Optical storage direct flexible micro-grid random optimization scheduling method considering demand response delay

Abstract

The invention belongs to the technical field of optical storage direct-flexible micro-grid dispatching, and particularly relates to an optical storage direct-flexible micro-grid random optimization dispatching method considering demand response delay. Comprising the following steps: constructing a double-layer collaborative optimization optical storage direct-flexible energy park micro-grid scheduling model comprising an energy operator EP and a load aggregator LA, and introducing delay response characteristics of three flexible loads, namely a translational load, a transferable load and a reducible load; an asymmetric electric energy transaction pricing model based on a Nash negotiation model is constructed, and the benefit distribution relation and bargaining capability difference of two parties are comprehensively considered; power fluctuation of a short time scale is obtained by adopting a random scene method, and distributed solution is realized by adopting an alternating direction multiplier method ADMM, so that a real-time scheduling scheme with relatively good convergence is obtained on the premise of protecting privacy. According to the strategy provided by the invention, load side economy and system power balance are considered, and unification of economy, fairness and privacy is realized.

Description

Technical Field

[0001] This invention belongs to the field of solar-storage-DC-flexible microgrid dispatching technology, and specifically relates to a stochastic optimization dispatching method for solar-storage-DC-flexible microgrids that takes into account demand response delay. Background Technology

[0002] Photovoltaic-storage-direct-current-flexible (PSDF) microgrids, with their advantages of high efficiency, flexibility, and cleanliness, can effectively improve energy utilization and promote the local consumption of renewable energy, making them an important development direction for building new power systems. However, with the increasing penetration of distributed generation (DG) sources such as wind and solar power in microgrids, the randomness and volatility of renewable energy output pose more severe challenges to the real-time power balance of microgrids. While traditional energy storage-based regulation can alleviate fluctuations to some extent, the limited capacity and high investment costs of energy storage systems make it difficult to independently handle power fluctuation mitigation. Therefore, demand-side response (DR), as a potentially economical regulation method, guides users to adjust their load power consumption to participate in power regulation, which is expected to improve the flexibility of source-storage-load coordinated operation and has significant research implications for improving the operational economy of photovoltaic-storage-direct-current-flexible microgrids. In recent years, demand-side response optimization has become a research hotspot in academia, and many fruitful research results have been achieved. However, existing research mainly explores the effect of demand-side response under ideal conditions, generally neglecting the demand-side response delay caused by user inertia, and rarely analyzing the differences in the enthusiasm of different users to participate in demand response, resulting in limited practical value. Furthermore, existing research typically optimizes microgrid operation based on day-ahead forecast data, lacking an exploration of the uncertainties affecting the generation and load sides. Moreover, most existing studies assume that power generation and consumption equipment belong to a single owner, neglecting the possibility of multiple stakeholders within a microgrid complex. In fact, the academic community has already focused on multi-stakeholder demand-side response optimization techniques for power grids and has made significant progress. However, existing research only ensures the lowest operating costs for each stakeholder within the consortium, failing to guarantee the maximization of the overall benefits of the consortium. Therefore, this invention proposes an energy microgrid scheduling method that considers the characteristics of demand response delay. Summary of the Invention

[0003] The purpose of this invention is to provide a stochastic optimal scheduling method for photovoltaic-storage-DC-flexible microgrids that considers demand response delays. This invention establishes a typical framework for photovoltaic-storage-DC-flexible microgrids and considers the delay characteristics of three types of flexible loads—shiftable loads, transferable loads, and loads that can be reduced—when participating in demand-side response. Combining Nash negotiation theory, an asymmetric electricity trading pricing mechanism is constructed based on the differences in bargaining power between the negotiating parties. A stochastic optimal scheduling method for photovoltaic-storage-DC-flexible microgrids is proposed, and Nash equilibrium is achieved between the cooperating parties.

[0004] To address the aforementioned technical issues, this invention provides a stochastic optimization scheduling method for photovoltaic-storage-DC-flexible microgrids that considers demand response delay. The method includes: constructing a two-layer collaborative optimization scheduling model for photovoltaic-storage-DC-flexible energy park microgrids involving energy operators (EP) and load aggregators (LA), introducing the delay response characteristics of three types of flexible loads—shiftable loads, transferable loads, and loads that can be reduced—to characterize the actual control behavior on the user side; constructing an asymmetric electricity trading pricing model based on the Nash negotiation model, comprehensively considering the interest distribution relationship and bargaining power differences between the two parties to achieve fair and efficient electricity trading pricing; and obtaining short-timescale power fluctuations using a stochastic scenario method and implementing distributed solution using the Alternating Directional Multiplier Method (ADMM) to ensure a real-time scheduling scheme with good convergence while protecting privacy.

[0005] Preferably, the system architecture of the solar-storage-direct-current-flexible energy park microgrid includes: an energy operator (EP) and a load aggregator (LA); the EP adopts a multi-energy complementary energy supply mode, consisting of distributed generation units composed of photovoltaic modules, wind turbine generators, energy storage batteries (BS), and gas turbines (GT); the LA is divided into shiftable loads, transferable loads, loads that can be reduced, and conventional loads; among which shiftable loads, transferable loads, and loads that can be reduced constitute adjustable flexible loads; the dispatch center issues flexible load adjustment instructions to the LA based on the forecast of the park's load and wind and solar power output data, and the dispatch center ensures power balance within the park by adjusting the power purchased and sold to the external grid, the power output of the GT, and the charging and discharging power of the BS.

[0006] Preferably, the scheduling model includes an EP model and an LA model; wherein the EP model constructs an EP comprehensive objective function with the objective of minimizing the daily comprehensive revenue of EP; wherein the daily comprehensive revenue of EP includes: electricity sales revenue, external grid electricity purchase cost, BS operating cost, and GT generation cost; the EP comprehensive objective function includes:

[0007] maxS L -C EG -C BS -C GT

[0008]

[0009] In the formula: S L For EP electricity sales revenue, C EG To reduce the cost of purchasing electricity from external power grids, C BS For BS operating costs, C GT For GT power generation costs, For EP electricity sales price, Let ω be the power output of EP at time t, and let s represent the scenario. s This represents the probability of scenario s occurring. Let represent the power purchased and sold to the external power grid at time t. The purchase and sale prices of electricity to the external power grid at time t are λ and t, respectively. BS The unit operating cost of BS, The charging and discharging power of BS at time t are respectively, and η is the charging and discharging power of BS c η d These represent the charge and discharge efficiencies of BS, Let be the output power of GT at time t. For GT start / stop status, a GT b GT and c GT These are the coefficients for the various factors related to the electricity generation cost of GT;

[0010] The LA model constructs an LA comprehensive objective function with the goal of minimizing the daily comprehensive cost of LA; wherein the daily comprehensive cost of LA includes: the cost of purchasing electricity from EP, the cost of offsettable load compensation, the cost of transferable load compensation, and the cost of load reduction compensation; the LA comprehensive objective function includes:

[0011] minC L +C sh +C tr +C re

[0012]

[0013]

[0014] In the formula: C L For LA electricity purchase cost, C sh For the total cost of compensating for the transferable load, C tr C is the total cost of compensating for transferable loads. re To compensate for the total cost of load reduction, For LA electricity purchase price, Let t be the total load demand of LA. The base load at time t, Let be the flexible load at time t. Let be the movable load at time t. Let be the transferable load at time t. Let c be the load that can be reduced at time t. sh P is the subsidy factor for the transferable load. t sh To optimize the power of the shiftable load at time t, C sh For the total cost of compensating for the transferable load, c tr P is the subsidy factor per unit power of transferable load. t trTo optimize the power of the shiftable load at time t, c re To reduce the subsidy price per unit of load power, P is the reduction factor. t re0 This represents the load power before reduction.

[0015] Preferably, the EP model includes the following constraints:

[0016] (1) Power balance constraint:

[0017]

[0018] In the formula: Let represent the power purchased and sold to the external power grid at time t. Let be the charging and discharging power of BS at time t. Let EP be the power sold at time t. For the photovoltaic output at time t, Let t be the output of the wind turbine. Let GT be the output power at time t;

[0019] (2) Generator power constraints:

[0020]

[0021] In the formula: Let t be the power purchased from the external power grid. To purchase the maximum amount of electricity from the grid, This represents the upper limit of photovoltaic power output. This is the upper limit of the wind turbine's output. Let be the output power of GT at time t. This represents the maximum output power of the GT.

[0022] (3) BS operation constraints:

[0023]

[0024] In the formula: Let be the charging and discharging power of BS at time t. This represents the maximum discharge power of BS. This represents the maximum charging power of BS. and These are all Boolean variables, serving as indicators of BS charging and discharging. Let η be the capacity of the BS at time t. c η d These represent the charge and discharge efficiencies of BS, Divided into upper and lower limits of BS capacity, The BS capacity at the initial moment within a scheduling cycle. This represents the BS capacity at the end of a scheduling cycle.

[0025] Preferably, the LA model includes the following constraints:

[0026] (1) Constraints for load transferable operation:

[0027]

[0028]

[0029] In the formula: All are Boolean variables, where This indicates the operating state of the load that can be shifted at time t. To issue a translation command, The actual response start flag for the transferable load is given by τ, where τ is the delay time and P is the load factor. t sh To optimize the power of the shiftable load at time t, P t sh0 To optimize the power of the shiftable load at time t, M is a large positive integer. T represents the load transferable interval within the scheduling cycle. sh The translational duration;

[0030] (2) Load transferability operation constraints:

[0031]

[0032] In the formula: All are Boolean variables, where This indicates the operating status of the load that can be transferred at time t. To issue a transfer instruction flag, P is the actual response start flag for transferable loads. t trx The power of the shiftable load at time t is the power that can be adjusted. P is the maximum power that the shiftable load can adjust at time t. t tr To optimize the power of the shiftable load at time t, P t tr0 To optimize the power of the transferable load at time t, The load transferable range within the scheduling cycle. M is the minimum transferable duration, where M is a large positive integer;

[0033] (3) Load constraints can be reduced:

[0034]

[0035]

[0036] In the formula: P t re0 The load power before reduction, For load reduction rate, Let M be the load reduction interval within the scheduling period, where M is a large positive integer. The reduction factor is... This is a signal to issue a translation command; τ is the delay time. This is a signal to issue a reduction order;

[0037] (4) Power balance constraints for EP and LA participating in power trading:

[0038]

[0039] In the formula: Let EP be the power sold at time t. Let t be the total load demand of LA at time t.

[0040] Preferably, obtaining short-timescale power fluctuations using the random scenario method includes: the prediction errors of photovoltaic and loads follow a normal distribution; scenarios are generated using Monte Carlo simulation; the obtained load and wind / solar output scenarios are clustered and reduced using k-means; and the statistical laws governing their power fluctuation characteristics include:

[0041]

[0042] In the formula: For the photovoltaic output at time t, Let t be the output of the wind turbine. The normal load output at time t, and These are the predicted values ​​for the conventional load, photovoltaic power output, and wind power output for hour t, respectively. and For the corresponding error distribution parameters, Let be the mean of the photovoltaic power output at time t, which follows a normal distribution. Let be the mean of the wind turbine output at time t, following a normal distribution. Let t be the mean of the load under a normal distribution. Let be the standard deviation of photovoltaic output at time t. Let be the standard deviation of the wind turbine output at time t. Let be the standard deviation of the load at time t.

[0043] Preferably, the asymmetric electricity trading pricing model includes:

[0044]

[0045] In the formula: α EP α LA These are the bargaining power values ​​for EP and LA, respectively. L For EP electricity sales revenue, C EG To reduce the cost of purchasing electricity from external power grids, C BS For BS operating costs, C GT For GT power generation costs, For EP's revenue before participating in cooperative transactions, For the costs incurred by LA before participating in the cooperative transaction, C L For LA electricity purchase cost, C sh For the total cost of compensating for the transferable load, C tr C is the total cost of compensating for transferable loads. re To compensate for the total cost of load reduction, For EP electricity sales price, The electricity purchase price in LA.

[0046] Preferably, the Nash negotiation model includes:

[0047]

[0048] In the formula: S m For the benefit of participant m after participating in the cooperation, The benefit of independent entity m before cooperation, i.e. the point at which negotiations break down.

[0049] Preferably, the method for solving the Nash negotiation model also includes: decoupling the Nash negotiation model into a microgrid park cost minimization subproblem and an electricity price setting subproblem; the objective function for minimizing the microgrid park cost includes:

[0050] minC=C EG +C BS +C GT +C sh +C tr +C re

[0051] In the formula: C represents the cost of the microgrid park, C EG To reduce the cost of purchasing electricity from external power grids, C BS For BS operating costs, C GT For GT power generation costs, C sh For the total cost of compensating for the transferable load, C tr C is the total cost of compensating for transferable loads. re The total cost of compensating for load reduction;

[0052] The constraints on the electricity price setting include: (1) the transaction price shall not be higher than the purchase price from the external power grid, nor lower than the sales price to the external power grid, and the constraints include:

[0053]

[0054] In the formula: For EP electricity sales price, For LA electricity purchase price, Let t be the electricity price sold to the external power grid. Let t be the price at which electricity is purchased from the external power grid.

[0055] (2) Ensure that the revenue of EP after participating in cooperative transactions is higher than its revenue when generating electricity at zero cost, and that the cost of LA after participating in cooperative transactions is lower than the cost of purchasing electricity from external grids in a single scenario. The constraints include:

[0056]

[0057] In the formula: S L For EP electricity sales revenue, C EG To reduce the cost of purchasing electricity from external power grids, C BS For BS operating costs, C GT For GT power generation costs, C sh For the total cost of compensating for the transferable load, C tr C is the total cost of compensating for transferable loads. re To compensate for the total cost of load reduction, For EP's revenue before participating in cooperative transactions, For the costs incurred by LA before participating in the cooperative transaction, C L For LA electricity purchase cost, For the photovoltaic output at time t, Let t be the output of the wind turbine. Let ω be the electricity price sold to the external power grid at time t. s This represents the probability of scenario s occurring. Let t be the total load demand of LA. Let t be the price at which electricity is purchased from the external power grid.

[0058] (3) The bargaining power constraints of EP and LA include:

[0059]

[0060] In the formula: α j To enhance EP's bargaining power with LA, The electrical energy that a certain entity can supply. The electrical energy that a particular entity can harvest. This represents the upper limit of electrical energy that can be supplied across all entities. This represents the upper limit of electrical energy that can be supplied across all entities. The power that a certain entity can provide. This refers to the power that a certain entity can absorb.

[0061] Preferably, the ADMM algorithm includes: (1) setting the maximum number of iterations and convergence accuracy, initializing the augmented Lagrange multipliers and the number of iterations, and calculating the bargaining power α between EP and LA. EP α LA (2) Introduce the penalty coefficient and Lagrange multiplier into the combined objective function of EP and LA respectively to form an augmented Lagrange optimization objective function; that is:

[0062]

[0063] In the formula: S L For EP electricity sales revenue, C EG To reduce the cost of purchasing electricity from external power grids, C BS For BS operating costs, C GT For GT power generation costs, L is the revenue that EP receives before participating in the cooperative transaction. EP L LA Let EP and LA be the augmented Lagrangian optimization objective functions, respectively. For EP electricity sales price, For LA electricity purchase price, For the costs incurred by LA before participating in the cooperative transaction, C L For LA electricity purchase cost, C sh C represents the total cost of compensating for the transferable load. tr C is the total cost of compensating for transferable loads. re To compensate for the total cost of load reduction, Lagrange multipliers for EP and LA respectively, ρ EP ρ LA The penalty coefficients for EP and LA are α and α, respectively. EP α LA (3) Update the electricity trading price decisions for EP and LA; EP and LA each calculate their electricity trading price strategies and exchange quotation information with each other. The electricity price strategy update method for both is shown in the following formula:

[0064]

[0065] In the formula: For EP electricity sales price, Let k be the electricity purchase price in LA, and k be the iteration number. The Lagrange multipliers for EP and LA are respectively; (4) Lagrange multiplier update; the Lagrange multiplier update method is shown in the following formula:

[0066]

[0067] In the formula: Let EP be the power sold at time t. Let ρ be the total load demand of LA at time t; EP ρ LA (5) The number of iterations k = k + 1; (6) The convergence condition; During the iteration process, when the original residual and the dual parameter residual are both less than their corresponding convergence accuracy, and the value of k is less than its maximum value, the algorithm is judged to be converged. The specific constraints are as follows:

[0068]

[0069] Where: ε1 is the convergence accuracy of the original residual; ε2 is the convergence accuracy of the dual residual; (7) Terminate the iteration; the iteration terminates when the above convergence conditions are met, otherwise return to step (2) until the iteration termination convergence condition is met or the maximum number of iterations k is reached. max Then stop.

[0070] Compared with existing technologies, this invention has the following advantages: First, it proposes a stochastic optimization scheduling strategy that considers demand response delay characteristics. Firstly, it constructs a two-layer collaborative optimization model involving energy operators (EPs) and load aggregators (LAs), introducing the delay response characteristics of three types of flexible loads: shiftable loads, transferable loads, and loads that can be reduced, to characterize the actual control behavior on the user side. Secondly, it establishes an asymmetric bargaining mechanism based on Nash negotiation theory, comprehensively considering the differences in bargaining power between the two parties to achieve fair and efficient electricity trading pricing. To address the uncertainties in wind and solar power output and load, it employs Monte Carlo scenario generation and K-means clustering reduction to construct typical operating scenarios, and utilizes the Alternating Direction Multiplier Method (ADMM) for distributed solution. Simulation results show that the proposed model can effectively reduce load aggregator costs by 0.79%, increase energy operator revenue by 1.28%, and reduce overall park operating costs by 6.76%. The proposed strategy balances load-side economy and system power balance, achieving a unity of economy, fairness and privacy. It provides a feasible solution for the optimized scheduling of microgrids with a high proportion of renewable energy access, and effectively reduces scheduling deviation by quantifying the delay probability distribution. Attached Figure Description

[0071] Figure 1 A flowchart of a stochastic optimization scheduling method for a photovoltaic-storage-DC-flexible microgrid that takes into account demand response delay, provided in an embodiment of the present invention.

[0072] Figure 2 The provided embodiment of the present invention provides a structural diagram of a park microgrid system.

[0073] Figure 3 The base load and wind and solar power output prediction curves are provided for embodiments of the present invention.

[0074] Figure 4 The flowchart of the cooperative game between EP and LA provided in the embodiment of the present invention.

[0075] Figure 5 This diagram illustrates the convergence of EP and LA negotiations in an embodiment of the present invention.

[0076] Figure 6 The following are the transaction electricity prices and comparison charts provided in the embodiments of the present invention; wherein (a) is a transaction electricity price chart after asymmetric negotiation in four scenarios; and (b) is a comparison chart of the impact of flexible load response delay on power.

[0077] Figure 7 The following diagrams illustrate the optimization results provided in the embodiments of the present invention: (a) is the energy supply and demand optimization result diagram for Scheme 1; and (b) is the energy supply and demand optimization result diagram for Scheme 2. Detailed Implementation

[0078] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. The advantages and features of the present invention will become clearer from the following description. It should be noted that the drawings are all in a very simplified form and use non-precise proportions, and are only used to facilitate and clarify the illustration of the embodiments of the present invention.

[0079] like Figure 1 As shown in the figure, this invention specifically provides a stochastic optimization scheduling method for photovoltaic-storage-DC-flexible microgrids that considers demand response delay. The method includes: First, a two-layer collaborative scheduling model is constructed, consisting of an Energy Producer (EP) and a Load Aggregator (LA), and a demand response delay model is introduced to improve the accuracy of scheduling decisions. Second, taking into account the distribution of interests and differences in bargaining power between the two parties, an electricity trading pricing mechanism based on asymmetric Nash negotiation is established. Finally, a stochastic scenario method is used to obtain short-timescale power fluctuations, and the Alternating Directional Multiplier Method (ADMM) is used to achieve distributed solution, ensuring a real-time scheduling scheme with good convergence while protecting privacy.

[0080] To provide a more detailed description of the present invention, the embodiments of the present invention also include: a microgrid scheduling model for photovoltaic-storage-DC-flexible energy parks. Figure 2This is a diagram of the microgrid system structure of the industrial park. The park is mainly divided into two parts: EP (Energy Utilization) and LA (Local Area). The EP adopts a multi-energy complementary energy supply mode, mainly composed of distributed generation units such as photovoltaic modules, wind turbine generators, BS (Battery Grid), and GT (Gross Grid). Based on the forecast of the park's load and wind and solar power output, the dispatch center issues flexible load adjustment instructions to the LA. Since flexible loads will have a delay when actually participating in load response, and there is a certain deviation between the actual and forecast values ​​of load and wind and solar power output within the park, the dispatch center can ensure power balance within the park by adjusting the power purchased and sold to the external grid, the power output of GT, and the charging and discharging power of BS.

[0081] As a preferred technical solution in this embodiment of the invention, the above-mentioned EP model includes: an EP comprehensive objective function; the daily comprehensive revenue of EP includes electricity sales revenue, external grid electricity purchase cost, BS operating cost and GT generation cost, and each objective function is shown in equations (1)-(5). Among them, equation (1) is the EP comprehensive objective function; equation (2) is the electricity sales revenue function to LA; equation (3) is the external grid electricity purchase cost function; equation (4) is the BS operating cost function; and equation (5) is the GT generation cost function.

[0082] max S L -C EG -C BS -C GT (1)

[0083]

[0084] In the formula: S L For EP electricity sales revenue, C EG To reduce the cost of purchasing electricity from external power grids, C BS For BS operating costs, C GT Cost of generating electricity for GT; For EP electricity sales price, Let ω be the power output of EP at time t, and let s represent the scenario. s This represents the probability of scenario s occurring; These represent the power purchased and sold to the external power grid at time t, respectively. λ represents the purchase and sale price of electricity to the external power grid at time t; BS This refers to the unit operating cost of the BS. The charging and discharging power of BS at time t are respectively; η c η d These are the charge and discharge efficiencies of BS, respectively. Let be the output power of GT at time t. For GT start / stop status, a GT b GT and c GT These are the coefficients for the various factors related to the power generation cost of GT.

[0085] Constraints; 1) Power balance constraint

[0086] In the formula: These represent the power purchased and sold to the external power grid at time t, respectively. , respectively, represent the charging and discharging power of BS at time t; Let EP be the power sold at time t. For the photovoltaic output at time t, Let t be the output of the wind turbine. Let GT be the output power at time t. 2) Generator power constraints:

[0087]

[0088] In the formula: Let be the power purchased from the external power grid at time t; To purchase the maximum amount of electricity from the grid, This represents the upper limit of photovoltaic power output. This is the upper limit of the wind turbine's output. Let be the output power of GT at time t. 3) BS operation constraints: When the BS performs charging and discharging operations, the charging and discharging power shall not exceed the corresponding maximum power limit, and the energy state shall be maintained within its capacity range. The relevant constraints are as shown in equations (11) to (16). Among them, equation (11) is the BS discharge power limit; equation (12) is the BS charging power limit; equation (13) indicates that the BS charging and discharging shall not be carried out simultaneously; equation (14) is the BS energy state calculation method; equation (15) is the BS energy state limit; equation (16) is the constraint for the BS to achieve energy cycle consistency within the cycle, ensuring that the energy state is equal at the beginning and end.

[0089]

[0090]

[0091] In the formula: , respectively, represent the charging and discharging power of BS at time t; This represents the maximum discharge power of BS. This represents the maximum charging power of BS. and These are all Boolean variables, serving as indicators of BS charging and discharging. Let η be the capacity of the BS at time t; c η d These are the charge and discharge efficiencies of BS, respectively. Divided into upper and lower limits of BS capacity; The BS capacity at the initial moment within a scheduling cycle; This represents the BS capacity at the end of a scheduling cycle.

[0092] As a preferred technical solution in this embodiment of the invention, the above-mentioned LA model includes: a cost function; based on the response characteristics of the load in demand response optimization, the load is divided into two categories: non-adjustable base load and adjustable flexible load. In the formula: Let t be the total load demand of LA. The base load at time t, Let t be the flexible load at time t. Adjustable flexible loads, when participating in load demand response, consider response delay. Based on their response characteristics, they are now classified into three categories: loads that can be moved (washing machines, dishwashers, etc.), loads that can be transferred (charging equipment such as electric vehicles), and loads that can be reduced (air conditioners, etc.). The relevant constraints are as follows: In the formula: Let be the flexible load at time t. Let be the movable load at time t. Let be the transferable load at time t. Let t be the load that can be reduced.

[0093] The daily comprehensive cost of LA includes the cost of purchasing electricity from EP, the cost of load shifting compensation, the cost of load transfer compensation, and the cost of load reduction compensation. The relevant functions are shown in equations (19) and (20). Equation (19) is the comprehensive objective function; equation (20) is the LA electricity purchase cost function.

[0094] min C L +C sh +C tr +C re (19)

[0095]

[0096] In the formula: C L For LA electricity purchase cost; C sh C represents the total cost of compensating for the transferable load. tr C is the total cost of compensating for transferable loads. re The total cost of compensating for load reduction; For LA electricity purchase price, Let t be the total load demand of LA at time t.

[0097] The load transfer constraint includes: transferable electrical loads can be flexibly adjusted across multiple time periods by setting allowable transfer time windows, thereby achieving load transfer across time periods. Assume the load transferable interval within the scheduling cycle is... The translation duration is T shThe relevant constraints are shown in equations (21) to (27). Among them, equations (21) to (22) are the operating power constraints of the linearized transferable load; equation (23) is the actual response start-up constraint; equation (24) is the operating time range of the transferable load; equation (25) is the operating time constraint; equation (26) indicates that there is only one start-up time; equation (27) indicates that the transferable load operates continuously.

[0098]

[0099] In the formula: All are Boolean variables, where This indicates the operating state of the load that can be shifted at time t. To issue a translation command, The actual response start flag for the transferable load; τ is the delay time; P t sh To calculate the power of the optimized transferable load at time t; P t sh0 To optimize the power of the shiftable load at time t, M is a large positive integer. T represents the load transferable interval within the scheduling cycle. sh The translational duration.

[0100] Shiftable load compensation includes: In the formula: c sh P is the subsidy factor for the transferable load. t sh To optimize the power of the shiftable load at time t, C sh The total cost of compensating for the transferable load.

[0101] Transferable load constraints include: high flexibility of transferable loads, allowing for arbitrary adjustment within a specified range throughout the entire scheduling cycle; and setting the transferable load range within the scheduling cycle as... To avoid frequent start-ups and shutdowns of the equipment, constraints are imposed as shown in equations (29) to (34). Among them, equations (29) to (30) are the operating power constraints of the transferable load after linearization; equation (31) is the actual response start constraint; equation (32) is the power limit of the transferable load; equation (33) indicates that the transferable load is continuously running; and equation (34) indicates that the total transferable load remains unchanged.

[0102]

[0103] In the formula: All are Boolean variables, where This indicates the operating status of the load that can be transferred at time t. To issue a transfer instruction flag, P is the actual response start flag for transferable loads;t trx The power of the shiftable load at time t is adjustable. P is the maximum power that the shiftable load can adjust at time t. t tr To calculate the power of the optimized transferable load at time t; P t tr0 To optimize the power of the transferable load at time t, The load transferable range within the scheduling cycle. M is the minimum transferable duration, where M is a large positive integer.

[0104] Transferable load compensation: In the formula: c tr P is the subsidy factor per unit power of transferable load. t tr To optimize the power of the shiftable load at time t, C tr The total cost of compensating for transferable loads.

[0105] The load reduction constraints include: the load reduction is mainly determined by the user's reasonable reduction of the corresponding load power according to the power consumption plan. In order to ensure the rationality of the reduction of the load reduction, the constraints are as shown in equations (36)-(38). Among them, equations (36)-(37) are the operating power constraints of the load reduction after linearization; equation (38) is the time limit for reduction.

[0106]

[0107] In the formula: P t re0 The load power before reduction; For load reduction rate, Let M be the load reduction interval within the scheduling period, where M is a large positive integer. This is a flag indicating that a reduction order has been issued.

[0108] Load compensation can be reduced: In the formula: c re To reduce the subsidy price per unit of load power, P is the reduction factor. t re0 C represents the load power before reduction; re The total cost of compensating for load reduction. Power balance constraints include: EP and LA participating in electricity trading must meet power balance constraints and trading price balance constraints.

[0109]

[0110] In the formula: Let EP be the power sold at time t. Let t be the total load demand of LA at time t.

[0111] As a preferred technical solution in this embodiment of the invention, since photovoltaic output, wind power output, and base load all have uncertainties, the predicted values ​​of photovoltaic output, wind power output, and base load deviate from the actual values ​​to a certain extent. The statistical law that their fluctuation characteristics follow is shown in equations (43)-(44). If the solution is directly based on the predicted values, the error between the predicted results and the actual results will be large. Therefore, this invention assumes that the prediction error of photovoltaic and load follows a normal distribution with a standard deviation of 10%, and uses the Monte Carlo simulation method to generate 1000 scenarios. Considering the computational complexity and speed, k-means is used to cluster and reduce the obtained load and wind and solar power output scenarios, and four typical scenarios are extracted for analysis.

[0112]

[0113] In the formula: For the photovoltaic output at time t, Let t be the output of the wind turbine. The normal load output at time t, and These are the predicted values ​​for the conventional load, photovoltaic power output, and wind power output for hour t, respectively. and These are the corresponding error distribution parameters. Let be the mean of the photovoltaic power output at time t, which follows a normal distribution. Let be the mean of the wind turbine output at time t, following a normal distribution. Let t be the mean of the load under a normal distribution. Let be the standard deviation of photovoltaic output at time t. Let be the standard deviation of the wind turbine output at time t. Let be the standard deviation of the load at time t.

[0114] As a preferred technical solution in this embodiment of the invention, the invention also includes a pricing mechanism based on asymmetric Nash negotiation; by proposing an EP-LA cooperative game framework based on Nash negotiation and constructing an asymmetric bargaining mechanism to realize energy transaction pricing. Nash negotiation theory; the constructed Nash negotiation model is a cooperative game, which, based on maximizing the interests of both participants, then sets the transaction price through a negotiation mechanism. The Nash negotiation model needs to satisfy symmetry and Pareto optimality, and its standard form is shown in equations (45)-(46).

[0115]

[0116] In the formula: S m For the benefit of participant m after participating in the cooperation, The benefit of independent entity m before cooperation is also the point of breakdown in negotiations. The solution that maximizes the Nash product is the equilibrium solution of the Nash negotiation game problem. Because the Nash negotiation model has strong multivariate coupling and non-convex nonlinear characteristics, it is decoupled into a microgrid park cost minimization subproblem (P1) and an electricity price setting subproblem (P2), and solved sequentially.

[0117] As a preferred technical solution in this embodiment of the invention, the solution method of the above-mentioned Nash negotiation model includes: the sub-problem of minimizing the park cost (P1): minimizing the microgrid park cost, i.e., minimizing the sum of the total daily comprehensive costs of EP and LA. In the process of solving sub-problem P1, due to the coupling of the power purchase and sale objective functions between the two-layer structure, power trading can be ignored. Specifically, it can be expressed as:

[0118]

[0119] In the formula: C represents the cost of the microgrid park, C EG To reduce the cost of purchasing electricity from external power grids, C BS For BS operating costs, C GT Cost of generating electricity for GT; C sh C represents the total cost of compensating for the transferable load. tr C is the total cost of compensating for transferable loads. re The total cost of compensation for load reduction. Subproblem P1 can be directly solved according to equation (47), and the optimal scheme for real-time interaction power between EP and LA is obtained. Electricity pricing subproblem (P2): When setting the electricity trading price of EP and LA, the distribution of benefits should be fully considered, that is, the trading price should not be higher than the purchase price of electricity from the external grid, nor lower than the sales price of electricity from the external grid. The relevant constraints are shown in equation (48).

[0120]

[0121] In the formula: For EP electricity sales price, For LA electricity purchase price, Let t be the electricity price sold to the external power grid. Let t be the purchase price of electricity from the external grid at time t. Simultaneously, it is necessary to ensure that the revenue of EP after participating in the cooperative transaction is higher than its revenue when generating electricity at zero cost, and that the cost of LA after participating in the cooperative transaction is lower than the cost under the single scenario of purchasing electricity from the external grid. The relevant constraints are shown in equations (49) to (52). Equation (49) guarantees an increase in the total revenue of EP after cooperation; equation (50) guarantees a decrease in the total cost of LA after cooperation; equation (51) is the revenue function of EP before cooperation; and equation (52) is the total cost function of LA before cooperation.

[0122]

[0123]

[0124] In the formula: S L For EP electricity sales revenue, C EG To reduce the cost of purchasing electricity from external power grids, C BS For BS operating costs, C GT Cost of generating electricity for GT; C sh C represents the total cost of compensating for the transferable load. tr C is the total cost of compensating for transferable loads. re To compensate for the total cost of load reduction, This refers to the revenue EP receives before participating in cooperative transactions; For the costs incurred by LA before participating in the cooperative transaction, C L For LA electricity purchase cost, For the photovoltaic output at time t, Let t be the output of the wind turbine. Let ω be the electricity price sold to the external power grid at time t. s This represents the probability of scenario s occurring; Let t be the total load demand of LA. Let t be the price at which electricity is purchased from the external power grid.

[0125] Asymmetric bargaining refers to a game state in which the bargaining power gradient is formed among the negotiating parties due to initial information or power asymmetry. This invention proposes a nonlinear function based on the natural logarithm to quantify the contributions of EP and LA, using equations (53)-(55) to quantify the bargaining power α of EP and LA. j .

[0126]

[0127] In the formula: α j To enhance EP's bargaining power with LA, Electrical energy that a particular entity can supply; The electrical energy that a particular entity can harvest; This represents the upper limit of electrical energy that can be supplied across all entities. This represents the upper limit of electrical energy that can be supplied across all entities. The power that a certain entity can provide; This refers to the power that a certain entity can absorb.

[0128] Based on the above analysis, the optimal real-time interactive power obtained in subproblem P1 is substituted into subproblem P2. Based on the standard formulas of Nash negotiation (45)-(46), and utilizing the strictly monotonically increasing convex function characteristics of the natural logarithm, the asymmetric pricing model of the microgrid park is constructed as shown in formula (56).

[0129]

[0130] The transaction balance constraint is as shown in equation (57):

[0131] When satisfied This indicates that EP and LA have reached an agreement on a trading price for electricity. Where: S L For EP electricity sales revenue, C EG To reduce the cost of purchasing electricity from external power grids, C BS For BS operating costs, C GT Cost of generating electricity for GT; This refers to the revenue EP receives before participating in cooperative transactions; For the costs incurred by LA before participating in the cooperative transaction, C L For LA electricity purchase cost; C sh C represents the total cost of compensating for the transferable load. tr C is the total cost of compensating for transferable loads. re To compensate for the total cost of load reduction, For EP electricity sales price, The electricity purchase price in LA.

[0132] As a preferred technical solution in this embodiment of the invention, the invention also includes the following ADMM solution algorithm; the above subproblem P2 is solved using the ADMM algorithm. The specific steps are as follows: 1) Initialization; set the maximum number of iterations and convergence accuracy, initialize the augmented Lagrange multipliers and the number of iterations, and calculate the bargaining power α of EP and LA according to equations (53)-(55). EP α LA 2) Introduce the penalty coefficient and Lagrange multiplier into the objective functions of EP and LA respectively, and write them as augmented Lagrange optimization objective functions as shown in equations (58)-(59).

[0133]

[0134] In the formula: S L For EP electricity sales revenue, C EG To reduce the cost of purchasing electricity from external power grids, C BS For BS operating costs, C GT For GT power generation costs, L is the revenue that EP receives before participating in the cooperative transaction. EP L LA Let EP and LA be the augmented Lagrangian optimization objective functions, respectively. For EP electricity sales price, For LA electricity purchase price, For the costs incurred by LA before participating in the cooperative transaction, C L For LA electricity purchase cost, C sh C represents the total cost of compensating for the transferable load. trC is the total cost of compensating for transferable loads. re To compensate for the total cost of load reduction, Lagrange multipliers for EP and LA respectively; ρ EP ρ LA The penalty coefficients for EP and LA are α and α, respectively. EP α LA These are the bargaining power values ​​of EP and LA, respectively. 3) Update the electricity price decision of EP and LA; EP and LA each calculate their own electricity price strategy and exchange quotation information with each other. The electricity price strategy update method of the two is as shown in equations (60)-(61).

[0135]

[0136] In the formula: For EP electricity sales price, Let k be the electricity purchase price in LA, and k be the iteration number. 4) Lagrange multipliers update; the Lagrange multiplier update method is as shown in equations (62)-(63).

[0137]

[0138] 5) Update iteration number k = k + 1. 6) Convergence condition; During the iteration process, when the original residual and the dual parameter residual are both less than their corresponding convergence accuracy, and the value of k is less than its maximum value, the algorithm can be judged to be converged. The specific constraints are as shown in equations (64) - (65).

[0139]

[0140] Where: ε1 is the convergence accuracy of the original residual; ε2 is the convergence accuracy of the dual residual. 7) Terminate the iteration; terminate the iteration when equations (64)-(65) are satisfied, otherwise return to step 2), until the iteration termination condition is met or the maximum number of iterations k is reached. max Then stop.

[0141] As a preferred technical solution in the embodiments of the present invention, the present invention also includes the following simulation analysis process: (1) Simulation parameters; In order to verify the effectiveness of the photovoltaic-storage-direct-flexible energy microgrid scheduling model that considers demand response delay proposed in the present invention, the present invention constructs a park including wind turbines, photovoltaic power plants, GT and BS, etc. Figure 3The 24-hour predicted power curves for load, wind power output, and photovoltaic output in the park are displayed. Based on this data, four typical scenarios are extracted and analyzed after Monte Carlo scenario generation and k-means clustering reduction. Parameters related to flexible loads are shown in Tables 1-3. The delay times for flexible loads participating in demand response are set to 15 min, 30 min, 45 min, and 60 min, with corresponding proportions of 0.12, 0.46, 0.2, and 0.16, respectively. The rated operating power of GT is 500 kW, and the consumption parameters are a = 0.0018025 yuan / (kWh)², b = 0.52575 yuan / (kWh), and c = 2.9 yuan. The rated charging and discharging power of BS is 240 kW, the capacity is 480 kWh, the initial SOC is set to 0.6, the charging and discharging efficiency is 0.95, and the lifetime loss cost is 0.5 yuan / kWh. The scheduling cycle is 24 hours, and the scheduling step size is 15 minutes. The optimization model was solved using Matlab with the help of Yalmip and Gurobi.

[0142] Table 1 Parameters of Transferable Loads

[0143]

[0144] Table 2 Transferable Load Parameters

[0145]

[0146] Table 3 Load Reduction Parameters

[0147]

[0148] This invention will compare two schemes to verify the superiority of the proposed strategy. Scheme 1: Scheduling decisions are made considering the delay in flexible load response, simulating real microgrid operation; Scheme 2: Scheduling decisions are made without considering the delay in flexible load response, simulating real microgrid operation.

[0149] (2) Simulation results analysis; Algorithm convergence analysis; The ADMM algorithm is used in this invention to solve subproblem P2 in a distributed manner. The convergence results of the asymmetric bargaining iteration between EP and LA are as follows: Figure 5 As shown, (a) is the iterative result of the original residual, and (b) is the iterative result of the dual residual. After 54 iterations, the algorithm proposed in this invention converges to 10 for both the original and dual residuals. -3 Within a timeframe of 310.26 seconds, the algorithm was successfully completed. This demonstrates that the distributed optimization algorithm proposed in this invention exhibits excellent convergence performance and computational efficiency, effectively balancing privacy protection requirements with the real-time requirements of optimized scheduling. Analysis of electricity trading results; Figure 7 Figure (a) shows the energy supply and demand optimization results for EP and LA in four scenarios of Scheme 1. Figure 7 As shown in (a), the main sources of electricity for EP are renewable energy output and purchases from the external grid. The gas turbine output time is from 7:15 to 20:15. The energy storage battery charges slowly when there is surplus wind and solar power and discharges promptly when there is insufficient electricity. The operating time of LA's load transferable load is from 5:00 to 6:45 and from 11:15 to 14:00, and the operating time of transferable load is from 10:15 to 16:00. The load that can be reduced decreases during the reductionable time. The electricity transaction prices between EP and LA under the four scenarios after optimization based on the asymmetric bargaining scheme proposed in Section 2.2 are as follows: Figure 6 As shown in Figure (a), the transaction prices in each scenario fall within the range between the purchase price from the external grid and the sales price. This demonstrates that the participation of EP and LA in power trading cooperation can bring more revenue to EP, while also significantly reducing LA's total electricity costs. The impact of different degrees of delay in flexible load response on power is compared below. Figure 6 As shown in (b), by Figure 6 It can be seen that varying degrees of delay in the participation of flexible loads in demand-side response will affect real-time power. Compared with on-time response, response delays of 15 min, 30 min, and 45 min all cause changes in peak power. For example, different degrees of delay before and after 20:00 all lead to an increase in load power. The EP and LA energy supply and demand optimization results for the four scenarios in Scheme 2 are as follows: Figure 7 As shown in (b), the change in charging and discharging power of the energy storage battery is most significant compared to Scheme 1. For example, in Scheme 2, the load energy storage battery exhibits charging at 11:45 and 17:30 due to the delayed response of the flexible load, while the energy storage battery in Scheme 1 has no charging power during this period. The delay effect causes the overall flexible load response in Scheme 2 to be delayed.

[0150] (3) Cost and Benefit Analysis; Table 4 shows a comparison of the operating costs and benefits of EP under the two schemes. The results show that the electricity purchase cost, BS cost, and GT cost of Scheme 1 are all lower than those of Scheme 2, and the corresponding transaction amount is also slightly lower. However, the daily comprehensive benefit of Scheme 1 is 6699.77 yuan, which is 85.91 yuan (1.28%) higher than that of Scheme 2. This indicates that introducing flexible load delay scheduling in Scheme 1 effectively improves the power optimization effect, thereby achieving higher overall economic benefits.

[0151] Table 4 Comparison of EP-side optimization results for each scheme

[0152]

[0153]

[0154] Table 5 shows the operating costs and compensation for the LA under the two schemes. The results indicate that the compensation amounts for movable and reducible loads are the same in both schemes. The compensation cost for movable loads in Scheme 1 is 378.78 yuan, slightly higher than Scheme 2 (4.43%). However, the daily comprehensive cost of the LA in Scheme 1 is 8857.63 yuan, a decrease of 68.35 yuan (0.79%) compared to Scheme 2. Therefore, Scheme 1 effectively reduces the operating costs of the LA and the overall park while ensuring flexible load scheduling, demonstrating superior economic efficiency.

[0155] Table 5 Comparison of LA-side optimization results for each scheme

[0156]

[0157] From a social welfare perspective, the total comprehensive costs of the park for Scheme 1 and Scheme 2 are RMB 2155.84 and RMB 2312.12, respectively, with Scheme 1 significantly reducing costs by RMB 156.28 (6.76%) compared to Scheme 2. This result demonstrates that the strategy proposed in this invention has a significant economic advantage over conventional scheduling strategies. It effectively reduces the comprehensive costs of the LA while increasing the daily comprehensive benefits of the EP, thus balancing the interests of both parties; it also effectively protects the privacy between the EP and the LA. Therefore, the microgrid scheduling strategy proposed in this invention achieves a balance between economy and security, and better reflects the optimization orientation aimed at maximizing social welfare.

[0158] In summary, this invention utilizes the ADMM algorithm to solve the electricity price trading model, exchanging only the electricity price and power information between the EP and LA, effectively protecting their privacy. Furthermore, the algorithm exhibits excellent convergence in this invention. The constructed asymmetric Nash negotiation model effectively avoids the EP's absolute dominance in price setting, maintaining the bargaining power of load operators. Flexible load response delays significantly affect the system's real-time power balance; traditional scheduling ignores the delay effect, leading to poor energy dispatching strategies. However, this invention's model effectively reduces dispatching bias by quantifying the delay probability distribution.

[0159] The above description is merely a description of preferred embodiments of the present invention and is not intended to limit the scope of the present invention in any way. Any changes or modifications made by those skilled in the art based on the above disclosure shall fall within the protection scope of the claims.

Claims

1. A stochastic optimization scheduling method for a photovoltaic-storage-DC-flexible microgrid considering demand response delay, characterized in that, include: By constructing a two-layer collaborative optimization microgrid dispatch model for photovoltaic-storage-direct-flexible energy parks that includes energy operators (EP) and load aggregators (LA), the delayed response characteristics of three types of flexible loads—shiftable loads, transferable loads, and loads that can be reduced—are introduced to characterize the actual control behavior on the user side. By constructing an asymmetric electricity trading pricing model based on the Nash negotiation model, and comprehensively considering the distribution of interests and differences in bargaining power between the two parties, a fair and efficient electricity trading pricing can be achieved. By employing a stochastic scenario method to obtain power fluctuations on a short timescale, and using the Alternating Direction Multiplier Method (ADMM) to achieve distributed solution, a real-time scheduling scheme with good convergence can be obtained while protecting privacy.

2. The stochastic optimization scheduling method for photovoltaic-storage-DC-flexible microgrids considering demand response delay as described in claim 1, characterized in that, The system architecture of the solar-storage-direct-current-flexible energy park microgrid includes: an energy operator (EP) and a load aggregator (LA); the EP adopts a multi-energy complementary energy supply mode, consisting of distributed generation units composed of photovoltaic modules, wind turbine generators, energy storage batteries (BS), and gas turbines (GT); the LA is divided into shiftable loads, transferable loads, loads that can be reduced, and conventional loads; among them, shiftable loads, transferable loads, and loads that can be reduced constitute adjustable flexible loads; the dispatch center issues flexible load adjustment instructions to the LA based on the forecast of the park's load and wind and solar power output data, and the dispatch center ensures power balance within the park by adjusting the power purchased and sold to the external grid, the power output of the GT, and the charging and discharging power of the BS.

3. The stochastic optimization scheduling method for photovoltaic-storage-DC-flexible microgrids considering demand response delay as described in claim 1, characterized in that, The scheduling model includes the EP model and the LA model; The EP model is constructed with the goal of minimizing the daily comprehensive revenue of EP; the daily comprehensive revenue of EP includes: electricity sales revenue, external grid electricity purchase cost, BS operating cost and GT generation cost; The EP comprehensive objective function includes: maxS L -C EG -C BS -C GT In the formula: S L For EP electricity sales revenue, C EG To reduce the cost of purchasing electricity from external power grids, C BS For BS operating costs, C GT For GT power generation costs, For EP electricity sales price, Let ω be the power output of EP at time t, and let s represent the scenario. s This represents the probability of scenario s occurring. Let represent the power purchased and sold to the external power grid at time t. The purchase and sale prices of electricity to the external power grid at time t are λ and t, respectively. BS The unit operating cost of BS, The charging and discharging power of BS at time t are respectively, and η is the charging and discharging power of BS c η d These represent the charge and discharge efficiencies of BS, Let be the output power of GT at time t. For GT start / stop status, a GT b GT and c GT These are the coefficients for the various factors related to the power generation cost of GT; The LA model constructs an LA comprehensive objective function with the goal of minimizing the daily comprehensive cost of LA; wherein the daily comprehensive cost of LA includes: the cost of purchasing electricity from EP, the cost of offsetting loads, the cost of transferring loads, and the cost of reducing loads; The LA synthesis objective function includes: minC L +C sh +C tr +C re In the formula: C L For LA electricity purchase cost, C sh For the total cost of compensating for the transferable load, C tr C is the total cost of compensating for transferable loads. re To compensate for the total cost of load reduction, For LA electricity purchase price, Let t be the total load demand of LA. The base load at time t, Let be the flexible load at time t. Let be the movable load at time t. Let be the transferable load at time t. Let c be the load that can be reduced at time t. sh P is the subsidy factor for the transferable load. t sh To optimize the power of the shiftable load at time t, C sh For the total cost of compensating for the transferable load, c tr P is the subsidy factor per unit power of transferable load. t tr To optimize the power of the shiftable load at time t, c re To reduce the subsidy price per unit of load power, P is the reduction factor. t re0 This represents the load power before reduction.

4. The stochastic optimization scheduling method for photovoltaic-storage-DC-flexible microgrids considering demand response delay as described in claim 3, characterized in that, The EP model includes the following constraints: (1) Power balance constraint: In the formula: Let represent the power purchased and sold to the external power grid at time t. Let be the charging and discharging power of BS at time t. Let EP be the power sold at time t. For the photovoltaic output at time t, Let t be the output of the wind turbine. Let GT be the output power at time t; (2) Generator power constraints: In the formula: Let t be the power purchased from the external power grid. To purchase the maximum amount of electricity from the grid, This represents the upper limit of photovoltaic power output. This is the upper limit of the wind turbine's output. Let be the output power of GT at time t. This represents the maximum output power of the GT. (3) BS operation constraints: In the formula: Let be the charging and discharging power of BS at time t. This represents the maximum discharge power of BS. This represents the maximum charging power of BS. and These are all Boolean variables, serving as indicators of BS charging and discharging. Let η be the capacity of the BS at time t. c η d These represent the charge and discharge efficiencies of BS, Divided into upper and lower limits of BS capacity, The BS capacity at the initial moment within a scheduling cycle. This represents the BS capacity at the end of a scheduling cycle.

5. The stochastic optimization scheduling method for photovoltaic-storage-DC-flexible microgrids considering demand response delay as described in claim 3, characterized in that, The LA model includes the following constraints: (1) Constraints for load transferable operation: In the formula: All are Boolean variables, where This indicates the operating state of the load that can be shifted at time t. To issue a translation command signal, The actual response start flag for the transferable load is given by τ, where τ is the delay time and P is the load factor. t sh To optimize the power of the shiftable load at time t, P t sh0 To optimize the power of the shiftable load at time t, M is a large positive integer. T represents the load transferable interval within the scheduling cycle. sh The translational duration; (2) Load transferability operation constraints: In the formula: All are Boolean variables, where This indicates the operating status of the load that can be transferred at time t. To issue a transfer instruction flag, P is the actual response start flag for transferable loads. t trx The power that the transferable load can adjust at time t. P represents the maximum power that the transferable load can adjust at time t. t tr To optimize the power of the transferable load at time t, P t tr0 To optimize the power of the transferable load at time t, The load transferable range within the scheduling cycle. M is the minimum transferable duration, where M is a large positive integer; (3) Load constraints can be reduced: In the formula: P t re0 The load power before reduction, For load reduction rate, Let M be the load reduction interval within the scheduling period, where M is a large positive integer. The reduction factor is... This is a signal to issue a translation command; τ is the delay time. This is a signal to issue a reduction order; (4) Power balance constraints for EP and LA participating in power trading: In the formula: Let EP be the power sold at time t. Let t be the total load demand of LA at time t.

6. The stochastic optimization scheduling method for photovoltaic-storage-DC-flexible microgrids considering demand response delay as described in claim 1, characterized in that, The method of obtaining short-timescale power fluctuations by using a random scenario method includes: the prediction errors of photovoltaic and load follow a normal distribution, and the Monte Carlo simulation method is used to generate scenarios; the obtained load and wind and solar power output scenarios are clustered and reduced using k-means. The statistical laws governing its power fluctuation characteristics include: In the formula: For the photovoltaic output at time t, Let t be the output of the wind turbine. The normal load output at time t, and These are the predicted values ​​for the conventional load, photovoltaic power output, and wind power output for hour t, respectively. and For the corresponding error distribution parameters, Let be the mean of the photovoltaic power output at time t, which follows a normal distribution. Let be the mean of the wind turbine output at time t, following a normal distribution. Let be the mean of the load under a normal distribution at time t. Let be the standard deviation of photovoltaic power output at time t. Let be the standard deviation of the wind turbine output at time t. Let be the standard deviation of the load at time t.

7. The stochastic optimization scheduling method for photovoltaic-storage-DC-flexible microgrids considering demand response delay as described in claim 1, characterized in that, The asymmetric electricity trading pricing model includes: In the formula: α EP α LA These are the bargaining power values ​​for EP and LA, respectively. L For EP electricity sales revenue, C EG To reduce the cost of purchasing electricity from external power grids, C BS For BS operating costs, C GT For GT power generation costs, For EP's revenue before participating in cooperative transactions, For the costs incurred by LA before participating in the cooperative transaction, C L For LA electricity purchase cost, C sh For the total cost of compensating for the transferable load, C tr C is the total cost of compensating for transferable loads. re To compensate for the total cost of load reduction, For EP electricity sales price, The purchase price of electricity in LA.

8. The stochastic optimization scheduling method for photovoltaic-storage-DC-flexible microgrids considering demand response delay as described in claim 7, characterized in that, The Nash negotiation model includes: In the formula: S m For the benefit of participant m after participating in the cooperation, The benefit of independent entity m before cooperation, i.e., the point at which negotiations break down.

9. The stochastic optimization scheduling method for photovoltaic-storage-DC-flexible microgrids considering demand response delay as described in claim 8, characterized in that, It also includes a solution method for the Nash negotiation model: decoupling the Nash negotiation model into a sub-problem of minimizing microgrid park costs and a sub-problem of electricity pricing; The objective function for minimizing the cost of the microgrid campus includes: minC=C EG +C BS +C GT +C sh +C tr +C re In the formula: C represents the cost of the microgrid park, C EG To reduce the cost of purchasing electricity from external power grids, C BS For BS operating costs, C GT For GT power generation costs, C sh For the total cost of compensating for the transferable load, C tr C is the total cost of compensating for transferable loads. re The total cost of compensating for load reduction; The constraints on electricity pricing include: (1) The transaction price of electricity shall not be higher than the purchase price of electricity from the external power grid, nor lower than the sales price of electricity from the external power grid. The constraints include: In the formula: For EP electricity sales price, For LA electricity purchase price, Let t be the electricity price sold to the external power grid. Let t be the purchase price of electricity from the external power grid at time t; (2) Ensure that the revenue of EP after participating in cooperative transactions is higher than its revenue when generating electricity at zero cost, and that the cost of LA after participating in cooperative transactions is lower than the cost of purchasing electricity from external grids in a single scenario. The constraints include: In the formula: S L For EP electricity sales revenue, C EG To reduce the cost of purchasing electricity from external power grids, C BS For BS operating costs, C GT For GT power generation costs, C sh For the total cost of compensating for the transferable load, C tr C is the total cost of compensating for transferable loads. re To compensate for the total cost of load reduction, For EP's revenue before participating in cooperative transactions, For the costs incurred by LA before participating in the cooperative transaction, C L For LA electricity purchase costs, For the photovoltaic output at time t, Let t be the output of the wind turbine. Let ω be the electricity price sold to the external power grid at time t. s This represents the probability of scenario s occurring. Let t be the total load demand of LA. Let t be the purchase price of electricity from the external power grid at time t; (3) The bargaining power constraints of EP and LA include: In the formula: α j To enhance EP's bargaining power with LA, The electrical energy that a certain entity can supply. The electrical energy that a particular entity can harvest. This represents the upper limit of electrical energy that can be supplied across all entities. This represents the upper limit of electrical energy that can be supplied across all entities. The power that a certain entity can provide. This refers to the power that a certain entity can absorb.

10. The stochastic optimization scheduling method for photovoltaic-storage-DC-flexible microgrids considering demand response delay as described in claim 1, characterized in that, The algorithm of the ADMM includes: (1) Set the maximum number of iterations and convergence accuracy, initialize the augmented Lagrange multipliers and the number of iterations, and calculate the bargaining power α between EP and LA. EP α LA ; (2) The combined objective functions of EP and LA are respectively introduced into the penalty coefficient and Lagrange multiplier to form the augmented Lagrange optimization objective function; that is: In the formula: S L For EP electricity sales revenue, C EG To reduce the cost of purchasing electricity from external power grids, C BS For BS operating costs, C GT For GT power generation costs, L is the revenue that EP receives before participating in the cooperative transaction. EP L LA Let EP and LA be the augmented Lagrangian optimization objective functions, respectively. For EP electricity sales price, For LA electricity purchase price, For the costs incurred by LA before participating in the cooperative transaction, C L For LA electricity purchase cost, C sh C represents the total cost of compensating for the transferable load. tr C is the total cost of compensating for transferable loads. re To compensate for the total cost of load reduction, Lagrange multipliers for EP and LA respectively, ρ EP ρ LA The penalty coefficients for EP and LA are α and α, respectively. EP α LA These are the bargaining power values ​​for EP and LA, respectively. (3) Update the electricity trading price decisions of EP and LA; EP and LA each calculate their electricity trading price strategies and exchange quotation information with each other. The update method of their electricity price strategies is shown in the following formula: In the formula: For EP electricity sales price, Let k be the electricity purchase price in LA, and k be the iteration number. These are the Lagrange multipliers for EP and LA, respectively; (4) Lagrange multiplier update; The Lagrange multiplier update method is shown in the following formula: In the formula: Let EP be the power sold at time t. Let ρ be the total load demand of LA at time t; EP ρ LA These are the penalty coefficients for EP and LA, respectively; (5) The number of update iterations is k = k + 1; (6) Convergence condition; During the iteration process, the algorithm is considered to have converged when both the original residual and the dual parameter residual are less than their corresponding convergence accuracy, and the value of k is less than its maximum value. The specific constraints are shown in the following formula: In the formula: ε1 is the convergence accuracy of the original residual; ε2 is the convergence accuracy of the dual residual; (7) Terminate the iteration; if the above convergence condition is met, terminate the iteration; otherwise, return to step (2) until the iteration termination convergence condition is met or the maximum number of iterations k is reached. max Then stop.

Images