Distributed adjustable capacity and time consistency control method for virtual power plant

Abstract

The invention discloses a distributed adjustable capacity and time consistency control method for a virtual power plant, and relates to the field of power system regulation, and the method comprises the steps: firstly collecting the consistency control information of internal resources of the virtual power plant, including adjustable capacity, sustainable adjustment time, communication cost, economic cost, capacity credible cost and time credible cost; then, constructing a consistency index of the distributed resources in the virtual power plant participating in regulation and control, and when a user cannot meet the consistency index of the adjustable capacity or the adjustable time, determining that the distributed resources in the virtual power plant participate in regulation and control; a virtual power plant distributed adjustable capacity consistency control method based on a PSO algorithm and a virtual power plant distributed adjustable time consistency control method based on an LSTM algorithm are adopted respectively. The method comprises the following steps: carrying out combinatorial optimization on distributed users through a PSO algorithm, solving an optimal adjustable capacity combination meeting a constraint condition, carrying out combinatorial optimization on each distributed user in a time dimension through an LSTM algorithm, and solving an optimal adjustable time combination meeting the constraint condition.

Description

Technical Field

[0001] The present invention relates to the field of power system regulation and control, and in particular to a method for controlling the consistency of distributed adjustable capacity and time of a virtual power plant. Background Art

[0002] Virtual Power Plant (VPP) is an innovative energy management system that integrates and optimizes multiple distributed energy resources (such as solar power generation, wind power generation, energy storage systems, controllable loads, etc.) distributed in different geographical locations to form a unified, dispatchable power supply or demand unit. Virtual power plants do not directly own these distributed resources, but through advanced communication technology and intelligent control strategies, they can remotely monitor, dispatch and optimize the configuration of these resources, thereby providing flexible, reliable and economical power services.

[0003] The consistency control of the distributed adjustable capacity and time of virtual power plants has the following significance: it can improve the regulation efficiency, realize the coordinated optimization of distributed resources, and increase the adjustable margin of resources; it can ensure the stability of the system, avoid timing mismatch, and enhance the reliability and safety of power grid operation; it can optimize economic benefits, enhance the competitiveness of virtual power plants in the market, and reduce the cost of regulation; it can promote energy transformation, adapt to the demand for renewable energy access, and promote the coordinated utilization of resources and low-carbon development. However, the existing methods have the following shortcomings:

[0004] 1) There is a lack of standardization in the consistency of distributed resources participating in regulation. There is no unified evaluation index for the regulation capability of distributed resources (such as load regulation and energy storage charging and discharging capability), making it difficult to quantify the degree of coordination between resources. There is a lack of analysis on margin matching in the process of distributed resource regulation, which may lead to overload or idleness of some resources. The response times of different resources vary significantly (such as the fast response of energy storage systems and the slow regulation of industrial loads), and there is a lack of coordination mechanism in the time dimension. Existing methods are mostly based on simple timing alignment or delay control, and fail to fully consider dynamic time consistency.

[0005] 2) Limitations of control methods. Traditional optimization algorithms (such as linear programming and genetic algorithms) find it difficult to balance convergence speed and global optimal solution when dealing with complex nonlinear distributed control problems. Control methods based on static time windows fail to adapt to dynamic demand changes and are prone to response lags or insufficient regulation. There is a lack of in-depth mining of historical data and time series characteristics, and the prediction ability is weak.

[0006] 3) Insufficient overall system performance and insufficient consistency in resource regulation may lead to unstable overall performance of the virtual power plant; failure to balance resource utilization efficiency and regulation costs will lead to a decline in the economic benefits of the virtual power plant.

[0007] In addition, the particle swarm optimization algorithm (PSO) is an optimization algorithm based on swarm intelligence, which has the advantages of fast convergence speed and strong global search ability, but the existing methods have not fully utilized the advantages of the PSO algorithm in solving complex optimization problems. The control method based on static time windows fails to adapt to dynamic demand changes, which easily leads to response lag or insufficient regulation; it lacks in-depth mining of historical data and time series characteristics, and has weak prediction capabilities. The long short-term memory network algorithm (LSTM) is a special recurrent neural network (RNN) that can capture long-term dependencies in time series and is suitable for time series prediction and pattern recognition, but the existing methods have not been able to apply LSTM to the control prediction of virtual power plants.

[0008] To this end, we designed a distributed adjustable capacity and time consistency control method for virtual power plants to solve the above problems. Summary of the invention

[0009] The purpose of this invention is to improve the consistency level, achieve efficient coordination of distributed resource regulation capacity through PSO algorithm optimization; and ensure the synchronization of the regulation process in time through LSTM algorithm prediction. Improve system stability and economy, improve resource utilization efficiency, reduce resource idleness and overload risks; optimize regulation costs and maximize the benefits of virtual power plants. Enhance adaptability and intelligence, dynamically adapt to changes in resources and demand, and promote the development of virtual power plants in the direction of intelligence and automation.

[0010] To achieve the above object, the present invention proposes a method for controlling the consistency of distributed adjustable capacity and time of a virtual power plant, comprising the following steps:

[0011] Step S1: Collect consistency control information of internal resources of the virtual power plant, wherein the consistency control information includes adjustable capacity of each user, sustainable adjustment time, user communication cost, user economic cost, user capacity trust cost and user time trust cost;

[0012] Step S2: Based on the consistency control information collected in step S1, a consistency index for the distributed resources in the virtual power plant to participate in the regulation is constructed. The consistency index stipulates that the user can participate in the regulation of the virtual power plant only when the user's adjustable capacity is greater than a preset capacity threshold and the sustainable adjustment time is greater than a preset time threshold; the consistency index includes an adjustable capacity consistency index and an adjustable time consistency index;

[0013] Step S3: In the process of virtual power plant control, when there are users who cannot meet the adjustable capacity consistency index in step S2, a distributed adjustable capacity consistency control method of virtual power plant based on particle swarm optimization algorithm is proposed; firstly, the distributed adjustable capacity consistency control objective function of virtual power plant is constructed, with the communication, economic and trusted capacity cost minimization as the goal, and corresponding constraints are set; then, based on the particle swarm optimization algorithm, each distributed user is combined and optimized to solve the optimal adjustable capacity combination that meets the constraints, and the consistent adjustable capacity of the distributed user combination is calculated according to the combination result;

[0014] Step S4: In the virtual power plant control process, when there are users who cannot meet the adjustable time consistency index in step S2, a distributed adjustable time consistency control method for virtual power plants based on the long short-term memory network algorithm is proposed; first, a distributed adjustable time consistency control objective function of the virtual power plant is constructed, with the goal of minimizing communication, economic, and trusted time costs, and corresponding constraints are set; then, based on the long short-term memory network algorithm, each distributed user is combined and optimized to solve the optimal adjustable time combination that meets the constraints, and the consistent adjustable time of the distributed user combination is calculated based on the combination result.

[0015] Preferably, in step S1, the collected consistency control information of the internal resources of the virtual power plant is as shown in the following formula:

[0016]

[0017] Where: X i (t) is the consistency control information of virtual power plant user i collected at time t; △P i (t) is the adjustable capacity of user i collected at time t, △T i (t) is the sustainable adjustment time of user i collected at time t, is the communication cost of user i collected at time t, ψ i (t) is the economic cost of user i collected at time t, is the capacity trust cost of user i collected at time t, is the time trust cost of user i collected at time t.

[0018] Preferably, in step S2, the consistency index of distributed resources participating in regulation within the virtual power plant is constructed as shown in the following formula:

[0019] △P i (t)≥△P min (2)

[0020] △T i (t)≥△T min (3) Where: △Pi (t) is the adjustable capacity of user i collected at time t, △T i (t) is the sustainable adjustment time of user i collected at time t, △P min is the adjustable capacity threshold, △T min is an adjustable time threshold, only the adjustable capacity △P i (t) is greater than the adjustable capacity threshold △P min , and can continuously adjust the time △T i (t) is greater than the adjustable time threshold △T min Only users can participate in the virtual power plant regulation.

[0021] Preferably, in step S3, during the control process of the virtual power plant, when there is a user who cannot meet the adjustable capacity consistency index in step S2, a virtual power plant distributed adjustable capacity consistency control method based on a particle swarm optimization algorithm is proposed; the method is expressed by the following formula:

[0022] min F C (4)

[0023]

[0024] ψ C (i,j)=ψ i (t)+ψ j (t) (7)

[0025]

[0026] △P i (t)+△P j (t)≥△P min (11)

[0027]

[0028] δ' C =f PSO (X C (t)) (13)

[0029] △P i ' ,j (t)=△P i (t)+△P j (t),δ' C (i,j)=1 (14)Where: F C The distributed adjustable capacity consistency control target of the virtual power plant is is the communication cost matrix of the capacity combination, ψ C is the economic cost matrix of capacity combination, λ C is the credible capacity cost matrix, δC is the capacity combination matrix, is the i-th row and j-th column element of the communication cost matrix, ψ C (i,j) is the i-th row and j-th column element of the economic cost matrix, λ C (i,j) is the i-th row and j-th column element of the trusted capacity cost matrix, is the communication cost of user i collected at time t, ψ i (t) is the economic cost of user i collected at time t, is the capacity trust cost of user i collected at time t, is the communication cost of user j collected at time t, ψ j (t) is the economic cost of user j collected at time t, is the capacity trust cost of user j collected at time t, m is the number of users participating in the adjustable capacity combination, △P i (t) is the adjustable capacity of user i collected at time t, △P j (t) is the adjustable capacity of user j collected at time t, △T i (t) is the sustainable adjustment time of user i collected at time t, △P min is the adjustable capacity threshold, △T min is an adjustable time threshold; X C (t) is the input sample of the particle swarm optimization algorithm, δ' C is the adjustable capacity user combination matrix output by the particle swarm optimization algorithm, δ' C (i, j) is the adjustable capacity user combination matrix element output by the particle swarm optimization algorithm, f PSO () is the distributed adjustable capacity consistency solution model of virtual power plant based on particle swarm optimization algorithm, △P i ' ,j (t) is the consistent adjustable capacity calculated based on the combined output results of the particle swarm optimization algorithm.

[0030] Preferably, in the virtual power plant control process, when there is a user who cannot meet the adjustable time consistency index in step S2, a distributed adjustable time consistency control method of a virtual power plant based on LSTM is proposed; the method is expressed by the following formula:

[0031] min F T (15)

[0032]

[0033] ψ T (i,j)=ψ i (t)+ψ j (t) (18)

[0034]

[0035] △T i (t)+△T j (t)≥△T min (twenty two)

[0036]

[0037] δ' T =f LSTM (X T (t)) (24)

[0038] △T i ' ,j (t) = △T i (t)+△T j (t),δ' T (i,j)=1 (25)

[0039] Where: F T The distributed adjustable time consistency control target of the virtual power plant is is the communication cost matrix of the time combination, ψ T is the economic cost matrix of the time combination, λ T is the credible time cost matrix, δ T is the time combination matrix, is the i-th row and j-th column element of the time-combined communication cost matrix, ψ T (i,j) is the i-th row and j-th column element of the economic cost matrix of the time combination, λ T (i,j) is the i-th row and j-th column element of the trusted time cost matrix, is the communication cost of user i collected at time t, ψ i (t) is the economic cost of user i collected at time t, is the time trust cost of user i collected at time t, is the communication cost of user j collected at time t, ψ j (t) is the economic cost of user j collected at time t, is the time trust cost of user j collected at time t, n is the number of users participating in the adjustable time combination, △P i (t) is the adjustable capacity of user i collected at time t, △P j (t) is the adjustable capacity of user j collected at time t, △T i (t) is the sustainable adjustment time of user i collected at time t, △P min is the adjustable capacity threshold, △T min is the adjustable time threshold, X T (t) is the input sample of the long short-term memory network algorithm, δ'T is the adjustable time user combination matrix output by the long short-term memory network algorithm, δ' T (i, j) is the adjustable time user combination matrix element output by the long short-term memory network algorithm, f LSTM () is the distributed adjustable time consistency solution model of virtual power plant based on long short-term memory network algorithm, △T i ' ,j (t) is the consistent adjustable time calculated based on the LSTM output combination result.

[0040] Compared with the prior art, the present invention has the following beneficial effects:

[0041] 1. By constructing the internal resource consistency control information collection technology of the virtual power plant, the present invention can comprehensively and accurately collect the adjustable information and control cost information of each user in the virtual power plant, providing a solid information foundation for consistency control; through the proposed distributed adjustable capacity and time consistency control method based on the PSO algorithm and LSTM algorithm, the present invention realizes the precise coordination and prediction of the distributed resource adjustment capacity and adjustment time, and significantly improves the consistency control level of the virtual power plant.

[0042] 2. The consistency index of distributed resources participating in regulation in the virtual power plant constructed by the present invention effectively screens distributed resources that meet the regulation requirements and reduces unnecessary calculation and regulation costs. By combining and optimizing distributed users through the PSO algorithm, the present invention finds the adjustable resource combination with the lowest communication, economic, and trusted capacity costs, achieving efficient resource utilization and cost minimization. The application of the LSTM algorithm further improves consistency control in the time dimension and reduces additional costs caused by time mismatch. Through precise prediction and collaborative control, the present invention reduces the risk of power grid operation caused by timing mismatch and improves the reliability and safety of the power grid.

[0043] 3. The present invention combines the advanced PSO algorithm and LSTM algorithm to realize the intelligent control and automatic scheduling of the distributed resources of the virtual power plant, and promotes the development of the virtual power plant towards intelligence and automation. This method not only improves the accuracy and efficiency of regulation, but also reduces the dependence on manual intervention, and improves the overall operation level of the virtual power plant. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 It is a flow chart of a method for controlling the consistency of distributed adjustable capacity and time of a virtual power plant in an embodiment of the present invention; DETAILED DESCRIPTION

[0045] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0046] See also Figure 1 ,This example proposes a distributed adjustable capacity and time consistency control method for a virtual power plant, including the following steps:

[0047] S1. Propose a technology for collecting information on the consistency control of resources within a virtual power plant, collect the adjustable information and control cost information of each user within the virtual power plant, and provide an information basis for the consistency control of internal distributed resources. The collected information mainly includes the adjustable capacity of each user, the sustainable adjustment time, the user's communication cost, the user's economic cost, the user's capacity trust cost, and the user's time trust cost.

[0048] The collected virtual power plant internal resource consistency control information is shown in formula (1), which mainly includes the adjustable capacity of each user, the sustainable adjustment time, the user communication cost, the user economic cost, the user capacity trust cost, and the user time trust cost.

[0049]

[0050] Where: X i (t) is the consistency control information of virtual power plant user i collected at time t; △P i (t) is the adjustable capacity of user i collected at time t, △T i (t) is the sustainable adjustment time of user i collected at time t, is the communication cost of user i collected at time t, ψ i (t) is the economic cost of user i collected at time t, is the capacity trust cost of user i collected at time t, is the time trust cost of user i collected at time t.

[0051] S2. Construct a consistency index for the participation of distributed resources in the regulation of the virtual power plant. In the regulation of the virtual power plant, different users have different adjustable capacities and sustainable regulation times. If all users are regulated, the amount of calculation will be greatly increased and the calculation efficiency will be affected. Therefore, the present invention constructs a consistency index for the participation of distributed resources in the regulation of the virtual power plant. The index stipulates that only users whose adjustable capacity is greater than the capacity threshold and whose sustainable regulation time is greater than the time threshold can participate in the regulation of the virtual power plant. This indicator aims to reduce the control cost of the virtual power plant and improve the control efficiency of the virtual power plant. The constructed consistency index includes an adjustable capacity consistency index and an adjustable time consistency index, wherein the adjustable capacity consistency index of the distributed resources of the virtual power plant is that when the user's adjustable capacity is greater than the capacity threshold, the user can participate in the control of the virtual power plant; the adjustable time consistency index of the distributed resources of the virtual power plant is that when the user's sustainable regulation time is greater than the time threshold, the user can participate in the control of the virtual power plant.

[0052] The consistency index for the participation of distributed resources in the regulation of a virtual power plant is constructed as shown in formulas (2)-(3), which provide a consistency index for the conditions for each distributed user to participate in the virtual power plant. In the regulation of a virtual power plant, different users have different adjustable capacities and sustainable regulation times. If all users are regulated, the amount of calculation will be greatly increased and the calculation efficiency will be affected. Therefore, the present invention constructs a consistency index for the participation of distributed resources in the regulation of a virtual power plant. This index stipulates that only users whose adjustable capacity is greater than the capacity threshold and whose sustainable regulation time is greater than the time threshold can participate in the regulation of a virtual power plant. This index aims to reduce the control cost of a virtual power plant and improve the control efficiency of a virtual power plant.

[0053] Formula (2) is the consistency index of the adjustable capacity of distributed resources of the virtual power plant constructed by the present invention. When the user's adjustable capacity is greater than the capacity threshold, the user can participate in the control of the virtual power plant.

[0054] Formula (3) is the adjustable time consistency index of the distributed resources of the virtual power plant constructed by the present invention. When the sustainable regulation time of the user is greater than the time threshold, the user can participate in the control of the virtual power plant.

[0055] △P i (t)≥△P min (2)

[0056] △T i (t)≥△T min (3)

[0057] Where: △P i (t) is the adjustable capacity of user i collected at time t, △T i (t) is the sustainable adjustment time of user i collected at time t, △P min is the adjustable capacity threshold, △T minis an adjustable time threshold.

[0058] S3. A distributed adjustable capacity consistency control method for a virtual power plant based on a PSO particle swarm optimization algorithm is proposed. In the virtual power plant control process, when all users cannot meet the adjustable capacity consistency index in step S2, the distributed users that do not meet the conditions can be combined to construct a combination of resources that meet the adjustable capacity consistency index. The present invention proposes to optimize distributed users based on the PSO particle swarm optimization algorithm to optimize the adjustable resource combination with the lowest communication, economic, and trusted capacity costs. First, a distributed adjustable capacity consistency control objective function for a virtual power plant is constructed. In the combination process, the goal is to minimize the communication, economic, and trusted capacity costs. Constraints include user participation in adjustable capacity combination constraints, that is, users whose adjustable capacity is less than the adjustable capacity threshold and whose adjustable time is greater than the time threshold can participate in the adjustable capacity combination; and adjustable capacity combination condition constraints, that is, the adjustable capacity of the combined user should not be lower than the adjustable capacity threshold. Then, the present invention proposes a method for solving the consistency of distributed adjustable capacity of a virtual power plant based on the PSO particle swarm optimization algorithm, performs combination optimization on each user based on the PSO particle swarm optimization algorithm, solves the optimal adjustable capacity combination that meets the constraints, and calculates the consistent adjustable capacity of the distributed user combination according to the optimal adjustable capacity combination result output by the PSO particle swarm optimization algorithm.

[0059] The distributed adjustable capacity consistency control method of the virtual power plant based on the PSO particle swarm optimization algorithm is shown in formulas (4)-(14). In the virtual power plant control process, when all users cannot meet the adjustable capacity consistency index in step S2, the distributed users that do not meet the conditions can be combined to construct a combined adjustable resource that meets the consistency index. The present invention proposes to optimize the distributed users based on the PSO particle swarm optimization algorithm to optimize the adjustable resource combination with the lowest communication, economic and trusted capacity costs.

[0060] Formula (4) is the objective function of the distributed adjustable capacity consistency control of the virtual power plant constructed by the present invention. In the combination process, the communication, economic, and trusted capacity costs are minimized. The calculation process is shown in formula (5), which is mainly calculated based on the communication, economic, and trusted capacity matrices and the capacity combination matrix.

[0061] Formula (6) is a communication cost matrix calculation model for capacity combination. The element in the i-th row and j-th column of the matrix is ​​calculated according to formula (6).

[0062] Formula (7) is the economic cost matrix calculation model of capacity combination. The element in the i-th row and j-th column of the matrix is ​​calculated according to formula (7).

[0063] Formula (8) is a calculation model for the trusted capacity cost matrix. The element in the i-th row and j-th column of the matrix is ​​calculated according to formula (8).

[0064] Formula (9) is a capacity combination matrix with m×m dimensions, where m is the number of users participating in the adjustable capacity combination, and an element of 1 indicates that the capacity combination is performed for the users in the row and column corresponding to the element.

[0065] Formula (10) is the constraint for users to participate in the adjustable capacity combination, that is, users whose adjustable capacity is less than the adjustable capacity threshold and whose adjustable time is greater than the time threshold can participate in the adjustable capacity combination.

[0066] Formula (11) is the constraint of the adjustable capacity combination condition, that is, the adjustable capacity of the combined user should not be lower than the adjustable capacity threshold.

[0067] Formulas (12)-(14) are a method for solving the consistency of distributed adjustable capacity of a virtual power plant based on the PSO particle swarm optimization algorithm. The present invention performs combinatorial optimization on each user based on the PSO particle swarm optimization algorithm to solve the optimal adjustable capacity combination that meets the constraints.

[0068] Formula (12) is the input sample of the PSO particle swarm optimization algorithm, including the adjustable capacity of each user, sustainable adjustment time, communication cost, economic cost, trusted capacity cost, trusted capacity consistency control objective function, adjustable capacity threshold, and time threshold.

[0069] Formula (13) is a training model of the PSO particle swarm optimization algorithm. The present invention trains input samples based on the PSO particle swarm optimization algorithm to solve the optimal adjustable capacity combination.

[0070] Formula (14) is a consistent capacity model based on the PSO particle swarm optimization algorithm. According to the optimal adjustable capacity combination result output by the PSO particle swarm optimization algorithm, the consistent adjustable capacity of the distributed user combination is calculated.

[0071] min F C (4)

[0072]

[0073] ψ C (i,j)=ψ i (t)+ψ j (t) (7)

[0074]

[0075] △P i (t)+△P j (t)≥△P min (11)

[0076]

[0077] δ' C =f PSO (X C (t)) (13)

[0078] △P i ' ,j (t)=△P i (t)+△P j (t),δ' C (i,j)=1 (14)

[0079] Where: F C The distributed adjustable capacity consistency control target of the virtual power plant is is the communication cost matrix of the capacity combination, ψ C is the economic cost matrix of capacity combination, λ C is the credible capacity cost matrix, δ C is the capacity combination matrix. is the i-th row and j-th column element of the communication cost matrix, ψ C (i,j) is the i-th row and j-th column element of the economic cost matrix, λ C (i,j) is the i-th row and j-th column element of the trusted capacity cost matrix, is the communication cost of user i collected at time t, ψ i (t) is the economic cost of user i collected at time t, is the capacity trust cost of user i collected at time t, is the communication cost of user j collected at time t, ψ j (t) is the economic cost of user j collected at time t, is the capacity trust cost of user j collected at time t. m is the number of users participating in the adjustable capacity combination, △P i (t) is the adjustable capacity of user i collected at time t, △P j (t) is the adjustable capacity of user j collected at time t, △T i (t) is the sustainable adjustment time of user i collected at time t, △P min is the adjustable capacity threshold, △T min is an adjustable time threshold. C (t) is the input sample of PSO particle swarm optimization algorithm, δ' C is the adjustable capacity user combination matrix output by the PSO particle swarm optimization algorithm, δ' C (i, j) is the adjustable capacity user combination matrix element output by the PSO particle swarm optimization algorithm, f PSO() Distributed adjustable capacity consistency solution model of virtual power plant based on PSO particle swarm optimization algorithm, △P i ' ,j (t) is the consistent adjustable capacity calculated based on the combined output results of the PSO particle swarm optimization algorithm.

[0080] S4. A distributed adjustable time consistency control method for a virtual power plant based on an LSTM algorithm is proposed. In the control process of the virtual power plant, when all users cannot meet the time consistency index in step S2, the distributed users that do not meet the conditions can be combined to construct a resource combination that meets the adjustable time consistency. The present invention proposes to optimize distributed users based on the LSTM algorithm to optimize the resource combination with the minimum communication, economic, and trusted time costs. First, a distributed adjustable time consistency control objective function for a virtual power plant is constructed. In the combination process, the communication, economic, and trusted time costs are minimized. Constraints include user participation in adjustable time combination constraints, that is, users whose adjustable capacity is greater than the adjustable capacity threshold and whose sustainable control time is less than the time threshold can participate in the adjustable time combination; adjustable time combination condition constraints, that is, the sustainable control time of the combined user should not be less than the adjustable time threshold. Then, the present invention proposes a method for solving the consistency of distributed adjustable time of a virtual power plant based on the LSTM algorithm, performs combination optimization on each user based on the LSTM algorithm, solves the optimal adjustable time combination that meets the constraints, and calculates the consistent adjustable time of the distributed user combination according to the optimal adjustable time combination result output by the LSTM algorithm.

[0081] The constructed virtual power plant distributed adjustable time consistency control method based on LSTM algorithm is shown in formulas (15)-(25). In the virtual power plant control process, users who do not meet the adjustable time consistency index in step S2 are combined. The present invention proposes to optimize the adjustable time of distributed users based on LSTM algorithm to optimize the adjustable resource combination with the minimum communication, economic and reliable time cost.

[0082] Formula (15) is the objective function of the distributed adjustable time consistency control of the virtual power plant constructed by the present invention. In the combination process, the communication, economic and trusted time costs are minimized. The calculation process is shown in formula (16), which is mainly calculated based on the communication, economic, trusted time matrix and time combination matrix.

[0083] Formula (17) is a calculation model for the time-combined communication cost matrix. The element in the i-th row and j-th column of the matrix is ​​calculated according to formula (17).

[0084] Formula (18) is the economic cost matrix calculation model of time combination. The element in the i-th row and j-th column of the matrix is ​​calculated according to formula (18).

[0085] Formula (19) is a calculation model for the trusted time cost matrix. The element in the i-th row and j-th column of the matrix is ​​calculated according to formula (19).

[0086] Formula (20) is a time combination matrix with n×n dimensions, where n is the number of users participating in the adjustable time combination, and an element of 1 indicates that the users in the row and column corresponding to the element are time-combined.

[0087] Formula (21) is the constraint for users to participate in the adjustable time combination, that is, users whose adjustable capacity is greater than the adjustable capacity threshold and whose adjustable time is less than the time threshold can participate in the adjustable time combination.

[0088] Formula (22) is the adjustable time combination condition constraint, that is, the combined user's adjustable time should not be less than the adjustable time threshold.

[0089] Formulas (23)-(25) are a method for solving the distributed adjustable time consistency of a virtual power plant based on the LSTM algorithm. The present invention performs combinatorial optimization on each user based on the LSTM algorithm to solve the optimal adjustable time combination that meets the constraints.

[0090] Formula (23) is the input sample of the LSTM algorithm, including the adjustable capacity of each user, sustainable adjustment time, communication cost, economic cost, trusted time cost, trusted time consistency control objective function, adjustable capacity threshold, and time threshold.

[0091] Formula (24) is a training model of the LSTM algorithm. The present invention trains input samples based on the LSTM algorithm to solve the optimal adjustable time combination.

[0092] Formula (25) is a consistency time model based on the LSTM algorithm. According to the optimal adjustable time combination result output by the LSTM algorithm, the consistent adjustable time of the distributed user combination is calculated.

[0093] min F T (15)

[0094]

[0095] ψ T (i,j)=ψ i (t)+ψ j (t) (18)

[0096]

[0097] △T i (t)+△T j (t)≥△T min (twenty two)

[0098]

[0099] δ' T =f LSTM (X T (t)) (24)

[0100] △T i ' ,j (t) = △T i (t)+△T j (t),δ' T (i,j)=1 (25)Where: F T The distributed adjustable time consistency control target of the virtual power plant is is the communication cost matrix of the time combination, ψ T is the economic cost matrix of the time combination, λ T is the credible time cost matrix, δ T is the time combination matrix. is the i-th row and j-th column element of the time-combined communication cost matrix, ψ T (i,j) is the i-th row and j-th column element of the economic cost matrix of the time combination, λ T (i,j) is the i-th row and j-th column element of the trusted time cost matrix, is the communication cost of user i collected at time t, ψ i (t) is the economic cost of user i collected at time t, is the time trust cost of user i collected at time t, is the communication cost of user j collected at time t, ψ j (t) is the economic cost of user j collected at time t, is the time trust cost of user j collected at time t. n is the number of users participating in the adjustable time combination, △P i (t) is the adjustable capacity of user i collected at time t, △P j (t) is the adjustable capacity of user j collected at time t, △T i (t) is the sustainable adjustment time of user i collected at time t, △P min is the adjustable capacity threshold, △T min is an adjustable time threshold. T (t) is the input sample of the LSTM algorithm, δ' T is the adjustable time user combination matrix output by the LSTM algorithm, δ' T (i, j) is the adjustable time user combination matrix element output by the LSTM model, f LSTM ()Distributed adjustable time consistency solution model of virtual power plant based on LSTM algorithm, △T i ' ,j(t) is the consistent adjustable time calculated based on the combined output results of the LSTM algorithm.

[0101] After the above steps, the adjustable capacity and adjustable time consistency control of each user within the virtual power plant can be achieved, and a resource combination that meets the consistency indicators can be combined at the lowest cost, thereby improving the effectiveness of virtual power plant control.

[0102] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A method for controlling the consistency of distributed adjustable capacity and time of a virtual power plant, characterized in that: The following steps are involved: Step S1: Collect consistency control information of internal resources of the virtual power plant, wherein the consistency control information includes adjustable capacity of each user, sustainable adjustment time, user communication cost, user economic cost, user capacity trust cost and user time trust cost; Step S2: Based on the consistency control information, a consistency index for the distributed resources in the virtual power plant to participate in the regulation is constructed. The consistency index stipulates that the user can participate in the regulation of the virtual power plant only when the user's adjustable capacity is greater than a preset capacity threshold and the sustainable adjustment time is greater than a preset time threshold; the consistency index includes an adjustable capacity consistency index and an adjustable time consistency index; Step S3: In the process of virtual power plant control, when there are users who cannot meet the adjustable capacity consistency index in step S2, a distributed adjustable capacity consistency control method of virtual power plant based on particle swarm optimization algorithm is proposed: construct the distributed adjustable capacity consistency control objective function of virtual power plant, with the communication, economic and trusted capacity cost minimization as the goal, and set constraints; based on the particle swarm optimization algorithm, each distributed user is combined and optimized to solve the optimal adjustable capacity combination that meets the constraints, and the consistent adjustable capacity of the distributed user combination is calculated according to the optimal adjustable capacity combination; Step S4: In the virtual power plant control process, when there are users who cannot meet the adjustable time consistency index in step S2, a distributed adjustable time consistency control method for virtual power plants based on the long short-term memory network algorithm is proposed: a distributed adjustable time consistency control objective function of the virtual power plant is constructed, with the goal of minimizing communication, economic, and trusted time costs, and setting constraints; based on the long short-term memory network algorithm, each distributed user is combined and optimized to solve the optimal adjustable time combination that meets the constraints, and the consistent adjustable time of the distributed user combination is calculated based on the optimal adjustable time combination.

2. A method for controlling the consistency of distributed adjustable capacity and time of a virtual power plant according to claim 1, characterized in that: In step S1, the collected consistency control information of the internal resources of the virtual power plant is shown in the following formula: Where: X i (t) is the consistency control information of virtual power plant user i collected at time t; △P i (t) is the adjustable capacity of user i collected at time t, △T i (t) is the sustainable adjustment time of user i collected at time t, is the communication cost of user i collected at time t, ψ i (t) is the economic cost of user i collected at time t, λ i C (t) is the capacity trust cost of user i collected at time t, λ i T (t) is the time trust cost of user i collected at time t.

3. A method for controlling the consistency of distributed adjustable capacity and time of a virtual power plant according to claim 1, characterized in that: In step S2, the consistency index of distributed resources participating in regulation within the virtual power plant is constructed as shown in the following formula: △P i (t)≥△P min (2) △T i (t)≥△T min (3) Where: △P i (t) is the adjustable capacity of user i collected at time t, △T i (t) is the sustainable adjustment time of user i collected at time t, △P min is the adjustable capacity threshold, △T min is an adjustable time threshold, only the adjustable capacity △P i (t) is greater than the adjustable capacity threshold △P min , and can continuously adjust the time △T i (t) is greater than the adjustable time threshold △T min Only users can participate in the virtual power plant regulation.

4. A method for controlling the consistency of distributed adjustable capacity and time of a virtual power plant according to claim 1, characterized in that: In step S3, during the virtual power plant control process, when there is a user who cannot meet the adjustable capacity consistency index in step S2, a virtual power plant distributed adjustable capacity consistency control method based on a particle swarm optimization algorithm is proposed; the method is expressed by the following formula: my F C (4) △P i (t)+△P j (t)≥△P min (11) δ' C =f PSO (X C (t)) (13) △P i ' ,j (t) = △P i (t)+△P j (t),δ' C (i,j)=1 (14)Where: F C The distributed adjustable capacity consistency control target of the virtual power plant is is the communication cost matrix of the capacity combination, ψ C is the economic cost matrix of capacity combination, λ C is the credible capacity cost matrix, δ C is the capacity combination matrix, is the i-th row and j-th column element of the communication cost matrix, ψ C (i,j) is the i-th row and j-th column element of the economic cost matrix, λ C (i,j) is the i-th row and j-th column element of the trusted capacity cost matrix, is the communication cost of user i collected at time t, ψ i (t) is the economic cost of user i collected at time t, λ i C (t) is the capacity trust cost of user i collected at time t, is the communication cost of user j collected at time t, ψ j (t) is the economic cost of user j collected at time t, is the capacity trust cost of user j collected at time t, m is the number of users participating in the adjustable capacity combination, △P i (t) is the adjustable capacity of user i collected at time t, △P j (t) is the adjustable capacity of user j collected at time t, △T i (t) is the sustainable adjustment time of user i collected at time t, △P min is the adjustable capacity threshold, △T min is an adjustable time threshold; X C (t) is the input sample of the particle swarm optimization algorithm, δ' C is the adjustable capacity user combination matrix output by the particle swarm optimization algorithm, δ' C (i, j) is the adjustable capacity user combination matrix element output by the particle swarm optimization algorithm, f PSO () is the distributed adjustable capacity consistency solution model of virtual power plant based on particle swarm optimization algorithm, △P i ' ,j (t) is the consistent adjustable capacity calculated based on the combined output results of the particle swarm optimization algorithm.

5. The method for controlling the consistency of distributed adjustable capacity and time of a virtual power plant according to claim 1, characterized in that: In step S3, during the virtual power plant control process, when there is a user who cannot meet the adjustable time consistency index in step S2, a distributed adjustable time consistency control method for a virtual power plant based on LSTM is proposed, and the method is expressed by the following formula: my F T (15) ψ T (i,j)=ψ i (t)+ψ j (t) (18) △T i (t)+△T j (t)≥△T min (22) δ' T =f LSTM (X T (t)) (24) △T i ' ,j (t) = △T i (t)+△T j (t),δ' T (i,j)=1 (25)Where: F T The distributed adjustable time consistency control target of the virtual power plant is is the communication cost matrix of the time combination, ψ T is the economic cost matrix of the time combination, λ T is the credible time cost matrix, δ T is the time combination matrix, is the i-th row and j-th column element of the time-combined communication cost matrix, ψ T (i,j) is the i-th row and j-th column element of the economic cost matrix of the time combination, λ T (i,j) is the i-th row and j-th column element of the trusted time cost matrix, is the communication cost of user i collected at time t, ψ i (t) is the economic cost of user i collected at time t, λ i T (t) is the time trust cost of user i collected at time t, is the communication cost of user j collected at time t, ψ j (t) is the economic cost of user j collected at time t, is the time trust cost of user j collected at time t, n is the number of users participating in the adjustable time combination, △P i (t) is the adjustable capacity of user i collected at time t, △P j (t) is the adjustable capacity of user j collected at time t, △T i (t) is the sustainable adjustment time of user i collected at time t, △P min is the adjustable capacity threshold, △T min is the adjustable time threshold, X T (t) is the input sample of the long short-term memory network algorithm, δ' T is the adjustable time user combination matrix output by the long short-term memory network algorithm, δ' T (i, j) is the adjustable time user combination matrix element output by the long short-term memory network algorithm, f LSTM () is the distributed adjustable time consistency solution model of virtual power plant based on long short-term memory network algorithm, △T i ' ,j (t) is the consistent adjustable time calculated based on the LSTM output combination result.

Images