A cloud-edge collaborative energy management method and system for intelligent power supply pile groups in ports

Abstract

This invention proposes a cloud-edge collaborative energy management method and system for intelligent power supply pile groups in ports, belonging to the field of port charging management technology. It addresses the problems of existing technologies, such as insufficient compatibility of port power supply facilities failing to meet the differentiated power supply needs of various types of vessels, lack of a global optimization perspective in energy management, and the absence of a cloud-edge collaborative mechanism making it difficult to balance global optimization and real-time response. This invention achieves global optimization and collaborative control of compatible access for various types of vessels, power supply costs, renewable energy consumption, and grid peak-valley differences by constructing an intelligent power supply pile group integrating multiple power supply modes and a cloud-edge collaborative energy management architecture.

Description

Technical Field

[0001] This invention relates to the field of port charging management technology, and in particular to a cloud-edge collaborative energy management method and system for intelligent power supply pile groups in ports. Background Technology

[0002] With the greening and decarbonization of ports worldwide, the energy supply methods for ships during berthing are undergoing profound changes. Traditional shore power systems primarily provide 50Hz or 60Hz AC power to conventional ships using AC power systems, replacing onboard diesel generators to reduce emissions. However, in recent years, the electrification and diversification of ship propulsion systems have accelerated significantly, with a rapid increase in the number of new types of ships such as electric ships, hybrid ships, and LNG-electric hybrid ships, placing complex power supply demands on port power supply facilities in multiple types, systems, and power levels. Specifically, existing port power supply facilities face the following challenges: 1) Single power supply mode: Traditional shore power facilities only support AC power output, which cannot meet the DC fast charging needs of electric ships and the bidirectional charging and discharging needs of energy storage ships; 2) Simple and crude energy management: Existing shore power systems are usually allocated according to the declared power of ships, lacking comprehensive optimization of port area renewable energy generation, grid time-of-use pricing and ship load characteristics, resulting in high power supply costs and low renewable energy absorption rate; 3) Lack of collaborative optimization mechanism: There is a lack of unified energy collaborative scheduling among multiple power supply piles, which easily leads to local overload and low overall efficiency, especially when multiple different types of ships are connected at the same time, the power allocation lacks a global optimization perspective; 4) Insufficient real-time response capability: Centralized cloud control relies on stable communication links, and in the event of communication delays or interruptions, it lacks local autonomous decision-making capability, making it difficult to ensure the continuity and security of power supply.

[0003] Existing research on ship shore power supply systems and methods based on cloud platforms and smart terminals mainly focuses on the proximity matching of ships and power supply piles and metering and billing functions, without addressing the identification of load characteristics and energy optimization scheduling for various types of ships. Multi-objective optimization configuration methods for port ship shore power systems consider the optimized construction of shore power system capacity, but the optimization objectives are focused on the planning level and do not involve real-time energy management at the operational level. In the field of smart charging piles, electric vehicle charging piles supporting multiple protocols and power levels are relatively mature, but their application scenario is for road vehicles, lacking adaptation solutions for the special electrical systems, high power, and diverse load characteristics of port ships, as well as the coordination of new energy sources in port areas. Cloud-edge collaborative technology has been applied in the industrial field, but in the charging and power supply scenario of port ships, how to effectively combine global optimization in the cloud with real-time control at the edge has yet to be seen.

[0004] In summary, existing technologies suffer from the following problems: insufficient compatibility of power supply facilities, failing to meet the differentiated power supply needs of various ship types; lack of a global optimization perspective in energy management, failing to comprehensively consider power supply costs, renewable energy consumption, and grid friendliness; and lack of cloud-edge collaboration mechanisms, making it difficult to balance the needs of global optimization and real-time response. Therefore, it is necessary to propose a new technical solution to achieve compatible access for charging and power supply to various ship types and intelligent energy management through cloud-edge collaboration. Summary of the Invention

[0005] To address the problems in existing technologies, such as insufficient compatibility of existing port power supply facilities to meet the differentiated power supply needs of various types of ships, lack of a global optimization perspective in energy management, and the absence of a cloud-edge collaboration mechanism that makes it difficult to balance global optimization and real-time response, this invention constructs an intelligent power supply pile group integrating multiple power supply modes and a cloud-edge collaborative energy management architecture. This achieves global optimization and collaborative control of compatible access for various types of ships, power supply costs, renewable energy consumption, and peak-valley differences in the power grid. Firstly, a cloud-edge collaborative energy management method for a port intelligent power supply pile group is proposed, specifically including the following steps: S1. Construct a smart power supply pile group for the port. The pile group includes multiple smart power supply piles and at least one edge computing node. Each smart power supply pile integrates an AC shore power module, a DC fast charging module, and a bidirectional charging and discharging module. The edge computing node enables pile group-level data aggregation and local control. S2. Collect multi-source operational status data from the ship, pile, and net sides. After data preprocessing and feature extraction by the edge computing node, upload the data to the cloud energy management platform. S3. The cloud-based pre-trained ship load characteristic classification model identifies the load type of connected ships and, in conjunction with port area new energy power generation forecasts and grid time-of-use electricity price information, establishes a multi-objective collaborative optimization model. S4. Solve the multi-objective collaborative optimization model to obtain the power allocation scheme, power supply mode switching strategy and charging and discharging strategy of each smart power supply pile in the pile group. S5. Edge computing nodes receive optimization strategies from the cloud and adjust the strategies locally based on real-time operating status, adjusting the output power, power supply mode, and energy storage converter control parameters of each smart power supply pile. S6. Power supply piles provide services. Each smart power supply pile provides charging and power supply services to the connected ships according to the received power instructions and power supply mode. At the same time, the port area energy storage system performs energy throughput according to the charging and discharging strategy.

[0006] Furthermore, the AC shore power module of the smart power supply pile achieves power system conversion through power electronic transformers and solid-state switches, supporting standard marine power systems of 6.6kV / 60Hz, 440V / 60Hz, and 400V / 50Hz; the DC fast charging module supports DC output with a wide voltage range of 400V to 1000V, and the output power level is 500kW to 2MW; the bidirectional charging and discharging module supports V2G function and is configured to receive energy fed back to the port power grid by energy storage electric ships.

[0007] Furthermore, the multi-source operational status data includes: access request signals from the ship side, communication data of the ship's battery management system, load demand power, and berthing time window; voltage, current, output power, operating mode, and interface temperature from the power supply pile side; predicted values ​​of new energy power generation and state of charge of the energy storage system from the port area side; and real-time electricity price, dispatch instructions, and voltage frequency from the grid side.

[0008] Furthermore, the pre-trained ship load feature classification model is built based on a deep convolutional neural network and a long short-term memory network. The output load type labels include: traditional marine AC load, electric ship DC fast charging load, hybrid ship intermittent charging load, energy storage ship load that can participate in V2G, and emergency rapid power supply load.

[0009] Furthermore, the objective function of the multi-objective collaborative optimization model is: ; in, For the overall power supply cost, Penalties for abandoned electricity from renewable energy sources The peak-valley difference penalty for the power grid is represented by w1, w2, and w3, which are the corresponding weighting coefficients.

[0010] Furthermore, the constraints of the multi-objective collaborative optimization model include power balance constraints, output power constraints of each power supply pile, ship charging demand constraints, port area energy storage system operation constraints, and power grid interaction constraints; the power balance constraint is the instantaneous power balance between the power grid, new energy sources, energy storage, and power supply piles.

[0011] Furthermore, the operational constraints of the port area energy storage system include: maintaining the state of charge between preset upper and lower limits, ensuring that the charging and discharging power does not exceed the maximum allowable value, and ensuring that the state of charge satisfies the energy conservation recursive relationship.

[0012] Furthermore, the local feedback adjustments performed by the edge computing nodes include: triggering emergency protection strategies when an anomaly is detected; enabling local offline optimization mode and using rolling time-domain optimization to autonomously generate a power allocation scheme when communication latency exceeds a threshold; and periodically synchronizing the correction results to the cloud to update model parameters.

[0013] Furthermore, the power supply mode switching strategies include: locking traditional AC loads into AC shore power mode; prioritizing DC fast charging loads and dynamically adjusting power; increasing charging power for intermittent charging loads during off-peak periods; allowing V2G vessels to feed back electricity during peak electricity prices or power shortages; and rapidly increasing emergency loads to their rated maximum power.

[0014] According to another aspect of the present invention, a port intelligent power supply pile group cloud-edge collaborative energy management system is also proposed, applied to the above-mentioned port intelligent power supply pile group cloud-edge collaborative energy management method, comprising: The physical layer of the pile group includes multiple smart power supply piles and port area energy storage system. The smart power supply piles integrate AC shore power modules, DC fast charging modules and bidirectional charging and discharging modules. The edge computing node layer is configured to collect multi-source operational status data from three sides: ship, pile, and network; perform data preprocessing and feature extraction; receive cloud optimization strategies and execute local feedback adjustments; and send control commands to each smart power supply pile. The cloud-based energy management platform is configured to operate a ship load characteristic classification model, a new energy power generation prediction model, and a multi-objective collaborative optimization model. It generates a pile group power allocation scheme and an energy storage charging and discharging strategy, and distributes them to edge computing nodes.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: Firstly, this invention enables compatibility and flexible power supply for multiple types of ships. By integrating an AC shore power module, a DC fast charging module, and a bidirectional charging and discharging module into each smart power supply pile, this invention can be compatible with various power supply needs such as traditional AC ships, DC fast charging for electric ships, intermittent charging for hybrid ships, and V2G for energy storage ships. This solves the problem of single-system and poor compatibility of existing port power supply facilities, and greatly improves the utilization rate and applicability of port power supply facilities. Secondly, regarding power supply cost optimization and efficient renewable energy consumption, a multi-objective collaborative optimization model was established in the cloud based on a ship load characteristic classification model and renewable energy power generation forecast. The model aims to minimize overall power supply costs, maximize renewable energy consumption rate, and minimize grid peak-valley differences, achieving globally optimal allocation of power to the pile group. This solution effectively reduces grid power purchase costs, increases the self-consumption ratio of renewable energy sources such as photovoltaic and wind power in the port area, and avoids energy waste. Third, cloud-edge collaboration ensures real-time reliability and grid-friendliness, constructing a collaborative architecture of global optimization in the cloud and local correction at the edge: the cloud is responsible for generating long-cycle strategies, while edge computing nodes are responsible for real-time data acquisition and feedback adjustment. In the event of communication delays or interruptions, edge nodes can autonomously activate local offline optimization mode to ensure power supply continuity and security. Simultaneously, by optimizing the peak-valley difference in the power grid and utilizing V2G functionality to assist in peak shaving, the level of friendly interaction between the port and the power grid is significantly improved. Attached Figure Description

[0016] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0017] Figure 1 This is a flowchart of a port smart power supply pile group and cloud-edge collaborative energy management method that supports charging and power supply for multiple types of ships, according to an embodiment of the present invention. Figure 2 This is a schematic diagram of the cloud-edge-device three-layer architecture of the method described in the embodiments of the present invention; Figure 3 This is a schematic diagram of the internal structure of the intelligent power supply pile in an embodiment of the present invention; Figure 4 This is a schematic diagram of the network structure of the ship load feature classification model in an embodiment of the present invention; Figure 5 This is a schematic diagram of the Pareto front solution set of the multi-objective collaborative optimization model in an embodiment of the present invention; Figure 6 This is a timing diagram of cloud-edge collaborative control in an embodiment of the present invention; Figure 7 This is a schematic diagram of the structure of a port intelligent power supply pile group and cloud-edge collaborative energy management system that supports charging and power supply for multiple types of ships in an embodiment of the present invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below. Obviously, the described embodiments are only a part of the embodiments of this invention, and not all of them. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0019] The specific embodiments of the present invention will be described below.

[0020] To address the problems in existing port power supply facilities, such as insufficient compatibility to meet the diverse power supply needs of various ship types, lack of a global optimization perspective in energy management, and the absence of a cloud-edge collaboration mechanism that makes it difficult to balance global optimization and real-time response, this invention proposes a cloud-edge collaborative energy management method and system for intelligent power supply pile groups in ports. By constructing an intelligent power supply pile group integrating multiple power supply modes and a cloud-edge collaborative energy management architecture, it achieves global optimization and collaborative control of compatible access for various ship types, power supply costs, renewable energy consumption, and peak-valley differences in the power grid.

[0021] Example 1 like Figure 1 , Figure 2 and Figure 3 As shown, this invention proposes a cloud-edge collaborative energy management method for intelligent power supply pile groups in ports. By constructing an intelligent power supply pile group integrating AC shore power, DC fast charging, and bidirectional charging and discharging, and combining a cloud-edge collaborative architecture of global multi-objective optimization in the cloud and local real-time feedback adjustment at the edge, it achieves compatible power supply for multiple types of ships, global optimization of power supply costs and renewable energy consumption, and autonomous and reliable control under communication anomalies, including: S1. Construct a smart power supply pile group for the port. The pile group includes multiple smart power supply piles and at least one edge computing node. Each smart power supply pile integrates an AC shore power module, a DC fast charging module, and a bidirectional charging and discharging module. The edge computing node enables pile group-level data aggregation and local control.

[0022] Specifically, the constructed port smart power supply network comprises N smart power supply piles and at least one edge computing node. Each smart power supply pile is connected to the edge computing node via an industrial Ethernet or fiber optic ring network. The edge computing node is connected to the cloud-based energy management platform via a 4G / 5G private network or wired broadband. Each smart power supply pile integrates three core modules. The AC shore power module uses a cascaded H-bridge topology power electronic transformer and solid-state switches to achieve power conversion, outputting multiple standard marine power systems to meet the flexible access requirements of different types of AC vessels, such as 6.6kV / 60Hz, 440V / 60Hz, and 400V / 50Hz standard marine power systems. The DC fast charging module, based on a modular multilevel converter topology, supports a wide voltage range of DC output, such as DC 400V to 1000V, with output power levels ranging from 500kW to 2MW. Power expansion is achieved through parallel connection of multiple modules, and it supports mainstream DC charging communication protocols for interaction with the electric vessel battery management system. The bidirectional charging / discharging module shares a power conversion circuit with the DC fast charging module and adds a V2G communication protocol stack. When the ship's battery SOC exceeds a preset feedback threshold and the grid electricity price exceeds a set trigger value, it automatically switches to inverter mode to feed electrical energy back to the port grid. The three modules are interconnected through an internal DC bus and a switching switch. The edge computing node controls the switching switch to select the currently active power supply mode based on cloud optimization strategies and real-time operating conditions, and sends power commands to the corresponding module.

[0023] S2. Collect multi-source operational status data from the ship, power pile, and grid sides. After data preprocessing and feature extraction at the edge computing node, the data is uploaded to the cloud-based energy management platform. The multi-source operational status data includes: access request signals, ship battery management system communication data, load demand power, and berthing time window on the ship side; voltage, current, output power, operating mode, and interface temperature on the power pile side; predicted values ​​of new energy power generation and state of charge of the energy storage system on the port side; and real-time electricity price, dispatch instructions, and voltage frequency on the grid side.

[0024] Specifically, the collected multi-source operational status data from the ship-pile-grid sides includes: ship-side access request signals, ship battery management system communication data, ship load demand power curves, and historical charging and power supply records; pile-side AC or DC output voltage, output current, output power, operating mode, charging interface temperature, communication status, protection action records, and internal module status self-check information; port-side photovoltaic and wind power generation forecasts, energy storage system SOC, maximum allowable charging and discharging power of energy storage, and energy storage battery temperature; and grid-side real-time electricity prices, grid dispatch instructions, and public grid voltage and frequency. Edge computing nodes are deployed in industrial-grade edge gateways within the port-side power station. First, outlier removal, missing value imputation, and timestamp alignment are performed on each collected data, and the multi-source data is resampled to a unified time grid. Then, the amplitude, rate of change, fluctuation variance, and main frequency component amplitude time-domain and frequency-domain characteristics of the connected ship load curve are extracted to generate a load feature vector. Finally, the feature vectors and key state data are packaged in a lightweight JSON format and uploaded to the cloud energy management platform via the MQTT protocol through a 4G / 5G private network.

[0025] S3. A cloud-based pre-trained ship load feature classification model identifies the load types of connected ships and, combined with port area renewable energy generation forecasts and grid time-of-use pricing information, establishes a multi-objective collaborative optimization model. The pre-trained ship load feature classification model is built upon a deep convolutional neural network and a long short-term memory network, outputting load type labels including: traditional marine AC loads, electric ship DC fast-charging loads, hybrid ship intermittent charging loads, energy storage ship loads capable of participating in V2G, and emergency rapid power supply loads. The cloud-based energy management platform receives feature vectors and state data uploaded from the edge and first runs the pre-trained ship load feature classification model, such as... Figure 4 As shown, the load type of the connected vessel is identified.

[0026] In this embodiment, the cloud-based ship load feature classification model is based on a hybrid architecture of deep convolutional neural network and long short-term memory network. The CNN layer is used to extract local features of load curve segments, and the LSTM layer is used to capture the temporal dependencies of load changes. The model input consists of load time-series curve segments before the ship connects and ship data reported when the ship first connects. The output consists of five load type labels, including Class I traditional marine AC load, Class II electric ship DC fast charging load, Class III hybrid ship intermittent charging load, Class IV energy storage ship load that can participate in V2G, and Class V emergency fast power supply load.

[0027] The objective function of the multi-objective collaborative optimization model is: ; in, For the overall power supply cost, Penalties for abandoned electricity from renewable energy sources The peak-valley difference penalty for the power grid is represented by w1, w2, and w3, which are the corresponding weight coefficients. The constraints of the multi-objective collaborative optimization model include power balance constraints, output power constraints of each power supply pile, ship charging demand constraints, port energy storage system operation constraints, and power grid interaction constraints; the power balance constraint is the instantaneous power balance between the power grid, new energy sources, energy storage, and power supply piles.

[0028] Specifically, the three sub-objectives mentioned above are defined as follows: Overall power supply cost: The total cost of purchasing electricity from the public grid for the charging pile group, including the electricity cost under time-of-use pricing: ; in, Let t be the time-of-use electricity price of the power grid. Let t be the total power absorbed by the pile group from the power grid at time t. This is the time interval between two adjacent optimized scheduling moments, i.e., the scheduling step size of the optimization model; in this embodiment, The timeframe can be set to 15 minutes, or it can be set to 5 minutes, 10 minutes, or 30 minutes depending on the scheduling cycle of the port area energy management platform; NT is the total number of discrete scheduling periods within the optimization cycle.

[0029] Penalty for Curtailed Renewable Energy: Curtailed solar and wind power caused by insufficient power supply demand in the port area, energy storage charging capacity, or grid interaction power limitations to absorb all renewable energy generation is expressed as follows: in, The optimization period includes a penalty for abandoned renewable energy; NT represents the total number of discrete scheduling periods within the optimization period; and t represents the discrete scheduling time number. Let be the amount of photovoltaic power curtailment at time t; Let be the amount of wind power curtailment at time t; This is the time interval between two adjacent optimized scheduling moments.

[0030] The curtailment power of photovoltaic power and the curtailment power of wind power respectively meet the following requirements: ; ; in, Let t be the predicted power output of the photovoltaic system. The photovoltaic power actually consumed by the port area power supply pile load, the port area energy storage system, or the internal power consumption of the port area at time t; Let t be the predicted generating capacity of the wind power system. The wind power actually consumed by the port area power supply pile load, port area energy storage system or internal power consumption at time t; This is used to ensure that the abandoned power is a non-negative value.

[0031] Peak-valley difference penalty: This term characterizes the peak-valley fluctuation amplitude of the power interaction between the port area and the public power grid during the optimization period. It aims to smooth the load curve of the charging pile group and reduce the impact of centralized charging and power supply behavior on the public power grid. Its expression is as follows: ; in, To optimize the peak-valley difference penalty term in the power grid within the cycle; t represents the interaction power between the port area's pile group and the public power grid at time t; t is the discrete scheduling time number; NT is the total number of discrete scheduling periods within the optimization period. This represents the maximum value of grid interaction power within the optimization period; This represents the minimum power interaction power within the optimization period. This is achieved by minimizing... This can reduce the peak-valley fluctuations of the port area's power supply pile group on the public power grid during the optimization period.

[0032] The three weighting coefficients w1, w2, and w3 are set by the port operator according to actual needs, with typical values ​​of w1=0.4, w2=0.35, and w3=0.25. Among them, w2 can be increased when the policy requires priority to consume new energy; w3 can be increased when a grid demand response event is triggered.

[0033] The operational constraints of the port area energy storage system include: maintaining the state of charge (SOC) between preset upper and lower limits, ensuring that the charging and discharging power does not exceed the maximum allowable value, and ensuring that the SOC satisfies the energy conservation recursive relationship. The constraints include: Power balance constraint: This constraint characterizes the power supply and demand balance between the port area power supply pile group, new energy power generation, port area energy storage system, and public power grid at each discrete scheduling moment. Its expression is: ; in, Number the discrete scheduling time; This refers to the total number of smart power supply piles in the port's smart power supply pile cluster. Number the smart power supply piles; for The total power exchanged between the port area and the public power grid at any given time is positive, with the power purchased by the port area from the public power grid being positive. for The actual power generated by photovoltaic power at any given moment and absorbed by the port area; for The actual power generated by wind power and absorbed by the port area at any given moment; for Discharge power of the energy storage system in the port area at any given time; for Time of the first The power output of each smart power supply pile to the connected ship; for The charging power of the energy storage system in the port area at any given time. All the above power parameters use a unified power unit and can represent the instantaneous power at the corresponding scheduling time or the average power within the scheduling step.

[0034] The actual absorption capacity of photovoltaic and wind power should also meet the following requirements respectively: ; ; in for The predicted power output of the photovoltaic system at any given time. for The predicted generating capacity of the wind power system at any given time. The above constraints are used to ensure that the actual power absorbed by the renewable energy does not exceed the predicted generating capacity at the corresponding time, and together with the renewable energy curtailment penalty, constrain the utilization process of renewable energy in the port area.

[0035] Output power constraints for each power supply pile: For the i-th power supply pile, its output power is limited by its rated power. And the load type restrictions for currently connected vessels: ; in, The power supply mode coefficients are as follows: AC mode is 1.0, DC fast charging mode is 1.1, and short-term overload of 10% is allowed; V2G mode is 1.0, and emergency mode is 1.2.

[0036] Ship charging demand constraints: For electric or hybrid vessels joining the network, their charging needs must be met within their valid berthing service window. The valid berthing service window is denoted as... , which is the first The arrival declaration information, berth scheduling information, AIS vessel identification information, confirmation time of power supply pile connection, and / or vessel departure plan for each vessel are determined. For the first Discrete scheduling time for ships to arrive at berth and be allowed to begin charging and power supply services. For the first The discrete scheduling times for a vessel's planned departure or request for charging / powering services together define the time range within which the vessel can receive charging / powering services. For vessels with uncertain departure times, The estimated departure time can be adjusted in real time based on the berth management system.

[0037] Within the aforementioned port call time window, the first The cumulative charging amount of a vessel should not be less than its required charging amount, subject to the following constraint: ; in, for Time of the first The charging power output of each power supply pile to the corresponding ship. The time interval between two adjacent optimized scheduling moments. For the first The total amount of charging required by the vessel during this port call. It can be obtained directly from the power demand reported by the ship, or it can be calculated based on the ship's battery rated capacity, state of charge at the time of connection, and target state of charge. The expression is as follows: ; in, For the first The target state of charge that the ship is expected to achieve. For the first The initial state of charge of a ship when it is connected to a power supply pile. For the first The rated battery capacity of the ship, For the first The equivalent charging efficiency during the charging process of a ship. When a ship directly reports its power demand to an edge computing node or a cloud energy management platform. The value reported by the ship can be used first; if no electricity demand is reported, the above formula should be used to calculate it.

[0038] Port area energy storage system operation constraints: These constraints limit the state of charge and charge / discharge power of the port area energy storage system within the optimization cycle, ensuring that the energy storage system operates within a safe range and satisfies the energy conservation recursive relationship. The constraints include: ; ; ; ; in, The state of charge of the port area energy storage system at time t; The state of charge at the next discrete scheduling time; This is the lower limit of the permissible state of charge for the energy storage system. This represents the upper limit of the allowed state of charge for the energy storage system. In this embodiment, 0.2 is acceptable. A value of 0.9 is acceptable. This range is used to avoid over-discharging or over-charging of energy storage batteries. The specific value can be adjusted according to the type of energy storage battery, the manufacturer's safety limits, and the port's operation strategy.

[0039] Let t be the charging power of the port area energy storage system. Let be the discharge power of the port area energy storage system at time t; The maximum charging power allowed by the energy storage system. The maximum allowable discharge power of the energy storage system can be determined by the rated power of the energy storage converter, the allowable power of the battery management system, and the safety margin of port operation. This refers to the rated energy storage capacity of the port area's energy storage system. To improve the charging efficiency of energy storage systems. The discharge efficiency of the energy storage system; t represents the time interval between two adjacent optimal scheduling times; t is the discrete scheduling time number.

[0040] In this embodiment, and All values ​​can be taken as 0.95, or can be adjusted based on the efficiency of the energy storage converter, the battery charge / discharge efficiency, and actual operational test results. The above recursive relationship of state of charge indicates that within each scheduling step, the state of charge of the energy storage system is jointly determined by the state of charge at the previous moment, the charging energy, the discharging energy, and the charge / discharge efficiency.

[0041] Power grid interaction constraint: This is used to limit the power exchange amplitude at the common connection point between the port area and the public power grid, to avoid the impact of centralized charging and power supply by the charging pile cluster or V2G feedback on the public power grid exceeding the contracted capacity or dispatch allowable range. Its expression is: ; in, Let t be the power exchange between the port area and the public power grid. When the port area purchases electricity from the public power grid, it takes a positive value. When the port area feeds back electricity to the public power grid through the energy storage system or V2G ships, it can take a negative value. This refers to the maximum permissible interaction power between the port area and the public power grid at the point of common connection. Its value is determined by the power grid company's approval for connection, the power supply contract capacity, the transformer capacity, the distribution line capacity, or dispatch instructions. The above two-sided inequality is equivalent to... This constraint is used to simultaneously limit the power output of the port area in both directions: purchasing electricity from the grid and feeding electricity back to the grid. If the port area's operation strategy only allows purchasing electricity from the public grid and does not allow reverse power feeding, then the above constraint can be equivalently set as follows: .

[0042] S4. Solve the multi-objective collaborative optimization model to obtain the power allocation scheme, power supply mode switching strategy, and charging and discharging strategy of each smart power supply pile in the pile group. The operating constraints of the port area energy storage system include: maintaining the state of charge between preset upper and lower limits, the charging and discharging power not exceeding the maximum allowable value, and the state of charge satisfying the energy conservation recursive relationship.

[0043] S5. Edge computing nodes receive optimization strategies from the cloud and adjust these strategies locally based on real-time operating status, modifying the output power, power supply mode, and energy storage converter control parameters of each smart power supply pile. The local feedback adjustments performed by the edge computing nodes include: triggering emergency protection strategies when an anomaly is detected; activating local offline optimization mode and autonomously generating a power allocation scheme using rolling time-domain optimization when communication latency exceeds a threshold; and periodically synchronizing the correction results to the cloud to update model parameters.

[0044] Edge computing nodes execute real-time closed-loop control: In normal mode, when the communication latency is less than 200ms, the edge computing node directly executes the optimization strategy issued by the cloud, collects real-time operating status data and calculates the deviation at a 1-second interval. If the power deviation is less than 5% and the SOC deviation is less than 2%, the original strategy is maintained; otherwise, the power command of the power supply pile is slightly modified. In communication latency mode, i.e., when the latency is greater than or equal to 200ms, the edge computing node enables local offline optimization mode. Based on the most recently received global optimization strategy and the current local real-time status, it uses a rolling time-domain optimization method to autonomously generate a power allocation scheme with the goal of minimizing short-term power supply costs and maximizing the consumption of new energy sources. The rolling optimization time domain is 1 hour and the step size is 5 minutes. In emergency mode, when voltage and frequency exceedances, abnormal power supply pile outputs, or protection actions are detected, the edge computing node triggers emergency protection strategies locally, including limiting power supply, switching to backup power supply mode, or triggering a safety shutdown, and reports the event to the cloud. Every 5 minutes, the edge computing node synchronizes the locally corrected execution results and operating data to the cloud. The cloud updates the optimization model parameters and ship load characteristic classification model accordingly, forming a continuous learning closed loop of cloud-edge collaboration.

[0045] S6. Power supply piles provide services. Each smart power supply pile provides charging and power supply services to the connected ships according to the received power instructions and power supply mode. At the same time, the port area energy storage system performs energy throughput according to the charging and discharging strategy.

[0046] Power supply mode switching strategies include: locking traditional AC loads to AC shore power mode; prioritizing DC fast charging loads and dynamically adjusting power; increasing charging power for intermittent charging loads during off-peak periods; allowing V2G vessels to feed back electricity during peak electricity prices or power shortages; and rapidly increasing emergency loads to their rated maximum power.

[0047] In this embodiment, multi-type ship compatibility and flexible power supply are achieved. By integrating AC shore power modules, DC fast charging modules, and bidirectional charging and discharging modules into each smart power pile, it also achieves compatibility with various power supply needs such as traditional AC ships, DC fast charging for electric ships, intermittent charging for hybrid ships, and V2G for energy storage ships. This solves the problem of the single system and poor compatibility of existing port power supply facilities, and significantly improves the utilization rate and applicability of port power supply facilities. Regarding power supply cost optimization and efficient renewable energy consumption, a multi-objective collaborative optimization model is established in the cloud based on ship load characteristic classification models and renewable energy generation predictions. The model aims to minimize overall power supply costs, maximize renewable energy consumption rates, and minimize grid peak-valley differences, achieving globally optimal allocation of power in the pile group. This solution can effectively reduce grid power purchase costs, increase the self-consumption ratio of renewable energy such as photovoltaic and wind power in the port area, and avoid energy waste.

[0048] Example 2 like Figure 5 , Figure 6 and Figure 7 As shown, this invention also proposes a port intelligent power supply pile group cloud-edge collaborative energy management system, using a port intelligent power supply pile group cloud-edge collaborative energy management method as described in Example 1, including the following: The physical layer of the pile group includes multiple smart power supply piles and a port area energy storage system. The smart power supply piles integrate AC shore power modules, DC fast charging modules and bidirectional charging and discharging modules. The edge computing node layer is configured to collect multi-source operational status data from three sides: ship, pile, and network; perform data preprocessing and feature extraction; receive cloud optimization strategies and execute local feedback adjustments; and send control commands to each smart power supply pile. The cloud-based energy management platform is configured to operate a ship load characteristic classification model, a new energy power generation prediction model, and a multi-objective collaborative optimization model. It generates a pile group power allocation scheme and an energy storage charging and discharging strategy, and distributes them to edge computing nodes.

[0049] In this embodiment, the physical layer of the pile group includes multiple smart power supply piles and a port area energy storage system. Each smart power supply pile integrates three core modules: an AC shore power module that achieves power system conversion through a cascaded H-bridge topology power electronic transformer and solid-state switch, supporting multiple standard marine power systems, with a rated output capacity of 1MVA to 5MVA, where MVA represents the rated apparent power capacity of the AC shore power module. A DC fast charging module based on a modular multilevel converter topology, supporting a wide voltage DC output from 400V to 1000V, with a single module of 250kW, expandable to 500kW to 2MW through parallel connection, where kW or MW represents the active power output of the DC fast charging module. The DC fast charging module is also compatible with CCS, CHAdeMO, or ship-specific DC charging communication protocols for information interaction with the ship's battery management system. A bidirectional charge / discharge module shares the power conversion circuit with the DC fast charging module, adding a bidirectional DC / DC converter and V2G protocol stack, with a rated power of 500kW and a peak efficiency of not less than 95%, supporting the ship's battery to feed power back to the port area power grid. The three modules are interconnected via an internal DC bus and a switching switch, with the edge computing node selecting the activation mode based on cloud policies and real-time operating conditions. The port area energy storage system is a 1MW / 2MWh lithium iron phosphate system, where 1MW represents the rated charge / discharge power of the energy storage system and 2MWh represents the rated energy storage capacity. The port area new energy system can be configured with a 2MWp photovoltaic power generation system and a 1.5MW wind power system, where MWp represents the peak installed capacity of the photovoltaic system and MW represents the rated installed capacity of the wind power system. The above capacity and power parameters are exemplary engineering configurations for this embodiment and can be adjusted according to berth grade, type of vessel access, port area transformer capacity, and new energy installed capacity.

[0050] The edge computing node layer is deployed in an industrial-grade edge gateway, connected to each power supply pile via gigabit industrial Ethernet or fiber optic ring network. It collects multi-source data from three sides: ship, pile, and network. This includes ship access requests, BMS communication data, load power, berthing window, power pile voltage, current, power, operating mode, temperature, communication and protection status, port area renewable energy generation forecasts, energy storage SOC and charge / discharge limits, as well as grid electricity prices, dispatch instructions, voltage and frequency. The edge nodes have data preprocessing and feature extraction capabilities, extracting load amplitude, rate of change, fluctuation variance, and frequency domain features, packaging them in JSON format, and uploading them to the cloud via the MQTT protocol. Local feedback adjustments include: triggering emergency protection and reporting to the cloud in case of anomalies; activating local offline optimization when communication latency exceeds 200ms; and adopting a rolling time-domain optimization self-generated scheme with a 1-hour time domain and a 5-minute step size, synchronizing the correction results to the cloud every 5 minutes to update model parameters.

[0051] The cloud-based energy management platform is deployed in the data center, running a ship load characteristic classification model, a new energy power generation prediction model, and a multi-objective collaborative optimization model. The ship load characteristic classification model is based on a CNN-LSTM hybrid architecture, taking load segments from the previous 30 minutes and historical data as input, and outputting five load labels. It is trained offline with no fewer than 2000 samples, using cross-entropy as the loss function and Adam as the optimizer. In the new energy power generation prediction, photovoltaic power outputs predicted values ​​every 15 minutes for the next 24 hours based on numerical weather forecasts and historical data, while wind power outputs predicted values ​​at the same time resolution based on WRF wind speed prediction, power curves, and wake effect models. The multi-objective collaborative optimization model aims to minimize power supply costs, maximize new energy absorption, and minimize peak-valley differences. Constraints cover power balance, power pile power, ship charging demand, energy storage operation, and grid interaction. Global optimization is performed every 15 minutes in the cloud, using an improved NSGA-II algorithm to obtain the Pareto front. After selecting the optimal compromise solution using the TOPSIS method, the solution is distributed to edge nodes. The optimization results include power commands for each power pile, power supply mode sequences, energy storage charging and discharging commands, and the total purchased power curve. The system achieves compatible access for multiple types of vessels and optimized energy management across the entire port area through a three-layer cloud-edge collaborative architecture.

[0052] By leveraging cloud-edge collaboration to ensure real-time reliability and grid-friendliness, a collaborative architecture has been constructed, integrating global optimization in the cloud and local correction at the edge: the cloud is responsible for generating long-term strategies, while edge computing nodes handle real-time data acquisition and feedback adjustments. In the event of communication delays or interruptions, edge nodes can autonomously activate local offline optimization modes to ensure power supply continuity and security. Simultaneously, by optimizing the peak-valley difference in the power grid and utilizing V2G functionality to assist in peak shaving, the level of friendly interaction between the port and the power grid has been significantly improved.

[0053] Example 3 An electronic device, comprising: Processor and memory; The processor executes the steps of the port smart power supply pile group cloud-edge collaborative energy management method as described in any of Embodiment 1 by calling programs or instructions stored in memory.

[0054] Example 4 A computer-readable storage medium includes computer program instructions that cause a computer to perform steps of cloud-edge collaborative energy management for a port smart power supply group as described in any of Embodiment 1.

[0055] Computer-readable storage media may take the form of any combination of one or more readable media. A readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may, for example, include, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatuses, or devices, or any combination thereof. More specific examples of readable storage media (a non-exhaustive list) include: electrical connections having one or more wires, portable disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0056] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the technical solutions of the embodiments of the present invention.

Claims

1. A cloud-edge collaborative energy management method for intelligent power supply pile groups in ports, characterized in that, Includes the following steps: S1. Construct a smart power supply pile group for the port. The pile group includes multiple smart power supply piles and at least one edge computing node. Each smart power supply pile integrates an AC shore power module, a DC fast charging module, and a bidirectional charging and discharging module. The edge computing node enables pile group-level data aggregation and local control. S2. Collect multi-source operational status data from the ship, pile, and net sides, and upload the data to the cloud energy management platform after data preprocessing and feature extraction by the edge computing node. S3. The cloud-based pre-trained ship load characteristic classification model identifies the load type of connected ships and, in conjunction with port area new energy power generation forecasts and grid time-of-use electricity price information, establishes a multi-objective collaborative optimization model. S4. Solve the multi-objective collaborative optimization model to obtain the power allocation scheme, power supply mode switching strategy and charging and discharging strategy of each smart power supply pile in the pile group. S5. The edge computing node receives the optimization strategy sent from the cloud and adjusts the strategy locally based on the real-time operating status, adjusting the output power, power supply mode and energy storage converter control parameters of each smart power supply pile. S6. Power supply piles provide services. Each smart power supply pile provides charging and power supply services to the connected ships according to the received power instructions and power supply mode. At the same time, the port area energy storage system performs energy throughput according to the charging and discharging strategy.

2. The port intelligent power supply pile group cloud-edge collaborative energy management method according to claim 1, characterized in that, The AC shore power module of the intelligent power supply pile achieves power system conversion through a power electronic transformer and a solid-state switch, supporting standard marine power systems of 6.6kV / 60Hz, 440V / 60Hz, and 400V / 50Hz; the DC fast charging module supports DC output with a wide voltage range of 400V to 1000V, and the output power level is 500kW to 2MW; the bidirectional charging and discharging module supports V2G function and is configured to receive energy fed back to the port power grid by energy storage electric ships.

3. The port intelligent power supply pile group cloud-edge collaborative energy management method according to claim 1, characterized in that, The multi-source operational status data includes: access request signals from the ship side, communication data of the ship's battery management system, load demand power, and berthing time window; voltage, current, output power, operating mode, and interface temperature from the power supply pile side; predicted values ​​of new energy power generation and state of charge of the energy storage system from the port area side; and real-time electricity price, dispatch instructions, and voltage frequency from the grid side.

4. The port intelligent power supply pile group cloud-edge collaborative energy management method according to claim 3, characterized in that, The pre-trained ship load feature classification model is constructed based on a deep convolutional neural network and a long short-term memory network. The output load type labels include: traditional marine AC load, electric ship DC fast charging load, hybrid ship intermittent charging load, energy storage ship load that can participate in V2G, and emergency rapid power supply load.

5. The port intelligent power supply pile group cloud-edge collaborative energy management method according to claim 1, characterized in that, The objective function of the multi-objective collaborative optimization model is: ； in, For the overall power supply cost, Penalties for abandoned electricity from renewable energy sources The peak-valley difference penalty for the power grid is represented by w1, w2, and w3, which are the corresponding weighting coefficients.

6. The port intelligent power supply pile group cloud-edge collaborative energy management method according to claim 5, characterized in that: The constraints of the multi-objective collaborative optimization model include power balance constraints, output power constraints of each power supply pile, ship charging demand constraints, port area energy storage system operation constraints, and power grid interaction constraints; the power balance constraints are the instantaneous power balance between the power grid, new energy sources, energy storage, and power supply piles.

7. The port intelligent power supply pile group cloud-edge collaborative energy management method according to claim 6, characterized in that, The operational constraints of the port area energy storage system include: maintaining the state of charge between preset upper and lower limits, ensuring that the charging and discharging power does not exceed the maximum allowable value, and ensuring that the state of charge satisfies the energy conservation recursive relationship.

8. The port intelligent power supply pile group cloud-edge collaborative energy management method according to claim 1, characterized in that, The local feedback adjustment performed by the edge computing node includes: triggering an emergency protection strategy when an anomaly is detected; enabling a local offline optimization mode when the communication latency exceeds a threshold, using rolling time-domain optimization to autonomously generate a power allocation scheme; and periodically synchronizing the correction results to the cloud to update model parameters.

9. The port intelligent power supply pile group cloud-edge collaborative energy management method according to claim 4, characterized in that, The power supply mode switching strategies include: locking traditional AC loads to AC shore power mode; prioritizing DC fast charging loads and dynamically adjusting power; increasing charging power for intermittent charging loads during off-peak periods; feeding back energy from V2G vessels during peak electricity prices or power shortages; and rapidly increasing emergency loads to their rated maximum power.

10. A cloud-edge collaborative energy management system for a port intelligent power supply pile group, characterized in that, The system is implemented based on the method of any one of claims 1 to 9, comprising: The physical layer of the pile group includes multiple smart power supply piles and a port area energy storage system. The smart power supply piles integrate AC shore power modules, DC fast charging modules and bidirectional charging and discharging modules. The edge computing node layer is configured to collect multi-source operational status data from three sides: ship, pile, and network; perform data preprocessing and feature extraction; receive cloud optimization strategies and execute local feedback adjustments; and send control commands to each smart power supply pile. The cloud-based energy management platform is configured to operate a ship load characteristic classification model, a new energy power generation prediction model, and a multi-objective collaborative optimization model. It generates a pile group power allocation scheme and an energy storage charging and discharging strategy, and distributes them to edge computing nodes.

Images