Source-load-storage coordinated two-stage industrial village and town energy scheduling method

Abstract

The invention discloses a source-load-storage coordinated two-stage industrial town energy scheduling method, which comprises the following steps of: constructing a mathematical model comprising photovoltaic power generation, an interruptible load, battery energy storage and a diesel generator based on data such as energy supply and demand states of an industrial town and a micro-grid and distributed source-load-storage operation; on this basis, a day-ahead-intra-day two-stage industrial town energy scheduling model is constructed, a layered energy scheduling optimization method is adopted in the day-ahead stage, and rolling optimization is introduced to determine an intra-day scheduling strategy with the industrial town extra cost caused by the minimum prediction deviation as the target in the intra-day stage; the two-stage industrial town energy scheduling model is solved through a linear optimization method, the industrial town and each micro-grid execute a scheduling strategy and update operation data, data are collected again after a scheduling period is finished, and then industrial town energy scheduling of the next scheduling period is carried out. According to the method, the key problems of difficulty in source-load-storage cooperative scheduling, lag in real-time supply and demand deviation adjustment, low complex model solving efficiency and the like are solved.

Description

Technical Field

[0001] The present invention belongs to the field of energy and power technology, and specifically relates to a two-stage industrial township energy scheduling method with source, load and storage coordination. Background Art

[0002] The coordinated scheduling of industrial townships and lower-level microgrids requires resolving the challenges of cross-level energy allocation and conflicting objectives. As regional energy hubs, industrial townships coordinate the energy storage, photovoltaics, diesel generators, and other resources of multiple microgrids with the core goal of reducing external electricity purchase costs. Microgrids, on the other hand, must prioritize local power supply reliability, leading to conflicts between the two in terms of energy allocation priorities and cost sharing. Limited cross-level information sharing makes it difficult for industrial townships to assess microgrid energy demand and scheduling potential in real time, resulting in a mismatch between scheduling instructions and the actual operating capabilities of microgrids. Furthermore, fluctuations in renewable energy output, transmitted through these hierarchical levels, further amplify the uncertainty of coordinated source-load-storage scheduling, making it difficult to achieve both economic efficiency and reliability.

[0003] Current research on energy scheduling between industrial townships and microgrids primarily focuses on centralized optimization and distributed coordination. Centralized optimization approaches aim for global economic efficiency, integrating the microgrid's source, load, and storage resources and formulating scheduling strategies using linear programming or heuristic algorithms. However, in practice, these approaches are susceptible to sudden meteorological changes or load fluctuations, leading to frequent deployment of system backups and increased operating costs. Distributed coordination strategies employ master-slave games or dynamic bidding mechanisms to coordinate the interests of both parties using electricity price signals or power interaction protocols. However, in the high-energy consumption scenarios of industrial townships, these strategies suffer from information exchange delays and asynchrony in policy responses, leading to supply-demand imbalances. Furthermore, researchers are using machine learning models to mine energy supply and demand patterns from historical data to generate dynamic scheduling strategies. While these approaches can improve adaptability to complex scenarios, they generally face the challenge of insufficiently embedding physical constraints. Some studies have attempted to improve model adaptability by introducing robust optimization or fuzzy logic. However, in the deeply coupled scenarios of large-scale energy demand in industrial townships and flexible energy supply from microgrids, comprehensive solutions that balance real-time response speed, multi-objective trade-offs, and cross-system stability are lacking. Summary of the Invention

[0004] To solve the above technical problems, the present invention provides a two-stage industrial township energy scheduling method with source-load-storage coordination. First, based on the energy supply and demand status of industrial townships and microgrids and the operation data of distributed sources, loads and storage, a mathematical model including photovoltaic power generation, interruptible loads, battery energy storage and diesel generators is constructed; secondly, a two-stage industrial township energy scheduling model, day-ahead and intraday, is constructed. The day-ahead stage adopts a hierarchical energy scheduling optimization method. The upper layer aims to minimize the total operating cost of the industrial township, and the lower layer aims to minimize the operating cost of each microgrid; the intraday stage aims to minimize the additional cost of the industrial township caused by the prediction deviation, and the rolling optimization is introduced to determine the intraday scheduling strategy; the optimal energy scheduling strategy is solved by the linear optimization method, the industrial township and each microgrid execute the scheduling strategy and update the operating data. After the scheduling cycle ends, the data is collected again and the next scheduling cycle begins.

[0005] The two-stage industrial township energy dispatching method with source-load-storage coordination described in the present invention comprises the following steps:

[0006] S1. Based on the energy supply and demand status of industrial towns and microgrids and the operation of distributed sources, loads and storage, a mathematical model including photovoltaic power generation, interruptible loads, battery energy storage and diesel generators is constructed;

[0007] S2. Construct a two-stage energy dispatch model for industrial townships: day-ahead and intraday. A hierarchical energy dispatch optimization method is adopted in the day-ahead stage. The upper layer aims to minimize the total operating cost of industrial townships, and the lower layer aims to minimize the operating cost of each microgrid. The intraday stage aims to minimize the additional cost of industrial townships caused by forecast deviations, and a rolling optimization is introduced to determine the intraday dispatch strategy.

[0008] S3. Solve the two-stage industrial township energy dispatch model through linear optimization method. The industrial township and each microgrid implement the dispatch strategy and update the operation data. After the dispatch cycle ends, return to S1 to collect data, and then carry out the industrial township energy dispatch for the next dispatch cycle.

[0009] Furthermore, S1 builds a mathematical model that includes photovoltaic power generation, interruptible loads, battery energy storage, and diesel generators based on data on the energy supply and demand status of industrial towns and microgrids and the operation of distributed sources, loads, and storage. Specifically:

[0010] a. Photovoltaic power generation modeling:

[0011] The output power of a photovoltaic array is affected by the size and efficiency of its photovoltaic modules and is mainly determined by the irradiance. Therefore, the photovoltaic system is modeled as shown in formula (1):

[0012] P PV =η PV ·A PV ·G PV{1+φ(τ-25)} (1)

[0013] Where η PV is the overall efficiency of the photovoltaic panel, A PV is the total area of ​​photovoltaic panels, G PV is the solar radiation, τ is the outdoor air temperature, and φ is the temperature coefficient corresponding to the maximum output power;

[0014] The photovoltaic operation and maintenance cost is shown in formula (2):

[0015] C PV =c PV ·P PV (2)

[0016] Where c PV PV operation and maintenance costs;

[0017] b. Interruptible load model:

[0018]

[0019] 0≤P t IL ≤εP t load (4)

[0020] Where c IL is the interruptible load cost coefficient, P t IL is the interruptible load power, ε is the interruptible load factor, P t load is the load power;

[0021] c.Battery energy storage model:

[0022] The operating cost of battery energy storage is shown in formula (5):

[0023]

[0024] Where c BESS is the battery energy storage operation cost coefficient, P t BESS,ch 、P t BESS,dis are the battery energy storage charging and discharging power at time t, is the stored energy of the battery at time t, η BESS,loss is the battery energy storage loss coefficient;

[0025]

[0026] Where α, β are binary variables to avoid simultaneous charging and discharging of battery energy storage, and α+β=1; α, β∈{0,1}, is the battery energy storage charge state, E BESS is the rated storage energy of the battery, η BESS,ch ,η BESS,dis are the battery energy storage charging and discharging efficiency;

[0027]

[0028] Where, are the initial and final state of charge of the battery energy storage system, They are the upper and lower limits of the state of charge of the battery energy storage system, are the maximum charge and discharge power allowed by the battery energy storage system;

[0029] d. Diesel generator model:

[0030] Diesel generator cost model:

[0031]

[0032] Where, P t DG is the output power of the diesel engine, c DG is the diesel generator cost coefficient;

[0033]

[0034] Where, They are the upper and lower limits of diesel generator output power respectively;

[0035] e. Uncertainty model:

[0036] The change in light intensity follows the Beta distribution, and its probability density function is:

[0037]

[0038] Where a * 、b * is the shape parameter of the Beta distribution, G, G max are the actual light intensity and maximum light intensity in each time period respectively;

[0039] The fluctuation of load and electricity price obeys normal distribution, and its probability density function is:

[0040]

[0041] In the formula, Δx is the fluctuation of load and electricity price, μ, δ 2 are the expectation and variance of the normal distribution respectively;

[0042] Furthermore, in S2, a hierarchical energy scheduling optimization method is adopted in the day-ahead stage. The upper layer aims to minimize the total operating cost of industrial townships, and the lower layer aims to minimize the operating cost of each microgrid. Specifically,

[0043] a. Upper model:

[0044] In the upper model, industrial towns act as leaders, optimizing both operating costs and system backup, with the objective function of minimizing the total cost of energy dispatch in industrial towns:

[0045]

[0046] Where, ρ s is the probability of scenario s, S is the total number of scenarios, T is the total scheduling period, n is the battery energy storage number, N is the total number of battery energy storage, is the energy dispatch cost coefficient between industrial towns and the upper-level power grid, is the energy dispatching power between industrial towns and the upper power grid, i is the microgrid number, I is the total number of microgrids, is the energy dispatch cost coefficient between the industrial township and microgrid i, is the energy dispatching power between the industrial township and microgrid i;

[0047]

[0048] Where, They are the upper and lower limits of energy dispatch power between industrial towns and the upper power grid, are the upper and lower limits of energy dispatch power between the industrial township and microgrid i, respectively;

[0049] b. Lower layer model:

[0050] In the lower model, each microgrid acts as a slave, with the objective function of minimizing the operating cost of each microgrid in the industrial township:

[0051]

[0052] c. Linearization of the two-layer model:

[0053] Use KKT conditions to transform the lower model:

[0054]

[0055] Where, are the constraints and the corresponding Lagrange multipliers respectively, M is a large constant, take M = 10 6 , is a binary variable;

[0056] Introduce the Lagrangian function of the two-layer model:

[0057]

[0058]

[0059] Using the equality constraints in KKT conditions and strong duality theory to solve the bilinear term Perform linearization:

[0060]

[0061] The strong duality theory equality constraint of the lower model is shown in formula (29):

[0062]

[0063] The original two-layer problem is transformed into a linearized single-layer problem:

[0064]

[0065] Furthermore, in S2, the intraday stage aims to minimize the additional costs of industrial towns caused by forecast deviations, and introduces rolling optimization to determine the intraday scheduling strategy, specifically:

[0066]

[0067] Where, The additional compensation power penalty coefficient provided to the upper power grid due to the forecast deviation, ΔP t in ,punish The additional compensation power provided to the upper grid due to the forecast deviation, ΔP t BESS,punish Additional compensation power provided by battery energy storage due to forecast deviations, are the diesel generator, photovoltaic power generation, interruptible load, and load forecast deviation values ​​in microgrid i respectively;

[0068] The steps of the intraday real-time rolling scheduling strategy are as follows:

[0069] a. Divide a day into 96 periods and perform optimization every 15 minutes. The control horizon ψ is four periods, i.e., 1 hour.

[0070] b. In the tth period, optimize the energy scheduling strategy for the industrial township by using ψ as a fixed intraday scheduling period. Based on the real-time data of the tth period and the forecast data of the next three periods, only the scheduling strategy for the tth period is implemented.

[0071] c. Scroll forward to the t+1th period and repeat step b until the optimization of 96 periods is completed;

[0072] Furthermore, in S3, a two-stage industrial township energy scheduling model is solved by a linear optimization method, specifically:

[0073] The two-stage industrial township energy scheduling optimization model with source-load-storage coordination established above is a mixed integer linear programming problem. In the Matlab environment, Yalmip is used to model it and the commercial solver LINPROG is called to solve it.

[0074] Furthermore, S3 is specifically:

[0075] S3-1. The industrial township dispatch center will distribute the optimal energy dispatch strategy to each microgrid, including coordinated photovoltaic output, energy storage charging and discharging strategies, load interruption control, and diesel generator output control, and dynamically adjust the energy interaction strategy with the upper-level power grid;

[0076] S3-2. After receiving the instructions, each entity performs energy dispatch according to the requirements of the strategy. When a dispatch cycle ends, the dispatch center collects the status information of each microgrid in the industrial township, including operating data, energy demand and network parameters;

[0077] S3-3. Synchronize the updated status information to the industrial township dispatch center. The system restarts data collection and analysis and enters the next dispatch cycle based on the latest data.

[0078] This paper proposes a two-stage energy dispatch method for industrial townships that integrates sources, loads, and storage. First, a mathematical model is constructed that includes photovoltaic power generation, interruptible loads, battery energy storage, and diesel generators to quantify the dynamic characteristics and operational constraints of sources, loads, and storage. Second, a two-stage energy dispatch model for industrial townships (day-ahead and intraday) is constructed. In the day-ahead stage, a two-tier decision-making mechanism is used to coordinate global economic efficiency and distributed autonomy. In the intraday stage, a rolling optimization approach dynamically corrects dispatch deviations caused by forecast deviations, minimizing additional operating costs to improve system robustness and real-time responsiveness. Finally, a linear optimization method is used to solve the energy dispatch strategy for industrial townships. The industrial townships and their microgrids execute the dispatch strategy and update operational data. After the dispatch cycle ends, data is collected again and the next dispatch cycle begins, balancing economic efficiency, reliability, and adaptability. BRIEF DESCRIPTION OF THE DRAWINGS

[0079] Figure 1 A flow chart of the two-stage industrial township energy dispatch method with source-load-storage coordination;

[0080] Figure 2 This is a schematic diagram of the energy dispatch architecture among the upper power grid, industrial townships, and microgrids;

[0081] Figure 3Detailed flow chart of the two-stage industrial township energy dispatch method with source-load-storage coordination DETAILED DESCRIPTION

[0082] In order to make the contents of the present invention more clearly understood, the following are specific embodiments in combination with the accompanying drawings: The following embodiments are only used to more clearly illustrate the technical solutions of the present invention and are not intended to limit the scope of protection of this application.

[0083] Figure 1 A flow chart of a two-stage industrial township energy dispatching method with source-load-storage coordination provided by an example of the present invention;

[0084] Figure 2 A schematic diagram of the energy dispatch architecture among a higher-level power grid, an industrial township, and a microgrid, provided in one embodiment of the present invention;

[0085] Figure 3 Detailed flow chart of a two-stage industrial township energy dispatching method with source-load-storage coordination provided by an example of the present invention

[0086] like Figure 1 As shown, the method includes:

[0087] Step S1: Based on the energy supply and demand status of the industrial township and microgrid and the distributed source-load-storage operation data, a mathematical model including photovoltaic power generation, interruptible loads, battery energy storage and diesel generators is constructed;

[0088] a. Photovoltaic power generation modeling:

[0089] The output power of a photovoltaic array is affected by the size and efficiency of its photovoltaic modules and is mainly determined by the irradiance. Therefore, the photovoltaic system is modeled as shown in formula (1):

[0090] P PV =η PV ·A PV ·G PV {1+φ(τ-25)} (1)

[0091] Where η PV is the overall efficiency of the photovoltaic panel, A PV is the total area of ​​photovoltaic panels, G PV is solar radiation, τ is the outdoor air temperature, φ is the temperature coefficient corresponding to the maximum output power;

[0092] The photovoltaic operation and maintenance cost is shown in formula (2):

[0093] C PV =c PV ·P PV (2)

[0094] Where cPV PV operation and maintenance costs;

[0095] b. Interruptible load model:

[0096]

[0097] 0≤P t IL ≤εP t load (4)

[0098] Where c IL is the interruptible load cost coefficient, P t IL is the interruptible load power, ε is the interruptible load factor, P t load is the load power;

[0099] c.Battery energy storage model:

[0100] The operating cost of battery energy storage is shown in formula (5):

[0101]

[0102] Where c BESS is the battery energy storage operation cost coefficient, P t BESS,ch 、P t BESS,dis are the battery energy storage charging and discharging power at time t, is the stored energy of the battery at time t, η BESS,loss is the battery energy storage loss coefficient;

[0103]

[0104] Where α, β are binary variables to avoid simultaneous charging and discharging of battery energy storage, and α+β=1; α, β∈{0,1}, is the battery energy storage charge state, E BESS is the rated storage energy of the battery, η BESS,ch ,η BESS,dis are the battery energy storage charging and discharging efficiency;

[0105]

[0106] Where, are the initial and final state of charge of the battery energy storage system, They are the upper and lower limits of the state of charge of the battery energy storage system, are the maximum charge and discharge power allowed by the battery energy storage system;

[0107] d. Diesel generator model:

[0108] Diesel generator cost model:

[0109]

[0110] Where, P t DG is the output power of the diesel engine, c DG is the diesel generator cost coefficient;

[0111]

[0112] Where, They are the upper and lower limits of diesel generator output power respectively;

[0113] e. Uncertainty model:

[0114] The change in light intensity follows the Beta distribution, and its probability density function is:

[0115]

[0116] Where a * 、b * is the shape parameter of the Beta distribution, G, G max are the actual light intensity and maximum light intensity in each time period respectively;

[0117] The fluctuation of load and electricity price obeys normal distribution, and its probability density function is:

[0118]

[0119] In the formula, Δx is the fluctuation of load and electricity price, μ, δ 2 are the expectation and variance of the normal distribution respectively;

[0120] Step S2: Construct a two-stage industrial township energy dispatch model: day-ahead and intraday. A hierarchical energy dispatch optimization method is adopted in the day-ahead stage. The upper layer aims to minimize the total operating cost of industrial townships, and the lower layer aims to minimize the operating cost of each microgrid. The intraday stage aims to minimize the additional cost of industrial townships caused by forecast deviations, and a rolling optimization is introduced to determine the intraday dispatch strategy.

[0121] In the upper model, industrial towns act as leaders, optimizing both operating costs and system backup, with the objective function of minimizing the total cost of energy dispatch in industrial towns:

[0122]

[0123] Where, ρ sis the probability of scenario s, S is the total number of scenarios, T is the total scheduling period, n is the battery energy storage number, N is the total number of battery energy storage, is the energy dispatch cost coefficient between industrial towns and the upper-level power grid, is the energy dispatching power between industrial towns and the upper power grid, i is the microgrid number, I is the total number of microgrids, is the energy dispatch cost coefficient between the industrial township and microgrid i, is the energy dispatching power between the industrial township and microgrid i;

[0124]

[0125] Where, They are the upper and lower limits of energy dispatch power between industrial towns and the upper power grid, are the upper and lower limits of energy dispatch power between the industrial township and microgrid i, respectively;

[0126] b. Lower layer model:

[0127] In the lower model, each microgrid acts as a slave, with the objective function of minimizing the operating cost of each microgrid in the industrial township:

[0128]

[0129] c. Linearization of the two-layer model:

[0130] Use KKT conditions to transform the lower model:

[0131]

[0132] Where, are the constraints and the corresponding Lagrange multipliers respectively, M is a large constant, take M = 10 6 , is a binary variable;

[0133] Introduce the Lagrangian function of the two-layer model:

[0134]

[0135]

[0136] Using the equality constraints in KKT conditions and strong duality theory to solve the bilinear term Perform linearization:

[0137]

[0138] The strong duality theory equality constraint of the lower model is shown in formula (29):

[0139]

[0140] The original two-layer problem is transformed into a linearized single-layer problem:

[0141]

[0142] Then, in the intraday stage, with the goal of minimizing the additional costs of industrial towns caused by forecast deviations, rolling optimization is introduced to determine the intraday scheduling strategy, specifically:

[0143]

[0144] Where, The additional compensation power penalty coefficient provided to the upper power grid due to the forecast deviation, ΔP t in ,punish The additional compensation power provided to the upper grid due to the forecast deviation, ΔP t BESS,punish Additional compensation power provided by battery energy storage due to forecast deviations, are the diesel generator, photovoltaic power generation, interruptible load, and load forecast deviation values ​​in microgrid i respectively;

[0145] The steps of the intraday real-time rolling scheduling strategy are as follows:

[0146] a. Divide a day into 96 periods and perform optimization every 15 minutes. The control horizon ψ is four periods, i.e., 1 hour.

[0147] b. In the tth period, optimize the energy scheduling strategy for the industrial township by using ψ as a fixed intraday scheduling period. Based on the real-time data of the tth period and the forecast data of the next three periods, only the scheduling strategy for the tth period is implemented.

[0148] c. Scroll forward to the t+1th period and repeat step b until the optimization of 96 periods is completed.

[0149] Step S3: Solve the two-stage industrial township energy scheduling model through the linear optimization method. The industrial township and each microgrid implement the scheduling strategy and update the operating data. After the scheduling cycle ends, return to S1 to collect data, and then perform the industrial township energy scheduling for the next scheduling cycle.

[0150] The two-stage industrial township energy scheduling optimization model with source-load-storage coordination established above is a mixed integer linear programming problem. In the Matlab environment, Yalmip is used to model it and the commercial solver LINPROG is called to solve it.

[0151] S3-1. The industrial township dispatch center will distribute the optimal energy dispatch strategy to each microgrid, including coordinated photovoltaic output, energy storage charging and discharging strategies, load interruption control, and diesel generator output control, and dynamically adjust the energy interaction strategy with the upper-level power grid;

[0152] S3-2. After receiving the instructions, each entity performs energy dispatch according to the requirements of the strategy. When a dispatch cycle ends, the dispatch center collects the status information of each microgrid in the industrial township, including operating data, energy demand and network parameters;

[0153] S3-3. Synchronize the updated status information to the industrial township dispatch center. The system restarts data collection and analysis and enters the next dispatch cycle based on the latest data.

[0154] It should be noted that the above embodiments are not intended to limit the scope of protection of the present invention, and equivalent changes or substitutions made on the basis of the above technical solutions fall within the scope of protection of the claims of the present invention.

Claims

1. A two-stage industrial township energy dispatching method with source-load-storage coordination, characterized by: The following steps are involved: S1. Based on the energy supply and demand status of industrial towns and microgrids and the operating data of distributed sources, loads and storage, a mathematical model including photovoltaic power generation, interruptible loads, battery energy storage and diesel generators is constructed; S2. Construct a two-stage energy dispatch model for industrial towns and villages: day-ahead and intraday. A hierarchical energy dispatch optimization method is used in the day-ahead stage. The upper layer aims to minimize the total operating cost of industrial towns and villages, while the lower layer aims to minimize the operating cost of each microgrid. The intraday stage aims to minimize the additional cost of industrial towns and villages caused by forecast deviations, and a rolling optimization method is introduced to determine the intraday dispatch strategy. S3. Solve the two-stage industrial township energy dispatch model through linear optimization method. The industrial township and each microgrid implement the dispatch strategy and update the operation data. After the dispatch cycle ends, return to S1 to collect data, and then carry out the industrial township energy dispatch for the next dispatch cycle.

2. A two-stage industrial township energy dispatching method with source-load-storage coordination according to claim 1, characterized in that: In S1, based on the energy supply and demand status of industrial towns and microgrids and the operation of distributed sources, loads and storage, a mathematical model including photovoltaic power generation, interruptible loads, battery energy storage and diesel generators was constructed. Specifically: a. Photovoltaic power generation modeling: The output power of a photovoltaic array is affected by the size and efficiency of its photovoltaic modules and is mainly determined by the irradiance. Therefore, the photovoltaic system is modeled as shown in formula (1): P PV =the PV ·A PV ·G PV {1+φ(τ-25)} (1) Where, η PV is the overall efficiency of the photovoltaic panel, A PV is the total area of ​​photovoltaic panels, G PV is solar radiation, τ is the outdoor air temperature, φ is the temperature coefficient corresponding to the maximum output power; The photovoltaic operation and maintenance cost is shown in formula (2): C PV =c PV ·P PV (2) Where c PV PV operation and maintenance costs; b. Interruptible load model: 0≤P t IL ≤εP t load (4) Where c IL is the interruptible load cost coefficient, P t IL is the interruptible load power, ε is the interruptible load factor, P t load is the load power; c.Battery energy storage model: The operating cost of battery energy storage is shown in formula (5): Where c BESS is the battery energy storage operation cost coefficient, P t BESS,ch 、P t BESS,dis are the battery energy storage charging and discharging power at time t, is the stored energy of the battery at time t, η BESS,loss is the battery energy storage loss coefficient; Where α, β are binary variables to avoid simultaneous charging and discharging of battery energy storage, and α+β=1; α, β∈{0,1}, is the battery energy storage charge state, E BESS is the rated storage energy of the battery, η BESS,ch ,η BESS,dis are the battery energy storage charging and discharging efficiency; Where, are the initial and final state of charge of the battery energy storage system, They are the upper and lower limits of the state of charge of the battery energy storage system, are the maximum charge and discharge power allowed by the battery energy storage system; d. Diesel generator model: Diesel generator cost model: C t DG =c DG ·P t DG (12) Where, P t DG is the output power of the diesel engine, c DG is the diesel generator cost coefficient; Where, They are the upper and lower limits of diesel generator output power respectively; e. Uncertainty model: The change in light intensity follows the Beta distribution, and its probability density function is: Where a * 、b * is the shape parameter of the Beta distribution, G, G max are the actual light intensity and maximum light intensity in each time period respectively; The fluctuation of load and electricity price obeys normal distribution, and its probability density function is: Where Δx is the fluctuation of load and electricity price, μ , δ 2 are the expectation and variance of the normal distribution, respectively.

3. The two-stage industrial township energy dispatching method with source-load-storage coordination according to claim 1 is characterized in that: In S2, a hierarchical energy scheduling optimization method is used in the day-ahead stage. The upper layer aims to minimize the total operating cost of industrial townships, while the lower layer aims to minimize the operating cost of each microgrid. Specifically, a. Upper model: In the upper model, industrial towns act as leaders, optimizing both operating costs and system backup, with the objective function of minimizing the total cost of energy dispatch in industrial towns: Where, ρ s is the probability of scenario s, S is the total number of scenarios, T is the total scheduling period, n is the battery energy storage number, N is the total number of battery energy storage, is the energy dispatch cost coefficient between industrial towns and the upper-level power grid, is the energy dispatching power between industrial towns and the upper power grid, i is the microgrid number, I is the total number of microgrids, is the energy dispatch cost coefficient between the industrial township and microgrid i, is the energy dispatching power between the industrial township and microgrid i; Where, They are the upper and lower limits of energy dispatch power between industrial towns and the upper power grid, are the upper and lower limits of energy dispatch power between the industrial township and microgrid i, respectively; b. Lower layer model: In the lower model, each microgrid acts as a slave, with the objective function of minimizing the operating cost of each microgrid in the industrial township: c. Linearization of the two-layer model: Use KKT conditions to transform the lower model: Where, are the constraints and the corresponding Lagrange multipliers respectively, M is a large constant, take M = 10 6 , is a binary variable; Introduce the Lagrangian function of the two-layer model: Using the equality constraints in KKT conditions and strong duality theory to solve the bilinear term Perform linearization: The strong duality theory equality constraint of the lower model is shown in formula (29): The original two-layer problem is transformed into a linearized single-layer problem:

4. The two-stage industrial township energy dispatching method with source-load-storage coordination according to claim 1 is characterized in that: In S2, In the intraday stage, the goal is to minimize the additional costs of industrial towns caused by forecast deviations. Rolling optimization is introduced to determine the intraday scheduling strategy, specifically: Where, The additional compensation power penalty coefficient provided to the upper power grid due to the forecast deviation, The additional compensation power provided to the upper grid due to the forecast deviation, ΔP t BESS,punish Additional compensation power provided by battery energy storage due to forecast deviations, are the diesel generator, photovoltaic power generation, interruptible load, and load forecast deviation values ​​in microgrid i respectively; The steps of the intraday real-time rolling scheduling strategy are as follows: a. Divide a day into 96 periods and perform optimization every 15 minutes. The control horizon ψ is four periods, i.e., 1 hour. b. In the tth period, optimize the energy scheduling strategy for the industrial township by using ψ as a fixed intraday scheduling period. Based on the real-time data of the tth period and the forecast data of the next three periods, only the scheduling strategy for the tth period is implemented. c. Scroll forward to the t+1th period and repeat step b until the optimization of 96 periods is completed.

5. The two-stage industrial township energy dispatching method with source-load-storage coordination according to claim 1 is characterized in that: In S3, the two-stage industrial township energy scheduling model is solved by linear optimization method, specifically: The two-stage industrial township energy scheduling optimization model with source-load-storage coordination established above is a mixed integer linear programming problem. In the Matlab environment, Yalmip is used to model it and the commercial solver LINPROG is called to solve it.

6. The two-stage industrial township energy dispatching method with source-load-storage coordination according to claim 1 is characterized in that: S3 specifically: S3-1. The industrial township dispatch center will distribute the optimal energy dispatch strategy to each microgrid, including coordinated photovoltaic output, energy storage charging and discharging strategies, load interruption control, and diesel generator output control, and dynamically adjust the energy interaction strategy with the upper-level power grid; S3-2. After receiving the instructions, each entity performs energy dispatch according to the requirements of the strategy. When a dispatch cycle ends, the dispatch center collects the status information of each microgrid in the industrial township, including operating data, energy demand and network parameters; S3-3. Synchronize the updated status information to the industrial township dispatch center. The system restarts data collection and analysis and enters the next dispatch cycle based on the latest data.

7. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, it implements a two-stage industrial township energy scheduling method with source, load and storage coordination as described in any one of claims 1 to 6 above.

8. A computer-readable storage medium having computer instructions stored thereon, characterized in that: When the computer instructions are executed by the processor, a two-stage industrial township energy scheduling method with source, load and storage coordination is implemented as described in any one of claims 1-6.

Images