A virtual power plant and distributed resource two-sided matching method based on gale-shapley algorithm

Abstract

The application discloses a kind of virtual power plant and distributed resource two-sided matching method based on Gale-Shapley algorithm, belong to electric power market technical field.Through the construction of electric power spot market environment, virtual power plant participates in market declaration optimization model, the declaration curve of wind, light, water storage load integration virtual power plant is optimized;Two-way selection decision model of virtual power plant and distributed resource is constructed, two-sided matching step based on Gale-Shapley algorithm is carried out, the two-sided optimization matching of virtual power plant and distributed resource is realized, so as to ensure the stability of configuration, guarantee that the matched two sides will not be disturbed by each other, realize the effective utilization rate of electric power resource, effectively solve the problem of optimal configuration of electric power resource, with good economic benefit and practical application value.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of electric power, more particularly, to a virtual power plant and distributed resource two-sided matching method based on Gale-Shapley algorithm. BACKGROUND

[0002] With a large number of renewable energy power generation, energy storage, controllable load and other distributed resources accessing the power grid, it is crucial to fully develop their regulation potential for the construction of a new power system. Under this background, virtual power plants aggregate multiple types of distributed resources through advanced communication, regulation and metering technologies to achieve stable and reliable overall output, which is a flexible management method for distributed resources. In real life, virtual power plants and distributed resources are both rational subjects, and both parties have the right to choose and refuse, but there is currently a lack of two-sided matching method for virtual power plants and distributed resources.

[0003] Virtual power plants differ from traditional power plant resource allocation, and the output of distributed resources varies, so virtual power plants are prone to cause unreasonable grid matching, affecting the effective utilization rate of power resources. The rationality of the matching between virtual power plants and distributed resources has become a problem that should be solved in the power system. SUMMARY

[0004] In view of the deficiencies in the prior art, the present application ensures the stability of the configuration by comparing different matching methods, so as to ensure that the two parties of the matching will not be disturbed by each other, thereby realizing the effective utilization rate of power resources.

[0005] The present application proposes a technical solution for the problem of rational matching of power resources, and the virtual power plant and distributed resource two-sided matching method based on Gale-Shapley algorithm helps to improve resource utilization, and has economic and practical application value.

[0006] The technical solution adopted by the present application to solve its technical problems is: a virtual power plant and distributed resource two-sided matching method based on Gale-Shapley algorithm, comprising the following steps: step 1: constructing a virtual power plant participating in market declaration optimization model under the environment of electric power spot market, optimizing the declaration curve of wind, light, water and load integrated virtual power plant; step 2: constructing a two-way selection decision model of virtual power plant and distributed resource, carrying out two-sided matching step based on Gale-Shapley algorithm, and realizing two-sided matching of virtual power plant and distributed resource.

[0007] In the technical solution, further, the step 1 constructs a virtual power plant participating in market declaration optimization model under the power spot market environment, and the declaration curve of the wind-solar-water-storage integrated virtual power plant is optimized. The virtual power plant operator combines distributed power generation, distributed energy storage and controllable load organically, and acts as an agent for participating in the power market. The virtual power plant declares the output curve in the day-ahead market, and participates in the day-ahead market as a price accepter. In real-time operation, due to the strong uncertainty of wind power and photovoltaic output, the deviation between the actual output of the virtual power plant and the cleared power will face examination, so when deciding the declaration curve in the day-ahead market, the virtual power plant needs to fully consider the output fluctuation of various renewable energy generation in addition to the uncertainty of market price.

[0008] The set S represents the uncertainty scenario set of wind, light, water output and market price, and the target function expression of the virtual power plant operating income in the power spot market is:

[0009] max R M -C IL -C PU

[0010]

[0011]

[0012]

[0013] In the formula, T represents the total period; π s represents the probability of scenario s; R M , C IL and C PU respectively represent the power spot market income, interruptible load adjustment cost and deviation power examination cost; represents the market price of period t under scenario s; λ IL represents the interruptible load adjustment cost; k pos and k neg respectively represent the positive / negative deviation power examination cost coefficient; P t M represents the output of the virtual power plant at time t declared in the day-ahead market; represents the interruptible load reduction output in period t under scenario s; and respectively represent the over / under generation in period t under scenario s.

[0014] Further, the constraint conditions include the power balance constraint:

[0015]

[0016]

[0017] In the formula: and respectively represent the output of virtual power plants, wind power, photovoltaic and hydropower in period t under scenario s; and respectively represent the charge / discharge power of energy storage; represents the load demand in period t under scenario s when interruptible load is not called.

[0018] Further, the constraint conditions include hydropower constraints:

[0019]

[0020]

[0021] V s,t+1 = V s,t + ΔV s,t

[0022]

[0023] V min ≤ V s,t ≤ V max

[0024] In the formula: P HV,min and P HV,max respectively represent the minimum / maximum output of hydropower; and respectively represent the inflow, outflow and abandoned water of reservoir-type hydropower stations in period t under scenario s; V s,t and ΔV s,t respectively represent the reservoir storage and water volume change in period t under scenario s; V min and V max respectively represent the minimum and maximum reservoir storage; η HV represents the hydropower generation efficiency.

[0025] Further, the constraint conditions include energy storage constraints:

[0026]

[0027]

[0028]

[0029]

[0030] In the formula: P ESS and E ESSrespectively represent the maximum charge-discharge power and the energy storage capacity of the energy storage; respectively represent the state of charge of the energy storage in the time period t under the scenario s; and ch and dis respectively represent the charge-discharge efficiency of the energy storage.

[0031] Further, the constraint conditions include an interruptible load constraint:

[0032]

[0033] wherein P IL,max represents the interruptible load capacity.

[0034] Further, the two-way selection decision model is a combination scheme of the distributed resource m aiming to maximize the maximum benefit generated after joining the virtual power plant n, and the matrix thereof is:

[0035]

[0036] wherein x n,k,m is a 0-1 variable, x n,k,m = 1 indicates that the distributed resource m is selected as a cooperative object in the kth combination scheme of the virtual power plant n.

[0037] Further, the optimization matching model is a target function of the virtual power plant selecting the distributed resource,

[0038]

[0039] wherein: represents the additional benefit generated after the aggregation of the distributed resource in the combination scheme k and the virtual power plant n; represents the total benefit of the virtual power plant n after selecting the combination scheme k; represents the benefit of the distributed resource i when it is operated alone; represents the aggregation cost of the combination scheme k, which mainly occurs in the installation of communication, control and metering devices involved in the aggregation of various resources.

[0040] Further, the bilateral matching step includes:

[0041] Step 1: Calculate the marginal benefit of M distributed resources joining N virtual power plants, and each distributed resource calculates the preference degree of each virtual power plant according to the marginal benefit, and initialize i to 1; Step 2: In the ith round of matching, all distributed resources not accepted by the virtual power plant select the virtual power plant with the highest preference degree; Step 3: The virtual power plant forms its own combination scheme matrix X1, X2, …, X NStep 1: Set i to 1; Step 2: Select the combination scheme with the maximum aggregated benefit for the virtual power plant i, and obtain the delay acceptance of the virtual power plant i to the distributed resources, and reject all the distributed resources which are not accepted by the virtual power plant i; Step 3: Increase n, and repeat Step 2 and Step 1 until the delay acceptance of all the virtual power plants is calculated, and the distributed resources which are rejected by the virtual power plant i are deleted from the preferences; Step 4: Increase i, and repeat Step 2 to Step 3 until all the distributed resources are accepted or there is no virtual power plant to be selected, and the delay acceptance of each virtual power plant is the final result of the bilateral matching.

[0042] The method can ensure the stability of the configuration, so that the two parties of the matching are not disturbed by each other, thereby realizing the effective utilization rate of the power resources.

[0043] The method can realize the matching of the virtual power plant and the distributed resources, and considers the two-way selection of the virtual power plant and the distributed resources, thereby helping the state energy to realize the optimal configuration according to the marginal contribution, ensuring the stability of the configuration, and effectively solving the problem of the optimal configuration of the power resources. BRIEF DESCRIPTION OF DRAWINGS

[0044] The drawings described herein are used to provide further understanding of the present application, and should not constitute improper limitation on the present application. The terms in the drawing numbers are only used to more conveniently describe and explain the present application, and should not be understood as any additional limitation.

[0045] Figure 1 It is a flowchart of the Gale-Shapley algorithm.

[0046] Figure 2 It is the result of the bilateral matching of the virtual power plant and the distributed resources. DETAILED DESCRIPTION

[0047] The present application is further described below in combination with the drawings and specific embodiments, so that the present application can be more clearly and directly understood.

[0048] Embodiment: As Figure 1As shown, the application is a virtual power plant and distributed resource two-sided matching method based on the Gale-Shapley algorithm, which operates according to the following steps: Step 1: In the power spot market environment, a virtual power plant participates in market declaration optimization model, and the declaration curve of the integrated virtual power plant of wind, light, and water storage load is optimized; Step 2: A two-way selection decision model of virtual power plant and distributed resource is constructed, and a two-sided matching step based on the Gale-Shapley algorithm is carried out to realize the two-sided matching of virtual power plant and distributed resource.

[0049] Referring to the power spot market trading rules of Zhejiang Province, the virtual power plant operator combines distributed power generation, distributed energy storage, and controllable load organically to proxy its participation in the power market. The virtual power plant declares the output curve on the day-ahead market and participates in the day-ahead market as a price accepter. In real-time operation, due to the strong uncertainty of wind power and photovoltaic output, the deviation between the actual output of the virtual power plant and the cleared power will face examination, so the virtual power plant needs to fully consider the output fluctuation of various renewable energy generation when deciding the day-ahead market declaration curve in addition to considering the uncertainty of market price. The set S represents the uncertainty scenario set of wind, light, and water output as well as market price. The target function expression of the virtual power plant operating income in the power spot market is optimized under multiple constraint conditions, including power balance constraint, hydroelectric constraint, energy storage constraint, and interruptible load constraint.

[0050] In constructing the two-way selection decision model of virtual power plant and distributed resource, and then proposing a two-sided matching method based on the Gale-Shapley algorithm, the specific method is as follows:

[0051] In the two-sided matching process of virtual power plant and distributed resource, the distributed resource first needs to select and sort the virtual power plant. The distributed resource should select with the goal of maximizing the income generated after joining the virtual power plant:

[0052]

[0053] In the formula: represents the income of virtual power plant n when it operates independently; represents the total income of distributed resource m after joining virtual power plant n.

[0054] By optimizing the matching strategy of M independent distributed resources, the virtual power plant can obtain a combination scheme matrix of distributed resources that can be selected. Assuming that in a round of matching, there are M n distributed resources, and the distributed resources that are delayed accepted in the last round and the distributed resources that have the willingness to cooperate in this round hope to cooperate with virtual power plant n, then the number of schemes that virtual power plant n can select is The combination scheme matrix is specifically represented as:

[0055]

[0056] In the formula: x n,k,m is a 0-1 variable, x n,k,m = 1 indicates that the distributed resource m is selected as a cooperative object in the kth combination scheme of the virtual power plant n. The virtual power plant selects the optimal combination scheme in each combination scheme, and the process is a nonlinear 0-1 mixed integer optimization problem. The objective function of the virtual power plant selecting the distributed resource is specifically represented as:

[0057]

[0058] In the formula: represents the additional income generated after the aggregation of the distributed resource in the combination scheme k and the virtual power plant n; represents the total income of the virtual power plant n after selecting the combination scheme k; represents the income of the distributed resource i when it operates alone; represents the aggregation cost of the combination scheme k, which mainly occurs in the installation of communication, control, and metering devices involved in the aggregation process of various resources.

[0059] As shown in the figure, finally, the steps of the bilateral matching method of the virtual power plant and the distributed resource are as follows: Figure 1

[0060] Step 1: Calculate the marginal income of M distributed resources joining N virtual power plants. Each distributed resource calculates the preference degree for each virtual power plant according to the marginal income, and initializes i to 1;

[0061] Step 2: In the ith round of matching, all distributed resources that are not accepted by the virtual power plant select the virtual power plant with the highest preference degree;

[0062] Step 3: The virtual power plant forms its own combination scheme matrix X1, X2, …, X N according to the selection of the distributed resource and the delayed acceptance in the last round, and initializes n to 1;

[0063] Step 4: For the virtual power plant n, select the combination scheme with the maximum aggregation income to obtain the delayed acceptance of the virtual power plant n to the distributed resource, and reject all distributed resources that are not delayed accepted;

[0064] Step 5: Continue to increase n, repeat Step 3 and Step 4 until the delayed acceptance of all virtual power plants is calculated, and the virtual power plant that is rejected to cooperate is deleted from the preference of each distributed resource; ​

[0065] Step 6: continuously increase i, repeat Step 2 to Step 5 until all distributed resources are accepted or there is no virtual power plant left to be selected. At this time, the delay acceptance result of each virtual power plant is the final bilateral matching result.

[0066] The inventors conducted a test in a certain region, and the region has a total of 6 virtual power plants, a total of 10 wind power stations, 5 photovoltaic power stations, 10 energy storage power stations and 5 power users participating in the virtual power plant, wherein the typical output curves of each wind power station and photovoltaic power station are as shown in Figure 1 The final aggregation of virtual power plants and distributed resources is as shown in Figure 2 The final aggregation of virtual power plants and distributed resources is as shown in

[0067] As described in the above embodiments, by constructing a virtual power plant participating in market declaration optimization model under the environment of power spot market, the declaration curve of wind, light, water storage and load integrated virtual power plant is optimized; a two-way selection decision model of virtual power plant and distributed resource is constructed, a bilateral matching step based on Gale-Shapley algorithm is carried out, and the bilateral optimization matching of virtual power plant and distributed resource is realized, so as to ensure the stability of the configuration, ensure that the matched parties will not be disturbed by each other, realize the effective utilization rate of power resources, effectively solve the problem of optimal configuration of power resources, and have good economic benefits and practical application value.

[0068] In addition to the above embodiments, within the scope disclosed by the claims and the specification of the present application, the technical features of the present application can be reselected and adjusted, thereby constituting new embodiments, which can be realized by those skilled in the art without creative labor, therefore the embodiments not described in detail by the present application should be regarded as specific embodiments of the present application and within the protection scope of the present application.

Claims

1. A virtual power plant and distributed resource two-sided matching method based on the Gale-Shapley algorithm, characterized in that, The method comprises the following steps: S1: constructing a virtual power plant participating in market declaration optimization model in a power spot market environment, and optimizing the declaration curve of a wind, light and water storage load integrated virtual power plant; In real-time operation, in addition to considering the uncertainty of market electricity price when deciding the day-ahead market declaration curve, the output fluctuation of renewable energy generation also needs to be fully considered; specifically, by considering the output of the virtual power plant in the t period of the day-ahead market, the interruptible load reduction output in the t period, the change of deviation electricity in the t period and the scene probability, the power spot market income is increased, the interruptible load adjustment cost and the deviation electricity assessment cost are reduced, and the power spot market operation income is maximized; S2: constructing a two-way selection decision model of the virtual power plant and the distributed resources, considering that the distributed resources maximize the generated maximum income after joining the virtual power plant, and performing a two-sided matching step based on the Gale-Shapley algorithm, first, the distributed resources need to select and sort the virtual power plants, and realize the two-sided optimization matching of the virtual power plant and the distributed resources.

2. The method according to claim 1, wherein, The declaration optimization model constructed in S1 is: The set S represents the wind, light and water output of the virtual power plant and the uncertainty scene set of the market electricity price, and the target function expression of the virtual power plant operation income in the power spot market is, max R M -C IL -C PU where: T represents total time period; π s represents the probability of scenario s; R M represents the income of power spot market, the cost of interruptible load adjustment and the cost of deviation electricity assessment, respectively; IL PU represents the income of power spot market, the cost of interruptible load adjustment and the cost of deviation electricity assessment, respectively; represents the market electricity price of time period t under scenario s; λ IL represents the cost of interruptible load adjustment; k pos represents the cost of interruptible load adjustment; k neg represents the positive / negative cost coefficient of deviation electricity assessment, respectively; represents the output of virtual power plant at time t declared in day-ahead market; represents the interruptible load reduction output in time period t under scenario s; represents the interruptible load reduction output in time period t under scenario s; ​ respectively represent the over / under generation in the t period under the scene s.

3. The method according to claim 2, wherein, The declaration optimization model comprises a power balance constraint method, In the formula: and respectively represent the output of the virtual power plant, wind power, photovoltaic and hydropower in period t under scenario s; and respectively represent the charge / discharge power of the energy storage; represents the load demand when the interruptible load is not called in period t under scenario s.

4. The method according to claim 2, wherein, The declaration optimization model comprises a hydropower constraint method, V s,t+1 = V s,t + ΔV s,t V min ≤V s,t ≤V max where P HV,min and P HV,max respectively represent the minimum / maximum output of the hydropower; and respectively represent the inflow, outflow and abandoned water of the reservoir-type hydropower station in period t under scenario s; V s,t and ΔV s,t respectively represent the reservoir storage and water volume change in period t under scenario s; V min and V max respectively represent the minimum and maximum reservoir storage. η HV represents the hydroelectric power generation efficiency.

5. The method according to claim 2, wherein, The declaration optimization model comprises an energy storage constraint method, wherein: P ESS and E ESS respectively represent the maximum charge-discharge power and the energy storage capacity of the energy storage; So represents the state of charge of the energy storage in period t under scenario s; η ch and η dis respectively represent the charge-discharge efficiency of the energy storage.

6. The method according to claim 2, wherein, The declaration optimization model comprises an interruptible load constraint method, where: P IL,max represents interruptible load capacity.

7. The method according to claim 1, wherein, In the two-sided matching process of the virtual power plant and the distributed resources, the distributed resources should select the maximum income generated after joining the virtual power plant as the target: In the formula: represents the income of the virtual power plant n when it is independently operated; represents the total income of the distributed resource M after joining the virtual power plant n.

8. The method according to claim 1, wherein, The two-way selection decision model is a combination scheme of the distributed resources M maximizing the maximum income generated after joining the virtual power plant n, and the matrix is: In the formula: is a 0-1 variable, represents selecting the distributed resource M as a cooperative object in the i-th combination scheme of the virtual power plant n. th combination scheme of the virtual power plant n.

9. The method according to claim 8, wherein, The optimization matching model is a target function of the virtual power plant selecting the distributed resources, wherein: represents the additional revenue generated by aggregating distributed resources in combination scheme k with virtual power plant n; represents the total revenue of virtual power plant n after selecting combination scheme k; represents the revenue of distributed resource i when it operates alone; represents the aggregation cost of combination scheme k, which mainly comes from the installation of communication, control and metering devices involved in the aggregation process of various resources.

10. The method according to claim 9, wherein, The two-sided matching step comprises: Step 1: Calculate the marginal income of M distributed resources joining n virtual power plants, and each distributed resource calculates the preference degree of each virtual power plant according to the marginal income, and initialize i as 1; Step 2: In the i-th round of matching, all distributed resources not accepted by the virtual power plant select the virtual power plant with the highest preference degree; Step 3: Virtual power plant forms its own combination scheme matrix X1, X2, …, Xn according to the selection of distributed resources and the delay acceptance of the last round N n is initialized to 1; Step 4: For the virtual power plant n, select the combination scheme with the maximum aggregated income to obtain the delayed acceptance of the virtual power plant n to the distributed resources, and reject all distributed resources not delayed accepted; Step 5: Increase n, and repeat Step 3 and Step 4 until the delayed acceptance of all virtual power plants is calculated, and each distributed resource deletes the virtual power plant not accepted for cooperation from the preference; Step 6: Increase i, and repeat Step 2 to Step 5 until all distributed resources are accepted or there is no virtual power plant to be selected, and the delayed acceptance result of each virtual power plant is the final two-sided matching result.

Images