A power distribution network carrying capacity evaluation and operation device with light storage and charging cooperation

Abstract

The application discloses a kind of photovoltaic energy storage and charging collaborative distribution network carrying capacity evaluation and operation device, it is related to photovoltaic energy storage and charging collaborative relevant field, including multi-source heterogeneous data acquisition module, collaborative scheduling module, edge communication module, the multi-source heterogeneous data acquisition module is used to collect photovoltaic, energy storage, charging pile and distribution network operation data, based on real-time acquisition data constructs the real-time digital twin of each branch, transformer, switch device of distribution network, dynamically calculates the limit carrying capacity of each branch, residual carrying margin, voltage safety margin, transformer overload risk level under three time scales of current time and future 15 minutes, 30 minutes, 60 minutes.This kind of photovoltaic energy storage and charging collaborative distribution network carrying capacity evaluation and operation device, by digital twin mapping, fault pre-play verification and active prediction scheduling trinity architecture;Significantly reduce infrastructure investment, significantly improve photovoltaic energy storage and charging resource utilization and distribution network operation stability.

Description

Technical Field

[0001] This invention relates to the field of photovoltaic-storage-charging synergy technology, specifically to a distribution network carrying capacity assessment and operation device for photovoltaic-storage-charging synergy. Background Technology

[0002] Against the backdrop of the dual-carbon strategy and the construction of a new power system, distributed photovoltaic, electric vehicle charging piles, and energy storage equipment are being integrated into the distribution network on a large scale. Integrated photovoltaic, energy storage, and charging stations are being widely deployed, and traditional passive distribution networks are gradually being transformed into active distribution networks with bidirectional power fluctuations. Photovoltaic output is highly intermittent due to weather conditions, and the load impact of disorderly charging of electric vehicles is significant. When these are combined and connected, they can easily cause problems such as distribution network voltage exceeding limits, equipment overload, and reverse power flow exceeding standards, which restrict the absorption of new energy and significantly reduce the load acceptance and carrying capacity of the distribution network. Therefore, relying on the synergy of photovoltaic, energy storage, and charging to improve the carrying capacity of the distribution network has become the mainstream management approach.

[0003] As disclosed in prior art 1 (Chinese patent application number CN202410519516.0, application date April 26, 2024), an energy storage off-grid system is provided. In this system, a first PTC thermistor and a second PTC thermistor are installed on the first and second live wires between the inverter and the grid. Since the resistance of a PTC thermistor increases with temperature, when the grid current instantaneously enters the first and second PTC thermistors, their temperatures rise instantaneously, causing their resistances to increase rapidly. This results in a smaller current entering the inverter, thus protecting the switching transistors in the inverter from damage. Furthermore, in this embodiment, the first PTC thermistor and the first relay are connected in parallel with two relays connected on the first live wire near the inverter side, and the second PTC thermistor and the second relay are connected in parallel with two relays connected on the second live wire near the inverter side. Therefore, only one additional set of relays is needed. (First and second relays) also eliminate the need for grid wiring, which helps protect the switching transistors in the inverter from damage while reducing hardware costs. For example, the collaborative control method and device for an integrated photovoltaic-storage-charging smart microgrid disclosed in prior art 2 (Chinese patent application number CN202511719104.2, application date 2025.11.21) significantly reduces the photovoltaic curtailment rate and improves the efficiency of new energy consumption by prioritizing load supply and storing surplus energy through a hierarchical response mechanism, combined with a joint prediction model for precise supply and demand matching. Simultaneously, it integrates peak-valley electricity prices and carbon trading data, storing electricity during off-peak hours and releasing it during peak hours, reducing reliance on grid power supply during high-price periods, balancing economic benefits and carbon emission reduction targets, and is suitable for cost optimization needs in different scenarios such as communities and industrial parks. The modular design and energy balancing algorithm of the energy storage unit avoid single-module overload, extend the overall service life, and reduce the replacement and maintenance costs of the energy storage system. Local computation by the edge control unit reduces data transmission losses and centralized control costs, further improving the economic efficiency of microgrid operation.

[0004] The existing patents also have the following problems: 1. The load-bearing capacity assessment is based on static rated capacity and empirical coefficient correction, without establishing a four-dimensional dynamic coupling relationship between source, grid, load and storage, resulting in large assessment errors under extreme conditions; 2. The collaborative control is a passive response type, which only triggers adjustment when the distribution network is overloaded or the voltage exceeds the limit, lacking forward-looking prediction and fault prediction capabilities; 3. The hardware architecture is a splicing of multiple device functions, with photovoltaic, energy storage and charging pile controllers operating independently, resulting in prolonged data interaction time and failing to meet the millisecond-level collaborative response requirements; especially for rural power grids, which generally have thin lines, high impedance and small transformer capacity margins, existing general-purpose devices cannot be adapted, resulting in a large number of photovoltaic, energy storage and charging stations being forced to operate at reduced capacity due to insufficient distribution network load-bearing capacity, resulting in low resource utilization.

[0005] Therefore, we propose a photovoltaic-storage-charging integrated distribution network capacity assessment and operation device to address the problems mentioned above. Summary of the Invention

[0006] The purpose of this invention is to provide a distribution network carrying capacity assessment and operation device that integrates photovoltaic, energy storage, and charging technologies. This addresses the problems mentioned in the background section where current market carrying capacity assessments are based on static rated capacity and empirical coefficient corrections, failing to establish a four-dimensional dynamic coupling relationship between source, grid, load, and storage, resulting in large assessment errors under extreme conditions. Furthermore, the collaborative control is passive, triggering adjustments only when the distribution network is overloaded or voltage exceeds limits, lacking forward-looking prediction and fault prediction capabilities. The hardware architecture involves multiple devices with separate controllers for photovoltaic, energy storage, and charging piles, leading to prolonged data interaction and failing to meet millisecond-level collaborative response requirements. Especially in rural areas where power grids typically have thin lines, high impedance, and small transformer capacity margins, existing general-purpose devices are unsuitable, causing many photovoltaic, energy storage, and charging stations to be forced to operate at reduced capacity due to insufficient distribution network carrying capacity, resulting in low resource utilization.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a photovoltaic-storage-charging coordinated distribution network carrying capacity assessment and operation device, comprising a multi-source heterogeneous data acquisition module, a collaborative scheduling module, and an edge communication module. The multi-source heterogeneous data acquisition module is used to collect photovoltaic, energy storage, charging pile, and distribution network operation data. The collaborative scheduling module is used to generate photovoltaic-storage-charging power adjustment commands. The edge communication module is used to interact with the cloud master station. It also includes a carrying capacity digital twin mapping layer, a fault pre-simulation verification layer, and a rural power grid weak feeder adaptive compensation layer. The carrying capacity digital twin mapping layer is connected to the multi-source heterogeneous data acquisition module through a high-speed PCIe bus. It has a built-in source-grid-load-storage four-dimensional dynamic decoupling operation kernel and a distribution network digital twin model library. Based on real-time acquired data, it constructs real-time digital twins of each branch, transformer, and switchgear of the distribution network, and dynamically calculates the ultimate carrying capacity, remaining carrying capacity margin, voltage safety margin, and transformer overload risk level of each branch at the current time and at three time scales: 15 minutes, 30 minutes, and 60 minutes.

[0008] The fault pre-simulation verification layer's input end connects to the collaborative scheduling module, and its output end sends back the carrying capacity digital twin mapping layer. Before the control command is issued and executed, the command to be executed is input into the digital twin for full-condition simulation pre-simulation. The simulation simulates the changes in voltage of each node of the distribution network, current of each branch, and transformer load rate within 10 minutes after adjustment. After verifying that there is no risk of secondary overload, secondary voltage exceeding the limit, or aggravation of three-phase imbalance, the collaborative scheduling module is then authorized to issue the actual command.

[0009] The adaptive compensation layer for weak feeders in rural power grids has a database of typical line parameters for rural power grids in multiple cities. It includes impedance parameters of overhead conductors of different diameters, current-carrying temperature correction coefficients, transformer short-time overload withstand duration curves, and localized constraint parameters for the allowable voltage fluctuation range of low-voltage distribution networks. It automatically identifies weak feeder sections in the distribution network, calculates dynamic reactive power compensation capacity based on real-time voltage drop, and corrects the actual ultimate carrying capacity of weak feeders.

[0010] Preferably, the multi-source heterogeneous data acquisition module integrates a high-precision clock synchronization unit and a multi-protocol adaptive communication subunit. It adopts a fixed short-cycle synchronous sampling mechanism to uniformly collect three-phase electrical quantities, power, frequency, and harmonic indicators of distribution network feeders; output of distributed photovoltaic strings and inverter operating status; charging and discharging power of energy storage converters, battery SOC, battery temperature, and cycle life parameters; charging power of DC charging piles on both the single pile and group sides, charging equipment status, and scheduled charging load queue information; voltage deviation at the grid connection point of the power station, three-phase imbalance, and other comprehensive power quality data, ensuring the time synchronization and alignment of data collected by multi-source devices.

[0011] Preferably, the multi-protocol adaptive communication subunit has the ability to be compatible with and resolve multiple industrial communication protocols, and is equipped with multiple differentiated hardware communication interfaces. It can be compatible with heterogeneous devices such as existing photovoltaic, energy storage, charging piles, and distribution network measurement transformers without modification, and realize parallel data access of multiple types of devices.

[0012] Preferably, the load-bearing capacity digital twin mapping layer adopts a multi-processor heterogeneous collaborative computing architecture, including a high-speed parallel preprocessing kernel, a model scheduling kernel, and an intelligent prediction and inference kernel. The high-speed parallel preprocessing kernel completes millisecond-level real-time signal filtering and preprocessing, the model scheduling kernel is responsible for digital twin model instantiation, system logic control, and multi-task timing scheduling, and the intelligent prediction and inference kernel runs machine learning algorithms to complete short-term prediction of photovoltaic output and regional charging load. The overall data processing and model operation latency is controlled at the millisecond level, and a single device can support the synchronous load-bearing capacity assessment and coordinated control of multiple medium-voltage feeders, multiple sets of distributed photovoltaic power stations, energy storage systems, and DC charging pile clusters in parallel.

[0013] Preferably, the four-dimensional dynamic decoupling computation kernel of source-grid-load-storage decomposes the overall carrying capacity of the distribution network into four independent dimensions: power supply carrying capacity, network transmission, load acceptance, and energy storage regulation, which are modeled and solved separately. The power supply carrying capacity model incorporates the random fluctuation characteristics of photovoltaic output and the reactive power regulation range constraints of the inverter; the network transmission model combines the constraints of line impedance, current carrying capacity, voltage drop, and transformer rated capacity; the load acceptance model focuses on characterizing the high-power impact load of charging piles and the time-series imbalance characteristics; and the energy storage regulation model integrates the upper and lower limits of battery SOC, the upper limit of charging and discharging power, and the constraints of battery health degradation state. After the four dimensions are solved independently, the optimal solution of the distribution network carrying capacity that satisfies all safety constraints is solved by a global coupling optimization algorithm.

[0014] Preferably, the collaborative scheduling module is a pre-predictive collaborative scheduling unit equipped with a rolling time window optimization algorithm based on model predictive control; a fixed-duration basic optimization window is set, and the window is divided into multiple short-duration independent control steps. The device updates all photovoltaic storage and charging control strategies within the future window at a fixed cycle, and outputs continuous and smooth power adjustment commands to avoid voltage and current shocks to the distribution network caused by large-amplitude step power adjustments.

[0015] Preferably, the pre-predictive collaborative scheduling unit has built-in resource call priority logic adapted to different abnormal operating conditions of the distribution network; under the condition of distribution network overload, priority is given to calling energy storage discharge power to support the grid, then limiting the charging power of charging piles, and finally reducing the active power output of photovoltaics; under the condition of low node voltage, priority is given to activating energy storage reactive power support, then adjusting the reactive power output of inverters, and finally releasing the active power discharge of energy storage; under the condition of high node voltage, priority is given to controlling energy storage charging to absorb excess power, then adjusting photovoltaic reactive power, and then limiting the charging load of charging piles; under the condition of three-phase imbalance, priority is given to balancing the charging load of charging piles in each phase, and then three-phase imbalance compensation is carried out through energy storage devices.

[0016] Preferably, the edge communication module is an edge-cloud collaborative interaction unit with local data caching, offline autonomy, and breakpoint resume functions. When the communication link between the device and the cloud master station is interrupted, it automatically switches to local independent operation mode, autonomously completing load capacity calculation, wind and solar load prediction, and photovoltaic-storage-charging collaborative control based on the locally stored digital twin model and historical load data. It can operate stably without the cloud for a long time. After the communication link is restored, it automatically retransmits all offline operation data, control records, and fault alarm information. At the same time, it supports remote cloud-based upgrade programs to update the local digital twin model and collaborative control algorithm online without the need for on-site disassembly and debugging.

[0017] Preferably, the hardware security protection unit establishes secure links with the multi-source heterogeneous data acquisition module, the edge-cloud collaborative interaction unit, and the local human-machine interaction unit respectively; the protection unit has a built-in national cryptographic hardware encryption chip to achieve full-link encryption protection for data transmission, local storage, and device identity authentication; it integrates hardware firewall, intrusion behavior detection, and hierarchical access control mechanisms to resist abnormal network attacks; and it sets up a multi-level user permission isolation mechanism, so that users with different permissions can only access data and operation functions at the corresponding level, meeting the requirements of the second-level and above security protection specifications of power monitoring systems.

[0018] Preferably, it integrates a wide-temperature adaptability operating unit to adapt to a wide range of outdoor temperature and humidity environments; it is equipped with an intelligent temperature control heat dissipation mechanism that automatically starts and stops the heat dissipation components according to the temperature of the core components of the motherboard, balancing heat dissipation effect and equipment energy consumption; it is equipped with a dustproof and moisture-proof protective structure to adapt to outdoor installation scenarios of transformer substations in high temperature, severe cold, dusty, rain and snow, ensuring stable and reliable long-term continuous operation of the equipment.

[0019] The device has a built-in wide-temperature operating module with an operating temperature range of -40℃ to 70℃ and an operating humidity range of 5% to 95%. It is equipped with an intelligent active cooling system, including a semiconductor cooling chip, an axial fan, and heat sink fins. The cooling system automatically starts when the motherboard temperature exceeds 55℃ and automatically stops when the temperature drops below 40℃. It is also equipped with a dust filter and a desiccant, making it suitable for outdoor transformer substation installation environments with high summer temperatures, severe winter temperatures, dust, and rain and snow. The mean time between failures (MTBF) is ≥100,000 hours.

[0020] Compared with existing technologies, the beneficial effects of this invention are as follows: This photovoltaic-storage-charging collaborative distribution network capacity assessment and operation device addresses the common industry problems of low capacity assessment accuracy, passive and delayed regulation, easy secondary limit exceedances during regulation, lack of regional adaptability, and easy paralysis due to network outages after large-scale access of photovoltaic, storage, and charging to the distribution network. It solves these problems by combining regional pain points such as distributed photovoltaic fluctuations, high-power impacts from charging piles, thin rural power lines, high impedance, and small transformer margins. Furthermore, through a three-in-one architecture of digital twin mapping, fault pre-simulation verification, and proactive predictive scheduling, it effectively assesses the distribution network capacity. The calculation method has been upgraded from static rated capacity calculation to four-dimensional dynamic decoupled calculation of source, grid, load and storage, reducing the assessment error from more than 30% to less than 5%. It realizes the innovation of the control mode from passive remediation after the fact to active prediction before the fact, predicting the carrying capacity change 15-60 minutes in advance and taking measures. At the same time, through full-condition pre-control simulation and verification, the risk of secondary over-limit is completely eliminated. With built-in parameter library and weak feeder adaptive compensation function, the carrying capacity of weak feeders can be increased by 15%-25% without large-scale grid transformation, which significantly reduces infrastructure investment and significantly improves the utilization rate of photovoltaic, energy storage and charging resources and the stability of distribution network operation.

[0021] 1. A digital twin of the distribution network is constructed in real time through a digital twin mapping layer for carrying capacity. A four-dimensional dynamic decoupling algorithm for source, grid, load and storage is used to accurately calculate the carrying capacity margin at multiple time scales. Combined with an active prediction and collaborative scheduling unit, advance control is achieved. With the help of a fault pre-simulation and verification layer, full-condition simulation verification is completed before the command is issued. This not only ensures the timeliness and accuracy of control, but also avoids secondary faults caused by over- or under-adjustment from the source, significantly improving the safety and reliability of distribution network operation.

[0022] 2. This device has a hardware-based database of typical rural power grid line parameters, which can automatically identify weak feeder sections and perform dynamic reactive power compensation and current carrying capacity correction, specifically addressing the unique problems of thin conductors, high impedance, and large voltage fluctuations in rural power grids. It supports six mainstream industrial communication protocols and is compatible with more than 98% of photovoltaic, energy storage, and charging equipment on the market, allowing for rapid deployment without large-scale replacement of existing hardware. It is also equipped with an edge cloud collaboration unit and IP65 wide-temperature protection design, enabling it to operate independently for 168 hours in remote counties with unstable communication, adapting to harsh outdoor environments and significantly reducing engineering modification and maintenance costs. Attached Figure Description

[0023] Figure 1 This is a schematic diagram of the overall process of the present invention;

[0024] Figure 2 This is a schematic diagram of the multi-source heterogeneous data acquisition module of the present invention;

[0025] Figure 3 This is a schematic diagram of the digital twin mapping layer process for load-bearing capacity in this invention;

[0026] Figure 4 This is a schematic diagram of the collaborative scheduling module of the present invention;

[0027] Figure 5 This is a schematic diagram of the fault prediction and verification layer process of the present invention;

[0028] Figure 6 This is a schematic diagram of the adaptive compensation layer process for weak feeders in rural power grids according to the present invention.

[0029] Figure 7 This is a schematic diagram of the edge cloud collaboration unit of the present invention;

[0030] Figure 8 This is a schematic diagram of the wide-temperature operation module of the present invention. Detailed Implementation

[0031] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0032] Please see Figures 1-8 The present invention provides the following technical solution:

[0033] A photovoltaic-storage-charging (PV-SGC-Charging) distributed network carrying capacity assessment and operation device includes a multi-source heterogeneous data acquisition module, a collaborative scheduling module, and an edge communication module. The multi-source heterogeneous data acquisition module collects PV, energy storage, charging pile, and distributed network operation data. The collaborative scheduling module generates PV-SGC-Charging power adjustment commands. The edge communication module interacts with the cloud master station. It also includes a carrying capacity digital twin mapping layer, a fault prediction and verification layer, and a rural power grid weak feeder adaptive compensation layer. The carrying capacity digital twin mapping layer is connected to the multi-source heterogeneous data acquisition module via a high-speed PCIe bus. It incorporates a four-dimensional dynamic decoupling operation kernel for source-grid-load-storage and a distributed network digital twin model library. Based on real-time acquired data, it constructs real-time digital twins of each branch, transformer, and switchgear in the distributed network, dynamically calculating the ultimate carrying capacity, remaining carrying margin, and voltage of each branch at the current moment and at three time scales: 15 minutes, 30 minutes, and 60 minutes. Safety margin and transformer overload risk level; the fault pre-simulation verification layer's input end connects to the collaborative scheduling module, and the output end transmits back the carrying capacity digital twin mapping layer. Before the control command is issued and executed, the command to be executed is input into the digital twin for full-condition simulation pre-simulation. The simulation shows the changes in voltage of each node, current of each branch, and transformer load rate of the distribution network within 10 minutes after adjustment. After verifying that there is no risk of secondary overload, secondary voltage exceeding the limit, or aggravation of three-phase imbalance, the collaborative scheduling module is authorized to issue the actual command. The rural power grid weak feeder adaptive compensation layer solidifies the database of typical line parameters of rural power grids in multiple cities, including impedance parameters of overhead conductors of different diameters, current carrying capacity temperature correction coefficient, transformer short-time overload tolerance duration curve, and localized constraint parameters of the allowable voltage fluctuation range of low-voltage distribution network. It automatically identifies weak feeder sections in the distribution network, calculates dynamic reactive power compensation capacity based on real-time voltage drop, and corrects the actual ultimate carrying capacity of weak feeders.

[0034] This device, built upon the three fundamental modules of traditional photovoltaic, energy storage, and charging control, employs a three-layer core innovative architecture to form a closed-loop control system. A multi-source heterogeneous data acquisition module transmits 10ms-level real-time data without delay to the load-bearing capacity digital twin mapping layer via a PCIe 3.0 x4 bus. This layer rapidly constructs a digital twin that maps 1:1 to the physical distribution network using a local model library, and synchronously calculates load-bearing capacity indicators for the current and future one-hour timescales using a four-dimensional decoupling algorithm. Preliminary instructions generated by the collaborative scheduling module are first sent to the fault pre-simulation verification layer, where the adjustment effect is simulated for 10 minutes within the twin. Only after verifying the absence of secondary risks is the instruction authorized for release. The rural power grid weak feeder adaptive compensation layer utilizes a fixed multi-location rural power grid parameter library to automatically identify high-impedance weak feeders, calculate dynamic reactive power compensation, and correct load-bearing limits, specifically addressing the regional bottlenecks of thin conductors and high impedance in rural power grids.

[0035] The multi-source heterogeneous data acquisition module integrates a high-precision clock synchronization unit and a multi-protocol adaptive communication subunit. It adopts a fixed short-cycle synchronous sampling mechanism to uniformly collect three-phase electrical quantities, power, frequency, and harmonic indicators of distribution network feeders; output of distributed photovoltaic strings and inverter operating status; charging and discharging power of energy storage converters, battery SOC, battery temperature, and cycle life parameters; charging power of DC charging piles on both the single pile and group sides; charging equipment status; and information on scheduled charging load queues. It also collects full-dimensional power quality data such as voltage deviation at the grid connection point of the power station and three-phase imbalance, ensuring the time synchronization and alignment of data collected by multi-source devices.

[0036] The multi-source heterogeneous data acquisition module uses a high-precision GPS / BeiDou dual-mode clock synchronization chip as the time reference, with a timing accuracy better than 1μs, ensuring that the timing of data from all acquisition channels is completely aligned, avoiding errors in power calculation and power flow analysis caused by data misalignment. The module adopts a fixed sampling period of 10ms, which can capture the millisecond-level fluctuations of high-power impact loads of charging piles without generating excessive redundant data. It synchronously collects key parameters from all dimensions of the power source, grid, load, and energy storage: current and voltage on the distribution network side are used for power flow calculation, photovoltaic output data is used for power supply capacity assessment, SOC and temperature on the energy storage side are used for regulation capacity calculation, charging power and reservation queue on the charging pile side are used for load forecasting, and power quality data is used to assess power supply reliability, providing accurate data sources for subsequent capacity assessment and coordinated scheduling.

[0037] The multi-protocol adaptive communication subunit has the ability to be compatible with and resolve multiple industrial communication protocols. It is equipped with multiple differentiated hardware communication interfaces, which can be compatible with heterogeneous devices such as existing photovoltaic, energy storage, charging piles, and distribution network measurement transformers without modification. It enables parallel data access of multiple types of devices. The multi-protocol compatible communication subunit supports six mainstream industrial communication protocols: Modbus-RTU, Modbus-TCP, IEC61850, CANopen, MQTT, and OPCUA. It is equipped with 8 RS485 interfaces, 4 Ethernet interfaces, 2 CAN bus interfaces, and 1 fiber optic interface, and can simultaneously connect photovoltaic inverters, energy storage PCS, DC charging piles, and distribution network transformers.

[0038] The multi-protocol compatible communication subunit adopts a hybrid architecture of hardware protocol stack and software plug-in, integrating hardware parsing chips for six mainstream industrial communication protocols with a protocol conversion latency of less than 2ms. Eight RS485 interfaces connect to photovoltaic inverters and charging piles using the traditional Modbus protocol, four Ethernet interfaces support TCP / IP communication for next-generation smart devices, two CAN bus interfaces interface with the battery management system, and one fiber optic interface enables long-distance, high-interference-resistance substation data transmission. This design can simultaneously connect to different heterogeneous devices, is compatible with photovoltaic, energy storage, and charging hardware in most regions, and allows for system upgrades without large-scale replacement of existing equipment, significantly reducing project implementation costs and timelines. It is particularly suitable for mixed new and old substation application scenarios in county-level areas.

[0039] The load-bearing capacity digital twin mapping layer adopts a multi-processor heterogeneous collaborative computing architecture, including a high-speed parallel preprocessing kernel, a model scheduling kernel, and an intelligent prediction and inference kernel. The high-speed parallel preprocessing kernel completes millisecond-level real-time signal filtering and preprocessing, the model scheduling kernel is responsible for digital twin model instantiation, system logic control, and multi-task timing scheduling, and the intelligent prediction and inference kernel runs machine learning algorithms to complete short-term prediction of photovoltaic power output and regional charging load. The overall data processing and model operation latency is controlled at the millisecond level, and a single device can support the synchronous load-bearing capacity assessment and coordinated control of multiple medium-voltage feeders, multiple sets of distributed photovoltaic power stations, energy storage systems, and DC charging pile clusters in parallel.

[0040] The load-bearing capacity digital twin mapping layer adopts a three-core heterogeneous computing architecture of FPGA+ARM+NPU, leveraging the advantages of each processor to achieve division of labor and collaboration. The FPGA core utilizes parallel computing characteristics to complete the filtering, noise reduction, and format conversion of multiple raw data within 10ms, while directly acquiring analog signals from the transformers to ensure real-time signal performance. The ARM core runs an embedded Linux system, responsible for the construction, updating, and overall system logic scheduling of the digital twin model. The NPU core is specifically designed to accelerate LSTM neural network inference, enabling high-precision prediction of photovoltaic output and load. The three-core parallel operation compresses the overall latency to within 50ms, meeting the real-time control requirements of the distribution network. A single motherboard with 8GB DDR4 memory and 64GB storage can simultaneously support the access of 8 10kV feeders and hundreds of devices, adapting to the centralized management and control needs of multiple substations in county-level distribution networks.

[0041] The four-dimensional dynamic decoupled computation kernel of power generation, grid, load, and energy storage decomposes the overall carrying capacity of the distribution network into four independent dimensions: power generation capacity, network transmission, load acceptance, and energy storage regulation. Each dimension is modeled and solved separately. The power generation capacity model incorporates the random fluctuation characteristics of photovoltaic output and the reactive power regulation range constraints of the inverter. The network transmission model combines the constraints of line impedance, current carrying capacity, voltage drop, and transformer rated capacity. The load acceptance model focuses on characterizing the high-power impact load of charging piles and the time-series imbalance characteristics. The energy storage regulation model integrates the upper and lower limits of battery SOC, the upper limit of charging and discharging power, and the constraints of battery health degradation state. After the four dimensions are solved independently, the optimal solution of the distribution network carrying capacity that satisfies all safety constraints is solved through a global coupled optimization algorithm.

[0042] The four-dimensional dynamic decoupled computational kernel for power generation, grid, load, and energy storage breaks through the limitations of traditional overall coupled calculations. It decomposes the complex distribution network carrying capacity problem into four independent sub-problems, which are modeled and solved separately. In the power generation dimension, a probability distribution function is used to describe the randomness of photovoltaic power output, while considering the reactive power regulation range of the inverter from -0.9 to +0.9. In the network dimension, based on measured line impedance and transformer parameters, the transmission limit is determined through power flow calculation. In the load dimension, a time-series model of the high-power impact load of charging piles is established to quantify the impact of three-phase imbalance. In the energy storage dimension, the available regulating capacity is calculated by comprehensively considering the upper and lower limits of SOC, the charge and discharge rate, and the SOH attenuation. After solving the four dimensions independently, a particle swarm optimization algorithm is used for global coupled optimization to maximize the distribution network carrying capacity under the premise of satisfying all constraints, reducing the evaluation error from more than 30% in the traditional method to less than 5%.

[0043] The collaborative scheduling module is a forward-predictive collaborative scheduling unit equipped with a rolling time window optimization algorithm based on model predictive control. It sets a fixed-duration basic optimization window, and the window is divided into multiple short-duration independent control steps. The device updates all photovoltaic, energy storage and charging control strategies within the future window at a fixed cycle, and outputs continuous and smooth power adjustment commands to avoid voltage and current shocks to the distribution network caused by large-scale step power adjustments.

[0044] The proactive predictive collaborative scheduling unit adopts a model predictive control algorithm to replace the traditional threshold-based passive control. It selects 15 minutes as the basic optimization time window to match the effective accuracy range of photovoltaic and load forecasts. The window is divided into three 5-minute control steps to match the equipment power regulation response time. Every 5 minutes, the algorithm re-solves the optimal power allocation scheme for the next 15 minutes based on the latest collected data and prediction results, and only executes the control action for the first 5 minutes. The subsequent step size is updated in the next cycle. This rolling optimization mechanism can continuously use the latest information to correct the strategy, achieving continuous and smooth control with each adjustment amplitude not exceeding 10% of the rated power. This avoids the drastic voltage fluctuations and equipment impacts caused by traditional step regulation, stabilizing the distribution network voltage fluctuation within ±2%.

[0045] The pre-predictive collaborative scheduling unit has built-in resource allocation priority logic adapted to different abnormal operating conditions of the distribution network. Under distribution network overload conditions, it prioritizes the use of energy storage discharge power to support the grid, then limits the charging power of charging piles, and finally reduces the active power output of photovoltaics. Under low node voltage conditions, it prioritizes the use of energy storage reactive power support, then adjusts the reactive power output of inverters, and finally releases the active power discharge of energy storage. Under high node voltage conditions, it prioritizes controlling the charging of energy storage to absorb excess power, then adjusts the reactive power of photovoltaics, and then limits the charging load of charging piles. Under three-phase imbalance conditions, it prioritizes balancing the charging load of charging piles in each phase, and then carries out three-phase imbalance compensation through energy storage devices.

[0046] The proactive predictive and collaborative scheduling unit incorporates multi-condition adaptive priority logic, automatically switching control targets and resource call order based on the real-time operating status of the distribution network. Under overload conditions, it prioritizes the discharge of energy storage due to its fast response speed and lack of impact on user experience. Secondly, it limits the power of charging piles, only reducing charging speed without interrupting service. Finally, it limits photovoltaic output to reduce the curtailment of renewable energy. Voltage issues are primarily addressed through reactive power regulation, as it has a more direct and economical impact on voltage, avoiding energy losses caused by active power regulation. Three-phase imbalances are addressed by adjusting the three-phase load distribution of charging piles at the source, supplemented by passive compensation from energy storage. This differentiated priority strategy maximizes user experience and renewable energy absorption rate while ensuring distribution network safety.

[0047] The edge communication module is an edge-cloud collaborative interaction unit with local data caching, offline autonomy, and breakpoint resume functions. When the communication link between the device and the cloud master station is interrupted, it automatically switches to local independent operation mode. Relying on the locally stored digital twin model and historical load data, it autonomously completes the carrying capacity calculation, wind and solar load prediction, and photovoltaic-storage-charging collaborative control, and can operate stably without the cloud for a long time. After the communication link is restored, it automatically retransmits all offline operation data, control records, and fault alarm information. At the same time, it supports remote cloud-based upgrade programs to update the local digital twin model and collaborative control algorithm online without the need for on-site disassembly and debugging.

[0048] The edge cloud collaborative unit adopts a hybrid architecture of cloud-based coordination and edge autonomy to solve the problem of unstable communication in remote areas. The unit monitors the heartbeat connection with the master station in real time. If no response is received for three consecutive times, the communication is considered interrupted, and the unit automatically switches to local independent operation mode. In offline mode, the device relies on 7 days of historical typical daily data stored locally and time series algorithms to perform load and photovoltaic forecasting. It completes all load capacity assessment and collaborative control functions through a local digital twin model and can operate independently for up to 168 hours. After communication is restored, it automatically resumes the transmission of all data during the offline period from the point of interruption to ensure data integrity. At the same time, it supports cloud-encrypted push of model update packages, and the background automatically completes the upgrade without the need for on-site debugging, which greatly reduces operation and maintenance costs.

[0049] The hardware security protection unit establishes secure links with the multi-source heterogeneous data acquisition module, the edge-cloud collaborative interaction unit, and the local human-machine interaction unit. The protection unit has a built-in national cryptographic hardware encryption chip to achieve full-link encryption protection for data transmission, local storage, and device identity authentication. It integrates a hardware firewall, intrusion detection, and hierarchical access control mechanisms to resist abnormal network attacks. It sets up a multi-level user permission isolation mechanism, so that users with different permissions can only access data and operation functions at the corresponding level, meeting the requirements of the second-level and above security protection specifications for power monitoring systems.

[0050] The hardware security protection unit constructs a full-link security protection system, covering the entire process of data collection, transmission, storage, and interaction. It incorporates national standard SM2 / SM3 / SM4 hardware encryption chips, using SM2 for identity authentication and key exchange, SM4 for end-to-end encryption of data transmission and storage, and SM3 for data integrity verification. The encryption speed is more than 100 times faster than software encryption. It also integrates hardware firewall, intrusion detection, and access control functions, filtering illegal traffic through a whitelist mechanism, monitoring abnormal behavior in real time, and automatically blocking it. It supports three-level permission management for administrators, operators, and visitors, ensuring that users with different permissions can only access corresponding functions and data.

[0051] It integrates a wide-temperature adaptability operating unit, suitable for a wide range of outdoor temperature and humidity environments; it is equipped with an intelligent temperature control heat dissipation mechanism that automatically starts and stops the heat dissipation components according to the temperature of the core components on the motherboard, balancing heat dissipation effect and equipment energy consumption; it is equipped with a dustproof and moisture-proof protective structure, suitable for outdoor installation scenarios of transformer substations in high temperature, extreme cold, dusty, rain and snow, ensuring stable and reliable long-term continuous operation of the equipment.

[0052] Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A photovoltaic-storage-charging coordinated distribution network carrying capacity assessment and operation device, comprising a multi-source heterogeneous data acquisition module, a coordinated scheduling module, and an edge communication module, wherein the multi-source heterogeneous data acquisition module is used to collect photovoltaic, energy storage, charging pile, and distribution network operation data; the coordinated scheduling module is used to generate photovoltaic-storage-charging power adjustment commands; and the edge communication module is used to interact with a cloud master station, characterized in that... It also includes a load-bearing capacity digital twin mapping layer, a fault prediction and verification layer, and a rural power grid weak feeder adaptive compensation layer. The load-bearing capacity digital twin mapping layer is connected to a multi-source heterogeneous data acquisition module through a high-speed PCIe bus. It has a built-in source-grid-load-storage four-dimensional dynamic decoupling calculation kernel and a distribution network digital twin model library. Based on real-time acquired data, it constructs real-time digital twins of each branch, transformer, and switchgear in the distribution network, and dynamically calculates the ultimate load-bearing capacity, remaining load-bearing margin, voltage safety margin, and transformer overload risk level of each branch at the current time and the next 15 minutes, 30 minutes, and 60 minutes. The fault pre-simulation verification layer's input end connects to the collaborative scheduling module, and its output end sends back the carrying capacity digital twin mapping layer. Before the control command is issued and executed, the command to be executed is input into the digital twin for full-condition simulation pre-simulation. The simulation simulates the changes in voltage of each node of the distribution network, current of each branch, and transformer load rate within 10 minutes after adjustment. After verifying that there is no risk of secondary overload, secondary voltage exceeding the limit, or aggravation of three-phase imbalance, the collaborative scheduling module is then authorized to issue the actual command. The adaptive compensation layer for weak feeders in rural power grids has a database of typical line parameters for rural power grids in multiple cities. It includes impedance parameters of overhead conductors of different diameters, current-carrying temperature correction coefficients, transformer short-time overload withstand duration curves, and localized constraint parameters for the allowable voltage fluctuation range of low-voltage distribution networks. It automatically identifies weak feeder sections in the distribution network, calculates dynamic reactive power compensation capacity based on real-time voltage drop, and corrects the actual ultimate carrying capacity of weak feeders.

2. The photovoltaic-storage-charging coordinated distribution network carrying capacity assessment and operation device according to claim 1, characterized in that: The multi-source heterogeneous data acquisition module integrates a high-precision clock synchronization unit and a multi-protocol adaptive communication subunit. It adopts a fixed short-cycle synchronous sampling mechanism to uniformly collect three-phase electrical quantities, power, frequency, and harmonic indicators of distribution network feeders; output of distributed photovoltaic strings and inverter operating status; charging and discharging power of energy storage converters, battery SOC, battery temperature, and cycle life parameters; charging power of DC charging piles on both the single-pile and group sides; charging equipment status; and information on scheduled charging load queues. It also collects full-dimensional power quality data such as voltage deviation at the grid connection point of the power station and three-phase imbalance, ensuring the time synchronization and alignment of data collected by multi-source devices.

3. The photovoltaic-storage-charging coordinated distribution network carrying capacity assessment and operation device according to claim 2, characterized in that: The multi-protocol adaptive communication subunit has the ability to be compatible with and resolve multiple industrial communication protocols, and is equipped with multiple differentiated hardware communication interfaces. It can be compatible with heterogeneous devices such as existing photovoltaic, energy storage, charging piles, and distribution network measurement transformers without modification, and realize parallel data access of multiple types of devices.

4. The photovoltaic-storage-charging coordinated distribution network carrying capacity assessment and operation device according to claim 1, characterized in that: The load-bearing capacity digital twin mapping layer adopts a multi-processor heterogeneous collaborative computing architecture, including a high-speed parallel preprocessing kernel, a model scheduling kernel, and an intelligent prediction and inference kernel. The high-speed parallel preprocessing kernel completes millisecond-level real-time signal filtering and preprocessing, the model scheduling kernel is responsible for digital twin model instantiation, system logic control, and multi-task timing scheduling, and the intelligent prediction and inference kernel runs machine learning algorithms to complete short-term prediction of photovoltaic power output and regional charging load. The overall data processing and model operation latency is controlled at the millisecond level, and a single device can support the synchronous load-bearing capacity assessment and coordinated control of multiple medium-voltage feeders, multiple sets of distributed photovoltaic power stations, energy storage systems, and DC charging pile clusters in parallel.

5. The photovoltaic-storage-charging coordinated distribution network carrying capacity assessment and operation device according to claim 4, characterized in that: The four-dimensional dynamic decoupling operation kernel of source-grid-load-storage decomposes the overall carrying capacity of the distribution network into four independent dimensions: power carrying capacity, network transmission, load acceptance, and energy storage regulation, and models and solves them separately. The power supply carrying capacity model incorporates the random fluctuation characteristics of photovoltaic power output and the reactive power regulation range constraints of the inverter. The network transmission model combines line impedance, current carrying capacity, voltage drop, and transformer rated capacity constraints. The load acceptance model focuses on characterizing the high-power impact load of charging piles and the time-series imbalance characteristics. The energy storage regulation model integrates the upper and lower limits of battery SOC, the upper limit of charging and discharging power, and the battery health degradation state constraints. After the four dimensions are solved independently, the optimal solution of distribution network carrying capacity that satisfies all safety constraints is solved through a global coupled optimization algorithm.

6. The photovoltaic-storage-charging coordinated distribution network carrying capacity assessment and operation device according to claim 1, characterized in that: The collaborative scheduling module is a pre-predictive collaborative scheduling unit equipped with a rolling time window optimization algorithm based on model predictive control. It sets a fixed-duration basic optimization window, and the window is divided into multiple short-duration independent control steps. The device updates all photovoltaic, energy storage and charging control strategies within the future window at a fixed cycle, and outputs continuous and smooth power adjustment commands to avoid voltage and current shocks to the distribution network caused by large-amplitude step power adjustments.

7. The photovoltaic-storage-charging coordinated distribution network carrying capacity assessment and operation device according to claim 6, characterized in that: The pre-predictive collaborative scheduling unit has built-in resource allocation priority logic adapted to different abnormal operating conditions of the distribution network. Under distribution network overload conditions, it prioritizes the use of energy storage discharge power to support the grid, then limits the charging power of charging piles, and finally reduces the active power output of photovoltaics. Under low node voltage conditions, it prioritizes the use of energy storage reactive power support, then adjusts the reactive power output of inverters, and finally releases the active power discharge of energy storage. Under high node voltage conditions, it prioritizes controlling the charging of energy storage to absorb excess power, then adjusts the reactive power of photovoltaics, and then limits the charging load of charging piles. Under three-phase imbalance conditions, it prioritizes balancing the charging load of charging piles in each phase, and then carries out three-phase imbalance compensation through energy storage devices.

8. The photovoltaic-storage-charging coordinated distribution network carrying capacity assessment and operation device according to claim 1, characterized in that: The edge communication module is an edge-cloud collaborative interaction unit with local data caching, offline autonomy, and breakpoint resume functions. When the communication link between the device and the cloud master station is interrupted, it automatically switches to local independent operation mode. Relying on the locally stored digital twin model and historical load data, it autonomously completes the carrying capacity calculation, wind and solar load prediction, and photovoltaic-storage-charging collaborative control, and can operate stably without the cloud for a long time. After the communication link is restored, it automatically retransmits all offline operation data, control records, and fault alarm information. At the same time, it supports remote upgrade program distribution from the cloud and online updates of the local digital twin model and collaborative control algorithm without the need for on-site disassembly and debugging.

9. The photovoltaic-storage-charging coordinated distribution network carrying capacity assessment and operation device according to claim 1, characterized in that: The hardware security protection unit establishes secure links with the multi-source heterogeneous data acquisition module, the edge-cloud collaborative interaction unit, and the local human-machine interaction unit. The protection unit has a built-in national cryptographic hardware encryption chip to achieve full-link encryption protection for data transmission, local storage, and device identity authentication. It integrates a hardware firewall, intrusion detection, and hierarchical access control mechanisms to resist abnormal network attacks. It sets up a multi-level user permission isolation mechanism, so that users with different permissions can only access data and operation functions at the corresponding level, meeting the requirements of the second-level and above security protection specifications for power monitoring systems.

10. The photovoltaic-storage-charging coordinated distribution network carrying capacity assessment and operation device according to claim 1, characterized in that: It integrates a wide-temperature adaptability operating unit, suitable for a wide range of outdoor temperature and humidity environments; it is equipped with an intelligent temperature control heat dissipation mechanism that automatically starts and stops the heat dissipation components according to the temperature of the core components on the motherboard, balancing heat dissipation effect and equipment energy consumption; it is equipped with a dustproof and moisture-proof protective structure, suitable for outdoor installation scenarios of transformer substations in high temperature, extreme cold, dusty, rain and snow, ensuring stable and reliable long-term continuous operation of the equipment.

Images