Multi-objective collaborative optimization method and system for integrated energy system, and storage medium

Abstract

The invention provides a multi-target collaborative optimization method and system for an integrated energy system and a storage medium, and the method comprises the steps: obtaining multi-source data of a target region where the integrated energy system is located, and generating a typical uncertainty scene set based on the multi-source data of the target region where the integrated energy system is located; constructing a dynamic behavior model library; establishing a multi-objective optimization model by taking the equipment capacity in the integrated energy system as a first-stage decision variable and taking the operation strategy of the equipment in the integrated energy system under each uncertainty scene as a second-stage decision variable; solving the multi-objective optimization model by adopting a hierarchical hybrid intelligent optimization algorithm to obtain a Pareto optimal frontier solution set; schemes in a Pareto optimal frontier solution set are used as alternative schemes, a fuzzy analytic hierarchy process is adopted, all the alternative schemes are comprehensively scored and sorted, the optimal scheme which most conforms to the intention of a decision maker is output, and the reliability of multi-target collaborative optimization of the comprehensive energy system is improved.

Description

Technical Field

[0001] This invention relates to the field of energy system optimization, and in particular to a multi-objective collaborative optimization method, system, and storage medium for integrated energy systems. Background Technology

[0002] As the main carriers of energy consumption, cities and industrial parks are the key battlegrounds for achieving this goal, while integrated energy systems (IES, which are energy systems that deeply couple multiple energy sources such as electricity, heat, hydrogen, and transportation) are regarded as the core technological path to achieve efficient, clean, and low-carbon energy transformation in regions.

[0003] However, existing integrated energy system planning and design methods generally have profound flaws. Currently, these methods typically separate investment planning from operational strategies, employing simplified "typical day" simulations and static linear equipment models. This approach essentially ignores the strong coupling relationship between investment decisions and dynamic annual operation—where "planning determines operation, and operation, in turn, influences planning"—and fails to reflect the nonlinear efficiency changes and lifespan degradation physical laws of key equipment such as heat pumps and energy storage batteries under real-world operating conditions.

[0004] Meanwhile, most existing methods are based on deterministic optimization and fail to effectively incorporate the inherent uncertainties of the system's operating environment, such as the fluctuations of renewable energy and the randomness of load demand. They also fail to fully integrate the complex dynamics brought about by emerging technologies, such as the flexibility of electric vehicles (V2G) and the cross-seasonal adjustment capabilities of hydrogen energy storage. This results in planning outcomes that lack a comprehensive consideration of the system's full life-cycle economics and operational robustness. The accuracy of the assessment results is low, failing to reflect the true performance under the influence of multiple uncertainties. Furthermore, they cannot provide decision-makers with a scientific and transparent basis for making complex trade-offs among multiple objectives such as cost, carbon reduction, and reliability. Consequently, the reliability of multi-objective collaborative optimization of integrated energy systems is not high.

[0005] To address the issue of low reliability in multi-objective collaborative optimization of integrated energy systems in existing technologies, this solution is proposed. Summary of the Invention

[0006] To address the problems existing in the prior art, this invention innovatively proposes a multi-objective collaborative optimization method, system, and storage medium for integrated energy systems. This effectively solves the problem of low reliability in multi-objective collaborative optimization of integrated energy systems caused by existing technologies, and effectively improves the reliability of multi-objective collaborative optimization of integrated energy systems.

[0007] The first aspect of this invention provides a multi-objective collaborative optimization method for an integrated energy system, comprising: Acquire multi-source data of the target area where the integrated energy system is located, and generate a set of typical uncertainty scenarios based on the multi-source data of the target area where the integrated energy system is located; Construct a dynamic behavior model library, which includes mathematical models for equipment in an integrated energy system that reflect actual operating characteristics; The equipment capacity in the integrated energy system is used as the first-stage decision variable, and the operating strategies of the equipment in the integrated energy system under various uncertainty scenarios are used as the second-stage decision variables. A multi-objective optimization model is established with the optimization objectives of minimizing the expected value of annualized total cost, minimizing the expected value of carbon emissions, and minimizing the equipment footprint. A hierarchical hybrid intelligent optimization algorithm is used to solve the multi-objective optimization model and obtain the Pareto optimal frontier solution set; Using the solutions in the Pareto optimal frontier as alternatives, the fuzzy hierarchical analysis method is used to comprehensively score and rank all alternatives, and output the optimal solution that best meets the decision-maker's intention.

[0008] A second aspect of the present invention provides a multi-objective collaborative optimization system for an integrated energy system, comprising: The generation module acquires multi-source data of the target area where the integrated energy system is located, and generates a set of typical uncertainty scenarios based on the multi-source data of the target area where the integrated energy system is located; The module constructs a dynamic behavior model library, which includes mathematical models that reflect the actual operating characteristics of equipment in an integrated energy system. A module is established, which takes the equipment capacity of the integrated energy system as the first-stage decision variable and the operation strategy of the equipment in the integrated energy system under various uncertain scenarios as the second-stage decision variable. A multi-objective optimization model is established with the optimization objectives of minimizing the expected value of annualized total cost, minimizing the expected value of carbon emissions, and minimizing the equipment footprint. The solution module employs a hierarchical hybrid intelligent optimization algorithm to solve the multi-objective optimization model and obtain the Pareto optimal frontier solution set. The output module uses the solutions in the Pareto optimal frontier as alternatives, and employs fuzzy hierarchical analysis to comprehensively score and rank all alternatives, outputting the optimal solution that best meets the decision-maker's intentions.

[0009] A third aspect of the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of a multi-objective collaborative optimization method for an integrated energy system as described in the first aspect of the present invention.

[0010] The technical solution adopted in this invention has the following technical effects: 1. The technical solution of this invention acquires multi-source data of the target area where the integrated energy system is located, and generates a set of typical uncertainty scenarios based on the multi-source data of the target area where the integrated energy system is located; constructs a dynamic behavior model library, which includes mathematical models that reflect the actual operating characteristics of the equipment in the integrated energy system; takes the equipment capacity in the integrated energy system as the first-stage decision variable, and takes the operating strategy of the equipment in the integrated energy system under each uncertainty scenario as the second-stage decision variable, and establishes a multi-objective optimization model with the optimization objectives of minimizing the expected value of annualized total cost, minimizing the expected value of carbon emissions, and minimizing the equipment footprint; uses a hierarchical hybrid intelligent optimization algorithm to solve the multi-objective optimization model and obtain the Pareto optimal front solution set; uses the solutions in the Pareto optimal front solution set as alternative solutions, and uses fuzzy hierarchical analysis to comprehensively score and rank all alternative solutions, and outputs the optimal solution that best meets the decision-maker's intention, effectively solving the problem of low reliability of multi-objective collaborative optimization of integrated energy systems caused by existing technologies, and effectively improving the reliability of multi-objective collaborative optimization of integrated energy systems.

[0011] 2. The technical solution of this invention obtains at least one full year of load time-series data and meteorological time-series data for the area to be planned where the integrated energy system is located; obtains data on the technical and economic parameters of equipment in the integrated energy system, time-of-use energy prices, and grid carbon emission factors; obtains travel pattern behavior data of electric vehicle users; identifies uncertain variables affecting the operation of the integrated energy system based on multi-source data of the area to be planned where the integrated energy system is located; performs probability distribution function statistical modeling and correlation analysis on the uncertain variables; generates initial uncertainty scenarios based on Latin hypercube sampling; reduces the generated initial scenario set based on dynamic time warping and K-medoids clustering to extract representative typical scenarios; determines the probability of occurrence of each typical scenario based on the number of each typical scenario and the total number of initial scenarios; and through the random scenario generation technology based on dynamic time warping and K-medoids clustering, the planning scheme can effectively resist the risks brought by multiple uncertainties in the future and has stronger robustness.

[0012] 3. The State of Health (SOH) decay model of the energy storage battery in the technical solution of this invention is: the correlation between the degree of decay of the energy storage battery and calendar decay and cycle decay. By introducing a high-fidelity dynamic model coupled with the state of health decay mechanism, the physical authenticity of the optimization results and the accuracy of economic evaluation are significantly improved. The dynamic behavior model library also includes V2G aggregation response model and hydrogen energy system dynamic model. By seamlessly integrating cutting-edge technologies such as V2G and hydrogen energy into the unified optimization framework, the flexibility potential of the integrated energy system is fully explored.

[0013] 4. The technical solution of this invention constructs a linguistic scale for a fuzzy judgment matrix; based on the linguistic scale of the fuzzy judgment matrix, a fuzzy judgment matrix is ​​constructed; the fuzzy weights of each optimization objective are calculated; the calculated fuzzy weights of each optimization objective are converted into final weight values; the objective values ​​of all alternative schemes are normalized; based on the final weight values ​​of each optimization objective and the normalized objective values ​​of all alternative schemes, the final score of each alternative scheme is determined; the alternative scheme with the highest final score is output as the optimal scheme, reducing discomfort and energy waste caused by the imbalance between air conditioning load supply and demand. By introducing a fuzzy multi-criteria decision-making tool, the subjective preferences of decision-makers are integrated into the decision-making process, realizing scientific and transparent decision-making from massive Pareto solution sets to the optimal compromise scheme; this invention significantly enhances the scientific nature and practical feasibility of engineering decisions while improving the optimality of planning schemes.

[0014] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit the invention. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a flowchart illustrating the method of Embodiment 1 in the present invention; Figure 2 This is a schematic diagram of the structure of an integrated energy system provided in Embodiment 1 of the present invention; Figure 3 This is a flowchart illustrating the hierarchical hybrid intelligent optimization algorithm in step S4 of the method in Embodiment 1 of the present invention. Figure 4 This is a flowchart illustrating the fuzzy hierarchical analysis method used for decision support in step S5 of the method in Embodiment 1 of the present invention. Figure 5 This is a schematic diagram of the system structure in Embodiment 2 of the present invention. Detailed Implementation

[0017] To clearly illustrate the technical features of this solution, the invention will be described in detail below through specific embodiments and in conjunction with the accompanying drawings. The following disclosure provides many different embodiments or examples for implementing different structures of the invention. To simplify the disclosure of the invention, components and arrangements of specific examples are described below. Furthermore, reference numerals and / or letters may be repeated in different examples. This repetition is for simplification and clarity and does not in itself indicate a relationship between the various embodiments and / or arrangements discussed. It should be noted that the components illustrated in the drawings are not necessarily drawn to scale. Descriptions of well-known components, processing techniques, and processes are omitted in this invention to avoid unnecessarily limiting the invention.

[0018] Example 1 like Figure 1 As shown, this invention provides a multi-objective collaborative optimization method for integrated energy systems, comprising: S1, acquire multi-source data of the target area where the integrated energy system is located, and generate a set of typical uncertainty scenarios based on the multi-source data of the target area where the integrated energy system is located; S2, Construct a dynamic behavior model library, which includes mathematical models for equipment in an integrated energy system that reflect actual operating characteristics; S3 takes the equipment capacity in the integrated energy system as the first-stage decision variable and the operation strategy of the equipment in the integrated energy system under various uncertain scenarios as the second-stage decision variable, and establishes a multi-objective optimization model with the optimization objectives of minimizing the expected value of annualized total cost, minimizing the expected value of carbon emissions, and minimizing the equipment footprint. S4 uses a hierarchical hybrid intelligent optimization algorithm to solve the multi-objective optimization model and obtain the Pareto optimal frontier solution set; S5 uses the solutions in the Pareto optimal frontier as alternatives, and employs fuzzy hierarchical analysis to comprehensively score and rank all alternatives, outputting the optimal solution that best meets the decision-maker's intention.

[0019] Specifically, step S1 includes: S101: Obtain at least one full year's load time-series data and meteorological time-series data for the area to be planned where the integrated energy system is located; obtain data on the technical and economic parameters of equipment, time-of-use energy prices, and grid carbon emission factors in the integrated energy system; obtain data on the travel patterns and behaviors of electric vehicle users. Specifically, high-resolution time-series data for at least one full year of the area to be planned will be acquired and processed, including but not limited to meteorological data such as electricity / heat load data, solar radiation intensity, and outdoor temperature; technical and economic parameters of equipment, time-of-use energy prices, and grid carbon emission factors will be acquired; and social behavior data such as the travel patterns of electric vehicle users will be obtained through surveys or historical data analysis. S102, Based on multi-source data of the planned area where the integrated energy system is located, identify the uncertain variables affecting the operation of the integrated energy system; the uncertain variables include the volatility of photovoltaic output, the randomness of load demand, and the grid availability of electric vehicle clusters; Specifically, based on data analysis, key uncertainty variables that have a significant impact on system operation are identified, such as the volatility of photovoltaic output, the randomness of load demand, and the grid availability of electric vehicle clusters.

[0020] S103, Perform statistical modeling and correlation analysis of probability distribution functions for uncertain variables; Specifically, marginal probability distribution fitting involves performing statistical analysis on each uncertain variable based on its historical data to fit its probability distribution function (PDF). For example, a Beta distribution is used to model the normalized solar radiation intensity, and a normal distribution is used to model the load forecast error; this step provides marginal distribution input for Latin hypercube sampling.

[0021] Correlation analysis: To avoid generating scenarios that do not conform to physical laws (e.g., extremely strong solar radiation on a cold winter day), it is necessary to calculate the correlation between various uncertain variables. Uncertain variables with low correlation are removed. For example, a Spearman rank correlation coefficient matrix is ​​calculated and constructed based on historical data. This matrix quantifies the nonlinear dependence between variables and serves as one of the core constraints for sampling.

[0022] S104, generating initial uncertainty scenarios based on Latin hypercube sampling; Specifically, the multiple marginal probability distribution functions and correlation matrices obtained in step S102 are used as inputs to the LHS algorithm (Latin Hypercube Sampling). Given a large initial number of scenarios (e.g., N=1000), the LHS algorithm performs the following operations: First, stratified sampling is performed in the multidimensional probability space to ensure uniform coverage of the probability distribution of each variable; then, correlations between variables are introduced through a transformation based on Cholesky decomposition; finally, the generated correlated probability samples are mapped to specific values ​​of uncertain variables using the inverse transformation sampling method. For time-series variables, this step typically involves first sampling to determine their key features (such as total daily power generation, peak-valley difference, etc.), and then reconstructing a complete 8760-hour time series for the entire year using a pre-defined generative model, thereby generating N high-fidelity initial annual operating scenarios. Each scenario represents a multidimensional time series.

[0023] S105, based on dynamic time warping and K-medoids clustering, reduces the generated initial scene set and extracts representative typical scenes; Specifically, to reduce the computational complexity of subsequent model optimization, the N initial scene sets generated by S104 are reduced to extract K representative typical scenes (e.g., K=20). This step uses the K-medoids clustering algorithm and Dynamic Time Warping (DTW) distance as the core metric for measuring the similarity between scenes. The specific execution flow is as follows: 1) Constructing the DTW distance matrix: Calculate the pairwise DTW distances between the N initial scenes, forming an N×N distance matrix D. For any two multidimensional time series scenes... and DTW distance between them The solution is obtained through dynamic programming. First, a T×T local cost matrix C is constructed, where the element C(i,j) is the vector distance between two scenarios at times i and j, typically the Euclidean distance, i.e. Then, the cumulative cost matrix γ is calculated using the following recursive relation, and its lower right element is the DTW distance:

[0024] After calculation, the complete distance matrix D is obtained, where D ab This value represents the similarity between scene a and scene b. The smaller the value, the more similar the shapes are.

[0025] 2) Perform K-medoids clustering: The K-medoids algorithm is a clustering method that is more suitable for this scenario than K-means because it selects the actual scene as the cluster center (i.e., "medoid").

[0026] a. Initialization: Randomly select K scenes from N initial scenes as the initial set of central bodies.

[0027] b. Assignment steps: For each non-central body scenario, based on the calculated DTW distance matrix D, assign it to the cluster containing the nearest central body.

[0028] c. Update steps: Within each cluster, recalculate the sum of distances from each member scene to all other scenes in that cluster. Select the scene that minimizes this sum of distances as the new centroid of that cluster.

[0029] d. Iteration: Repeat the assignment and update steps until the centroids of all clusters no longer change, or the preset maximum number of iterations is reached, at which point the clustering process converges.

[0030] S106. Determine the probability of occurrence of each typical scenario based on the number of each typical scenario and the total number of initial scenarios.

[0031] Specifically, after clustering convergence, the final K medoids are selected as typical scenarios. The probability of each typical scenario occurring is calculated by dividing the number of scenarios in its cluster by the initial total number of scenarios N. For example, the probability of typical scenario s... for:

[0032] Finally, output these K typical scenarios and their corresponding probabilities. Used to build subsequent two-stage stochastic programming models.

[0033] In step S2, a high-fidelity dynamic behavior model library is constructed, namely, as follows: Figure 2 As shown, mathematical models reflecting the actual operating characteristics of key equipment in an integrated energy system are established, namely solar photovoltaic arrays, electrochemical energy storage stations (BESS), ground source heat pumps, electrolyzers, high-pressure hydrogen storage tanks, and fuel cells.

[0034] Specifically, the dynamic behavior model library includes: heat pump dynamic COP model, energy storage battery state of health (SOH) decay model, V2G aggregation response model, and hydrogen energy system dynamic model; The dynamic COP model of the heat pump is a nonlinear function that describes the performance coefficient COP of the heat pump as dependent on the heat source temperature, the load-side temperature, and the load factor PLR. Specifically, the coefficient of performance (COP) of a heat pump is its core economic indicator, but it is not a constant. This invention establishes a multivariate nonlinear function model to describe its dynamic changes:

[0035] in, Let T be the coefficient of performance of the heat pump at time t. source (t) is the temperature on the heat source side at time t (such as the circulating water temperature in a buried pipe or the outdoor air temperature), T load (t) is the outlet water temperature on the load side at time t, and PLR(t) is the ratio of the current output of the heat pump to its rated power at time t, i.e., the part load factor. The specific form of this function f can be obtained in two ways: (1) Empirical Fitting: Based on performance curve data provided by the equipment manufacturer under different operating conditions, multiple regression analysis (such as quadratic polynomial fitting) is used to obtain the results. For example:

[0036] (2) Thermodynamic modeling: Based on the modified Carnot cycle theory, the model is derived from the physical mechanism. In the optimization model, this nonlinear relationship is processed by the piecewise linearization technique and transformed into a set of linear inequality constraints to facilitate processing by the MILP solver.

[0037] The State of Health (SOH) degradation model for energy storage batteries is as follows: the correlation between the degree of degradation of energy storage batteries and calendar degradation and cycle degradation; Specifically, the State of Health (SOH) degradation model for energy storage batteries is as follows: ; in, This represents the maximum available capacity of the energy storage batteries for the next scheduling cycle. This represents the maximum available capacity of the energy storage batteries during the current scheduling cycle. This represents the annualized calendar degradation rate of the energy storage batteries for the current scheduling cycle. This refers to the capacity loss caused by each full cycle of an energy storage battery. The annual equivalent full cycle count is calculated using the rainflow counting method for energy storage batteries in the current scheduling cycle.

[0038] The BESS full life-cycle capacity degradation model is a major innovation of this invention, as it incorporates the "State of Health" (SOH) of the energy storage battery into the optimization model. The total degradation of the energy storage battery consists of two parts: Calendar aging: Even when a battery is idle, slow chemical reactions occur internally, leading to a decrease in capacity. This is mainly related to temperature and state of charge (SOC). This invention uses a model based on the Arrhenius equation to describe this:

[0039] Where T1 is the battery operating temperature (Kelvin), t is time, A refers to the pre-factor, and E... a R is the activation energy, and R is the ideal gas constant. This model quantifies the capacity loss under idle conditions.

[0040] Cycle Aging: Battery charge-discharge cycles are the primary cause of capacity degradation, and the degree of degradation is strongly correlated with depth of cycle (DOD) and charge-discharge rate (C-rate). To accurately quantify this, this invention introduces a rainflow counting algorithm in the lower-level optimization evaluation. This algorithm can identify and extract all complete and half-cycles and their corresponding DODs from an irregular charge-discharge power or SOC time series. Then, based on the DOD-lifetime curve provided by the battery manufacturer, which is typically in power-law form, for example...

[0041] Where α and β are battery characteristic coefficients, which convert the number of cycles at different depths into the equivalent full cycle number (EFC) at 100% DOD.

[0042] Status Update: At the end of each scheduling cycle (e.g., one year), the battery's maximum available capacity Cmax (i.e., SOH) will be updated based on calendar degradation and accumulated EFC. ; Here, `Losscyc_per_EFC` represents the capacity loss caused by each full loop. This dynamically updated C... max This will serve as a core basis for evaluating the battery replacement cycle in the next year's scheduling or investment cost calculation, thereby closely linking the operating strategy with long-term investment costs.

[0043] It should be noted that the annualized calendar decay rate Loss in the formula cal_rate Compared with the calendar decay model Loss described above cal (%) are not the same parameter, but they are closely related. Loss cal It is the fundamental model describing the physical process of decay, while Loss cal_rate This is an annual decay rate parameter calculated based on this model. Specifically, it is obtained by calculating the annual decay rate within the Loss range. cal In the (%) model, time t is set to one year, and a representative annual average temperature T1 is substituted to calculate the cumulative capacity loss percentage within that year's time step. This calculation result is the Loss. cal_rate This approach discretizes the continuous physical decay process into an annual decay rate, making it compatible with the upper-level planning model in this invention, which uses an annual step size.

[0044] The key point here is that, for battery energy storage systems (BESS), the device lifespan L... BESS It is not a fixed input parameter, but rather C as described in step S200.max The dynamic update process is intrinsically determined. Specifically, C max The dynamic update process simulates the battery capacity degradation trajectory under year-on-year operation. In the optimization model of this invention, when the simulated C... max The number of years elapsed until the value first drops to a preset scrap threshold (e.g., 80% of the initial capacity) is determined as the actual lifespan L of the energy storage battery under this operating strategy. BESS .

[0045] Therefore, "dynamically updated C" max "It is precisely by determining the actual lifespan (LBESS) of the energy storage battery that the value of the capital recovery factor (CRFi) is affected, and ultimately, through the calculation formula of CREP, its cost impact is quantitatively reflected in the total cost objective function f1. This clearly demonstrates that the dynamic behavior model of Cmax directly participates in the calculation of the objective function, constituting the core technical means of this invention that tightly couples the operating strategy with long-term investment costs."

[0046] In the post-processing stage of lower-level scheduling optimization, the scheduling results for each scenario are used to accurately calculate the annual equivalent full cycle count (EFC(y)) of the battery using the rainflow counting method. The battery replacement cycle (actual lifespan of the energy storage battery) L BESS Based on calendar lifespan L cal and cycle life L cyc Dynamic calculation, that is:

[0047] Where L cyc The calculation formula is:

[0048] in, , That is, the annual equivalent full cycle count for scene s in scene set S, and This is the total design cycle life. The dynamic life L... BESS It is directly used for calculating the annualized replacement cost of the upper-level planning model.

[0049] Regarding parameter L cal The calendar life of an energy storage battery is explained as follows: it refers to the end of the battery's lifespan due to the passage of time and the natural aging of its internal materials, without considering or with minimal consideration of charge-discharge cycles. This parameter is a standard technical parameter provided by energy storage battery manufacturers in their product specifications or technical manuals. It is usually expressed in "years," for example, a battery's calendar life might be 10 or 15 years.

[0050] The V2G aggregation response model is an aggregation model that describes the dynamic charging and discharging power and energy capacity that a V2G cluster can provide at any given time. Specifically, the behavior of a single EV (electric vehicle) is random, but a cluster of EVs exhibits predictable statistical patterns.

[0051] Spatiotemporal distribution generation: Using the Markov Chain Monte Carlo (MCMC) method, based on the EV user survey data (vehicle type, battery capacity, commuting time, daily mileage, etc.) collected in step S101, the spatiotemporal distribution of a large-scale EV fleet (e.g., 1000 vehicles) is simulated for each day of the year. The simulation process defines vehicle states (e.g., "on the way," "at home," "in the park") and establishes a state transition probability matrix based on statistical data. The MCMC simulation will generate, at any time t, whether each vehicle is on the way, at home, or in the park, and its initial SOC upon arrival at the park.

[0052] Aggregate Model Construction: Based on the simulation results above, the set of vehicles Vconn(t) connected to the V2G charging piles in the park at any time t, and their SOC distribution, can be statistically determined. Therefore, an aggregated battery model can be established, which describes the total up / down power capacity and energy capacity that the entire V2G cluster can provide at time t. Maximum charging power:

[0053] Maximum discharge power:

[0054] Maximum rechargeable capacity:

[0055] Maximum discharge capacity:

[0056] Among them, SOC min_user These are the minimum SOC thresholds set by the user to ensure subsequent travel. These dynamically changing upper and lower limits will be incorporated as constraints into the IES's optimization scheduling model. It is the maximum charging power allowed by the electric vehicle itself or by the charging station it is connected to. The maximum allowable discharge power of the on-board inverter of the electric vehicle itself; It is the highest state of charge (SOC) target value that is allowed to be charged, set to protect battery health or according to grid requirements; It is the minimum state of charge (SOC) threshold set by electric vehicle users to ensure their subsequent travel needs; This refers to the state of charge of the electric vehicle v at time t. This refers to the rated capacity of the battery in an electric vehicle, representing the total amount of electricity that its battery can store. Let v be the set of electric vehicles.

[0057] The dynamic model of the hydrogen energy system is as follows: the hydrogen production efficiency of the electrolyzer is a nonlinear function model with respect to the input power, and the power generation efficiency of the fuel cell is a nonlinear function model with respect to the output power.

[0058] Specifically, the electrolyzer's hydrogen production efficiency is not constant, but rather a nonlinear function of the input power. This invention uses a polynomial function to fit its efficiency curve:

[0059] The hydrogen production rate is

[0060] Where b0, b1, and b3 are the quadratic fitting coefficients of the electrolyzer efficiency model, which are constants determined based on the equipment characteristics; LHV refers to the input electrical power consumed by the electrolyzer at scheduling time t; H2 It is the low calorific value of hydrogen. Let be the hydrogen production efficiency of the electrolyzer at scheduling time t.

[0061] Hydrogen Tank: Its state is determined by the internal hydrogen pressure p H2 (t) Characterization. Pressure changes follow the discrete form of the ideal gas law and are related to the mass flow rate of hydrogen being stored / retrieved:

[0062] And is subject to maximum / minimum safety pressure limits.

[0063] in, , These are the absolute pressures of hydrogen inside the hydrogen storage tank at the start and end of scheduling time t (i.e., at the start of time t); It is the mass flow rate of hydrogen injected into the hydrogen storage tank at scheduling time t; It is the mass flow rate of hydrogen gas taken out from the hydrogen storage tank at scheduling time t; It is the specific gas constant of hydrogen; This is the temperature inside the hydrogen storage tank, which is assumed to be a constant value in this model. It is the fixed volume of the hydrogen storage tank.

[0064] Fuel Cell: Power Generation Efficiency Similarly, since it is a nonlinear function of output power, it is modeled in a manner similar to that of an electrolytic cell:

[0065] hydrogen consumption rate for

[0066] This refers to the output electrical power of the fuel cell at scheduling time t; This refers to the power generation efficiency of the fuel cell at the scheduling time t; c0, c1, and c2 are the quadratic fitting coefficients of the fuel cell efficiency model.

[0067] In step S3, a two-stage multi-objective stochastic programming model is constructed. Equipment capacity is used as the first-stage decision variable, and the operating strategies under various uncertain scenarios are used as the second-stage decision variables. A multi-objective mathematical model is established with the optimization objectives of minimizing the expected annualized total cost, minimizing the expected carbon emissions, and minimizing the equipment footprint.

[0068] Decision Variables: First-stage decision variable (Capacity Planning): X e (The installed capacity of various types of equipment is a continuous or integer variable, for example, X) pv , Xbess_power, Xbess_energy.

[0069] The second stage variables (Operation Scheduling) include P(t,s), Q(t,s), G(t,s), and u(t,s), which represent the power, heat flow, gas flow rate, and equipment start-up / shutdown status (0-1 variables) at time t in scenario s, respectively.

[0070] It should be noted that the scenario indices s, a, and b used in the article all belong to the same set of uncertain scenarios, but their usage is slightly different: s is a general index that refers to any scenario; while a and b are usually used in pairs to describe the unexpected constraint relationship between different scenarios.

[0071] It should be noted that in this invention, i, j, and t are all used as time indexes, where t usually refers to the current or independent scheduling time, while i and j are mainly used as auxiliary indexes for traversing or summing time periods in mathematical expressions.

[0072] Specifically, the multi-objective optimization model includes: minimizing the expected total cost, minimizing the expected carbon emissions, and minimizing the equipment footprint; The specific method for calculating the minimum expected value of total cost is as follows:

[0073] in: The expected value of total cost. This represents the annualized initial investment cost. The annualized replacement cost is... To fix maintenance costs, The expected value of variable operating costs; The specific method for calculating the minimum expected value of carbon emissions is as follows:

[0074] in, To minimize the expected carbon emissions, Let be the probability of occurrence of typical scenario s in the set of typical scenarios S. This represents the power purchased by scenario s at time t. Let T be the real-time marginal carbon emission factor of the power grid at time t, where T is the set of time steps throughout the year. This represents the natural gas consumption at time t in a typical scenario. Carbon emission factors of natural gas; The specific method for calculating the minimum equipment footprint is as follows:

[0075] in, For the equipment's footprint, A e Let $e$ be the area occupied per unit capacity of device $e$ in device set $E$. Let e ​​be the capacity of device e in device set E.

[0076] Specifically, annualized initial investment cost The specific calculation method is as follows: ; in, It is the capital recovery factor of equipment e, I e X is the unit investment cost of equipment e. e This refers to the installation capacity of device e. The calculation formula is: ; Where d is the discount rate. This refers to the actual lifespan of the equipment; Annualized replacement cost The specific calculation method is as follows: ; in, It is the total present value of all future reset events. It refers to the entire project planning cycle. Capital recovery factor; total present value of all future reset events. The calculation formula is:

[0077] in, It is a collection of equipment that needs to be replaced during the project cycle. It is the sequence number of the reset event. This is the total number of times equipment e is replaced during the project cycle; the calculation formula is: (Round down).

[0078] For the entire project planning cycle Capital recovery factor The calculation formula is: ; Fixed maintenance costs The specific calculation method is as follows: ; in, The annual fixed operation and maintenance cost per unit capacity of equipment e; Expected variable operating costs The specific calculation method is as follows:

[0079] in, Let be the fuel cost at time t in scenario s. Let t be the grid interaction cost (electricity purchase cost - electricity sales revenue) at time t in scenario s. For the variable operation and maintenance cost at time t in scenario s, The cost of V2G scheduling compensation at time t in scenario s.

[0080] In step S4, a hierarchical hybrid intelligent optimization algorithm is used to solve the problem. A master-slave algorithm framework is designed: the upper layer (master problem) uses the NSGA-III multi-objective evolutionary algorithm to search for the optimal equipment capacity configuration scheme; the lower layer (sub-problem) uses commercial solvers (such as Gurobi, CPLEX) to solve for the optimal operating strategy and performance indicators under all typical uncertainty scenarios for each capacity scheme given by the upper layer, and feeds the results back to the upper layer as fitness for iterative evolution until convergence to the Pareto optimal frontier.

[0081] The constraints of the multi-objective optimization model include the balance constraints of electricity, heat, and hydrogen multi-energy flow under various typical scenarios, the constraints of the dynamic behavior model of equipment, and the annual periodic constraints of the energy storage system.

[0082] Among them, the multi-energy balance constraint is that for each scenario s at each time t, the supply and demand of the electricity, heat, and hydrogen networks must be balanced.

[0083] Power balance:

[0084] in, This represents the basic power load demand under scenario s at time t. Let be the electrical power consumed by the heat pump in scenario s at time t; It is the charging power of the battery energy storage system (BESS) under scenario s at time t; It is the charging power of the electric vehicle (V2G) cluster under scenario s at time t; It is the electrical power consumed by the electrolyzer for hydrogen production in scenario s at time t; It is the power generation of the photovoltaic (PV) system under scenario s at time t; It is the electrical power purchased from the main grid at time t in scenario s; It is the electrical power sold to the main grid at time t in scenario s; It is the discharge power of the battery energy storage system (BESS) under scenario s at time t; P is the discharge power of the electric vehicle (V2G) cluster in scenario s at time t; fc (t,s) represents the power output of the fuel cell at time t in scenario s.

[0085] Thermal equilibrium:

[0086] in, It is the heat power generated by the heat pump in scenario s at time t; It is the thermal power generated by the gas-fired boiler in scenario s at time t; It is the heat release power of the thermal storage device under scenario s at time t; It is the thermal storage power of the thermal storage device under scenario s at time t.

[0087] Hydrogen balance:

[0088] in, It is the hydrogen production rate of the fuel cell in scenario s at time t. It is the hydrogen consumption rate of the fuel cell under scenario s at time t; The rate at which hydrogen flows into the hydrogen storage tank at time t in scenario s; It is the rate at which hydrogen flows out of the hydrogen storage tank at time t in scenario s.

[0089] Device Operation Constraints: The output of all equipment must not exceed its installed capacity (Xe) and must meet its minimum output, ramp rate, and other restrictions.

[0090]

[0091] in, It is the actual output power of the photovoltaic system under scenario s at time t; It is the available power factor of a unit capacity photovoltaic system under scenario s at time t; It is the total installed capacity of the photovoltaic system under scenario s at time t.

[0092]

[0093]

[0094]

[0095] in, and These are the actual charging / discharging power of the BESS energy storage battery under scenario s at time t; and These are the binary variables representing the charging / discharging states of the BESS at time t and scenario s, respectively. and These are the minimum and maximum charging power of BESS in scenario s at time t, respectively. and These represent the minimum and maximum discharge power of BESS at time t in scenario s, respectively.

[0096] P bess_ch_max and P bess_dis_max With capacity variable X bess_power Related.

[0097] Dynamic model constraints: The mathematical models (such as changing COP and decaying battery capacity) of the equipment in the integrated energy system established in the dynamic behavior model in step S2 that reflect the actual operating characteristics are embedded as constraints.

[0098]

[0099] in, It is the heat power generated by the heat pump in scenario s at time t. Let be the electrical power consumed by the heat pump in scenario s at time t. Let be the coefficient of performance (COP) of the heat pump in scenario s at time t.

[0100] In step S4, such as Figure 3 To solve the multi-objective optimization model, this step designs a master-slave algorithm framework. Optionally, the upper layer (master problem) uses the NSGA-III multi-objective evolutionary algorithm to search for the optimal equipment capacity configuration scheme (individual). For each capacity scheme given by the upper layer, the lower layer (sub-problem) uses a commercial solver (such as Gurobi) to solve in parallel for its optimal operating strategy and performance indicators under all 20 typical uncertainty scenarios, and feeds back the expected values ​​of the three objective functions as fitness to the upper layer.

[0101] Optionally, by setting the number of iterations (e.g., 200 generations) or a convergence criterion, the algorithm will eventually output a set containing Pareto optimal solutions.

[0102] In step S5, for each solution in the set of Pareto optimal solutions, it is neither "all satisfy" nor "satisfy any one" of the objective functions, but rather achieves an optimal trade-off among the three objective functions (f1: cost, f2: carbon emissions, f3: area).

[0103] Specifically, step S5 includes: S501, Construct the linguistic scale of the fuzzy judgment matrix; Specifically, such as Figure 4 As shown, step S5 aims to select the solution that best suits the decision-maker's preferences from the Pareto optimal solution set. Fuzzy AHP can be used. First, solutions on the Pareto optimal front are considered as alternatives. Then, university decision-makers are invited to compare each optimization objective pairwise (e.g., using fuzzy language such as "cost is slightly more important than environmental protection"). This process follows the standard steps of Fuzzy AHP: First, a linguistic scale is defined to map a decision-maker's fuzzy linguistic judgment to a specific "triangular fuzzy number." The triangular fuzzy number is denoted by (l, m, u), representing the minimum, most likely, and maximum possible values ​​of a fuzzy concept, respectively. A commonly used scale is shown in Table 1 below: Table 1: Commonly Used Scales

[0104] If the judgment is "j1 is more important than i1", then its fuzzy number is the reciprocal of the corresponding value above, i.e. (1 / u, 1 / m, 1 / l).

[0105] S502, constructing a fuzzy judgment matrix based on the linguistic scale of the fuzzy judgment matrix; Decision-makers are invited to make pairwise comparisons of the three optimization objectives of this invention: (f1: cost, f2: carbon emissions, f3: area). For example: Comparing "cost" and "carbon emissions": Policymakers consider "cost slightly more important than carbon emissions." According to the table above, this corresponds to the triangular fuzzy number (1, 3, 5).

[0106] Comparing "cost" and "area": ​​The decision-maker believes that "cost is significantly more important than area." This corresponds to the triangular fuzzy number (3, 5, 7). It is important to emphasize that in this method, fuzzy language such as "slightly important" or "significantly important" is the raw input for the decision-maker's judgment, rather than a conclusion drawn from pre-set weighted differences. The purpose of this method is precisely to transform these qualitative, fuzzy judgments into quantitative weighting coefficients through subsequent mathematical calculations.

[0107] Comparing "carbon emissions" and "area": ​​policymakers consider "both equally important." This corresponds to the triangular fuzzy number (1, 1, 1).

[0108] Based on these comparison results, a 3x3 fuzzy judgment matrix can be constructed.

[0109] S503, calculate the fuzzy weights of each optimization objective; Based on the constructed fuzzy judgment matrix, standard calculation methods in fuzzy hierarchical analysis (such as "fuzzy extended analysis" or "geometric mean") are used to calculate the fuzzy weights of each optimization objective. (o represents the target, and i1 and j1 represent two different targets when comparing weights). This calculation process is a mature mathematical step, the core idea of ​​which is to synthesize all the judgment information in the matrix to obtain a quantitative result of the relative importance of each target.

[0110] S504, convert the calculated fuzzy weights of each optimization objective into final weight values; The calculated fuzzy weights for each target Transform it into a clear, single weight value. The most commonly used method is the "center of gravity method," and the calculation formula is: The deblurred weight values ​​are normalized until their sum equals 1, resulting in the final comprehensive weight vector. After normalization, a set of explicit weight coefficients is obtained, for example: W. 成本 =0.5, W 碳排 =0.3, W 面积 =0.2.

[0111] S505, normalizes the target values ​​of all alternative solutions; The objective values ​​of the alternative solutions refer to the three objective function values ​​corresponding to each solution in the Pareto optimal solution set calculated by the multi-objective optimization algorithm in step S4.

[0112] The specific explanation is as follows: 1) Source of "alternative solutions": Each "alternative solution" is an independent, non-dominated system planning scheme from the Pareto optimal solution set. For example, assuming step S4 outputs 50 Pareto optimal solutions, then there are 50 alternative solutions.

[0113] 2) Composition of the “Target Value”: For each alternative (let k represent the k-th alternative), its “target value” is a vector containing three specific values ​​(f1(k), f2(k), f3(k)). These three values ​​are: the expected annualized total cost of the alternative; the expected annualized total carbon emissions of the alternative; and the total equipment footprint of the alternative.

[0114] 3) In the process of finding the optimal planning scheme, these three objective values ​​have been calculated for each scheme solution.

[0115] The specific calculation method for obtaining the comprehensive score of each scheme by weighted summation based on the weight vector follows the Simple Additive Weighting (SAW) method, which includes the following two key steps: 1) Because the three objectives (cost, carbon emissions, and area) have different dimensions and numerical ranges, direct weighting would lead to the result being dominated by the objective with the larger value. Therefore, it is necessary to first normalize the objective value of each alternative. For the cost-type objective in this invention (i.e., the smaller the value, the better), the following linear normalization formula is used to normalize each objective value f. ko (Solution k, objective o) is transformed into a dimensionless score between 0 and 1. The higher the score, the better:

[0116] in It is the normalized score of scheme k on target o; f ko It is the original value of scheme k on target o; f o,max and f o,min These are the maximum and minimum values ​​among all possible solutions.

[0117] S506, Determine the final score of each alternative solution based on the final weight value of each optimization objective and the normalized objective value of all alternative solutions; Specifically, the normalized score obtained from S505 Compared with the comprehensive weight vector calculated previously using fuzzy hierarchical analysis. Combined, calculate the final score S for each alternative solution k. k :

[0118] Where S k W is the final score of alternative solution k; o It is the weight of the target o.

[0119] S507 outputs the alternative with the highest final score as the optimal solution.

[0120] Preferably, the entire Pareto frontier can be presented to users in the form of a three-dimensional scatter plot or a parallel coordinate graph, allowing users to intuitively see the trade-offs between different solutions in terms of cost, carbon reduction, and land occupation.

[0121] Next, users input their subjective preferences through a visual interface (e.g., by dragging a slider or selecting language options such as "cost is slightly more important than carbon reduction"). The built-in Fuzzy AHP model then transforms the user's qualitative preferences into quantitative target weights and provides a comprehensive score for all alternatives.

[0122] The highest-rated solution is marked as the "Recommended Solution," and its detailed information, including specific equipment selection and capacity, expected annualized cost, carbon emissions, footprint, and operational strategies in typical scenarios, is sent back to the visualization interface. The visualization interface then presents the final recommended solution to the user in a well-structured, illustrated comprehensive report, providing a scientific basis for their practical engineering decisions.

[0123] This invention acquires multi-source data of the target area where the integrated energy system is located, and generates a set of typical uncertainty scenarios based on this data. A dynamic behavior model library is constructed, including mathematical models reflecting the actual operating characteristics of equipment within the integrated energy system. The equipment capacity in the integrated energy system is used as the first-stage decision variable, and the operating strategies of the equipment under each uncertainty scenario are used as the second-stage decision variables. A multi-objective optimization model is established with the optimization objectives of minimizing the expected annualized total cost, minimizing the expected carbon emissions, and minimizing the equipment footprint. A hierarchical hybrid intelligent optimization algorithm is used to solve the multi-objective optimization model, obtaining a Pareto optimal front solution set. The solutions in the Pareto optimal front solution set are used as candidate solutions. Fuzzy hierarchical analysis is used to comprehensively score and rank all candidate solutions, outputting the optimal solution that best meets the decision-maker's intentions. This effectively solves the problem of low reliability in multi-objective collaborative optimization of integrated energy systems caused by existing technologies, and effectively improves the reliability of multi-objective collaborative optimization of integrated energy systems.

[0124] The technical solution of this invention acquires at least one full year of load time-series data and meteorological time-series data for the area to be planned where the integrated energy system is located; acquires data on the technical and economic parameters of equipment in the integrated energy system, time-of-use energy prices, and grid carbon emission factors; acquires travel pattern behavior data of electric vehicle users; identifies uncertain variables affecting the operation of the integrated energy system based on multi-source data of the area to be planned where the integrated energy system is located; performs probability distribution function statistical modeling and correlation analysis on the uncertain variables; generates initial uncertainty scenarios based on Latin hypercube sampling; reduces the generated initial scenario set based on dynamic time warping and K-medoids clustering to extract representative typical scenarios; determines the probability of occurrence of each typical scenario based on the number of each typical scenario and the total number of initial scenarios; and through the random scenario generation technology based on dynamic time warping and K-medoids clustering, the planning scheme can effectively resist the risks brought by multiple uncertainties in the future and has stronger robustness.

[0125] The state of health (SOH) degradation model of the energy storage battery in this invention is as follows: the degree of degradation of the energy storage battery is related to calendar degradation and cycle degradation. By introducing a high-fidelity dynamic model coupled with the state of health degradation mechanism, the physical authenticity of the optimization results and the accuracy of economic evaluation are significantly improved. The dynamic behavior model library also includes a V2G aggregation response model and a hydrogen energy system dynamic model. By seamlessly integrating cutting-edge technologies such as V2G and hydrogen energy into a unified optimization framework, the flexibility potential of the integrated energy system is fully explored.

[0126] This invention constructs a linguistic scale for a fuzzy judgment matrix; based on the linguistic scale, a fuzzy judgment matrix is ​​constructed; fuzzy weights for each optimization objective are calculated; the calculated fuzzy weights for each optimization objective are converted into final weight values; the objective values ​​of all candidate schemes are normalized; based on the final weight values ​​of each optimization objective and the normalized objective values ​​of all candidate schemes, the final score of each candidate scheme is determined; the candidate scheme with the highest final score is output as the optimal scheme. This reduces discomfort and energy waste caused by the imbalance between air conditioning load supply and demand. By introducing a fuzzy multi-criteria decision-making tool, the decision-maker's subjective preferences are integrated into the decision-making process, achieving scientific and transparent decision-making from massive Pareto solution sets to the optimal compromise scheme. This invention significantly enhances the scientific nature and practical feasibility of engineering decisions while improving the optimality of planning schemes.

[0127] Example 2 like Figure 5 As shown, the present invention also provides a multi-objective collaborative optimization system for integrated energy systems, comprising: The generation module 101 acquires multi-source data of the target area where the integrated energy system is located, and generates a set of typical uncertainty scenarios based on the multi-source data of the target area where the integrated energy system is located; Module 102 is used to build a dynamic behavior model library, which includes mathematical models that reflect the actual operating characteristics of equipment in an integrated energy system. Module 103 is established, taking the equipment capacity of the integrated energy system as the first-stage decision variable and the operation strategy of the equipment in the integrated energy system under various uncertain scenarios as the second-stage decision variable. A multi-objective optimization model is established with the optimization objectives of minimizing the expected value of annualized total cost, minimizing the expected value of carbon emissions, and minimizing the equipment footprint. Solver module 104 uses a hierarchical hybrid intelligent optimization algorithm to solve the multi-objective optimization model and obtain the Pareto optimal frontier solution set; Output module 105 uses the solutions in the Pareto optimal frontier solution set as alternatives, and employs fuzzy hierarchical analysis to comprehensively score and rank all alternatives, outputting the optimal solution that best meets the decision-maker's intention.

[0128] It should be noted that the implementation process of the generation module 101, construction module 102, establishment module 103, solution module 104, and output module 105 in this embodiment corresponds to the method steps in embodiment one, and will not be repeated here.

[0129] This invention acquires multi-source data of the target area where the integrated energy system is located, and generates a set of typical uncertainty scenarios based on this data. A dynamic behavior model library is constructed, including mathematical models reflecting the actual operating characteristics of equipment within the integrated energy system. The equipment capacity in the integrated energy system is used as the first-stage decision variable, and the operating strategies of the equipment under each uncertainty scenario are used as the second-stage decision variables. A multi-objective optimization model is established with the optimization objectives of minimizing the expected annualized total cost, minimizing the expected carbon emissions, and minimizing the equipment footprint. A hierarchical hybrid intelligent optimization algorithm is used to solve the multi-objective optimization model, obtaining a Pareto optimal front solution set. The solutions in the Pareto optimal front solution set are used as candidate solutions. Fuzzy hierarchical analysis is used to comprehensively score and rank all candidate solutions, outputting the optimal solution that best meets the decision-maker's intentions. This effectively solves the problem of low reliability in multi-objective collaborative optimization of integrated energy systems caused by existing technologies, and effectively improves the reliability of multi-objective collaborative optimization of integrated energy systems.

[0130] The technical solution of this invention acquires at least one full year of load time-series data and meteorological time-series data for the area to be planned where the integrated energy system is located; acquires data on the technical and economic parameters of equipment in the integrated energy system, time-of-use energy prices, and grid carbon emission factors; acquires travel pattern behavior data of electric vehicle users; identifies uncertain variables affecting the operation of the integrated energy system based on multi-source data of the area to be planned where the integrated energy system is located; performs probability distribution function statistical modeling and correlation analysis on the uncertain variables; generates initial uncertainty scenarios based on Latin hypercube sampling; reduces the generated initial scenario set based on dynamic time warping and K-medoids clustering to extract representative typical scenarios; determines the probability of occurrence of each typical scenario based on the number of each typical scenario and the total number of initial scenarios; and through the random scenario generation technology based on dynamic time warping and K-medoids clustering, the planning scheme can effectively resist the risks brought by multiple uncertainties in the future and has stronger robustness.

[0131] The state of health (SOH) degradation model of the energy storage battery in this invention is as follows: the degree of degradation of the energy storage battery is related to calendar degradation and cycle degradation. By introducing a high-fidelity dynamic model coupled with the state of health degradation mechanism, the physical authenticity of the optimization results and the accuracy of economic evaluation are significantly improved. The dynamic behavior model library also includes a V2G aggregation response model and a hydrogen energy system dynamic model. By seamlessly integrating cutting-edge technologies such as V2G and hydrogen energy into a unified optimization framework, the flexibility potential of the integrated energy system is fully explored.

[0132] This invention constructs a linguistic scale for a fuzzy judgment matrix; based on the linguistic scale, a fuzzy judgment matrix is ​​constructed; fuzzy weights for each optimization objective are calculated; the calculated fuzzy weights for each optimization objective are converted into final weight values; the objective values ​​of all candidate schemes are normalized; based on the final weight values ​​of each optimization objective and the normalized objective values ​​of all candidate schemes, the final score of each candidate scheme is determined; the candidate scheme with the highest final score is output as the optimal scheme. This reduces discomfort and energy waste caused by the imbalance between air conditioning load supply and demand. By introducing a fuzzy multi-criteria decision-making tool, the decision-maker's subjective preferences are integrated into the decision-making process, achieving scientific and transparent decision-making from massive Pareto solution sets to the optimal compromise scheme. This invention significantly enhances the scientific nature and practical feasibility of engineering decisions while improving the optimality of planning schemes.

[0133] Example 3 The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of a multi-objective collaborative optimization method for an integrated energy system as described in Embodiment 1.

[0134] This invention acquires multi-source data of the target area where the integrated energy system is located, and generates a set of typical uncertainty scenarios based on this data. A dynamic behavior model library is constructed, including mathematical models reflecting the actual operating characteristics of equipment within the integrated energy system. The equipment capacity in the integrated energy system is used as the first-stage decision variable, and the operating strategies of the equipment under each uncertainty scenario are used as the second-stage decision variables. A multi-objective optimization model is established with the optimization objectives of minimizing the expected annualized total cost, minimizing the expected carbon emissions, and minimizing the equipment footprint. A hierarchical hybrid intelligent optimization algorithm is used to solve the multi-objective optimization model, obtaining a Pareto optimal front solution set. The solutions in the Pareto optimal front solution set are used as candidate solutions. Fuzzy hierarchical analysis is used to comprehensively score and rank all candidate solutions, outputting the optimal solution that best meets the decision-maker's intentions. This effectively solves the problem of low reliability in multi-objective collaborative optimization of integrated energy systems caused by existing technologies, and effectively improves the reliability of multi-objective collaborative optimization of integrated energy systems.

[0135] The technical solution of this invention acquires at least one full year of load time-series data and meteorological time-series data for the area to be planned where the integrated energy system is located; acquires data on the technical and economic parameters of equipment in the integrated energy system, time-of-use energy prices, and grid carbon emission factors; acquires travel pattern behavior data of electric vehicle users; identifies uncertain variables affecting the operation of the integrated energy system based on multi-source data of the area to be planned where the integrated energy system is located; performs probability distribution function statistical modeling and correlation analysis on the uncertain variables; generates initial uncertainty scenarios based on Latin hypercube sampling; reduces the generated initial scenario set based on dynamic time warping and K-medoids clustering to extract representative typical scenarios; determines the probability of occurrence of each typical scenario based on the number of each typical scenario and the total number of initial scenarios; and through the random scenario generation technology based on dynamic time warping and K-medoids clustering, the planning scheme can effectively resist the risks brought by multiple uncertainties in the future and has stronger robustness.

[0136] The state of health (SOH) degradation model of the energy storage battery in this invention is as follows: the degree of degradation of the energy storage battery is related to calendar degradation and cycle degradation. By introducing a high-fidelity dynamic model coupled with the state of health degradation mechanism, the physical authenticity of the optimization results and the accuracy of economic evaluation are significantly improved. The dynamic behavior model library also includes a V2G aggregation response model and a hydrogen energy system dynamic model. By seamlessly integrating cutting-edge technologies such as V2G and hydrogen energy into a unified optimization framework, the flexibility potential of the integrated energy system is fully explored.

[0137] This invention constructs a linguistic scale for a fuzzy judgment matrix; based on the linguistic scale, a fuzzy judgment matrix is ​​constructed; fuzzy weights for each optimization objective are calculated; the calculated fuzzy weights for each optimization objective are converted into final weight values; the objective values ​​of all candidate schemes are normalized; based on the final weight values ​​of each optimization objective and the normalized objective values ​​of all candidate schemes, the final score of each candidate scheme is determined; the candidate scheme with the highest final score is output as the optimal scheme. This reduces discomfort and energy waste caused by the imbalance between air conditioning load supply and demand. By introducing a fuzzy multi-criteria decision-making tool, the decision-maker's subjective preferences are integrated into the decision-making process, achieving scientific and transparent decision-making from massive Pareto solution sets to the optimal compromise scheme. This invention significantly enhances the scientific nature and practical feasibility of engineering decisions while improving the optimality of planning schemes.

[0138] While the specific embodiments of the present invention have been described above in conjunction with the accompanying drawings, this is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art without creative effort based on the technical solutions of the present invention are still within the scope of protection of the present invention.

Claims

1. A multi-objective collaborative optimization method for an integrated energy system, characterized in that, include: Acquire multi-source data of the target area where the integrated energy system is located, and generate a set of typical uncertainty scenarios based on the multi-source data of the target area where the integrated energy system is located; Construct a dynamic behavior model library, which includes mathematical models for equipment in an integrated energy system that reflect actual operating characteristics; The equipment capacity in the integrated energy system is used as the first-stage decision variable, and the operating strategies of the equipment in the integrated energy system under various uncertainty scenarios are used as the second-stage decision variables. A multi-objective optimization model is established with the optimization objectives of minimizing the expected value of annualized total cost, minimizing the expected value of carbon emissions, and minimizing the equipment footprint. A hierarchical hybrid intelligent optimization algorithm is used to solve the multi-objective optimization model and obtain the Pareto optimal frontier solution set; Using the solutions in the Pareto optimal frontier as alternatives, the fuzzy hierarchical analysis method is used to comprehensively score and rank all alternatives, and output the optimal solution that best meets the decision-maker's intention.

2. The multi-objective collaborative optimization method for an integrated energy system according to claim 1, characterized in that, Acquire multi-source data of the target area where the integrated energy system is located, and based on the multi-source data of the target area where the integrated energy system is located, generate a set of typical uncertainty scenarios, specifically including: Acquire at least one full year's worth of load time-series data and meteorological time-series data for the area to be planned where the integrated energy system is located; acquire data on the technical and economic parameters of equipment, time-of-use energy prices, and grid carbon emission factors in the integrated energy system; acquire data on the travel patterns and behaviors of electric vehicle users. Based on multi-source data of the planned area where the integrated energy system is located, uncertainties affecting the operation of the integrated energy system are identified; these uncertainties include the volatility of photovoltaic output, the randomness of load demand, and the grid availability of electric vehicle clusters. Perform statistical modeling and correlation analysis of probability distribution functions for uncertain variables; Initial uncertainty scenarios are generated based on Latin hypercube sampling; Based on dynamic time warping and K-medoids clustering, the generated initial scene set is reduced to extract representative typical scenes; The probability of occurrence of each typical scenario is determined based on the number of typical scenarios and the total number of initial scenarios.

3. The multi-objective collaborative optimization method for an integrated energy system according to claim 1, characterized in that, The dynamic behavior model library includes: heat pump dynamic COP model, energy storage battery state of health (SOH) decay model, V2G polymerization response model, and hydrogen energy system dynamic model; The dynamic COP model of the heat pump is a nonlinear function that describes the performance coefficient COP of the heat pump as dependent on the heat source temperature, the load-side temperature, and the load factor PLR. The State of Health (SOH) degradation model for energy storage batteries is as follows: the correlation between the degree of degradation of energy storage batteries and calendar degradation and cycle degradation; The V2G aggregation response model is an aggregation model that describes the dynamic charging and discharging power and energy capacity that a V2G cluster can provide at any given time. The dynamic model of the hydrogen energy system is as follows: the hydrogen production efficiency of the electrolyzer is a nonlinear function model with respect to the input power, and the power generation efficiency of the fuel cell is a nonlinear function model with respect to the output power.

4. The multi-objective collaborative optimization method for an integrated energy system according to claim 3, characterized in that, The specific SOH (State of Health) degradation model for energy storage batteries is as follows: ； in, This represents the maximum available capacity of the energy storage batteries for the next scheduling cycle. This represents the maximum available capacity of the energy storage batteries during the current scheduling cycle. This represents the annualized calendar degradation rate of the energy storage batteries for the current scheduling cycle. This refers to the capacity loss caused by each full cycle of an energy storage battery. The annual equivalent full cycle count is calculated using the rainflow counting method for energy storage batteries in the current scheduling cycle.

5. The multi-objective collaborative optimization method for an integrated energy system according to claim 1, characterized in that, The multi-objective optimization model specifically includes: minimizing the expected total cost, minimizing the expected carbon emissions, and minimizing the equipment footprint; The specific method for calculating the minimum expected value of total cost is as follows: in: The expected value of total cost. This represents the annualized initial investment cost. The annualized replacement cost is... To fix maintenance costs, The expected value of variable operating costs; The specific method for calculating the minimum expected value of carbon emissions is as follows: in, To minimize the expected carbon emissions, Let be the probability of occurrence of typical scenario s in the set of typical scenarios S. This represents the power purchased by scenario s at time t. Let T be the real-time marginal carbon emission factor of the power grid at time t, where T is the set of time steps throughout the year. This represents the natural gas consumption at time t in a typical scenario. Carbon emission factors of natural gas; The specific method for calculating the minimum equipment footprint is as follows: in, For the equipment's footprint, A e Let $e$ be the area occupied per unit capacity of device $e$ in device set $E$. Let e ​​be the capacity of device e in device set E.

6. The multi-objective collaborative optimization method for an integrated energy system according to claim 5, characterized in that, Annualized initial investment cost The specific calculation method is as follows: ； in, It is the capital recovery factor of equipment e, I e X is the unit investment cost of equipment e. e This refers to the installation capacity of device e. The calculation formula is: ； Where d is the discount rate. This refers to the actual lifespan of the equipment; Annualized replacement cost The specific calculation method is as follows: ； in, It is the total present value of all future reset events. It refers to the entire project planning cycle. Capital recovery factor; total present value of all future reset events. The calculation formula is: in, It is a collection of equipment that needs to be replaced during the project cycle. It is the sequence number of the reset event. This represents the total number of times equipment e was replaced during the project cycle. For the entire project planning cycle Capital recovery factor The calculation formula is: ； Fixed maintenance costs The specific calculation method is as follows: ； in, The annual fixed operation and maintenance cost per unit capacity of equipment e; Expected variable operating costs The specific calculation method is as follows: ； in, Let be the fuel cost at time t in scenario s. Let be the power grid interaction cost at time t in scenario s. For the variable operation and maintenance cost at time t in scenario s, The cost of V2G scheduling compensation at time t in scenario s.

7. The multi-objective collaborative optimization method for an integrated energy system according to claim 1, characterized in that, The constraints of the multi-objective optimization model include the balance constraints of electricity, heat, and hydrogen multi-energy flow under various typical scenarios, the constraints of the dynamic behavior model of equipment, and the annual periodic constraints of the energy storage system.

8. The multi-objective collaborative optimization method for an integrated energy system according to claim 1, characterized in that, Using solutions in the Pareto optimal frontier as alternatives, fuzzy hierarchical analysis is employed to comprehensively score and rank all alternatives, outputting the optimal solution that best reflects the decision-maker's intent. Specifically, this includes: Constructing a linguistic scale for a fuzzy judgment matrix; Construct a fuzzy judgment matrix based on the linguistic scale of the fuzzy judgment matrix; Calculate the fuzzy weights for each optimization objective; The calculated fuzzy weights for each optimization objective are converted into final weight values. The target values ​​of all alternative solutions are normalized. The final score of each alternative is determined based on the final weight value of each optimization objective and the normalized objective value of all alternatives. The option with the highest final score is output as the optimal solution.

9. A multi-objective collaborative optimization system for an integrated energy system, characterized in that, include: The generation module acquires multi-source data of the target area where the integrated energy system is located, and generates a set of typical uncertainty scenarios based on the multi-source data of the target area where the integrated energy system is located; The module constructs a dynamic behavior model library, which includes mathematical models that reflect the actual operating characteristics of equipment in an integrated energy system. A module is established, which takes the equipment capacity of the integrated energy system as the first-stage decision variable and the operation strategy of the equipment in the integrated energy system under various uncertain scenarios as the second-stage decision variable. A multi-objective optimization model is established with the optimization objectives of minimizing the expected value of annualized total cost, minimizing the expected value of carbon emissions, and minimizing the equipment footprint. The solution module employs a hierarchical hybrid intelligent optimization algorithm to solve the multi-objective optimization model and obtain the Pareto optimal frontier solution set. The output module uses the solutions in the Pareto optimal frontier as alternatives, and employs fuzzy hierarchical analysis to comprehensively score and rank all alternatives, outputting the optimal solution that best meets the decision-maker's intentions.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of a multi-objective collaborative optimization method for an integrated energy system as described in any one of claims 1 to 8.

Images