Comprehensive energy system optimization scheduling method based on improved particle swarm optimization

Abstract

The invention relates to a comprehensive energy system optimization scheduling method based on an improved particle swarm algorithm, and belongs to the technical field of comprehensive energy system optimization scheduling. The method comprises the steps of constructing an operation model of an integrated energy system with electrical cold and heat as main energy flows, constructing double-target integrated energy system optimization scheduling with the lowest operation cost and the optimal power supply reliability, and solving a multi-target problem through an improved particle swarm algorithm on the premise of meeting power balance constraints and equipment operation constraints. Self-adaptive variation and variable inertia factors are introduced on the basis of a traditional particle swarm algorithm, the self-adaptive variation refers to a variation thought in a genetic algorithm, and variation operation expands a population search space which is continuously reduced in iteration, so that particles can jump out of the position of an optimal value which is searched previously, search is carried out in a larger space, and the search efficiency is improved. The population diversity is maintained, and the possibility of searching the optimal value by the algorithm is improved. And the comprehensive energy system can give full play to economy and reliability under constraint conditions.

Description

Technical Field

[0001] The present invention belongs to the technical field of integrated energy system optimization and scheduling, and in particular relates to an integrated energy system optimization and scheduling method based on an improved particle swarm algorithm, which is used to reduce the total operating cost of the integrated energy system and improve the power supply reliability of the system. Background Art

[0002] The goal of an integrated energy system is to achieve coordinated planning, optimized operation, collaborative management, interactive response, and mutual complementarity among various heterogeneous energy subsystems. The scheduling problem of an integrated energy system involves the integrated utilization and scheduling management of various energy resources to achieve efficient energy utilization and supply and demand balance. This includes the unified scheduling and optimal configuration of various renewable energy sources such as wind energy and solar energy, as well as traditional energy. When multiple objective functions are involved, there are often conflicts in objective functions and multiple equivalent optimal solutions, which increase the complexity and diversity of the solution. Currently, traditional particle swarm algorithms and genetic algorithms have certain limitations. Therefore, continuous optimization is needed to improve the algorithm's search ability and convergence speed, so as to solve larger-scale and complex multi-objective optimization problems. An integrated energy system optimization scheduling based on an improved particle swarm algorithm can effectively improve energy utilization efficiency and promote sustainable energy development while meeting the diversified energy demand within the system. It is a new type of integrated energy system optimization scheduling. Summary of the Invention

[0003] This invention provides a method for optimizing and scheduling an integrated energy system based on an improved particle swarm algorithm. This method reduces the total operating cost of the energy system and improves system reliability based on the conversion between electricity, gas, cooling, and heat energy. The specific steps are as follows:

[0004] Step 1: Build an integrated energy system with electricity, gas, cooling and heating as energy flows;

[0005] Step 2: Construct a multi-objective objective function with total operating cost and power supply reliability as the goals:

[0006]

[0007] C pb 、C ps 、C bess 、C GB 、C G are the main grid electricity purchase price, electricity sales price, battery cost, gas turbine operation and maintenance cost, and natural gas price; P Gin 、P Gout 、P bess 、P GB 、P MT They are power purchased from the main grid, power sold, battery operating power, gas turbine operating power, and micro gas turbine operating power; L pcis the penalty coefficient when the power constraint is not met; H sum is the total value of the current equilibrium constraint.

[0008] Energy reliability costs:

[0009]

[0010] C1 and C2 are the penalty prices for unit curtailment of solar and wind power, respectively; P pre_p 、P pre_w are the predicted photovoltaic output power and wind power output power respectively; P PV 、P WT are the actual photovoltaic output power and wind power output power respectively.

[0011] The objective function 1 is to reduce the system operating cost and improve the equipment operating efficiency by minimizing the total operating cost function, and the objective function 2 is to describe the power supply reliability by minimizing the wind power consumption;

[0012] Step 3: Taking equipment operation constraints, ramp constraints, and power balance constraints as constraints, the objective function and constraints are used to construct the optimal scheduling of the integrated energy system;

[0013] Equipment operation constraints:

[0014] 1) Photovoltaic unit output constraints

[0015] P PV_min ≤P PV ≤P PV_max

[0016] P PV_max 、P PV_min It is the upper and lower limits of the output of photovoltaic generator set.

[0017] 2) Wind turbine output constraints

[0018] P WT_min ≤P WT ≤P WT_max

[0019] P WT_max 、P WT_min The upper and lower limits of wind turbine output.

[0020] 3) Micro gas turbine output constraints:

[0021] P MT_min (t)≤P MT (t)≤P MT_max (t)

[0022] P MT (t)-P MT (t-1)≤L MT

[0023] P MT is the output power of the micro gas turbine; P MT_max 、P MT_min L is the upper and lower limits of the micro gas turbine output; MT is the ramp constraint limit of the micro gas turbine.

[0024] 4) Waste heat boiler output constraints:

[0025] P HRB_min (t)≤P HRB (t)≤P HRB_max (t)

[0026] P HRB is the output power of waste heat boiler; P HRB_max 、P HRB_min The upper and lower limits of waste heat boiler output.

[0027] 5) Gas boiler output constraints

[0028] P GB_min (t)≤P GB (t)≤P GB_max (t)

[0029] P GB is the output power of the gas boiler; P GB_max 、P GB_min The upper and lower limits of gas boiler output.

[0030] 6) Absorption chiller output constraints

[0031] P AC_min (t)≤P AC (t)≤P AC_max (t)

[0032] P AC is the output power of the absorption chiller; P AC_min 、P AC_max The upper and lower limits of the absorption chiller output.

[0033] 7) Electric refrigerator output constraints:

[0034] P ER_min (t)≤P ER (t)≤P ER_max (t)

[0035] P ER is the output power of the electric refrigerator; P ER_min 、P ER_max The upper and lower limits of the electric refrigerator output.

[0036] 8) P2G output constraints

[0037] P P2G_min (t)≤P P2G (t)≤P P2G_max (t)

[0038] P P2G is the P2G output power; P P2G_max 、P P2G_min The upper and lower limits of P2G output.

[0039] 9) Interaction constraints of the main network of the tie line

[0040] P Grid_min (t)≤P Grid (t)≤P Grid_max (t)

[0041] P Grid is the main network interaction power of the tie line; P Grid_max 、P Grid_min It is the upper and lower limits of the main network interaction of the interconnection line.

[0042] 10) Battery output limit:

[0043]

[0044] P BESS_min (t)≤P BESS (t)≤P BESS_max (t)

[0045] SOC min (t)≤SOC(t)≤SOC max (t)

[0046] are the capacity of the energy storage device at the initial moment and the capacity at the final moment, P BESS is the battery output power; P BESS_max 、P BESS_min The upper and lower limits of battery output; SOC is the battery capacity state; SOC max , SOC min They are the upper and lower limits of the battery capacity status respectively.

[0047] (2) Power balance constraints

[0048] 1) Power balance constraints of power grid

[0049] P PV (t)+P WT (t)+P MT (t)+P BESS (t)+P Grid (t) = P ER (t)+P P2G (t)+Lele (t)

[0050] L ele is the electrical load value.

[0051] 2) Gas network power balance constraints:

[0052] P P2G (t) = P WT (t)+P GB (t)+L gas (t)

[0053] L gas is the gas load value.

[0054] 3) Cold network power balance constraints:

[0055] P AC (t)+P ER (t) = L cold (t)

[0056] L cold is the cooling load value.

[0057] 4) Thermal network power balance constraints:

[0058] P GB (t)+P HRB (t) = P ER (t)+L hot (t)

[0059] L hot is the heat load value.

[0060] Step 4: Under the constraints of multiple objectives and equipment constraints, the improved particle swarm algorithm is used to solve the integrated energy system and obtain the Pareto frontier and the equipment scheduling after the electricity, gas, cooling and heat networks are balanced. Adaptive mutation is introduced on the basis of the traditional particle swarm algorithm. Adaptive mutation refers to the mutation idea in the genetic algorithm and introduces mutation operation in the PSO algorithm, that is, reinitializing certain variables with a certain probability. The mutation operation expands the population search space that is shrinking in the iteration, allowing particles to jump out of the optimal value position previously searched and search in a larger space, while maintaining the diversity of the population and increasing the possibility of the algorithm finding the optimal value. Therefore, a simple mutation operator is introduced on the basis of the ordinary particle swarm algorithm. After each particle update, the particle is reinitialized with a certain probability, reflecting a strong global search capability. The inertia weight factor of the traditional PSO algorithm is fixed and easy to fall into the local optimal value. To address this shortcoming, the particle swarm algorithm is improved from two aspects: inertia weight factor and genetic mutation:

[0061]

[0062] IT is the current iteration number; MI is the total iteration number; w s and w e are the initial and final values ​​of the inertia weight factor. In the early stages of the iteration, a larger w prevents the algorithm from falling into local minima, facilitating global search. In the later stages of the iteration, a smaller w facilitates local search and facilitates convergence of the algorithm.

[0063] The basic steps of MOPSO are:

[0064] (1) Initialization: Generate N in the D-dimensional search space s A population of particles, initializing the speed, position and external files of each particle in the population;

[0065] (2) Calculate the objective function value of each particle in the population and save the non-inferior solution to the external storage file according to the dominance relationship; initialize the individual optimal P of the particle best and the global optimal G best , is the initial position of the particle, G best is the best position of the initial population;

[0066] (3) Update the particle's velocity and position according to the position and velocity update formula;

[0067] (4) Calculate the objective function value of the particle and compare it with the P of the previous iteration best For comparison, re-adjust P best ;

[0068] (5) Perform hierarchical sorting on the population, store the optimal non-dominated Pareto solution into an external archive set, remove non-Pareto solutions, and determine whether the external archive set exceeds the specified capacity. If so, select m particles according to the crowding distance;

[0069] (6) Update the external storage set and determine the current global optimal G best , using the Pareto optimal solution saved in the external archive set, this paper uses the roulette method to select G from the external set according to the crowding distance of the optimal solution best .

[0070] (7) Small probability mutation: In order to prevent the MOPSO algorithm from converging to the local optimal frontier rather than the global optimal frontier too early, a small probability mutation mechanism is introduced to make the position of the particle at the initial position undergo a small probability perturbation of ±30%, thereby improving the search ability of the particle and increasing the particle's ability to find the global optimal frontier;

[0071] (8) Determine whether the termination condition is met. If so, output all optimal solutions in the external file. Otherwise, jump to step (3).

[0072] Assume that the total number of particles is N s , then the update formula for the position and velocity of the nth particle in dimension D is as follows:

[0073]

[0074] Where: ω' is the inertia weight factor, c1 and c2 are learning factors, reflecting the self-learning ability and social learning ability respectively, and r1 and r2 are random numbers uniformly distributed in [0,1]. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] The drawings in the specification are part of the present invention and are intended to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0076] Figure 1 A schematic diagram of an application of an integrated energy system in an optimized scheduling of an integrated energy system according to an embodiment of the present invention;

[0077] Figure 2 This is a flow chart of an improved particle swarm algorithm for optimizing and scheduling an integrated energy system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0078] To more clearly illustrate the technical solutions of the embodiments of the present invention, the following will provide a clear and complete description of the technical solutions of the embodiments of the present invention in conjunction with the accompanying drawings of the embodiments of the present invention. In addition, the drawings are merely schematic illustrations of the present invention and are not necessarily drawn to scale. Some of the block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software, or in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.

[0079] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is intended to include the plural form. In addition, it should be understood that when the terms "comprise" and / or "include" are used in this specification, they indicate the presence of features, steps, operations, devices, components and / or combinations thereof.

[0080] In the absence of conflict, the embodiments of the present invention and the features thereof may be combined with each other.

[0081] Example 1: The integrated energy system model in this example is established as follows Figure 1As shown, the main energy flow sources are electricity, gas, cooling, and heat. Electricity flows from the upstream power grid, photovoltaic power generation, and wind power generation, coupled with micro-turbines, P2G devices, electrical energy storage, and electrical loads. Gas flows from the natural gas grid to supply micro-turbines and gas boilers, which power gas and heat loads, respectively. Electric chillers and absorption chillers power cooling loads. This integrated electric cooling and heating energy system model not only effectively facilitates the consumption of wind power resources but also utilizes the thermal energy generated during system operation, balancing the system's economic and environmental benefits.

[0082] Example 2:

[0083] This example optimizes scheduling based on Example 1, with the goal of minimizing total operating costs and maximizing power supply reliability. This can minimize the operating costs of the energy system, including fuel costs, operation and maintenance costs, etc., ensure the power supply reliability of the energy system, reduce power outages and failures, improve power supply reliability, meet user electricity needs, and reasonably allocate various energy resources, improve energy utilization efficiency, reduce energy waste, and maximize the utilization of energy resources.

[0084] Step 1: Build an integrated energy system with electricity, gas, cooling, and heating as energy flows, as shown in Case 1;

[0085] Step 2: Construct a multi-objective objective function with total operating cost and power supply reliability as the goals:

[0086] A multi-objective objective function is constructed with total operating cost and power supply reliability as the goals:

[0087]

[0088] C pb 、C ps 、C bess 、C GB 、C G are the main grid electricity purchase price, electricity sales price, battery cost, gas turbine operation and maintenance cost, and natural gas price; P Gin 、P Gout 、P bess 、P GB 、P MT They are power purchased from the main grid, power sold, battery operating power, gas turbine operating power, and micro gas turbine operating power; L pc is the penalty coefficient when the power constraint is not met; H sum is the total value of the current equilibrium constraint.

[0089] Energy reliability costs:

[0090]

[0091] C1 and C2 are the penalty prices for unit curtailment of solar and wind power, respectively; P pre_p 、P pre_w are the predicted photovoltaic output power and wind power output power respectively; P PV 、P WT are the actual photovoltaic output power and wind power output power respectively.

[0092] The objective function 1 is to reduce the system operating cost and improve the equipment operating efficiency by minimizing the total operating cost function, and the objective function 2 is to describe the power supply reliability by minimizing the wind power consumption;

[0093] Step 3: Taking equipment operation constraints, ramp constraints, and power balance constraints as constraints, the objective function and constraints are used to construct the optimal scheduling of the integrated energy system;

[0094] Equipment operation constraints:

[0095] 1) Photovoltaic unit output constraints

[0096] P PV_min ≤P PV ≤P PV_max

[0097] P PV_max 、P PV_min It is the upper and lower limits of the output of photovoltaic generator set.

[0098] 2) Wind turbine output constraints

[0099] P WT_min ≤P WT ≤P WT_max

[0100] P WT_max 、P WT_min The upper and lower limits of wind turbine output.

[0101] 3) Micro gas turbine output constraints:

[0102] P MT_min (t)≤P MT (t)≤P MT_max (t)

[0103] P MT (t)-P MT (t-1)≤L MT

[0104] P MT is the output power of the micro gas turbine; P MT_max 、P MT_min L is the upper and lower limits of the micro gas turbine output; MT is the ramp constraint limit of the micro gas turbine.

[0105] 4) Waste heat boiler output constraints:

[0106] P HRB_min (t)≤P HRB (t)≤P HRB_max (t)

[0107] P HRB is the output power of waste heat boiler; P HRB_max 、P HRB_min The upper and lower limits of waste heat boiler output.

[0108] 5) Gas boiler output constraints

[0109] P GB_min (t)≤P GB (t)≤P GB_max (t)

[0110] P GB is the output power of the gas boiler; P GB_max 、P GB_min The upper and lower limits of gas boiler output.

[0111] 6) Absorption chiller output constraints

[0112] P AC_min (t)≤P AC (t)≤P AC_max (t)

[0113] P AC is the output power of the absorption chiller; P AC_min 、P AC_max The upper and lower limits of the absorption chiller output.

[0114] 7) Electric refrigerator output constraints:

[0115] P ER_min (t)≤P ER (t)≤P ER_max (t)

[0116] P ER is the output power of the electric refrigerator; P ER_min 、P ER_max The upper and lower limits of the electric refrigerator output.

[0117] 8) P2G output constraints

[0118] P P2G_min (t)≤P P2G (t)≤P P2G_max (t)

[0119] P P2G is the P2G output power; P P2G_max 、PP2G_min The upper and lower limits of P2G output.

[0120] 9) Interaction constraints of the main network of the tie line

[0121] P Grid_min (t)≤P Grid (t)≤P Grid_max (t)

[0122] P Grid is the main network interaction power of the tie line; P Grid_max 、P Grid_min It is the upper and lower limits of the main network interaction of the interconnection line.

[0123] 10) Battery output limit:

[0124]

[0125] P BESS_min (t)≤P BESS (t)≤P BESS_max (t)

[0126] SOC min (t)≤SOC(t)≤SOC max (t)

[0127] are the capacity of the energy storage device at the initial moment and the capacity at the final moment, P BESS is the battery output power; P BESS_max 、P BESS_min The upper and lower limits of battery output; SOC is the battery capacity state; SOC max , SOC min They are the upper and lower limits of the battery capacity status respectively.

[0128] (2) Power balance constraints

[0129] 1) Power balance constraints of power grid

[0130] P PV (t)+P WT (t)+P MT (t)+P BESS (t)+P Grid (t) = P ER (t)+P P2G (t)+L ele (t)

[0131] L ele is the electrical load value.

[0132] 2) Gas network power balance constraints:

[0133] PP2G (t) = P WT (t)+P GB (t)+L gas (t)

[0134] L gas is the gas load value.

[0135] 3) Cold network power balance constraints:

[0136] P AC (t)+P ER (t) = L cold (t)

[0137] L cold is the cooling load value.

[0138] 4) Thermal network power balance constraints:

[0139] P GB (t)+P HRB (t) = P ER (t)+L hot (t)

[0140] L hot is the heat load value.

[0141] Step 4: Under the constraints of multiple objectives and equipment constraints, the improved particle swarm algorithm is used to solve the Pareto frontier and the equipment scheduling after the balance of electricity, gas, cooling and heating networks. The flow chart of the improved particle swarm algorithm in the optimization scheduling is as follows Figure 2 As shown in the figure, adaptive mutation is introduced to the traditional particle swarm algorithm. Adaptive mutation draws on the idea of ​​mutation in genetic algorithms and introduces a mutation operation into the PSO algorithm, which reinitializes certain variables with a certain probability. This operation expands the population search space, which is shrinking during iterations, allowing particles to move beyond the previously found optimal value and search in a larger space. This maintains population diversity and increases the algorithm's probability of finding the optimal value. Therefore, by introducing a simple mutation operator on top of the conventional particle swarm algorithm, particles are reinitialized with a certain probability after each update, demonstrating a strong global search capability.

[0142] The inertia weight factor of the traditional PSO algorithm is fixed and easy to fall into the local optimal value. To address this shortcoming, the particle swarm algorithm is improved from two aspects: the inertia weight factor and genetic variation:

[0143]

[0144] IT is the current iteration number; MI is the total iteration number; w s and w eare the initial and final values ​​of the inertia weight factor. In the early stages of the iteration, a larger w prevents the algorithm from falling into local minima, facilitating global search. In the later stages of the iteration, a smaller w facilitates local search and facilitates convergence of the algorithm.

[0145] The basic steps of MOPSO are:

[0146] (1) Initialization: Generate N in the D-dimensional search space s A population of particles, initializing the speed, position and external files of each particle in the population;

[0147] (2) Calculate the objective function value of each particle in the population and save the non-inferior solution to the external storage file according to the dominance relationship; initialize the individual optimal P of the particle best and the global optimal G best , is the initial position of the particle, G best is the best position of the initial population;

[0148] (3) Update the particle's velocity and position according to the position and velocity update formula;

[0149] (4) Calculate the objective function value of the particle and compare it with the P of the previous iteration best For comparison, re-adjust P best ;

[0150] (5) Perform hierarchical sorting on the population, store the optimal non-dominated Pareto solution into an external archive set, remove non-Pareto solutions, and determine whether the external archive set exceeds the specified capacity. If so, select m particles according to the crowding distance;

[0151] (6) Update the external storage set and determine the current global optimal G best , using the Pareto optimal solution saved in the external archive set, this paper uses the roulette method to select G from the external set according to the crowding distance of the optimal solution best .

[0152] (7) Small probability mutation: In order to prevent the MOPSO algorithm from converging to the local optimal frontier rather than the global optimal frontier too early, a small probability mutation mechanism is introduced to make the position of the particle at the initial position undergo a small probability perturbation of ±30%, thereby improving the search ability of the particle and increasing the particle's ability to find the global optimal frontier;

[0153] (8) Determine whether the termination condition is met. If so, output all optimal solutions in the external file. Otherwise, jump to step (3).

[0154] Assume that the total number of particles is N s , then the update formula for the position and velocity of the nth particle in dimension D is as follows:

[0155]

[0156] Where: ω' is the inertia weight factor, c1 and c2 are learning factors, reflecting the self-learning ability and social learning ability respectively, and r1 and r2 are random numbers uniformly distributed in [0,1].

Claims

1. A method for optimizing and scheduling an integrated energy system based on an improved particle swarm optimization algorithm, characterized by: An integrated energy system with electricity, gas, cooling, and heat as its energy flows is constructed, with the objective functions of minimizing operating costs and maximizing wind power consumption. The integrated energy system is optimized and dispatched. An improved particle swarm algorithm is used to solve the multi-objective problem, yielding a set of non-inferior solutions that minimize operating costs and maximize wind power consumption. These solutions lack significant room for improvement across multiple objective functions, meaning it is impossible to improve the value of one objective function without affecting the values ​​of others. The Pareto solution set encompasses multiple optimal solutions to a problem, demonstrating the trade-offs and compromises between different objectives, helping decision makers choose between them. Solving multi-objective optimization problems with the improved particle swarm algorithm yields a better Pareto solution set, helping to solve multi-objective optimization problems. The integrated energy system consists of an energy supply side, an equipment conversion side, and an energy consumption side. Scheduling primarily involves the output power of wind and solar energy, as well as the output power of the conversion equipment.

2. The integrated energy system according to claim 1, characterized in that Objective function and constraints of the integrated energy system: A multi-objective objective function is constructed with total operating cost and power supply reliability as the goals: C pb 、C ps 、C bess 、C GB 、C G are the main grid electricity purchase price, electricity sales price, battery cost, gas turbine operation and maintenance cost, and natural gas price; P Gin 、P Gout 、P bess 、P GB 、P MT They are power purchased from the main grid, power sold, battery operating power, gas turbine operating power, and micro gas turbine operating power; L pc is the penalty coefficient when the power constraint is not met; H sum is the total value of the current equilibrium constraint. Energy reliability costs: C1 and C2 are the penalty prices for unit curtailment of solar and wind power, respectively; P pre_p 、P pre_w are the predicted photovoltaic output power and wind power output power respectively; P PV 、P WT are the actual photovoltaic output power and wind power output power respectively. The objective function 1 is to reduce the system operating cost and improve the equipment operating efficiency by minimizing the total operating cost function, and the objective function 2 is to describe the power supply reliability by minimizing the wind power consumption.

3. The new integrated energy system and objective function according to claims 1 and 2, characterized in that: Taking equipment operation constraints, ramp constraints, and power balance constraints as constraints, the objective function and constraints are used to construct the optimal scheduling of the integrated energy system; (1) Equipment operation constraints: P PV_min ≤P PV ≤P PV_max P WT_min ≤P WT ≤P WT_max P MT_min (t)≤P MT (t)≤P MT_max (t) P MT (t)-P MT (t-1)≤L MT P HRB_min (t)≤P HRB (t)≤P HRB_max (t) P GB_min (t)≤P GB (t)≤P GB_max (t) P AC_min (t)≤P AC (t)≤P AC_max (t) P ER_min (t)≤P ER (t)≤P ER_max (t) P P2G_min (t)≤P P2G (t)≤P P2G_max (t) P Grid_min (t)≤P Grid (t)≤P Grid_max (t) P BESS_min (t)≤P BESS (t)≤P BESS_max (t) SOC min (t)≤SOC(t)≤SOC max (t) P PV_max 、P PV_min It is the upper and lower limits of the output of photovoltaic generator set. WT_max 、P WT_min It is the upper and lower limits of wind turbine output. MT is the output power of the micro gas turbine; P MT_max 、P MT_min L is the upper and lower limits of the micro gas turbine output; MT P is the ramp constraint limit of the micro gas turbine. HRB is the output power of waste heat boiler; P HRB_max 、P HRB_min It is the upper and lower limits of waste heat boiler output. GB is the output power of the gas boiler; P GB_max 、P GB_min It is the upper and lower limits of gas boiler output. AC is the output power of the absorption chiller; P AC_min 、P AC_max It is the upper and lower limits of the absorption chiller output. ER is the output power of the electric refrigerator; P ER_min 、P ER_max It is the upper and lower limits of the electric refrigerator output. P2G is the P2G output power; P P2G_max 、P P2G_min It is the upper and lower limits of P2G output. Grid is the main network interaction power of the tie line; P Grid_max 、P Grid_min It is the upper and lower limits of the main network interaction of the tie line. BESS is the battery output power; P BESS_max 、P BESS_min The upper and lower limits of battery output; SOC is the battery capacity status; SOC max , SOC min are the upper and lower limits of the battery capacity status, are the capacity of the energy storage device at the initial moment and the capacity at the final moment respectively. (2) Power balance constraints: mainly power balance constraints of the electric network, gas network, cooling network, and heating network: P PV (t)+P WT (t)+P MT (t)+P BESS (t)+P Grid (t)=P ER (t)+P P2G (t)+L ele (t) P P2G (t)=P WT (t)+P GB (t)+L gas (t) P AC (t)+P ER (t)=L cold (t) P GB (t)+P HRB (t)=P ER (t)+L hot (t) L ele is the electric load value, L gas is the gas load value, L cold is the cooling load value, L hot is the heat load value.

4. According to the above claim, it is characterized in that Improved particle swarm optimization algorithm: The inertia weight factor of the traditional PSO algorithm is fixed and prone to falling into local optimal values. To address this shortcoming, the particle swarm algorithm is improved from two aspects: inertia weight factor and genetic variation. The improved weight factor strategy is as follows: IT is the current iteration number; MI is the total number of iterations; w s and w e are the initial and final values ​​of the inertia weight factor. In the early stages of the iteration, a larger w prevents the algorithm from falling into local minima, facilitating global search. In the later stages of the iteration, a smaller w facilitates local search and facilitates convergence of the algorithm. The basic steps of the multi-objective particle swarm optimization algorithm are: (1) Initialization: Generate N in the D-dimensional search space s A population of particles, initializing the speed, position and external files of each particle in the population; (2) Calculate the objective function value of each particle in the population and save the non-inferior solution to the external storage file according to the dominance relationship; initialize the individual optimal P of the particle best and the global optimal G best , is the initial position of the particle, G best is the best position of the initial population; (3) Update the particle's velocity and position according to the position and velocity update formula; (4) Calculate the objective function value of the particle and compare it with the P of the previous iteration best For comparison, re-adjust P best ; (5) Perform hierarchical sorting on the population, store the optimal non-dominated Pareto solution into an external archive set, remove non-Pareto solutions, and determine whether the external archive set exceeds the specified capacity. If so, select m particles according to the crowding distance; (6) Update the external storage set and determine the current global optimal G best , using the Pareto optimal solution saved in the external archive set, this paper uses the roulette method to select G from the external set according to the crowding distance of the optimal solution best . (7) Small probability mutation: In order to prevent the MOPSO algorithm from converging to the local optimal frontier rather than the global optimal frontier too early, a small probability mutation mechanism is introduced to make the position of the particle at the initial position undergo a small probability perturbation of ±30%, thereby improving the search ability of the particle and increasing the particle's ability to find the global optimal frontier; (8) Determine whether the termination condition is met. If so, output all optimal solutions in the external file. Otherwise, jump to step (3). Assume that the total number of particles is N s , then the update formula for the position and velocity of the nth particle in dimension D is as follows: Where: ω' is the inertia weight factor, c1 and c2 are learning factors, reflecting the self-learning ability and social learning ability respectively, and r1 and r2 are random numbers uniformly distributed in [0,1].

Images