Wharf multi-source energy collaborative management system and method based on digital twinning

Abstract

The present application relates to the field of wharf energy management, and discloses a wharf multi-source energy collaborative management system and method based on digital twinning, comprising: a business state mapping module for obtaining logistics operation plan data; a load prediction conversion module for mapping the operation plan into power load prediction data; a commercial parameter acquisition module for determining energy storage depreciation cost and time-of-use electricity price; a resource scheduling optimization module for determining power distribution instructions in combination with prediction data, electricity price and cost; and a residual error compensation correction module for correcting instructions according to load residual error. The present application feeds forward logistics instructions to the energy system, so that the supply matrix adjusts the output in advance according to the loading and unloading plan, eliminates scheduling response lag, and optimizes electricity price and asset loss, thereby maintaining system stability and improving comprehensive economic benefits.

Description

Technical Field

[0001] This invention relates to a digital twin-based multi-source energy collaborative management and control system and method for wharves, belonging to the field of wharf energy management technology. Background Technology

[0002] Currently, in the operation and management system of large ports, integrating renewable energy and energy storage facilities is a common way to optimize energy purchase costs. The current control system mainly relies on historical power load data for statistical modeling and uses the electrical data collected from physical nodes to provide feedback and adjust the energy output status. The energy consumption characteristics of port loading and unloading operations are driven by the business plan of the terminal operating system. The operation logic of large, high-energy-consuming equipment such as quay cranes and yard cranes depends entirely on the ship berthing sequence and container transfer instructions. The existing control method usually treats physical power load as an independent random variable. Due to the lack of logical connection between the energy dispatch side and the business administration side, the system has difficulty predicting the load jump caused by sudden loading and unloading tasks, resulting in the adjustment action lagging behind the time of occurrence of objective physical energy consumption, thus generating high demand rate expenditures.

[0003] Besides the physical limitations of hardware facilities, scheduling methods at the control level also have optimization blind spots. For example, Chinese invention patent application CN120106321A discloses a port energy dynamic optimization method and system based on digital twins. Although it improves computational efficiency through physical information neural networks and quantum-inspired optimization algorithms, its core logic still focuses on the correlation between grid-side parameters and environmental carbon signals. Because this technical solution fails to penetrate to the lowest level of production instructions at the terminal, it cannot establish a real-time atomized mapping between discrete container movement, load changes, and energy consumption. As a result, when dealing with high-frequency, highly random loading and unloading operation pulses, the prediction model still has a time lag in capturing physical load fluctuations, making it difficult to accurately suppress energy demand from the source of production. Even by simply increasing energy storage capacity or introducing high-order mathematical prediction models, it is difficult to eliminate the information asymmetry between business decision flow and physical load flow. Blindly expanding hardware facilities will result in high construction costs and equipment maintenance pressure. In addition, load forecasting without the support of logistics operation logic often leads to ambiguity in the calculation results when facing dynamic changes in loading and unloading plans or simultaneous operations of multiple ships.

[0004] Therefore, the technical problem to be solved by this invention is how to reconstruct the mapping link between port administrative logistics instructions and physical energy consumption demand, and to build a collaborative mechanism that can deeply link business timing and energy supply in order to improve the predictability of the control system for dynamic loading and unloading tasks. Summary of the Invention

[0005] To address the information asymmetry between port administrative logistics instructions and physical energy dispatching, as mentioned in the background art, and the resulting delays in dispatching response and deterioration in operating costs, the technical solution of this invention is as follows: A port multi-source energy collaborative management and control system based on digital twins, comprising:

[0006] The business status mapping module is used to obtain logistics operation plan data from the terminal operating system. The logistics operation plan data includes container position movement instructions with timestamps, container weight data, yard coordinate data, and corresponding operation equipment numbers.

[0007] The load forecasting and conversion module, connected to the business status mapping module, is used to convert logistics operation plan data into energy efficiency models for each operation link according to the equipment power mapping rules built into the digital twin platform, and calculate the power loss coefficient according to the displacement distance of the operation equipment number under the logistics operation plan data, so as to generate power load forecast data corresponding to the preset time window.

[0008] The commercial parameter acquisition module is used to acquire time-of-use electricity price data of the power system and the cumulative charge-discharge cycle number of each energy storage unit. Based on the electrochemical capacity decay slope preset in the battery aging model, the current unit depreciation cost of the energy storage unit is determined.

[0009] The resource scheduling optimization module is connected to the load forecasting conversion module and the business parameter acquisition module, respectively. It is used to take the electricity load forecasting data as the feedforward input, combine it with the time-of-use electricity price data and the unit depreciation cost, and calculate and generate the power allocation instructions corresponding to the multi-source energy system with the goal of minimizing the total operating cost.

[0010] The residual compensation correction module is used to monitor the real-time load power of the grid connection point, compare the real-time load power with the predicted load power generated synchronously by the digital twin platform to generate a residual value, and adjust the amplitude of the power distribution command based on the product of the residual value and the preset feedback gain.

[0011] Preferably, when generating electricity load forecast data, the load forecasting conversion module determines the nonlinear frictional resistance of the operating equipment during the displacement process based on the container weight data and the yard coordinate data, and adds it to the power loss coefficient.

[0012] Preferably, the equipment power mapping rules include power consumption benchmark data of quay cranes, yard cranes, and container trucks under different loads; the load prediction conversion module obtains the instantaneous power curve of the operating equipment within a preset time window by simulating the movement trajectory of the operating equipment under the logistics operation plan data in the digital twin platform.

[0013] Preferably, the commercial parameter acquisition module obtains the electrochemical impedance data of each energy storage unit and calculates the real-time correction value of the capacity decay slope to update the unit depreciation cost.

[0014] Preferably, the power allocation instruction includes a power-limiting component of the distributed photovoltaic system, a charging and discharging scheduling component of the energy storage unit, and a power supply regulation component of the port shore power system; the resource scheduling optimization module balances commercial revenue by establishing a multi-objective constraint set while maintaining the frequency fluctuation of the multi-source energy system within the range of 0.2Hz.

[0015] Preferably, the residual compensation correction module also includes a residual feature analysis unit, which is used to analyze the deviation characteristics between real-time load power and predicted load power in the frequency domain; when the deviation characteristics exceed the preset burst load threshold, the residual compensation correction module increases the power limit of the energy storage unit by 20% to 30%.

[0016] Preferably, when correcting power allocation instructions, the residual compensation correction module superimposes compensation power based on the first-order differential term of the load according to the progress deviation of the logistics operation plan data, so as to make the expected trajectory of the digital twin platform converge with the actual measured operation trajectory of the physical site.

[0017] Preferably, the system also includes a data governance module, which is used to normalize the heterogeneous data from the terminal operating system, multi-source energy system and power grid sensors, and establish a unified data bus with timestamp alignment function.

[0018] Preferably, before outputting the power allocation command, the resource scheduling optimization module calculates the power distribution network security verification result based on the simulated power topology in the digital twin platform. When the verification result shows that the node voltage offset is less than 5%, the power allocation command is released to the multi-source energy system.

[0019] A digital twin-based method for collaborative management and control of multi-source energy at a port, used to implement a digital twin-based collaborative management and control system for multi-source energy at a port, includes the following steps:

[0020] Step S1001: Obtain logistics operation plan data from the terminal terminal operating system. The logistics operation plan data includes container position movement instructions with timestamps, container weight data, yard coordinate data, and corresponding operation equipment numbers.

[0021] Step S1002: Based on the preset equipment power mapping rules in the digital twin platform, the logistics operation plan data is converted into the energy efficiency model of each operation link, and the power loss coefficient is calculated according to the displacement distance of the operation equipment number under the logistics operation plan data, so as to generate the power load prediction data corresponding to the preset time window.

[0022] Step S1003: Obtain the time-of-use electricity price data of the power system and the cumulative charge-discharge cycle number of each energy storage unit, and determine the current unit depreciation cost of the energy storage unit based on the electrochemical capacity decay slope preset in the battery aging model.

[0023] Step S1004: Using the electricity load forecast data as the feedforward input, combined with the time-of-use electricity price data and the unit depreciation cost, the power allocation instruction corresponding to the multi-source energy system is calculated and generated with the goal of minimizing the total operating cost.

[0024] Step S1005: Monitor the real-time load power of the grid connection point, compare the real-time load power with the predicted load power generated synchronously by the digital twin platform to generate a residual value, and adjust the amplitude of the power distribution command based on the product of the residual value and the preset feedback gain.

[0025] Compared with the prior art, the beneficial effects of the present invention are:

[0026] 1. In the collaborative management and control of multi-source energy at the port, by parsing the logistics instructions of the port terminal operating system into discrete action atomic clusters and converting them into a directed load demand map containing time-series variables and spatial node variables in the digital space, the time-series coupling between the business flow in the management dimension and the energy flow in the physical dimension is realized. This feedforward data processing mechanism based on administrative instructions enables the multi-source energy supply matrix to adjust the output boundary in advance according to the loading and unloading operation plan, thereby avoiding the inherent response lag problem of traditional scheduling based on data collection threshold triggering.

[0027] 2. By incorporating the time-of-use electricity rate and the micro-aging conversion parameters of the energy storage unit into the objective function of multi-step rolling optimization, an evaluation benchmark that deeply integrates commercial operating costs and physical asset depreciation is constructed. During the scheduling calculation process, the system constrains the discrete change rate of energy storage power, enabling the energy throughput path to automatically adapt to the nonlinear electrochemical characteristics of the battery while maximizing economic benefits. This prevents overcharging and discharging of energy storage devices under frequent business pulse drives, ensuring the comprehensive economic benefits of the system throughout its entire life cycle.

[0028] 3. By synchronously monitoring the actual load scalar at the physical power grid junction point and performing dynamic residual calculations between it and the predicted load in the digital twin space, a control closed loop with adaptive compensation capability is constructed. When the work progress defined by the administrative order deviates from the actual loading and unloading trajectory at the physical site in time and space, the system suppresses the execution weight of the reference power order by triggering the attenuation factor and superimposes the compensation power based on the load differential term to achieve real-time convergence between the digital model expectation and the actual situation at the physical site, ensuring the robustness of the management and dispatch system under fluctuating operating conditions. Attached Figure Description

[0029] Figure 1This is a flowchart illustrating the multi-source energy collaborative management logic of the logistics instruction feedforward of the present invention.

[0030] Figure 2 This is a diagram of the multi-dimensional state evolution and residual compensation closed-loop architecture of the collaborative control system of this invention.

[0031] The objectives, features, and advantages of this invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0032] The technical solutions of the embodiments of this application will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.

[0033] A digital twin-based multi-source energy collaborative management and control system for a port includes:

[0034] The business status mapping module is used to obtain logistics operation plan data from the terminal operating system. The logistics operation plan data includes container position movement instructions with timestamps, container weight data, yard coordinate data, and corresponding operation equipment numbers.

[0035] The load forecasting and conversion module, connected to the business status mapping module, is used to convert logistics operation plan data into energy efficiency models for each operation link according to the equipment power mapping rules built into the digital twin platform, and calculate the power loss coefficient according to the displacement distance of the operation equipment number under the logistics operation plan data, so as to generate power load forecast data corresponding to the preset time window.

[0036] The commercial parameter acquisition module is used to acquire time-of-use electricity price data of the power system and the cumulative charge-discharge cycle number of each energy storage unit. Based on the electrochemical capacity decay slope preset in the battery aging model, the current unit depreciation cost of the energy storage unit is determined.

[0037] The resource scheduling optimization module is connected to the load forecasting conversion module and the business parameter acquisition module, respectively. It is used to take the electricity load forecasting data as the feedforward input, combine it with the time-of-use electricity price data and the unit depreciation cost, and calculate and generate the power allocation instructions corresponding to the multi-source energy system with the goal of minimizing the total operating cost.

[0038] The residual compensation correction module is used to monitor the real-time load power of the grid connection point, compare the real-time load power with the predicted load power generated synchronously by the digital twin platform to generate a residual value, and adjust the amplitude of the power distribution command based on the product of the residual value and the preset feedback gain.

[0039] Preferably, when generating electricity load forecast data, the load forecasting conversion module determines the nonlinear frictional resistance of the operating equipment during the displacement process based on the container weight data and the yard coordinate data, and adds it to the power loss coefficient.

[0040] Preferably, the equipment power mapping rules include power consumption benchmark data of quay cranes, yard cranes, and container trucks under different loads; the load prediction conversion module obtains the instantaneous power curve of the operating equipment within a preset time window by simulating the movement trajectory of the operating equipment under the logistics operation plan data in the digital twin platform.

[0041] Preferably, the commercial parameter acquisition module obtains the electrochemical impedance data of each energy storage unit and calculates the real-time correction value of the capacity decay slope to update the unit depreciation cost.

[0042] Preferably, the power allocation instruction includes a power-limiting component of the distributed photovoltaic system, a charging and discharging scheduling component of the energy storage unit, and a power supply regulation component of the port shore power system; the resource scheduling optimization module balances commercial revenue by establishing a multi-objective constraint set while maintaining the frequency fluctuation of the multi-source energy system within the range of 0.2Hz.

[0043] Preferably, the residual compensation correction module also includes a residual feature analysis unit, which is used to analyze the deviation characteristics between real-time load power and predicted load power in the frequency domain; when the deviation characteristics exceed the preset burst load threshold, the residual compensation correction module increases the power limit of the energy storage unit by 20% to 30%.

[0044] Preferably, when correcting power allocation instructions, the residual compensation correction module superimposes compensation power based on the first-order differential term of the load according to the progress deviation of the logistics operation plan data, so as to make the expected trajectory of the digital twin platform converge with the actual measured operation trajectory of the physical site.

[0045] Preferably, the system also includes a data governance module, which is used to normalize the heterogeneous data from the terminal operating system, multi-source energy system and power grid sensors, and establish a unified data bus with timestamp alignment function.

[0046] Preferably, before outputting the power allocation command, the resource scheduling optimization module calculates the power distribution network security verification result based on the simulated power topology in the digital twin platform. When the verification result shows that the node voltage offset is less than 5%, the power allocation command is released to the multi-source energy system.

[0047] A digital twin-based method for coordinated management and control of multi-source energy at a port includes the following steps:

[0048] Step S1001: Obtain logistics operation plan data from the terminal terminal operating system. The logistics operation plan data includes container position movement instructions with timestamps, container weight data, yard coordinate data, and corresponding operation equipment numbers.

[0049] Step S1002: Based on the preset equipment power mapping rules in the digital twin platform, the logistics operation plan data is converted into the energy efficiency model of each operation link, and the power loss coefficient is calculated according to the displacement distance of the operation equipment number under the logistics operation plan data, so as to generate the power load prediction data corresponding to the preset time window.

[0050] Step S1003: Obtain the time-of-use electricity price data of the power system and the cumulative charge-discharge cycle number of each energy storage unit, and determine the current unit depreciation cost of the energy storage unit based on the electrochemical capacity decay slope preset in the battery aging model.

[0051] Step S1004: Using the electricity load forecast data as the feedforward input, combined with the time-of-use electricity price data and the unit depreciation cost, the power allocation instruction corresponding to the multi-source energy system is calculated and generated with the goal of minimizing the total operating cost.

[0052] Step S1005: Monitor the real-time load power of the grid connection point, compare the real-time load power with the predicted load power generated synchronously by the digital twin platform to generate a residual value, and adjust the amplitude of the power distribution command based on the product of the residual value and the preset feedback gain.

[0053] Example 1: In a seaport container terminal equipped with 12 quay cranes with a rated power of 800kW and 25 automated guided vehicles (AGVs), the system must handle the random berthing operation flow and the energy network composed of distributed photovoltaic, small wind farms, and multimodal energy storage systems. When multiple 10,000-TEU container ships berth simultaneously and begin parallel loading and unloading operations, the terminal's instantaneous power load exhibits pulse characteristics, with its variability exceeding 5MW within seconds. The control method based on electrical threshold triggering, by separating commercial administrative dispatch instructions from physical energy supply at the information level, causes the energy storage system to only initiate feedback regulation after detecting fluctuations in grid voltage or frequency. This time lag between perception and execution results in frequent demand rate expenditures at the terminal's grid connection point. To address the spatiotemporal information asymmetry between the management and physical dimensions, a business status mapping module connects to the terminal's terminal operating system in real time to obtain logistics operation plan data containing timestamps. This logistics operation plan data includes container... Based on the location movement instructions, container weight data, and yard coordinate data, the load forecasting and conversion module decouples the logistics operation plan data into discrete action atom clusters according to the equipment power mapping rules built into the digital twin platform. This data is then converted into a directed load demand map containing temporal and spatial node variables within the digital space. This approach introduces business data flow from the management dimension, transforming random electrical pulses into deterministic performance consumption scalars that are extrapolated 2 to 4 hours in advance within the digital twin space. The directed load demand map is a topological structure with each operation equipment number in the logistics operation plan as nodes and timestamped container location movement instruction sequences as directed edges. Each directed edge carries an attribute feature vector containing the operation load, estimated displacement energy consumption, and time window span. By searching the topological path along the operation time sequence within the digital space, the system aggregates discrete logistics action atoms into vector chains with energy flow significance, thereby achieving a structured transformation from business logic to physical energy consumption logic.

[0054] After obtaining the directed load demand map, the resource scheduling optimization module simultaneously acquires the time-of-use electricity price data of the power system and the cumulative charge-discharge cycle count of each energy storage unit. The resource scheduling optimization module takes the minimum comprehensive energy purchase cost and the minimum equipment cycle life loss within the management cycle as the constraint objectives, and performs multi-step rolling time series solution. Its objective function for scheduling optimization is as follows: ,in, Total operating costs; This is the current sampling time; The total length of the preset time window; for Time-of-use electricity rates for the power grid at any given time; This is a scalar value of active power obtained by the system from the power grid; The unit depreciation cost is determined for the energy storage unit based on the current cumulative number of charge-discharge cycles and the corresponding electrochemical capacity decay slope. For energy storage units in The rate of change of power output at any given time; This is a peak over-limit penalty term derived from the evolution of the directed load demand pattern; As the weighting coefficient for the penalty term, in the actual calculation process, the resource scheduling optimization module uses the mixed integer linear programming (MILP) algorithm to perform a time-series rolling solution to the above objective function. The system uses a preset time window. The process is discretized into several decision steps, with the state of charge (SOC) and power output of the energy storage unit within each step serving as state variables. A linear constraint space is constructed, incorporating power balance constraints, energy storage charge / discharge depth constraints, and grid demand quota constraints. The branch-and-bound method is then used to search for the factor that minimizes the total operating cost. The optimal decision sequence that obtains the minimum value is finally output as the power allocation instruction matrix corresponding to each time step. This formula couples business management variables with physical constraints and calculates the equipment reference power distribution matrix corresponding to each time node.

[0055] Because there may be deviations between the digital twin prediction model and the actual measured data at the industrial site, the residual compensation correction module synchronously monitors the actual load scalar at the physical power grid junction point and calculates the dynamic residual between this scalar and the predicted load of the corresponding node in the digital twin space. When a process interruption occurs at the work site due to weather or equipment failure, and the dynamic residual exceeds a preset threshold of 15%, the residual compensation correction module triggers an attenuation factor to suppress the execution weight of the reference power output instruction set. The system then superimposes compensation power based on the differential term of the local actual load scalar, switching the control logic from feedforward prediction based on administrative instructions to self-regulation based on physical feedback. To adapt to the damped scheduling mode and avoid scheduling oscillations, the port energy center completes the pre-charging and discharging layout of energy storage units 15 minutes in advance based on the logistics loading and unloading instructions with administrative attributes. When dealing with the load peak caused by multiple ships berthing together, the system smooths out power fluctuations through pre-scheduled energy storage output, keeping demand rate expenditures within the preset contract range. The charging and discharging depth of energy storage units is controlled, avoiding battery asset overdraft due to power compensation. The port's multi-source energy system realizes the convergence of heterogeneous data from the administrative logistics network, the underlying electrical network to the enterprise financial network at the logical level, improving the port energy management center's ability to predict and coordinate the operation scenario.

[0056] Example 2: In a container terminal equipped with distributed photovoltaic and high-power frequency conversion loading and unloading equipment, the adjustment status of the terminal's multi-source energy collaborative management system in response to the impact of highly randomized operational instructions was examined. The experiment used anonymized logistics operation plan data generated from 24 hours of continuous operation as input. This data included 860 loading and unloading instructions with timestamps and the movement coordinates of 15 gantry cranes. The experimental platform was built based on digital twin technology, and its computing environment met the specifications of a processor with a main frequency of not less than 3.0 GHz and 64 GB of RAM to support parallel tensor operations of business data streams in digital space. To determine the parameters that balance the continuous requirements of loading and unloading operations with the load on computing resources, the experiment used time windows... The settings were configured so that, given the instruction density in the logistics operation plan data ranged from 40 to 60 times per hour, the complete berthing operation cycle of a single container ship could be covered and energy purchase costs calculated. Low-value path, time window The time was set to 240 minutes. Based on this, Gaussian random noise with a signal-to-noise ratio of 25dB and 8% job execution delay jitter were superimposed on the original input data to examine the data correction status of the residual compensation correction module under non-ideal working conditions.

[0057] The experimental design included control group A, experimental group B, and control group C. Control group A adopted a feedback scheduling mode based on grid-side power threshold triggering. Control group C, as a partially missing feature group, removed the residual compensation correction module. Experimental group B adopted the system claimed in this invention. Under the condition of an initial average base load of 3.12MW accompanied by a 5.85MW pulse impact, the average container weight in the logistics operation plan data was 28.64t. The expected load spatiotemporal tensor generated by the load prediction conversion module showed a load ramp of 2.45MW within the next 15 minutes. This data served as the input to the resource scheduling optimization module, driving the energy storage unit to perform charging and discharging actions. Due to the superimposed execution delay noise, a dynamic residual of 0.62MW was generated between the real-time load power monitored at the physical junction point and the predicted load power. The residual compensation correction module adjusted the amplitude of the power allocation command based on this residual value. Under the same business flow impact, the peak power of the grid connection point of control group A reached 6.18MW, frequently triggering the demand limit, leading to increased energy purchase costs. The cost was 14,582.45 yuan. Through feedforward guidance of administrative directives, the peak power at the grid connection point of Experimental Group B was reduced to 4.42MW, thus lowering the energy purchase cost. The cost was reduced to 12781.18 yuan, achieving a cost optimization rate of 12.35%. However, control group C, lacking correction for delay noise, experienced scheduling oscillations at its grid connection point, with a peak value reaching 5.26MW. To determine the numerical range boundaries, an out-of-range control group D was set up, extending the time window... Set to 30 minutes, because The timeframe was shortened to 30 minutes, but the system failed to detect the long-cycle trend of loading and unloading operations, resulting in the energy storage units failing to complete sufficient pre-charging during periods of low electricity prices, thus increasing their energy purchase costs. Compared to test group B, the increase of 9.56% confirms that the time parameter range defined by the present invention is the optimal working window that balances prediction accuracy and scheduling depth. This experiment demonstrates, through the correlation changes of gradient data, that the architecture that maps administrative logistics management logic to physical energy space is stable under interference conditions, enabling the terminal energy system to cope with uncertain business shocks in a predictive manner.

[0058] Example 3: In a container terminal operation with 12 quay cranes with a rated power of 800kW and 25 automated guided vehicles (AGVs), the system needs to address load prediction deviations caused by differences in mechanical performance between equipment from different manufacturing batches. To determine the physical parameters of the energy efficiency model in the load prediction conversion module, the business state mapping module retrieves a test command with a 30t standard load, driving the target equipment at the displacement starting point. and the end point of displacement The system performs round-trip movements, synchronously reading the real-time power scalar output from the inverter inside the operating equipment via sensors. The load forecasting and conversion module calculates the power loss factor using the least squares method based on the measured power sequence and corresponding operating time. The physical transformation logic upon which its calculation is based is as follows: ,in, This is a real-time power scalar for the operating equipment. This is the preset no-load standby power constant in the device power mapping rules. For container weight data, The displacement distance is the Euclidean distance determined by the displacement start point data and the displacement end point data. To determine the duration of the action based on the timestamp, this procedure calculates the power loss coefficient by performing calculations on 10 sets of work cycles with different load gradients. The average value is used to eliminate logical gaps in the energy efficiency mapping process. To accurately characterize the dynamic characteristics of lifting equipment and automated guided vehicles under high-speed operation, the power loss coefficient is... In its implementation, it is refined into a composite function containing both mechanical friction and aerodynamic drag components, where the mechanical friction component is related to the container weight. The aerodynamic drag component exhibits a linear correlation, while it is modeled nonlinearly based on the instantaneous velocity simulated by the digital twin platform. Its value is proportional to the square of the displacement velocity. By superimposing this nonlinear correction term in the digital space, the power loss coefficient is adjusted. It can adapt to the nonlinear power loss characteristics under different motion intensities, thereby correcting the prediction bias of a single linear mapping under high-speed conditions.

[0059] During the real-time system control phase, the residual compensation and correction module employs a dynamic residual evaluation procedure based on a sliding time window to determine the judgment threshold for load deviation. The module extracts the real-time load power sequence and the predicted electricity load data sequence from the previous 300 seconds and calculates the standard deviation of their difference. and the mean of the residuals The confidence threshold is calculated as follows: When the absolute value of the monitored dynamic residual exceeds the confidence threshold for five consecutive sampling periods, the deviation correction logic within the residual compensation correction module is triggered. This step uses the first-order difference term of the real-time load sequence to adjust the output amplitude of the power allocation command, ensuring that when interference occurs at the terminal due to plan changes or sudden equipment speed limits, the control system can maintain power smoothness on the grid side based on physical feedback. In this process, the system establishes a cross-scale logic bridging mechanism. The residual compensation correction module converts minute-level logistics operation progress deviations into accumulated energy gaps within the corresponding time window as low-frequency correction biases. By extracting the first-order difference term of the second-level real-time load sequence, the instantaneous change trend of physical equipment actions is identified as a high-frequency bias. To compensate for the gain, the system uses an adaptive damping filter to linearly weight and fuse low-frequency energy bias and high-frequency power gain. This allows management and physical signals, which originally had different properties, to converge under a unified power reference framework. This ensures that the compensated command amplitude can track long-term changes in business progress while suppressing short-term disturbances caused by physical pulses. Under short-term abnormal operational disturbances with high-frequency jump characteristics, the residual feature analysis unit is activated in parallel to execute the frequency domain deviation extraction procedure. Based on the discrete Fourier transform principle, the residual feature analysis unit extracts the frequency domain amplitude spectrum of the difference sequence between real-time load power and predicted load power within a preset time window, setting the cutoff frequency threshold to 0.05Hz. The reason is that the business cycle of routine loading and unloading operations at the dock is usually greater than 20 seconds, and the corresponding power change components are concentrated in the low frequency band. However, the frequency of abnormal disturbances caused by equipment startup or mechanical impact is usually above 0.1Hz. By using a frequency cutoff of 0.05Hz, the system can effectively decouple the planning deviation at the business level from the transient abnormal disturbances at the physical level. This ensures that the residual characteristic analysis unit only performs power compensation decisions for unexpected sudden pulse loads. The cutoff frequency threshold is set to 0.05Hz, and the sum of the squares of the amplitudes of harmonic components with frequencies greater than the cutoff frequency threshold is accumulated. The sudden load characteristic parameters are output. If the sudden load characteristic parameters exceed the sudden load threshold set by the system, the physical situation is determined. When a high-frequency loading / unloading pulse occurs, the residual compensation correction module sends an upward adjustment command to the energy storage control unit, increasing the charging and discharging power limit by 20% to 30%. This is achieved by neutralizing the high-frequency power gap through the instantaneous response of the energy storage device. The adjustment range is determined based on the short-time overload characteristics of the energy storage inverter (PCS) and the pulse discharge rate limit of the power battery pack. Under the millisecond-level surge current generated at the moment of starting the port crane, the inverter power devices are allowed to inject 1.2 to 1.3 times the short-time current within the thermal safety range. The system confirms that the power tube junction temperature still has redundancy by retrieving the thermal management sensor data of the energy storage unit, and thus temporarily releases the power limit within this dynamic range to cope with the impact of the high-frequency loading / unloading pulse on the bus frequency.

[0060] The application of the aforementioned calibration and judgment procedures ensures that the system achieves a prediction accuracy of no less than 95% when processing loading and unloading instructions under specific equipment numbers, without relying on manual experience for setting. In the operational test of a 65t quay crane, the load prediction and conversion module, based on the calibration results… The predicted load for a single hoisting cycle is calculated. Due to a 12-second operation delay caused by the alignment deviation of the on-site guide vehicle, the residual compensation correction module automatically adjusts the power output component of the energy storage unit after detecting that the residual value exceeds the dynamically calibrated 0.18MW threshold, ensuring that the real-time load at the grid connection point is always controlled within the power quota boundary. This implementation method provides a stable path for the port energy collaborative management and control scheme to be implemented in a heterogeneous equipment environment by restoring the energy efficiency conversion logic to a measurable parameter calculation process. The commercial parameter acquisition module determines the unit depreciation cost through the asset loss characteristic mapping procedure, and retrieves the cumulative charge and discharge cycle number of the energy storage unit. The system uses the depth of discharge data within the corresponding time window and the capacity decay function built into the battery aging model to calculate the current capacity retention rate (SOH). Based on the electrochemical empirical degradation model, the capacity decay function is obtained by subtracting the degradation increment from the initial state baseline value. The capacity retention rate (SOH) is equal to a constant 1 minus the product of the aging rate factor and the cumulative charge-discharge cycle count N raised to the power of 1 / 2. The aging rate factor is a dimensionless empirical constant with a value range of 0.001 to 0.005. The aging rate factor is calibrated using the charge-discharge cycle baseline data from the energy storage unit's factory test file. Specifically, the commercial parameter acquisition module uses online electrochemical impedance spectroscopy (EIS) measurement technology to acquire the real and imaginary parts of the complex impedance of the energy storage unit at different frequencies in real time and extracts the charge transfer resistance, which characterizes the electrode dynamics. The system pre-stores the charge transfer resistance change rate and aging rate factor. A mapping function table between drift values ​​is used to compare the current drift values. The real-time correction value of the aging rate factor is calculated based on the resistance difference compared to the baseline condition. This is then superimposed on the basic aging rate factor to achieve dynamic calibration of the capacity decay slope, ensuring that the depreciation cost calculation can truly reflect the actual decay trajectory of the electrochemical system. The basic depreciation coefficient is established by using the ratio of the initial investment cost of the energy storage unit to the preset final value of the cycle life, and the electrochemical capacity decay slope is determined by combining the first derivative of the capacity retention rate (SOH) with the number of cycles. This transforms the physical loss per unit cycle into a quantitative component in the optimization of energy purchase cost, so that the power allocation instructions generated by the resource scheduling optimization module include protective constraints for the long-term life of energy storage assets.

[0061] Example 4: Deploy a control system in a newly built automated container terminal area to address the hardware differences between the energy efficiency performance of heterogeneous loading and unloading equipment under operating conditions and the built-in standard model of the digital twin platform. The business status mapping module issues instruction sets to the target equipment, driving the operating equipment to generate a complete operation cycle under no-load conditions, and simultaneously acquiring the instantaneous power data sequence within this cycle through power monitoring sensors. The load forecasting conversion module will obtain By fitting the standard no-load power curve within the digital twin space, an initial deviation correction factor is calculated for that specific device number. and the deviation correction factor Stored as a global bias value for the device power mapping rules, where... for Instantaneous no-load power of the equipment at all times. These are the correction coefficients for the energy efficiency model.

[0062] After the system enters real-time monitoring mode, it needs to address the clock asynchrony between sensor-collected data and control system command data. The residual compensation correction module determines the time synchronization procedure between spatial coordinates and energy data, and the business status mapping module reads the start timestamp from the logistics operation plan data. The transmission delay between the information stream and the physical stream is determined by cross-correlating the information with the rising edge of the real-time load power sequence. The residual compensation correction module is based on the determined Synchronizing and aligning electricity load forecast data ensures that the calculated dynamic residuals represent power fluctuations caused by logistics loading and unloading operations, thus keeping the real-time load at the grid connection point stable and controlled within the electricity quota boundaries. This is the logical start time corresponding to the logistics instruction. This refers to the combined latency of system communication and sensing.

[0063] Example 5: In situations where mechanical wear of the quay crane's drive mechanism causes deviations in the energy efficiency model due to high-frequency heavy-load loading and unloading operations, the system employs a periodic verification procedure based on the equipment power mapping rules. The business status mapping module retrieves displacement data from historical operation cycles and extracts real-time power scalars collected by power monitoring sensors. The load prediction and conversion module calculates the energy efficiency residual by comparing the real-time power scalars with the electricity load prediction data within the corresponding time window. In the energy efficiency residual item When the average value deviates from the preset value by 8% over 20 consecutive operating cycles, the load forecasting and conversion module corrects the power loss coefficient using a recursive least squares algorithm. The corrected parameters are then updated to the digital twin platform to maintain the correspondence between the prediction model and the physical entity. This is the dimensionless residual ratio of the measured power to the predicted power.

[0064] When the system faces a situation where the frequency of the grid connection point becomes unstable due to fluctuations in the output of the distributed photovoltaic system, and it is necessary to balance the energy purchase cost, the resource scheduling optimization module determines the weighting coefficients. The calibration procedure of the digital twin platform involves constructing a dynamic frequency response model based on the synchronous generator rotor motion equation. It performs calculations according to the physical relationship that the AC bus frequency change rate equals the grid imbalance power divided by the equivalent inertia constant. The equivalent inertia constant characterizes the mechanical characteristics of the power system in maintaining frequency stability, and its value is set by the sum of the nameplate parameters of the synchronous generators and virtual synchronous generators within the distribution network. The grid imbalance power is the difference between the injected power from the multi-source energy system and the expected electricity load. The digital twin platform dynamically calculates the grid imbalance power, solves the first-order differential equation, and outputs the expected frequency deviation as a constraint verification index. In the pre-deployment stage, the resource scheduling optimization module adjusts the penalty term weight coefficient in steps of 0.1. Simultaneously record the frequency fluctuation values ​​of the multi-source energy system at each step size, and lock the current value when it reaches the warning boundary of 0.15Hz. To control the baseline value, the resource scheduling optimization module obtains the real-time available power scalar of the distributed photovoltaic system and adjusts the baseline power output instruction set and the power limit component of the distributed photovoltaic system according to the weighted allocation logic, so that the real-time load at the grid connection point is under control within the power quota boundary. This is a weighting adjustment factor used to measure the relationship between business returns and the strength of frequency constraints.

[0065] When performing collaborative management and control, the resource scheduling optimization module adopts a hierarchical load absorption decision procedure. The resource scheduling optimization module compares the electricity load forecast data with the real-time available power scalar of the distributed photovoltaic system to determine the primary absorption component. If the real-time load at the grid connection point still exceeds the power quota boundary after absorption, the energy storage unit is triggered to perform a discharge action based on the power allocation command. When the output power of the energy storage unit reaches the physical extreme value and the system frequency fluctuation amplitude approaches the constraint boundary of 0.2Hz, the system automatically adjusts the adjustment component of the port shore power system to maintain the power balance of the terminal energy network. This process is transmitted through the physical constraint of the logic threshold, reducing the demand electricity cost while ensuring the stability of the system frequency.

[0066] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0067] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A multi-source energy collaborative management and control system for a port based on digital twins, characterized in that, include: The business status mapping module is used to obtain logistics operation plan data from the terminal operating system. The logistics operation plan data includes container position movement instructions with timestamps, container weight data, yard coordinate data, and corresponding operation equipment numbers. The load forecasting and conversion module, connected to the business status mapping module, is used to convert logistics operation plan data into energy efficiency models for each operation link according to the equipment power mapping rules built into the digital twin platform, and calculate the power loss coefficient according to the displacement distance of the operation equipment number under the logistics operation plan data, so as to generate power load forecast data corresponding to the preset time window. The commercial parameter acquisition module is used to acquire time-of-use electricity price data of the power system and the cumulative charge-discharge cycle number of each energy storage unit. Based on the electrochemical capacity decay slope preset in the battery aging model, the current unit depreciation cost of the energy storage unit is determined. The resource scheduling optimization module is connected to the load forecasting conversion module and the business parameter acquisition module, respectively. It is used to take the electricity load forecasting data as the feedforward input, combine it with the time-of-use electricity price data and the unit depreciation cost, and calculate and generate the power allocation instructions corresponding to the multi-source energy system with the goal of minimizing the total operating cost. The residual compensation correction module is used to monitor the real-time load power of the grid connection point, compare the real-time load power with the predicted load power generated synchronously by the digital twin platform to generate a residual value, and adjust the amplitude of the power distribution command based on the product of the residual value and the preset feedback gain.

2. The multi-source energy collaborative management and control system for a port based on digital twins according to claim 1, characterized in that, When generating electricity load forecast data, the load forecasting conversion module determines the nonlinear frictional resistance of the operating equipment during the displacement process based on container weight data and yard coordinate data, and adds it to the power loss coefficient.

3. The multi-source energy collaborative management and control system for a port based on digital twins according to claim 1, characterized in that, The equipment power mapping rules include power consumption benchmark data for quay cranes, yard cranes, and container trucks under different loads; the load prediction conversion module obtains the instantaneous power curve of the operating equipment within a preset time window by simulating the movement trajectory of the operating equipment under the logistics operation plan data in the digital twin platform.

4. The multi-source energy collaborative management and control system for a port based on digital twins according to claim 1, characterized in that, The commercial parameter acquisition module obtains the electrochemical impedance data of each energy storage unit and calculates the real-time correction value of the capacity decay slope to update the unit depreciation cost.

5. A multi-source energy collaborative management and control system for a port based on digital twins as described in claim 1, characterized in that, The power allocation instructions include the power limiting component of the distributed photovoltaic system, the charging and discharging scheduling component of the energy storage unit, and the power supply regulation component of the port shore power system; the resource scheduling optimization module balances commercial revenue by establishing a multi-objective constraint set, while maintaining the frequency fluctuation of the multi-source energy system within the range of 0.2Hz.

6. A multi-source energy collaborative management and control system for a port based on digital twins as described in claim 1, characterized in that, The residual compensation correction module also includes a residual characteristic analysis unit, which is used to analyze the deviation characteristics between real-time load power and predicted load power in the frequency domain.

7. A multi-source energy collaborative management and control system for a port based on digital twins as described in claim 1, characterized in that, When correcting power allocation instructions, the residual compensation correction module adds compensation power based on the first derivative of the load according to the progress deviation of the logistics operation plan data, so as to make the expected trajectory of the digital twin platform converge with the actual measured operation trajectory of the physical site.

8. A multi-source energy collaborative management and control system for a port based on digital twins as described in claim 1, characterized in that, The system also includes a data governance module, which is used to normalize heterogeneous data from the terminal operating system, multi-source energy system and power grid sensors, and establish a unified data bus with timestamp alignment function.

9. A multi-source energy collaborative management and control system for a port based on digital twins according to claim 1, characterized in that, Before outputting power allocation commands, the resource scheduling optimization module calculates the distribution network security verification results based on the simulated power topology in the digital twin platform. When the verification results show that the node voltage offset is less than 5%, the power allocation command is released to the multi-source energy system.

10. A method for coordinated management and control of multi-source energy at a port based on digital twins, used to implement the coordinated management and control system for multi-source energy at a port based on digital twins as described in claim 1, characterized in that, Includes the following steps: Step S1001: Obtain logistics operation plan data from the terminal terminal operating system. The logistics operation plan data includes container position movement instructions with timestamps, container weight data, yard coordinate data, and corresponding operation equipment numbers. Step S1002: Based on the preset equipment power mapping rules in the digital twin platform, the logistics operation plan data is converted into the energy efficiency model of each operation link, and the power loss coefficient is calculated according to the displacement distance of the operation equipment number under the logistics operation plan data, so as to generate the power load prediction data corresponding to the preset time window. Step S1003: Obtain the time-of-use electricity price data of the power system and the cumulative charge-discharge cycle number of each energy storage unit, and determine the current unit depreciation cost of the energy storage unit based on the electrochemical capacity decay slope preset in the battery aging model. Step S1004: Using the electricity load forecast data as the feedforward input, combined with the time-of-use electricity price data and the unit depreciation cost, the power allocation instruction corresponding to the multi-source energy system is calculated and generated with the goal of minimizing the total operating cost. Step S1005: Monitor the real-time load power of the grid connection point, compare the real-time load power with the predicted load power generated synchronously by the digital twin platform to generate a residual value, and adjust the amplitude of the power distribution command based on the product of the residual value and the preset feedback gain.

Images