Power fluctuation stabilizing control method and system

Abstract

The invention belongs to the technical field of power distribution network flexible load resource cluster optimization control, and particularly relates to a power fluctuation stabilization control method and system, aiming at a distributed new energy power fluctuation stabilization control scene, a multi-dimensional clustering index is established, and an improved K-means + + algorithm based on index weighting is adopted to realize mass dispersed heterogeneous flexible load resource grouping clustering; constructing an aggregation evaluation model, evaluating the response capability and economical efficiency of the resource polymer, and judging to obtain a polymer regulation priority to realize resource hierarchical regulation; and based on a particle swarm algorithm, with optimal economy as a target and a power fluctuation rate as a coupling constraint, constructing a power distribution network power fluctuation stabilization control model, and optimizing a resource aggregate regulation quantity in real time. According to the method, a clustering model is built, and regulation and control resources with small capacity and large quantity are subjected to dimensionality reduction to form resource clusters with large capacity and small quantity; the control cost is reduced, and the solving speed is increased.

Description

Technical Field

[0001] The present invention belongs to the technical field of distribution network flexible load resource cluster optimization control, and in particular relates to a power fluctuation smoothing control method and system. Background Art

[0002] With the access of large-scale distributed energy to the power grid, new power generation systems with high penetration rates have put forward new requirements for the participation of demand-side flexible load resources in regulation. The rapid changes in system state variables will affect the supply and demand balance of the power grid, increase the system operating costs, and directly affect the power quality and stability of the power grid. In order to avoid the output volatility and randomness of high-proportion distributed new energy restricting the rapid development of new energy, traditional distribution networks are in urgent need of rich and diverse and more flexible power regulation methods. A large number of diversified flexible load resources with different types and parameters in the distribution network have been proven to have great potential to provide power regulation services, and have gradually been incorporated into the interactive objects of the active distribution network, such as air conditioners, water heaters, and distributed energy storage. However, due to its large number and small single-unit capacity, the direct participation of flexible loads in power fluctuation smoothing will bring many problems such as the explosion of decision variable dimensions and difficulty in convergence of solution results. At the same time, most traditional clustering algorithms have poor reliability and convergence when applied in the field of power systems. There is little research on the clustering of flexible load resources serving the power fluctuation smoothing control of distribution networks, which makes actual scheduling lack of reference and is not conducive to the optimization and control of regulation resources. Summary of the invention

[0003] The present invention provides a power fluctuation smoothing control method and system to solve the defects in the prior art that flexible loads directly participate in power fluctuation smoothing, which leads to an explosion of decision variable dimensions and difficulty in converging solution results, and realize power fluctuation smoothing based on resource aggregation.

[0004] The present invention provides a power fluctuation smoothing control method, comprising: Establish multi-dimensional clustering indexes and use the improved K-means++ algorithm based on index weighting to realize the clustering of dispersed heterogeneous flexible load resources; Based on the clustering results, an aggregation evaluation model is constructed to evaluate the responsiveness and economic efficiency of resource aggregates, and the aggregate regulation priority is obtained to achieve hierarchical regulation; Based on the particle swarm algorithm, with the optimal economy as the goal and the power fluctuation rate as the coupling constraint, a power fluctuation smoothing control model for the distribution network is constructed. The resource aggregate regulation quantity is optimized in real time, and resources are regulated in a hierarchical manner according to the regulation priority to achieve power fluctuation smoothing control of distributed renewable energy grid-connected.

[0005] According to a power fluctuation smoothing control method provided by the present invention, the improved K-means++ algorithm based on index weighting is used to realize the clustering of massive distributed heterogeneous flexible load resources, including the following steps: 1) According to the optimization goal served by clustering, establish a resource active leveling clustering index system and calculate the value of each index in the index system; 2) Calculate the weight of active power smoothing clustering index based on subjective and objective factors, and dynamically adjust the weight based on the real-time grid smoothing demand; 3) The improved K-means++ algorithm based on index weighting realizes dynamic clustering of adjustable flexible load resources.

[0006] According to a power fluctuation smoothing control method provided by the present invention, based on the clustering results, an aggregation evaluation model is constructed to evaluate the responsiveness and economy of resource aggregates, and the aggregate regulation priority is obtained to achieve hierarchical regulation, including: 1) Based on the weight-improved K-means++ algorithm, the flexible load resources are divided into K Clusters are formed, and the response capability index of each cluster center is used as the common characteristic parameter of the aggregate to determine the regulation priority of the flexible load aggregate in the smoothing control and realize hierarchical regulation. 2) For each cluster, an aggregate evaluation model is established that takes into account the flexible load aggregate response capability, including: Aggregate the adjustable performance of each cluster,

[0007]

[0008] In the formula, For polymer k The number of individual flexible loads included; It means t Moment Aggregate k The adjustable capacity available; Represents an aggregate k The control accuracy rate; Aggregate the unit control cost of each cluster,

[0009] In the formula, For polymer t The aggregate control cost at the moment represents the sum of the control costs of each flexible load monomer in the aggregate; The priority of resource regulation is determined according to the results of the aggregate evaluation model, and the regulation amount is sent to several distributed resources in the resource aggregate according to the priority order of resource regulation in the subsequent optimization control.

[0010] According to a power fluctuation smoothing control method provided by the present invention, based on a particle swarm algorithm, with economic optimization as the goal and power fluctuation rate as a coupling constraint, a distribution network power fluctuation smoothing control model is constructed, the resource aggregate control quantity is optimized in real time, and resources are hierarchically controlled according to the control priority, so as to realize the power fluctuation smoothing control of distributed new energy grid connection, including: 1) Establish a minute-level AGC control framework model based on resource clustering and optimization algorithm; 2) Establish the economic objective function and constraints of the optimization control model; 3.2.1. Based on the priority of resource regulation determined by the aggregate evaluation model, the economic optimization goal at the aggregate level is established:

[0011] 3.2.2. The flexible load operation constraints based on virtual energy storage are: The charging and discharging power constraints and capacity constraints are the basic conditions that the flexible load aggregate needs to meet during actual operation:

[0012]

[0013] 3.2.3. The power balance constraint is,

[0014] In the formula, For the Total distributed photovoltaic power at the moment; For the The wind farm output power at the moment; For the Power of other loads except flexible load at any time; For the Power loss at each moment; 3.2.4. The power fluctuation rate constraint of the lower grid point is: Defining the power fluctuation rate at the downstream point based on the specific structure and resource parameters of the distribution network The constraint condition is that volatility Control within the expected target volatility Inside;

[0015] 3.2.5. Economic-based polymer decomposition regulation, The objective function and constraints are as follows: ; 3) Establish an optimization control model based on particle swarm algorithm; The present invention also provides a power fluctuation smoothing control system, comprising: Multidimensional clustering module, used to establish multidimensional clustering indicators; using the improved K-means++ algorithm based on indicator weighting to achieve clustering of dispersed heterogeneous flexible load resources; Aggregation evaluation module, which is used to build an aggregation evaluation model, evaluate the responsiveness and economic efficiency of resource aggregates, and derive the aggregate regulation priority to achieve hierarchical regulation; The smoothing control module is used to build a distribution network power fluctuation smoothing control model based on the particle swarm algorithm, with economic optimization as the goal and power fluctuation rate as the coupling constraint, optimize the resource aggregate control quantity in real time, and hierarchically control resources according to the control priority to achieve distributed new energy grid-connected power fluctuation smoothing control.

[0016] A power fluctuation smoothing control system provided in accordance with the present invention includes: a multidimensional clustering module, which is specifically used to establish a resource active power smoothing clustering index system for the optimization target served by clustering, and calculate the value of each index in the index system; calculate the clustering index weights in combination with subjective and objective factors, and dynamically adjust the weights based on the real-time power grid smoothing demand; and realize dynamic clustering of adjustable flexible load resources based on the improved K-means++ algorithm based on the index weights.

[0017] A power fluctuation smoothing control system provided by the present invention comprises: an aggregation evaluation module, which is specifically used to divide the massive flexible load resources into K Clusters are formed, and the response capability index of each cluster center is used as the common characteristic parameter of the aggregate to determine the regulation priority of the flexible load aggregate in smoothing control and realize hierarchical regulation. An aggregation evaluation model considering the response capability of flexible load aggregation is established for each cluster, including the adjustable capacity, regulation accuracy and regulation cost of the cluster after clustering.

[0018] A power fluctuation smoothing control system provided according to the present invention includes: a smoothing control module, which is specifically used to establish a minute-level AGC control framework model based on resource clustering and optimization algorithm; set up the economic objective function and constraint conditions of the optimization control model; and establish an optimization control model based on a particle swarm algorithm.

[0019] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, any of the power fluctuation smoothing control methods described above is implemented.

[0020] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the power fluctuation smoothing control method as described above is implemented.

[0021] Compared with the prior art, the present invention has the following beneficial effects: The present invention applies the weight-based improved K-means++ clustering algorithm to the clustering of flexible load resources, builds a real-time clustering model, reduces the dimension of small-capacity and large-number control resources into large-capacity and small-number resource clusters, and performs hierarchical control on load aggregates based on flexible load response capabilities, providing support for a high proportion of resources participating in the control. Compared with traditional methods, this method has the comprehensive advantages of being more economical, reliable, and easy to schedule in meeting real-time control needs.

[0022] The present invention studies and improves the K-means++ clustering algorithm based on index weighting, establishes a multidimensional clustering index including multiple response capability indicators and economic indicators, and realizes the clustering of massive distributed heterogeneous flexible load resources. Taking into account the flexible load response capability and response cost, the cluster center of each flexible load resource aggregate is used as the common characteristic parameter of the aggregate, and the regulation priority of the flexible load aggregate in the smoothing control is determined to achieve hierarchical regulation. Finally, with the optimal economy as the goal and the power fluctuation rate as the coupling constraint, a distribution network power fluctuation smoothing control model is constructed based on the particle swarm algorithm, and the resource aggregate regulation quantity is optimized in real time to solve the problem of power fluctuation smoothing control of new energy grid connection. In combination with the above-mentioned technical solutions and the technical problems to be solved, the advantages and positive effects of the technical solution to be protected by the present invention are analyzed from the following aspects: 1. Make full use of the regulation potential of flexible load resources and give full play to the controllability of the distribution network containing multiple flexible load resources.

[0023] 2. Propose and adopt a clustering algorithm that considers the actual meaning of resource indicators to improve the feasibility of cluster control of flexible load resources.

[0024] 3. Based on the resource clustering results, the flexible load aggregates are hierarchically controlled, and the PSO algorithm is used for rolling optimization and solution to improve the smoothing effect, reduce the control cost, and speed up the solution. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0026] Figure 1 The flexible load resource active power leveling clustering index system provided by the embodiment of the present invention; Figure 2The flexible load resource active power leveling clustering index system provided by the embodiment of the present invention; Figure 3 is a flow chart of an improved K-means++ algorithm based on index weighting provided by an embodiment of the present invention; Figure 4 It is a minute-level AGC control framework model based on resource clustering and optimization algorithm provided by an embodiment of the present invention; Figure 5 It is a flow chart of solving an optimization control model based on a particle swarm algorithm provided by an embodiment of the present invention; Figure 6 It is a schematic diagram of an IEEE-33 distribution network system including variable frequency air conditioner flexible load resources provided by an embodiment of the present invention; Figure 7 is a curve diagram showing the relationship between clustering target and cluster number based on the elbow method provided in an embodiment of the present invention; Figure 8 It is a three-dimensional scatter plot of clustering results based on the improved K-means++ algorithm provided in an embodiment of the present invention; FIG. 9 ( a ) is a two-dimensional scatter plot of the clustering results of control accuracy and unit control cost provided by an embodiment of the present invention; FIG9( b ) is a two-dimensional scatter plot of adjustable capacity-control accuracy clustering results provided by an embodiment of the present invention; Fig.10 : is a comparison diagram of the power of the lower network point before and after the leveling control provided by the embodiment of the present invention; Fig.11 is a comparison diagram of power fluctuation rates before and after smoothing control provided by an embodiment of the present invention; Fig.12 is the total operating power and equivalent SOC of the flexible load cluster provided by the embodiment of the present invention; Fig.13 is a comparison chart of the outputs of different polymers provided by an embodiment of the present invention; Fig.14 It is a polymer control amount-control cost curve before and after economic optimization provided by an embodiment of the present invention; Fig.15 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0027] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0028] In order to meet the multiple requirements of accuracy, economy and reliability of flexible load response to rapid fluctuations in system state changes, this embodiment proposes a power fluctuation smoothing control method, which takes into account the effectiveness of power fluctuation smoothing and the economy of regulation, ensures that the flexible load can quickly and accurately respond to the power changes at the downstream grid point, realizes power fluctuation smoothing based on resource aggregation, and provides strong support for the safe operation and reliable power supply of the new power system.

[0029] This embodiment is implemented by the following technical solution, a power fluctuation smoothing control method, including: Establish multi-dimensional clustering indicators, and use the improved K-means++ algorithm based on indicator weighting to realize the clustering of massive distributed heterogeneous flexible load resources; build an aggregation evaluation model to evaluate the responsiveness and economy of resource aggregates, determine the aggregate regulation priority and realize hierarchical regulation; based on the particle swarm algorithm, with the optimal economy as the goal and the power fluctuation rate as the coupling constraint, build a distribution network power fluctuation smoothing control model, optimize the resource aggregate regulation quantity in real time, and solve the problem of distributed new energy grid-connected power fluctuation smoothing control.

[0030] When implementing it, Figure 1 As shown, an embodiment of the present invention provides a power fluctuation smoothing control method, comprising the following steps: S1, based on the power fluctuation smoothing scenario, establishes multi-dimensional clustering indicators, and uses the improved K-means++ algorithm based on indicator weighting to achieve clustering of massive distributed heterogeneous flexible load resources; S1.1, such as Figure 2 As shown, a resource active leveling clustering index system is established; and each index value in the index system is calculated; Set the controllability index and economic index as the first-level evaluation index of the resource active power leveling index system; Based on the controllability index, the adjustable capacity and the control accuracy are used as evaluation indicators to construct a second layer of evaluation indicators corresponding to the controllability index; Based on the economic index, taking the unit control cost as the evaluation index, constructing a second-level evaluation index corresponding to the economic index; Furthermore, the controllable capacity in the controllability index is divided into adjustable capacity and adjustable capacity according to the leveling control scenario. The adjustable capacity refers to increasing the output level of the source-side resources, and the adjustable capacity refers to reducing the load demand of the load-side resources. The calculation formula is:

[0031]

[0032] In the formula, , Respectively represent the maximum and minimum output of the flexible load; Represents the real-time operating power of the flexible load; The calculation formula of the control accuracy is:

[0033] In the formula, Respectively represent The output and actual output of the flexible load unit under the control strategy at each moment, is the total number of control periods, For adjustable flexible load capacity, T The duration of the effect of adjustable resources.

[0034] Based on the economic index, the unit control cost is used as the evaluation index to construct the second-level evaluation index corresponding to the economic index; The economic index is the unit control cost of different flexible load units, and the calculation formula is:

[0035] In the formula, for t The unit electricity price of the flexible load at any moment, is the actual control power of the flexible load control unit at time t; It is the duration of a single control period, i.e. 1 minute.

[0036] S1.2, calculate the weights of clustering indicators by combining subjective and objective factors; and dynamically adjust the weights based on the real-time grid smoothing demand; S1.2.1 Use the analytic hierarchy process to quantify the relative importance of each indicator in the evaluation system and determine the subjective weight of the relevant indicators; S1.2.1.1 Construct an evaluation index judgment matrix, which is:

[0037] in, Represents the number of indicators, Indicators and indicators The importance of comparing hour, ,when hour, ; S1.2.1.2Consistency check, judgment matrix The consistency ratio , defined as:

[0038] In the formula, is a consistency indicator, , is the maximum eigenvalue of the judgment matrix; For the order The associated constant represents the random consistency indicator; satisfying the condition Explain that the judgment matrix is ​​reasonable; S1.2.1.3 Calculate the subjective weight using the geometric mean method:

[0039] in, Indicates i The subjective weight of the evaluation index, For the The product of each element in the row square root, .

[0040] S1.2.2 Use the anti-entropy weight method to determine the objective weight of each evaluation indicator; S1.2.2.1 Standardize the evaluation indicators and calculate the first In the group data Indicators The proportion is:

[0041] In the formula, Represent the number of indicators and the number of evaluation objects, respectively. For the In group data index, for The proportion of S1.2.2.2 Calculate the The anti-entropy value of the indicator:

[0042] S1.2.2.3 Calculate the Entropy weight of each indicator:

[0043] In the formula, Indicates i The objective weight of an evaluation indicator.

[0044] S1.2.3 Calculate the comprehensive weight of the evaluation index based on the subjective weight and the objective weight using a linear weighting method; S1.2.3 Calculate the comprehensive weight using the linear superposition method. The calculation formula is:

[0045] in, is the subjective and objective weight distribution coefficient, Indicates i The comprehensive weight of the evaluation indicators.

[0046] S1.3, such as Figure 3 As shown in the figure, the flexible load resources are clustered based on the K-means++ algorithm improved based on the index weight; S1.3.1 Normalize the clustering index, and the calculation formula is:

[0047] In the formula, and are the normalized positive and inverse indicators, respectively. is the indicator data before normalization, , Respectively represent the minimum and maximum values ​​of the indicator data; S1.3.2 Determine the number of clusters K The elbow method is a simple and effective method to determine the number of clusters. K The core indicator of the method is the sum of squared error (SSE), that is:

[0048] In the formula, For the i clusters; p for The sample points in ; yes The center of mass ( is the mean of all samples in ); SSE is the clustering error of all samples, which represents the quality of clustering effect. SSE and k The relationship curve is like the shape of an elbow, and the elbow corresponds to k The value is the optimal number of clusters of the data; S1.3.3 Dynamic selection of initial cluster centers; S1.3.3.1 Randomly select a point as the initial aggregation center; S1.3.3.2 For each point in the sample, calculate its Euclidean distance to the nearest initial aggregation center D ( x ), the Euclidean distance D ( x ) is calculated as:

[0049] S1.3.3.3 Calculate the probability of each sample point being selected as the next cluster center; the calculation formula for the probability of being selected as the cluster center is:

[0050] in X is the sample set of all sample points; S1.3.3.4 Select the sample point corresponding to the maximum probability value as the next cluster center; S1.3.3.5 Repeat S1.3.3.2-S1.3.3.4 until the number of selected cluster centers reaches the optimal number of clusters K , the initial cluster center selection is completed.

[0051] S1.3.4 Dynamic weighting of clustering indicators; The weights reflecting the importance of the indicators are introduced into the K-means++ algorithm to obtain the weighted Euclidean distance, and the calculation formula is:

[0052] In the formula, u i Representative i The characteristic vector of the distribution unit, u k represents the selected cluster center feature vector, x ij Representative i The distribution unit feature vector j Cluster features, x kj The first character vector representing the cluster center j Cluster features, ω j The comprehensive weight coefficient corresponding to the representative indicator; S1.3.5 Execute K-means++ algorithm to cluster the active power leveling index; Calculate the weighted Euclidean distance from the feature value to the cluster center. The calculation formula is:

[0053] In the formula, Represents cluster m i The characteristic vector of the distribution unit, Represents the cluster center feature vector selected by m clusters, Represents cluster m i The distribution unit feature vector j Cluster features, represents the first feature vector of the cluster center selected by the m clusters j Clustering features.

[0054] Each feature vector is assigned to the cluster corresponding to the cluster center with the smallest Euclidean distance; in each cluster, the update formula of the cluster center is:

[0055] In the formula, represent The number of eigenvectors in the cluster; Repeat S1.3.5 until the average error criterion function converges and clustering is completed. The calculation formula of the average error criterion function is:

[0056] In the formula, Represents the cluster center feature vector selected by m clusters, Represents the updated cluster center feature vector of m clusters.

[0057] S2, based on the clustering results, build an aggregation evaluation model to evaluate the responsiveness and economy of resource aggregates, derive the aggregate regulation priority and realize hierarchical regulation; S2.1, hierarchical control of flexible load resources based on cluster center control priority; The K-means++ algorithm based on weight improvement divides the massive flexible load resources into K Clusters are formed, and the response capability index of each cluster center is used as the common characteristic parameter of the aggregate to determine the regulation priority of the flexible load aggregate in the smoothing control and realize hierarchical regulation. S2.2, Aggregate assessment model considering the aggregate response capability of flexible loads; Aggregate the adjustable performance of each cluster:

[0058]

[0059] In the formula, For polymer k The number of individual flexible loads included; It means t Moment Aggregate k The adjustable capacity; In addition, Represents an aggregate k The control accuracy.

[0060] Aggregate the unit control cost of each cluster:

[0061] In the formula, For polymer tThe aggregate control cost at the moment represents the sum of the control costs of each flexible load monomer in the aggregate.

[0062] The priority of resource regulation is determined according to the results of the aggregate evaluation model, and the regulation amount is sent to several distributed resources in the resource aggregate according to the priority order of resource regulation in the subsequent optimization control.

[0063] S3, based on the particle swarm algorithm, takes economic optimization as the goal and power fluctuation rate as the coupling constraint, builds a distribution network power fluctuation smoothing control model, optimizes the resource aggregate regulation quantity in real time, and solves the problem of power fluctuation smoothing control of distributed renewable energy grid connection.

[0064] S3.1, such as Figure 4 As shown, a minute-level AGC control framework model based on resource clustering and optimization algorithm is established; In the first stage of the control framework, the control demand judgment and resource grouping are completed. First, the power reference value of the downstream network is determined based on real-time data to generate the target value in the current control cycle; then, in the same control cycle, the power fluctuation rate of the downstream network is calculated according to the control step length, and it is used as the coupling constraint of the minute-level AGC optimization control and the power fluctuation smoothing target, and then the control demand of the distribution network in the current period is calculated. In each control cycle, the power of the downstream network at the first moment is selected as the control reference value, and the resources are grouped at the same time. The priority order of resource control is determined by the aggregation model, which serves as a reliable basis for the next stage of optimization control.

[0065] In the second stage of the control framework, according to the results of the first stage control demand judgment, the control amount is sent to several distributed resources in the resource aggregate according to the priority order of resource control, and the output size of each resource aggregate is used as the decision variable, taking into account the effectiveness of power fluctuation smoothing and control economy, and using the particle swarm algorithm to solve the optimization control model. After each optimization model solution is completed, it is determined whether the current control cycle is over based on the control step size and the number of cycle solutions, and then it is determined whether the clustering index needs to be updated and the resource clustering and optimization control of the next cycle needs to be performed.

[0066] S3.2, establish the economic objective function and constraint conditions of the optimization control model; S3.2.1 Based on the aggregate assessment model, the priority of resource regulation is determined and the economic optimization goals at the aggregate level are established:

[0067] S3.2.2 Flexible load operation constraints based on virtual energy storage, The charging and discharging power constraints and capacity constraints are the basic conditions that the flexible load aggregate needs to meet during actual operation:

[0068]

[0069] S3.2.3 Power balance constraints,

[0070] In the formula, For the Total distributed photovoltaic power at the moment; For the The wind farm output power at the moment; For the Power of other loads except flexible load at any time; For the Power loss at any time.

[0071] S3.2.4 Constraints on power fluctuation rate at lower grid points Power fluctuation rate at the lower network point , as an indicator to measure power change, reflects the difference between the power at the current moment and the power at the previous moment. Based on the specific structure and resource parameters of the distribution network, the constraints of the power fluctuation rate of the network point are defined, and the fluctuation rate is Control within the expected target volatility Inside.

[0072]

[0073] S3.2.5 Economic-based polymer decomposition and regulation strategy By solving the aggregate evaluation model, the total adjustable capacity of each resource aggregate is obtained. In order to quickly decompose the control amount, a decomposition strategy for aggregate control instructions is proposed under the premise of considering economic efficiency. This model abandons the traditional method of allocating capacity in proportion, and aims to quickly respond to the demand for smoothing the fluctuation of new energy power based on economic optimization. The objective function and constraints are shown in the following formula:

[0074] S3.3, such as Figure 5 As shown, an optimization control model based on particle swarm algorithm is constructed; S3.3.1 Initialize the particle swarm: First, set the number of particles in the particle swarm and randomly initialize the position, speed and fitness value for each particle, including: S3.3.1.1 Determine the variable dimension D of the optimal feasible solution based on the clustering results of the distributed resources in the distribution network, corresponding to the spatial dimension of the particles in the PSO, and set the number of particles in the population N; S3.3.1.2 Set the upper limit of the search area according to the aggregation index of the aggregate and lower bound , giving the maximum value of the particle search speed and minimum value , and the convergence accuracy ε or the maximum number of iterations , and take the value of the initial state of the particles within the constraints of position and velocity to complete the initialization of the position and velocity of the population particles; S3.3.1.3 At the beginning of the algorithm, the individual learning factors of the population , social learning factor and inertia weight ω Set up to update particle position and velocity.

[0075] S3.3.2 Calculate the fitness value: For each particle, calculate the fitness value according to its current position. The fitness function is designed according to the economic objective function of the optimization model to evaluate the quality of the particle in the solution space.

[0076] S3.3.3 Update individual optimal solution and global optimal solution: In each iteration, re-evaluate the fitness of each particle in the population and the corresponding objective function value, update and calculate the individual optimal solution and the population optimal solution to achieve the purpose of optimization; S3.3.4 Determine the termination condition: Check whether the termination condition is met, such as reaching the maximum number of iterations, the fitness value reaching the preset threshold, etc. If the termination condition is met, the global optimal solution is output; otherwise, return to step (3.3) to continue iterating, including: S3.3.4.1 After each iteration, the difference between the optimal particle fitness value of the population obtained in this iteration and the previous iteration is used to determine whether the optimization solution has converged. The convergence conditions are as follows:

[0077] In the formula, ε is the convergence accuracy; For the n The optimal fitness value of the population calculated by the iteration; S3.3.4.2 If the convergence condition is met or the maximum number of iterations has been reached , then the optimal particle of the current population corresponds to the optimal solution of the decision variable, realizing the optimization control process; if the convergence condition cannot be met, execute S3.3.5 to update the particle position and velocity.

[0078] S3.3.5 Update of particle position and speed and boundary processing: At the beginning of each iteration, the particle optimization speed and position are updated and calculated according to the following formulas based on the individual optimal fitness value and the population optimal fitness value obtained in the previous iteration:

[0079] In the formula, is the individual optimal solution of the particle; is the optimal solution for the particle population; is the individual learning factor, with a value of 0.5; is the social learning factor, with a value of 0.5; , is a random number in the range of [0,1], which increases the randomness of particle optimization; ω The weight of the particle to maintain its historical velocity is 0.9; After completing the update of particle velocity and position, check the update results according to the following formula:

[0080] In the formula, For particles i No. d dimensional position variable; For particles i No. d dimensional velocity variable; For particles that exceed the constraint range, a feasible solution is randomly generated within the constraint range to replace them.

[0081] This embodiment also provides a power fluctuation smoothing control system, including: Multidimensional clustering module, used to establish multidimensional clustering indicators; using the improved K-means++ algorithm based on indicator weighting to achieve clustering of dispersed heterogeneous flexible load resources; Aggregation evaluation module, which is used to build an aggregation evaluation model, evaluate the responsiveness and economic efficiency of resource aggregates, and derive the aggregate regulation priority to achieve hierarchical regulation; The smoothing control module is used to build a distribution network power fluctuation smoothing control model based on the particle swarm algorithm, with economic optimization as the goal and power fluctuation rate as the coupling constraint, optimize the resource aggregate control quantity in real time, and hierarchically control resources according to the control priority to achieve distributed new energy grid-connected power fluctuation smoothing control.

[0082] In addition, the multidimensional clustering module is specifically used to establish a resource active power smoothing clustering index system for the optimization goal served by clustering, and calculate the value of each index in the index system; calculate the clustering index weights based on subjective and objective factors, and dynamically adjust the weights based on the real-time grid smoothing demand; and realize the dynamic clustering of adjustable flexible load resources based on the improved K-means++ algorithm based on the index weights.

[0083] In addition, the aggregation evaluation module is specifically used to divide the massive flexible load resources into KClusters are formed, and the response capability index of each cluster center is used as the common characteristic parameter of the aggregate to determine the regulation priority of the flexible load aggregate in smoothing control and realize hierarchical regulation. An aggregation evaluation model considering the response capability of flexible load aggregation is established for each cluster, including the adjustable capacity, regulation accuracy and regulation cost of the cluster after clustering.

[0084] In addition, the smoothing control module is specifically used to establish a minute-level AGC control framework model based on resource clustering and optimization algorithm; to set up the economic objective function and constraint conditions of the optimization control model; and to establish an optimization control model based on particle swarm algorithm.

[0085] Example: Select as Figure 6 The IEEE-33 node distribution network system shown in the figure is taken as the research object, and the variable frequency air conditioning load cluster is connected at node 6 as a flexible load control resource, and its overall rated operating power is set to 800kW; in addition, the distributed photovoltaic power source is connected to node 17 and node 32 respectively, and the wind farm is connected to node 13. Thus, a distribution network topology structure including flexible load resources is constructed.

[0086] This embodiment takes the flexible load resources contained in the distribution network in a certain regulation cycle within a day as the analysis object, performs cluster analysis and aggregation evaluation on the flexible load regulation resources, and hierarchically regulates the resource aggregates to participate in power fluctuation smoothing control, including: 1) Normalization processing; The indicators selected in this embodiment involve positive indicators and inverse indicators, wherein the positive indicators include: adjustable capacity, regulation accuracy; the inverse indicators include: unit regulation cost. The above indicators are normalized respectively to obtain normalized data that can participate in clustering processing, as shown in Table 1: Table 1 Flexible load clustering index considering active power smoothing capability

[0087] 2) Selection of the optimal number of clusters; Calculate the clustering target when the number of clusters changes The value of The curve that changes with the number of clusters is used to obtain the optimal number of clusters for flexible load resources using the elbow method, such as Figure 7 As shown: Figure 7 The descending curve of the unitized clustering target as the number of clusters changes is described. When the number of clusters is 4, the slope of the descending curve of the clustering target reaches the maximum value, and the return brought by the increase in the number of clusters decreases rapidly. The number of clusters here is the optimal number of clusters. That is, in this flexible load cluster, the optimal number of clusters for 960 flexible load units is 4.

[0088] 3) Determine the weight of indicators; Since the normalized indicator data needs to conform to the actual meaning of the indicator when clustering, it is necessary to weight the indicator. In the weight configuration process, the present invention is based on the AHP-anti-entropy weight method, and the weights of the indicators are assigned in objective and subjective dimensions, so as to better meet the actual needs of smooth control. In this process, the importance of regulation accuracy is particularly emphasized, and it is placed in the first place, followed by economic indicators, while adjustable capacity is regarded as a relatively minor consideration. Subsequently, the judgment matrix of the indicator layer is specifically calculated. Finally, through the precise calculation of the AHP method, the indicator weight configuration results from a subjective perspective are obtained, and the specific data are shown in Table 2; combined with the objective weights calculated by the anti-entropy weight method, the comprehensive weight configuration results are obtained by linearly superimposing two weighting methods, as shown in Table 3.

[0089] Table 2 Clustering index subjective weight judgment matrix

[0090] Table 3 Results of comprehensive weight configuration of clustering indicators

[0091] As can be seen from Table 3, in terms of weight distribution, the control accuracy R Occupies the primary position, significantly higher than the unit control cost C And adjustable capacity α The weight of C The weight is slightly higher than the adjustable capacity α , which is in line with the priority considerations of the accuracy of flexible load regulation and the control of economic costs in power fluctuation smoothing control.

[0092] Proof of the relevant effects of the embodiments. The embodiments of the present invention have achieved some positive effects during the development or use process, and indeed have great advantages over the prior art. The following content is described in conjunction with the data, charts, etc. of the test process.

[0093] 4) Analysis of resource clustering effects; The Euclidean distance of K-means++ is optimized based on the index weight, so that the clustering effect is more in line with the demand for balancing. At the beginning of a certain control cycle, an aggregation analysis was conducted on 960 variable frequency air conditioner load resources. The results are as follows: Figure 8 As shown in Figure 9(a) and Figure 9(b). Figure 8 The three-dimensional distribution of the aggregation results is intuitively displayed. The 960 flexible load resources are divided into four different clusters, represented by red, blue, green and pink scatter points, respectively. At the same time, the circles of the corresponding colors also mark their respective cluster centers. Figure 9 (a) shows the two-dimensional scatter plot of the clustering results of regulation accuracy-unit regulation cost, and Figure 9 (b) shows the two-dimensional scatter plot of the clustering results of adjustable capacity-regulation accuracy.

[0094] The specific aggregate evaluation results of the flexible load resources in this regulation cycle are shown in Table 4: Table 4 Flexible load resource clustering results

[0095] As shown in Table 4, the ranking of flexible load aggregates considering controllability and economy is: ① Large controllable capacity, high control accuracy, and low unit control cost; ② Small controllable capacity, high control accuracy, and high unit control cost; ③ Large controllable capacity, low control accuracy, and high unit control cost; ④ Small controllable capacity, low control accuracy, and low unit control cost. Therefore, the priority ranking of each cluster when participating in active power leveling is: Aggregate 1> Aggregate 2> Aggregate 3> Aggregate 4. The control order of the aggregates is basically consistent with the control positive index representing controllability and the control inverse index representing economy. It can improve the economy and feasibility of hierarchical control of flexible load resources participating in active power leveling within a certain range, and further verify the effectiveness of clustering method for resource dimensionality reduction.

[0096] 5) Analysis of power fluctuation control effect; In this control period, the power of the down-grid point at the start time is used as the reference value for power leveling. , and based on this, the optimization control is carried out. The optimization control model is solved by particle swarm algorithm, and the power and fluctuation rate of the down-grid point before and after optimization are compared and analyzed. Fig.10 and Fig.11 As shown in the figure, the minute fluctuation rate of power at the downstream network points within the control period was kept below 2%, which not only met the requirements of the national standards, but also showed significant improvement compared with before control. The absolute value of the maximum fluctuation rate after control was 1.28%, which was 0.31% lower than before control, achieving good control effect.

[0097] 6) Analysis of polymer output effect; Fig.12 The total operating power and equivalent SOC of the variable frequency air conditioning cluster are shown. From the equivalent energy storage model, it can be seen that when the SOC is a positive value, it represents virtual energy storage charging, corresponding to the power increase scenario; when the SOC is a negative value, it represents energy storage discharge, corresponding to the power reduction scenario. Fig.13 The output curves of the four aggregates after hierarchical regulation are compared. In power fluctuation smoothing control, resource clustering aims to reduce the dimension of decision variables, evaluate the priority order of adjustable resources, and minimize the regulation cost. Fig.13It can be observed that in this control cycle, the control amount of aggregate 1 significantly exceeds that of other aggregates. At multiple control moments, the output of aggregates 2, 3, and 4 is almost zero, which is attributed to the small fluctuation of the power of the downstream grid at that moment, and the required total control amount does not exceed the adjustable capacity of aggregate 1, which indirectly reflects that aggregate 1 has a higher control priority; at a few control moments, the output of aggregate 4 is greater than that of aggregate 3, which is because the adjustable capacity of aggregate 3 is smaller than that of aggregate 4. When the required total control amount is greater than the sum of the adjustable capacities of the four aggregates, the phenomenon of aggregate 4 having more output than aggregate 3 will appear, which is also in line with the actual output of the aggregates following the control priority.

[0098] This embodiment focuses on the principle of economy and optimizes the resource control amount. By comparing the total economic cost of cluster optimization control before and after cluster optimization, the controllable amount-control cost curve of all clusters before and after optimization control is drawn, such as Fig.14 As shown. Fig.14 It can be clearly seen that compared with the optimization control, the total control cost required to control the same capacity of the aggregate has been reduced. In the figure, when the power control amount is 125 kW, the control cost is reduced from 58.2 yuan before optimization to 46.1 yuan after optimization.

[0099] Since the flexible load aggregation response model can accurately determine the regulation priority of flexible load resources and greatly reduce the variable dimension, the optimization control after resource clustering shows significant advantages in response speed and regulation cost.

[0100] This embodiment also provides a power fluctuation smoothing control system considering the flexible load aggregation response, including: a multi-dimensional clustering module, an aggregation evaluation module and a smoothing control module; The multidimensional clustering module is used to establish multidimensional clustering indicators; The aggregation evaluation module is used to build an aggregation evaluation model, evaluate the responsiveness and economic efficiency of resource aggregates, determine the priority of aggregate regulation and implement hierarchical regulation; The smoothing control module is used to construct a power fluctuation smoothing control model for the distribution network, optimize the resource aggregate regulation quantity in real time, and realize the power fluctuation smoothing control of distributed renewable energy grid-connected power.

[0101] Fig.15 An example of a physical structure diagram of an electronic device is shown in FIG. Fig.15 As shown, the electronic device may include: a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus. The processor may call the logic instructions in the memory to execute the power fluctuation smoothing control method.

[0102] In addition, the logic instructions in the above-mentioned memory can be implemented in the form of software functional units and can be stored in a computer-readable storage medium when sold or used as an independent product. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program code.

[0103] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the power fluctuation smoothing control system method considering the flexible load aggregation response provided by the above methods.

[0104] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0105] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0106] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A power fluctuation smoothing control method, characterized in that: include: Establish multi-dimensional clustering indexes and use the improved K-means++ algorithm based on index weighting to realize the clustering of dispersed heterogeneous flexible load resources; Based on the clustering results, an aggregation evaluation model is constructed to evaluate the responsiveness and economic efficiency of resource aggregates, and the aggregate regulation priority is obtained to achieve hierarchical regulation; Based on the particle swarm algorithm, with the optimal economy as the goal and the power fluctuation rate as the coupling constraint, a power fluctuation smoothing control model for the distribution network is constructed. The resource aggregate regulation quantity is optimized in real time, and resources are regulated in a hierarchical manner according to the regulation priority to achieve power fluctuation smoothing control of distributed renewable energy grid-connected.

2. The power fluctuation smoothing control method according to claim 1, characterized in that: The multi-dimensional clustering index is established, and the improved K-means++ algorithm based on index weighting is used to realize the clustering of dispersed heterogeneous flexible load resources, including: 1) According to the optimization goal served by clustering, establish a resource active leveling clustering index system and calculate the value of each index in the index system; 2) Calculate the weight of active power smoothing clustering index based on subjective and objective factors, and dynamically adjust the weight based on the real-time grid smoothing demand; 3) The improved K-means++ algorithm based on index weighting realizes dynamic clustering of adjustable flexible load resources.

3. The power fluctuation smoothing control method according to claim 1, characterized in that: Based on the clustering results, an aggregation evaluation model is constructed to evaluate the responsiveness and economic efficiency of resource aggregates, and the aggregate regulation priority is obtained to achieve hierarchical regulation, including: 1) Based on the weight-improved K-means++ algorithm, the flexible load resources are divided into K Clusters are formed, and the response capability index of each cluster center is used as the common characteristic parameter of the aggregate to determine the regulation priority of the flexible load aggregate in the smoothing control and realize hierarchical regulation. 2) For each cluster, an aggregate evaluation model is established that takes into account the aggregate response capability of flexible loads, including: Aggregate the adjustable performance of each cluster, In the formula, For polymer k The number of individual flexible loads included; It means t Moment Aggregate k The adjustable capacity available; Represents an aggregate k The control accuracy rate; Aggregate the unit control cost of each cluster, In the formula, For polymer t The aggregate control cost at the moment represents the sum of the control costs of each flexible load monomer in the aggregate; The priority of resource regulation is determined according to the results of the aggregate evaluation model, and the regulation amount is sent to several distributed resources in the resource aggregate according to the priority order of resource regulation in the subsequent optimization control.

4. The power fluctuation smoothing control method according to claim 1, characterized in that: The particle swarm algorithm is based on the goal of economic optimization and the power fluctuation rate is used as a coupling constraint to build a distribution network power fluctuation control model, optimize the resource aggregate control quantity in real time, and hierarchically control resources according to the control priority to achieve distributed new energy grid-connected power fluctuation control, including: 1) Establish a minute-level AGC control framework model based on resource clustering and optimization algorithm; 2) Establish the economic objective function and constraints of the optimization control model; 3) Establish an optimization control model based on particle swarm algorithm.

5. A power fluctuation smoothing control system, characterized in that: include: Multidimensional clustering module, used to establish multidimensional clustering indicators; using the improved K-means++ algorithm based on indicator weighting to achieve clustering of dispersed heterogeneous flexible load resources; Aggregation evaluation module, which is used to build an aggregation evaluation model, evaluate the responsiveness and economic efficiency of resource aggregates, and derive the aggregate regulation priority to achieve hierarchical regulation; The smoothing control module is used to build a distribution network power fluctuation smoothing control model based on the particle swarm algorithm, with economy as the goal and power fluctuation rate as the coupling constraint, to optimize the resource aggregate control quantity in real time, and to hierarchically control resources according to the control priority, so as to realize the power fluctuation smoothing control of distributed new energy grid-connected.

6. The power fluctuation smoothing control system according to claim 5, characterized in that: include: The multidimensional clustering module is specifically used to establish a resource active leveling clustering index system for the optimization goal served by clustering, and calculate the values ​​of each index in the index system; The clustering index weights are calculated based on subjective and objective factors, and the weights are dynamically adjusted based on the real-time grid smoothing demand. The dynamic clustering of adjustable flexible load resources is realized based on the improved K-means++ algorithm based on index weights.

7. The power fluctuation smoothing control system according to claim 5, characterized in that: include: The aggregation evaluation module is specifically used to divide the massive flexible load resources into K Clusters are formed, and the response capability index of each cluster center is used as the common characteristic parameter of the aggregate to determine the regulation priority of the flexible load aggregate in smoothing control and realize hierarchical regulation. An aggregation evaluation model considering the response capability of flexible load aggregation is established for each cluster, including the adjustable capacity, regulation accuracy and regulation cost of the cluster after clustering.

8. The power fluctuation smoothing control system according to claim 5, characterized in that: include: The smoothing control module is specifically used to establish a minute-level AGC control framework model based on resource clustering and optimization algorithm; to set up the economic objective function and constraint conditions of the optimization control model; and to establish an optimization control model based on particle swarm algorithm.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the power fluctuation smoothing control method according to any one of claims 1 to 4 is implemented.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the power fluctuation smoothing control method according to any one of claims 1 to 4 is implemented.

Images