Multi-virtual power plant cooperative operation method considering differentiated flexible adjustment capability

Abstract

A multi-virtual power plant cooperative operation method considering differentiated flexible adjustment capability comprises the following steps: step 1, constructing a virtual power plant adjustment capability model, and calculating the adjustment capability of each virtual power plant; 2, constructing a virtual power plant alliance body according to a calculation result of the adjustment capability of the virtual power plant; 3, constructing a target function of the virtual power plant alliance body; 4, setting constraint conditions of the virtual power plant alliance body; and 5, under the constraint condition, linearization and optimization solution are carried out on the target function. Through the surplus and deficiency mutual aid of the flexible adjustment capability among multiple virtual power plants, the deviation punishment possibly faced by a single virtual power plant is reduced, the total operation cost is reduced, and the stability and the overall benefit of a power system are improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of power system optimization and dispatching, and specifically relates to a method for collaborative operation of multiple virtual power plants taking into account differentiated flexible adjustment capabilities. Background Art

[0002] With the large-scale integration of renewable energy, power grids face numerous new challenges. Energy sources such as solar and wind power are subject to natural constraints, resulting in significant randomness and volatility in power generation. This significantly increases uncertainty in grid operation, placing pressure on power system operational security and system balance. Against this backdrop, virtual power plants, as a key component of the emerging power system, demonstrate significant application value. Leveraging advanced information and communication technologies, virtual power plants aggregate and efficiently utilize a wide range of resources, including distributed power sources, distributed energy storage, and user-side adjustable loads, creating an intelligent management platform. With their fast and flexible regulation capabilities and excellent economic efficiency, virtual power plants not only enable efficient participation in the electricity market but also play a vital role in grid operation. Through intelligent resource aggregation and optimized scheduling, virtual power plants help promote the local consumption of renewable energy, improve the safety and economic efficiency of grid operation, and provide strong support for building a more reliable, clean, and efficient power system.

[0003] On the other hand, the uncertainty of renewable energy sources increases system uncertainty, leading to potential discrepancies between actual and reported output during real-time operation and resulting in supply-demand imbalances, posing a threat to grid security. To prevent this phenomenon on a large scale, the grid has implemented deviation penalties, charging additional costs to virtual power plants that exhibit excessive deviations. Therefore, to avoid unnecessary economic losses, virtual power plants with varying flexibility within the same virtual power plant alliance can be re-matched to their respective demands, achieving a balanced balance of surplus and deficit in their regulation capabilities. This fully leverages the flexibility of virtual power plants and provides reliable support for the safe and economical operation of the power system.

[0004] However, existing collaborative optimization strategies for virtual power plants typically consider independent operation and only single interactions with the external power grid. However, with the gradual establishment of virtual power plants of varying types and regions, the scale of heterogeneous distributed resources such as wind and solar power within geographically diverse virtual power plants varies, leading to random output. Furthermore, user electricity consumption behaviors and patterns within different virtual power plants are diverse and uncertain. Due to the lack of information exchange and energy interaction between multiple virtual power plants, power deviations are highly likely to occur when a single virtual power plant participates in grid scheduling, leading to the risk of revenue loss. Summary of the Invention

[0005] The purpose of the present invention is to address the deficiencies of the above-mentioned prior art, thereby providing a method for the coordinated operation of multiple virtual power plants taking into account differentiated flexible adjustment capabilities. By making full use of the surplus and shortage of flexible adjustment capabilities among multiple virtual power plants, the deviation penalty that a single virtual power plant may face is reduced, the overall operating cost is reduced, and the stability and overall benefits of the power system are improved.

[0006] A method for collaborative operation of multiple virtual power plants considering differentiated flexible regulation capabilities includes the following steps:

[0007] Step 1: Build a virtual power plant regulation capacity model and calculate the regulation capacity of each virtual power plant;

[0008] Step 2: Based on the calculation results of the virtual power plant regulation capacity, a virtual power plant alliance is established;

[0009] Step 3: Construct the objective function of the virtual power plant alliance;

[0010] Step 4: Set the constraints of the virtual power plant alliance;

[0011] Step 5: Under the constraints, linearize the objective function and optimize the solution.

[0012] The virtual power plant regulation capability model in step 1 is:

[0013] ΔP=P adj -P bid

[0014] Among them, ΔP is the regulation capacity of the virtual power plant, P adj is the actual adjustable load of the virtual power plant, P bid It is the winning load of the virtual power plant after reporting the quantity in the day-ahead electricity energy market. If ΔP≥0, it means that the virtual power plant has the ability to adjust upward; if ΔP≤0, it means that it has the ability to adjust downward.

[0015] Step 2 specifically includes the following steps:

[0016] First, based on the calculation results of the regulation capacity of each virtual power plant in step 1, the power plants with upward regulation capacity and the power plants with downward regulation capacity are distinguished;

[0017] Subsequently, virtual power plants are combined according to the direction of regulation capabilities to build a virtual power plant alliance including regulation capacity providers and recipients.

[0018] Step three specifically includes:

[0019] Each virtual power plant in the virtual power plant alliance has different regulation capabilities. Assuming that the regulation capacity of the virtual power plant is ΔP, the actual regulation capacity of the virtual power plants in the virtual power plant alliance can be grouped into sufficient regulation capacity and upward regulation capacity ΔPm and the downward adjustment capacity ΔP due to insufficient adjustment capacity n ,Right now:

[0020] ΔP m =max(ΔP,0),ΔP n =max(-ΔP,0)

[0021] Assume that the set of virtual power plants with sufficient regulation capacity within the alliance is N b , the set of virtual power plants with insufficient regulation capacity is N s , the provision and acceptance of regulatory capabilities meet the following constraints:

[0022]

[0023] Introducing adjustment costs to optimize the allocation of adjustment capacity:

[0024] The regulation cost is proportional to the amount of regulation capacity used. Assume that the regulation cost of a virtual power plant with insufficient regulation capacity is a m , the regulation cost of a virtual power plant with sufficient regulation capacity is a n , then the total adjustment cost is:

[0025]

[0026] In order to reduce the deviation of grid load, assuming that each virtual power plant has a certain load deviation, the deviation penalty is:

[0027] μ|P devi -ΔP agg |

[0028] Considering the regulation cost, deviation penalty cost and real-time market purchase regulation capacity cost of each virtual power plant in the alliance, the objective function of constructing the virtual power plant alliance is:

[0029]

[0030] where a m is the regulation cost of the virtual power plant with insufficient regulation capacity; a n is the regulation cost of a virtual power plant with sufficient regulation capacity; P devi N is the deviation between the actual operating load of the virtual power plant alliance and the net load reported on the day before; s is the number of virtual power plants with insufficient regulation capacity; N b is the number of virtual power plants with sufficient regulation capacity; μ is the unit deviation penalty cost; ΔP≤0 is the regulation capacity required by virtual power plants with insufficient regulation capacity; ΔP n ≥0 is the regulation capacity that a virtual power plant with sufficient regulation capacity can provide; ΔP aggThe regulation capacity purchased by the virtual power plant alliance in the real-time market.

[0031] In step 4, the constraints include:

[0032] 1) Regulation capacity balance constraint: The regulation capacity of the alliance in the real-time market needs to achieve a balance between supply and demand, that is, the regulation capacity purchased by the virtual power plant alliance is equivalent to the net value of the regulation capacity of each virtual power plant after the surplus and shortage of each virtual power plant; specifically, the upward regulation capacity ΔP provided by the virtual power plant with sufficient regulation capacity is m and the downward regulation capacity ΔP required by the virtual power plant with insufficient regulation capacity n The regulatory capacity ΔP that must be purchased with the alliance in the real-time market agg Phase equilibrium:

[0033]

[0034] 2) Upper and lower limit constraints:

[0035] P m min ≤P m +ΔP m ≤P m max (3)

[0036] P n min ≤P n +ΔP n ≤P n max (4)

[0037] -P devi ≤ΔP agg ≤P devi (5)

[0038] Among them, P m min is the lower limit of electricity consumption of virtual power plants with insufficient regulation capacity; P m max P is the upper limit of electricity consumption of virtual power plants with insufficient regulation capacity; n min is the lower limit of electricity consumption of a virtual power plant with sufficient regulation capacity; P n max is the lower limit of electricity consumption of a virtual power plant with sufficient regulation capacity; P m The amount reported by the virtual power plant with insufficient regulation capacity on the previous day; P n The capacity reported by virtual power plants with sufficient regulation capacity on the previous day;

[0039] 3) Power flow constraints:

[0040] P line =H -1 P inj (6)

[0041] V=H -T D r P line +H-T D x Q line -V0H -T H0 (7)

[0042] V min ≤V≤V max (8)

[0043]

[0044] Where: H is the node-branch correlation matrix of the distribution network; P line is the line current, P inj Inject power into the node, M i is the distribution network node-regulation capacity insufficient party correlation matrix, N i is the distribution network node-regulation capacity sufficient party association matrix, I is the total number of nodes in the distribution network; V is the node voltage, V=[V1,...,V I ] T ; V min is the lower limit of node voltage; V max is the upper limit of node voltage; is the lower limit of the line flow; is the upper limit of the line flow; D r is the node resistance matrix; D x is the node reactance matrix; H0 is the first row of the system node-branch correlation matrix; H -T is the conjugate transpose of the node-branch correlation matrix of the distribution network, Q line is the line reactive power flow; V0 is the equilibrium node voltage.

[0045] Step 5 specifically includes:

[0046] The auxiliary variable s is introduced to linearize the objective function. The linearized objective function is as follows:

[0047]

[0048] st(2)-(9)

[0049] -s≤-(P devi -ΔP dso )(11)

[0050] -s≤P devi -ΔP dso (12)

[0051] Where: P devi is the error between the actual output of the virtual power plant and the day-ahead clearing result during the real-time operation of the distribution network, ΔP dsois the net deviation of DSO after adjustment by the virtual power plant;

[0052] Based on the linearization processing results and the setting of constraint conditions, the objective function is optimized and solved using MATLAB combined with the Gurobi solver to obtain the regulation capacity allocation plan for each virtual power plant and the regulation capacity value that the virtual power plant alliance needs to purchase in the real-time market.

[0053] The present invention takes into account the differences in adjustable capabilities between different virtual power plants and the grid regulation needs, and with the goal of minimizing the regulation cost and deviation penalty cost, redistributes the regulation capabilities of virtual power plants to achieve a balance between the regulation capabilities of different virtual power plants. The present invention can better mobilize the flexibility of virtual power plants, achieve coordinated scheduling between the supply side and the demand side, and thus enhance the flexibility and economy of the power system. In addition, by balancing the flexible regulation capabilities between multiple virtual power plants, the present invention reduces the deviation penalties that a single virtual power plant may face, reduces overall operating costs, and improves the stability and overall benefits of the power system. DETAILED DESCRIPTION

[0054] Step 1: Build a regulation capacity model for each virtual power plant and calculate the regulation capacity of each virtual power plant

[0055] First, the regulation capacity of each virtual power plant is calculated, which is expressed as the difference between the actual adjustable load and the reported load:

[0056] ΔP=P adj -P bid

[0057] Among them, ΔP is the regulation capacity of the virtual power plant, P adj is the actual adjustable load of the virtual power plant, P bid =ΔP is the winning bid load of the virtual power plant after reporting its load in the day-ahead electricity market. If ΔP ≥ 0, it indicates that the virtual power plant has the ability to adjust upward; if ΔP ≤ 0, it indicates that it has the ability to adjust downward.

[0058] Step 2: Build a virtual power plant alliance based on the calculation results of the virtual power plant regulation capacity

[0059] First, based on the calculation results of the regulation capacity of each virtual power plant in step 1, the power plants with upward regulation capacity and the power plants with downward regulation capacity are distinguished;

[0060] Subsequently, virtual power plants are combined according to regulation capabilities to build a virtual power plant alliance that includes regulation capacity providers and recipients;

[0061] Next, based on the regulation capacity value and corresponding regulation cost of each virtual power plant, a matching relationship between regulation capacity supply and demand is established within the alliance, and power plants with lower regulation costs are given priority to undertake regulation tasks, thereby reducing overall regulation costs.

[0062] Finally, a distribution plan for the internal regulation capacity of the alliance is formed, and the regulation output or demand value of each virtual power plant is clarified, providing input for the subsequent objective function optimization, achieving the mutual assistance of surplus and shortage of regulation capacity and minimizing the deviation cost.

[0063] The present invention proposes a collaborative operation strategy based on a virtual power plant alliance, which aims to give full play to the integrated complementary advantages of heterogeneous resources through the alliance of multiple virtual power plants to form a virtual power plant alliance. In this alliance, each virtual power plant has different regulation capabilities and regulation needs. Specifically, virtual power plants with sufficient regulation capabilities (i.e., virtual power plants with upward regulation capabilities) serve as providers of flexible regulation capabilities, while virtual power plants with insufficient regulation capabilities (i.e., virtual power plants with downward regulation capabilities) serve as recipients of flexible regulation capabilities. Through internal coordination within the alliance, virtual power plants with sufficient regulation capabilities provide regulation capabilities to virtual power plants with insufficient regulation capabilities, and at the same time give priority to virtual power plants with lower regulation costs to replace virtual power plants with higher regulation costs, thereby achieving mutual assistance in regulation capabilities. This mechanism ultimately minimizes the total regulation cost and regulation deviation of the alliance.

[0064] The primary purpose of this coordinated operation strategy is to reduce the deviation penalty costs incurred by virtual power plants in the real-time market. Therefore, the strategy's execution time falls between the day-ahead and real-time markets, effectively realigning regulatory capacity before the real-time market, over a short period of time. Through this strategy, the virtual power plant alliance can optimize the allocation of regulatory capacity before the real-time market, reducing deviation penalties caused by fluctuations in renewable energy output and improving the economic efficiency and stability of the power system.

[0065] Step 3: Construct the objective function of the virtual power plant alliance

[0066] First, multiple virtual power plants form a virtual power plant alliance through alliance, and each virtual power plant has different regulation capabilities. Assuming that the regulation capability of a virtual power plant is ΔP, its actual regulation capability can be divided into upward regulation capability and downward regulation capability. That is:

[0067] ΔP m =max(ΔP,0),ΔP n =max(-ΔP,0)

[0068] Among them, virtual power plants with sufficient regulation capacity are those with ΔP≥0, and virtual power plants with insufficient regulation capacity are those with ΔP≤0.

[0069] Within the alliance, virtual power plants with sufficient regulation capacity provide regulation capacity to virtual power plants with insufficient regulation capacity, forming a "mutual assistance of regulation capacity surplus and shortage" mechanism. To this end, let the set of virtual power plants with sufficient regulation capacity within the alliance be N b , the set of virtual power plants with insufficient regulation capacity is N s Then, the provision and acceptance of regulatory capabilities satisfy the following constraints:

[0070]

[0071] This constraint indicates that the total regulating capacity of virtual power plants with sufficient regulating capacity should be equal to the total regulating demand of virtual power plants with insufficient regulating capacity.

[0072] Next, we introduce regulation cost to optimize the allocation of regulation capacity. The regulation cost is proportional to the amount of regulation capacity used. Let the regulation cost of each virtual power plant be m , the regulation cost of a virtual power plant with sufficient regulation capacity is a n , then the total adjustment cost is:

[0073]

[0074] At the same time, in order to reduce the deviation of the power grid load, the power grid has certain requirements for the regulation capacity of each virtual power plant, which needs to be met through regulation capacity. The power grid usually penalizes this deviation to encourage participants to minimize the deviation and ensure the stable operation of the power grid. Assuming that each virtual power plant has a certain load deviation, the deviation penalty is:

[0075] μ|P devi -ΔP agg |

[0076] Finally, the objective function considering the regulation cost of each virtual power plant in the alliance, the deviation penalty cost, and the real-time market purchase regulation capacity cost is obtained:

[0077]

[0078] where a m is the regulation cost of the virtual power plant with insufficient regulation capacity; a n is the regulation cost of a virtual power plant with sufficient regulation capacity; P devi N is the deviation between the actual operating load of the virtual power plant alliance and the net load reported on the day before; s is the number of virtual power plants with insufficient regulation capacity; N b is the number of virtual power plants with sufficient regulation capacity; μ is the unit deviation penalty cost; ΔP≤0 is the regulation capacity required by virtual power plants with insufficient regulation capacity; ΔP n≥0 is the regulation capacity that a virtual power plant with sufficient regulation capacity can provide; ΔP agg The regulation capacity purchased by the virtual power plant alliance in the real-time market.

[0079] Step 4: Set constraints for the virtual power plant alliance

[0080] 1) Regulation capacity balance constraint: The regulation capacity of the alliance in the real-time market needs to achieve a balance between supply and demand, that is, the regulation capacity purchased by the virtual power plant alliance is equivalent to the net value after the surplus and shortage of the regulation capacity of each virtual power plant. Specifically, the upward regulation capacity ΔP provided by the virtual power plant with sufficient regulation capacity is m and the downward regulation capacity ΔP required by the virtual power plant with insufficient regulation capacity n The regulatory capacity ΔP that must be purchased with the alliance in the real-time market agg Phase equilibrium.

[0081]

[0082] This constraint ensures the balance between supply and demand of regulation capacity within the alliance and avoids power system instability caused by unreasonable allocation of regulation capacity.

[0083] 2) Upper and lower limit constraints: The total electricity consumption of virtual power plants with sufficient and insufficient regulation capabilities is subject to upper and lower limit constraints, and the net load deviation of the virtual power plant alliance after surplus and shortage assistance is also subject to upper and lower limit constraints.

[0084] P m min ≤P m +ΔP m ≤P m max (3)

[0085] P n min ≤P n +ΔP n ≤P n max (4)

[0086] -P devi ≤ΔP agg ≤P devi (5)

[0087] Among them, P m min is the lower limit of electricity consumption of virtual power plants with insufficient regulation capacity; P m max P is the upper limit of electricity consumption of virtual power plants with insufficient regulation capacity; n min is the lower limit of electricity consumption of a virtual power plant with sufficient regulation capacity; P n max is the lower limit of electricity consumption of a virtual power plant with sufficient regulation capacity; P m The amount reported by the virtual power plant with insufficient regulation capacity on the previous day; P n The reported capacity was recently submitted by a virtual power plant with sufficient regulation capacity.

[0088] This constraint ensures that the electricity demand of each virtual power plant is within its technical capabilities, avoiding equipment damage or system instability caused by over-regulation.

[0089] 3) Power flow constraint: To ensure the normal operation of the power system, the power flow P line Within the safe operating range.

[0090] P line =H -1 P inj (6)

[0091] V=H -T D r P line +H -T D x Q line -V0H -T H0 (7)

[0092] V min ≤V≤V max (8)

[0093]

[0094] Where H is the node-branch correlation matrix of the distribution network; P line is the line current, P inj Inject power into the node, M i is the distribution network node-regulation capacity insufficient party correlation matrix, N i is the distribution network node-regulation capacity sufficient party association matrix, I is the total number of nodes in the distribution network; V is the node voltage, V=[V1,...,V I ] T ; V min is the lower limit of node voltage; V max is the upper limit of node voltage; is the lower limit of the line flow; is the upper limit of the line flow; D r is the node resistance matrix; D x is the node reactance matrix; H is the system node-branch correlation matrix; H0 is the first row of the matrix. The present invention assumes that the reactive injection power remains unchanged.

[0095] This constraint ensures that the power flow of the power system does not exceed the limit, avoiding line overload or system collapse caused by excessive power flow.

[0096] Step 5: Linearize and optimize the objective function under the constraints

[0097] Since the objective function contains an absolute value function, it is more complicated to solve it directly. To simplify the calculation, we introduce an auxiliary variable s to linearize the objective function. The linearized objective function is as follows:

[0098]

[0099] st(2)-(9)

[0100] -s≤-(P devi -ΔP dso ) (11)

[0101] -s≤P devi -ΔP dso (12)

[0102] Through the linearization process and constraint setting described above, the objective function is optimized and solved using MATLAB combined with the Gurobi solver. The solution provides a plan for allocating the regulation capacity of each virtual power plant, as well as the regulation capacity that the virtual power plant alliance needs to purchase in the real-time market.

[0103] Case Analysis

[0104] This section uses data from the IEEE-33 node example and makes the following assumptions:

[0105] 1) Assume that a total of 32 virtual power plants form a consortium, and each node except the reference node is connected to a virtual power plant;

[0106] 2) Assuming the first node is the reference node, virtual power plants with insufficient regulation capacity are connected to nodes 2-17, and virtual power plants with sufficient regulation capacity are connected to nodes 18-33;

[0107] 3) Assume that the original injected power of each node in the IEEE-33 node system is the day-ahead reported capacity of each virtual power plant;

[0108] 4) Assume that the deviation between the net load of the virtual power plant alliance and the day-ahead load in actual operation is 30kW, and the unit penalty cost is 1 yuan / kW;

[0109] 5) Assume that the upper and lower limits of electricity demand of each virtual power plant are as shown in Table 1.

[0110] Table 1 Virtual power plant example settings

[0111]

[0112]

[0113] According to the above, by using formulas (2)-(12), MATLAB combined with Gurobi is used to solve the surplus and shortage of the regulation capacity of each virtual power plant, as shown in Table 2.

[0114] Table 2: Adjustable capacity trading surplus and shortage mutual assistance

[0115]

[0116] A negative objective function indicates that market transactions will bring certain benefits to the virtual power plant alliance;

[0117] The regulation capabilities of the 1st to 16th virtual power plants are positive, indicating that these virtual power plants provide corresponding flexible regulation capabilities;

[0118] The winning bid values ​​for the regulation capabilities of the 17th to 32nd virtual power plants are negative, indicating that these virtual power plants require corresponding flexible regulation capabilities.

[0119] The purchase adjustment capacity value of the virtual power plant alliance indicates that the virtual power plant alliance still needs to purchase flexible adjustment capacity in the real-time market to make up for the adjustment deviation value that still exists after surplus and shortage mutual assistance.

[0120] By comparing the regulation capacity value purchased by the virtual power plant alliance in the transaction results and the net load deviation value of the virtual power plant alliance, it can be seen that the virtual power plant alliance effectively compensated for the net load deviation after adopting a collaborative operation strategy, avoided the deviation assessment costs caused by the deviation of new energy output, and reduced operating costs.

[0121] The present invention proposes a collaborative operation strategy for multiple virtual power plants that takes into account differentiated flexible regulation capabilities. This strategy distinguishes the regulation capabilities of different virtual power plants, matches virtual power plants with sufficient and insufficient regulation capabilities, and achieves a balance between regulation capabilities, thereby compensating for deviations in renewable energy output. The present invention proposes a clearing model for trading the adjustable capacity of differentiated demand-side flexible resources. Based on the regulation capabilities of each virtual power plant in the virtual power plant alliance, deviations in renewable energy output are compensated by internally adjusting regulation capabilities and demand allocation. This reduces the risk of deviation assessments faced by a single virtual power plant, reduces the total regulation cost of multiple virtual power plants, and improves the safety and economy of the power system.

Claims

1. A method for collaborative operation of multiple virtual power plants taking into account differentiated flexible adjustment capabilities, characterized by: The following steps are involved: Step 1: Build a virtual power plant regulation capacity model and calculate the regulation capacity of each virtual power plant; Step 2: Based on the calculation results of the virtual power plant regulation capacity, a virtual power plant alliance is established; Step 3: Construct the objective function of the virtual power plant alliance; Step 4: Set the constraints of the virtual power plant alliance; Step 5: Under the constraints, linearize the objective function and optimize the solution.

2. The method for collaborative operation of multiple virtual power plants considering differentiated flexible adjustment capabilities according to claim 1, characterized in that: The regulation capability model of the virtual power plant in step 1 is: ΔP=P adj -P bid Among them, ΔP is the regulation capacity of the virtual power plant, P adj is the actual adjustable load of the virtual power plant, P bid It is the winning load of the virtual power plant after reporting the quantity in the day-ahead electricity energy market. If ΔP≥0, it means that the virtual power plant has the ability to adjust upward; if ΔP≤0, it means that it has the ability to adjust downward.

3. The method for collaborative operation of multiple virtual power plants considering differentiated flexible adjustment capabilities according to claim 1, characterized in that: Step 2 specifically includes the following steps: First, based on the calculation results of the regulation capacity of each virtual power plant in step 1, the power plants with upward regulation capacity and the power plants with downward regulation capacity are distinguished; Subsequently, virtual power plants are combined according to the direction of regulation capabilities to build a virtual power plant alliance including regulation capacity providers and recipients.

4. The method for collaborative operation of multiple virtual power plants considering differentiated flexible adjustment capabilities according to claim 1, characterized in that: Step three specifically includes: Each virtual power plant in the virtual power plant alliance has different regulation capabilities. Assuming that the regulation capacity of the virtual power plant is ΔP, the actual regulation capacity of the virtual power plants in the virtual power plant alliance can be grouped into sufficient regulation capacity and upward regulation capacity ΔP m and the downward adjustment capacity ΔP due to insufficient adjustment capacity n ,Right now: ΔP m =max(ΔP,0),ΔP n =max(-ΔP,0) Assume that the set of virtual power plants with sufficient regulation capacity within the alliance is N b , the set of virtual power plants with insufficient regulation capacity is N s , the provision and acceptance of regulatory capabilities meet the following constraints: Introducing adjustment costs to optimize the allocation of adjustment capacity: The regulation cost is proportional to the amount of regulation capacity used. Assume that the regulation cost of a virtual power plant with insufficient regulation capacity is a m , the regulation cost of a virtual power plant with sufficient regulation capacity is a n , then the total adjustment cost is: In order to reduce the deviation of grid load, assuming that each virtual power plant has a certain load deviation, the deviation penalty is: m|P devi -ΔP agg | Considering the regulation cost, deviation penalty cost and real-time market purchase regulation capacity cost of each virtual power plant in the alliance, the objective function of constructing the virtual power plant alliance is: where a m is the regulation cost of the virtual power plant with insufficient regulation capacity; a n is the regulation cost of a virtual power plant with sufficient regulation capacity; P devi N is the deviation between the actual operating load of the virtual power plant alliance and the net load reported on the day before; s is the number of virtual power plants with insufficient regulation capacity; N b is the number of virtual power plants with sufficient regulation capacity; μ is the unit deviation penalty cost; ΔP≤0 is the regulation capacity required by virtual power plants with insufficient regulation capacity; ΔP n ≥0 is the regulation capacity that a virtual power plant with sufficient regulation capacity can provide; ΔP agg The regulation capacity purchased by the virtual power plant alliance in the real-time market.

5. The method for collaborative operation of multiple virtual power plants considering differentiated flexible adjustment capabilities according to claim 1, characterized in that: In step 4, the constraints include: 1) Regulation capacity balance constraint: The alliance’s regulation capacity in the real-time market needs to achieve a balance between supply and demand. That is, the regulation capacity purchased by the virtual power plant alliance is equivalent to the net value of the regulation capacity after the surplus and shortage of each virtual power plant are compensated. Specifically, the upward regulation capacity ΔP provided by the virtual power plant with sufficient regulation capacity m and the downward regulation capacity ΔP required by the virtual power plant with insufficient regulation capacity n The regulatory capacity ΔP that must be purchased with the alliance in the real-time market agg Phase equilibrium: 2) Upper and lower limit constraints: P mmin ≤P m +ΔP m ≤P mmax (3) P nmin ≤P n +ΔP n ≤P nmax (4) -P devi ≤ΔP agg ≤P devi (5) Among them, P mmin is the lower limit of electricity consumption of virtual power plants with insufficient regulation capacity; P mmax P is the upper limit of electricity consumption of virtual power plants with insufficient regulation capacity; nmin is the lower limit of electricity consumption of a virtual power plant with sufficient regulation capacity; P nmax is the lower limit of electricity consumption of a virtual power plant with sufficient regulation capacity; P m The amount reported by the virtual power plant with insufficient regulation capacity on the previous day; n The daily quota submitted by virtual power plants with sufficient regulation capacity; 3) Power flow constraints: P line =H -1 P inj (6) V=H -T D r P line +H -T D x Q line -V0H -T H0 (7) In min ≤V≤V max (8) Where: H is the node-branch correlation matrix of the distribution network; P line is the line current, P inj Inject power into the node, M i is the distribution network node-regulation capacity insufficient party correlation matrix, N i is the distribution network node-regulation capacity sufficient party association matrix, I is the total number of nodes in the distribution network; V is the node voltage, V=[V1,...,V I ] T ; V min is the lower limit of node voltage; V max is the upper limit of node voltage; is the lower limit of the line flow; is the upper limit of the line flow; D r is the node resistance matrix; D x is the node reactance matrix; H0 is the first row of the system node-branch correlation matrix; H -T is the conjugate transpose of the node-branch correlation matrix of the distribution network, Q line is the line reactive power flow; V0 is the equilibrium node voltage.

6. The method for collaborative operation of multiple virtual power plants considering differentiated flexible adjustment capabilities according to claim 1, characterized in that: Step 5 specifically includes: The auxiliary variable s is introduced to linearize the objective function. The linearized objective function is as follows: st(2)-(9) -s≤-(P devi -ΔP dso ) (11) -s≤P devi -ΔP dso (12) Where: P devi is the error between the actual output of the virtual power plant and the day-ahead clearing result during the real-time operation of the distribution network, ΔP dso is the net deviation of DSO after adjustment by the virtual power plant; Based on the linearization processing results and the setting of constraint conditions, the objective function is optimized and solved using MATLAB combined with the Gurobi solver to obtain the regulation capacity allocation plan for each virtual power plant and the regulation capacity value that the virtual power plant alliance needs to purchase in the real-time market.

Images