Multi-time scale based microgrid multi-element energy storage collaborative optimization method and system

Abstract

The present application proposes a multi-time scale based micro-grid multi-element energy storage collaborative optimization method and system, which takes the lowest life cycle cost as the core optimization target, considers the flexible adjustment ability of hydrogen energy storage system, electrochemical energy storage system and demand response, and constructs a multi-time scale micro-grid multi-element energy storage collaborative optimization configuration model; the first stage and the second stage constraint conditions of the micro-grid multi-element energy storage collaborative optimization configuration model are constructed; the improved column generation algorithm is used to solve the micro-grid multi-element energy storage collaborative optimization configuration model, and the optimal capacity configuration scheme and real-time operation optimization strategy of the micro-grid are obtained. Through multi-time scale modeling and two-stage collaborative optimization solution, the present application quantifies the internal correlation between multi-element energy storage configuration and micro-grid energy balance, significantly reduces the system operation cost, and takes into account the effectiveness and economy of the configuration scheme, thereby providing technical support for stable and efficient operation of the micro-grid.

Description

Technical Field

[0001] This invention belongs to the field of microgrid optimization scheduling and energy storage configuration technology, specifically involving a method and system for collaborative optimization of multi-element energy storage in microgrids based on multiple time scales. Background Technology

[0002] Currently, the integration of high-proportion renewable energy sources such as wind and solar power into microgrids has become a core direction of energy transformation. However, their output exhibits significant multi-timescale volatility—including short-term fluctuations at the minute and hourly levels within a day, as well as long-term output imbalances due to seasonal changes. Coupled with the random variations in user load, this leads to key issues such as source-load supply-demand imbalance and low renewable energy absorption rates commonly faced by microgrids. Energy storage systems, as a core means of smoothing fluctuations and improving system flexibility, directly determine the economic efficiency and operational reliability of the microgrid throughout its entire lifecycle through the rationality of their capacity and power configuration. However, traditional microgrid energy storage configuration schemes still have significant shortcomings: most schemes focus only on a single timescale, lacking multi-timescale collaborative design; at the same time, the functional complementarity of diverse energy storage solutions has not been fully explored, easily leading to problems such as wasted energy storage investment, high operating costs, or severe renewable energy curtailment, making it difficult to adapt to the complex source-load characteristics and efficient operation requirements of microgrids.

[0003] Existing microgrid energy storage configuration schemes include: research on multi-timescale optimization operation of integrated energy systems based on multi-element energy storage, proposing a three-stage multi-timescale optimization method of day-ahead, intraday, and real-time, which successfully reduces the number of lithium battery charge-discharge cycles from 131 to 13 by synergistically smoothing high-frequency and low-frequency fluctuations through supercapacitors and lithium batteries, significantly reducing operating costs. However, this research focuses on output optimization during system operation and does not involve the capacity and power synergistic configuration of multi-element energy storage, thus failing to provide guidance for energy storage hardware selection and parameter determination in the early stages of engineering, exhibiting a limitation of emphasizing operation while neglecting configuration; and power-capacity decoupling configuration of hydrogen energy storage and battery energy storage. One approach employs an improved Benders decomposition algorithm to solve the configuration problem on an annual scale, effectively reducing the total system cost. However, this method does not construct closed-loop constraints for long-term and short-term energy storage states, which can easily lead to a disconnect between long-term and short-term operations. Another approach is to establish a microgrid dispatch model that coordinates wind, solar, hydrogen, and batteries. This model improves the utilization rate of renewable energy by adjusting hydrogen energy storage across seasons and balancing the daily power of batteries. However, this model aims to minimize daily operating costs and does not integrate the economics of the entire life cycle. Furthermore, it does not clarify the intrinsic relationship between the characteristics of source and load fluctuations (such as fluctuation frequency and amplitude) and energy storage configuration parameters, resulting in insufficient balance between the engineering applicability and economic efficiency of the configuration scheme.

[0004] In summary, although existing research on microgrid energy storage configuration has made progress in single-time-scale optimization or single-energy storage type application, the following key shortcomings still exist. Therefore, it is urgent to develop a new energy storage configuration method to overcome the existing technical bottlenecks. Summary of the Invention

[0005] The purpose of this invention is to address the aforementioned problems in the existing technology by providing a method and system for collaborative optimization of multi-timescale microgrid energy storage.

[0006] To achieve the above objectives, the technical solution of the present invention is as follows:

[0007] Firstly, this invention proposes a multi-time-scale microgrid multi-element energy storage collaborative optimization method, including:

[0008] S1. With the lowest life-cycle cost as the core optimization objective, and considering the hydrogen energy storage system, electrochemical energy storage system, and the flexible adjustment capability of demand response, a multi-time-scale microgrid multi-energy storage collaborative optimization configuration model is constructed.

[0009] S2. An improved column generation algorithm is used to solve the microgrid multi-energy storage collaborative optimization configuration model, and obtain the optimal capacity configuration scheme and real-time operation optimization strategy of the microgrid. The improved column generation algorithm realizes two iterations of the algorithm through dynamic column filtering.

[0010] In S1, the objective function of the microgrid multi-energy storage collaborative optimization configuration model is:

[0011] ;

[0012] ;

[0013] ;

[0014] ;

[0015] ;

[0016] In the above formula, For total lifecycle cost, For investment costs, For operation and maintenance costs, For replacement costs, For grid interaction costs, For electrochemical energy storage unit power, This is the rated power of electrochemical energy storage. The energy investment cost for electrochemical energy storage, For the rated energy capacity of electrochemical energy storage, For hydrogen energy storage unit power, This is the rated power for hydrogen energy storage. The energy investment cost for hydrogen energy storage, For the rated energy capacity of hydrogen energy storage, For the total lifespan of the energy storage system, The discount rate for funds. The annual operating and maintenance cost per unit energy of electrochemical energy storage. , The annual operating and maintenance cost per unit energy of hydrogen energy storage. This refers to the total number of replacements for the energy storage system throughout its entire lifecycle. For a single replacement cycle, The total initial investment cost for electrochemical energy storage, The total initial investment cost for hydrogen energy storage. , The first The year's first sky The power that a microgrid purchases from the main grid and sells back to the main grid at any given time. , These refer to the electricity purchase price and the electricity sales price from the power grid.

[0017] The constraints of the microgrid multi-energy storage collaborative optimization configuration model include hourly power balance constraints, daily rolling annual energy balance constraints, and source-load uncertainty quantification constraints.

[0018] The power balance constraint for hourly operation is:

[0019] ;

[0020] In the above formula, for The power exchange of the power grid at any given time. , They are respectively The constant charging and discharging power of electrochemical energy storage , They are respectively The charging and discharging power of hydrogen energy storage at all times. , They are respectively The load reduction power and load transfer power required for real-time demand response. for Total load power of the microgrid at any given time , They are respectively Real-time photovoltaic and wind power output;

[0021] The annual energy balance constraint for daily rolling is:

[0022] ;

[0023] In the above formula, For the number of days in a year, , The first sky The power input and output of the microgrid to the mains at any given time;

[0024] The source load uncertainty quantification constraint is:

[0025] ;

[0026] ;

[0027] ;

[0028] ;

[0029] In the above formula, This is the quantification range for source-load uncertainty power, which includes photovoltaic, wind power, and load. The source load uncertainty fluctuation coefficient, For the first sky Predicted power for time-source load uncertainty , , The first sky Always consider the actual power range of photovoltaic, wind power, and load after taking into account uncertainties. These are the uncertainty fluctuation coefficients for photovoltaic, wind power, and load, respectively. , , The first sky Predicted power output of photovoltaic, wind power, and load at all times.

[0030] S2 includes:

[0031] S21. Based on representative scenarios of source-load uncertainty, an initial column is randomly generated within the feasible region of variables in the microgrid multi-electrode energy storage collaborative optimization configuration model to form an initial set. ,in The integer capacity column corresponds to the first-stage capacity configuration and is generated. Integer capacity vectors that satisfy capacity and allocation constraints ; For continuous operation columns, corresponding to the second stage of operation optimization, each group of integer columns is matched. A group of continuous running vectors that satisfy the running constraints ;

[0032] S22. Based on integer capacity vectors and continuously running vectors, introduce combination coefficients to construct the main problem:

[0033] ;

[0034] ;

[0035] ;

[0036] ;

[0037] ;

[0038] In the above formula, The total lifecycle cost of the main problem For the first The combination coefficient of a set of integer sequences For the first The investment cost corresponding to the set of integers. In order to be with the first The first group of integer columns matched The combination coefficient of a group of consecutive columns For the first The operation and maintenance costs corresponding to a group of consecutive columns, For the first The replacement cost corresponding to the set of integer columns, For the first The power grid interaction cost corresponding to a group of consecutive columns;

[0039] S23. Solve the main problem and obtain the current objective value of the main problem. And extract the dual variables of the constraints of the main problem. ,in The number of constraints;

[0040] S24. Based on the dual variables of the main problem constraints, construct sub-problems with the objective of maximizing column contribution values:

[0041] ;

[0042] In the above formula, For the total lifecycle cost of the sub-problem, For the new integer column, For the new continuous column, The constraints of the main problem are in vector form. The inner product of the dual variable and the constraint;

[0043] S25. Solve the subproblems to obtain the objective function values ​​corresponding to the new integer column and the consecutive columns. and judge Whether it is satisfied, among which, A contribution threshold is set for each column; if this threshold is met, the new integer column and the consecutive columns are added to the column set. If the conditions are not met, then column generation will stop.

[0044] S26. Determine the column set at this time. If the number of integer columns exceeds the maximum threshold for the number of integer columns and the number of consecutive columns exceeds the maximum threshold for the number of consecutive columns, and if not, return to step S24 to generate new columns and add them to the column set. If the requirements are met, use the following formula to calculate the cumulative contribution value of each column, perform dynamic column filtering, and obtain a high-quality column set:

[0045] ;

[0046] In the above formula, For the first The cumulative contribution value of the column, The initial objective value for the main problem. To join the The target value of the main problem following the column. This is a set of filtered high-quality integer columns. Filter thresholds for contribution values. This is the set of high-quality continuous columns after filtering.

[0047] S27. After filtering, update the column combination range of the main problem to the set of high-quality columns. Subsequent iterations will only be based on... , Generate new columns and solve the problem, then revise the subproblem as follows:

[0048] ;

[0049] In the above formula, As a representative scene, This refers to the total lifecycle cost in a single scenario.

[0050] S28. Solve for the objective value of the corrected subproblem. If satisfied ,but , The newly generated column is valid. Add the new column to the main problem and solve the main problem again to obtain the new objective value. If the difference between the objective values ​​of the main problem in two consecutive iterations satisfies or the number of times the column is generated reaches If the iteration terminates, otherwise return to S24 to continue generating new columns.

[0051] The method further includes:

[0052] The first and second stage constraints of constructing a multi-energy storage collaborative optimization configuration model for microgrids are defined. The first stage is the capacity configuration optimization stage, and the second stage is the real-time operation optimization stage.

[0053] The constraints in the first stage include energy storage capacity constraints and capacity allocation constraints;

[0054] The energy storage capacity constraint is:

[0055] ;

[0056] ;

[0057] ;

[0058] ;

[0059] In the above formula, This is the rated power of electrochemical energy storage. This represents the upper limit of electrochemical energy storage power. For the rated energy capacity of electrochemical energy storage, This represents the upper limit of electrochemical energy storage capacity. This is the rated power for hydrogen energy storage. This represents the upper limit of hydrogen energy storage capacity. For the rated energy capacity of hydrogen energy storage, This represents the upper limit of hydrogen energy storage capacity.

[0060] The capacity allocation constraint is as follows:

[0061] ;

[0062] ;

[0063] In the above formula, , These are the lower and upper limits of the electrochemical energy storage charge / discharge duration, respectively. , These are the lower and upper limits of hydrogen energy storage charge / discharge duration, respectively.

[0064] The second stage constraints include energy storage charge and discharge power constraints, energy storage state of charge constraints, and demand response constraints.

[0065] The energy storage charging and discharging power constraint is:

[0066] ;

[0067] ;

[0068] ;

[0069] ;

[0070] In the above formula, , The first sky The charging and discharging power of electrochemical energy storage at all times. This is the rated power of electrochemical energy storage. , The first sky The charging and discharging power of the hydrogen energy storage system at all times. This is the rated power for hydrogen energy storage;

[0071] The energy storage state of charge constraint is:

[0072] ;

[0073] ;

[0074] In the above formula, For the first sky The state of charge of electrochemical energy storage at any given time. , These represent the lower and upper limits of charge for electrochemical energy storage, respectively. , , These are the lower and upper limits of charge for hydrogen energy storage, respectively.

[0075] The demand response constraint is:

[0076] ;

[0077] ;

[0078] In the above formula, For the first sky Real-time demand response load reduction power To reduce power to the maximum load in demand response, For the first sky real-time demand response load transfer power This represents the maximum load transfer power in demand response.

[0079] Secondly, this invention proposes a multi-time-scale microgrid multi-element energy storage collaborative optimization system, including a collaborative optimization configuration model construction module and a collaborative optimization configuration model solving module;

[0080] The collaborative optimization configuration model construction module is used to construct a multi-timescale microgrid multi-energy storage collaborative optimization configuration model with the lowest life cycle cost as the core optimization objective, taking into account hydrogen energy storage system, electrochemical energy storage system, and the flexible adjustment capability of demand response.

[0081] The collaborative optimization configuration model solving module is used to solve the collaborative optimization configuration model of microgrid multi-element energy storage using an improved column generation algorithm, so as to obtain the optimal capacity configuration scheme and real-time operation optimization strategy of microgrid. The improved column generation algorithm realizes two iterations of the algorithm through dynamic column filtering.

[0082] The collaborative optimization configuration model construction module includes an objective function construction unit and a constraint condition construction unit;

[0083] The objective function construction unit is used to construct the objective function of the following microgrid multi-element energy storage collaborative optimization configuration model:

[0084] ;

[0085] ;

[0086] ;

[0087] ;

[0088] ;

[0089] In the above formula, For total lifecycle cost, For investment costs, For operation and maintenance costs, For replacement costs, For grid interaction costs, For electrochemical energy storage unit power, This is the rated power of electrochemical energy storage. The energy investment cost for electrochemical energy storage, For the rated energy capacity of electrochemical energy storage, For hydrogen energy storage unit power, This is the rated power for hydrogen energy storage. The energy investment cost for hydrogen energy storage, For the rated energy capacity of hydrogen energy storage, For the total lifespan of the energy storage system, The discount rate for funds. The annual operating and maintenance cost per unit energy of electrochemical energy storage. , The annual operating and maintenance cost per unit energy of hydrogen energy storage. This refers to the total number of replacements for the energy storage system throughout its entire lifecycle. For a single replacement cycle, The total initial investment cost for electrochemical energy storage, The total initial investment cost for hydrogen energy storage. , The first The year's first sky The power that a microgrid purchases from the main grid and sells back to the main grid at any given time. , These refer to the electricity purchase price and the electricity sales price from the power grid.

[0090] The constraint construction unit is used to construct the constraint conditions of the microgrid multi-energy storage collaborative optimization configuration model, including hourly operation power balance constraints, daily rolling annual energy balance constraints, and source-load uncertainty quantification constraints.

[0091] The power balance constraint for hourly operation is:

[0092] ;

[0093] In the above formula, for The power exchange of the power grid at any given time. , They are respectively The constant charging and discharging power of electrochemical energy storage , They are respectively The charging and discharging power of hydrogen energy storage at all times. , They are respectively The load reduction power and load transfer power required for real-time demand response. for Total load power of the microgrid at any given time , They are respectively Real-time photovoltaic and wind power output;

[0094] The annual energy balance constraint for daily rolling is:

[0095] ;

[0096] In the above formula, For the number of days in a year, , The first sky The power input and output of the microgrid to the mains at any given time;

[0097] The source load uncertainty quantification constraint is:

[0098] ;

[0099] ;

[0100] ;

[0101] ;

[0102] In the above formula, This is the quantification range for source-load uncertainty power, which includes photovoltaic, wind power, and load. The source load uncertainty fluctuation coefficient, For the first sky Predicted power for time-source load uncertainty , , The first sky Always consider the actual power range of photovoltaic, wind power, and load after taking into account uncertainties. These are the uncertainty fluctuation coefficients for photovoltaic, wind power, and load, respectively. , , The first sky Predicted power output of photovoltaic, wind power, and load at all times.

[0103] The collaborative optimization configuration model solution module includes an initial set formation unit, a main problem construction unit, a main problem solving unit, a sub-problem construction unit, a sub-problem solving unit, a dynamic column filtering unit, a column combination range update unit, and an iterative discrimination unit.

[0104] The initial set forming unit is used to randomly generate initial columns within the feasible domain of the variables in the microgrid multi-energy storage collaborative optimization configuration model based on representative scenarios of source-load uncertainty, thus forming an initial set. ,in The integer capacity column corresponds to the first-stage capacity configuration and is generated. Integer capacity vectors that satisfy capacity and allocation constraints ; For continuous operation columns, corresponding to the second stage of operation optimization, each group of integer columns is matched. A group of continuous running vectors that satisfy the running constraints ;

[0105] The main problem construction unit is used to construct the main problem based on an integer capacity vector and a continuously running vector, introducing combination coefficients:

[0106] ;

[0107] ;

[0108] ;

[0109] ;

[0110] ;

[0111] In the above formula, The total lifecycle cost of the main problem For the first The combination coefficient of a set of integer sequences For the first The investment cost corresponding to the set of integers. In order to be with the first The first group of integer columns matched The combination coefficient of a group of consecutive columns For the first The operation and maintenance costs corresponding to a group of consecutive columns, For the first The replacement cost corresponding to the set of integer columns, For the first The power grid interaction cost corresponding to a group of consecutive columns;

[0112] The main problem solving unit is used to solve the main problem and obtain the current objective value of the main problem. And extract the dual variables of the constraints of the main problem. ,in The number of constraints;

[0113] The subproblem construction unit is used to construct subproblems based on the dual variables of the main problem constraints, with the objective of maximizing column contribution values:

[0114] ;

[0115] In the above formula, For the total lifecycle cost of the sub-problem, For the new integer column, For the new continuous column, The constraints of the main problem are in vector form. The inner product of the dual variable and the constraint;

[0116] The subproblem-solving unit is used to solve subproblems and obtain the objective function values ​​corresponding to the new integer column and the continuous column. and judge Whether it is satisfied, among which, A contribution threshold is set for each column; if this threshold is met, the new integer column and the consecutive columns are added to the column set. If the conditions are not met, then column generation will stop.

[0117] The dynamic column filtering unit is used to determine the column set at this time. If the number of integer columns exceeds the maximum threshold for the number of integer columns and the number of consecutive columns exceeds the maximum threshold for the number of consecutive columns, and if not, return to the sub-problem building unit to continue generating new columns and adding them to the column set. If the conditions are met, use the following formula to calculate the cumulative contribution value of each column, perform dynamic column filtering, and obtain a high-quality column set:

[0118] ;

[0119] In the above formula, For the first The cumulative contribution value of the column, The initial objective value for the main problem. To join the The target value of the main problem following the column. This is a set of filtered high-quality integer columns. Filter thresholds for contribution values. This is the set of high-quality continuous columns after filtering.

[0120] The column combination range update unit is used to update the column combination range of the main problem to a high-quality column set. Subsequent iterations are based solely on... , Generate new columns and solve the problem, then modify the objective function of the subproblem as follows:

[0121] ;

[0122] In the above formula, As a representative scene, This refers to the total lifecycle cost in a single scenario.

[0123] The iterative discrimination unit is used to solve the corrected subproblem objective value. If satisfied ,but , The newly generated column is valid. Add the new column to the main problem and solve the main problem again to obtain the new objective value. If the difference between the objective values ​​of the main problem in two consecutive iterations satisfies or the number of times the column is generated reaches If the iteration terminates, otherwise return to the subproblem building unit to continue generating new columns.

[0124] The system also includes a phase constraint construction module;

[0125] The stage constraint construction module is used to construct the first and second stage constraints of the microgrid multi-energy storage collaborative optimization configuration model. The first stage is the capacity configuration optimization stage, and the second stage is the real-time operation optimization stage.

[0126] The constraints in the first stage include energy storage capacity constraints and capacity allocation constraints;

[0127] The energy storage capacity constraint is:

[0128] ;

[0129] ;

[0130] ;

[0131] ;

[0132] In the above formula, This is the rated power of electrochemical energy storage. This represents the upper limit of electrochemical energy storage power. For the rated energy capacity of electrochemical energy storage, This represents the upper limit of electrochemical energy storage capacity. This is the rated power for hydrogen energy storage. This represents the upper limit of hydrogen energy storage capacity. For the rated energy capacity of hydrogen energy storage, This represents the upper limit of hydrogen energy storage capacity.

[0133] The capacity allocation constraint is as follows:

[0134] ;

[0135] ;

[0136] In the above formula, , These are the lower and upper limits of the electrochemical energy storage charge / discharge duration, respectively. , These are the lower and upper limits of hydrogen energy storage charge / discharge duration, respectively.

[0137] The second stage constraints include energy storage charge and discharge power constraints, energy storage state of charge constraints, and demand response constraints.

[0138] The energy storage charging and discharging power constraint is:

[0139] ;

[0140] ;

[0141] ;

[0142] ;

[0143] In the above formula, , The first sky The charging and discharging power of electrochemical energy storage at all times. This is the rated power of electrochemical energy storage. , The first sky The charging and discharging power of the hydrogen energy storage system at all times. This is the rated power for hydrogen energy storage;

[0144] The energy storage state of charge constraint is:

[0145] ;

[0146] ;

[0147] In the above formula, For the first sky The state of charge of electrochemical energy storage at any given time. , These represent the lower and upper limits of charge for electrochemical energy storage, respectively. , , These are the lower and upper limits of charge for hydrogen energy storage, respectively.

[0148] The demand response constraint is:

[0149] ;

[0150] ;

[0151] In the above formula, For the first sky Real-time demand response load reduction power To reduce power to the maximum load in demand response, For the first sky real-time demand response load transfer power This represents the maximum load transfer power in demand response.

[0152] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0153] 1. This invention proposes a multi-timescale microgrid multi-energy storage collaborative optimization method and system. This method takes the lowest life-cycle cost as its core optimization objective, considering hydrogen energy storage systems, electrochemical energy storage systems, and the flexible adjustment capability of demand response, and constructs a multi-timescale microgrid multi-energy storage collaborative optimization configuration model. An improved column generation algorithm is used to solve the microgrid multi-energy storage collaborative optimization configuration model, obtaining the optimal capacity configuration scheme and real-time operation optimization strategy for the microgrid. The improved column generation algorithm achieves two iterations through dynamic column filtering. On the one hand, this method considers multi-timescale collaboration, constructing a closed-loop constraint between long-term and short-term energy storage states, effectively connecting long-term configuration with short-term operation, enabling energy storage to adapt to both seasonal and intraday fluctuations. On the other hand, this method integrates the full life-cycle economy of the microgrid, fully explores the functional coupling of multi-energy storage, clarifies the intrinsic relationship between source-load fluctuation characteristics and energy storage configuration parameters, and balances the effectiveness and economy of the configuration scheme, making it more suitable for the complex source-load characteristics and high-efficiency operation requirements of the microgrid.

[0154] 2. This invention proposes a multi-timescale microgrid multi-element energy storage collaborative optimization method and system. The method adopts an improved column generation algorithm to solve the microgrid multi-element energy storage collaborative optimization configuration model. In view of the high-dimensional calculation and integer variable problems in the optimization process, the improved column generation algorithm realizes two iterations of the algorithm through the improved core strategy of dynamic column screening, effectively decomposes the two-stage mixed integer programming problem, avoids computational redundancy, and improves the solution efficiency and robustness.

[0155] 3. This invention proposes a multi-timescale microgrid multi-energy storage collaborative optimization method and system. Since energy storage configuration is a long-term decision, and the operating status needs to be dynamically adjusted according to the real-time scenario, this method constructs the first and second stage constraints of the microgrid multi-energy storage collaborative optimization configuration model. The first stage is the capacity configuration optimization stage, and the second stage is the real-time operation optimization stage. By adopting a two-stage optimization approach, the adjustment potential of multi-energy storage and demand response is further integrated, taking into account both configuration rationality and operating economy. Attached Figure Description

[0156] Figure 1 This is an overall flowchart of the method described in this invention.

[0157] Figure 2 This is a flowchart of the solution process for the improved column algorithm described in Example 1.

[0158] Figure 3 This is a structural diagram of the system described in this invention. Detailed Implementation

[0159] The present invention will now be described in further detail with reference to specific embodiments and accompanying drawings.

[0160] This invention proposes a multi-timescale microgrid multi-energy storage collaborative optimization method and system. Considering the decoupling characteristics of power and energy capacity in hydrogen energy storage systems, the short-time response advantages of electrochemical energy storage systems, and the flexible adjustment capability of demand response, a microgrid multi-energy storage optimization planning model is constructed. Addressing the multi-timescale power and energy balance needs of microgrids, and combining the uncertainty characteristics of source and load, an hourly power balance model and a daily rolling annual energy balance model are established. A two-stage optimization approach is adopted, constructing the first and second stage constraints to form a complete optimization system. To address the challenges of high-dimensional computation and integer variables in the optimization process, an improved column generation algorithm is used to effectively decompose and solve the two-stage mixed integer programming problem, improving solution efficiency and accuracy. Finally, the optimization results are verified, and the optimal power and energy capacity configuration scheme of the microgrid multi-energy storage system is output. Through multi-timescale modeling and two-stage collaborative optimization, the intrinsic correlation between multi-energy storage configuration and microgrid energy balance is quantified, significantly reducing system operating costs and balancing the effectiveness and economy of the configuration scheme, providing technical support for the stable and efficient operation of microgrids.

[0161] Example 1:

[0162] like Figure 1 As shown, the multi-time-scale microgrid multi-element energy storage collaborative optimization method is carried out in the following steps:

[0163] 1. With the lowest life-cycle cost as the core optimization objective, and considering the hydrogen energy storage system, electrochemical energy storage system, and the flexible adjustment capability of demand response, a multi-time-scale microgrid multi-energy storage collaborative optimization configuration model is constructed.

[0164] During microgrid operation, fluctuations in output from distributed power sources such as photovoltaics and wind power, along with random changes in load consumption, further exacerbate the risk of supply-demand imbalance across multiple time scales. A single time scale model alone cannot fully cover the balance demand. Therefore, a systematic model framework needs to be constructed, centered on the balance objective across multiple time scales, integrating multi-dimensional basic data and quantifying the impact of uncertainties.

[0165] Considering the decoupling characteristics of power and energy capacity of hydrogen energy storage system (HES), the short-time response advantage of electrochemical energy storage system (EES), and the flexible adjustment capability of demand response (DR), a multi-timescale microgrid multi-element energy storage collaborative optimization configuration model is constructed.

[0166] With the lowest total lifecycle cost as the core optimization objective, covering the costs of investment, operation and maintenance, replacement, and grid interaction, the objective function of the microgrid multi-energy storage collaborative optimization configuration model is:

[0167] ;

[0168] Investment cost is the one-time capital investment in the initial construction of a microgrid multi-energy storage system. It mainly covers the power capacity and energy capacity purchase expenditure of electrochemical energy storage (EES) and hydrogen energy storage (HES), and directly determines the initial construction scale and technology configuration level of the energy storage system.

[0169] ;

[0170] Operation and maintenance costs are the ongoing expenditures incurred throughout the entire lifecycle of an energy storage system to ensure the normal operation of the equipment. These costs include routine equipment inspections, condition monitoring, consumable replacement, and energy losses. They need to be discounted to present value using a discount rate to reflect the long-term operational burden.

[0171] ;

[0172] Replacement cost is the expenditure incurred to replace equipment after the energy storage system reaches its design lifespan in order to maintain the continuity of microgrid energy storage services. It needs to consider the number of replacements and the replacement cycle throughout the entire life cycle, and is also discounted to present value using a discount rate:

[0173] ;

[0174] ;

[0175] ;

[0176] Grid interaction costs are the net revenue and expenditure incurred during power exchange between a microgrid and the main grid. These costs include expenditures on purchasing electricity from the main grid and revenues from selling electricity to the main grid. They are directly related to the amount of power interaction across multiple time scales and the grid purchase and sale prices of electricity.

[0177] ;

[0178] In the above formula, For total lifecycle cost, For investment costs, For operation and maintenance costs, For replacement costs, For grid interaction costs, For electrochemical energy storage unit power, This is the rated power of electrochemical energy storage. The energy investment cost for electrochemical energy storage, For the rated energy capacity of electrochemical energy storage, For hydrogen energy storage unit power, This is the rated power for hydrogen energy storage. The energy investment cost for hydrogen energy storage, For the rated energy capacity of hydrogen energy storage, For the total lifespan of the energy storage system, The discount rate for funds. The annual operating and maintenance cost per unit energy of electrochemical energy storage. , The annual operating and maintenance cost per unit energy of hydrogen energy storage. This refers to the total number of replacements for the energy storage system throughout its entire lifecycle. For a single replacement cycle, The total initial investment cost for electrochemical energy storage, The total initial investment cost for hydrogen energy storage. , The first The year's first sky The power that a microgrid purchases from the main grid and sells back to the main grid at any given time. , These refer to the electricity purchase price and the electricity sales price from the power grid.

[0179] Using distributed power generation output forecast data, load consumption data, energy storage equipment technical parameters, and grid access constraints as core basic data, key constraints such as power balance constraints, energy conservation constraints, and source-load uncertainty confidence interval constraints are clarified to provide data support and boundary constraints for model construction.

[0180] To address the multi-timescale power and energy balance requirements of microgrids, and considering the uncertainty characteristics of source and load, constraints are constructed for a collaborative optimization configuration model of multi-element energy storage in microgrids. These constraints include hourly power balance constraints, daily rolling annual energy balance constraints, and source-load uncertainty quantification constraints. These three constraints are fundamental operating rules that microgrids must meet at any time scale, and are basic constraints jointly observed by the first and second stages. Simultaneously, the first and second stages also have their own dedicated constraints, used to achieve specific optimization objectives for their respective stages.

[0181] To address the rapid power fluctuations and the need for real-time supply-demand matching in microgrids within an hour, a power balance model is constructed with a 1-hour time step to ensure instantaneous power supply-demand balance. The power balance constraints for hourly operation are as follows:

[0182] ;

[0183] In the above formula, for The power exchange of the power grid at any given time is represented by positive values ​​for input power and negative values ​​for output power. , They are respectively The constant charging and discharging power of electrochemical energy storage , They are respectively The charging and discharging power of hydrogen energy storage at all times. , They are respectively The load reduction power and load transfer power required for real-time demand response. for Total load power of the microgrid at any given time , They are respectively Real-time photovoltaic and wind power output;

[0184] To ensure energy conservation during long-term operation and avoid energy deficits or waste on an annual scale, the daily rolling annual energy balance constraint, based on a one-day rolling cycle and combined with the total annual energy consumption demand, is as follows:

[0185] ;

[0186] In the above formula, For the number of days in a year, , The first sky The power input and output of the microgrid to the mains at any given time;

[0187] The deviation in source load prediction directly affects the reliability of the model. Therefore, Monte Carlo simulation is used to generate multiple sets of source load uncertainty scenarios. Combined with confidence interval analysis, the boundary range of source load power fluctuation is determined, providing a basis for robust optimization of the model. The source load uncertainty quantification constraint is as follows:

[0188] ;

[0189] ;

[0190] ;

[0191] ;

[0192] In the above formula, This is the quantification range for source-load uncertainty power, which includes photovoltaic, wind power, and load. The source load uncertainty fluctuation coefficient, For the first sky Predicted power for time-source load uncertainty , , The first sky Always consider the actual power range of photovoltaic, wind power, and load after taking into account uncertainties. These are the uncertainty fluctuation coefficients for photovoltaic, wind power, and load, respectively. , , The first sky Predicted power output of photovoltaic, wind power, and load at all times.

[0193] 2. Construct the first and second stage constraints of the microgrid multi-energy storage collaborative optimization configuration model. The first stage is the capacity configuration optimization stage, and the second stage is the real-time operation optimization stage.

[0194] After clarifying the boundaries of balancing demand and source-load uncertainty across multiple time scales, it is necessary to further integrate the adjustment potential of diversified energy storage and demand response to construct an optimization framework that balances allocation rationality and operational economy. Since energy storage allocation is a long-term decision, while operational status needs to be dynamically adjusted according to real-time scenarios, a two-stage optimization approach is adopted. The first stage focuses on capacity allocation optimization, and the second stage focuses on real-time operation optimization, forming a complete optimization system.

[0195] The first phase of constraints includes energy storage capacity constraints and capacity allocation constraints.

[0196] To ensure the technical feasibility of the configuration scheme, it is necessary to clarify the boundary limits and ratio requirements of the energy storage capacity. The energy storage capacity constraints are as follows:

[0197] ;

[0198] ;

[0199] ;

[0200] ;

[0201] In the above formula, This is the rated power of electrochemical energy storage. This represents the upper limit of electrochemical energy storage power. For the rated energy capacity of electrochemical energy storage, This represents the upper limit of electrochemical energy storage capacity. This is the rated power for hydrogen energy storage. This represents the upper limit of hydrogen energy storage capacity. For the rated energy capacity of hydrogen energy storage, This represents the upper limit of hydrogen energy storage capacity.

[0202] In addition, capacity allocation constraints must be met, which are as follows:

[0203] ;

[0204] ;

[0205] In the above formula, , These are the lower and upper limits of the electrochemical energy storage charge / discharge duration, respectively. , These are the lower and upper limits of hydrogen energy storage charge / discharge duration, respectively.

[0206] Based on the capacity configuration determined in the first phase, the technical boundaries during operation are clarified to ensure safe and stable operation. The constraints in the second phase include energy storage charging and discharging power constraints, energy storage state of charge constraints, and demand response constraints.

[0207] To avoid overload damage to energy storage devices, the charging and discharging power must be limited to the rated range. The energy storage charging and discharging power constraint is as follows:

[0208] ;

[0209] ;

[0210] ;

[0211] ;

[0212] In the above formula, , The first sky The charging and discharging power of electrochemical energy storage at all times. This is the rated power of electrochemical energy storage. , The first sky The charging and discharging power of the hydrogen energy storage system at all times. This is the rated power for hydrogen energy storage;

[0213] A safe operating range for the energy storage state of charge (SOC) is defined to prevent overcharging and over-discharging of the energy storage system from affecting the equipment's lifespan. The energy storage state of charge constraint is as follows:

[0214] ;

[0215] ;

[0216] In the above formula, For the first sky The state of charge of electrochemical energy storage at any given time. , These represent the lower and upper limits of charge for electrochemical energy storage, respectively. , , These are the lower and upper limits of charge for hydrogen energy storage, respectively.

[0217] To balance microgrid regulation and user electricity rights, the maximum regulation amount of demand response is limited. The demand response constraint is as follows:

[0218] ;

[0219] ;

[0220] In the above formula, For the first sky Real-time demand response load reduction power To reduce power to the maximum load in demand response, For the first sky real-time demand response load transfer power This represents the maximum load transfer power in demand response.

[0221] 3. An improved column generation algorithm is adopted to solve the microgrid multi-energy storage collaborative optimization configuration model, so as to obtain the optimal capacity configuration scheme and real-time operation optimization strategy of the microgrid. The improved column generation algorithm realizes two iterations of the algorithm through dynamic column filtering.

[0222] To address the challenges of high-dimensional computation and integer variables in the optimization process, an improved column generation algorithm is adopted to effectively decompose and solve the two-stage mixed integer programming problem, thereby improving the solution efficiency and accuracy.

[0223] Because the microgrid multi-element energy storage collaborative optimization configuration model contains a large number of high-dimensional continuous variables and integer decision variables, traditional solution algorithms are prone to slow convergence and getting trapped in local optima, making it difficult to meet the efficiency and accuracy requirements of engineering applications. Therefore, this invention proposes an improved column generation algorithm, which adapts to the two-stage decision logic and uncertainty characteristics through intelligent initial column generation, adaptive column selection, and hybrid variable collaborative optimization, achieving efficient acquisition of the global optimal solution. The specific implementation steps are as follows (see...). Figure 2 ):

[0224] First, an intelligent initial column is generated. The quality of the initial column directly determines the convergence speed of the algorithm. Traditional column generation relies on manually setting the initial column, which can easily lead to incomplete coverage of the feasible region. This invention combines the characteristics of source load uncertainty scenarios and generates initial columns through scenario clustering and random sampling to ensure coverage of the core feasible region.

[0225] Generated by Monte Carlo simulation In scenarios with uncertain source and load conditions, and with the objective of minimizing the sum of squares within clusters, the K-means clustering algorithm is used to reduce the dimensionality of the scenario, preserving... Representative scenarios:

[0226] ;

[0227] In the above formula, For the sum of squares within the cluster, For scene feature vectors, For the first A cluster, As the cluster center, this formula ensures that representative scenarios cover the core characteristics of source load fluctuations.

[0228] For each representative scenario, initial columns are randomly generated within the feasible region of the variables in the microgrid multi-energy storage collaborative optimization configuration model to form an initial set. ;

[0229] in, The integer capacity column corresponds to the first-stage capacity configuration and is generated. Integer capacity vectors that satisfy capacity and allocation constraints ;

[0230] For continuous column execution, corresponding to the second stage of optimization, ensure the matching of integer variables and continuous variables in the initial column set, with each set of integer columns matching. A group of continuous running vectors that satisfy the running constraints .

[0231] Second, based on the initial set of columns, a main problem is constructed, the core of which is to find an approximate optimal solution through column combination optimization, while providing dual information for the subproblems:

[0232] The main problem aims to minimize the total lifecycle cost, and introduces a combination coefficient to link an integer capacity vector with a continuously running vector:

[0233] ;

[0234] ;

[0235] ;

[0236] ;

[0237] ;

[0238] In the above formula, The total lifecycle cost of the main problem For the first The combination coefficient of a set of integer sequences For the first The investment cost corresponding to the set of integers. In order to be with the first The first group of integer columns matched The combination coefficient of a group of consecutive columns For the first The operation and maintenance costs corresponding to a group of consecutive columns, For the first The replacement cost corresponding to the set of integer columns, For the first The power grid interaction cost corresponding to a group of consecutive columns;

[0239] Solve the main problem to obtain the current objective value of the main problem. And extract the dual variables of the constraints of the main problem. ,in The number of constraints is represented by the dual variable, which reflects the weight of each constraint on the objective function.

[0240] Third, construct sub-problems. The core of sub-problems is to find feasible solutions that can significantly reduce the objective value of the main problem based on the dual variables of the main problem, thus avoiding computational redundancy caused by blindly generating columns.

[0241] Based on the dual variables of the main problem constraints, and with the objective of maximizing the column contribution value (i.e., finding a new column that reduces the main problem's objective value the most), a subproblem is constructed:

[0242] ;

[0243] In the above formula, For the total lifecycle cost of the sub-problem, For the new integer column, For the new continuous column, The constraints of the main problem are in vector form. The inner product of the dual variable and the constraints is used to measure the degree to which the new column satisfies the constraints of the main problem.

[0244] The feasibility of adding a new column is assessed, and the interior-point method is used to solve the subproblems, yielding the objective function values ​​for the new integer column and the consecutive columns. and judge Whether it is satisfied, among which, Column contribution threshold (determines whether a new column can be added to the column set). If this condition is met, it means that the new column can significantly optimize the solution to the main problem. Therefore, add the new integer column and the continuous column to the column set. (Add integer column) Add consecutive columns If the condition is not met, column generation will stop.

[0245] Among them, the new integer sequence in the initial iteration and new consecutive columns These correspond to the energy storage capacity configuration optimization variables in the first stage and the energy storage operation optimization variables in the second stage, respectively. The third step is the detailed generation process of the new column at this point. Based on the current dual variables of the main problem, it finds the feasible combination of variables that will reduce the objective value of the main problem the most by solving the subproblems. Each time such a set of variables is found, it is called "generating a new column" and is added to the column set of the main problem for optimization in the next iteration.

[0246] Fourth, to further address the computational redundancy and mixed variable adaptation issues in traditional column generation, the following improvement strategy is designed based on the algorithm's characteristics:

[0247] First, determine the column set at this point. Does the number of integer columns exceed the maximum number of integer columns threshold? And whether the number of consecutive columns exceeds the maximum threshold for the number of consecutive columns. If the conditions are not met, return to step three to generate new columns and add them to the column set. If the conditions are met, calculate the cumulative contribution value of each column using the following formula, perform dynamic column filtering, obtain a high-quality column set, retain high-quality columns whose contribution values ​​meet the threshold requirements, remove redundant columns, and update the column set:

[0248] ;

[0249] In the above formula, For the first The cumulative contribution value of the column, the first The column is only used in the dynamic column filtering process and includes integer columns and consecutive columns. The initial objective value of the main problem (in this case, and The values ​​are not exactly the same. Initial main problem objective value This refers to the objective value obtained by the algorithm in its first solution to the main problem; the current objective value of the main problem. It is the real-time target value obtained by iterative algorithm solution, and the two are different. To join the The target value of the main problem following the column. This is a set of filtered high-quality integer columns. A threshold for contribution values ​​is used to filter high-quality column sets, ensuring that the number of column sets is kept within a reasonable range. This is the set of high-quality continuous columns after filtering.

[0250] After filtering, the range of column combinations for the main question is updated to a set of high-quality columns, resulting in the new main question:

[0251] ;

[0252] The constraints in the formula remain unchanged, except that the column index range is limited to the filtered set of high-quality columns. Subsequent iterations are based solely on... , Generate new columns and solve the problem to avoid redundant calculations;

[0253] Based on high-quality column sets , The steps to generate a new column include:

[0254] locking Medium cost optimal column ;

[0255] by Centered on a neighborhood step size within the feasible integer domain, a new integer column is generated:

[0256] ;

[0257] In the above formula, For integer domain step size, A coefficient is selected for the step size to control the generation of new columns in the upper and lower neighborhoods of the optimal column;

[0258] right Match each integer column Optimal continuous column of cost for single-scenario applications:

[0259] ;

[0260] by Centered on the continuous feasible region, set the continuous neighborhood step size to generate a new continuous column:

[0261] ;

[0262] In the above formula, This is the relative neighborhood step size, with a value of 0.05.

[0263] Then, in solving the subproblem, the scenario probability determined by the proportion of cluster sizes is introduced, and the subproblem is modified to minimize the cost of a single scenario and maximize the column contribution value, with the objective of minimizing the cost of a single scenario and maximizing the column contribution value:

[0264] ;

[0265] In the above formula, As a representative scene, The total lifecycle cost under a single scenario; the robustness of the solution is improved by prioritizing the generation of the optimal column for high-probability scenarios through scenario weighting.

[0266] The validity of the new target value is first determined by the column contribution threshold. If valid, the column set is updated. The updated column set is then substituted into the main problem for solution, thereby driving the iterative optimization of the main problem.

[0267] By introducing a sub-problem to correct the source-load uncertainty representative scenario, the original global lifecycle cost optimization is transformed into cost optimization under a single scenario. This makes the generated new columns more consistent with the actual operating scenario of microgrid source-load fluctuations and reduces optimization deviation. Furthermore, the column generation range is expanded from the initial global limitation to a set of high-quality columns after dynamic screening, avoiding redundant calculations of invalid columns.

[0268] Finally, to address the matching requirements between integer capacity variables and continuously running variables, variable matching constraints are added when generating new columns: , The matching coefficients ensure that the new column satisfies the association constraints of the two-stage variables.

[0269] Fifth, convergence check and correction of integer variables:

[0270] Solve the corrected subproblem objective value If satisfied ,but , The generated new column is valid and will , After the newly generated column is added to the main problem, the main problem is solved again to obtain a new objective value. If the difference between the objective values ​​of the main problem in two consecutive iterations satisfies , or the number of times the column is generated reaches If the iteration terminates, otherwise return to step 3 to continue generating new columns.

[0271] Since integer variables in the main problem can take non-integer values, the combination coefficients of the integer series need to be calculated after the iteration terminates. Perform integerization correction: for Sort in descending order (usually only a few integer columns) (It will be significantly greater than 0, and the rest are close to 0), if it exists If the integer column is a given value, it is directly selected as the optimal capacity configuration for output, and the correction process ends; if there is no such column, it is directly selected as the optimal capacity configuration for output. When selecting an integer column, choose the first... A candidate set consists of a sequence of integers. ( Generally, 3-5 are chosen, and the integer column that contributes the most to the objective function is retained (if...). (Since its marginal contribution to the objective function is small, it is not included as a candidate integer column), the candidate set... Integer columns in Perform a weighted average and round the newly obtained capacity configuration to the nearest integer as the optimal capacity configuration output.

[0272] 4. Verify the optimization results and output the optimal power and energy capacity configuration scheme for the microgrid multi-element energy storage system;

[0273] The capacity configuration scheme obtained by the improved column generation algorithm needs to undergo comprehensive verification to ensure its feasibility and superiority in practical applications, avoiding configuration failure due to differences between model assumptions and actual scenarios. Therefore, verification analysis is carried out from multiple dimensions to ultimately output a reliable optimal capacity configuration scheme. The specific process is as follows:

[0274] First, the feasibility of the optimization results is verified. The rationality of the configuration scheme during technical implementation is verified from the following three perspectives:

[0275] Power balance verification: Under typical scenarios, simulate the operation of a microgrid to verify whether the hourly power balance model meets real-time supply and demand matching and whether power fluctuations are within the allowable range;

[0276] Energy balance verification: Statistically analyze the total annual energy inflow and outflow to verify whether the daily rolling annual energy balance model satisfies long-term energy conservation and has no obvious energy deficit or waste;

[0277] Constraint fulfillment verification: Check whether parameters such as energy storage capacity, charging and discharging power, and SOC meet the constraints, and whether the demand response adjustment is within the allowable range to ensure the technical feasibility of the configuration scheme.

[0278] Then, the superiority of the optimization results is compared with existing methods, highlighting the advantages of the present invention.

[0279] This invention aims to minimize the total lifecycle cost and uses an improved column generation algorithm to solve a two-stage mixed integer programming problem. The test conditions cover typical source-load uncertainty scenarios.

[0280] Because microgrid sources and loads exhibit multi-timescale fluctuations and uncertainties, this invention quantitatively compares the proposed method with the Benders decomposition algorithm and column generation algorithm from three core dimensions: economy, robustness, and solution efficiency (results are shown in Table 1).

[0281] Economic Comparison: Compare economic indicators such as total life cycle cost and unit energy storage cost;

[0282] Robustness comparison: Under different source load uncertainty fluctuation coefficients, the power deficit rate and energy waste rate of the configuration schemes are compared to verify the robustness of the model in complex scenarios;

[0283] Efficiency Comparison: Compare the solution time and convergence speed of the algorithms to demonstrate the efficiency of the improved algorithm;

[0284] Table 1. Performance Comparison of Multidimensional Indicators

[0285] ;

[0286] As shown in Table 1, in terms of economics, the total lifecycle cost of this invention is reduced to 650 million yuan, which is 24.4% and 9.7% lower than the two comparative schemes, respectively, demonstrating a significant cost control advantage. Regarding robustness, the power deficit rate of this invention is 1.2%, and the energy waste rate is 2.0%, exhibiting stronger scenario adaptability and resistance to fluctuations. In terms of solution efficiency, the solution time of this invention is shortened to 22.0 minutes, which is 48.4% and 38.5% higher than the two comparative schemes, respectively, effectively adapting to the needs of actual engineering applications. In summary, all core indicators of the proposed solution are superior to traditional methods, fully demonstrating the comprehensive advantages of this invention in multi-timescale coordination, multi-element energy storage complementarity, and algorithm optimization.

[0287] Finally, the optimal configuration scheme is output;

[0288] By integrating the verification results, eliminating candidate solutions that do not meet the requirements, and outputting the optimal configuration parameters of the microgrid multi-element energy storage system, a complete configuration solution report is generated, providing direct guidance for engineering applications.

[0289] Example 2:

[0290] like Figure 3 As shown, a multi-time-scale microgrid multi-element energy storage collaborative optimization system includes a collaborative optimization configuration model construction module and a collaborative optimization configuration model solution module.

[0291] The collaborative optimization configuration model construction module is used to construct a multi-timescale microgrid multi-energy storage collaborative optimization configuration model with the lowest life cycle cost as the core optimization objective, taking into account hydrogen energy storage system, electrochemical energy storage system, and the flexible adjustment capability of demand response.

[0292] The collaborative optimization configuration model solving module is used to solve the collaborative optimization configuration model of microgrid multi-element energy storage using an improved column generation algorithm, so as to obtain the optimal capacity configuration scheme and real-time operation optimization strategy of microgrid. The improved column generation algorithm realizes two iterations of the algorithm through dynamic column filtering.

[0293] The collaborative optimization configuration model construction module includes an objective function construction unit and a constraint condition construction unit;

[0294] The objective function construction unit is used to construct the objective function of the following microgrid multi-element energy storage collaborative optimization configuration model:

[0295] ;

[0296] ;

[0297] ;

[0298] ;

[0299] ;

[0300] In the above formula, For total lifecycle cost, For investment costs, For operation and maintenance costs, For replacement costs, For grid interaction costs, For electrochemical energy storage unit power, This is the rated power of electrochemical energy storage. The energy investment cost for electrochemical energy storage, For the rated energy capacity of electrochemical energy storage, For hydrogen energy storage unit power, This is the rated power for hydrogen energy storage. The energy investment cost for hydrogen energy storage, For the rated energy capacity of hydrogen energy storage, For the total lifespan of the energy storage system, The discount rate for funds. The annual operating and maintenance cost per unit energy of electrochemical energy storage. , The annual operating and maintenance cost per unit energy of hydrogen energy storage. This refers to the total number of replacements for the energy storage system throughout its entire lifecycle. For a single replacement cycle, The total initial investment cost for electrochemical energy storage, The total initial investment cost for hydrogen energy storage. , The first The year's first sky The power that a microgrid purchases from the main grid and sells back to the main grid at any given time. , These refer to the electricity purchase price and the electricity sales price from the power grid.

[0301] The constraint construction unit is used to construct the constraint conditions of the microgrid multi-energy storage collaborative optimization configuration model, including hourly operation power balance constraints, daily rolling annual energy balance constraints, and source-load uncertainty quantification constraints.

[0302] The power balance constraint for hourly operation is:

[0303] ;

[0304] In the above formula, for The power exchange of the power grid at any given time. , They are respectively The constant charging and discharging power of electrochemical energy storage , They are respectively The charging and discharging power of hydrogen energy storage at all times. , They are respectively The load reduction power and load transfer power required for real-time demand response. for Total load power of the microgrid at any given time , They are respectively Real-time photovoltaic and wind power output;

[0305] The annual energy balance constraint for daily rolling is:

[0306] ;

[0307] In the above formula, For the number of days in a year, , The first sky The power input and output of the microgrid to the mains at any given time;

[0308] The source load uncertainty quantification constraint is:

[0309] ;

[0310] ;

[0311] ;

[0312] ;

[0313] In the above formula, This is the quantification range for source-load uncertainty power, which includes photovoltaic, wind power, and load. The source load uncertainty fluctuation coefficient, For the first sky Predicted power for time-source load uncertainty , , The first sky Always consider the actual power range of photovoltaic, wind power, and load after taking into account uncertainties. These are the uncertainty fluctuation coefficients for photovoltaic, wind power, and load, respectively. , , The first sky Predicted power output of photovoltaic, wind power, and load at all times.

[0314] The collaborative optimization configuration model solution module includes an initial set formation unit, a main problem construction unit, a main problem solving unit, a sub-problem construction unit, a sub-problem solving unit, a dynamic column filtering unit, a column combination range update unit, and an iterative discrimination unit.

[0315] The initial set forming unit is used to randomly generate initial columns within the feasible domain of the variables in the microgrid multi-energy storage collaborative optimization configuration model based on representative scenarios of source-load uncertainty, thus forming an initial set. ,in The integer capacity column corresponds to the first-stage capacity configuration and is generated. Integer capacity vectors that satisfy capacity and allocation constraints ; For continuous operation columns, corresponding to the second stage of operation optimization, each group of integer columns is matched. A group of continuous running vectors that satisfy the running constraints ;

[0316] The main problem construction unit is used to construct the main problem based on an integer capacity vector and a continuously running vector, introducing combination coefficients:

[0317] ;

[0318] ;

[0319] ;

[0320] ;

[0321] ;

[0322] In the above formula, The total lifecycle cost of the main problem For the first The combination coefficient of a set of integer sequences For the first The investment cost corresponding to the set of integers. In order to be with the first The first group of integer columns matched The combination coefficient of a group of consecutive columns For the first The operation and maintenance costs corresponding to a group of consecutive columns, For the first The replacement cost corresponding to the set of integer columns, For the first The power grid interaction cost corresponding to a group of consecutive columns;

[0323] The main problem solving unit is used to solve the main problem and obtain the current objective value of the main problem. And extract the dual variables of the constraints of the main problem. ,in The number of constraints;

[0324] The subproblem construction unit is used to construct subproblems based on the dual variables of the main problem constraints, with the objective of maximizing column contribution values:

[0325] ;

[0326] In the above formula, For the total lifecycle cost of the sub-problem, For the new integer column, For the new continuous column, The constraints of the main problem are in vector form. The inner product of the dual variable and the constraint;

[0327] The subproblem-solving unit is used to solve subproblems and obtain the objective function values ​​corresponding to the new integer column and the continuous column. and judge Whether it is satisfied, among which, A contribution threshold is set for each column; if this threshold is met, the new integer column and the consecutive columns are added to the column set. If the conditions are not met, then column generation will stop.

[0328] The dynamic column filtering unit is used to determine the column set at this time. If the number of integer columns exceeds the maximum threshold for the number of integer columns and the number of consecutive columns exceeds the maximum threshold for the number of consecutive columns, and if not, return to the sub-problem building unit to continue generating new columns and adding them to the column set. If the conditions are met, use the following formula to calculate the cumulative contribution value of each column, perform dynamic column filtering, and obtain a high-quality column set:

[0329] ;

[0330] In the above formula, For the first The cumulative contribution value of the column, The initial objective value for the main problem. To join the The target value of the main problem following the column. This is a set of filtered high-quality integer columns. Filter thresholds for contribution values. This is the set of high-quality continuous columns after filtering.

[0331] The column combination range update unit is used to update the column combination range of the main problem to a high-quality column set. Subsequent iterations are based solely on... , Generate new columns and solve the problem, then modify the objective function of the subproblem as follows:

[0332] ;

[0333] In the above formula, As a representative scene, This refers to the total lifecycle cost in a single scenario.

[0334] The iterative discrimination unit is used to solve the corrected subproblem objective value. If satisfied ,but , The newly generated column is valid. Add the new column to the main problem and solve the main problem again to obtain the new objective value. If the difference between the objective values ​​of the main problem in two consecutive iterations satisfies or the number of times the column is generated reaches If the iteration terminates, otherwise return to the subproblem building unit to continue generating new columns.

[0335] The system also includes a phase constraint construction module;

[0336] The stage constraint construction module is used to construct the first and second stage constraints of the microgrid multi-energy storage collaborative optimization configuration model. The first stage is the capacity configuration optimization stage, and the second stage is the real-time operation optimization stage.

[0337] The constraints in the first stage include energy storage capacity constraints and capacity allocation constraints;

[0338] The energy storage capacity constraint is:

[0339] ;

[0340] ;

[0341] ;

[0342] ;

[0343] In the above formula, This is the rated power of electrochemical energy storage. This represents the upper limit of electrochemical energy storage power. For the rated energy capacity of electrochemical energy storage, This represents the upper limit of electrochemical energy storage capacity. This is the rated power for hydrogen energy storage. This represents the upper limit of hydrogen energy storage capacity. For the rated energy capacity of hydrogen energy storage, This represents the upper limit of hydrogen energy storage capacity.

[0344] The capacity allocation constraint is as follows:

[0345] ;

[0346] ;

[0347] In the above formula, , These are the lower and upper limits of the electrochemical energy storage charge / discharge duration, respectively. , These are the lower and upper limits of hydrogen energy storage charge / discharge duration, respectively.

[0348] The second stage constraints include energy storage charge and discharge power constraints, energy storage state of charge constraints, and demand response constraints.

[0349] The energy storage charging and discharging power constraint is:

[0350] ;

[0351] ;

[0352] ;

[0353] ;

[0354] In the above formula, , The first sky The charging and discharging power of electrochemical energy storage at all times. This is the rated power of electrochemical energy storage. , The first sky The charging and discharging power of the hydrogen energy storage system at all times. This is the rated power for hydrogen energy storage;

[0355] The energy storage state of charge constraint is:

[0356] ;

[0357] ;

[0358] In the above formula, For the first sky The state of charge of electrochemical energy storage at any given time. , These represent the lower and upper limits of charge for electrochemical energy storage, respectively. , , These are the lower and upper limits of charge for hydrogen energy storage, respectively.

[0359] The demand response constraint is:

[0360] ;

[0361] ;

[0362] In the above formula, For the first sky Real-time demand response load reduction power To reduce power to the maximum load in demand response, For the first sky real-time demand response load transfer power This represents the maximum load transfer power in demand response.

Claims

1. A multi-time-scale microgrid multi-element energy storage collaborative optimization method, characterized in that, The method includes: S1. With the lowest life-cycle cost as the core optimization objective, and considering the hydrogen energy storage system, electrochemical energy storage system, and the flexible adjustment capability of demand response, a multi-time-scale microgrid multi-energy storage collaborative optimization configuration model is constructed. S2. An improved column generation algorithm is used to solve the microgrid multi-energy storage collaborative optimization configuration model, and obtain the optimal capacity configuration scheme and real-time operation optimization strategy of the microgrid. The improved column generation algorithm realizes two iterations of the algorithm through dynamic column filtering.

2. The multi-time-scale microgrid multi-element energy storage synergistic optimization method according to claim 1, characterized in that, In S1, the objective function of the microgrid multi-energy storage collaborative optimization configuration model is: ； ； ； ； ； In the above formula, For total lifecycle cost, For investment costs, For operation and maintenance costs, For replacement costs, For grid interaction costs, For electrochemical energy storage unit power, This is the rated power of electrochemical energy storage. The energy investment cost for electrochemical energy storage, For the rated energy capacity of electrochemical energy storage, For hydrogen energy storage unit power, This is the rated power for hydrogen energy storage. The energy investment cost for hydrogen energy storage, For the rated energy capacity of hydrogen energy storage, For the total lifespan of the energy storage system, The discount rate for funds. The annual operating and maintenance cost per unit energy of electrochemical energy storage. , The annual operating and maintenance cost per unit energy of hydrogen energy storage. This refers to the total number of replacements for the energy storage system throughout its entire lifecycle. For a single replacement cycle, The total initial investment cost for electrochemical energy storage, The total initial investment cost for hydrogen energy storage. , The first The year's first sky The power that a microgrid purchases from the main grid and sells back to the main grid at any given time. , These refer to the electricity purchase price and the electricity sales price from the power grid. The constraints of the microgrid multi-energy storage collaborative optimization configuration model include hourly power balance constraints, daily rolling annual energy balance constraints, and source-load uncertainty quantification constraints. The power balance constraint for hourly operation is: ； In the above formula, for The power exchange of the power grid at any given time. , They are respectively The constant charging and discharging power of electrochemical energy storage , They are respectively The charging and discharging power of hydrogen energy storage at all times. , They are respectively The load reduction power and load transfer power required for real-time demand response. for Total load power of the microgrid at any given time , They are respectively Real-time photovoltaic and wind power output; The annual energy balance constraint for the daily rolling cycle is: ； In the above formula, For the number of days in a year, , The first sky The power input and output of the microgrid to the mains at any given time; The source load uncertainty quantification constraint is: ； ； ； ； In the above formula, This is the quantification range for source-load uncertainty power, which includes photovoltaic, wind power, and load. The source load uncertainty fluctuation coefficient, For the first sky Predicted power for time-source load uncertainty , , The first sky Always consider the actual power range of photovoltaic, wind power, and load after taking into account uncertainties. These are the uncertainty fluctuation coefficients for photovoltaic, wind power, and load, respectively. , , The first sky Predicted power output of photovoltaic, wind power, and load at all times.

3. The multi-time-scale microgrid multi-element energy storage synergistic optimization method according to claim 1, characterized in that, S2 includes: S21. Based on representative scenarios of source-load uncertainty, an initial column is randomly generated within the feasible region of variables in the microgrid multi-electrode energy storage collaborative optimization configuration model to form an initial set. ,in The integer capacity column corresponds to the first-stage capacity configuration and is generated. Integer capacity vectors that satisfy capacity and allocation constraints ; For continuous operation columns, corresponding to the second stage of operation optimization, each group of integer columns is matched. A group of continuous running vectors that satisfy the running constraints ; S22. Based on integer capacity vectors and continuously running vectors, introduce combination coefficients to construct the main problem: ； ； ； ； ； In the above formula, The total lifecycle cost of the main problem For the first The combination coefficient of a set of integer sequences For the first The investment cost corresponding to the set of integers. In order to be with the first The first group of integer columns matched The combination coefficient of a group of consecutive columns For the first The operation and maintenance costs corresponding to a group of consecutive columns, For the first The replacement cost corresponding to the set of integer columns, For the first The power grid interaction cost corresponding to a group of consecutive columns; S23. Solve the main problem and obtain the current objective value of the main problem. And extract the dual variables of the constraints of the main problem. ,in The number of constraints; S24. Based on the dual variables of the main problem constraints, construct sub-problems with the objective of maximizing column contribution values: ； In the above formula, For the total lifecycle cost of the subproblem, For the new integer column, For the new continuous column, The constraints of the main problem are in vector form. The inner product of the dual variable and the constraint; S25. Solve the subproblems to obtain the objective function values ​​corresponding to the new integer column and the consecutive columns. and judge Whether it is satisfied, among which, A contribution threshold is set for each column; if this threshold is met, the new integer column and the consecutive columns are added to the column set. If the conditions are not met, then column generation will stop. S26. Determine the column set at this time. If the number of integer columns exceeds the maximum threshold for the number of integer columns and the number of consecutive columns exceeds the maximum threshold for the number of consecutive columns, then return to step S24 to generate new columns and add them to the column set. If the requirements are met, then use the following formula to calculate the cumulative contribution value of each column, perform dynamic column filtering, and obtain a high-quality column set: ； In the above formula, For the first The cumulative contribution value of the column, The initial objective value for the main problem. To join the The target value of the main problem following the column. This is a set of filtered high-quality integer columns. Filter thresholds for contribution values. This is the set of high-quality continuous columns after filtering. S27. After filtering, update the column combination range of the main problem to the set of high-quality columns. Subsequent iterations will only be based on... , Generate new columns and solve the problem, then revise the subproblem as follows: ； In the above formula, As a representative scene, This refers to the total lifecycle cost in a single scenario. S28. Solve for the objective value of the corrected subproblem. If satisfied ,but , The newly generated column is valid. Add the newly generated column to the main problem and solve the main problem again to obtain the new objective value. If the difference between the objective values ​​of the main problem in two consecutive iterations satisfies or the number of times the column is generated reaches If the iteration terminates, otherwise return to S24 to continue generating new columns.

4. The multi-time-scale microgrid multi-element energy storage synergistic optimization method according to claim 1, characterized in that, The method further includes: The first and second stage constraints of constructing a multi-element energy storage collaborative optimization configuration model for microgrids are defined. The first stage is the capacity configuration optimization stage, and the second stage is the real-time operation optimization stage.

5. The multi-time-scale microgrid multi-element energy storage synergistic optimization method according to claim 4, characterized in that, The constraints in the first stage include energy storage capacity constraints and capacity allocation constraints; The energy storage capacity constraint is: ； ； ； ； In the above formula, This is the rated power of electrochemical energy storage. This represents the upper limit of electrochemical energy storage power. For the rated energy capacity of electrochemical energy storage, This represents the upper limit of electrochemical energy storage capacity. This is the rated power for hydrogen energy storage. This represents the upper limit of hydrogen energy storage capacity. For the rated energy capacity of hydrogen energy storage, This represents the upper limit of hydrogen energy storage capacity; The capacity ratio constraint is: ； ； In the above formula, , These are the lower and upper limits of the electrochemical energy storage charge / discharge duration, respectively. , These are the lower and upper limits of hydrogen energy storage charge / discharge duration, respectively. The second stage constraints include energy storage charge and discharge power constraints, energy storage state of charge constraints, and demand response constraints. The energy storage charging and discharging power constraint is: ； ； ； ； In the above formula, , The first sky The charging and discharging power of electrochemical energy storage at all times. This is the rated power of electrochemical energy storage. , The first sky The charging and discharging power of the hydrogen energy storage system at all times. This is the rated power for hydrogen energy storage; The energy storage state of charge constraint is: ； ； In the above formula, For the first sky The state of charge of electrochemical energy storage at any given time. , These represent the lower and upper limits of charge for electrochemical energy storage, respectively. , , These are the lower and upper limits of charge for hydrogen energy storage, respectively. The demand response constraint is: ； ； In the above formula, For the first sky Real-time demand response load reduction power To reduce power to the maximum load in demand response, For the first sky real-time demand response load transfer power This represents the maximum load transfer power in demand response.

6. A multi-time-scale microgrid multi-element energy storage collaborative optimization system, characterized in that, The system includes a collaborative optimization configuration model construction module and a collaborative optimization configuration model solving module; The collaborative optimization configuration model construction module is used to construct a multi-timescale microgrid multi-energy storage collaborative optimization configuration model with the lowest life cycle cost as the core optimization objective, taking into account hydrogen energy storage system, electrochemical energy storage system, and the flexible adjustment capability of demand response. The collaborative optimization configuration model solving module is used to solve the collaborative optimization configuration model of microgrid multi-element energy storage using an improved column generation algorithm, so as to obtain the optimal capacity configuration scheme and real-time operation optimization strategy of microgrid. The improved column generation algorithm realizes two iterations of the algorithm through dynamic column filtering.

7. The microgrid multi-element energy storage collaborative optimization system based on multiple time scales according to claim 6, characterized in that, The collaborative optimization configuration model construction module includes an objective function construction unit and a constraint condition construction unit; The objective function construction unit is used to construct the objective function of the following microgrid multi-element energy storage collaborative optimization configuration model: ； ； ； ； ； In the above formula, For total lifecycle cost, For investment costs, For operation and maintenance costs, For replacement costs, For grid interaction costs, For electrochemical energy storage unit power, This is the rated power of electrochemical energy storage. The energy investment cost for electrochemical energy storage, For the rated energy capacity of electrochemical energy storage, For hydrogen energy storage unit power, This is the rated power for hydrogen energy storage. The energy investment cost for hydrogen energy storage, For the rated energy capacity of hydrogen energy storage, For the total lifespan of the energy storage system, The discount rate for funds. The annual operating and maintenance cost per unit energy of electrochemical energy storage. , The annual operating and maintenance cost per unit energy of hydrogen energy storage. This refers to the total number of replacements for the energy storage system throughout its entire lifecycle. For a single replacement cycle, The total initial investment cost for electrochemical energy storage, The total initial investment cost for hydrogen energy storage. , The first The year's first sky The power that a microgrid purchases from the main grid and sells back to the main grid at any given time. , These refer to the electricity purchase price and the electricity sales price from the power grid. The constraint construction unit is used to construct the constraint conditions of the microgrid multi-energy storage collaborative optimization configuration model, including hourly operation power balance constraints, daily rolling annual energy balance constraints, and source-load uncertainty quantification constraints. The power balance constraint for hourly operation is: ； In the above formula, for The power exchange of the power grid at any given time. , They are respectively The constant charging and discharging power of electrochemical energy storage , They are respectively The charging and discharging power of hydrogen energy storage at all times. , They are respectively The load reduction power and load transfer power required for real-time demand response. for Total load power of the microgrid at any given time , They are respectively Real-time photovoltaic and wind power output; The annual energy balance constraint for the daily rolling cycle is: ； In the above formula, For the number of days in a year, , The first sky The power input and output of the microgrid to the mains at any given time; The source load uncertainty quantification constraint is: ； ； ； ； In the above formula, This is the quantification range for source-load uncertainty power, which includes photovoltaic, wind power, and load. The source load uncertainty fluctuation coefficient, For the first sky Predicted power for time-source load uncertainty , , The first sky Always consider the actual power range of photovoltaic, wind power, and load after taking into account uncertainties. These are the uncertainty fluctuation coefficients for photovoltaic, wind power, and load, respectively. , , The first sky Predicted power output of photovoltaic, wind power, and load at all times.

8. The microgrid multi-element energy storage collaborative optimization system based on multiple time scales according to claim 6, characterized in that, The collaborative optimization configuration model solution module includes an initial set formation unit, a main problem construction unit, a main problem solving unit, a sub-problem construction unit, a sub-problem solving unit, a dynamic column filtering unit, a column combination range update unit, and an iterative discrimination unit. The initial set forming unit is used to randomly generate initial columns within the feasible domain of the variables in the microgrid multi-electrode energy storage collaborative optimization configuration model based on representative scenarios of source-load uncertainty, thus forming an initial set. ,in The integer capacity column corresponds to the first-stage capacity configuration and is generated. Integer capacity vectors that satisfy capacity and allocation constraints ; For continuous operation columns, corresponding to the second stage of operation optimization, each group of integer columns is matched. A group of continuous running vectors that satisfy the running constraints ; The main problem construction unit is used to construct the main problem based on an integer capacity vector and a continuously running vector, introducing combination coefficients: ； ； ； ； ； In the above formula, The total lifecycle cost of the main problem For the first The combination coefficient of a set of integer sequences For the first The investment cost corresponding to the set of integers. In order to be with the first The first group of integer columns matched The combination coefficient of a group of consecutive columns For the first The operation and maintenance costs corresponding to a group of consecutive columns, For the first The replacement cost corresponding to the set of integer columns, For the first The power grid interaction cost corresponding to a group of consecutive columns; The main problem solving unit is used to solve the main problem and obtain the current objective value of the main problem. And extract the dual variables of the constraints of the main problem. ,in The number of constraints; The subproblem construction unit is used to construct subproblems based on the dual variables of the main problem constraints, with the objective of maximizing column contribution values: ； In the above formula, For the total lifecycle cost of the subproblem, For the new integer column, For the new continuous column, The constraints of the main problem are in vector form. The inner product of the dual variable and the constraint; The subproblem-solving unit is used to solve subproblems and obtain the objective function values ​​corresponding to the new integer column and the continuous column. and judge Whether it is satisfied, among which, A contribution threshold is set for each column; if this threshold is met, the new integer column and the consecutive columns are added to the column set. If the conditions are not met, then column generation will stop. The dynamic column filtering unit is used to determine the column set at this time. If the number of integer columns exceeds the maximum threshold for the number of integer columns and the number of consecutive columns exceeds the maximum threshold for the number of consecutive columns, and if not, return to the sub-problem building unit to continue generating new columns and adding them to the column set. If the conditions are met, use the following formula to calculate the cumulative contribution value of each column, perform dynamic column filtering, and obtain a high-quality column set: ； In the above formula, For the first The cumulative contribution value of the column, The initial objective value for the main problem. To join the The target value of the main problem following the column. This is a set of filtered high-quality integer columns. Filter thresholds for contribution values. This is the set of high-quality continuous columns after filtering. The column combination range update unit is used to update the column combination range of the main problem to a high-quality column set. Subsequent iterations are based solely on... , Generate new columns and solve the problem, then modify the objective function of the subproblem as follows: ； In the above formula, As a representative scene, This refers to the total lifecycle cost in a single scenario. The iterative discrimination unit is used to solve the corrected subproblem objective value. If satisfied ,but , The newly generated column is valid. Add the newly generated column to the main problem and solve the main problem again to obtain the new objective value. If the difference between the objective values ​​of the main problem in two consecutive iterations satisfies or the number of times the column is generated reaches If the iteration terminates, otherwise return to the subproblem building unit to continue generating new columns.

9. The microgrid multi-element energy storage collaborative optimization system based on multiple time scales according to claim 6, characterized in that, The system also includes a phase constraint construction module; The stage constraint construction module is used to construct the first and second stage constraints of the microgrid multi-energy storage collaborative optimization configuration model. The first stage is the capacity configuration optimization stage, and the second stage is the real-time operation optimization stage.

10. The multi-time-scale microgrid multi-element energy storage collaborative optimization system according to claim 9, characterized in that, The constraints in the first stage include energy storage capacity constraints and capacity allocation constraints; The energy storage capacity constraint is: ； ； ； ； In the above formula, This is the rated power of electrochemical energy storage. This represents the upper limit of electrochemical energy storage power. For the rated energy capacity of electrochemical energy storage, This represents the upper limit of electrochemical energy storage capacity. This is the rated power for hydrogen energy storage. This represents the upper limit of hydrogen energy storage capacity. For the rated energy capacity of hydrogen energy storage, This represents the upper limit of hydrogen energy storage capacity; The capacity ratio constraint is: ； ； In the above formula, , These are the lower and upper limits of the electrochemical energy storage charge / discharge duration, respectively. , These are the lower and upper limits of hydrogen energy storage charge / discharge duration, respectively. The second stage constraints include energy storage charge and discharge power constraints, energy storage state of charge constraints, and demand response constraints. The energy storage charging and discharging power constraint is: ； ； ； ； In the above formula, , The first sky The charging and discharging power of electrochemical energy storage at all times. This is the rated power of electrochemical energy storage. , The first sky The charging and discharging power of the hydrogen energy storage system at all times. This is the rated power for hydrogen energy storage; The energy storage state of charge constraint is: ； ； In the above formula, For the first sky The state of charge of electrochemical energy storage at any given time. , These represent the lower and upper limits of charge for electrochemical energy storage, respectively. , , These are the lower and upper limits of charge for hydrogen energy storage, respectively. The demand response constraint is: ； ； In the above formula, For the first sky Real-time demand response load reduction power To reduce power to the maximum load in demand response, For the first sky real-time demand response load transfer power This represents the maximum load transfer power in demand response.

Images