Comprehensive energy system multi-time scale optimization scheduling method considering source load uncertainty

Abstract

The invention discloses an integrated energy system multi-time-scale optimization scheduling method considering source load uncertainty, and the method comprises the steps: formulating a multi-time-scale optimization operation strategy for coordinating economical efficiency and low-carbon performance under the interference of the source load uncertainty; therefore, the flexibility margin of the comprehensive energy system is improved, the reasonable balance between economy and environment is achieved, and the consumption of renewable energy sources is promoted. The method comprises the following steps: firstly, aiming at the difference between the day-ahead precision and the day-ahead precision of source load prediction, adopting different description methods for the day-ahead uncertainty and the day-ahead uncertainty; secondly, taking economic and carbon emission optimization as an objective function, and constructing a day-ahead-day coordination optimization framework and strategy of the integrated energy system; and finally, designing an optimization method combining multiple targets and uncertainty to solve the multi-time scale optimization scheduling strategy. Simulation results show that the strategy provided by the invention not only ensures the flexible operation of the system and deals with the deviation between the intra-day scheduling and the day-ahead plan, but also effectively coordinates the economic cost and carbon emission of the system.

Description

Technical Field

[0001] This invention belongs to the field of integrated energy system operation and scheduling technology, specifically relating to a multi-timescale optimization scheduling method for integrated energy systems that considers source-load uncertainty. Background Technology

[0002] With the depletion of fossil fuels and increasing environmental pollution, countries around the world have formulated corresponding energy transition policies, and building a modern energy system that is economically low-carbon, flexible, and stable has become an important direction for the energy strategies of many countries. Integrated energy systems that combine electricity, heat, and gas are a crucial component of energy transition and sustainable development. Through multi-energy coupling and complementarity, integrated energy systems achieve flexible energy interaction, effectively improving energy efficiency and reducing environmental pollution.

[0003] However, renewable energy is characterized by randomness, volatility, and intermittency. As the penetration rate of renewable energy continues to increase, system uncertainty intensifies, causing integrated energy systems to be unable to meet the requirements for flexible operation during certain periods. Compared to traditional energy systems, integrated energy systems possess richer flexibility resources, which can effectively improve the system's flexible operation capabilities in multi-energy transmission, storage, and conversion processes. Different energy conversion devices have different adjustment capabilities, and therefore different abilities to cope with system uncertainty. Exploiting flexibility resources is a key technical means to solve the above problems. Furthermore, the operation of integrated energy systems involves a conflict between economics and the environment. Traditional or high-carbon energy sources are often cheaper and more readily available; excessive pursuit of low economic costs can lead to a sharp increase in carbon emissions.

[0004] Therefore, to support the economical, low-carbon, and flexible operation of the system, it is crucial to propose a comprehensive energy system optimization operation method that can ensure system flexibility and balance economic and environmental aspects under the interference of source and load uncertainties. Current research on the optimization operation of comprehensive energy systems mainly focuses on the day-ahead long-term timescale, failing to effectively coordinate the inconsistencies between intraday scheduling and day-ahead planning, and relatively few studies consider the uncertainties of source and load forecast data in intraday scheduling. Summary of the Invention

[0005] The technical problem this invention aims to solve is the lack of economical, low-carbon, and flexible operating schemes to address the inconsistencies between source load uncertainty and intraday scheduling and day-ahead planning, which prevent the system from meeting flexible operating requirements. This invention proposes a multi-timescale optimization scheduling method for integrated energy systems that considers source load uncertainty. Using economy and low carbon emissions as objective functions, it coordinates day-ahead and intraday scheduling within the system to improve scheduling accuracy, operating efficiency, and energy utilization in practical applications.

[0006] The above-mentioned objectives of the present invention are achieved through the following technical solutions:

[0007] This invention provides a multi-timescale optimal scheduling method for integrated energy systems considering source-load uncertainty, comprising:

[0008] (1) Modeling steps for uncertainties in day-ahead and intraday source-load forecasts:

[0009] To address the uncertainties in source-load forecasting during the day-ahead phase, a multi-discrete scenario-based partial Bruker optimization method is employed to model the uncertainties in renewable energy output, including:

[0010] a) Obtain M original scenes through Latin hypercube sampling, and then use the Kantorovich distance scene reduction technique to obtain K discrete typical scenes and scene simplification. probability distribution And use this probability distribution as the initial probability distribution;

[0011] b) Using the initial probability distribution as the center, construct confidence intervals for the 1-norm and ∞-norm to form a comprehensive norm fuzzy set, thereby limiting the fluctuation range of the probability distribution;

[0012] To address the uncertainty in source-load forecasting during the intraday phase, a fuzzy chance constraint method based on credibility measures is introduced to fuzzify the uncertain variables in the power balance constraint. The fuzzy chance constraint is then transformed into a deterministic constraint through a clear equivalence class transformation.

[0013] (2) Steps for constructing a day-ahead and intraday coordinated optimization scheduling framework for an integrated energy system:

[0014] a) Day-ahead scheduling phase: Using 1 hour as the time scale, the next day's operation plan is formulated 24 hours in advance, and multi-energy flexibility supply and demand balance constraints are introduced to enhance the system's adaptability to the uncertainty of renewable energy output and load demand. A day-ahead sub-Blu-ray optimization model based on multiple discrete scenarios is established.

[0015] b) Intraday scheduling phase: At 15-minute intervals, based on the daily scheduling plan and ultra-short-term forecast data, the scheduling plan for the next 4 hours is calculated each time, and a rolling update strategy is implemented. Only the optimization results of the first period of each rolling window are executed to correct the deviation between the daily plan and the actual operation, and an intraday rolling optimization model based on fuzzy chance constraints is established.

[0016] Both phases aim to minimize operating economic costs and carbon emissions.

[0017] (3) Steps for solving multiple objectives and uncertainties in a coordinated manner:

[0018] a) The recent solution for the Brussels bar optimization problem includes:

[0019] The column and constraint generation algorithm is used to decompose the day-to-day Bruker optimization model into main and sub-problems:

[0020] The main problem is to take economic efficiency as the primary objective and carbon emissions as the constraint. A uniformly distributed Pareto front is generated through the normalization method and the plane constraint method, and a compromise scheduling scheme for the day-ahead is comprehensively decided.

[0021] The sub-problems are solved under the worst-case probability distribution to determine the output of the remaining equipment and the energy purchase plan from the higher level, forming the day-ahead "skeleton" plan;

[0022] b) Intraday rolling optimization solution, including:

[0023] By using the point corresponding to the day-ahead compromise solution on the Utopia line as a fixed parameter, the intraday rolling optimization model based on fuzzy chance constraints is transformed into a single-objective optimization model, and only the scheduling instructions for the first period of the rolling window are executed to achieve rapid response and deviation correction.

[0024] Furthermore, the probability distribution in step (1) The confidence constraint is expressed as:

[0025] ,

[0026] ,

[0027] In the formula: Pr(•) is the probability function; ||•||1 and ||•|| ∞ These are the 1-norm and ∞-norm, respectively; K is the number of discrete typical scenes; M is the number of original scenes; It is a probability distribution; These are the initial scene probability distribution values; and This is the allowable deviation value of the probability distribution. The larger the value, the more conservative the robust model, and vice versa.

[0028] In probability distribution The confidence levels satisfied by the 1-norm and ∞-norm are respectively given by and Indicate, then and Represented as:

[0029] ;

[0030] Confidence set of probability distribution for:

[0031] ;

[0032] The uncertainty variables in the power balance constraint are fuzzy processed as follows:

[0033] ,

[0034] In the formula: Cr{•} is the confidence measure of event {•}; α is the confidence level; The energy fuzzy parameter for the heat power collected by the concentrating system in the solar thermal power plant at time t is the energy transferred to the heat flow fluid. and These are the fuzzy output parameters of wind power and photovoltaic power at time t, respectively. , , These are the fuzzy load parameters of the electric, heat, and gas energy networks at time t, respectively. Let t be the electrical power supplied by the power generation system to the grid at time t; Let t be the heat power input from the heating network to the thermal storage system. Let t be the thermal power output of the thermal storage system to the heating network at time t; Let t be the energy transferred from the heat transfer fluid in the solar thermal power plant to the thermal storage system at time t. Let t be the energy transferred from the heat-conducting fluid in the solar thermal power plant to the power generation system at time t; Let t be the energy transferred by the thermal storage system to the heat-conducting fluid in the solar thermal power plant at time t. , These represent the purchased electricity and gas power at time t, respectively. , , , These represent the electrical power generated or consumed by the electric-to-gas conversion device, electric boiler, combined heat and power unit, and gas turbine at time t, respectively. , These represent the charging and discharging power of the battery at time t, respectively. , These represent the thermal power generated by the cogeneration unit and the electric boiler at time t, respectively. , , These represent the gas power generated or consumed by the cogeneration unit, gas turbine, and power-to-gas conversion unit at time t, respectively. , These represent the storage and venting power of the gas storage tank at time t, respectively.

[0035] The fuzzy opportunity constraint is transformed into a deterministic constraint through a clear equivalence class transformation, expressed as:

[0036] ,

[0037] ,

[0038] ,

[0039] ,

[0040] In the formula: , This is the predicted proportion coefficient for solar thermal power plants; , This is the predicted proportion coefficient for wind power; , This is the predicted proportion coefficient for photovoltaic power. , This is the load forecasting ratio coefficient; Let be the predicted power value of the gas load at time t; Let be the predicted power value of the electrical load at time t; and These are the predicted output values ​​of wind power and photovoltaic power at time t, respectively.

[0041] Furthermore, the day-ahead distributed bar optimization model based on multiple discrete scenarios in step (2) includes a two-stage, three-layer optimization problem; the first stage, the min problem, obtains the output value of the heating network supply-side equipment; the second stage, the max-min problem, finds the worst-case scenario probability distribution that maximizes the balance between the minimum operating economic cost and carbon emissions of the integrated energy system, and obtains the output of the remaining operating equipment and the amount of energy purchased from the upper-level energy network; the expression is as follows:

[0042] ,

[0043] ,

[0044] ,

[0045] In the formula: x is the decision variable for the first stage, which is the output value of the electric boiler and the combined heat and power unit; and These are continuous and integer variables in the second stage of scenario k, respectively. The former refers to the output value of each device except for the electric boiler and cogeneration unit, as well as the amount of energy purchased from the upper level, while the latter refers to the start-up and shutdown variables of the energy storage device. and These represent the operating costs and carbon emissions generated by variable x, respectively. and y and z represent the economic costs and carbon emissions generated by variables y and z, respectively; K is the number of discrete scenarios. Let k be the probability of the k-th scene occurring. Let Ψ be the k-th scene; Ψ is the fuzzy set with comprehensive norm. Given a set Time optimization variables and The feasible domain; and The constraints are divided into Phase 1 and Phase 2, and the constraints include multi-energy flexibility supply and demand balance constraints, equipment model constraints, upstream energy purchase constraints, and power balance constraints of each energy bus of electric, heat and gas. For energy purchase costs; and The operating and maintenance costs are for electric boilers and combined heat and power units, respectively. The operating and maintenance costs for equipment other than electric boilers and combined heat and power units; Carbon emissions generated from purchasing energy; and These are the carbon emissions generated by electric boilers and combined heat and power units during operation, respectively. This refers to the carbon emissions generated during the operation of equipment other than electric boilers and combined heat and power units.

[0046] The intraday rolling optimization model based on fuzzy chance constraints is expressed as:

[0047] ,

[0048] ,

[0049] ,

[0050] In the formula: The total economic cost of intraday scheduling; For the operation and maintenance costs of all equipment; To adjust costs, including costs adjusted for energy purchase plans. Adjustment cost with equipment output ; This represents the total carbon emissions scheduled within the day. Carbon emissions generated by all equipment; and These are the equality constraints and inequality constraints in the model, including equipment model constraints, upstream energy purchase constraints, power balance constraints of the electric, heat and gas energy bus, and fuzzy opportunity constraints processed by clear equivalence classes.

[0051] Furthermore, in step (3), during the day-ahead scheduling phase, with economic efficiency as the primary objective and carbon emissions as the constraint, the two-stage model is decomposed to obtain the main problem:

[0052] ,

[0053]

[0054] In the formula: η is the relaxation amount of the subproblem; q(•) is the simplified carbon emission scheduling objective function; l is the number of iterations; and This is an auxiliary variable in the l-th iteration; Let be the scene probability value in the l-th iteration; this value is known.

[0055] At this point, the main problem becomes a multi-objective optimization problem, which can be transformed into the following using the normalization method and the plane constraint method:

[0056] ,

[0057] In the formula: The direction of the Utopia Line; Let w be the normalized objective function; w be all variables in the principal problem, including x, y, and z. This invention will... Divide the solution into m equal parts, and finally obtain the uniformly distributed Pareto front.

[0058] After solving the multi-objective optimization problem and obtaining the Pareto front, a compromise solution that balances the two objectives of economic efficiency and carbon emissions is determined by comprehensive decision-making.

[0059] The goal of the subproblem is to find the worst-case scenario probability distribution that maximizes the economic costs of equipment other than electric boilers and cogeneration units, as well as the costs of purchasing energy from upstream suppliers. The decomposed subproblems are:

[0060] ,

[0061] In the formula: and All are known quantities; among them, The main problem is a compromise solution obtained based on the normalization method, the plane constraint method, and the comprehensive decision method; The point on the Utopian line that represents the compromise solution mainly serves to constrain the carbon emission values ​​in the subproblem, thereby ensuring that the low-carbon objective of the optimization model is not excessively sacrificed during the second-stage solution process.

[0062] During the intraday scheduling phase, the intraday rolling optimization model based on fuzzy chance constraints is transformed into a single-objective optimization model using the normalization method and the plane constraint method, resulting in:

[0063]

[0064] In the formula: This represents the direction of the utopian line in the normalized objective function during intraday scheduling. The objective function is the normalized objective function; Let be a point on the Utopia line, and be a known quantity.

[0065] Furthermore, the multi-energy flexibility supply and demand balance constraint includes: establishing upward / downward flexibility supply and demand balance inequalities for the three energy subsystems of electricity, heat, and gas, and requiring that the flexibility margin of each subsystem is non-negative in each time period.

[0066] Furthermore, the equipment model constraints include a solar thermal power plant, energy conversion equipment, and energy storage equipment. The solar thermal power plant adopts a three-stage model of heat collection, heat storage, and power generation, and the uncertainties of its heat collection and conversion are characterized by scenario variables and fuzzy variables at the day-ahead and intraday stages, respectively. The energy storage equipment includes batteries, gas storage tanks, and the solar thermal power plant's heat storage system, and its charging / discharging power and state of charge participate in both day-ahead sub-blob optimization and intraday rolling optimization.

[0067] It should be noted that intraday scheduling has a significantly reduced time scale compared to day-ahead scheduling, requiring fast solution algorithms to meet real-time requirements under high-frequency optimization. The multi-objective optimization method based on normalized plane constraints solves multiple single-objective optimization problems by using a discrete set of points on a utopia line, performs multi-objective trade-off analysis on the Pareto front, and ultimately determines the comprehensive optimal scheduling scheme. This algorithm needs to handle the complex coupling relationships between multi-dimensional objectives and perform multiple iterative calculations, leading to a significant increase in computational complexity and time. Therefore, in the intraday scheduling phase, [the following is missing from the original text]. The parameters are preset to fixed values, which correspond to the points on the Utopian line representing the trade-off between economy and low carbon emissions in the day-ahead scheduling plan. This simplification significantly improves the computational efficiency of the intraday optimization model while ensuring the system's low-carbon operating baseline.

[0068] In summary, this invention first describes the uncertainties in source-load forecasting during day-ahead and intraday scheduling by using biblical optimization and fuzzy chance constraints, respectively. Next, under the influence of renewable energy uncertainty, a day-ahead and intraday coordinated optimization scheduling framework and strategy for the integrated energy system are established to address the inconsistencies between intraday scheduling and day-ahead planning. Finally, a collaborative solution method combining multi-objective optimization and uncertainty optimization is designed to solve the constructed multi-timescale optimization scheduling model, effectively supporting the economical, low-carbon, and flexible operation of the system.

[0069] Compared with the prior art, the present invention has the following advantages:

[0070] (1) This invention proposes a multi-objective day-ahead-intraday coordinated optimization scheduling strategy for integrated energy systems. By introducing multi-energy flexibility supply and demand balance constraints, when dealing with the interference of uncertainties in renewable energy output and load demand, it not only ensures the flexible operation of integrated energy systems and handles the deviation between intraday scheduling and day-ahead planning, but also effectively coordinates economic costs and carbon emissions, thereby expanding the research framework of integrated energy systems;

[0071] (2) This invention designs a solution method that combines multi-objective optimization and uncertainty optimization. In day-ahead scheduling, it combines multi-discrete scenario-based sub-bar optimization with multi-objective optimization. Through optimization scheduling under multi-objective and worst-case probability distributions, the integrated energy system achieves a reasonable trade-off between economy and environment in the face of renewable energy uncertainty. In intraday scheduling, it combines credibility fuzzy theory with multi-objective optimization. While fully considering source-load uncertainty, it balances economic costs and carbon emissions, thereby effectively supporting the economic, low-carbon, and flexible operation of the integrated energy system. Attached Figure Description

[0072] Figure 1 This is a framework diagram for day-ahead and intraday coordinated optimization scheduling of the integrated energy system;

[0073] Figure 2 This is a flowchart of the day-ahead and intraday coordinated optimization scheduling process for the integrated energy system;

[0074] Figure 3 It is a diagram of the integrated energy system structure;

[0075] Figure 4 This is a chart showing time-of-use electricity and gas prices;

[0076] Figure 5 This is a recent graph showing renewable energy and load data;

[0077] Figure 6 Renewable energy and load data for the day;

[0078] Figure 7 This is a typical wind power scenario diagram;

[0079] Figure 8 This is a typical photovoltaic scenario diagram;

[0080] Figure 9 This is a typical scene diagram of a solar thermal power plant collecting heat power;

[0081] Figure 10 This is a diagram showing the day-ahead scheduling results of electricity, heat, and gas energy based on the model proposed in this invention;

[0082] Figure 11 This is a multi-functional flexibility supply and demand diagram that does not consider flexibility constraints;

[0083] Figure 12 This is a supply and demand relationship diagram of the multi-functional flexibility model proposed in this invention;

[0084] Figure 13 A comparison chart of the scheduling results for the day and the previous day. Detailed Implementation

[0085] To more thoroughly explain the purpose, characteristics, and advantages of this invention, the following content will provide a detailed analysis of the technical solution based on specific embodiments.

[0086] The present invention provides a multi-timescale optimization scheduling method for integrated energy systems considering source-load uncertainty, comprising the following steps:

[0087] Step 1: Addressing Uncertainties in Day-ahead and Intraday Source Load Forecasts

[0088] 1. Handling of Renewable Energy Uncertainties

[0089] To address the uncertainty in source-load forecasting during the day-ahead phase, a multi-discrete scenario-based debulking optimization method is employed to model the uncertainty in renewable energy output. First, M original scenarios are obtained using Latin hypercube sampling. Then, a scenario reduction technique based on Kantorovich distance is used to obtain K discrete typical scenarios and scenario-based scenarios. probability distribution This probability distribution is then used as the initial probability distribution. Given the potential deviations in the numerical values ​​of different scenarios during the discretization process, a 1-norm and ∞-norm confidence interval centered on the initial probability distribution is constructed, forming a comprehensive norm fuzzy set to limit the probability distribution range of uncertain scenarios.

[0090] probability distribution The confidence constraint can be expressed as:

[0091]

[0092]

[0093] In the formula: Pr(•) is the probability function; ||•||1 and ||•|| ∞ These are the 1-norm and ∞-norm, respectively; K is the number of discrete typical scenes; M is the number of original scenes; It is a probability distribution; These are the initial scene probability distribution values; and This is the allowable deviation value of the probability distribution. The larger the value, the more conservative the robust model, and vice versa.

[0094] In probability distribution The confidence levels satisfied by the 1-norm and ∞-norm are respectively given by and Indicate, then and Represented as:

[0095]

[0096] In summary, the confidence set of the probability distribution can be derived. for:

[0097]

[0098] 2. Handling of intraday source-load uncertainty

[0099] Given the differences in the accuracy of intraday forecasting of renewable energy output and load demand compared to day-ahead forecasting, and the different response speed requirements between intraday and day-ahead dispatching, a fuzzy chance constraint method based on credibility measure is introduced to address the uncertainty in source-load forecasting during the intraday phase. This method fuzzifies the uncertain variables in the power balance constraint and transforms the fuzzy chance constraint into a deterministic constraint through clear equivalence class transformation.

[0100] The present invention performs fuzzy processing on the uncertain variables included in the power balance constraint as follows:

[0101]

[0102] In the formula: Cr{•} is the confidence measure of event {•}; α is the confidence level; The energy fuzzy parameter for the heat power collected by the concentrating system in the solar thermal power plant at time t is the energy transferred to the heat flow fluid. and These are the fuzzy output parameters of wind power and photovoltaic power at time t, respectively. , , These are the fuzzy load parameters of the electric, heat, and gas energy networks at time t, respectively. Let t be the electrical power supplied by the power generation system to the grid at time t; Let t be the heat power input from the heating network to the thermal storage system. Let t be the thermal power output of the thermal storage system to the heating network at time t; Let t be the energy transferred from the heat transfer fluid in the solar thermal power plant to the thermal storage system at time t. Let t be the energy transferred from the heat-conducting fluid in the solar thermal power plant to the power generation system at time t; Let t be the energy transferred by the thermal storage system to the heat-conducting fluid in the solar thermal power plant at time t. , These represent the purchased electricity and gas power at time t, respectively. , , , These represent the electrical power generated or consumed by the electric-to-gas conversion device, electric boiler, combined heat and power unit, and gas turbine at time t, respectively. , These represent the charging and discharging power of the battery at time t, respectively. , These represent the thermal power generated by the cogeneration unit and the electric boiler at time t, respectively. , , These represent the gas power generated or consumed by the cogeneration unit, gas turbine, and power-to-gas conversion unit at time t, respectively. , t represents the storage and venting power of the gas storage tank at time t.

[0103] The above fuzzy opportunity constraints are transformed into corresponding clear equivalence classes, represented as:

[0104]

[0105]

[0106]

[0107]

[0108] In the formula: Let t be the energy transferred by the heat collection system to the heat conduction fluid in the solar thermal power plant at time t. , , These are the predicted power values ​​for the electrical, thermal, and gas loads at time t, respectively. , These are the predicted output values ​​of wind power and photovoltaic power at time t, respectively. , This is the predicted proportion coefficient for solar thermal power plants; , This is the predicted proportion coefficient for wind power; , This is the predicted proportion coefficient for photovoltaic power. , This is the load forecasting ratio.

[0109] Step 2: Establish a day-ahead and intraday coordinated optimization scheduling framework for the integrated energy system

[0110] Multi-timescale optimization scheduling of integrated energy systems encompasses day-ahead and intraday scheduling to support the system's economical, low-carbon, and flexible operation. Day-ahead scheduling develops the next day's operating plan 24 hours in advance, providing the system with a "skeleton" plan that rationally balances overall economic efficiency and low-carbon performance, ensuring flexible system operation while reducing long-term operating costs and carbon emissions. Intraday scheduling, acting as the "muscle" of regulation, makes short-term adjustments based on the latest system status and forecast data, building upon the day-ahead scheduling plan. This allows for rapid response to uncertainties and addresses localized real-time changes. The framework for day-ahead-intraday coordinated optimization is as follows: Figure 1 As shown.

[0111] Based on the above analysis, day-ahead scheduling uses a 1-hour time scale to calculate the scheduling plan for the next 24 hours. Intraday scheduling combines the day-ahead scheduling plan with ultra-short-term forecast data, updating the forecast information at 15-minute intervals, and calculating a scheduling plan covering the next 4 hours each time. In each update, only the optimized scheduling for the first time period is executed, and then the above steps are repeated to implement a rolling update strategy to finally determine the system scheduling scheme. The multi-time-scale optimized scheduling process proposed in this invention is as follows: Figure 2 As shown.

[0112] 1. Day-to-day Brussels Bar Optimization Model Based on Multiple Discrete Scenarios

[0113] The day-ahead distributed broom optimization model based on multiple discrete scenarios includes a two-stage, three-layer optimization problem. The first stage, the min problem, obtains the output values ​​of the supply-side equipment (electric boilers and cogeneration units) of the heating network. The second stage, the max-min problem, seeks the probability distribution of the worst-case scenario that maximizes the balance between the economic cost of integrated energy system operation and carbon emissions, obtaining the output of the remaining operating equipment and the amount of energy purchased from the upstream energy network. The objective function can be expressed as:

[0114]

[0115]

[0116] In the formula: x is the decision variable for the first stage, which is the output value of the electric boiler and the combined heat and power unit; and These are continuous and integer variables in the second stage of scenario k, respectively. The former refers to the output value of each device except for the electric boiler and cogeneration unit, as well as the amount of energy purchased from the upper level, while the latter refers to the start-up and shutdown variables of the energy storage device. and These represent the operating costs and carbon emissions generated by variable x, respectively. and y and z represent the operating costs and carbon emissions generated by variables y and z, respectively; K is the number of discrete scenarios. Let k be the probability of the k-th scene occurring. Let Ψ be the k-th scene; Ψ is the fuzzy set with comprehensive norm. Given a set Time optimization variables and The feasible domain; For energy purchase costs; and The operating and maintenance costs are for electric boilers and combined heat and power units, respectively. The operating and maintenance costs for equipment other than electric boilers and combined heat and power units; Carbon emissions generated from purchasing energy; and These are the carbon emissions generated by electric boilers and combined heat and power units during operation, respectively. This refers to the carbon emissions generated during the operation of equipment other than electric boilers and combined heat and power units.

[0117] The constraints can be expressed as:

[0118]

[0119] In the formula: These are constraints that only relate to the first stage and do not consider the uncertainty of renewable energy scenario probabilities; therefore, they are deterministic constraints. This is the second stage of constraints. The constraints include multi-energy flexibility supply and demand balance constraints, equipment model constraints, upstream energy purchase constraints, and power balance constraints of each energy bus.

[0120] (1) Multi-functional flexibility supply and demand balance constraints

[0121] The multi-energy flexibility supply and demand balance of an integrated energy system refers to the system's ability to meet flexibility demands through the flexible supply of electricity, heat, and gas. Integrating the flexibility supply and demand of each component of the system yields the following:

[0122]

[0123] In the formula: , , , and , These represent the upward and downward flexible supply of the power grid, heating network, and gas network at time t, respectively; u represents the equipment that generates flexible power supply, including wind turbines, photovoltaics, gas turbines, combined heat and power units, batteries, and solar thermal power plants; and The upward and downward flexibility of the electrical energy supplied by the equipment at time t are respectively; v represents the equipment that generates flexible supply of thermal energy, including electric boilers, combined heat and power units, and solar thermal power plants; and The upward and downward flexibility of the thermal energy supplied by the equipment at time t are respectively; w represents the equipment that generates the flexible supply of gas energy, including the electro-gas converter and the gas storage tank; and The upward and downward flexibility of the gas energy provided by the device at time t are respectively. , , , and , t represents the upward and downward flexibility requirements of the power grid, heating network, and gas network at time t; x represents the equipment that generates the power flexibility requirement, including electrical loads, electric-to-gas conversion devices, and electric boilers; and t represents the upward and downward flexibility of the electrical energy consumed by the equipment at time t; y represents the equipment that generates thermal flexibility requirements, i.e., the heat load. and The upward and downward flexibility of the heat energy consumed by the heat load at time t are respectively; z represents the equipment that generates the gas energy flexibility demand, including gas load, gas turbine and cogeneration unit; and The upward and downward flexibility of the equipment at time t represent the gas energy consumed.

[0124] To ensure a certain degree of flexibility in the integrated energy system, the flexibility supply and demand balance constraint can be expressed as:

[0125]

[0126] In the formula: α represents the energy category, α∈{e,h,g} where e represents electrical energy, h represents thermal energy, and g represents gas energy; and These represent the upward and downward flexibility margins of the energy subsystem at time t, respectively.

[0127] (2) Equipment model constraints

[0128] 1) Solar thermal power plant

[0129] Abstracting the heat-conducting fluid inside a solar thermal power plant as an energy hub node and ignoring energy losses during its flow, the power plant can be considered as consisting of a solar concentrator, a thermal storage, and a power generation system. In addition to the above three subsystems, the model constraints also include internal thermal power balance constraints.

[0130] 2) Energy conversion equipment

[0131] Energy conversion equipment includes gas turbines, electric-to-gas conversion devices, electric boilers, and combined heat and power units. Their mathematical models and operating constraints are similar to those of the power generation system in a solar thermal power plant.

[0132] 3) Energy storage equipment

[0133] Energy storage devices include batteries and gas storage tanks, and their mathematical models and operating constraints are similar to those of thermal storage systems in solar thermal power plants.

[0134] (3) Constraints on energy purchases by higher authorities

[0135]

[0136] In the formula: and These represent the electricity and gas purchased from the superior energy grid at time t, respectively. and These are the maximum electricity and gas purchase capacities, respectively.

[0137] (4) Power balance constraints of each energy bus for electricity, heat and gas

[0138]

[0139] In the formula: , These are the predicted power values ​​of heat and gas loads at time t, respectively. , , , These represent the electrical power generated or consumed by the electric-to-gas conversion device, electric boiler, combined heat and power unit, and gas turbine at time t, respectively. , These represent the charging and discharging power of the battery at time t, respectively. , These represent the thermal power generated by the cogeneration unit and the electric boiler at time t, respectively. , , These represent the gas power generated or consumed by the cogeneration unit, gas turbine, and power-to-gas conversion unit at time t, respectively. , t represents the storage and venting power of the gas storage tank at time t.

[0140] 2. Intraday Rolling Optimization Model Based on Fuzzy Chance Constraints

[0141] The objective function is expressed as:

[0142]

[0143]

[0144]

[0145] In the formula: The total economic cost of intraday scheduling; For the operation and maintenance costs of all equipment; To adjust costs, including costs adjusted for energy purchase plans. Adjustment cost with equipment output ; This represents the total carbon emissions scheduled within the day. Carbon emissions generated by all equipment.

[0146] The constraints are expressed as follows:

[0147]

[0148] In the formula: and These are the equality constraints and inequality constraints in the model, including equipment model constraints, upstream energy purchase constraints, power balance constraints of the electric, heat and gas energy bus, and fuzzy opportunity constraints processed by clear equivalence classes.

[0149] Step 3: Design a collaborative solution method that integrates multi-objective optimization and uncertainty optimization.

[0150] To address the multi-objective optimization problem in the day-ahead-intraday coordinated scheduling strategy of integrated energy systems, this invention transforms it into a single-objective optimization problem using the normalization method and the plane constraint method.

[0151] In the day-ahead scheduling phase, a column and constraint generation algorithm is used to iteratively solve the day-ahead sub-Bruker optimization model based on multiple discrete scenarios. While minimizing impact on the scheduling scheme, economic scheduling is taken as the main objective function, simplifying the carbon emission scheduling objective function. The two-stage model is then decomposed, resulting in the main problem:

[0152]

[0153]

[0154] In the formula: η is the relaxation amount of the subproblem; q(•) is the simplified carbon emission scheduling objective function; l is the number of iterations; and This is an auxiliary variable in the l-th iteration; Let be the scene probability value in the l-th iteration, which is known.

[0155] At this point, the main problem becomes a multi-objective optimization problem, which can be transformed into the following using the normalization method and the plane constraint method:

[0156]

[0157] In the formula: The direction of the Utopia Line; Let w be the normalized objective function; w be all variables in the principal problem, including x, y, and z. This invention will... Divide the solution into m equal parts, and finally obtain the uniformly distributed Pareto front.

[0158] After solving the multi-objective optimization problem and obtaining the Pareto front, a compromise solution that balances the two objectives of economic efficiency and carbon emissions is determined through comprehensive decision-making.

[0159] The goal of the subproblem is to find the worst-case scenario probability distribution that maximizes the economic costs of equipment other than electric boilers and combined heat and power units, as well as the costs of purchasing energy from upstream suppliers. The decomposed subproblems are:

[0160]

[0161] In the formula: and All are known quantities. The main problem is a compromise solution obtained based on the normalization method, the plane constraint method, and the comprehensive decision method; The point corresponding to the compromise solution on the Utopia Line serves primarily to constrain the carbon emission values ​​in the subproblem, thereby ensuring that the low-carbon objective of the optimization model is not excessively sacrificed during the second-stage solution process.

[0162] During the intraday scheduling phase, the intraday rolling optimization model based on fuzzy chance constraints is transformed into a single-objective optimization model using the normalization method and the plane constraint method, resulting in:

[0163]

[0164] In the formula: This represents the direction of the utopian line in the normalized objective function during intraday scheduling. The objective function is the normalized objective function; Let be a point on the Utopia line, and be a known quantity.

[0165] It should be noted that intraday scheduling has a significantly reduced time scale compared to day-ahead scheduling, requiring fast solution algorithms to meet real-time requirements under high-frequency optimization. The multi-objective optimization method based on normalized plane constraints solves multiple single-objective optimization problems by using a discrete set of points on a utopia line, performs multi-objective trade-off analysis on the Pareto front, and ultimately determines the comprehensive optimal scheduling scheme. This algorithm needs to handle the complex coupling relationships between multi-dimensional objectives and perform multiple iterative calculations, leading to a significant increase in computational complexity and time. Therefore, in the intraday scheduling phase, [the following is missing from the original text]. The parameters are preset to fixed values, which correspond to the points on the Utopian line representing the trade-off between economy and low carbon emissions in the day-ahead scheduling plan. This simplification significantly improves the computational efficiency of the intraday optimization model while ensuring the system's low-carbon operating baseline.

[0166] Case Analysis

[0167] This invention focuses on an integrated energy system that includes solar thermal power plants and multi-element energy storage devices. Its structure is shown below. Figure 3 The parameters of the solar thermal power plant in the example are shown in Table 1, and the parameters of the other equipment are shown in Table 2. The time-of-use electricity price and gas price are shown in Table 3. Figure 4 Renewable energy and load data for the current day and intraday forecasts can be found here. Figure 5 , Figure 6 Furthermore, the objective function of this invention includes economic efficiency and carbon emissions. To obtain a uniformly distributed Pareto front and avoid excessive computation time, the Utopia line is divided into 10 equal parts.

[0168] In recent dispatching operations, considering the impact of renewable energy uncertainties on economic costs and carbon emissions, a confidence level for the uncertainty probability was set. and With values ​​of 0.5 and 0.99 respectively, Latin hypercube sampling was used to randomly generate scenarios from normally distributed renewable energy data samples. The number of scenarios, M, was set to 1000. The random scenarios were then reduced by considering the Kantorovich distance method to obtain discrete typical scenarios and their corresponding probability distributions. The number of discrete typical scenarios, K, was set to 5. Simultaneously, the volatility of wind power, photovoltaic power, and various loads was also taken into account, and their prediction error coefficients were... , , , , Based on recent forecasts and actual data, the values ​​are 0.25, 0.10, 0.15, 0.08, and 0.08. Discrete typical scenario curves for heat power collection by wind power, photovoltaic, and solar thermal power plants are shown below. Figures 7-9 In addition, the current scheduling time scale τ is set to 1 hour, and the total scheduling cycle T is set to 24 hours.

[0169] In intraday dispatching, this invention sets the unit adjustment cost for changes in purchased electricity and gas power to 0.045 yuan / (kW·h) and 0.025 yuan / (kW·h), respectively. Simultaneously, intraday dispatching needs to adjust the output of each device based on the day-ahead plan to smooth out power fluctuations. Given that energy storage devices respond faster than energy conversion devices, this invention sets the unit adjustment cost for power changes in energy conversion devices and energy storage devices to 0.03 yuan / (kW·h) and 0.01 yuan / (kW·h), respectively, to prioritize the use of energy storage devices to meet the needs of adjusting intraday output, ensuring the stability and economy of system operation. The intraday real-time dispatching timescale τ is set to 15 minutes, and the total dispatching cycle T is set to 24 hours. Furthermore, triangular fuzzy parameters are used to represent the intraday forecast data of renewable energy output and various loads, with a confidence level α of 0.8. The values ​​of the triangular membership parameters are shown in Table 3.

[0170] Table 1 Basic parameters of solar thermal power plant

[0171]

[0172] Table 2 Other equipment parameters

[0173]

[0174] Table 3. Membership parameters of triangles

[0175]

[0176] The following analysis and explanation will focus on the results, effects, and advantages of the model of this invention.

[0177] (1) Results analysis of the proposed model

[0178] The day-ahead scheduling results of the electricity, heat, and gas of the model proposed in this invention are as follows: Figure 10 As shown.

[0179] Depend on Figure 10 It can be seen that the proposed multi-objective distributed bar optimization day-ahead scheduling model can achieve coordinated complementarity among multiple energy sources. Considering the uncertainties of renewable energy, all renewable energy sources in the example have been utilized. Therefore, the proposed model can effectively support the stable operation of integrated energy systems including solar thermal power plants.

[0180] (2) Analysis of the effectiveness of the proposed model

[0181] On the one hand, to verify the effectiveness of considering the supply and demand balance constraint of multi-energy flexibility, an additional scenario without considering flexibility constraints is set up. The multi-energy flexibility supply and demand relationships of the model without flexibility constraints and the model proposed in this invention are as follows: Figure 11 , Figure 12 As shown.

[0182] Figure 11 (b) shows that at times 10:00 and 12:00, the supply of power downlink flexibility is less than the demand for power downlink flexibility; Figure 11 In (e), at times 07:00, 10:00, and 23:00, the supply of gas upward flexibility failed to meet the demand for gas upward flexibility. Figure 11 In scenario (f), at time 12:00, the gas-powered downlink flexibility supply failed to meet the gas-powered downlink flexibility demand. Because this scenario does not consider multi-energy flexibility supply and demand balance constraints, when load and renewable energy fluctuate, the flexibility supply may be unable to meet the flexibility demand at certain times. Figure 12 It can be seen that the flexibility supply of each energy subsystem at any given time can meet the corresponding flexibility requirements. This indicates that after considering the multi-energy flexibility supply and demand balance constraints, the system can ensure its flexibility by adjusting the output of each device, thereby effectively coping with the uncertainties brought about by load and renewable energy fluctuations.

[0183] On the other hand, intraday scheduling is based on the results of the previous day's scheduling, comparing the intraday scheduling calculation results with the previous day's scheduling plan, including solar thermal power plants, energy conversion equipment, energy storage equipment and the energy purchase plan of the superior.

[0184] from Figure 13 It can be seen that the intraday scheduling trends of various devices are basically consistent with the day-ahead scheduling. Compared with energy conversion equipment, the intraday scheduling of energy storage equipment differs significantly from the day-ahead plan in some periods, such as... Figure 13 (b) The thermal energy storage system of the solar thermal power plant is shown. This is because the adjustment cost of energy storage devices is lower than that of energy conversion devices when setting parameters, and energy storage generates almost no carbon emissions during operation; therefore, the system prioritizes the scheduling of energy storage devices. Furthermore, the intraday energy purchase plan largely aligns with the day-ahead energy purchase plan, but at certain times the intraday plan makes necessary adjustments to the day-ahead plan. For example, in… Figure 13 (i) During the period from 18:00 to 22:00, intraday dispatch increased the purchased power to meet the evening peak demand of the system's electricity load during this period. This is because intraday dispatch provides a more accurate description of electricity load changes, and adjustments to the purchased power and the output of other equipment are needed to address shortcomings in the day-ahead dispatch plan. Figure 13 (j) shows that during the period from 00:00 to 04:00, the daily gas purchase power was lower than the day-ahead plan. This is because, in order to cope with the strong uncertainty in the day-ahead forecast, the day-ahead gas purchase power was higher than the day-ahead gas load. The day-ahead and daily forecast fluctuations of gas load are relatively mild, indicating that the intraday scheduling takes into account less uncertainty than the day-ahead optimization, and the resulting strategy is more in line with the actual application scenario of the system.

[0185] In summary, incorporating multi-energy flexibility supply and demand balance constraints into the model ensures the flexible operation of the system, enabling the integrated energy system to withstand the risks of uncertain fluctuations at various times. Under the fluctuation conditions set in the case study of this invention, intraday scheduling can be planned based on the output of each device during the day-ahead scheduling phase, effectively mitigating power fluctuations by flexibly adjusting the power output of each device. This demonstrates that considering multi-energy flexibility supply and demand balance constraints during the day-ahead scheduling phase has a significant positive impact on intraday scheduling in response to power fluctuations, providing accurate and reliable guidance for the optimized operation of the integrated energy system.

[0186] (3) Analysis of the advantages of the proposed model

[0187] To further investigate the impact of different objective functions on the operation of the integrated energy system, two additional scenarios were proposed based on the model of this invention, as shown in Table 4. The calculation settings remained unchanged, and all scenarios considered multi-energy flexibility constraints. The results are shown in Table 5.

[0188] Table 4 Experimental Scenarios

[0189]

[0190] Table 5 Comparison of optimization results for different objective functions

[0191]

[0192] As shown in Table 5, Scenario 1 only considers the economic objective. Although the economic cost is 0.89 times that of the model in this invention, the carbon emissions are 1.30 times higher. Therefore, compared to a single economic objective, it is necessary for the model in this invention to consider multiple objectives. It can achieve greater environmental benefits by increasing a small portion of the economic cost, which aligns with the trend of low-carbon development in the economy and society. On the other hand, Scenario 2 only considers the carbon emission target. Although the carbon emissions are 0.85 times that of the model in this invention, the economic cost is 1.22 times higher. Therefore, sacrificing economic efficiency in pursuit of low carbon emissions is not advisable, as it may lead to a surge in economic costs, thereby hindering the development momentum of integrated energy systems including solar thermal power plants.

[0193] In summary, the model proposed in this invention, after considering multiple objectives, enables the operation of the integrated energy system to achieve a reasonable balance between economy and environment.

Claims

1. A multi-timescale optimal scheduling method for a comprehensive energy system considering source-load uncertainty, characterized in that, include: (1) Modeling steps for uncertainties in day-ahead and intraday source-load forecasts: To address the uncertainties in source-load forecasting during the day-ahead phase, a multi-discrete scenario-based partial Bruker optimization method is employed to model the uncertainties in renewable energy output, including: a) Obtain M original scenes through Latin hypercube sampling, and then use the Kantorovich distance scene reduction technique to obtain K discrete typical scenes and scene simplification. probability distribution And use this probability distribution as the initial probability distribution; b) Using the initial probability distribution as the center, construct confidence intervals for the 1-norm and ∞-norm to form a comprehensive norm fuzzy set, thereby limiting the fluctuation range of the probability distribution; To address the uncertainty in source-load forecasting during the intraday phase, a fuzzy chance constraint method based on credibility measures is introduced to fuzzify the uncertain variables in the power balance constraint. The fuzzy chance constraint is then transformed into a deterministic constraint through a clear equivalence class transformation. (2) Steps for constructing a day-ahead and intraday coordinated optimization scheduling framework for an integrated energy system: a) Day-ahead scheduling phase: Using 1 hour as the time scale, the next day's operation plan is formulated 24 hours in advance, and multi-energy flexibility supply and demand balance constraints are introduced to enhance the system's adaptability to the uncertainty of renewable energy output and load demand. A day-ahead sub-Blu-ray optimization model based on multiple discrete scenarios is established. b) Intraday scheduling phase: At 15-minute intervals, based on the daily scheduling plan and ultra-short-term forecast data, the scheduling plan for the next 4 hours is calculated each time, and a rolling update strategy is implemented. Only the optimization results of the first period of each rolling window are executed to correct the deviation between the daily plan and the actual operation, and an intraday rolling optimization model based on fuzzy chance constraints is established. Both phases aim to minimize operating economic costs and carbon emissions. (3) Steps for solving multiple objectives and uncertainties in a coordinated manner: a) The recent solution for the Brussels bar optimization problem includes: The column and constraint generation algorithm is used to decompose the day-to-day Bruker optimization model into main and sub-problems: The main problem is to take economic efficiency as the primary objective and carbon emissions as the constraint. A uniformly distributed Pareto front is generated through the normalization method and the plane constraint method, and a compromise scheduling scheme for the day-ahead is comprehensively decided. The sub-problems are solved under the worst-case probability distribution to determine the output of the remaining equipment and the energy purchase plan from the higher level, forming the day-ahead "skeleton" plan; b) Intraday rolling optimization solution, including: By using the point corresponding to the day-ahead compromise solution on the Utopia line as a fixed parameter, the intraday rolling optimization model based on fuzzy chance constraints is transformed into a single-objective optimization model, and only the scheduling instructions for the first period of the rolling window are executed to achieve rapid response and deviation correction.

2. The method according to claim 1, characterized in that, Probability distribution in step (1) The confidence constraint is expressed as: ， ， In the formula: Pr(•) is the probability function; ||•||1 and ||•|| ∞ These are the 1-norm and ∞-norm, respectively; K is the number of discrete typical scenes; M is the number of original scenes; It is a probability distribution; These are the initial scene probability distribution values; and This is the allowable deviation value of the probability distribution. The larger the value, the more conservative the robust model, and vice versa. In probability distribution The confidence levels satisfied by the 1-norm and ∞-norm are respectively given by and Indicate, then and Represented as: ； Confidence set of probability distribution for: ； The uncertainty variables in the power balance constraint are fuzzy processed as follows: ， In the formula: Cr{•} is the confidence measure of event {•}; α is the confidence level; The energy fuzzy parameter for the heat power collected by the concentrating system in the solar thermal power plant at time t is the energy transferred to the heat flow fluid. and These are the fuzzy output parameters of wind power and photovoltaic power at time t, respectively. , , These are the fuzzy load parameters of the electric, heat, and gas energy networks at time t, respectively. Let t be the electrical power supplied by the power generation system to the grid at time t; Let t be the heat power input from the heating network to the thermal storage system. Let t be the thermal power output of the thermal storage system to the heating network at time t; Let t be the energy transferred from the heat transfer fluid in the solar thermal power plant to the thermal storage system at time t. Let t be the energy transferred from the heat-conducting fluid in the solar thermal power plant to the power generation system at time t; Let t be the energy transferred by the thermal storage system to the heat-conducting fluid in the solar thermal power plant at time t. , These represent the purchased electricity and gas power at time t, respectively. , , , These represent the electrical power generated or consumed by the electric-to-gas conversion device, electric boiler, combined heat and power unit, and gas turbine at time t, respectively. , These represent the charging and discharging power of the battery at time t, respectively. , These represent the thermal power generated by the cogeneration unit and the electric boiler at time t, respectively. , , These represent the gas power generated or consumed by the cogeneration unit, gas turbine, and power-to-gas conversion unit at time t, respectively. , These represent the storage and venting power of the gas storage tank at time t, respectively. The fuzzy opportunity constraint is transformed into a deterministic constraint through a clear equivalence class transformation, expressed as: ， ， ， ， In the formula: , This is the predicted proportion coefficient for solar thermal power plants; , This is the predicted proportion coefficient for wind power; , This is the predicted proportion coefficient for photovoltaic power. , This is the load forecasting ratio coefficient; Let be the predicted power value of the gas load at time t; Let be the predicted power value of the electrical load at time t; and These are the predicted output values ​​of wind power and photovoltaic power at time t, respectively.

3. The method according to claim 1, characterized in that, The day-ahead distributed bar optimization model based on multiple discrete scenarios in step (2) includes a two-stage, three-layer optimization problem; the first stage, the min problem, obtains the output value of the heating network supply-side equipment; the second stage, the max-min problem, finds the worst-case probability distribution that maximizes the balance between the economic cost of integrated energy system operation and carbon emissions, and obtains the output of the remaining operating equipment and the amount of energy purchased from the upper-level energy network; the expression is as follows: ， ， ， In the formula: x is the decision variable for the first stage, which is the output value of the electric boiler and the combined heat and power unit; and These are continuous and integer variables in the second stage of scenario k, respectively. The former refers to the output value of each device except for the electric boiler and cogeneration unit, as well as the amount of energy purchased from the upper level, while the latter refers to the start-up and shutdown variables of the energy storage device. and These represent the operating costs and carbon emissions generated by variable x, respectively. and y and z represent the economic costs and carbon emissions generated by variables y and z, respectively; K is the number of discrete scenarios. Let k be the probability of the k-th scene occurring. For the k-th scene; Ψ is a fuzzy set with a comprehensive norm; Given a set Time optimization variables and The feasible domain; and The constraints are divided into Phase 1 and Phase 2, and the constraints include multi-energy flexibility supply and demand balance constraints, equipment model constraints, upstream energy purchase constraints, and power balance constraints of each energy bus of electric, heat and gas. For energy purchase costs; and The operating and maintenance costs are for electric boilers and combined heat and power units, respectively. The operating and maintenance costs for equipment other than electric boilers and combined heat and power units; Carbon emissions generated from purchasing energy; and These are the carbon emissions generated by electric boilers and combined heat and power units during operation, respectively. This refers to the carbon emissions generated during the operation of equipment other than electric boilers and combined heat and power units. The intraday rolling optimization model based on fuzzy chance constraints is expressed as: ， ， ， In the formula: The total economic cost of intraday scheduling; For the operation and maintenance costs of all equipment; To adjust costs, including costs adjusted for energy purchase plans. Adjustment cost with equipment output ; This represents the total carbon emissions scheduled within the day. Carbon emissions generated by all equipment; and These are the equality constraints and inequality constraints in the model, including equipment model constraints, upstream energy purchase constraints, power balance constraints of the electric, heat and gas energy bus, and fuzzy opportunity constraints processed by clear equivalence classes.

4. The scheduling method according to claim 1, characterized in that, In step (3), during the day-ahead scheduling phase, with economic efficiency as the primary objective and carbon emissions as the constraint, the two-stage model is decomposed to obtain the main problem: ， ， In the formula: η is the relaxation amount of the subproblem; q(•) is the simplified carbon emission scheduling objective function; l is the number of iterations; and This is an auxiliary variable in the l-th iteration; Let be the scene probability value in the l-th iteration; this value is known. At this point, the main problem becomes a multi-objective optimization problem, which can be transformed into the following using the normalization method and the plane constraint method: ， In the formula: The direction of the Utopia Line; The normalized objective function; w represents all variables in the principal problem, including x, y, and z; Divide the solution into m equal parts, and finally obtain the uniformly distributed Pareto front. After solving the multi-objective optimization problem and obtaining the Pareto front, a compromise solution that balances the two objectives of economic efficiency and carbon emissions is determined by comprehensive decision-making. The goal of the subproblem is to find the worst-case scenario probability distribution that maximizes the economic costs of equipment other than electric boilers and cogeneration units, as well as the costs of purchasing energy from upstream suppliers. The decomposed subproblems are: ， In the formula: and All are known quantities; among them, The main problem is a compromise solution obtained based on the normalization method, the plane constraint method, and the comprehensive decision method; The point on the Utopian line that represents the compromise solution mainly serves to constrain the carbon emission values ​​in the subproblem, thereby ensuring that the low-carbon objective of the optimization model is not excessively sacrificed during the second-stage solution process. During the intraday scheduling phase, the intraday rolling optimization model based on fuzzy chance constraints is transformed into a single-objective optimization model using the normalization method and the plane constraint method, resulting in: ， In the formula: This represents the direction of the utopian line in the normalized objective function during intraday scheduling. The objective function is the normalized objective function; Let be a point on the Utopia line, and be a known quantity.

5. The scheduling method according to claim 3, characterized in that, The multi-energy flexibility supply and demand balance constraint includes: establishing upward / downward flexibility supply and demand balance inequalities for the three energy subsystems of electricity, heat, and gas, and requiring that the flexibility margin of each subsystem is non-negative in each time period.

6. The scheduling method according to claim 3, characterized in that, The equipment model constraints include a solar thermal power plant, energy conversion equipment, and energy storage equipment. The solar thermal power plant adopts a three-stage model of heat collection, heat storage, and power generation, and the uncertainties of its heat collection and conversion are characterized by scenario variables and fuzzy variables at the day-ahead and intraday stages, respectively. The energy storage equipment includes batteries, gas tanks, and the solar thermal power plant's heat storage system. Its charging / discharging power and state of charge participate in both day-ahead sub-blob optimization and intraday rolling optimization.

Images