Micro-grid energy optimization scheduling system fusing digital twinning

Abstract

The invention provides a micro-grid energy optimization scheduling system fusing digital twinning, and the system comprises a physical micro-grid layer which is used for collecting real-time operation data through a sensor network, and adjusting the operation state of equipment according to an equipment control instruction issued by an execution control layer; the data sensing layer is used for receiving, preprocessing and transmitting the real-time operation data; the digital twinborn platform layer is used for constructing and updating a digital twinborn model based on the preprocessed real-time operation data, performing real-time dynamic simulation, extreme scene simulation and operation state prediction based on the digital twinborn model, and outputting a simulation result and prediction data; the energy optimization decision-making layer is used for receiving the simulation result and the prediction data, solving through an optimization algorithm by taking low carbon and stability as a planning target, and generating an optimal scheduling scheme; and the execution control layer is used for converting the optimal scheduling scheme into an equipment control instruction and issuing the equipment control instruction to the physical micro-grid layer so as to better adapt to a complex micro-grid operation environment.

Description

Technical Field

[0001] This disclosure relates to the field of new energy technology, specifically to a microgrid energy optimization scheduling system integrating digital twins. Background Technology

[0002] The smart park microgrid energy optimization and dispatching system is an intelligent management system that relies on the collaborative control architecture of source-grid-load-storage. By integrating the Internet of Things, intelligent algorithms and digital twin technology, it realizes dynamic optimization of the entire chain of energy production, storage and consumption in the park. Its core objective is to maximize the consumption of clean energy and minimize carbon emissions and operating costs while ensuring the reliability of power supply.

[0003] However, existing microgrid energy management methods still have the following shortcomings: (i) lack of high-fidelity dynamic simulation capability; due to the failure to deeply integrate digital twin technology, existing methods cannot build high-fidelity virtual models, making it difficult to perform real-time accurate mapping and state inference of physical entities. This results in insufficient dynamic response and forward-looking control capabilities when the system faces complex operating conditions such as load changes, equipment failures, and fluctuations in new energy output; (ii) relatively simple optimization dimensions; in the optimization process, existing methods do not comprehensively consider the synergistic optimization of key objectives such as "low carbon" and "stability", making it difficult to achieve comprehensive optimization of multiple objectives; (iii) poor scenario adaptability; in extreme operating scenarios under many complex constraints, existing methods are difficult to effectively cope with and cannot meet actual operating needs.

[0004] Therefore, there is an urgent need to build a low-carbon smart park microgrid energy optimization and scheduling scheme that integrates digital twins to better adapt to the complex microgrid operating environment. Summary of the Invention

[0005] This disclosure provides a microgrid energy optimization scheduling system that integrates digital twins to better adapt to the complex operating environment of microgrids.

[0006] This disclosure provides a microgrid energy optimization scheduling system integrating digital twins, including:

[0007] The physical microgrid layer (101) is used to collect real-time operating data through the sensor network and adjust the operating status of the equipment according to the equipment control commands issued by the execution control layer.

[0008] The data sensing layer (102) is connected to the physical microgrid layer (101) and is used to receive, preprocess and transmit the real-time operation data;

[0009] a digital twin platform layer (103) connected with the data perception layer (102), configured to construct and update a digital twin model based on the preprocessed real-time operation data, and perform real-time dynamic simulation, extreme scenario simulation and operation state prediction based on the digital twin model, and output simulation results and prediction data;

[0010] an energy optimization decision layer (104) connected with the digital twin platform layer (103), configured to receive the simulation results and prediction data, and generate an optimal scheduling scheme by solving through an optimization algorithm with low carbon and stability as planning targets;

[0011] an execution control layer (105) connected with the energy optimization decision layer (104) and the physical microgrid layer (101) respectively, configured to convert the optimal scheduling scheme into device control instructions and issue the device control instructions to the physical microgrid layer (101).

[0012] In some embodiments, the energy optimization decision layer (104) comprises a target function module;

[0013] a low carbon target function is:

[0014]

[0015] wherein is the total carbon emission, is the scheduling period, is the power purchase amount from the large power grid at time t, is the power generation amount of the diesel generator at time t, is the average carbon emission factor of the park connected to the regional power grid at time t, is the carbon emission factor of the diesel generator, and are dynamically updated by a prediction analysis module;

[0016] a stability target function is:

[0017]

[0018]

[0019]

[0020] wherein, is a voltage fluctuation constraint term, is a frequency deviation constraint term, is a power supply reliability constraint term; the node voltage at time t , the system frequency are real-time voltage amplitude and system frequency output by a dynamic simulation module; The rated voltage of the park's microgrid. Rated frequency; power outage duration at time t. Load power loss The simulation values ​​output from the scene simulation module; The scheduling period; The total load at time t is the load demand forecast from the predictive analysis module; These are the corresponding weighting coefficients.

[0021] In some embodiments, the energy optimization decision layer (104) includes a constraint module, the constraints of which include:

[0022] The power balance constraint is:

[0023]

[0024] in , These represent the photovoltaic power output and wind power output at time t, respectively. For energy storage charging and discharging power, Let t be the amount of electricity purchased from the main power grid. Let t be the total load at time t;

[0025] The constraint formula for energy storage devices is:

[0026]

[0027]

[0028] in It is currently showing as charging. The time display shows discharge. Let t be the state of charge of the energy storage device. Energy storage status data updated in real time from the digital twin model; , The charging and discharging efficiency is determined based on the characteristics of the energy storage battery; Δt is the scheduling time interval. , These are the lower and upper limits of the charge level for energy storage devices, respectively.

[0029] The carbon emission constraint formula is: ,in t represents the upper limit of carbon emissions at time t, which is the simulated value of carbon emissions output from the scenario simulation module.

[0030] In some embodiments, the digital twin platform layer (103) includes:

[0031] Virtual model module, dynamic simulation module, scene simulation module and predictive analysis module.

[0032] In some embodiments, the virtual model module is configured to construct a three-dimensional geometric model of the micro-grid, import device characteristic parameters, and finally load operation constraint rules to form a digital twin model.

[0033] In some embodiments, the dynamic simulation module is configured to receive a candidate scheduling scheme from the energy optimization decision layer (104) as input, run the digital twin model in a real-time simulation environment, and calculate and update state quantities of each node of the micro-grid in real time, the state quantities including at least real-time voltage amplitude and system frequency.

[0034] In some embodiments, the scenario simulation module is configured to generate a large number of random scenarios based on the Monte Carlo method, filter out the top N high-risk scenarios, drive the digital twin model to perform deep simulation on these scenarios, and statistically aggregate the simulation results to output simulation values including carbon emissions, power outage duration, and load outage amount.

[0035] In some embodiments, the prediction analysis module integrates a hybrid prediction model of LSTM and attention mechanism, and is configured to predict the total load demand of the park, the carbon emission factor of the regional power grid, and the carbon emission factor of the diesel generator based on historical time series data.

[0036] In some embodiments, the energy optimization decision layer (104) includes an optimization algorithm module.

[0037] The optimization algorithm module generates a candidate scheduling scheme by calling an optimization algorithm under given objective functions and constraint conditions.

[0038] In some embodiments, the energy optimization decision layer (104) includes a decision evaluation module.

[0039] The decision evaluation module performs multi-dimensional evaluation on the candidate scheduling scheme generated by the optimization algorithm, constructs an evaluation index system including low carbon and stability, determines the index weight by using the analytic hierarchy process (AHP), and sorts the candidate scheduling schemes by using the technique for order preference by similarity to ideal solution (TOPSIS) method to select the optimal scheduling scheme.

[0040] The present disclosure has the following advantages compared with the prior art:

[0041] 1. Realize perception-decision-execution closed-loop control: introduce digital twin driving technology, construct a virtual model with high precision mapping of the physical micro-grid, achieve real-time simulation, accurate prediction and dynamic correction of system operation state, form a complete closed loop of "virtual-real interaction and continuous iteration", and greatly improve the dynamic response speed and adaptive ability of the system.

[0042] 2. Enhance the system's resilience to complex scenarios: By using a digital twin platform to perform Monte Carlo simulations and risk screening for extreme scenarios, optimization decisions can fully consider various potential risks, thereby setting scientific safety boundaries (such as dynamic carbon emission limits), significantly enhancing the system's resilience and power supply reliability under uncertain conditions.

[0043] 3. Achieve multi-objective collaborative optimization goals: Construct a multi-objective optimization framework focusing on "low carbon" and "stability", and use algorithms to automatically generate comprehensive optimal scheduling strategies that take into account clean energy utilization, carbon emission control, voltage and frequency stability and power supply reliability. This overcomes the shortcomings of traditional methods with a single optimization dimension and achieves a coordinated unity of economic, environmental and safety benefits. Attached Figure Description

[0044] The accompanying drawings, which are incorporated in and form a part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure.

[0045] Figure 1 A schematic diagram of a microgrid energy optimization scheduling system integrating digital twins is provided as an embodiment of this disclosure;

[0046] Figure 2 A schematic diagram of a digital twin platform layer structure provided in this embodiment of the disclosure;

[0047] Figure 3 This is a schematic diagram of an energy optimization decision layer structure provided in an embodiment of this disclosure.

[0048] The accompanying drawings have illustrated specific embodiments of this disclosure, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concepts of this disclosure to those skilled in the art through reference to particular embodiments. Detailed Implementation

[0049] The present disclosure will be further described below with reference to the accompanying drawings. The following embodiments are only used to illustrate the technical solutions of the present disclosure more clearly, and should not be used to limit the scope of protection of the present disclosure.

[0050] like Figure 1 As shown, this disclosure proposes a microgrid energy optimization scheduling system 100 integrating digital twins, comprising:

[0051] The physical microgrid layer 101 is used to collect real-time operating data through the sensor network and adjust the operating status of the equipment according to the equipment control commands issued by the execution control layer.

[0052] The data sensing layer 102 is connected to the physical microgrid layer 101 and is used to receive, preprocess and transmit the real-time operating data.

[0053] The digital twin platform layer 103 is connected with the data perception layer 102, and is configured to construct and update a digital twin model based on preprocessed real-time operation data, and perform real-time dynamic simulation, extreme scenario simulation and operation state prediction based on the digital twin model, and output simulation results and prediction data;

[0054] The energy optimization decision layer 104 is connected with the digital twin platform layer 103, and is configured to receive the simulation results and prediction data, and generate an optimal scheduling scheme by solving through an optimization algorithm with low carbon and stability as planning targets;

[0055] The execution control layer 105 is connected with the energy optimization decision layer 104 and the physical micro-grid layer 101 respectively, and is configured to convert the optimal scheduling scheme into device control instructions and issue the device control instructions to the physical micro-grid layer 101.

[0056] In an optional embodiment, the physical micro-grid layer 101 includes a distributed power generation unit, an energy storage module, a device cluster control module and a sensor network, the distributed power generation unit includes a photovoltaic power generation system, a wind power generation system and a diesel power generation system, and the energy storage module stores energy converted from the photovoltaic power generation system, the wind power generation system and the diesel power generation system in a lithium battery pack;

[0057] The device cluster control module includes a photovoltaic inverter controller and a generator controller configured in the distributed power generation unit, and an energy storage BMS controller configured in the energy storage module.

[0058] The sensor network includes smart meters, voltage and current transformers and power sensors in the device cluster control module, device state sensors arranged in each node of the distributed power generation unit, and temperature sensors, humidity sensors, light sensors and wind speed sensors distributed in the monitored environment.

[0059] In an optional embodiment, the data perception layer 102 includes a data communication module connected with the sensor network, the data communication module is divided into a wired transmission unit and a wireless transmission unit, the wired transmission unit uses industrial Ethernet and optical fiber communication in nodes, and the wireless transmission unit uses short-range wireless transmission and medium- and long-range wireless transmission to realize data communication, and specifically includes:

[0060] 1) LoRa modules and ZigBee modules are used for distributed sensor data transmission;

[0061] 2) NB-IoT modules and industrial routers are configured for large-scale data aggregation transmission;

[0062] The data perception layer 102 further comprises a data gateway, which is configured with an edge computing node for realizing data local caching and protocol conversion, and realizing data preprocessing according to the edge computing node, specifically including:

[0063] 1) Kalman filtering algorithm is adopted to smooth the high-frequency fluctuation data from the sensor device network module, so as to filter data noise;

[0064] 2) The missing values in the sensor data are predicted and completed based on an LSTM neural network;

[0065] 3) z-score standardization processing is adopted to convert different dimensional sensor data into standardized values in the interval [0, 1].

[0066] In an optional embodiment, the digital twin platform layer 103 comprises a virtual model module 1031, a dynamic simulation module 1032, a scenario simulation module 1033 and a prediction analysis module 1034.

[0067] The virtual model module 1031 is used to construct a three-dimensional geometric model of the microgrid, import device characteristic parameters, and finally load and run constraint rules to form a digital twin model, which is stored in a blockchain to ensure that the model cannot be tampered with;

[0068] Preferably, the virtual model module 1031 is divided into a physical modeling unit, a behavior modeling unit and a rule modeling unit, specifically including:

[0069] 1) The physical modeling unit is based on the actual topological structure of the park microgrid, combined with building information model (BIM) and geographic information system (GIS) technology, to map the size, position and connection relationship of physical entities such as photovoltaic arrays, fans, energy storage stations, transformers and load nodes;

[0070] 2) The behavior modeling unit is based on mechanism analysis and data-driven combination to establish device operation characteristic models, including photovoltaic panel I-V characteristic curve models, lithium battery charge and discharge efficiency models, and load time-varying characteristic models, etc.

[0071] 3) The rule modeling unit embeds microgrid operation constraint rules, including power balance constraints, energy storage device constraints and carbon emission constraints. The constraint formula is the same as the constraint condition stored in the constraint condition module 1042 of the energy decision optimization layer.

[0072] The dynamic simulation module 1032 is used to receive a candidate scheduling scheme from the energy optimization decision layer 104 as input, run the digital twin model in a real-time simulation environment, and real-time calculate and update the state quantity of each node of the microgrid, the state quantity at least including real-time voltage amplitude and system frequency.

[0073] In this embodiment, the calculated state quantity will be fed back to the energy optimization decision layer 104 as an input parameter of the objective function.

[0074] The scenario simulation module 1033 is configured to generate a large number of random scenarios based on the Monte Carlo method, screen out the top N high-risk scenarios, drive the digital twin model to perform deep simulation on these scenarios, and statistically aggregate the simulation results to output various simulation values including carbon emissions, power outage duration, and load outage amount.

[0075] The statistical aggregation is a weighted average or maximum value based on the influence degree of each scenario.

[0076] The various simulation values including carbon emissions, power outage duration, and load outage amount will be input parameters of the objective function and constraint conditions.

[0077] In a specific implementation, the scenario simulation module 1033 is configured to perform random sampling of thousands of levels on uncertain parameters such as extreme wind speed, device failure rate, and load mutation amplitude based on the Monte Carlo probability simulation algorithm, to generate multiple sets of scenario samples covering single extreme events and superimposed extreme events. The digital thread technology is used to track the energy transmission path under each scenario, and the influence degree of each scenario is calculated, including one or more of the carbon emission exceeding rate and the voltage out-of-limit probability. The top 20% of high-risk scenarios in terms of influence degree are screened out for deep simulation, and the simulation results are statistically aggregated to output various simulation values including carbon emissions, power outage duration, and load outage amount.

[0078] The prediction analysis module 1034 integrates a hybrid prediction model combining LSTM and attention mechanism, and is configured to predict the total load demand of the park, the regional power grid carbon emission factor, and the diesel generator carbon emission factor based on historical time series data.

[0079] In this embodiment, the prediction analysis module 1034 uses a sliding window mechanism for prediction update. The historical time series data includes one or more of historical load curves, historical power grid carbon emission factors, and device historical operation records.

[0080] The output of this module will be input parameters of the objective function and constraint conditions.

[0081] It should be understood that the digital twin model is a real-time running state mapping of the physical microgrid layer 101, and the outputs of the dynamic simulation module 1032, the scenario simulation module 1033, and the prediction analysis module 1034 are simulation results and prediction data of future running states.

[0082] Optionally, the digital twin platform layer 103 further includes a data fusion module 1035, which receives preprocessed data from the data gateway in the data perception layer 102 and uses federated learning to fuse local data from edge nodes with global data from the cloud to ensure the integrity and timeliness of the model input data.

[0083] In one optional embodiment, the energy optimization decision layer 104 includes an objective function module 1041, a constraint condition module 1042, an optimization algorithm module 1043, and a decision evaluation module 1044.

[0084] The objective function module 1041 adopts a multi-objective optimization framework, which includes low-carbon objectives and stability objectives, specifically:

[0085] The low-carbon goal aims to minimize total carbon emissions. The objective function for calculating the total carbon emissions generated during the operation of the park's microgrid is as follows:

[0086]

[0087] in Total carbon emissions For the scheduling period, Let t be the amount of electricity purchased from the main power grid. Let t be the power output of the diesel generator. Let t be the average carbon emission factor of the park connected to the regional power grid. For the carbon emission factors of diesel generators, and Dynamically updated by predictive analysis module 1034;

[0088] The stability objective uses voltage fluctuation range, frequency deviation, and power supply reliability as constraints, and its objective function is:

[0089]

[0090]

[0091]

[0092]

[0093] in, For voltage fluctuation constraint terms, For frequency deviation constraint terms, For power supply reliability constraints; node voltage at time t. System frequency The real-time voltage amplitude and system frequency are output using the dynamic simulation module 1032; The rated voltage of the park's microgrid. Rated frequency; power outage duration at time t. Load power loss The simulation values ​​output from the scene simulation module 1033; The scheduling period; The total load at time t is the load demand forecast from the predictive analysis module 1034; These are the corresponding weighting coefficients.

[0094] Understandably, the voltage fluctuation constraint formula is expressed as follows: when the voltage deviation exceeds the ±5% deviation limit specified in the national standard, it needs to be multiplied by a penalty coefficient of 10 to force the optimization algorithm to prioritize controlling the voltage within the constraint range. The rated voltage of the park's microgrid (such as 380V for three-phase low-voltage distribution or 220V for single-phase low-voltage distribution) corresponds to the power distribution standard of physical microgrid layer 101.

[0095] The formula for the frequency deviation constraint is as follows: when the frequency deviation exceeds the ±0.5Hz deviation limit specified in the national standard, it is multiplied by a penalty coefficient of 15 to strengthen the frequency stability constraint. The rated frequency is used to match the power frequency standard of the park's microgrid.

[0096] Among them, load loss This indicates the load demand that was not met due to a power outage.

[0097] The constraint module 1042 has a built-in constraint database that stores various constraints for microgrid operation, including power balance constraints, energy storage device constraints, and carbon emission constraints. The constraint parameters are dynamically updated according to the type of industrial park and policy requirements, specifically including:

[0098] 1) The power balance constraint expression is:

[0099]

[0100] in , These represent the photovoltaic power output and wind power output at time t, respectively. For energy storage charging and discharging power, Let t be the amount of electricity purchased from the main power grid. Let t be the total load at time t.

[0101] 2) The constraint expression for the energy storage device is:

[0102]

[0103]

[0104] in It is currently showing as charging. The time display shows discharge. Let t be the state of charge of the energy storage device. Energy storage status data updated in real time from the digital twin model; , The charging and discharging efficiency is determined based on the characteristics of the energy storage battery; Δt is the scheduling time interval. , These are the lower and upper limits for the charge of energy storage devices, respectively.

[0105] In this embodiment, Δt is the scheduling time interval, which is usually set to 15 minutes in a typical park microgrid to meet the requirements of distribution network scheduling granularity.

[0106] generally 92%-95%, The value is 90%-93%, with the specific value determined based on the current level of commercial energy storage battery technology.

[0107] , The value needs to be set according to the battery type. If it is a lithium iron phosphate battery, then =20% =90%; if it is a ternary lithium battery, then =15% =85%.

[0108] 3) The carbon emission constraint expression is: ,in t represents the upper limit of carbon emissions at time t, which is the simulated value of carbon emissions output from the scenario simulation module.

[0109] The optimization algorithm module 1043 generates candidate scheduling schemes by calling the optimization algorithm under given objective function and constraints.

[0110] The optimization algorithm can be a multi-objective genetic algorithm (NSGA-II) or a multi-objective particle swarm optimization (PSO) algorithm, etc.

[0111] The decision evaluation module 1044 evaluates the candidate scheduling schemes generated by the optimization algorithm from multiple dimensions, constructs an evaluation index system including low carbon emissions and stability, uses the analytic hierarchy process (AHP) to determine the index weights, and combines the TOPSIS method to rank the candidate scheduling schemes to select the optimal scheduling scheme.

[0112] The optimal scheduling scheme includes the output plan of each distributed energy source, the energy storage charging and discharging power curve, and the power purchase and sale period and power with the main power grid.

[0113] In one optional implementation, the execution control layer 105 includes a strategy output module, a distributed energy control module, an energy storage system control module, and a grid interaction control module;

[0114] The strategy output module transforms the optimal scheduling scheme into a standardized instruction format, which includes the output plan of each distributed energy source, the energy storage charging and discharging power curve, and the control instructions for the power purchase and sale period and power of the main power grid. The output data can be pushed in real time and traced back to the past. At the same time, a decision report is generated, which includes the basis of the scheme, expected effects and risk warnings.

[0115] The distributed energy control module is divided into a photovoltaic controller and a wind power controller;

[0116] The execution control layer 105 includes an energy storage system control module, which is divided into a charge and discharge control unit and a battery management system.

[0117] The execution control layer 105 includes a grid interaction control module, which is divided into a grid-connected switching unit and a power control module.

[0118] For example, photovoltaic power output plan: the target output power of the photovoltaic inverter for each time period (but the actual output is limited by the amount of sunshine, which refers to "planned use").

[0119] Wind power output plan: the target output power of wind turbines for each time period.

[0120] Diesel generator output plan: specify the time periods for starting and the output power.

[0121] Energy storage charge / discharge power curve: whether the energy storage system is charging (negative power) or discharging (positive power) at each time period, and the specific power value.

[0122] Electricity purchase and sale plan with the main power grid: the power value of purchasing electricity from the grid (positive power) or selling electricity to the grid (negative power) in each time period.

[0123] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0124] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0125] It should be understood that the above embodiments are only used to illustrate the technical solutions of this disclosure, and not to limit them; although this disclosure has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or make equivalent substitutions for some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this disclosure.

Claims

1. A microgrid energy optimization scheduling system integrating digital twins, characterized in that, include: The physical microgrid layer (101) is used to collect real-time operating data through the sensor network and adjust the operating status of the equipment according to the equipment control commands issued by the execution control layer. The data sensing layer (102) is connected to the physical microgrid layer (101) and is used to receive, preprocess and transmit the real-time operation data; The digital twin platform layer (103) is connected to the data perception layer (102) and is used to build and update the digital twin model based on the preprocessed real-time running data, and to perform real-time dynamic simulation, extreme scenario simulation and running status prediction based on the digital twin model, and output simulation results and prediction data. The energy optimization decision layer (104) is connected to the digital twin platform layer (103) and is used to receive the simulation results and prediction data, and to solve the optimal scheduling scheme by means of optimization algorithm with low carbon and stability as planning objectives. The execution control layer (105) is connected to the energy optimization decision layer (104) and the physical microgrid layer (101) respectively, and is used to convert the optimal scheduling scheme into equipment control commands and send them to the physical microgrid layer (101).

2. The microgrid energy optimization and scheduling system integrating digital twins according to claim 1, characterized in that, The energy optimization decision layer (104) includes an objective function module; The low-carbon objective function is: in Total carbon emissions For the scheduling period, Let t be the amount of electricity purchased from the main power grid. Let t be the power output of the diesel generator. Let t be the average carbon emission factor of the park connected to the regional power grid. For the carbon emission factors of diesel generators, and Updated dynamically by the predictive analytics module; The stable objective function is: in, For voltage fluctuation constraint terms, For frequency deviation constraint terms, For power supply reliability constraints; node voltage at time t. System frequency The real-time voltage amplitude and system frequency are output using the dynamic simulation module. The rated voltage of the park's microgrid. The rated frequency; the duration of the power outage at time t. Load loss The simulation values ​​output from the scene simulation module; The scheduling period; The total load at time t is the load demand forecast from the predictive analysis module; These are the corresponding weighting coefficients.

3. The microgrid energy optimization and scheduling system integrating digital twins according to claim 1, characterized in that, The energy optimization decision layer (104) includes a constraint module, and the constraints include: The power balance constraint is: in , These represent the photovoltaic power output and wind power output at time t, respectively. For energy storage charging and discharging power, Let t be the amount of electricity purchased from the main power grid. Let t be the total load at time t; The constraint formula for energy storage devices is: in It is currently showing as charging. The time display shows discharge. Let t be the state of charge of the energy storage device. Energy storage status data updated in real time from the digital twin model; , The charging and discharging efficiency is determined based on the characteristics of the energy storage battery; Δt is the scheduling time interval. , These are the lower and upper limits of the charge level for energy storage devices, respectively. The carbon emission constraint formula is: ,in t represents the upper limit of carbon emissions at time t, which is the simulated value of carbon emissions output from the scenario simulation module.

4. The microgrid energy optimization and scheduling system integrating digital twins according to claim 1, characterized in that, The digital twin platform layer (103) includes: Virtual model module, dynamic simulation module, scene simulation module and predictive analysis module.

5. The microgrid energy optimization and scheduling system integrating digital twins according to claim 4, characterized in that, The virtual model module is used to construct a three-dimensional geometric model of the microgrid, import equipment characteristic parameters, and finally load and run constraint rules to form a digital twin model.

6. The microgrid energy optimization and scheduling system integrating digital twins according to claim 4, characterized in that, The dynamic simulation module is used to receive candidate scheduling schemes from the energy optimization decision layer (104) as input, run the digital twin model in a real-time simulation environment, and calculate and update the state variables of each node of the microgrid in real time. The state variables include at least the real-time voltage amplitude and the system frequency.

7. The microgrid energy optimization scheduling system integrating digital twins according to claim 4, characterized in that, The scenario simulation module is used to generate a large number of random scenarios based on the Monte Carlo method, select the top N high-risk scenarios, drive the digital twin model to perform in-depth simulation of these scenarios, and statistically aggregate the simulation results to output various simulation values ​​including carbon emissions, power outage duration, and load loss.

8. The microgrid energy optimization and scheduling system integrating digital twins according to claim 4, characterized in that, The predictive analysis module integrates a hybrid predictive model of LSTM and attention mechanism, and is configured to predict the total load demand of the park, the carbon emission factor of the regional power grid, and the carbon emission factor of the diesel generator based on historical time series data.

9. The microgrid energy optimization scheduling system integrating digital twins according to claim 1, characterized in that, The energy optimization decision layer (104) includes an optimization algorithm module; The optimization algorithm module generates candidate scheduling schemes by calling optimization algorithms under given objective functions and constraints.

10. The microgrid energy optimization scheduling system integrating digital twins according to claim 1, characterized in that, The energy optimization decision layer (104) includes a decision evaluation module; The decision evaluation module evaluates the candidate scheduling schemes generated by the optimization algorithm from multiple dimensions, constructs an evaluation index system including low carbon emissions and stability, uses the analytic hierarchy process (AHP) to determine the index weights, and combines the TOPSIS method to rank the candidate scheduling schemes to select the optimal scheduling scheme.

Images