An electric energy management system and method suitable for photovoltaic storage and charging

Abstract

The present application relates to the technical field of electric energy management, and particularly relates to an electric energy management system and method suitable for photovoltaic storage and charging. The method comprises the following steps: obtaining a multi-source physical parameter set of a photovoltaic storage and charging station, and performing multi-source electric energy access physical topology analysis to obtain a physical access topology matrix; performing electric energy back-feeding charging gun classification based on the physical access topology matrix to obtain an access charging gun classified electric parameter data frame; performing grid-connected peak shaving regulation simulation on the access charging gun classified electric parameter data frame according to the physical access topology matrix to obtain a feasible power channel table; performing channel-by-channel regulation instruction structuring on the feasible power channel table to obtain a structured regulation instruction set; obtaining execution equipment feedback data, detecting an asymmetric load offset channel; performing storage and charging station safety reset on the asymmetric load offset channel, and performing dynamic balance redistribution to obtain a structured reset regulation instruction set. The present application helps to improve the operation stability and energy utilization rate of the photovoltaic storage and charging system.

Description

Technical Field

[0001] This invention relates to the field of power management technology, and in particular to a power management system and method suitable for photovoltaic energy storage and charging. Background Technology

[0002] The organic integration of photovoltaic power generation, energy storage systems, and electric vehicle charging facilities to construct an integrated "photovoltaic-energy storage-charging" system has become an important development direction for green energy utilization. In such systems, power management methods, as the core supporting technology, primarily undertake the key tasks of coordinating the operating status of various energy subsystems, allocating power flow in real time, ensuring stable system operation, and improving energy efficiency. The quality of the power management method directly determines the comprehensive performance of photovoltaic energy utilization, electric vehicle charging efficiency, and energy storage unit lifespan. Therefore, researching an efficient power management method suitable for integrated photovoltaic-energy storage-charging scenarios has become an important topic in smart energy control technology.

[0003] In traditional technologies, photovoltaic systems, energy storage systems, and charging piles typically employ a decentralized control approach, with each control logic operating independently and lacking a unified scheduling and coordination mechanism. This approach often exhibits lag in response to complex real-time power fluctuations, load changes, and dynamic electricity price adjustments, sometimes leading to issues such as power backflow, grid harmonic pollution, and frequent charging and discharging of energy storage, severely restricting the overall operational efficiency and economic viability of the system. Especially in environments with concurrent charging demands from multiple users and vehicles, traditional methods struggle to dynamically predict load demand and fail to fully account for the uncertainties of photovoltaic power generation, resulting in frequent problems such as blind energy allocation, misuse of energy storage, and energy waste. Summary of the Invention

[0004] Therefore, it is necessary for the present invention to provide a power management system and method suitable for photovoltaic energy storage and charging, in order to solve at least one of the above-mentioned technical problems.

[0005] To achieve the above objectives, a power management method suitable for photovoltaic energy storage and charging includes the following steps: Step S1: Obtain the multi-source physical parameter set of the photovoltaic energy storage and charging station, and perform multi-source power access physical topology analysis on the multi-source physical parameter set of the photovoltaic energy storage and charging station to obtain the physical access topology matrix; Step S2: Perform high-frequency sampling of physical connections based on the physical access topology matrix to obtain electrical status data frames; classify the electrical status data frames for reverse power charging guns to obtain access charging gun classification electrical parameter data frames; Step S3: Select high-power charging and discharging path channels for the classified electrical parameter data frames of the access charging guns according to the physical access topology matrix to obtain the photovoltaic charging and discharging path set; perform grid-connected peak shaving regulation simulation on the photovoltaic charging and discharging path set to obtain the feasible power channel table; Step S4: Structure the channel-by-channel control instructions in the feasible power channel table to obtain a structured control instruction set, and send the structured control instruction set to the execution equipment one by one through the photovoltaic energy storage and charging station management platform to execute the control instructions; Step S5: Obtain feedback data from the execution device; identify execution deviation based on the feedback data from the execution device, and detect the asymmetric load offset channel based on the execution deviation identification result; Step S6: Perform a safety reset of the asymmetric load offset channel for the energy storage and charging station, and perform dynamic equalization redistribution based on the safety reset result to obtain a structured reset control instruction set; send the structured reset control instruction set to the execution equipment one by one through the photovoltaic energy storage and charging station management platform to execute the control instructions.

[0006] This invention constructs a closed-loop control system to achieve a complete energy management process, from physical parameter sensing, path selection, control issuance to deviation identification and reset / reset. This significantly improves the system's adaptability and response speed to complex operating conditions. Specifically, by unifying the encapsulation of multi-source physical parameters and modeling the topology of photovoltaic energy storage stations, accurate identification of physical access relationships across devices and channels is achieved, effectively solving the path selection confusion problem caused by dispersed control logic in traditional methods. Introducing parameters such as voltage fluctuation amplitude greater than 10V, duration not less than 0.5 seconds, and current direction consistency as screening criteria in the high-frequency sampling stage of physical connections enables more accurate identification of reverse feed behavior, avoiding misclassification or omissions and improving the reliability of charging gun status identification. During path selection, injecting reverse feed capability tags and updating port status in real time, combined with the rated load parameters of each path, high-power path screening ensures efficient use of high-power channels while avoiding path overload. In the grid-connected control simulation stage, setting a transformer capacity upper limit, such as 13... For 6 kW and single-channel power limits of 20 kW, dynamic peak shaving judgment and path backoff mechanisms are implemented to effectively suppress the disturbance of instantaneous power surges to the power grid and ensure grid connection stability. In the subsequent structured steps of control commands, the method of binding path numbers with physical control templates enables precise control of the execution equipment and improves the response efficiency of command issuance and execution. In the deviation identification stage, a load imbalance coefficient ≥0.75 is introduced as an asymmetric load screening threshold, and structural parameters such as branch density >0.35 and channel depth ≥4 are superimposed for comprehensive evaluation, which helps to identify critical paths that have a significant impact on the system power balance. Finally, a power distribution compensation map is formed by disconnection and differential pressure reconstruction, which not only realizes the rapid diversion of power from disconnected paths, but also improves the utilization efficiency of backup channels and the balance of power allocation of the entire station. Therefore, this invention has the advantages of strong real-time performance, high accuracy, and good controllability in complex multi-source power environments, and significantly improves the operational stability and energy utilization rate of photovoltaic energy storage and charging systems.

[0007] Optionally, step S1 specifically includes: Step S11: Obtain the multi-source physical parameter set of the photovoltaic energy storage and charging station, and encapsulate it in a structured manner to obtain a structured physical parameter dataset; Step S12: Perform port mapping analysis and bind physical connection identifiers on the structured physical parameter dataset to obtain the physical port information table; Step S13: Construct a multi-source device interconnection channel relationship diagram based on the physical port information table; Step S14: Analyze the multi-source access path combination and logical dependency relationship between each type of device based on the multi-source device interconnection channel relationship diagram to obtain the photovoltaic energy storage and charging station access path structure; Step S15: Based on the access path structure of the photovoltaic energy storage and charging station, perform topology integration to obtain the physical access topology matrix.

[0008] This invention achieves unified management of underlying parameters such as voltage, current, power direction, and port conductance status from heterogeneous devices like photovoltaic modules, energy storage units, and charging piles through structured encapsulation of multi-source physical parameters, improving the efficiency and accuracy of data parsing. Subsequently, a port mapping analysis and physical connection identifier binding mechanism are introduced to ensure the unique identification of each device interface and its corresponding connection relationship, effectively avoiding path identification confusion caused by port misalignment or multiple access points in traditional methods. By constructing a multi-source device interconnection channel relationship diagram, the system can comprehensively grasp the actual physical connection status between photovoltaic, energy storage, and load, providing reliable support for subsequent path regulation. During access path analysis, by identifying path combinations and logical dependencies, the sequential relationship and power transmission direction of each device during charging and discharging can be revealed, improving the logical consistency of the regulation strategy. The final physical access topology matrix covers key indicators such as the structural depth, number of connection hops, and branch density of each path, effectively improving the system's adaptability to high-load conditions and path fault tolerance.

[0009] Optionally, step S14 specifically includes: Step S141: Perform path identification and classification on the starting and ending devices of each path in the interconnection channel relationship diagram to obtain a path identification index list; Step S142: Traverse the path identification index list to identify the relay devices for each path, and extract the electrical parameters of the relay devices to obtain the path relay dependency table; Step S143: Based on the minimum allowable current and maximum supported voltage limits of each relay device in the path relay dependency table, calculate the maximum throughput capacity of the path, and filter and remove paths with low throughput capacity to obtain a set of feasible paths; Step S144: Reclassify feasible paths according to the origin device source type of the feasible access path set to construct a multi-source access path set; Step S145: Model the logical dependency relationship based on the multi-source access path set to generate the access path structure of the photovoltaic energy storage and charging station.

[0010] This invention improves the logical clarity and scheduling efficiency of photovoltaic (PV) energy storage and charging station path control by systematically sorting out the relationships between interconnected channels. Identifying and classifying starting and ending points of equipment helps clarify the origin and destination of power flow, avoiding ambiguity in path identification. Traversing and extracting relay equipment allows the system to monitor electrical bottlenecks in the path in real time. Combined with parameter constraints of minimum allowable current and maximum supported voltage, it effectively filters paths with insufficient carrying capacity, avoiding power congestion or backflow risks caused by equipment limitations during actual control. Path reclassification further refines the attribution relationships of multi-source access paths, facilitating the development of differentiated control strategies based on different energy sources. Finally, logical dependency modeling ensures that the path structure has a resolvable control sequence, providing a structured basis for hierarchical distribution of scheduling commands, priority setting, and conflict handling, thereby comprehensively improving the stability, response speed, and energy utilization efficiency of the PV energy storage and charging system.

[0011] Optionally, the specific classification of the reverse-flow charging guns in step S2 is as follows: Real-time decoding of the charging gun interface layer is performed based on the electrical status data frame to obtain the interface electrical parameter identification data frame. When the voltage fluctuation amplitude is greater than 10V within 0.5 seconds and the current current direction is consistent with the direction of at least two of the first three frames in the interface electrical parameter identification data frame, the current interface electrical parameter of the charging gun is determined to be valid, so as to filter the valid interface electrical parameter data points in the interface electrical parameter identification data frame. When a valid interface electrical parameter data point is determined to be valid more than 3 times within 2 seconds, the data frame corresponding to the valid interface electrical parameter data point is marked as a valid interface electrical parameter data frame. The signal voltage, communication status and power direction of each charging gun in the current data frame of the effective interface electrical parameter data frame are jointly interpreted to extract the reverse feed-in judgment feature group; Based on the reverse feed-in judgment feature group, the valid interface electrical parameter data frames are classified for access charging guns to generate a reverse feed-in capability tag set. The reverse irrigation capability tag set is bound and integrated with the valid interface electrical parameter data frame to generate the access charging gun classification electrical parameter data frame.

[0012] This invention improves the real-time performance and accuracy of interface identification by decoding high-frequency data of the charging gun's electrical status, avoiding control errors caused by access identification delays. Setting a voltage fluctuation amplitude exceeding 10V within 0.5 seconds as an identification condition effectively distinguishes between transient disturbances and actual electrical parameter changes during charging / discharging state switching, improving the robustness of data filtering. Requiring the current direction to be consistent with at least two of the previous three frames helps enhance the stability of direction determination, preventing the misidentification of repeatedly oscillating direction fluctuations as valid paths. The requirement of more than three consecutive valid determinations within two seconds enhances the continuity requirement for reverse flow trend identification, preventing occasional fluctuations from being misjudged as continuous reverse flow, thereby improving tag accuracy. By introducing a joint interpretation mechanism of signal voltage, communication status, and power direction, a more comprehensive reverse flow behavior feature group can be constructed, significantly enhancing the accuracy and resolution of charging gun classification and providing a highly reliable capability tag foundation for subsequent power control. Overall, this strengthens the system's ability to capture the dynamic behavior of interface electrical parameters, significantly improving the sensitivity and control accuracy of photovoltaic energy storage systems in identifying and responding to reverse flow paths.

[0013] Optionally, the selection of the high-power charging and discharging path channel in step S3 is specifically as follows: The reverse feed capability label from the charging gun classification electrical parameter data frame is injected into the terminal node dimension of the physical access topology matrix. The rated carrying current, voltage level and port status of each edge in the physical access topology matrix after the label injection are updated in real time to generate a power structure topology map. High-power transmission path selection is performed based on the power structure topology diagram to obtain a set of candidate high-power transmission paths. For each path in the high-power transmission candidate path set, calculate the maximum power that can be carried on the entire path and generate a path power capability feature set; The optimal path selection is performed based on the path power capability feature set to obtain the photovoltaic charging and discharging path set.

[0014] This invention embeds reverse-feeding capability tags into the terminal nodes of the topology structure, effectively integrating the physical location of the equipment with its energy behavior attributes. This allows path assessment to move beyond relying on a single electrical parameter and instead possess dynamic capability identification characteristics. Simultaneously, the rated carrying current and voltage level of the connection edges are updated synchronously, dynamically reflecting the actual operating status and avoiding the misidentification problem of "nominal connectivity, but actual bottleneck" during path selection. When selecting candidate paths, the maximum carrying power index of the entire circuit is introduced, avoiding local overload caused by traditional shortest path selection. Its calculation process uniformly considers voltage level, resistance, and current limitations, enhancing the reliability and safety redundancy under system control. The optimal path selection mechanism achieves dynamic optimal allocation of charging and discharging paths, improving overall transmission efficiency, reducing power loss, and providing a clear and structurally sound path foundation for subsequent grid-connected control.

[0015] Optionally, the grid-connected peak shaving and regulation simulation in step S3 is specifically as follows: Based on the electrical status data frame, the grid-connected power assessment is performed on the power status of the terminal equipment connected to each path in the photovoltaic charging and discharging path set, thereby constructing a grid-connected power assessment matrix; Based on the multi-source physical parameters of the photovoltaic energy storage station and the capacity of the centralized station-end access transformer, grid connection control constraints and control cycles are set to obtain the grid connection control constraint table. Based on the grid-connected power assessment matrix and grid-connected regulation constraint table, the grid-connected power curve changes of all photovoltaic charging and discharging paths in each regulation cycle are simulated to predict the instantaneous total power change trend at the grid connection point. Paths that do not meet the grid-connected regulation constraint conditions are judged to fall back, thereby obtaining a peak shaving regulation simulation record set. The paths identified as being in a sustainable control state in the peak shaving and control simulation record set are extracted. The path numbers and device identifiers in the physical access topology matrix are bound to the paths in the sustainable control state, and the control priority, maximum adjustable power value and expected occupancy time are marked to obtain the feasible power channel table.

[0016] This invention significantly improves the real-time performance and accuracy of path power status perception by introducing electrical state data to periodically evaluate the power behavior of terminal equipment. Grid-connected control constraints based on transformer capacity (e.g., an upper limit of 136kW) effectively avoid overload impacts and ensure the power safety of grid-connected nodes. The setting of the control period provides the system with a predictive window, enabling dynamic capture of power fluctuation trends, early peak shaving, and reduction of the impact of sudden power surges on the system. Path rollback determination helps to divert peak shaving tasks at a fine granular level, improving the granularity and flexibility of path control. Path selection by simulating power curve changes ensures that paths entering subsequent control processes have stability and responsiveness, effectively improving the overall allocation efficiency of control resources. The final feasible channel table not only includes binding information between equipment identifiers and path numbers but also introduces fields such as maximum adjustable power value and occupation duration, providing a highly operable control basis for subsequent instruction generation and scheduling priority allocation.

[0017] Optionally, step S4 specifically includes: Step S41: Extract the characteristics of the electrical execution unit along the path based on the feasible power channel table, and generate the equipment control mapping table; Step S42: Perform structural analysis on the physical control parameters of the equipment for each path in the equipment control mapping table to obtain a set of control parameter templates; Step S43: Combine the equipment control mapping table and the control parameter template set, encapsulate them according to the path number, and construct a structured control instruction set; Step S44: Through the local control communication link of the photovoltaic energy storage station management platform, each instruction in the structured control instruction set is sent to the corresponding execution device channel by channel to execute the control instructions.

[0018] This invention extracts the features of electrical execution units involved in the path, enabling a refined extension of the control process from the path layer to the device layer. This achieves a closed-loop mapping from logical channels to physical devices, enhancing the targeting and response efficiency of command issuance. Device control parameter structure parsing standardizes key physical attributes such as control voltage range, trigger delay, and action time constant, generating parameter template sets and improving the adaptability of the control structure. Structured encapsulation using path numbers as indexes ensures that control commands do not cross-interfere between different paths, constructing a unified control semantic system. The use of local control links further reduces communication latency, adapting to control requirements with short-cycle responses within 100ms.

[0019] Optionally, the detection of the asymmetric load offset channel in step S5 specifically involves: Based on the feedback data from the execution device, the real-time load characteristics of the device channel are extracted, and a channel feedback status matrix is ​​generated. Calculate the power flow amplitude offset value of each path in the current scheduling cycle based on the channel feedback state matrix, and record the offset duration, maximum offset amount and minimum current threshold trigger record to obtain the power offset rate feature set. Based on the physical access topology matrix, the physical channel depth, branch convergence density and hop count features of the offset path corresponding to the power offset rate feature set are extracted to obtain the offset path feature set. By combining the control priority field in the structured control instruction set, a candidate set of asymmetric load paths in the offset path feature set is screened; The load imbalance coefficient is evaluated for the candidate set of asymmetric load paths. When the load imbalance coefficient of any path is ≥0.75 and it is activated in two consecutive control cycles, the physical channel number and the device number of the affected node are extracted and encapsulated to generate an asymmetric load offset channel set.

[0020] This invention introduces a load feature recognition mechanism based on feedback, effectively achieving dynamic correction of path operating status and improving the timeliness of the control closed loop response. By calculating the power flow offset amplitude and combining the duration and current lower limit trigger records, abnormal paths with poor stability or frequent power fluctuations can be identified, providing data support for control risk warning. Setting a minimum current threshold to trigger low-load identification helps to identify potential energy consumption redundancy in a timely manner under light load conditions. Introducing topology parameters such as physical channel depth, branch convergence density, and hop count can accurately locate potential bottleneck sections with complex structures and long power flow paths, improving the accuracy of offset tracing. By controlling the priority field to filter paths, it ensures that critical paths with a greater impact on system power balance are identified first. Setting the imbalance coefficient threshold to 0.75 and combining it with the status within two consecutive cycles avoids misjudgments caused by short-term fluctuations, improving the stability and robustness of anomaly detection. The final generated asymmetric load offset channel set can be used for subsequent reset and control optimization, improving the accuracy and safety of the overall system power scheduling.

[0021] Optionally, step S6 specifically includes: Step S61: Extract the path number and device number from the asymmetric load offset channel set to generate a reset candidate control sequence; Step S62: Extract the relay device number of each path to be reset in the reset candidate control sequence, and retrieve the communication protocol, response time and isolation control parameters to construct a channel-by-channel disconnection command; Step S63: Encapsulate the channel-by-channel disconnection instructions to generate a disconnection control instruction set, and send each instruction in the disconnection control instruction set to the corresponding execution device for each channel to execute the disconnection task; Step S64: Obtain the real-time electrical parameters of the energy storage and charging station, and perform path channel differential pressure reconstruction based on the real-time electrical parameters of the energy storage and charging station to generate a power distribution compensation map; Step S65: Perform dynamic load redistribution based on the power distribution compensation map to generate a balanced control target matrix; Step S66: Encapsulate the equalization control target matrix into structured instructions to obtain a structured reset control instruction set, and send each instruction in the structured reset control instruction set to the corresponding execution device channel by channel to execute the control instructions.

[0022] This invention achieves dynamic self-stabilization and equalization control of a photovoltaic energy storage system by rapidly isolating and reconstructing abnormal channels. By extracting relay devices using path and device numbers and invoking communication protocols and response time parameters, it ensures that commands are physically reachable and time-sensitive. Setting isolation control parameters allows for precise triggering of disconnection operations, preventing accidental disturbances to adjacent channels. After disconnection, electrical parameters are collected for differential pressure reconstruction, which helps identify the physical feasibility of load migration in backup paths and improves power compensation allocation efficiency. When setting the power allocation compensation map, a redundancy and load margin scoring mechanism is incorporated to make the allocation more robust and provide a safety margin. In equalization control, a target matrix is ​​constructed and structured commands are encapsulated, reducing the risk of control logic confusion under complex physical connections. The overall process achieves closed-loop control from offset identification, isolation, evaluation to rebalancing, effectively mitigating the impact of asymmetrical loads on the stability and energy efficiency of the energy storage system.

[0023] Optionally, this specification also provides a power management system suitable for photovoltaic energy storage and charging, for executing the power management method for photovoltaic energy storage and charging as described above, the power management system for photovoltaic energy storage and charging comprising: The topology analysis module is used to obtain the multi-source physical parameter set of the photovoltaic energy storage and charging station, and to perform multi-source power access physical topology analysis on the multi-source physical parameter set of the photovoltaic energy storage and charging station to obtain the physical access topology matrix; The electrical status sampling module is used to perform high-frequency sampling of physical connections based on the physical access topology matrix to obtain electrical status data frames; and to classify the electrical status data frames into power reverse charging guns to obtain access charging gun classification electrical parameter data frames. The grid-connected peak shaving and control module is used to select high-power charging and discharging path channels for the classified electrical parameter data frames of the access charging guns according to the physical access topology matrix, thereby obtaining a photovoltaic charging and discharging path set; and to perform grid-connected peak shaving and control simulation on the photovoltaic charging and discharging path set to obtain a feasible power channel table. The instruction structuring module is used to structure the feasible power channel table channel by channel to obtain a structured control instruction set, and then send the structured control instruction set to the execution device one by one through the photovoltaic energy storage and charging station management platform to execute the control instructions. The execution deviation identification module is used to acquire feedback data from the execution device; to identify execution deviations based on the feedback data from the execution device; and to detect asymmetric load offset channels based on the execution deviation identification results. The safety reset module is used to perform a safety reset of the energy storage and charging station for asymmetrical load offset channels, and to perform dynamic balance redistribution based on the safety reset results to obtain a structured reset control instruction set. The structured reset control instruction set is then sent to the execution equipment one by one through the photovoltaic energy storage and charging station management platform to execute the control instructions.

[0024] The present invention provides an energy management system applicable to photovoltaic energy storage and charging. This system can implement any of the energy management methods applicable to photovoltaic energy storage and charging of the present invention. It serves as a medium for coordinating the operation and signal transmission between various modules to complete the energy management method applicable to photovoltaic energy storage and charging. The internal modules of the system cooperate with each other, thereby significantly improving the operational stability and energy utilization rate of the photovoltaic energy storage and charging system. Attached Figure Description

[0025] Other features, objects, and advantages of the invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings: Figure 1 This is a flowchart illustrating the steps of the energy management method for photovoltaic energy storage and charging according to the present invention. Figure 2 This is a detailed flowchart of step S1 in the present invention; The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0026] The technical method of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0027] Furthermore, the accompanying drawings are merely illustrative of the invention and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor methods and / or microcontroller methods.

[0028] It should be understood that although the terms "first," "second," etc., may be used herein to describe various units, these units should not be limited by these terms. These terms are used merely to distinguish one unit from another. For example, without departing from the scope of the exemplary embodiments, a first unit may be referred to as a second unit, and similarly, a second unit may be referred to as a first unit. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0029] To achieve the above objectives, please refer to Figures 1 to 2This invention provides a power management method suitable for photovoltaic energy storage and charging, the method comprising the following steps: Step S1: Obtain the multi-source physical parameter set of the photovoltaic energy storage and charging station, and perform multi-source power access physical topology analysis on the multi-source physical parameter set of the photovoltaic energy storage and charging station to obtain the physical access topology matrix; In this embodiment, a multi-channel acquisition device deployed at the photovoltaic energy storage and charging station collects parameters such as voltage, current, power direction, frequency, and energy metering from key physical nodes including the output terminals of photovoltaic panels, the input and output terminals of energy storage battery packs, and electric vehicle charging interfaces, constructing a multi-source physical parameter set. The acquisition frequency of each channel is uniformly set to 100Hz and encapsulated using a structured data format. Each record includes fields such as device number, interface number, sampling time, voltage, current, and power direction, forming a structured physical parameter dataset. Subsequently, by parsing the physical connection configuration file of this dataset, a mapping relationship between port numbers and device physical locations is established, and this mapping relationship is used to construct an interconnection channel relationship diagram. Based on this diagram, the physical connection channels and their topological constraints between each type of device are extracted, redundant branches are eliminated, and finally, a physical access topology matrix describing the multi-source access situation of the entire energy storage and charging system is generated. This matrix is ​​in the form of an adjacency matrix, with rows and columns corresponding to different electrical nodes. The matrix element values ​​represent connection weights and directions, with units of Ω. -1 .

[0030] Step S2: Perform high-frequency sampling of physical connections based on the physical access topology matrix to obtain electrical status data frames; classify the electrical status data frames for reverse power charging guns to obtain access charging gun classification electrical parameter data frames; In this embodiment, based on the physical access topology matrix, photovoltaic output paths, energy storage input / output paths, and corresponding paths for various charging guns are selected from the established physical channels. Current transformers and voltage sampling ports are configured for high-frequency sampling, with the sampling frequency for each channel set to 1kHz and the sampling duration to 60 seconds, resulting in an electrical status data frame. The data frame fields include: path number, voltage, current, active power, reactive power, power direction, and signal quality indicators. Subsequently, reverse feedback feature analysis is performed on the data segments corresponding to each charging gun path in this data frame. Thresholds are set as voltage fluctuation ΔV > 10V and current direction reversal occurring ≥ 3 times / 2 seconds. Records meeting these conditions are marked as "suspected reverse feedback." Valid channel data is further filtered by combining communication status fields (such as handshake completion flag and ACK feedback valid identifier) ​​to generate valid interface electrical parameter data frames. The signal voltage, communication status, and power direction of each channel in this frame are jointly judged to extract reverse feedback judgment feature groups. Finally, based on the saliency of the feature groups, channels with uplink energy feedback capability are distinguished and labeled, such as "forward power supply type", "reverse feedback type" or "one-way non-feedback type", forming a classification electrical parameter data frame for the access charging gun.

[0031] Step S3: Select high-power charging and discharging path channels for the classified electrical parameter data frames of the access charging guns according to the physical access topology matrix to obtain the photovoltaic charging and discharging path set; perform grid-connected peak shaving regulation simulation on the photovoltaic charging and discharging path set to obtain the feasible power channel table; In this embodiment, the reverse current carrying capacity label in the access charging gun classification electrical parameter data frame is mapped to each terminal node in the physical access topology matrix, and the path information, such as the maximum current carrying capacity (I) on each path segment, is updated according to the label. max ), Maximum voltage level (V) max Based on this update, a power structure topology graph is constructed. For each physical path, the maximum power carrying capacity of the entire path is calculated using the formula: P_max = min(I_max_i × V_max_i), ∀i∈path; where i is each connection segment in the path, thus obtaining the path power capacity feature set. A high-power threshold P is set. min For a power of 5kW, select those that meet P max ≥P min The paths are then used to form a photovoltaic charging and discharging path set. Then, grid-connected power regulation simulation is performed on each path in the path set. The capacity C of the main transformer of the energy storage and charging station is read. max =400kW, the upper limit of grid-connected power is set at 0.9×C max For each path P, the power prediction sequence P(t) within the current control period T is accumulated and judged to be ∑P(t) ≤ 0.9 × C. max The paths that meet the criteria are recorded, along with their regulation duration, priority, maximum regulation capacity, and other parameters, forming a feasible power channel table.

[0032] Step S4: Structure the channel-by-channel control instructions in the feasible power channel table to obtain a structured control instruction set, and send the structured control instruction set to the execution equipment one by one through the photovoltaic energy storage and charging station management platform to execute the control instructions; In this embodiment, the path numbers in the feasible power channel table are mapped one-to-one with the corresponding device numbers. Combined with the control instruction formats (such as Modbus instruction formats) stored in the device control template library, a device control mapping table is generated. For each control device (such as a relay, power controller, or bidirectional converter) in each path, its control template fields, such as start address, opcode, and response delay, are read and uniformly encapsulated into a structured control instruction set. Each instruction includes the following fields: path number, control device number, control direction, execution duration, control amplitude, and instruction CRC checksum. This instruction set is sent channel by channel through the PLC communication network (RS-485 bus protocol) deployed inside the photovoltaic energy storage station, controlling up to 12 channels per second, and a retransmission mechanism of 3 times is set to ensure the reliability of instruction execution.

[0033] Step S5: Obtain feedback data from the execution device; identify execution deviation based on the feedback data from the execution device, and detect the asymmetric load offset channel based on the execution deviation identification result; In this embodiment, a multi-channel acquisition device deployed at the photovoltaic energy storage station acquires the status feedback data of all executing devices after the end of the current control cycle, extracts parameters such as real-time current, voltage, and switch status, and generates a channel feedback status matrix. The power flow offset rate ΔP(t) and the theoretical value P are then calculated for each path in this matrix. expected The deviation ratio is calculated as: Deviation Rate = |P_measured - P_expected| / P_expected. When the deviation rate of any path is ≥20% in two consecutive control cycles and the maximum current is less than 2A, the path is recorded as a high deviation record. Combining the priority field of each path in the structured control instruction set, information such as physical channel depth (path hop count) and branch convergence density (number of branches / total number of trunk segments) is extracted to establish a deviation path feature set. This feature set is then filtered, with paths having a branch convergence density >0.35, hop count ≥4, and priority ≤3 selected as candidate asymmetric load paths. Further, load imbalance coefficients are applied to these paths; when the coefficient is ≥0.75 and deviation occurs in two consecutive cycles, the path is marked as an asymmetric load deviation channel.

[0034] Step S6: Perform a safety reset of the asymmetric load offset channel for the energy storage and charging station, and perform dynamic equalization redistribution based on the safety reset result to obtain a structured reset control instruction set; send the structured reset control instruction set to the execution equipment one by one through the photovoltaic energy storage and charging station management platform to execute the control instructions.

[0035] In this embodiment, the path numbers and relay device numbers in the asymmetric load offset channel set are organized to generate a reset candidate control sequence. Relay device control parameters are extracted, and channel-by-channel disconnection control instructions are constructed. Each instruction includes: target address, disconnection execution time window (typically set to 100ms), and isolation control confirmation code. This disconnection instruction set is sent through the station's PLC bus system to execute the disconnection operation. Within one second after disconnection, electrical parameter data of the storage and charging station is collected, generating a raw data table of channel voltage drop. The voltage difference ΔV is calculated and normalized to the 0-1 range to obtain a voltage drop feature matrix. Based on this matrix, the power migration capability and redundancy carrying capacity of all backup paths are identified, and a compensation map is generated through channel scoring rules. The optimal channel combination is selected according to the compensation map, and the target power of each device is recalculated according to the average power allocation method to form a balanced control target matrix. Finally, this matrix is ​​encapsulated into a structured reset control instruction set and sent to the devices to complete the balanced control task.

[0036] Optionally, step S1 specifically includes: Step S11: Obtain the multi-source physical parameter set of the photovoltaic energy storage and charging station, and encapsulate it in a structured manner to obtain a structured physical parameter dataset; In this embodiment, multi-source data acquisition terminals deployed at key nodes of the photovoltaic energy storage and charging station, including voltage monitors, current sensors, active / reactive power acquisition devices, and environmental parameter recorders, are used to synchronously sample the photovoltaic panels, energy storage battery clusters, DC charging guns, and inverters. The sampling period is uniformly set to 50ms. The results of each round of acquisition are encapsulated into structured record items by timestamp and device number. Fields include device number (ID), interface number (PortID), sampling time (t), voltage (V), current (I), active power (P), temperature (T), etc., generating an initial data table. Subsequently, the format is standardized according to field standardization rules, missing values ​​are filled using three-point linear interpolation, and outliers are removed based on the upper and lower limits of the device's nominal parameters. Finally, all standardized record items are summarized to generate a structured physical parameter dataset. The data structure is in the form of a two-dimensional matrix, where each row represents a data snapshot at a sampling time point, and each column is a physical parameter item. The overall structure satisfies: Structured dataset = [t, ID, PortID, V, I, P, T, SOC, Freq].

[0037] Step S12: Perform port mapping analysis and bind physical connection identifiers on the structured physical parameter dataset to obtain the physical port information table; In this embodiment, the device number and its corresponding interface number fields are extracted from the structured physical parameter dataset. The mapping relationship between the logical names and physical numbers of the device ports is extracted using a predefined wiring topology configuration file (stored in JSON format). The configuration file structure is in key-value pair format, such as "inv_1":{"PortA":"PV_in_01","PortB":"Battery_out_03"}. The port numbers in the structured dataset are mapped one-to-one with this configuration, and their physical connection relationships are parsed. During the operation, interface verification rules are set, such as: input and output corresponding identifiers must be present, and "0" is used when the current directions are different; otherwise, it is considered a configuration conflict. After mapping is completed, each port relationship is encapsulated as a physical port record. The record fields include: device number, port number, physical location, connection direction (in / out), associated device and interface electrical parameter threshold (such as rated current I). rated =32A). All record items are aggregated by device number to generate a physical port information table. The structure is a key-value dictionary, where the key is the device ID and the value is all port connection attributes of the device.

[0038] Step S13: Construct a multi-source device interconnection channel relationship diagram based on the physical port information table; In this embodiment, the physical port information table is used as input to extract the correspondence between all devices connected via ports, constructing an interconnection channel edge set. Each edge represents a group of devices with electrical connection capabilities, structured as a triple: (device A number, device B number, channel attribute), where the channel attribute includes parameters such as connection type (AC / DC), rated current, and cable distance. The threshold for determining channel validity is set as follows: connection length ≤ 80 meters, voltage level matching, and current capability ≥ 20A. If the conditions are met, the connection is marked as "communication and power supply available"; otherwise, it is not included in the channel graph. All valid edges are expressed as an adjacency matrix G in graph structure, where G[i][j] = 1 indicates that there is a valid channel between devices i and j. To preserve the attributes of each channel in the topology, the edge weight matrix is ​​defined as W[i][j] = rated current. The final interconnection channel relationship graph supports structural traversal and contains complete electrical parameter information, which can be used for subsequent path dependency modeling.

[0039] Step S14: Analyze the multi-source access path combination and logical dependency relationship between each type of device based on the multi-source device interconnection channel relationship diagram to obtain the photovoltaic energy storage and charging station access path structure; In this embodiment, based on the existing multi-source device interconnection channel relationship diagram, various types of devices (such as photovoltaic strings, inverters, battery packs, and charging guns) are first categorized. The starting point of the path is set as the photovoltaic panel / energy storage device, and the ending point is the load or grid-connected node. A breadth-first path search method is used to identify all feasible paths between the start and end pairs, and the sequence of relay devices is recorded. To enhance physical constraint judgment, the maximum supported voltage V of all relay devices in the path is extracted from the physical port information table. max and minimum allowable current I min Calculate the maximum capacity P for each path. path As follows: P_path=min(I_min_i×V_max_i), Where P_path is the maximum throughput capacity, I_min_i is the minimum allowable current of the i-th relay device, and V_max_i is the maximum supported voltage of the i-th relay device. Set P... min =3kW was used as the path retention threshold to filter out low-capacity paths. Then, based on the source identifier (e.g., PV, Battery) of the starting device, reclassification was performed, and multi-source access path sets were generated by clustering by type. Finally, based on the interconnection logic and power supply direction of each path relay, a logical dependency matrix between paths was established, marking relational fields such as sequence, interlocking, and sharing, forming the complete access path structure of this energy storage and charging station for the next step of physical structure integration modeling.

[0040] Step S15: Based on the access path structure of the photovoltaic energy storage and charging station, perform topology integration to obtain the physical access topology matrix.

[0041] In this embodiment, all multi-source paths with labeled logical dependencies are extracted from the access path structure of the photovoltaic energy storage station, and abstracted into a graph structure consisting of a set of edges and nodes. Nodes correspond to device numbers, and edges represent effective path channels. After constructing a unified path graph, topology regularization is performed on the graph, mainly including redundant branch merging, repeated relay node folding, and symmetric path reduction. Paths with a redundant path distance threshold of less than 10m and consistent power direction are considered mergingable, ultimately constructing a regularized physical access topology structure. This structure graph is further converted into a standard adjacency matrix form A, where A[i][j]=1 indicates that device i and device j have a logical path connection. A power direction matrix D is also established, where D[i][j]=1 indicates an effective power path flowing from i to j, and 0 otherwise. The physical access topology matrix generated by combining A and D has network analysis capabilities and serves as the basic structural input for subsequent power regulation and path selection, providing a foundation for real-time regulation and analysis.

[0042] Optionally, step S14 specifically includes: Step S141: Perform path identification and classification on the starting and ending devices of each path in the interconnection channel relationship diagram to obtain a path identification index list; In this embodiment, based on the constructed interconnection channel relationship diagram, the functional attributes of each device are first labeled. For example, photovoltaic strings are labeled as "PV source end", energy storage batteries are labeled as "energy storage end", charging piles are labeled as "load terminal", and grid-connected inverters are labeled as "grid-connected terminal". Then, a path search is performed with the source end device as the starting point and the terminal device as the ending point, limiting the maximum number of hops in the path to 6 to avoid excessive line loss due to crossing too many relay devices. During the path search process, for each path from the starting point to the ending point identified, the starting device number, the ending device number, and the list of relay devices in the path are immediately recorded as index items to form a path identification index list. The data structure of this list is a two-dimensional array, with each row corresponding to a path, and the field format is as follows: Path identification item = [Start_ID, End_ID, [Mid_ID_1, Mid_ID_2, ..., Mid_ID_n]].

[0043] Step S142: Traverse the path identification index list to identify the relay devices for each path, and extract the electrical parameters of the relay devices to obtain the path relay dependency table; In this embodiment, each path recorded in the path identification index list is parsed one by one, extracting all relay device numbers and retrieving the corresponding electrical parameters from the previously constructed physical port information table, including maximum supported voltage (V_max), minimum allowable current (I_min), and rated operating frequency (Hz). The electrical parameters of each relay device are uniformly standardized as a dictionary structure, for example: {"ID":"inv_05","V_max":750,"I_min":10}. The relay device parameters on each path are aggregated into one item to construct a path relay dependency table. Each record item includes the path number, a list of path relay device IDs, and their corresponding electrical parameter fields, thus providing a parameter basis for subsequent traffic capacity calculations. To ensure data integrity, a maximum device parameter missing rate of 5% is set. Paths exceeding this rate will be directly marked as invalid paths and removed from the dependency table.

[0044] Step S143: Based on the minimum allowable current and maximum supported voltage limits of each relay device in the path relay dependency table, calculate the maximum throughput capacity of the path, and filter and remove paths with low throughput capacity to obtain a set of feasible paths; In this embodiment, based on the existing path relay dependency table, the minimum allowable current and maximum supported voltage of all relay devices in each path are read sequentially, and the maximum carrying power (passage capacity) of the path is calculated using the following expression: P_path=min(I_min_i×V_max_i). Where V_max_i and I_min_i are the maximum supported voltage and minimum allowable current of the i-th relay device in the path, respectively. For ease of filtering, a threshold P_min = 3.5kW is uniformly set; that is, if P_max_path < 3.5kW, a path is marked as a low-capacity path and removed. Paths that meet the criteria are retained by their index numbers, forming a set of feasible paths. In actual deployment, to improve scheduling stability, a backup capacity score for the path (such as device temperature rise margin and hop count) can be further introduced as a secondary filtering reference.

[0045] Step S144: Reclassify feasible paths according to the origin device source type of the feasible access path set to construct a multi-source access path set; In this embodiment, for each path retained in the feasible access path set, the functional type label of the starting device is extracted. For example, if the starting point is "PV", it is labeled as a "photovoltaic path"; if the starting point is "Battery", it is labeled as an "energy storage path". The paths are regrouped and classified according to device function to form a multi-source access path set. Each path set records its path number, starting and ending device numbers, relay device sequence, and the energy type to which the path belongs. During the classification process, to prevent cross-classification errors, starting devices are only allowed to have a single source type label. If a device is classified repeatedly, the primary type must be manually reviewed and determined. The final output structure is a path list clustered by energy source, providing the path set foundation for the logical dependency modeling stage.

[0046] Step S145: Model the logical dependency relationship based on the multi-source access path set to generate the access path structure of the photovoltaic energy storage and charging station.

[0047] In this embodiment, after obtaining the multi-source access path set, the dependency relationship of each path is modeled based on the device connection order, device type identifier, and control priority fields recorded in the photovoltaic energy storage station management platform, constructing a control logic dependency graph between paths. Logical dependencies can be divided into three categories: sequential dependency (path A must be activated before path B), interlock dependency (path A and path B cannot be activated simultaneously), and shared dependency (path A and path B share the same relay device). These logical dependencies are modeled as an adjacency list structure and recorded using key-value pairs, for example: {"path_01":["before:path_03","lock:path_05"]}. Simultaneously, a path dependency matrix D is established, where D[i][j]=1 indicates that path i has a dependency relationship with path j. To control the dependency strength, the maximum number of hops is set to 4; if this is exceeded, no logical chain is established. Finally, based on the path dependency matrix and adjacency graph, the path set is structured into an access path structure, serving as input for physical access topology integration. The construction of path logic dependencies is not arbitrary, but rather integrates the engineering principles of physical connection structure, equipment control characteristics, and scheduling priority settings in photovoltaic energy storage and charging systems. It is achieved through structured data extraction and dependency graph construction, combined with existing domestic and international standards and control strategies, and has a clear theoretical basis and engineering operability.

[0048] Optionally, the specific classification of the reverse-flow charging guns in step S2 is as follows: Real-time decoding of the charging gun interface layer is performed based on the electrical status data frame to obtain the interface electrical parameter identification data frame. In this embodiment, based on the physical access topology matrix, photovoltaic output paths, energy storage input / output paths, and corresponding paths for various charging guns are selected from the established physical channels. Current transformers and voltage sampling ports are configured for high-frequency sampling, with each channel's sampling frequency set to 1kHz and a sampling duration of 60 seconds, resulting in electrical status data frames. Interface voltage and current channel data are extracted by comparing interface numbers with data frame indexes, and hierarchical mapping is performed according to the control numbers based on the station port hierarchy table recorded in the photovoltaic energy storage and charging station management platform. During interface hierarchy decoding, if a port corresponding to a certain number is identified as "L2-FEED-IN", it is classified as a secondary AC reverse-feed interface. High-frequency noise data is removed through analysis of the access signal, and real-time hierarchy decoding is completed by combining the port characteristic parameter library preset in the photovoltaic energy storage and charging station equipment configuration table of the photovoltaic energy storage and charging station management platform, thereby outputting structured interface electrical parameter identification data frames.

[0049] When the voltage fluctuation amplitude is greater than 10V within 0.5 seconds and the current current direction is consistent with the direction of at least two of the first three frames in the interface electrical parameter identification data frame, the current interface electrical parameter of the charging gun is determined to be valid, so as to filter the valid interface electrical parameter data points in the interface electrical parameter identification data frame. When a valid interface electrical parameter data point is determined to be valid more than 3 times within 2 seconds, the data frame corresponding to the valid interface electrical parameter data point is marked as a valid interface electrical parameter data frame. In this embodiment, the judgment threshold is set as follows: when the voltage fluctuation amplitude (difference between the maximum and minimum voltage) of a certain interface exceeds 10V within any 0.5-second window, and the current direction recorded in the current frame is consistent with at least two of the previous three frames, the interface electrical parameters of the current frame can be considered valid. The current direction is determined by sign judgment: positive value indicates charging, negative value indicates reverse charging, and consistent direction means the same sign. Continuous valid frames are recorded, and the sliding window is counted every 2 seconds during operation. When the number of valid frames for a certain interface in the window exceeds 3, the data at the time point of that frame is marked as a "valid interface electrical parameter data frame". This filtering logic relies on the real-time data caching system of the photovoltaic energy storage and charging station management platform. The data is stored in memory in a FIFO queue structure and synchronously indexed.

[0050] The signal voltage, communication status and power direction of each charging gun in the current data frame of the effective interface electrical parameter data frame are jointly interpreted to extract the reverse feed-in judgment feature group; In this embodiment, the field group {Voltage, Current, StatusByte} is extracted from the valid interface electrical parameter data frame. The reference normal value for the signal voltage is set to 220V±15%. If it exceeds this range, it will be considered abnormal data and excluded. The power direction is determined by calculating P=V×I. If P<0, it indicates reverse current. The communication status is controlled by the lower four bits of the status byte: "0001" indicates charging, "0010" indicates idle, and "0011" indicates connection error. The three together constitute a feature triplet, such as [235V, -4.3A, "Charging"], which is determined to be a possible reverse current event. Among multiple triplets, if two sets of triplets that meet the reverse current direction appear consecutively on the same port, the system records the feature combination and adds it to the "Reverse Current Judgment Feature Set".

[0051] Based on the reverse feed-in judgment feature group, the valid interface electrical parameter data frames are classified for access charging guns to generate a reverse feed-in capability tag set. In this embodiment, based on the feature groups, the frequency of the triplet for the backflow direction of each interface is counted within one control cycle (default setting is 5 seconds), and evaluated in conjunction with the maximum backflow capability of the interface current recorded in the station configuration file (e.g., maximum allowable backflow current for DC interface = 10A). The classification criteria are as follows: frequency ≥ 3 times and backflow current ≥ 60% maximum capability is recorded as "strong backflow"; frequency 1~2 times or current between 30% and 60% is "weak backflow"; and failure to meet the criteria is "no backflow". A JSON-structured tag set is generated, such as: {"PortID":"CHG-06","BackflowCapability":"Strong"}, and written into the backflow capability tag set for binding electrical parameter frames.

[0052] The reverse irrigation capability tag set is bound and integrated with the valid interface electrical parameter data frame to generate the access charging gun classification electrical parameter data frame.

[0053] In this embodiment, for each record in the "Valid Interface Electrical Parameter Data Frame Table," the corresponding reverse flow capability tag is searched by the PortID field to complete the field appending and integration. The final generated access charging gun classification electrical parameter data frame contains the following field structure: [Time, PortID, Voltage, Current, DirectionFlag, Validity, BackflowCapability]. For example, the sample data is: [14:01:12.400, CHG-06, 234.1, -3.6,"Reverse", TRUE,"Strong"]. The integrated data frame will be synchronously passed to the subsequent power path filtering process as key input information for evaluating the charging and discharging channel capability.

[0054] Optionally, the selection of the high-power charging and discharging path channel in step S3 is specifically as follows: The reverse feed capability label from the charging gun classification electrical parameter data frame is injected into the terminal node dimension of the physical access topology matrix. The rated carrying current, voltage level and port status of each edge in the physical access topology matrix after the label injection are updated in real time to generate a power structure topology map. In this embodiment, the generated access charging gun classification parameter data frames are read, where each frame contains port number, voltage, current, power direction, and reverse feed capability label. The reverse feed capability label field is injected as an additional node attribute into the physical access topology matrix. The topology matrix is ​​a three-dimensional structure defined as T[i][j]={I_max,V_level,Status}, where i and j represent node numbers, I_max is the maximum carrying current, V_level is the voltage level, and Status indicates the on / off state. After the label is injected, the "reverse feed direction capability" field of the terminal node attribute is assigned a value according to the label type (such as "Strong" or "Weak"), and the edge status is updated synchronously. For example, for an edge from node N5 to N12, if the terminal node N12 is a charging gun port with "Strong" reverse feed capability, the I_max value of its connected edge will be increased to the maximum reverse feed current allowed by the device (such as set to 18A), and the V_level will be updated to 400V. After all injections are completed, a structured power structure topology map is generated as the basis for the next step of path evaluation.

[0055] High-power transmission path selection is performed based on the power structure topology diagram to obtain a set of candidate high-power transmission paths. In this embodiment, a path traversal search is performed on the updated power structure topology to identify transmission links with continuous high carrying capacity. The starting point of each path is set as the energy storage inverter node (e.g., B01), and the ending point is the charging port node with a reverse-feed capability label of "Strong" or "Weak." The minimum effective current threshold for each path is set to 12A. During the traversal, only paths where the I_max of each edge from the starting point to the ending point is not lower than this threshold are retained. A queue-based approach is used for layer-by-layer expansion. For each expanded path, the voltage level consistency of the current segment is calculated in real time (requiring that the V_level deviation between adjacent edges of the relay equipment does not exceed ±20V). Paths that do not meet this requirement are excluded. Finally, the candidate paths are uniformly recorded as a path sequence list, with each path represented by a node number sequence, such as [B01→N4→N8→CHG-06], and the corresponding edge attribute information is retained. All paths that meet the conditions constitute a high-power transmission candidate path set and proceed to the capability characteristic analysis stage.

[0056] For each path in the high-power transmission candidate path set, calculate the maximum power that can be carried on the entire path and generate a path power capability feature set; In this embodiment, for each candidate path, the rated carrying current and voltage level of all edges on the path are jointly calculated. Assuming a path has n edges, numbered e_1, e_2, ..., e_n, with corresponding carrying currents I_1, I_2, ..., I_n and voltages V_1, V_2, ..., V_n, then the maximum carrying power P_path of the entire path is defined by the following formula: P_path = min(I_1, I_2, ..., I_n) × min(V1, V2, ..., V_n). n This formula ensures that the power transmission capacity of the weakest link is selected among all repeater segments, avoiding misjudgment due to overload. For example, if the current capacity along a path is [18A, 16A, 20A] and the voltage is [410V, 395V, 400V], then its maximum carrying power is min(18, 16, 20) * min(410, 395, 400) = 16 × 395 = 6320W. A record is generated for each path, including the path number, node sequence, maximum carrying power value, total path length, and hop count. These are summarized into a path power capacity feature set and stored in a JSON structure for subsequent filtering.

[0057] The optimal path selection is performed based on the path power capability feature set to obtain the photovoltaic charging and discharging path set.

[0058] In this embodiment, after reading the path power capacity feature set, the path optimization and screening logic is executed, and the screening rules are set as follows: the maximum load capacity must be ≥5000W, the number of path hops ≤4, and the starting point and ending point of the path must be in different power supply branches (distinguished by transformer group). During the optimization and screening process, each record is judged by field, and the power density index (defined as: maximum load capacity divided by path length, unit W / hop) is calculated for compliant paths, and recorded as Power_Density=P_path / Hops. Finally, the top N paths in terms of power density (N can be set to 10) are selected, and these paths are encapsulated into a final photovoltaic charging and discharging path set in the following format: [{"PathID":"P001","Nodes":["B01","N3","N7","CHG-02"],"P_max":6480,"Hops":3},{"PathID":"P002","Nodes":["B01","N4","CHG-05"],"P_max":5920,"Hops":2}]; This path set will serve as the core input for the next step of grid-connected control simulation and structured instruction generation.

[0059] Optionally, the grid-connected peak shaving and regulation simulation in step S3 is specifically as follows: Based on the electrical status data frame, the grid-connected power assessment is performed on the power status of the terminal equipment connected to each path in the photovoltaic charging and discharging path set, thereby constructing a grid-connected power assessment matrix; In this embodiment, for each path in the photovoltaic charging and discharging path set, the active power value of the corresponding device is matched from the electrical status data frame according to the path terminal device number, and then archived according to the path number. The power value is calculated by the following formula: ;in This is the voltage value (in V). This is the current value (in amperes, A). The default power factor is set to 0.95. All paths are traversed and the terminal power is summarized to generate a grid-connected power assessment matrix. In this matrix, negative values ​​represent reverse power transmission, and positive values ​​represent absorbed power. This structure is used for grid-connected power prediction and analysis.

[0060] Based on the multi-source physical parameters of the photovoltaic energy storage station and the capacity of the centralized station-end access transformer, grid connection control constraints and control cycles are set to obtain the grid connection control constraint table. In this embodiment, the relevant configuration of the station-end transformer is extracted from the multi-source physical parameters of the photovoltaic energy storage station. The total load limit, single-path maximum adjustable power limit, and control time period parameters for each transformer are set in the control model. Taking transformer T1 as an example, its rated capacity is 160kVA. Considering a maximum utilization factor of 0.85, the upper limit is set to 136kW. Furthermore, the maximum adjustable power per path must not exceed 20kW. The control cycle is set to update every 30 seconds. The system constructs a grid-connected control constraint table with the following structure: Based on the grid-connected power assessment matrix and grid-connected regulation constraint table, the grid-connected power curve changes of all photovoltaic charging and discharging paths in each regulation cycle are simulated to predict the instantaneous total power change trend at the grid connection point. Paths that do not meet the grid-connected regulation constraint conditions are judged to fall back, thereby obtaining a peak shaving regulation simulation record set. In this embodiment, the control cycle is set to 30 seconds in the photovoltaic energy storage station management platform, and the simulation time window can be set to a continuous 5-minute period, divided into 10 cycles. Within each cycle, based on the grid-connected power assessment matrix constructed in the previous cycle, the system predicts the power change trend of all photovoltaic charging and discharging paths. The prediction method adopts linear trend extrapolation based on two consecutive cycles, that is, using the current power value of the path... Power value of the previous cycle Based on this, calculate the trend slope. : ;in This constitutes one control cycle (30 seconds), and the power prediction value for the next cycle is derived from this. : Taking path P021 as an example, if the power in the (n-1)th cycle is 20.5kW and the power in the nth cycle is 21.2kW, then the predicted power for the next cycle is 21.9kW. If the power in two consecutive predicted cycles exceeds the single-path upper limit (20kW) set in the grid-connected control constraint table, the path is marked as a "backoff candidate path". In addition, the platform calculates the total predicted grid-connected power after superimposing all paths in each cycle. If this value exceeds the upper limit of 136kW set by transformer T1, for example, if the predicted result for a certain cycle is 141.2kW, then the excess part needs to be reduced. The reduction strategy is based on the absolute value of the power change slope K of each path, prioritizing the path with the larger change rate for backoff. In this example, if the power growth rate of P021 is the largest, it is instructed to enter the power withdrawal state until the total power recovers to below the threshold. After each round of simulation, the number of each path, the current power value, the predicted value, the over-limit mark, the number of backoffs, and the status label ("sustainable" or "backoff") are recorded and packaged into a structured simulation record dataset. If a path remains in a stable power state (with a change of ≤1.0kW) for two consecutive cycles without triggering any limiting conditions, the system considers it a "sustainable path" and reserves it for subsequent instruction generation. The paths identified as being in a sustainable control state in the peak shaving and control simulation record set are extracted. The path numbers and device identifiers in the physical access topology matrix are bound to the paths in the sustainable control state, and the control priority, maximum adjustable power value and expected occupancy time are marked to obtain the feasible power channel table.

[0061] In this embodiment, all paths with a "sustainable" status are selected from the peak shaving and control simulation records, and their starting and ending node numbers in the physical access topology matrix are traced back. Taking path P017 as an example, its starting point is the energy storage converter INV-02, and its ending point is the charging port CHG-06. Its maximum adjustable power is obtained from its historical operation records as 16.8 kW, and its control priority is set to 1 (according to the rule that ≥15 kW is priority 1), with an expected control period of 15 minutes. Information on all such paths, including path number, starting and ending devices, maximum adjustable power, priority, and expected usage time, will be uniformly bound and encapsulated to form a feasible power channel set. This channel set will be called in the subsequent structured instruction generation step to construct device execution instructions and realize a closed loop for path power control.

[0062] Optionally, step S4 specifically includes: Step S41: Extract the characteristics of the electrical execution unit along the path based on the feasible power channel table, and generate the equipment control mapping table; In this embodiment, the photovoltaic energy storage station management platform calls the generated feasible power channel table to extract the number of each path, the identification of the starting device and the terminal device, and further retrieves the electrical execution unit information involved in the path. Taking path number P018 as an example, its starting point is inverter INV-03 and its ending point is vehicle charging interface CHG-09. By parsing the execution unit structure of each device through the device interface description document, it is confirmed that INV-03 has three sets of controllable power output terminals, numbered OUT1~OUT3, each port supports a maximum output power of 25 kW and a voltage level of DC 750V; CHG-09 has CAN bus access and BMS feedback reception capabilities. Key fields such as path number, device type, electrical port number, control interface address, and supported power limit are encapsulated into control entries, and assembled path by path to form a device control mapping table for subsequent parameter structure injection.

[0063] Step S42: Perform structural analysis on the physical control parameters of the equipment for each path in the equipment control mapping table to obtain a set of control parameter templates; In this embodiment, based on the generated device control mapping table, the executable control parameters are categorized according to device type. Taking the energy storage inverter INV-03 as an example, supported control parameters include: power setpoint (unit: kW), start / stop status (0 / 1), output voltage setpoint (unit: V), and control response time (unit: ms). The charging interface CHG-09 supports control fields including: power access request, communication handshake status, and BMS data response channel number. The platform extracts the physical meaning, unit range, and data structure format of each parameter field according to the device interface protocol configuration document. For example, the power setpoint is a 16-bit unsigned integer with a maximum setting of 30kW, and the control response time field must be an integer multiple of 10ms. Finally, the control parameter structures required for each type of device are assembled into a template format and encapsulated in the manner of "device type - parameter field - data format" to construct a control parameter template set.

[0064] Step S43: Combine the equipment control mapping table and the control parameter template set, encapsulate them according to the path number, and construct a structured control instruction set; In this embodiment, the extracted device control mapping table and the constructed control parameter template set are fused together by field alignment, and structured instruction entries are encapsulated one by one according to the path number. Taking path P018 as an example, its goal is to control INV-03 to output 18.6kW of power to port CHG-09. The power setting field is extracted from the control parameter template set, the value is converted into an unsigned integer format, and recorded as "Set power: 18600 (unit W)"; the start / stop status field is set to "1", the output voltage field is set to "750", and the response time is set to "60ms". The final encapsulation format of the instruction structure is: {PathID: P018, StartDevice:INV-03, EndDevice: CHG-09, Commands: {Power: 18600, Enable: 1, Voltage: 750,ResponseTime: 60}}; after all path instructions are encapsulated, all path structured instructions are combined into a control instruction set. The data structure adopts JSON style and has a channel verification field to ensure the integrity and reliability of the data before execution.

[0065] Step S44: Through the local control communication link of the photovoltaic energy storage station management platform, each instruction in the structured control instruction set is sent to the corresponding execution device channel by channel to execute the control instructions.

[0066] In this embodiment, a control channel is established with each execution device through an industrial Ethernet communication link deployed inside the photovoltaic energy storage station. The platform scheduling service reads the control content corresponding to each path in the structured control instruction set and calls the communication controller interface bound to the corresponding device address to send the instructions channel by channel. Taking path P018 as an example, a control instruction data packet with output port OUT2 is sent to INV-03. This data packet contains the device address field, command field, and checksum. After sending, the system waits for an acknowledgment frame to return within 30ms. If an ACK frame response is received, the platform confirms successful transmission; if no response is received within the set maximum number of retries (default 3 times), the device is marked as "communication abnormal," and the path number and reason for failure are recorded. After all paths have completed instruction transmission, a control completion status log is generated and synchronized to the upper-level monitoring platform for subsequent feedback identification and processing.

[0067] Optionally, the detection of the asymmetric load offset channel in step S5 specifically involves: Based on the feedback data from the execution device, the real-time load characteristics of the device channel are extracted, and a channel feedback status matrix is ​​generated. In this embodiment, after the control cycle is completed, load feedback data of all controlled devices during that cycle is collected, including current, voltage, power factor, port on / off status, and response delay. Each feedback packet is uploaded via CAN communication and includes a timestamp. Taking path P042 as an example, the feedback values ​​of its terminal device CHG-11 are: current 18.2A, voltage 728V, and feedback delay 45ms. The platform categorizes all feedback fields by path number and device port, organizing them into a "channel feedback status matrix". Each row of the matrix represents a power path, and each column corresponds to a specific feedback field. The structured format of this matrix is ​​as follows: [PathID, Current(A), Voltage(V), PF, Status, Delay(ms), Timestamp]; this matrix provides a real-time measurement basis for subsequent offset analysis, and all fields are uniformly formatted as floating-point or integer types for easy timing calculation.

[0068] Calculate the power flow amplitude offset value of each path in the current scheduling cycle based on the channel feedback state matrix, and record the offset duration, maximum offset amount and minimum current threshold trigger record to obtain the power offset rate feature set. In this embodiment, the offset values ​​of all paths relative to the preset target power are calculated in each control cycle, and the power offset is defined as the predicted power. With feedback power Difference: Taking path P042 as an example, with a target power of 13.6kW, and feedback current and voltage of 18.2A and 728V respectively, the feedback power is: = 18.2 × 728 =13249.6W ≈ 13.25kW; = 13.25 - 13.6 = -0.35kW; When the absolute value of the offset exceeds ±0.3kW and lasts for more than two cycles (1 minute), the system records the offset duration, maximum offset value, and whether it is below the minimum current threshold (default 10A) for that path. The offset statistics of all paths are encapsulated into a "power offset rate feature set", with fields including PathID, ΔP, duration, maximum offset, and whether a low current alarm is triggered, for subsequent unbalanced path identification.

[0069] Based on the physical access topology matrix, the physical channel depth, branch convergence density and hop count features of the offset path corresponding to the power offset rate feature set are extracted to obtain the offset path feature set. In this embodiment, the platform extracts structural features from the paths marked in the offset feature set of the previous step, based on the physical access topology matrix. For each path, it tracks the number of port hops, the number of relay devices, and the number of access nodes in each branch in the topology map, thereby calculating indicators such as the physical channel depth and branch convergence density of the path. For example, path P042 has 4 levels of relay hops in its topology (DC box → combiner box → converter → distribution cabinet), and its total number of access nodes in its branches is 7. Then, its branch convergence density is calculated as: branch node number / channel length = 7 / 4 = 1.75. This information, together with the path number and offset features, is encapsulated to generate an "offset path feature set," which is used to help identify potential areas of concentrated load imbalance.

[0070] By combining the control priority field in the structured control instruction set, a candidate set of asymmetric load paths in the offset path feature set is screened; In this embodiment, the generated structured control instruction set records the control priority field for each path, with a value range of 1 to 5, where a smaller value indicates a higher priority. The platform extracts the corresponding priority for each path in the "Offset Path Feature Set" by number and establishes a matching table. The screening rules are as follows: branch convergence density > 1.5, channel depth ≥ 4, hop count ≥ 4, priority ≤ 3. Paths that meet all conditions are included in the "Asymmetric Load Path Candidate Set". Taking path P042 as an example, its branch density is 1.75, channel depth is 4, and priority is 2, which fully meets the screening rules, and the platform marks this path as a candidate path.

[0071] The load imbalance coefficient is evaluated for the candidate set of asymmetric load paths. When the load imbalance coefficient of any path is ≥0.75 and it is activated in two consecutive control cycles, the physical channel number and the device number of the affected node are extracted and encapsulated to generate an asymmetric load offset channel set.

[0072] In this embodiment, the imbalance coefficient is evaluated for candidate paths in each control cycle. Imbalance coefficient Defined as: That is, the ratio of the absolute value of the offset power to the target power. When If the power index is ≥0.75 and the path has been activated by regulation for two consecutive cycles (i.e., there is an effective power command issued and fed back), it is judged as an asymmetric load path. Taking path P042 as an example, its... =-0.35kW, =13.6kW, =0.35 / 13.6≈0.026, which does not constitute an anomaly; if the other path P056 is: =-10.5kW, =13.2kW, The value is approximately 0.795, and the system is activated for two consecutive cycles. Therefore, it is added to the "Asymmetric Load Offset Channel Set". This set includes path number, affected device ID, offset characteristics, and structural indicators for subsequent reset and control.

[0073] Of particular importance is the specific process of selecting the asymmetric load path candidate set from the offset path feature set: Extract the path number, physical channel depth and branch convergence density from the offset path feature set to generate a path structure attribute table; In this embodiment, the path ID, physical channel depth, and branch convergence density fields are extracted from the offset path feature set. These fields are encapsulated as key-value pairs and uniformly formatted as structured records with the path ID (PathID) as the primary key. For example, path P037 has a physical channel depth of 5 layers and a branch convergence density of 0.42; the path ID serves as the index field for this record. All path structure information is aggregated to construct a "path structure attribute table." This table has explicit data type restrictions: channel depth and hop count are integers, convergence density is a floating-point number, and precision is controlled to two decimal places to ensure calculation accuracy.

[0074] Extract the control priority field corresponding to each path number in the structured control instruction set and generate a path control level table; Using the path number as an index, the path structure attribute table and the path control level table are matched and aligned to generate a joint path control evaluation table. In this embodiment, the control priority field corresponding to all path numbers in the structured control instruction set is extracted, and a "Path Control Level Table" indexed by the path number is established. This table uses a fixed enumeration value method to limit the priority field (Level) to between 1 and 5, ensuring the control level division is controllable. For example, the control level of path P037 is 2, and that of path P041 is 4. The two tables are aligned in the path number dimension, and the structural attributes and control attributes are horizontally joined through the path ID index to generate a "Joint Path Control Evaluation Table".

[0075] The filtering rules are set as follows: branch convergence density > 0.35, physical channel depth, number of hops ≥ 4, and control priority ≤ 3. The paths in the joint path regulation and evaluation table are screened under multiple conditions using screening rules, and the paths that meet the screening rules are marked as candidate sets of asymmetric load paths.

[0076] In this embodiment, all records in the joint path control evaluation table are filtered. The filtering conditions are set as follows: branch convergence density > 0.35, channel depth ≥ 4, path hop count ≥ 4, and control priority ≤ 3. All four conditions must be met simultaneously. During the filtering process, all fields undergo data validity verification before judgment to ensure that there are no null values ​​or data in abnormal ranges. Taking path P037 as an example, its branch convergence density is 0.42, channel depth is 5, hop count is 4, and priority is 2, which fully meets the filtering rules. Therefore, it is marked as a member of the asymmetric load path candidate set, and the path number and its corresponding feature fields are packaged and output as candidates for subsequent reset control.

[0077] Optionally, step S6 specifically includes: Step S61: Extract the path number and device number from the asymmetric load offset channel set to generate a reset candidate control sequence; In this embodiment, paths and devices requiring reset operations are selected based on the path number and affected device number fields. Specifically, the path number is used as the primary key to quickly locate the corresponding physical channel and device unit in the path database, while the unique identifier and operating status of the device are obtained from the device asset management system. Taking path P056 as an example, the relay device IDs involved in this path include DC-Box-12, Inverter-5, and Switch-8. A reset candidate control sequence is automatically generated based on the collected path and device information. The sequence consists of pairs of path IDs and device IDs, with a sequence length equal to the number of asymmetric load offset paths. Each record includes a reset priority and timestamp field to ensure the timeliness and order of the control sequence. This sequence is used for the accurate transmission of subsequent disconnection commands, ensuring the targeted and complete nature of the reset operation.

[0078] Step S62: Extract the relay device number of each path to be reset in the reset candidate control sequence, and retrieve the communication protocol, response time and isolation control parameters to construct a channel-by-channel disconnection command; In this embodiment, relay device numbers are extracted one by one from the reset candidate control sequence. Combined with pre-stored communication protocols (such as Modbus TCP / IP, IEC 61850, etc.), device response time parameters (typically set within the range of 100ms to 500ms), and isolation control parameters (such as switch disconnection command codes, operation duration, etc.) in the device management library, a detailed disconnection instruction set is constructed. Taking device DC-Box-12 as an example, its communication protocol is IEC 61850, the disconnection instruction uses the command code "DISCONNECT", the response time is configured to 300ms, and the isolation parameter is set to disconnect all downstream port current outputs. During the construction process, the instruction format follows the execution device standard, including fields such as device ID, operation type, execution parameters, and checksum, ensuring that the instruction is complete and can be correctly parsed by the device. This process achieves fine-grained customization of the disconnection instruction by automatically matching the reset sequence with device characteristics.

[0079] Step S63: Encapsulate the channel-by-channel disconnection instructions to generate a disconnection control instruction set, and send each instruction in the disconnection control instruction set to the corresponding execution device for each channel to execute the disconnection task; In this embodiment, the disconnection command is formatted and encapsulated to generate a standardized disconnection control command set, including the command sequence number, target device ID, command type, and execution parameters. A specific communication protocol frame format, such as a TCP / IP-based data packet, is used during encapsulation to ensure the integrity and security of transmitted data. The commands are sent sequentially to the corresponding execution devices via the local control communication link of the photovoltaic energy storage station management platform. The sending frequency is set to an interval of no less than 200ms between each command to avoid network congestion or command conflicts. The platform monitors the execution status of each command through a feedback channel to ensure successful disconnection operation and records execution logs for subsequent analysis.

[0080] Step S64: Obtain the real-time electrical parameters of the energy storage and charging station, and perform path channel differential pressure reconstruction based on the real-time electrical parameters of the energy storage and charging station to generate a power distribution compensation map; In this embodiment, after the disconnection operation is completed, the photovoltaic energy storage and charging station management platform initiates real-time electrical parameter acquisition of the station, mainly including key electrical parameters such as voltage, current, power factor, and power flow direction. The acquisition utilizes a distributed sensor network with a sampling frequency of 100Hz to ensure high-timeliness data. By measuring the voltage drop of each path channel after disconnection, the voltage drop difference of each channel is calculated and normalized to form a channel voltage drop characteristic matrix. Based on this matrix, combined with the physical access topology, the voltage difference structure of the path channels is reconstructed, the possibility and limitations of power migration are quantitatively analyzed, and a power allocation compensation map is output. The map expresses the current carrying capacity and redundancy margin of each backup path in graph structure form, assisting in the formulation of scheduling strategies.

[0081] Step S65: Perform dynamic load redistribution based on the power distribution compensation map to generate a balanced control target matrix; In this embodiment, the path current carrying capacity provided by the power distribution compensation map is used to dynamically adjust the load distribution of each path, ensuring a balanced power flow. During the load redistribution process, the target power adjustment amount for each path is calculated based on real-time electrical parameters. The following equation is satisfied: ;in For the current path power, For path Maximum load capacity, The number of paths is specified. The adjusted power allocation results are integrated into a "balanced control target matrix". The matrix columns include path number, target power, adjustment range and corresponding execution priority. It is stored in floating-point format with a precision controlled within 0.1kW, and supports subsequent instruction encapsulation and issuance.

[0082] Step S66: Encapsulate the equalization control target matrix into structured instructions to obtain a structured reset control instruction set, and send each instruction in the structured reset control instruction set to the corresponding execution device channel by channel to execute the control instructions.

[0083] In this embodiment, a structured reset control instruction set is constructed based on the target power and path information in the balanced control target matrix, using an instruction template compatible with the execution device. The instruction set includes path number, device ID, control parameters (such as target power value, adjustment time window, etc.), and verification fields. The encapsulation process adopts a hierarchical structure, supporting batch processing and single instruction tracking to ensure instruction integrity and traceability. Through the photovoltaic energy storage station management platform, the instruction set is sequentially distributed to the corresponding execution devices according to the path order, with the transmission interval controlled between 150ms and 300ms to ensure communication quality. The device uploads its status feedback in real time after execution, and the platform adjusts the distribution strategy based on the feedback to achieve dynamic closed-loop control.

[0084] Most importantly, the path channel pressure differential reconstruction is specifically as follows: Step S641: Collect real-time electrical parameters of the storage and charging station after disconnection and generate a raw data table of channel voltage drop; In this embodiment, after the disconnection command is executed, the electrical parameters of each channel within the energy storage and charging station are collected in real time. The collected data includes key indicators such as voltage, current, power factor, and phase angle for each channel. A distributed sampling device is used, with the sensor sampling frequency set to 200Hz to ensure the timeliness and accuracy of the data. The collected data is synchronized with timestamps and then aggregated to form a channel electrical parameter dataset. The data format adopts a two-dimensional matrix structure, with rows corresponding to each channel number and columns representing sampling time, voltage value (in V), current value (in A), etc. This raw data table serves as the basis for subsequent voltage drop analysis, ensuring a complete reflection of the electrical state changes of each channel under the influence of the disconnection path.

[0085] Step S642: Calculate and normalize the differential pressure in the original channel voltage drop data table to generate the channel voltage drop feature matrix; In this embodiment, based on the collected raw channel electrical parameter data, the channel voltage drop difference is calculated by comparing the voltage values ​​of the corresponding channels at the same time before and after the disconnection. The voltage drop difference is calculated using the following formula: ;in, For the first The voltage drop difference of the channel, and These are the voltage values ​​collected before and after the disconnection operation. After calculation, the minimum-maximum normalization method is used to normalize the voltage drop difference of all channels. The normalized data constitutes the channel voltage drop feature matrix, with the rows corresponding to the channel number and the columns representing the normalized voltage drop value and the corresponding sampling time, ensuring that the data is within the [0,1] interval for easy subsequent comparison and fusion processing.

[0086] Step S643: Based on the channel voltage drop feature matrix, reconstruct the physical access topology matrix, determine the power migration feasibility index of each backup path after the interruption of the interrupted connection path, and generate a path channel voltage difference structure diagram. In this embodiment, the power migration capability of backup paths after a disconnection is evaluated by combining the channel voltage drop feature matrix with the physical access topology matrix of the photovoltaic energy storage station. First, the normalized channel voltage drop value corresponding to each backup path is mapped to the weight dimension of the corresponding edge in the topology matrix. The weight value reflects the degree of voltage drop impact on the path channel. Second, the connectivity and current-carrying capacity of each backup path are analyzed based on the electrical connection status between nodes in the topology matrix. Combining the voltage drop weights, a path migration feasibility index is calculated. The closer the index value is to 0, the smaller the voltage drop impact and the stronger the migration capability; the closer it is to 1, the larger the voltage drop impact and the more limited the migration. The final generated path channel voltage difference structure diagram displays the electrical performance of backup paths in the energy storage station in the form of graph nodes and weighted edges, providing a basis for subsequent power allocation.

[0087] Step S644: Score the redundancy and load margin of each backup path in the path channel pressure difference structure diagram, and generate a power distribution compensation map.

[0088] In this embodiment, based on the path channel pressure difference structure diagram, the channel redundancy and load margin of each backup path are further analyzed. Redundancy refers to the number of backup channels included in the path and their electrical independence. The redundancy index uses an integer score, ranging from 0 to 10, with higher values ​​indicating more redundant channels in the path. The load margin is calculated by taking the ratio of the maximum carrying current of each channel in the path to the actual load current, and then averaging the ratio across the entire path. The margin score uses a percentage system, and the specific formula is as follows: ;in, To accommodate margin, This represents the number of channels in the path. and The first The actual load current and maximum carrying current of the channel are compared. Combining redundancy and carrying margin scores, path current carrying indicators are generated, forming a power allocation compensation map. This map indicates the reliability and safety margin of backup paths during power migration, assisting in balanced dispatch decisions. The compensation map data structure adopts a weighted directed graph model, where nodes represent electrical equipment and edges have scoring attributes, supporting dynamic querying and updating.

[0089] Optionally, this specification also provides a power management system suitable for photovoltaic energy storage and charging, for executing the power management method for photovoltaic energy storage and charging as described above, the power management system for photovoltaic energy storage and charging comprising: The topology analysis module is used to obtain the multi-source physical parameter set of the photovoltaic energy storage and charging station, and to perform multi-source power access physical topology analysis on the multi-source physical parameter set of the photovoltaic energy storage and charging station to obtain the physical access topology matrix; The electrical status sampling module is used to perform high-frequency sampling of physical connections based on the physical access topology matrix to obtain electrical status data frames; and to classify the electrical status data frames into power reverse charging guns to obtain access charging gun classification electrical parameter data frames. The grid-connected peak shaving and control module is used to select high-power charging and discharging path channels for the classified electrical parameter data frames of the access charging guns according to the physical access topology matrix, thereby obtaining a photovoltaic charging and discharging path set; and to perform grid-connected peak shaving and control simulation on the photovoltaic charging and discharging path set to obtain a feasible power channel table. The instruction structuring module is used to structure the feasible power channel table channel by channel to obtain a structured control instruction set, and then send the structured control instruction set to the execution device one by one through the photovoltaic energy storage and charging station management platform to execute the control instructions. The execution deviation identification module is used to acquire feedback data from the execution device; to identify execution deviations based on the feedback data from the execution device; and to detect asymmetric load offset channels based on the execution deviation identification results. The safety reset module is used to perform a safety reset of the energy storage and charging station for asymmetrical load offset channels, and to perform dynamic balance redistribution based on the safety reset results to obtain a structured reset control instruction set. The structured reset control instruction set is then sent to the execution equipment one by one through the photovoltaic energy storage and charging station management platform to execute the control instructions.

[0090] Therefore, the embodiments should be considered as exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of the equivalents of the application are intended to be included within the invention.

[0091] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features of the invention herein.

Claims

1. A power management method suitable for photovoltaic energy storage and charging, characterized in that, Includes the following steps: Step S1: Obtain the multi-source physical parameter set of the photovoltaic energy storage and charging station, and perform multi-source power access physical topology analysis on the multi-source physical parameter set of the photovoltaic energy storage and charging station to obtain the physical access topology matrix; Step S2: Perform high-frequency sampling of physical connections based on the physical access topology matrix to obtain electrical status data frames; The electrical status data frames are classified into power reverse charging guns to obtain the access charging gun classification electrical parameter data frames. Step S3: Select high-power charging and discharging path channels for the classified electrical parameter data frames of the access charging guns according to the physical access topology matrix to obtain the photovoltaic charging and discharging path set; A grid-connected peak shaving regulation simulation was performed on the photovoltaic charging and discharging path set to obtain a feasible power channel table; Step S4: Structure the channel-by-channel control instructions in the feasible power channel table to obtain a structured control instruction set, and send the structured control instruction set to the execution equipment one by one through the photovoltaic energy storage and charging station management platform to execute the control instructions; Step S5: Obtain feedback data from the execution device; identify execution deviation based on the feedback data from the execution device, and detect the asymmetric load offset channel based on the execution deviation identification result; Step S6: Perform a safety reset of the storage and charging station on the asymmetric load offset channel, and perform dynamic balance redistribution based on the safety reset result to obtain a structured reset control instruction set; The structured reset control command set is sent out one by one to the execution equipment through the photovoltaic energy storage station management platform to execute the control command.

2. The power management method applicable to photovoltaic energy storage and charging according to claim 1, characterized in that, Step S1 is as follows: Step S11: Obtain the multi-source physical parameter set of the photovoltaic energy storage and charging station, and encapsulate it in a structured manner to obtain a structured physical parameter dataset; Step S12: Perform port mapping analysis and bind physical connection identifiers on the structured physical parameter dataset to obtain the physical port information table; Step S13: Construct a multi-source device interconnection channel relationship diagram based on the physical port information table; Step S14: Analyze the multi-source access path combination and logical dependency relationship between each type of device based on the multi-source device interconnection channel relationship diagram to obtain the photovoltaic energy storage and charging station access path structure; Step S15: Based on the access path structure of the photovoltaic energy storage and charging station, perform topology integration to obtain the physical access topology matrix.

3. The power management method applicable to photovoltaic energy storage and charging according to claim 2, characterized in that, Step S14 is as follows: Step S141: Perform path identification and classification on the starting and ending devices of each path in the interconnection channel relationship diagram to obtain a path identification index list; Step S142: Traverse the path identification index list to identify the relay devices for each path, and extract the electrical parameters of the relay devices to obtain the path relay dependency table; Step S143: Based on the minimum allowable current and maximum supported voltage limits of each relay device in the path relay dependency table, calculate the maximum throughput capacity of the path, and filter and remove paths with low throughput capacity to obtain a set of feasible paths; Step S144: Reclassify feasible paths according to the origin device source type of the feasible access path set to construct a multi-source access path set; Step S145: Model the logical dependency relationship based on the multi-source access path set to generate the access path structure of the photovoltaic energy storage and charging station.

4. The power management method applicable to photovoltaic energy storage and charging according to claim 1, characterized in that, The specific classification of the reverse-flow charging guns in step S2 is as follows: Real-time decoding of the charging gun interface layer is performed based on the electrical status data frame to obtain the interface electrical parameter identification data frame. When the voltage fluctuation amplitude is greater than 10V within 0.5 seconds and the current current direction is consistent with the direction of at least two of the first three frames in the interface electrical parameter identification data frame, the current interface electrical parameter of the charging gun is determined to be valid, so as to filter the valid interface electrical parameter data points in the interface electrical parameter identification data frame. When a valid interface electrical parameter data point is determined to be valid more than 3 times within 2 seconds, the data frame corresponding to the valid interface electrical parameter data point is marked as a valid interface electrical parameter data frame. The signal voltage, communication status and power direction of each charging gun in the current data frame of the effective interface electrical parameter data frame are jointly interpreted to extract the reverse feed-in judgment feature group; Based on the reverse feed-in judgment feature group, the valid interface electrical parameter data frames are classified for access charging guns to generate a reverse feed-in capability tag set. The reverse irrigation capability tag set is bound and integrated with the valid interface electrical parameter data frame to generate the access charging gun classification electrical parameter data frame.

5. The power management method applicable to photovoltaic energy storage and charging according to claim 1, characterized in that, The specific steps for selecting the high-power charging and discharging path channel in step S3 are as follows: The reverse feed capability label from the charging gun classification electrical parameter data frame is injected into the terminal node dimension of the physical access topology matrix. The rated carrying current, voltage level and port status of each edge in the physical access topology matrix after the label injection are updated in real time to generate a power structure topology map. High-power transmission path selection is performed based on the power structure topology diagram to obtain a set of candidate high-power transmission paths. For each path in the high-power transmission candidate path set, calculate the maximum power that can be carried on the entire path and generate a path power capability feature set; The optimal path selection is performed based on the path power capability feature set to obtain the photovoltaic charging and discharging path set.

6. The power management method applicable to photovoltaic energy storage and charging according to claim 1, characterized in that, The grid-connected peak shaving and regulation simulation in step S3 is specifically as follows: Based on the electrical status data frame, the grid-connected power assessment is performed on the power status of the terminal equipment connected to each path in the photovoltaic charging and discharging path set, thereby constructing a grid-connected power assessment matrix; Based on the multi-source physical parameters of the photovoltaic energy storage station and the capacity of the centralized station-end access transformer, grid connection control constraints and control cycles are set to obtain the grid connection control constraint table. Based on the grid-connected power assessment matrix and grid-connected regulation constraint table, the grid-connected power curve changes of all photovoltaic charging and discharging paths in each regulation cycle are simulated to predict the instantaneous total power change trend at the grid connection point. Paths that do not meet the grid-connected regulation constraint conditions are judged to fall back, thereby obtaining a peak shaving regulation simulation record set. The paths identified as being in a sustainable control state in the peak shaving and control simulation record set are extracted. The path numbers and device identifiers in the physical access topology matrix are bound to the paths in the sustainable control state, and the control priority, maximum adjustable power value and expected occupancy time are marked to obtain the feasible power channel table.

7. The power management method applicable to photovoltaic energy storage and charging according to claim 1, characterized in that, Step S4 is as follows: Step S41: Extract the characteristics of the electrical execution unit along the path based on the feasible power channel table, and generate the equipment control mapping table; Step S42: Perform structural analysis on the physical control parameters of the equipment for each path in the equipment control mapping table to obtain a set of control parameter templates; Step S43: Combine the equipment control mapping table and the control parameter template set, encapsulate them according to the path number, and construct a structured control instruction set; Step S44: Through the local control communication link of the photovoltaic energy storage station management platform, each instruction in the structured control instruction set is sent to the corresponding execution device channel by channel to execute the control instructions.

8. The power management method applicable to photovoltaic energy storage and charging according to claim 1, characterized in that, The specific steps for detecting the asymmetric load offset channel in step S5 are as follows: Based on the feedback data from the execution device, the real-time load characteristics of the device channel are extracted, and a channel feedback status matrix is ​​generated. Calculate the power flow amplitude offset value of each path in the current scheduling cycle based on the channel feedback state matrix, and record the offset duration, maximum offset amount and minimum current threshold trigger record to obtain the power offset rate feature set. Based on the physical access topology matrix, the physical channel depth, branch convergence density and hop count features of the offset path corresponding to the power offset rate feature set are extracted to obtain the offset path feature set. By combining the control priority field in the structured control instruction set, a candidate set of asymmetric load paths in the offset path feature set is screened; The load imbalance coefficient is evaluated for the candidate set of asymmetric load paths. When the load imbalance coefficient of any path is ≥0.75 and it is activated in two consecutive control cycles, the physical channel number and the device number of the affected node are extracted and encapsulated to generate an asymmetric load offset channel set.

9. The power management method applicable to photovoltaic energy storage and charging according to claim 1, characterized in that, Step S6 is as follows: Step S61: Extract the path number and device number from the asymmetric load offset channel set to generate a reset candidate control sequence; Step S62: Extract the relay device number of each path to be reset in the reset candidate control sequence, and retrieve the communication protocol, response time and isolation control parameters to construct a channel-by-channel disconnection command; Step S63: Encapsulate the channel-by-channel disconnection instructions to generate a disconnection control instruction set, and send each instruction in the disconnection control instruction set to the corresponding execution device for each channel to execute the disconnection task; Step S64: Obtain the real-time electrical parameters of the energy storage and charging station, and perform path channel differential pressure reconstruction based on the real-time electrical parameters of the energy storage and charging station to generate a power distribution compensation map; Step S65: Perform dynamic load redistribution based on the power distribution compensation map to generate a balanced control target matrix; Step S66: Encapsulate the equalization control target matrix into structured instructions to obtain a structured reset control instruction set, and send each instruction in the structured reset control instruction set to the corresponding execution device channel by channel to execute the control instructions.

10. A power management system suitable for photovoltaic energy storage and charging, characterized in that, For executing the power management method applicable to photovoltaic energy storage and charging as described in claim 1, the power management system applicable to photovoltaic energy storage and charging includes: The topology analysis module is used to obtain the multi-source physical parameter set of the photovoltaic energy storage and charging station, and to perform multi-source power access physical topology analysis on the multi-source physical parameter set of the photovoltaic energy storage and charging station to obtain the physical access topology matrix; The electrical status sampling module is used to perform high-frequency sampling of physical connections based on the physical access topology matrix to obtain electrical status data frames; and to classify the electrical status data frames into power reverse charging guns to obtain access charging gun classification electrical parameter data frames. The grid-connected peak shaving and control module is used to select high-power charging and discharging path channels for the classified electrical parameter data frames of the access charging guns according to the physical access topology matrix, thereby obtaining a photovoltaic charging and discharging path set; and to perform grid-connected peak shaving and control simulation on the photovoltaic charging and discharging path set to obtain a feasible power channel table. The instruction structuring module is used to structure the feasible power channel table channel by channel to obtain a structured control instruction set, and then send the structured control instruction set to the execution device one by one through the photovoltaic energy storage and charging station management platform to execute the control instructions. The execution deviation identification module is used to acquire feedback data from the execution device; to identify execution deviations based on the feedback data from the execution device; and to detect asymmetric load offset channels based on the execution deviation identification results. The safety reset module is used to perform a safety reset of the energy storage and charging station for asymmetrical load offset channels, and to perform dynamic balance redistribution based on the safety reset results to obtain a structured reset control instruction set. The structured reset control instruction set is then sent to the execution equipment one by one through the photovoltaic energy storage and charging station management platform to execute the control instructions.

Images