Reactive resource aggregation method under multi-time scale voltage support criterion

Abstract

The invention discloses a reactive power resource aggregation method under a multi-time scale voltage support criterion, and belongs to the technical field of power system voltage stability control. The core innovation of the method lies in that a refined voltage support capability criterion system covering three different levels of long time, medium time and short time is constructed, and is used for accurately quantifying the dynamic regulation capability of heterogeneous reactive power resources such as an on-load tap changer (OLTC), a capacitor bank (CBs), a static var compensator (SVC) and a photovoltaic inverter. On the basis, the method introduces a DBSCAN clustering algorithm, combines the time scale criterion with a power grid topological structure, and performs comprehensive clustering analysis on reactive resource access nodes, thereby forming a hierarchical aggregation coordination strategy which realizes optimization in both time dimension and space dimension. According to the invention, the voltage stability, the reactive power resource utilization efficiency and the new energy consumption capability of the power distribution network under the high-proportion new energy access background are remarkably improved.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of power system steady-state analysis, and relates to a clustering method considering the weighted voltage support capability of photovoltaic access to a distribution network. BACKGROUND

[0002] With the rapid increase of the penetration rate of renewable energy, especially photovoltaic power generation in the distribution network, the voltage stability problem caused by high proportion of photovoltaic access is increasingly prominent. The intermittency and volatility of photovoltaic output lead to fundamental changes in the operation characteristics of the distribution network: when the irradiance changes rapidly, the second-to-minute-level fluctuation of photovoltaic power generation will cause node voltage to fluctuate dramatically, exceed the limit, and even cause voltage sag. When photovoltaic power is generated in a light load period, the reverse power flow phenomenon may change the voltage distribution pattern, making the action characteristics of traditional voltage regulation devices (such as OLTC) ineffective. More complex is that the voltage problems at different time scales are intertwined, and the traditional reactive power regulation method is difficult to achieve effective global optimization due to the single time scale and insufficient device coordination. Existing researches focus on single device or local area regulation strategies, lack of system description of "source-network" coupling characteristics, and do not fully consider the differences in response speed, regulation characteristics, etc. of OLTC, CBs, SVC and photovoltaic inverters, resulting in low voltage support efficiency of the system in complex conditions.

[0003] In view of the above challenges, the present research innovatively proposes a hierarchical aggregation scheme of multi-form reactive power resources based on multi-time scale voltage support capability criterion. By constructing a complete criterion system: the long-time scale criterion focuses on the support capability of OLTC and CBs for the basic voltage, the medium-time scale criterion quantifies the suppression effect of SVC on voltage fluctuation, and the short-time scale criterion represents the rapid response characteristics of photovoltaic inverters, and combines the DBSCAN clustering algorithm to find a reasonable multi-form reactive power resource aggregation strategy. This "criterion quantization + intelligent clustering" method realizes the whole-cycle voltage optimization from planning to control, and provides a systematic theoretical basis and technical support for solving the stable operation of high proportion of photovoltaic distribution network. SUMMARY

[0004] The purpose of the present application is to overcome the problems existing in the prior art, and to provide a more practical clustering decision, a hierarchical aggregation method of multi-form reactive power resources based on multi-time scale voltage support capability criterion.

[0005] To achieve the above technical purpose, the technical scheme adopted by the present application is as follows:

[0006] A reactive power resource aggregation method under a multi-time scale voltage support criterion, the method comprising the following steps:

[0007] 1. Based on the difference of the operating characteristics of reactive power resources in different time scales, a multi-time scale voltage support capability criterion can be constructed. The long time scale criterion is obtained by multiplying the voltage safety margin normalization term and the static reactive power reserve ratio term.

[0008] The long time scale voltage support capability criterion can be divided into two parts, namely the voltage term and the reactive power term. The voltage term reflects the operating conditions of the voltage level of the distribution network region in the long time scale, and the reactive power term reflects the reactive power support capability of the on-load voltage regulating transformer and the capacitor bank connected to the distribution network. The specific voltage term formula is as follows:

[0009]

[0010] In the formula, is the voltage of node , is the minimum value of the node voltage, is the reference value of the node voltage.

[0011] The formula converts the complex voltage stability problem into an intuitive quantitative evaluation through mathematical construction. The core idea is to evaluate the voltage safety margin of the system by calculating the relative distance between each node voltage and the minimum allowed voltage.

[0012] The reactive power support term is a judgment of the regional reactive power support capability, which is used to evaluate the coverage capability of the system reactive power compensation resources to the load demand. This index directly reflects the reactive power balance state of the power grid by quantitatively comparing the relationship between the available reactive power capacity and the system demand. The specific reactive power term formula is as follows:

[0013]

[0014] In the formula, , are the installation nodes of the capacitor bank and the on-load voltage regulating transformer, , are the capacities of the capacitor bank and the on-load voltage regulating transformer compensation equipment, is the regional reactive power load demand.

[0015] The formula quantifies the coverage degree of the static reactive power resources to the system demand. When the ratio is greater than 1, it indicates that the configured reactive power capacity of the system is sufficient; when it is less than 1, it indicates that there may be a risk of reactive power shortage. This index is directly related to the static reactive power balance of the system, and provides a decision basis for the operation personnel to make reactive power compensation strategies.

[0016] The numerator part The available capacity of traditional reactive power compensation equipment is represented, in which the capacitor bank capacity represents discrete reactive power compensation capability, and the output size thereof varies with the square of the operating voltage; and the OLTC reactive power regulation capability represents continuous adjustable reactive power support characteristics, and the denominator part reflects the system reactive load demand.

[0017] Therefore, the long-time-scale voltage support capability index criterion considers the voltage item and the reactive power item at the same time. First, the voltage safety margin of each node is calculated, and the minimum value of the whole network is taken to reflect the weakest link. At the same time, the ratio of the total capacity of traditional reactive power equipment to the system demand is calculated, and finally the two items are multiplied to obtain the comprehensive criterion. Through the coupling calculation of the voltage item and the reactive power item, the specific formula is as follows:

[0018]

[0019] 2. The medium-time-scale voltage support capability criterion considers the reactive power capacity and the voltage fluctuation degree of the distribution network. Voltage stability depends on both the current voltage level and the dynamic reactive power regulation capability of the system, and both are indispensable. Coupling calculation precisely describes this dual dependence through mathematical form.

[0020] The specific formula is as follows:

[0021]

[0022] In the formula, represents the maximum voltage fluctuation rate of the nodes in the distribution network, is a penalty coefficient (0.15), represents the reactive power capacity of the SVC, is the regional reactive load demand, represents the node of the distribution network to which the SVC is connected.

[0023] Among them, and The ratio of the two is consistent with the reactive power item effect of the long-time-scale voltage support capability criterion, and quantifies the coverage degree of the static reactive power compensator to the system demand. When the ratio is greater than 1, it indicates that the configured reactive power capacity of the system is sufficient; and when the ratio is less than 1, it indicates that there may be a risk of reactive power shortage. This index is directly related to the static reactive power balance of the system, and provides a decision basis for the operation personnel to make reactive power compensation strategies.

[0024] The formula is as follows: wherein, the voltage fluctuation rate refers to the fluctuation intensity of the grid node voltage in a certain time period, the maximum voltage fluctuation rate refers to the fluctuation value of the node with the most intense voltage fluctuation selected from the nodes, that is, the key position with the most unstable voltage in the system. The criterion is dynamically adjusted by the voltage fluctuation rate. The standard deviation of the voltage is calculated to reflect the stability of the system voltage. When the voltage fluctuation intensifies, the value decreases, thereby reducing the overall criterion value and timely reflecting the system risk.

[0025] The formula of the maximum voltage fluctuation rate is as follows:

[0026]

[0027]

[0028] wherein, is the number of nodes, is the time window (the sampling time point is 15 minutes, so 96 is taken), is the voltage sampling value of the i-th node in the k-th 15 minutes, is the average value of the voltage of the node i in a time length.

[0029] 3. The short time scale criterion is based on the combination of real-time reactive power regulation margin and voltage-reactive power sensitivity information, and the specific function is as follows:

[0030]

[0031] wherein, Ω C is a set of candidate nodes for installing photovoltaic power stations, represents the real-time reactive power regulation margin of the photovoltaic power station installed at the node j, and the device can additionally provide reactive power support.

[0032] The formula of the real-time reactive power regulation margin is as follows

[0033]

[0034]

[0035] represents the rated capacity of the photovoltaic power station installed at the node j, , is the active power and reactive power output by the photovoltaic inverter of the node j.

[0036] Based on the expansion of the power flow equation near the system operating point, the numerical relationship among the active power deviation, the reactive power deviation, the voltage amplitude deviation and the voltage phase angle deviation, that is, the Jacobian matrix, can be obtained, as follows:

[0037]

[0038] ∆P and ∆Q are the active and reactive power variation matrices injected at the node, respectively; ∆θ and ∆U are the voltage phase angle magnitude variation matrices at the node, respectively. Inverting the above equation yields...

[0039]

[0040] Among them, sensitivity factor Each value in the equation corresponds to a set of differential relations. .

[0041] This criterion, by combining real-time reactive power regulation margin and sensitivity analysis, intuitively reflects the average voltage boost effect that each unit of reactive power resources in the current system can generate, and can accurately determine whether the existing reactive power reserves are distributed at highly sensitive nodes.

[0042] 4. The DBSCAN clustering method is used for multi-morphological reactive resource clustering. The specific process is as follows:

[0043] (1) DBSCAN starts with an arbitrary starting data point that has never been visited. The neighborhood of this point is extracted using distance ε (all points within ε are neighbors).

[0044] (2) If there are enough points in this neighborhood, the clustering process begins, and the current data point becomes the first point in the new cluster. Otherwise, the point will be marked as noise (this noise point may later become part of a cluster). In both cases, the point is marked as "visited".

[0045] (3) For the first point in the new cluster, the points near its ε distance will also become part of the same cluster. This process makes all points in the ε neighborhood belong to the same cluster, and then repeats for all the new points that have just been added to the cluster group.

[0046] (4) The process of steps (2) and (3) will be repeated until all points in the cluster are determined, that is, all points near the cluster have been visited and labeled.

[0047] (5) Once we have completed the current clustering, we will retrieve and process a new unvisited point, which will lead to the discovery of further clustering or noise. This process is repeated until all points are marked as visited. This is because after all points have been visited, each point is marked as belonging to a cluster or noise. Attached Figure Description

[0048] Figure 1This is a flowchart illustrating the present invention;

[0049] Figure 2 This is a diagram of the IEEE 33-node system;

[0050] Figures 3-5 These are cluster diagrams for reactive power resources of OLTC and CBs, SVC, and photovoltaic inverters, respectively. Detailed Implementation

[0051] The technical solution of the present invention will be described in detail below with reference to the embodiments and the accompanying drawings.

[0052] Example 1: The IEEE 33-node system is selected as the implementation case. Multi-mode reactive power resources are set up on the nodes. The specific layout scheme is as follows.

[0053] The OLTC access nodes are: [2,5,7,10,13,15,17,21,23,24,26,28,30,32];

[0054] The CBs access nodes are: [4,6,9,11,16,18,22,25]

[0055] The SVC access nodes are: [2,5,7,10,13,15,17,21,23,24,26,28,30,32]

[0056] The access nodes for the photovoltaic inverters are: [3,6,8,9,11,12,14,16,22,25,31]

[0057] The specific steps are as follows:

[0058] 1) Calculate long-time scale criteria based on voltage safety margin and reactive power capacity coverage to reflect the static stability of the system. That is, construct a feature matrix under long-time scale based on the long-time voltage support capability criteria index to provide relevant data for DBSCAN clustering. The specific formula is as follows:

[0059]

[0060] 2) Construct a mid-time-scale criterion using voltage fluctuation rate and SVC reactive power support capability to capture minute-level dynamic characteristics. Specifically, based on the mid-time-scale voltage support capability criterion index, construct a feature matrix at the mid-time scale to provide relevant data for DBSCAN clustering. The specific formula is shown below:

[0061]

[0062] 3) Combining the real-time adjustment reactive power margin and the short-timescale criteria derived from voltage-reactive power sensitivity analysis, the second-level rapid response capability is evaluated. That is, based on the short-timescale voltage support capability criteria, a feature matrix under the short-timescale is constructed to provide relevant data for DBSCAN clustering. The specific formula is shown below:

[0063]

[0064] 4) Inputting multidimensional features into the DBSCAN algorithm, density clustering is used to identify reactive resources with similar characteristics, forming a control strategy of "partition autonomy and global coordination". The partitioning results are shown below:

[0065] Over a long time scale, a series of calculations showed that OLTC and CBs were ultimately clustered into 2 clusters.

[0066] Cluster 1: 9 nodes, clustering results are shown in the table below:

[0067] Table 3-5 Clustering results for Cluster 1

[0068] Reactive resource type Node OLTC 10，13，15，17，21 CBs 11，16，18，22

[0069] Cluster 2: 13 nodes, clustering details are shown in the table below.

[0070] Table 3-6 Cluster 2 Clustering Results

[0071] Reactive resource type Node OLTC 2，5，7，23，24，26，28，30，32 CBs 4，6，9，25

[0072] At a medium timescale, after a series of calculations, the final clustering result of SVC was 3 clusters.

[0073] The specific results are shown in the table below:

[0074] Table 3-7 SVC Clustering Results

[0075] Cluster Node Cluster 1 10，13，15，17，32 Cluster 2 2，21，23，24 Cluster 3 5，7，26，28，30

[0076] On a short timescale, after a series of calculations, the photovoltaic inverters were finally clustered into 2 clusters.

[0077] The specific results are shown in the table below:

[0078] Table 3-8 Clustering Results of Photovoltaic Inverters

[0079] Cluster Node Cluster 1 3，6，8，22，25 Cluster 2 9，11，12，14，16，31

[0080] Based on the IEEE 33-bus system, this study proposes a hierarchical aggregation strategy for multi-form reactive power resources to address reactive power regulation needs at different time scales in power systems. This strategy fully considers the dynamic response characteristics of reactive power equipment, dividing reactive power resources into three levels according to time scale: long-term, medium-term, and short-term. OLTCs, CBs, SVCs, and photovoltaic inverters are used for coordinated control in these levels, respectively. At the long-term time scale, OLTCs and CBs, due to their slower regulation speed, are used for supporting the system base voltage. At the medium-term time scale, SVCs participate in dynamic reactive power compensation due to their fast response capability. At the short-term time scale, photovoltaic inverters utilize their millisecond-level response characteristics to provide instantaneous reactive power regulation.

[0081] To optimize the allocation and coordinated operation of reactive power resources, this study employs the DBSCAN clustering method, combined with voltage support capability criteria, to perform cluster analysis on the access nodes of various reactive power devices. On a long timescale, OLTC and CBs nodes are divided into two clusters, containing 9 and 13 nodes respectively, reflecting their voltage regulation roles in different regions of the system. On a medium timescale, SVC nodes are further subdivided into three clusters to accommodate dynamic reactive power demand changes in different regions. On a short timescale, photovoltaic inverter nodes are divided into two clusters to optimize their rapid reactive power response capability.

[0082] The above are merely specific embodiments of the present invention. Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention patent application and not to limit its protection scope. Although the present invention patent application has been described in detail with reference to the above embodiments, those skilled in the art should understand that after reading the present invention patent application, they can still make various changes, modifications or equivalent substitutions to the specific embodiments of the application, but these changes, modifications or equivalent substitutions are all within the protection scope of the claims pending approval.

Claims

1. A method for reactive resource aggregation under multi-time scale voltage support criterion, characterized in that, The method comprises the following steps: 1) combing the difference of the operation characteristics of the multi-form reactive resources at different time scales to determine the reactive resources adapted at different time scales; 2) based on the reactive-voltage interaction mechanism, that is, considering the correlation between voltage and reactive power, a voltage support capability criterion of multi-time scale of multi-form reactive resources is constructed; 3) based on the voltage support capability criterion of multi-time scale, a feature matrix is constructed, and a DBSCAN clustering algorithm is used to cluster and divide the same kind of reactive resources connected to the distribution network, that is, to cluster and analyze each kind of reactive resources connected to the distribution network.

2. The method of claim 1, wherein, In the step 2), based on the difference of the operation characteristics of the reactive resources at different time scales, the voltage support capability criterion of multi-time scale can be constructed, and the long-time scale criterion is obtained by multiplying the voltage safety margin normalization term and the static reactive reserve ratio term. The specific voltage term formula is as follows: In the formula, For nodes voltage, This represents the minimum node voltage. This is a reference value for the node voltage. The specific reactive term formula is as follows: wherein , are capacitor bank and on-load tap changer installation nodes, respectively, , is the capacity of the capacitor bank and on-load tap changer compensation device, is the regional reactive load demand. The coupling calculation of the voltage term and the reactive term is as follows:

3. The method of claim 1, wherein, In the step 2), the medium-time scale criterion is composed of the node maximum voltage fluctuation rate penalty term and the SVC capacity ratio, and the specific function is as follows: The specific formula is as follows: wherein, represents the maximum voltage fluctuation rate of the node in the distribution network, is a penalty coefficient (0.15), represents the reactive power capacity of the SVC, is the regional reactive power load demand, represents the node in the distribution network where the SVC is connected. The formula of the maximum voltage fluctuation rate is as follows: In the formula, N is the number of nodes, T is the time window (the sampling time point is 15 minutes, so take 96), Vi(k) is the voltage sampling value of the i-th node in the k-th 15 minutes, Vi is the voltage average value of node i in a period of time.

4. The method of claim 1, wherein, In the step 2), the short-time scale criterion is based on the combination of reactive margin and voltage-reactive sensitivity information, and the specific function is as follows: wherein Ω C is a set of candidate nodes for installing photovoltaic power stations, represents the real-time reactive power regulation margin of the photovoltaic power station installed at node j, and the amount of additional reactive power support that the device can provide. representative of the rated capacity of the photovoltaic power plant installed at node j, , Pj, G, Qj, active and reactive power output by the photovoltaic inverter for node j. And based on the expansion of the power flow equation near the system operating point, the numerical relationship among the active power deviation, the reactive power deviation, the voltage amplitude deviation and the voltage phase angle deviation, that is, the Jacobian matrix, can be obtained, as follows: And, the inverse of the above formula can be obtained as follows: wherein each of the numerical values in the sensitivity factor corresponds to a set of differential relationships .

5. The method of claim 1, wherein, In the step 3), the DBSCAN clustering algorithm is used to cluster and analyze the voltage support capability criterion, and the multi-form reactive resources connected to the distribution network are clustered, and the same kind of reactive resources is divided into multiple sets, wherein the specific steps of the DBSCAN algorithm are as follows: (1) DBSCAN starts with an arbitrary starting data point that has not been visited. The neighborhood of this point is extracted with a distance ε (all points within ε distance are neighbors). (2) If there are enough points in this neighborhood, the clustering process begins, and the current data point becomes the first point in the new cluster. Otherwise, the point will be marked as noise (this noise point may become part of a cluster later). In both cases, this point is marked as "visited". (3) For the first point in the new cluster, the points within ε distance of it also become part of the same cluster. This process makes all points within ε neighborhood belong to the same cluster, and then repeats for all new points just added to the cluster group. (4) Steps 2 and 3 will be repeated until all points in the cluster are determined, that is, all points within the cluster neighborhood have been visited and marked. (5) Once we have completed the current clustering, we retrieve and process a new unvisited point, which will lead to further clustering or noise discovery. This process is repeated until all points are marked as visited. Because after all points are visited, each point is marked as belonging to a cluster or as noise.

Images