Micro-grid energy multi-objective layered optimization scheduling method based on dual-carbon constraint

Abstract

The invention discloses a micro-grid energy multi-target layered optimization scheduling method based on dual-carbon constraint, relates to the technical field of automatic intelligent scheduling of a power system, and is used for solving the problems of single scheduling target, weak dynamic response capability and insufficient carbon emission control of a micro-grid under a dual-carbon background. The method comprises the following steps: firstly, modeling wind energy, photovoltaic and electric vehicles and various energy storage units to construct a multi-energy body modeling system; and constructing a layered dynamic scheduling model by taking the operation economy, the power volatility and the carbon emission level as optimization targets. An improved algorithm is combined with model prediction control, day-ahead prediction and intra-day rolling optimization are divided, and a multi-target scheduling strategy is achieved. Aiming at system structure change caused by temporary access of the novel energy storage unit, a dynamic adjustment mechanism based on a disturbance coefficient and a structure state level is provided, the stability and consistency of a scheduling model are ensured, and the robustness and the response capability of the system are improved in an extreme meteorological scene.

Description

Technical Field

[0001] The present invention relates to the technical field of automated intelligent dispatching of power systems, and more specifically, to a multi-objective hierarchical optimization dispatching method for microgrid energy based on dual-carbon constraints. Background Art

[0002] As the global energy structure transitions toward clean, low-carbon energy, my country's "dual carbon" strategic goals are placing greater demands on the flexibility and intelligence of its power systems. Microgrids, as an important vehicle for integrating distributed power sources, energy storage systems, and smart loads, have the ability to increase renewable energy utilization, alleviate pressure on the main grid, and ensure local power supply security. They have become an essential component of building a new power system. However, current microgrid dispatching methods generally face the following technical bottlenecks: On the one hand, existing dispatch strategies are often driven by economic efficiency, neglecting comprehensive objectives such as carbon emission optimization, multi-energy coordination, and system volatility control, making them difficult to adapt to the low-carbon, high-efficiency operational demands of the "dual carbon" era. On the other hand, with the widespread integration of intermittent energy sources such as photovoltaic and wind power, the operational structure of microgrids has become dynamically uncertain. This is especially true when flexible user-side energy storage devices (such as on-board energy storage and battery clusters) and variable loads are added. Traditional static dispatch models struggle to reflect system changes in a timely manner, resulting in the risk of scheduling lags and operational instability.

[0003] In addition, the frequent occurrence of extreme climate events poses severe challenges to power supply forecasting, load response and system security. Traditional single-cycle scheduling or fixed time granularity optimization methods lack sufficient rolling correction capabilities and cannot achieve true dynamic response.

[0004] In view of the above problems, the present invention proposes a solution. Summary of the Invention

[0005] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a microgrid energy multi-objective hierarchical optimization scheduling method based on dual-carbon constraints to solve the problems raised in the above-mentioned background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions: In a preferred embodiment, it comprises: Step 1: Model the dynamic response characteristics and interactive coupling relationships of key energy units in the microgrid based on wind power, photovoltaics, electric vehicles, and multi-form energy storage; Step 2: Based on the completion of the multi-energy modeling system, an optimization scheduling framework with multi-objective coordination capabilities, a hierarchical decoupling structure, and a dynamic feedback adjustment mechanism is constructed; Step 3: Combine extreme weather events, power grid fluctuations, and abnormal user loads to build a typical extreme scenario event database; Step 4: Construct a time series data set that integrates multi-dimensional state information to dynamically track and model the operation status.

[0007] In a preferred embodiment, in step 1, the collected wind speed data is statistically modeled to construct a wind power forecast output curve; the intraday photovoltaic processing forecast curve is formed based on the irradiance and temperature data; the time series SOC curve and the scheduling available power boundary data table are formed based on the charging efficiency parameter nch and the discharge efficiency parameter nds of each energy storage unit and the upper and lower power limits; based on the urban traffic behavior statistical data, a travel probability distribution model of electric vehicles is constructed.

[0008] In a preferred embodiment, in step 2, decision variables are defined, a multi-objective function is constructed, and constraints are defined; the scheduling process is divided into two levels: day-ahead forecast optimization and intra-day rolling adjustment.

[0009] In a preferred embodiment, in step 2, for the day-ahead stage, the population is first initialized, and the multi-objective function is evaluated, and a feasible scheduling strategy solution set is output through a multi-objective Pareto frontier solution search.

[0010] In a preferred embodiment, in step 2, for the intraday stage, a rolling cycle is set, future output is predicted, a short-term prediction is performed, and a current rolling cycle scheduling strategy is output; Obtain the types and corresponding quantities of various energy storage units, calculate the newly added proportion of each type of energy storage unit, and calculate the disturbance coefficient of the overall energy storage structure; weighted sum the newly added proportion of new energy storage units and the energy storage structure disturbance coefficient after the change of the multi-energy body modeling system to determine the structural state level reference value Djc of the changed multi-energy body modeling system, set the level reference value threshold Ydjc, when the structural state level reference value Djc of the changed multi-energy body modeling system is greater than or equal to the level reference value threshold Ydjc, it indicates that the scheduling structure mapping model of the changed multi-energy body modeling system at this level does not need to be adjusted; when the structural state level reference value Djc of the changed multi-energy body modeling system is less than the level reference value threshold Ydjc, it indicates that the scheduling structure mapping model of the changed multi-energy body modeling system at this level needs to be adjusted.

[0011] In a preferred embodiment, in step 2, the number of scheduling structure changes occurring within a unit time window is obtained, the scheduling structure change frequency is calculated, and the structure change frequency fchg is compared with the change frequency threshold Yth. When the structure change frequency fchg is greater than or equal to the change frequency threshold Yth, the standard rolling MPC is still used; when the structure change frequency fchg is less than the change frequency threshold Yth, adjustments need to be made.

[0012] In a preferred embodiment, in step 3, typical extreme event indicator data are obtained to construct a scenario sample set; Introducing a two-stage robust optimization approach: Phase 1: Generate a basic scheduling plan Phase 2: Optimization of response strategies under extreme disturbances.

[0013] In a preferred embodiment, in step 4, the collected energy flow data, mains power and carbon emission data, energy storage and V2G status data, and environmental meteorological data are used to generate a data matrix and perform dimensionality reduction processing; the rate of change and similarity of the operating status trajectory of the reduced data are analyzed.

[0014] In a preferred embodiment, the method for obtaining the change frequency threshold is as follows: Calculate the mean and standard deviation of the historical change frequency of the scheduling structure, analyze the mean and standard deviation of the historical change frequency of the scheduling structure using K-means clustering, and determine the change frequency threshold.

[0015] The technical effects and advantages of the microgrid energy multi-objective hierarchical optimization scheduling method based on dual carbon constraints of the present invention are as follows: This method constructs a multi-objective function, synergistically incorporating factors such as economic efficiency, carbon emission control, and renewable energy absorption into its optimization objectives. It also employs a hierarchical scheduling architecture and robust optimization strategies, effectively improving the system's scheduling efficiency and operational stability. Furthermore, the method exhibits excellent dynamic response capabilities, adapting to fluctuations in renewable energy output and enabling refined management of carbon emissions, thus meeting the requirements for low-carbon, efficient, and intelligent operation of microgrids under the dual carbon goals. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 This is an operational flow chart of the microgrid energy multi-objective hierarchical optimization scheduling method based on dual carbon constraints of the present invention. DETAILED DESCRIPTION

[0017] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0018] The present invention is collected. Example

[0019] The present invention discloses a multi-objective hierarchical optimization scheduling method for microgrid energy based on dual carbon constraints, such as Figure 1 Shown, including: Step 1: Model the key energy units in the microgrid based on wind energy, photovoltaics, electric vehicles, and multi-form energy storage, and build a multi-energy modeling system: First, wind speed data is collected and a daily wind speed curve is constructed based on the time series. A Weibull wind speed distribution fitting algorithm is used to statistically model the collected wind speed data, extracting shape and scale parameters from the wind speed data. A three-stage wind power output model is then used to divide the wind speed data into shutdown, partial power, and rated power stages based on the wind speed range in the data, thereby mapping wind speed to output power. Furthermore, a linear interpolation algorithm is used to construct the output time series, and the wind speed gap values ​​in the wind speed data are supplemented to construct a complete wind power prediction output curve.

[0020] Next, the irradiance G and temperature T data are collected in real time through the light intensity and battery module temperature sensors, and the current I-voltage V characteristic curve of the photovoltaic cell unit is constructed based on the five-parameter equivalent circuit model. The specific basis is: ; in, It represents the current value generated by the photovoltaic cell under given lighting conditions; It represents the leakage current of the diode in the absence of light; q represents the basic charge carried by an electron; Rs represents the series impedance inside the battery and the connecting line, n represents the ideality factor or quality factor; K represents the Boltzmann constant, and Rsh represents the leakage path impedance inside the photovoltaic cell.

[0021] The maximum power point (MPP) of the photovoltaic cell unit is calculated based on the sampling points at each T1 time, forming a daily photovoltaic processing prediction curve. The maximum power point tracking (MPPT) strategy is further simulated using the perturbation and observation method (P&O) or incremental derivative method to achieve tracking response behavior modeling under dynamic lighting, that is, photovoltaic power generation modeling; Among various energy storage devices, for energy storage units such as lithium batteries and supercapacitors, a state space SOC model is established to express the charging and discharging process of each energy storage unit. The specific basis is: ; in, Indicates the current remaining energy percentage; represents the state of charge at the next moment t+1, Indicates the effective conversion ratio of energy during the charging process; represents the charging power of the energy storage unit at time t, Indicates the length of the scheduling cycle, represents the discharge power of the energy storage unit at time t, Indicates the output conversion rate of energy during the discharge process, Indicates the maximum energy storage capacity of the energy storage unit.

[0022] Furthermore, the charging efficiency parameter nch and discharging efficiency parameter nds as well as the upper and lower power limits of each energy storage unit are set, and dynamic efficiency curves are set for different types of energy storage equipment units. The Rainflow cycle count + capacity decay rate model is introduced, and the energy storage health status SOH of the energy storage equipment is updated according to the number of charge and discharge cycles and the discharge depth DOD of each energy storage unit. Finally, a time series SOC curve and a scheduling available power boundary data table are formed.

[0023] Furthermore, we obtain statistical data on urban traffic behavior and, based on this data, construct a travel probability distribution model for electric vehicles. We use the Poisson distribution and Gaussian mixture model (GMM) to predict the time period when electric vehicles enter the microgrid. We then establish bidirectional power exchange models (V2G and G2V), using the battery SOC as the state variable and adding power constraints and a willingness function: ; Through fuzzy control method, the user participation intention is converted into a control variable to influence the participation power of V2G; Step 2: Based on the completion of the multi-energy modeling system, build a hierarchical, dynamic, and scalable optimization scheduling system; Based on a multi-energy modeling system, operating economy, power volatility, and carbon emission levels are used as multi-objective optimization indicators. By constructing an objective function, key factors such as system power purchase costs, wind and solar curtailment penalties, electric vehicle behavior constraints, energy storage efficiency, and carbon trading prices are jointly considered. The specific steps are as follows: Step S1: define decision variables; Set the scheduling period to T and the time granularity to hours. The defined decision variables include: P(g,t): power purchased from the main grid at time t; P(w,t): wind power output; P(pv,t): photovoltaic output power; P(chev,t),P(disev,t): charging and discharging power of electric vehicles; P(chs,t), P(diss,t): charging and discharging power of energy storage unit; SOC(s,t): energy storage state of charge; P(curt,t): curtailed wind and solar power; P(load,t): total load demand; C(co2,t):carbon emission cost; λt: electricity price; γt: Carbon trading unit price.

[0024] Step S2: constructing a multi-objective function; ; in, Indicates the difference between all power output and load; Indicates the difference between the power output and the load at the previous moment t-1.

[0025] Among them, f1 represents the operating economy objective function; f2 represents the power volatility objective function, P(net, t) represents the net dispatch power; f3 represents the carbon emission objective function; α represents the penalty factor for curtailment of solar and wind power; β represents the penalty factor for the dispatch behavior willingness of electric vehicles; θ(ev, t) represents the willingness function of whether electric vehicles respond to dispatch at time t, and its value is 0 or 1, determined by the actual situation; C(co2, t) = ϵg⋅P(g, t), where ϵg is the carbon emission factor per unit of electricity purchase.

[0026] Step S3: construct constraint conditions; Power balance constraints: P(w,t)+P(pv,t)+P(g,t)+P(diss,t)+P(disev,t)=P(load,t)+P(chs,t)+P(chev,t)+P(cur,t); Energy storage energy balance and SOC evolution: ; Electric vehicle charging and discharging constraints: ; and meet the user connection window; Furthermore, the scheduling process is divided into two levels: day-ahead forecast optimization and intraday rolling adjustment. In the day-ahead stage, the improved NSGA-II algorithm is used to initialize the population, including the scheduling variable set {P(g, t), P(chs, t), P(diss, t), P(chev, t), P(disev, t)}; and the multi-objective functions f1, f2, and f3 are evaluated. Then, the Pareto frontier is obtained based on the congestion distance and fast non-dominated sorting, and a multi-objective Pareto frontier solution search is performed to output a set of feasible scheduling strategy solutions for selection in the intraday stage.

[0027] The intraday phase is based on the rolling time domain optimization of the model predictive control (MPC) framework, setting each rolling cycle to ΔT hours, and predicting the compliance and new energy output for the next H hours. Then, the candidate scheduling solution output by NSGA-II is used as the initialization input, and a short-term linear prediction model is constructed or an LSTM neural network is introduced to perform short-term predictions on P(load, t), P(pv, t), and P(w, t). The local optimization problem is solved, the scheduling variables at the current moment are updated, and finally the final scheduling strategy for the current rolling cycle is output, and the next cycle is promoted.

[0028] However, when a large number of new energy storage units are temporarily incorporated into the microgrid, the original scheduling variable set, SOC state matrix, and energy balance constraints of the multi-energy modeling system will change. MPC cannot synchronize the changes of the multi-energy modeling system, resulting in missing variables in the optimization objective function, causing non-differentiability or divergence of the solution. Therefore, this embodiment first calls the energy device access interface to obtain the current device overview list Snow = {S1, S2, ..., Sn}, records the types and corresponding quantities of various energy storage units such as lithium batteries, sodium-sulfur batteries, and supercapacitors, and obtains the energy storage structure of the previous T2 period from the historical configuration record: Furthermore, the new proportion ρk of each type of energy storage unit Sk is calculated: ; in, represents the number of the kth type of energy storage unit, Indicates the number of energy storage units of type k in the historical records.

[0029] Among them, if the type of energy storage unit is a newly added category, the ρk of this type of energy storage unit is 1; Furthermore, the disturbance coefficient of the overall energy storage structure is calculated as δs: ; Where K represents the total number of energy storage unit categories, wk represents the dispatch impact weight coefficient of the k-th type of energy storage unit; It should be noted that the dispatch impact weight coefficient is determined based on the maximum power capacity and regulation response speed of the energy storage unit of this category. The acquisition method is conventional technical means and will not be elaborated here.

[0030] Furthermore, the weighted sum of the newly added proportion of the new energy storage unit and the energy storage structure disturbance coefficient after the change of the multi-energy body modeling system is used to determine the structural state level reference value Djc of the changed multi-energy body modeling system, specifically according to the formula: Djc=Q1×ρk+Q2×δs, where Q1 represents the weight value of the newly added proportion ρk of the energy storage unit Sk, and Q2 represents the weight value of the disturbance coefficient δs of the energy storage structure; Then, the level reference value threshold Ydjc is set, and the structural state level reference value Djc of the changed multi-energy body modeling system is compared with the level reference value threshold Ydjc. When the structural state level reference value Djc of the changed multi-energy body modeling system is greater than or equal to the level reference value threshold Ydjc, it indicates that the scheduling structure mapping model of the changed multi-energy body modeling system at this level does not need to be adjusted; when the structural state level reference value Djc of the changed multi-energy body modeling system is less than the level reference value threshold Ydjc, it indicates that the scheduling structure mapping model of the changed multi-energy body modeling system at this level needs to use the dynamic mapping function Fmap to regenerate the constraint matrix and the objective function expression to ensure that the optimization model and the scheduling structure are synchronized; It should be noted that the specific application of the dynamic mapping function Fmap in the scheduling structure mapping model is a conventional technical means and will not be described in detail here.

[0031] Furthermore, the number of scheduling structure changes Nchg that occur within the unit time window is obtained, and the scheduling structure change frequency fchg is calculated: ; in, Indicates the number of structural changes within a unit time window; Indicates the statistics window length.

[0032] The change frequency threshold Yth is determined based on the mean and standard deviation of the historical change frequency of the scheduling structure. The structure change frequency fchg is compared with the change frequency threshold Yth. When the structure change frequency fchg is greater than or equal to the change frequency threshold Yth, the standard rolling MPC is still used. When the structure change frequency fchg is less than the change frequency threshold Yth, the linear prediction + streamlined constraint emergency simplification model switch is initiated to ensure that the optimization model remains synchronized with the scheduling structure. Step 3: Build a library of typical extreme scenarios; Using historical meteorological and load data, the following typical extreme event indicators are extracted: Duration of low radiation exposure; frequency of sudden wind stops; Number of hours with daily maximum temperature ≥35°C; Hours of electricity load ≥ 95% of capacity; The Monte Carlo sampling method is used to construct a scenario sample set in combination with the above indicators: ; Among them, each scenario w includes: prediction error disturbance, load abnormal value sequence and ambient temperature influence coefficient; Furthermore, a two-stage robust optimization approach is introduced into the optimization solution: Phase 1: Generate a basic scheduling plan Define the objective function: ; Among them, Cop represents the cost of electricity purchase, start-up and shutdown costs, and the cost of wind and solar power curtailment; Ccarbon represents carbon emissions × carbon trading unit price; Define the constraints: Power balance constraints; Energy storage output constraints; V2G regulation range constraints; Carbon emission cap constraints; And use the improved NSGA-II multi-objective evolutionary algorithm to output the scheduling solution set X1; Phase 2: Optimizing response strategies under extreme disturbances; The CVaR risk control indicator is introduced to evaluate the profitability stability of each basic scheduling solution x∗∈X under the scenario library Ω. First, calculate the average worst-case cost of the conditional value at risk (CVaR) control solution in a high-risk scenario: ; Among them, L(x, ) represents the actual operating cost of the scheduling scheme x in scenario ω, α1 represents the confidence level, η represents the value of the minimum loss bound VaR, and N represents the total number of scenario samples; The CVaR calculation steps are as follows: Step C1: Calculate the operating cost Li under all scenarios ωi∈Ω; Step C2: Sort all Li in ascending order; Step C3: Determine the VaR point corresponding to the confidence level α: η=L[α,N]; Step C4: Calculate the expectation of the part greater than η: ; Furthermore, if severe scenarios such as energy storage failure, photovoltaic failure, and load peak occur under robust optimization, it will have a drastic impact on the robust optimization process. Therefore, in this embodiment, the orderly power consumption mechanism based on load levels first divides the loads into levels according to the importance of power consumption in extreme cases: P (load) = {L1, L2, L3}; Then, a load priority power supply constraint matrix Mload is constructed for constraining response decisions, and energy storage resources and electric vehicles are coordinated to participate in emergency discharge to improve operational resilience.

[0033] Step 4: Operation status evolution analysis; Use InfluxDB or TimescaleDB to build a multi-dimensional time series database platform that supports high-concurrency data writing, historical backtracking, timestamp alignment, and multi-scale aggregation. This platform collects energy flow data such as the output of each energy unit, load demand, and power balance; electricity price, electricity purchase volume, carbon factor, carbon price, and other city electricity and carbon emission data; energy storage and V2G status data such as SOC, charge and discharge status, battery temperature, and battery health; and environmental meteorological data such as light, wind speed, temperature, and humidity. Synchronize the data collected by all acquisition nodes into a unified time series database to generate a data matrix; Use principal component analysis (PCA) to reduce the dimensionality of the data matrix: ; Among them, Z represents the feature matrix after dimensionality reduction, X represents the original data matrix, W represents the principal component projection matrix, X∈Rn×m: original observation data matrix; W∈Rm×k: PCA projection matrix, Z∈Rn×k: data representation after dimensionality reduction; The t-SNE algorithm is used to perform nonlinear visual clustering on the dimensionality reduction data, and different operating states are clustered according to the projection results. At the same time, the DTW (Dynamic Time Warping) algorithm is introduced to analyze the rate of change and similarity of the operating state trajectory: ; Determine whether the current operating state is deviating from the past typical stable operating path and capture the microgrid state mutation signal.

[0034] The above formulas are all dimensionless and numerical calculations. The formulas are obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters in the formulas are set by technicians in this field according to actual conditions.

[0035] The above embodiments may be implemented in whole or in part through software, hardware, firmware or any other combination. When implemented using software, the above embodiments may be implemented in whole or in part in the form of a computer program product.

[0036] Those skilled in the art will appreciate that the modules and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application of the technical solution and the invention constraints. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0037] In addition, each functional module in each embodiment of the present application may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0038] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

[0039] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A multi-objective hierarchical optimization scheduling method for microgrid energy based on dual carbon constraints. It is characterized by including: Step 1: Model the dynamic response characteristics and interactive coupling relationships of key energy units in the microgrid based on wind power, photovoltaics, electric vehicles, and multi-form energy storage; Step 2: Based on the completion of the multi-energy modeling system, an optimization scheduling framework with multi-objective coordination capabilities, a hierarchical decoupling structure, and a dynamic feedback adjustment mechanism is constructed; Step 3: Combine extreme weather events, power grid fluctuations, and abnormal user loads to build a typical extreme scenario event database; Step 4: Construct a time series data set that integrates multi-dimensional state information to dynamically track and model the operation status.

2. The microgrid energy multi-objective hierarchical optimization scheduling method based on dual carbon constraints according to claim 1 is characterized by: In step 1, the collected wind speed data is statistically modeled to construct a wind power forecast output curve; the intraday photovoltaic processing forecast curve is formed through irradiance and temperature data; the charging efficiency parameter nch and discharge efficiency parameter nds of each energy storage unit and the upper and lower power limits are used to form a time series SOC curve and a scheduling available power boundary data table; based on urban traffic behavior statistical data, a travel probability distribution model for electric vehicles is constructed.

3. The microgrid energy multi-objective hierarchical optimization scheduling method based on dual carbon constraints according to claim 2 is characterized by: In step 2, the decision variables are defined, the multi-objective function is constructed, and the constraints are defined; the scheduling process is divided into two levels: day-ahead forecast optimization and intraday rolling adjustment.

4. The microgrid energy multi-objective hierarchical optimization scheduling method based on dual carbon constraints according to claim 3 is characterized by: In step 2, for the day-ahead stage, the population is first initialized, and the multi-objective function is evaluated, and a feasible scheduling strategy solution set is output through multi-objective frontier solution search.

5. The microgrid energy multi-objective hierarchical optimization scheduling method based on dual carbon constraints according to claim 4 is characterized by: In step 2, for the intraday stage, a rolling cycle is set, future output is predicted, short-term prediction is performed, and the current rolling cycle scheduling strategy is output; Obtain the types and corresponding quantities of various energy storage units, calculate the new proportion of each type of energy storage unit, and calculate the disturbance coefficient of the overall energy storage structure; The weighted sum of the newly added proportion of new energy storage units and the energy storage structure disturbance coefficient of the changed multi-energy body modeling system determines the structural state level reference value Djc of the changed multi-energy body modeling system, and sets the level reference value threshold Ydjc. When the structural state level reference value Djc of the changed multi-energy body modeling system is greater than or equal to the level reference value threshold Ydjc, it indicates that the scheduling structure mapping model of the changed multi-energy body modeling system at this level does not need to be adjusted; when the structural state level reference value Djc of the changed multi-energy body modeling system is less than the level reference value threshold Ydjc, it indicates that the scheduling structure mapping model of the changed multi-energy body modeling system at this level needs to be adjusted.

6. The microgrid energy multi-objective hierarchical optimization scheduling method based on dual carbon constraints according to claim 5 is characterized by: In step 2, the number of scheduling structure changes that occur within the unit time window is obtained, the scheduling structure change frequency is calculated, and the structure change frequency fchg is compared with the change frequency threshold Yth. When the structure change frequency fchg is greater than or equal to the change frequency threshold Yth, the standard rolling MPC is still used; When the structure change frequency fchg is less than the change frequency threshold Yth, adjustment is required.

7. The microgrid energy multi-objective hierarchical optimization scheduling method based on dual carbon constraints according to claim 6 is characterized by: In step 3, typical extreme event indicator data are obtained to construct a scenario sample set; Introducing a two-stage robust optimization approach: Phase 1: Generate a basic scheduling plan Phase 2: Optimization of response strategies under extreme disturbances.

8. The microgrid energy multi-objective hierarchical optimization scheduling method based on dual carbon constraints according to claim 7 is characterized by: In step 4, the collected energy flow data, mains power and carbon emission data, energy storage and V2G status data, and environmental meteorological data are used to generate a data matrix and perform dimensionality reduction processing; the rate of change and similarity of the operating status trajectory of the reduced data are analyzed.

9. The microgrid energy multi-objective hierarchical optimization scheduling method based on dual carbon constraints according to claim 6 is characterized by: The method for obtaining the change frequency threshold is as follows: Calculate the mean and standard deviation of the historical change frequency of the scheduling structure, analyze the mean and standard deviation of the historical change frequency of the scheduling structure using K-means clustering, and determine the change frequency threshold.

Images