates two-dimensional coordination control coordinates and selects power receiving candidate transformer areas. As a result, it is ranked lower when sorting by coupling instability risk index from low to high, and is more likely to be excluded from the power receiving candidate transformer area qualification screening because it exceeds the low-risk threshold. Ultimately, this achieves a more stringent safety screening of impact source transformer areas when receiving power.

[0078] In this embodiment, the load change rate is detected and the impact source area is marked by time-series differential operation. Simultaneously, the peak moment characteristics within the impact event window are recorded, achieving accurate capture and feature characterization of the complete transient process of the impact load event. Based on this, a positive offset is applied to the basic risk indicators of the impact source area, causing the impact source area to be automatically subject to stricter safety screening when receiving power. This reduces the risk of cascading instability induced by injecting power into areas with uncertain voltage support capabilities during the impact transient process.

[0079] like Figure 5 As shown, in an exemplary embodiment, step 400 includes:

[0080] Step 410: Identify the transformer areas whose load rate exceeds the preset heavy overload threshold as target transformer areas, and identify the transformer areas whose coupling instability risk index is lower than the preset low risk threshold as candidate receiving transformer areas. Sort the candidate receiving transformer areas according to the coupling instability risk index, and preferentially allocate the total heavy overload power to be transferred to the candidate receiving transformer areas ranked higher, until the cumulative power allocation reaches the total heavy overload power to be transferred or all candidate receiving transformer areas reach the allocation limit. Based on the power allocation results, generate power coordination instructions for each converter.

[0081] In this embodiment, the heavy overload threshold is a load rate threshold determined based on the rated capacity of the transformers in each distribution area and the operating procedures. When the real-time load rate of a distribution area exceeds this threshold, the distribution area is determined to be in a heavy overload state, requiring power support to be activated. Candidate receiving distribution areas refer to a set of distribution areas whose coupling instability risk index is lower than a preset low-risk threshold and which meet the basic conditions for safe power reception. The total heavy overload power to be transferred refers to the portion of the real-time load power of the distribution area identified as the target distribution area that exceeds the product of the rated capacity of the transformer and the heavy overload threshold, expressed in kilowatts. Specifically, it is determined by subtracting the product of the rated capacity of the transformer and the heavy overload threshold from the real-time load power of the power support distribution area. If multiple power support distribution areas exist, the total heavy overload power to be transferred for each area is calculated separately and then combined into the total power to be transferred by the system. The allocation upper limit refers to the maximum power value that a candidate receiving distribution area can safely accept under the current power redundancy gap coefficient constraint.

[0082] In practice, the server identifies distribution areas with load rates exceeding a preset heavy overload threshold as target distribution areas. In an exemplary parameter configuration, the heavy overload threshold is set to 0.8 times the rated capacity. For each target distribution area, the server calculates its real-time load power minus the product of the distribution transformer's rated capacity and the heavy overload threshold, as the total heavy overload power to be transferred. The server identifies distribution areas with a coupling instability risk index lower than a preset low-risk threshold (set to 1.2 in the exemplary parameter configuration) as candidate receiving distribution areas, sorting them from low to high according to the coupling instability risk index. The candidate receiving distribution area with the lowest coupling instability risk index is designated as the priority receiving distribution area, whose converter target power command is numerically equal to the total heavy overload power to be transferred but does not exceed the maximum acceptable power corresponding to the current power redundancy gap coefficient of that priority receiving distribution area. If the maximum acceptable power of the priority receiving distribution area is less than the total heavy overload power to be transferred, the converter target power command of the priority receiving distribution area is set to its maximum acceptable power value, and the remaining unallocated power is then allocated sequentially according to the ranking of the remaining candidate receiving distribution areas. This process continues until the cumulative power allocation reaches the total amount to be transferred or all candidate receiving areas reach their respective allocation limits. The server then generates power coordination instructions for each converter based on the final allocation results.

[0083] In this embodiment, by sorting candidate receiving areas according to their coupling instability risk index from low to high and prioritizing power allocation to the candidate receiving area with the lowest risk index, the power allocation efficiency is maximized while meeting safety constraints. The receiving area with the strongest voltage support capability is selected as the priority receiving area to ensure the highest safety of power support. The remaining power is allocated sequentially in ascending order of risk index, thus ensuring a balanced power allocation.

[0084] In an exemplary embodiment, step 500 includes: executing a power coordination command through a fuzzy dual-loop controller to adjust the output power of the converter connected between the DC bus and each distribution area. The fuzzy dual-loop control architecture includes a fuzzy voltage outer loop controller and a fuzzy current inner loop controller.

[0085] Among them, the fuzzy voltage outer loop controller takes the DC bus voltage deviation and voltage deviation change rate as input and generates the current inner loop reference value through fuzzy inference; the fuzzy current inner loop controller takes the deviation and deviation change rate between the current inner loop reference value and the actual output current of the converter as input and generates the converter drive signal through fuzzy inference.

[0086] A fuzzy dual-loop controller refers to a closed-loop control device that employs a nested structure of an outer voltage loop and an inner current loop, with both loops using fuzzy inference to generate control quantities. This differs from conventional dual-loop proportional-integral controllers. The fuzzy voltage outer loop controller is located in the outer loop, taking the DC bus voltage deviation and voltage deviation change rate as inputs, and directly outputs the reference value for the inner current loop through fuzzy inference. The DC bus voltage deviation is the difference between the real-time acquired DC bus voltage and the system's preset safe voltage. The voltage deviation change rate is the difference between the current voltage deviation and the previous voltage deviation divided by the sampling period. These two quantities reflect the magnitude and rate of voltage deviation from the safe value, respectively. When the DC bus voltage deviates from the safe voltage due to power transfer or sudden load changes, the voltage deviation and deviation change rate reflect the magnitude and rate of deviation, respectively, serving as the dual input information for the fuzzy voltage outer loop controller.

[0087] The fuzzy current inner loop controller, located within the inner loop, takes the deviation and rate of change of the deviation between the current inner loop reference value and the actual output current of the converter as input. Through fuzzy inference, it directly outputs the converter drive signal. The current inner loop reference value represents the magnitude of the current that the converter should supply to or absorb from the DC bus to restore the voltage to a safe level under the current voltage deviation condition, measured in amperes. The actual output current of the converter is the current value collected in real time by a current sensor installed at the converter output terminal. This current value serves as the feedback input to the fuzzy current inner loop controller, and is compared with the current inner loop reference value to form the current deviation. The converter drive signal refers to the pulse-width modulation signal ultimately output by the converter control module. It includes the on-time and on-sequence of the switching transistors, directly controlling the on and off of each power switching device in the converter, ensuring that the converter outputs voltage and current corresponding to the target power command.

[0088] In practical implementation, the server executes power coordination commands through a fuzzy dual-loop controller to adjust the output power of each converter. In the fuzzy voltage outer loop controller, the Mamdani fuzzy inference model is used. Its internal structure includes three parts: a fuzzification interface, a fuzzy inference rule base, and a defuzzification interface. The DC bus voltage deviation and voltage deviation rate of change are divided into multiple fuzzy subsets, and a fuzzy inference rule base is established. When the voltage deviation and deviation rate of change enter the system, after fuzzification, rule matching and inference, and centroid defuzzification, the reference value for the inner current loop is directly output. In the fuzzy current inner loop controller, the Mamdani fuzzy inference model is also used. The current deviation and current deviation rate of change are divided into multiple fuzzy subsets. After fuzzification, rule matching, inference, and defuzzification, the duty cycle value of the drive signal is directly output, and then converted into a specific converter drive signal by a pulse width modulation generator.

[0089] Neither the fuzzy voltage outer loop controller nor the fuzzy current inner loop controller contains a proportional-integral (PI) control loop at their outputs. Their internal structure consists only of a fuzzification interface, a fuzzy inference rule base, and a defuzzification interface. There are no proportional gain, integral time constant, or integral accumulator; the output control quantity is directly calculated by the fuzzy inference rules based on the input deviation and the rate of change of deviation. Because there is no integral accumulator, the control output stabilizes once the deviation is eliminated, avoiding the output overshoot, oscillation, and control quantity jumps during mode switching that occur in conventional PI controllers due to integral accumulation and saturation.

[0090] Building upon this, the fuzzy inference rule base of the fuzzy voltage outer loop controller can be further configured to include two rule subsets. The first rule subset is used for fine-tuning when both the DC bus voltage deviation and the rate of change of voltage deviation are within the steady-state small deviation range. It subdivides the voltage deviation and the rate of change of deviation within the steady-state small deviation range into multiple small-interval fuzzy subsets, with a relatively large overlap between adjacent subsets. The membership function uses a triangular function, and the rule conclusions emphasize small-amplitude current corrections to achieve precise maintenance of the DC bus voltage near the safe value. The steady-state small deviation range refers to a voltage deviation absolute value less than or equal to ±2% of the system's preset safe voltage and a voltage deviation rate of change absolute value less than or equal to ±0.5% / s of the system's preset safe voltage. The second rule subset is used for rapid response when the DC bus voltage deviation or voltage deviation rate of change reaches the transient large deviation condition. It divides the voltage deviation and deviation rate of change within the transient large deviation range into a smaller number of fuzzy subsets, each covering a wider numerical range. The membership function uses a trapezoidal function, and the rule conclusion emphasizes the magnitude of current correction, prioritizing the fastest possible return of the voltage to the safe range, while temporarily relaxing the requirements for steady-state accuracy. The transient large deviation condition refers to either the absolute value of the voltage deviation exceeding ±5% of the system's preset safe voltage or the absolute value of the voltage deviation rate of change exceeding ±2% / s of the system's preset safe voltage; either condition is considered met.

[0091] When the coupling instability risk index exceeds a preset risk threshold, the fuzzy voltage outer loop controller compares the received coupling instability risk index with the preset risk threshold in each control cycle. Once the index exceeds the threshold, it immediately switches the currently used first rule subset to the second rule subset, and subsequent fuzzy inference is performed based on the second rule subset. When the risk index falls below the threshold, it automatically switches back to the first rule subset. The rule subset switching mechanism works synergistically with the non-proportional integral control loop: since there is no integral accumulation stage at the output, the controller's output during switching is directly calculated from the switched rule subset based on the current voltage deviation and deviation change rate, unaffected by historical integral accumulation values. This avoids output overshoot or oscillation caused by integral saturation during mode switching in conventional PI controllers, resulting in a smooth and rapid transition from fine adjustment mode to fast response mode.

[0092] In this embodiment, by adopting a pure fuzzy dual-closed-loop control architecture without proportional-integral control loops, both the fuzzy voltage outer loop controller and the fuzzy current inner loop controller directly calculate the output value based on the input deviation and the rate of change of deviation using fuzzy inference rules. This eliminates the output overshoot and control quantity jump problems caused by integral accumulation and integral saturation in conventional proportional-integral controllers, enabling the controller to adaptively adjust the output intensity according to the magnitude and rate of change of the deviation, and smoothly transition between steady-state and transient operating conditions.

[0093] In one exemplary embodiment, after generating the inner current loop reference value through fuzzy inference, the method further includes:

[0094] When the coupling instability risk index of the current transformer area exceeds the preset risk threshold, determine the risk deviation amount and the direction of change of the risk deviation amount.

[0095] Based on the risk deviation amount and the direction of change indicators, the feedforward compensation coefficient is determined through a nonlinear mapping relationship.

[0096] The current increment is determined based on the feedforward compensation coefficient, and the current increment is superimposed on the current inner loop reference value output by the fuzzy voltage outer loop controller to obtain the corrected current inner loop reference value.

[0097] The corrected current inner loop reference value is used as the input of the fuzzy current inner loop controller, and the converter drive signal is generated through fuzzy inference.

[0098] Specifically, when the direction of change indicator indicates an increase in risk, the feedforward compensation coefficient increases according to the first slope curve; when the direction of change indicator indicates a decrease in risk, the feedforward compensation coefficient decreases according to the second slope curve, and the absolute value of the decrease rate of the second slope curve is less than the absolute value of the increase rate of the first slope curve.

[0099] In this embodiment, the preset risk threshold is a threshold value used to determine whether the current coupling instability risk of the transformer area has become severe enough to require additional compensation intervention. When the coupling instability risk index exceeds this threshold value, it indicates that the voltage support capability of the transformer area has weakened to the point that conventional closed-loop regulation alone is insufficient to suppress the voltage drop trend. For example, the risk threshold can be set to 2.5. The risk deviation is the difference between the coupling instability risk index and the preset risk threshold. When the risk index exceeds the threshold, the deviation is positive. The change direction index refers to the change direction of the risk deviation within the continuous control cycle, determined by the first-order difference sign of the risk deviation, reflecting whether the risk is rising or falling. The nonlinear mapping relationship refers to a two-dimensional lookup table used to determine the feedforward compensation coefficient based on the risk deviation and the direction of change indicator. Its core characteristic is that when the direction of change indicator indicates an increase in risk, the feedforward compensation coefficient increases according to the first slope curve; when the direction of change indicator indicates a decrease in risk, the feedforward compensation coefficient decreases according to the second slope curve. Furthermore, the absolute value of the decrease rate of the second slope curve is less than the absolute value of the increase rate of the first slope curve, forming an asymmetric response of rapid compensation when risk increases and slow withdrawal when risk decreases. The feedforward compensation coefficient is a dimensionless coefficient obtained through the nonlinear mapping relationship lookup table, with a value ranging from 0 to 1, comprehensively reflecting the severity and trend of the current coupling instability risk. The current increment refers to the value converted from the feedforward compensation coefficient to the current compensation amount, which is then superimposed on the current inner loop reference value output by the fuzzy voltage outer loop controller to obtain the corrected inner loop reference value.

[0100] In practical implementation, the control cycle is the time interval between the converter control module performing a complete voltage and current dual closed-loop calculation and drive signal update. The duration is determined based on the converter switching frequency and the processing capability of the control chip. Within each control cycle, the control module sequentially executes four steps: voltage outer loop fuzzy inference, current inner loop fuzzy inference, feedforward compensation calculation, and drive signal generation. After the fuzzy voltage outer loop controller generates the current inner loop reference value through fuzzy inference, the server obtains the coupling instability risk index of the current distribution area and compares it with a preset risk threshold. When the risk index exceeds the threshold, the server calculates the risk deviation and the direction of change. The risk deviation is the difference between the coupling instability risk index of the current control cycle and the preset risk threshold. The direction of change index is determined by subtracting the risk deviation of the previous control cycle from the risk deviation of the current control cycle. If the difference is positive, it is set to +1, indicating that the risk is increasing; if the difference is negative, it is set to -1, indicating that the risk is decreasing; and if the difference is 0, it is set to 0, indicating that the risk remains unchanged.

[0101] Subsequently, the server uses the risk deviation and the direction of change index as dual inputs to determine the feedforward compensation coefficients by looking up a pre-built offline two-dimensional nonlinear mapping table. The nonlinear mapping table is a pre-built offline two-dimensional data table constructed by determining the corresponding feedforward compensation coefficients through simulation and experimentation under different combinations of risk deviation values ​​and direction of change index values, storing the three sets of data in rows and columns. The risk deviation is used as the row index, and the direction of change index is used as the column index. The table structure is as follows: rows are evenly divided into multiple intervals based on the risk deviation from 0 to the maximum allowable value; columns are divided into three columns based on the direction of change index: positive 1, negative 1, and 0.

[0102] When the direction of change indicator is positive (risk increasing), the feedforward compensation coefficient increases according to the first slope curve as the risk deviation increases. The first slope curve has a small slope when the risk deviation is small, and gradually increases as the risk deviation increases, showing an overall non-linear increasing trend. This reflects the need to rapidly increase compensation when the risk is rising. The growth rate of the first slope curve reaches its maximum when the risk deviation is large, enabling feedforward compensation to respond at the fastest rate when the risk is severe. When the direction of change indicator is negative (risk decreasing), the feedforward compensation coefficient decays according to the second slope curve. The second slope curve shows a slow decreasing trend, gradually falling from a higher compensation coefficient to a lower value. The decay rate of the second slope curve refers to the rate at which the feedforward compensation coefficient decreases with the risk deviation at each point on the second slope curve, expressed as the decrease in the feedforward compensation coefficient divided by the decrease in the risk deviation. The decay rate of the second slope curve is generally less than the growth rate of the first slope curve, achieving rapid compensation when the risk increases and slow exit when the risk decreases. When used online, if the current risk deviation value is not exactly equal to the row index value in the table, the server uses piecewise linear interpolation to calculate the corresponding feedforward compensation coefficient. The discrete data points in the table are approximated by a continuous broken line representing the first and second slope curves after interpolation. In an exemplary table lookup result: when the risk deviation is 0.2 and the direction of change index is positive 1, the feedforward compensation coefficient is approximately 0.15; when the risk deviation is 0.5 and the direction of change index is positive 1, the feedforward compensation coefficient is approximately 0.40; when the risk deviation is 0.5 and the direction of change index is negative 1, the feedforward compensation coefficient is approximately 0.30, reflecting the asymmetric characteristic that the compensation strength is greater when the risk increases than when the risk decreases under the same risk deviation value.

[0103] After obtaining the feedforward compensation coefficient, the server uses the product of the feedforward compensation coefficient and the rated power of the converter as the power feedforward compensation amount. The rated power of the converter is the maximum output power specified on the nameplate of the bidirectional DC / DC converter under rated operating conditions, in kilowatts. It is the power value on the DC side corresponding to the rated capacity of the distribution transformer in the area connected to the converter, and serves as the benchmark for calculating the power feedforward compensation amount. The power feedforward compensation amount is superimposed on the current inner loop reference value output by the fuzzy voltage outer loop controller in the form of a current increment. The current increment is the power feedforward compensation amount divided by the system preset safety voltage, so that the converter obtains an additional output boost component before the conventional closed-loop regulation has responded, thus compensating for the voltage drop that may be caused by power injection in advance.

[0104] To ensure that the entire process from load change perception to risk index update to feedforward compensation generation is completed within a single control cycle, the load rate sampling period of the acquisition module is less than or equal to half of the control cycle, and the data interaction period of the communication module is synchronized with the control cycle, ensuring that the feedforward compensation unit can generate compensation based on the latest risk index in each control cycle.

[0105] In this embodiment, the feedforward compensation coefficient is determined through a nonlinear mapping relationship based on the risk deviation and the direction of change. An asymmetric mapping strategy, employing a rate of increase of the first slope curve greater than the rate of decrease of the second slope curve, achieves a feedforward compensation effect: rapidly injecting strong compensation to quickly suppress voltage drop trends when the risk of coupling instability increases, and slowly decreasing compensation to maintain a smooth system transition when the risk decreases. The feedforward compensation current increment is superimposed on the inner current loop reference value before the fuzzy closed-loop regulation responds, effectively shortening the delay from risk perception to compensation effectiveness.

[0106] In one exemplary embodiment, the method further includes:

[0107] If the number of times the coupling instability risk index of any transformer area exceeds the preset optimization trigger threshold reaches the preset trigger number, online joint optimization is initiated to adaptively adjust the membership function parameters, fuzzy rule weights, and risk index mapping parameters of the fuzzy dual closed-loop controller.

[0108] The preset optimization trigger threshold is a threshold value for the coupling instability risk index used to determine whether online joint optimization needs to be initiated. Its value is lower than the preset emergency risk threshold but higher than the preset risk threshold, forming a risk-level response mechanism. In an exemplary parameter configuration, it can be set to 4.0. The preset trigger count refers to the cumulative number of control cycles in which the coupling instability risk index continuously exceeds the optimization trigger threshold. Its purpose is to avoid frequent activation of the optimization algorithm due to instantaneous interference or short-term risk fluctuations. Online joint optimization refers to the optimization process that adaptively adjusts the membership function parameters, fuzzy rule weights, and risk index mapping parameters of the fuzzy dual-loop controller simultaneously without interrupting normal control functions during system operation. The membership function parameters refer to the shape parameters of the membership functions of each input-output fuzzy subset in the Mamdani fuzzy inference model, including the vertex position and base width of the triangular membership function and the four inflection points of the trapezoidal membership function. The fuzzy rule weight is the confidence coefficient of the conclusion of each fuzzy inference rule. The risk index mapping parameter refers to the base value of the nonlinear mapping function in the calculation of the coupling instability risk index.

[0109] In practice, the server continuously monitors the coupling instability risk index of each transformer area during each control cycle. When the number of times the coupling instability risk index of any transformer area exceeds the optimization trigger threshold reaches the preset trigger number, online joint optimization is initiated. In an exemplary parameter configuration, the optimization trigger threshold is set to 4.0, which is lower than the preset emergency risk threshold (example value 6.0) but higher than the preset risk threshold (example value 2.5), forming a risk-level response mechanism: when the risk index exceeds 2.5, feedforward compensation is triggered; when it exceeds 4.0, online optimization is performed by adding parameters on the basis of feedforward compensation; and when it exceeds 6.0, emergency protection action is triggered. The server uses a particle swarm optimization algorithm for adaptive adjustment, encoding all parameters to be optimized into a multi-dimensional vector, with each vector as a particle. During initialization, a group of particles is randomly generated to form an initial particle swarm. Each particle updates its own position based on its own historical best position and the group's historical best position. After multiple iterations, the algorithm converges to the optimal parameter combination. The optimized inputs are the coupling instability risk index and DC bus voltage steady-state error for each transformer area in each control cycle. The optimization objective is to minimize the integral value of the DC bus voltage steady-state error and the weighted sum of the power feedforward compensation response delay. The outputs are the optimized membership function parameters, fuzzy rule weights, and risk index mapping parameters. During the optimization process, the optimized parameters replace the current operating parameters in real time, achieving coordinated adaptive adjustment of the evaluation model, control model, and compensation model.

[0110] During the online optimization process, after each generation of particles generates a new parameter combination, the server first inputs the parameter combination into a stability prediction model pre-trained based on historical operating data for verification. This stability prediction model is a limit gradient boosting tree model trained on historical operating data. It is constructed by using historical data on the coupling instability risk index of each transformer area under different operating conditions as the training set, fuzzy controller membership function parameters, fuzzy rule weights, and risk index mapping parameters as input features, and the corresponding trend of coupling instability risk index changes as the output label, trained using the gradient boosting tree algorithm. When used online, after inputting a new parameter combination, the model directly outputs the predicted trend of the coupling instability risk index for each transformer area. Only when the prediction results show that the risk index of all transformer areas does not exceed the preset emergency risk threshold will the server update the parameter combination to the corresponding module. If the prediction results show that the risk index of any transformer area will exceed the preset emergency risk threshold, it is determined that the parameter combination has an instability risk, the parameter combination is abandoned, the current operating parameters are continued, and the optimization failure event is recorded. When the number of consecutive optimization failures reaches the preset limit, the server pauses optimization, maintains the current parameters, and restarts after the system operation status returns to stability.

[0111] In this embodiment, by initiating a particle swarm optimization algorithm to jointly optimize the fuzzy controller membership function parameters, fuzzy rule weights, and risk index mapping parameters online when the coupling instability risk index continuously exceeds the optimization trigger threshold, adaptive adjustment of control parameters to adapt to changes in load characteristics or operating condition drift is achieved during system operation. During the optimization process, a stability prediction model is used to pre-verify new parameter combinations, and parameters are only updated when a safe scenario is predicted, ensuring that adaptive parameter optimization always occurs within the safety boundary.

[0112] To provide a clearer explanation of the multi-transformer area load balancing and heavy overload handling method embodiments provided in this application, a specific embodiment is described below, using a distribution network with three transformer areas (Transformer Area A, Transformer Area B, and Transformer Area C) as an example. The system deploys a DC interconnection device consisting of three bidirectional DC / DC converters. The DC side of each converter is connected in parallel to a common DC bus, and the other side is connected to the low-voltage DC ports of the three distribution transformers respectively.

[0113] Assume it is a weekday morning during peak electricity consumption. Distribution area A experiences a surge in load due to the connection of high-capacity industrial equipment, distribution area B experiences moderate load, and distribution area C experiences light load. The control cycle is set to 10 milliseconds, and the load rate sampling period is 5 milliseconds.

[0114] Within each 5-millisecond sampling period, the server collects real-time load current and voltage through current transformers and voltage transformers installed at the high-voltage side incoming terminals of each distribution transformer, calculating the real-time load power and load factor. Simultaneously, it collects the inter-pole voltage through a DC voltage sensor on the common DC bus. Current sampling data shows that the load factor for distribution transformer A is 92% (rated capacity 500kVA, real-time load power 460kW), the load factor for distribution transformer B is 60% (rated capacity 400kVA, real-time load power 240kW), and the load factor for distribution transformer C is 30% (rated capacity 400kVA, real-time load power 120kW). The common DC bus voltage is 720V (the system's preset safe voltage is 750V). The server stores this data in chronological order in the load factor time-series data.

[0115] First, the server calculates the power redundancy gap factor and voltage hold-up rate for each of the three transformer substations. For substation A, the available power margin is 500kVA minus 460kW, which equals 40kW (adjusted for the rated power factor). The power redundancy gap factor is 40kW divided by 500kVA, approximately 0.08, indicating that substation A is extremely close to full load. For substation B, the power redundancy gap factor is approximately 0.40; for substation C, the power redundancy gap factor is approximately 0.70, indicating that substation C has sufficient power capacity. The voltage hold-up rate is 720V divided by 750V, which equals 0.96. A value below 1 indicates that the current DC bus voltage is slightly below the safe voltage level.

[0116] Next, the server invokes the coupling instability risk index calculation unit to calculate the risk index for each of the three transformer areas. First, it calculates the ratio of the power redundancy gap coefficient to the voltage holding rate (0.96): for transformer area A, it's approximately 0.083 (0.08 / 0.96); for transformer area B, it's approximately 0.417 (0.40 / 0.96); and for transformer area C, it's approximately 0.729 (0.70 / 0.96). All three ratios are less than the preset unit reference ratio of 1, with initial deviations of -0.917 (transformer area A), -0.583 (transformer area B), and -0.271 (transformer area C), respectively. After exponential function mapping, the coupling instability risk index for transformer area A is approximately 2.50, for transformer area B it's approximately 1.79, and for transformer area C it's approximately 1.31. During this period, the server simultaneously performs differential calculations on the load rate time-series data to obtain the load change rate. Differential calculations over recent sampling periods show that the load change rate of transformer area A is +15% per minute (exceeding the preset mutation threshold of 1.2 times the rated change rate). The server determines that transformer area A is experiencing an impact load event, marks it as the impact source transformer area, and records the impact event window characteristics. Based on the impact source marking, the server adds a positive offset of 0.15 to the risk index of 2.50 for transformer area A, resulting in a final coupling instability risk index of 2.65 for transformer area A.

[0117] Subsequently, the power coordination control module in the server performs the division of substation roles and the generation of power coordination instructions. The preset heavy overload threshold is 0.8 times the rated capacity. Substation A's load rate of 92% exceeds this threshold and is identified as the target substation for power support. The total heavy overload power to be transferred is 460kW minus 500kVA multiplied by 0.8, which equals 60kW. Substation B's load rate is 60%, and Substation C's load rate is 30%, both of which are within the threshold. The preset low-risk threshold is 1.2. The risk indices of Substation B (1.79) and Substation C (1.31) both exceed the threshold. Under the current weak condition of a voltage hold-up rate of 0.96, although Substation C does not meet the standard receiving qualification, its risk index is closest to the low-risk threshold. After comprehensive evaluation, the server allows Substation C to receive power with a limited share, sets its allocation share to 40kW, and generates the corresponding power coordination instructions. The remaining 20kW will not be allocated in the current cycle and will be maintained by Substation A relying on its own short-term overload capacity. It will be reallocated after the voltage conditions improve.

[0118] After receiving the power coordination command, the converter control module adjusts the output power of the converter in transformer area C through a fuzzy dual-loop controller. The fuzzy voltage outer loop controller takes the current voltage deviation of -30V (720V minus 750V) and the voltage deviation change rate as input, and generates a current inner loop reference value through Mamdani fuzzy inference. During this period, the server detects that the coupling instability risk index of transformer area A (2.65) exceeds the preset risk threshold of 2.5. The feedforward compensation unit calculates the risk deviation of 0.15 and the direction of change (positive 1, risk increasing). The feedforward compensation coefficient is obtained from the nonlinear mapping table as approximately 0.12. The feedforward compensation is superimposed on the current inner loop reference value output by the fuzzy voltage outer loop controller in the form of current increments. The corrected inner loop reference value enters the fuzzy current inner loop controller, which generates a converter drive signal through fuzzy inference. This signal is then controlled by pulse width modulation to control the conduction timing of the converter switching transistors, enabling the converter to draw 40kW of power from the common DC bus and inject it into transformer area C.

[0119] Upon entering the next control cycle, the load factor of transformer area A decreased from 92% to 84% (420kW divided by 500kVA) due to the 40kW power being transferred to transformer area C, still exceeding the allowable load by 20kW. Transformer area C, receiving 40kW of power, saw its load factor rise from 30% to 40% (160kW divided by 400kVA). Under the combined effect of fuzzy control regulation and feedforward compensation, the DC bus voltage recovered from 720V to 735V, and the voltage hold-up rate recovered to 0.98. The server recalculated the coupling instability risk index for each transformer area: the power redundancy gap coefficient of transformer area C decreased to approximately 0.60, the voltage hold-up rate increased to 0.98, the ratio 0.60 / 0.98 was approximately 0.612, the original deviation was approximately -0.388, and the mapped risk index was approximately 1.47. The risk index for transformer area B was approximately 1.65. In the current cycle, the risk index of area C is relatively low at 1.47. The server will allocate the remaining 20kW to area C, generate a new power coordination command, and complete the complete transfer of the excess power of 60kW from area A.

[0120] The entire scheme described above, from the synchronous acquisition of raw data, the calculation of power redundancy gap coefficient and voltage holding rate, the generation of coupling instability risk index (including impact source marking and offset addition), the safety constraint decision of target area and receiving area, the generation of power coordination command, to the entire process of converter drive with fuzzy dual closed-loop control and feedforward compensation, is completed collaboratively within a single control cycle (10 milliseconds).

[0121] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages in other steps. It is understood that the steps in different embodiments can be freely combined as needed, and all non-contradictory solutions formed by such combinations are within the scope of protection of this application.

[0122] Based on the same inventive concept, this application also provides a multi-region load balancing and heavy overload processing device for implementing the multi-region load balancing and heavy overload processing method described above. The solution provided by this device is similar to the solution described in the above method. Therefore, the specific limitations of the one or more multi-region load balancing and heavy overload processing device embodiments provided below can be found in the limitations of the multi-region load balancing and heavy overload processing method described above, and will not be repeated here.

[0123] In one exemplary embodiment, such as Figure 6 As shown, a multi-zone load balancing and overload handling device 600 is provided, including: a data acquisition module 610, a parameter determination module 620, an index determination module 630, an instruction generation module 640, and an adjustment module 650, wherein:

[0124] The data acquisition module 610 is used to collect load operation data of multiple distribution transformers and voltage data of DC bus.

[0125] The parameter determination module 620 is used to determine the power redundancy gap coefficient based on the load operation data of each distribution transformer area, and to determine the voltage holding rate based on the voltage data. The power redundancy gap coefficient is used to characterize the relative ability of the distribution transformer area to carry or release additional power without triggering overload protection.

[0126] The index determination module 630 is used to generate a coupling instability risk index based on the power redundancy gap coefficient and voltage retention rate.

[0127] The instruction generation module 640 is used to determine the target substation that needs power support and the receiving substation that can accept power from multiple substations based on the load status and coupling instability risk index of each substation, and to generate power coordination instructions.

[0128] The regulation module 650 is used to regulate the output power of the converter connected between the DC bus and each distribution area through a fuzzy control strategy based on power coordination commands.

[0129] In an exemplary embodiment, the parameter determination module 620 is further configured to, for each transformer area, obtain the rated capacity and current real-time active power of the transformer in the transformer area; determine the available power margin of the transformer area based on the difference between the rated capacity and the current real-time active power; and determine the ratio of the available power margin to the rated capacity as the power redundancy gap coefficient.

[0130] In an exemplary embodiment, the parameter determination module 620 is further configured to determine the ratio of the power redundancy gap coefficient to the voltage retention rate, and to determine the deviation of the ratio from the preset unit reference ratio as the original deviation.

[0131] The original deviation is nonlinearly mapped to obtain the coupling instability risk index; among which, when the voltage retention rate is insufficient relative to the power redundancy gap coefficient, the coupling instability risk index increases nonlinearly.

[0132] In an exemplary embodiment, the parameter determination module 620 is further configured to perform time-series differential calculation on the load operation data of each transformer area to obtain the load change rate; when the load change rate exceeds a preset sudden change threshold, the corresponding transformer area is marked as an impact source transformer area, and the impact event window characteristics are recorded; based on the load operation data and voltage data, a basic risk index is determined; for the transformer areas marked as impact source transformer areas, a preset positive offset is applied to the basic risk index based on the impact event window characteristics to obtain the coupling instability risk index.

[0133] In an exemplary embodiment, the instruction generation module 640 is further configured to: identify the transformer areas whose load rate exceeds a preset heavy overload threshold as target transformer areas; identify the transformer areas whose coupling instability risk index is lower than a preset low-risk threshold as candidate receiving transformer areas; sort the candidate receiving transformer areas according to the coupling instability risk index; preferentially allocate the total heavy overload power to be transferred to the candidate receiving transformer areas ranked higher, until the cumulative power allocation reaches the total heavy overload power to be transferred or all candidate receiving transformer areas reach the allocation limit; and generate power coordination instructions for each converter based on the power allocation results.

[0134] In an exemplary embodiment, the adjustment module 650 is further configured to execute power coordination commands through a fuzzy dual-loop controller to adjust the output power of the converter connected between the DC bus and each distribution area. The fuzzy dual-loop control architecture includes a fuzzy voltage outer loop controller and a fuzzy current inner loop controller. The fuzzy voltage outer loop controller takes the DC bus voltage deviation and voltage deviation change rate as inputs and generates a current inner loop reference value through fuzzy inference. The fuzzy current inner loop controller takes the deviation between the current inner loop reference value and the actual output current of the converter and the deviation change rate as inputs and generates a converter drive signal through fuzzy inference.

[0135] In an exemplary embodiment, the adjustment module 650 is further configured to: determine the risk deviation amount and the direction of change of the risk deviation amount when the coupling instability risk index of the current transformer area exceeds a preset risk threshold; determine the feedforward compensation coefficient based on the risk deviation amount and the direction of change of the risk deviation amount through a nonlinear mapping relationship; determine the current increment based on the feedforward compensation coefficient, and superimpose the current increment onto the current inner loop reference value output by the fuzzy voltage outer loop controller to obtain a corrected current inner loop reference value; use the corrected current inner loop reference value as the input of the fuzzy current inner loop controller, and generate a converter drive signal through fuzzy inference; wherein, when the direction of change of the risk indicator indicates an increase in risk, the feedforward compensation coefficient increases according to a first slope curve; when the direction of change of the risk indicator indicates a decrease in risk, the feedforward compensation coefficient decreases according to a second slope curve, and the absolute value of the decrease rate of the second slope curve is less than the absolute value of the increase rate of the first slope curve.

[0136] In an exemplary embodiment, the adjustment module 650 is further configured to initiate online joint optimization when the number of times the coupling instability risk index of any station area exceeds the preset optimization trigger threshold reaches the preset trigger number, and adaptively adjust the membership function parameters, fuzzy rule weights, and risk index mapping parameters of the fuzzy dual closed-loop controller.

[0137] Each module in the aforementioned multi-zone load balancing and heavy overload handling device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of a computer device in software form, so that the processor can call and execute the corresponding operations of each module.

[0138] In one exemplary embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 7As shown, this computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs stored in the non-volatile storage media. The database stores load operation data for multiple distribution transformers and DC bus voltage data. The I / O interfaces are used for information exchange between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a multi-region load balancing and overload handling method.

[0139] Those skilled in the art will understand that Figure 7 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0140] In one exemplary embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps in any of the above embodiments of the multi-area load balancing and heavy overload processing method.

[0141] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored, which, when executed by a processor, implements the steps in any of the above embodiments of the multi-area load balancing and heavy overload processing method.

[0142] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in any of the above embodiments of the multi-area load balancing and heavy overload processing method.

[0143] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0144] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0145] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0146] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for load balancing and heavy overload handling across multiple transformer areas, characterized in that, include: Collect load operation data and DC bus voltage data from multiple distribution transformers in the area; The power redundancy gap coefficient is determined based on the load operation data of each distribution transformer area, and the voltage holding rate is determined based on the voltage data. The power redundancy gap coefficient is used to characterize the relative ability of the distribution transformer area to carry or release additional power without triggering overload protection. Based on the power redundancy gap coefficient and the voltage holding rate, a coupling instability risk index is generated; Based on the load status of each distribution area and the coupling instability risk index, the target distribution area requiring power support and the receiving distribution area that can accept power are determined from multiple distribution areas, and a power coordination command is generated. Based on the power coordination command, the output power of the converter connected between the DC bus and each distribution area is adjusted through a fuzzy control strategy.

2. The method according to claim 1, characterized in that, The determination of the power redundancy gap coefficient based on the load operation data of each transformer area includes: For each transformer substation, obtain the rated capacity and current real-time active power of the transformer in that substation. Based on the difference between the rated capacity and the current real-time active power, the available power margin of the transformer area is determined; The ratio of the available power margin to the rated capacity is determined as the power redundancy gap coefficient.

3. The method according to claim 1, characterized in that, The generation of the coupling instability risk index based on the power redundancy gap coefficient and the voltage holding rate includes: Determine the ratio of the power redundancy gap coefficient to the voltage retention rate, and define the deviation of the ratio from the preset unit reference ratio as the original deviation; The original deviation is nonlinearly mapped to obtain the coupling instability risk index; When the voltage retention rate is insufficient relative to the power redundancy gap coefficient, the coupling instability risk index increases nonlinearly.

4. The method according to claim 1, characterized in that, After collecting load operation data from multiple distribution transformers and DC bus voltage data, the process also includes: The load change rate is obtained by performing time-series differential calculation on the load operation data of each transformer area; If the load change rate exceeds a preset sudden change threshold, the corresponding transformer area is marked as an impact source transformer area, and the impact event window characteristics are recorded. Based on the load operation data and the voltage data, determine the basic risk indicators; For a transformer area marked as an impact source area, a preset positive offset is applied to the basic risk index based on the impact event window characteristics to obtain the coupling instability risk index.

5. The method according to claim 1, characterized in that, Based on the load status of each distribution area and the coupling instability risk index, the system determines the target distribution area requiring power support and the receiving distribution area with acceptable power from multiple distribution areas, and generates a power coordination command, including: The transformer area whose load rate exceeds the preset heavy overload threshold is identified as the target transformer area; Substations with a coupling instability risk index below a preset low-risk threshold are identified as candidate receiving substations. The candidate receiving stations are sorted according to the coupling instability risk index. The total amount of heavy overload power to be transferred is preferentially allocated to the candidate receiving stations ranked higher, until the cumulative power allocation reaches the total amount of heavy overload power to be transferred or all candidate receiving stations reach the allocation limit. Based on the power allocation results, power coordination instructions for each converter are generated.

6. The method according to any one of claims 1 to 5, characterized in that, The adjustment of the output power of the converter connected between the DC bus and each distribution area using a fuzzy control strategy based on the power coordination command includes: The power coordination command is executed by a fuzzy dual-loop controller to adjust the output power of the converter connected between the DC bus and each distribution area. The fuzzy dual-loop control architecture includes a fuzzy voltage outer loop controller and a fuzzy current inner loop controller. The fuzzy voltage outer loop controller takes the DC bus voltage deviation and voltage deviation change rate as input and generates the current inner loop reference value through fuzzy inference; the fuzzy current inner loop controller takes the deviation and deviation change rate between the current inner loop reference value and the actual output current of the converter as input and generates the converter drive signal through fuzzy inference.

7. The method according to claim 6, characterized in that, After generating the inner current loop reference value through fuzzy inference, the method further includes: When the coupling instability risk index of the current transformer area exceeds a preset risk threshold, determine the risk deviation amount and the direction index of the change of the risk deviation amount; The feedforward compensation coefficient is determined through a nonlinear mapping relationship based on the risk deviation and the direction of change index. The current increment is determined based on the feedforward compensation coefficient, and the current increment is superimposed on the current inner loop reference value output by the fuzzy voltage outer loop controller to obtain the corrected current inner loop reference value. The corrected current inner loop reference value is used as the input of the fuzzy current inner loop controller, and the converter drive signal is generated through fuzzy inference. Specifically, when the change direction indicator indicates an increase in risk, the feedforward compensation coefficient increases according to the first slope curve; when the change direction indicator indicates a decrease in risk, the feedforward compensation coefficient decreases according to the second slope curve, and the absolute value of the decrease rate of the second slope curve is less than the absolute value of the increase rate of the first slope curve.

8. The method according to claim 6, characterized in that, The method further includes: If the number of times the coupling instability risk index of any transformer area exceeds the preset optimization trigger threshold reaches the preset trigger number, online joint optimization is initiated to adaptively adjust the membership function parameters, fuzzy rule weights, and risk index mapping parameters of the fuzzy dual closed-loop controller.

9. A multi-zone load balancing and overload handling device, characterized in that, The device includes: The data acquisition module is used to collect load operation data of multiple distribution transformers and voltage data of the DC bus. The parameter determination module is used to determine the power redundancy gap coefficient based on the load operation data of each distribution transformer area, and to determine the voltage holding rate based on the voltage data. The power redundancy gap coefficient is used to characterize the relative ability of the distribution transformer area to carry or release additional power without triggering overload protection. An index determination module is used to generate a coupling instability risk index based on the power redundancy gap coefficient and the voltage holding rate; The instruction generation module is used to determine the target substation that needs power support and the receiving substation that can accept power from multiple substations based on the load status of each substation and the coupling instability risk index, and to generate power coordination instructions. The adjustment module is used to adjust the output power of the converter connected between the DC bus and each distribution area through a fuzzy control strategy based on the power coordination command.

10. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 8.

Images