A method for source-grid-load coordinated expansion planning considering node inertia vulnerability assessment

Abstract

The present invention relates to a source-grid-load coordinated expansion planning method that considers node inertia vulnerability assessment. The method comprises the following steps: obtaining annual load and renewable energy output data, clustering the data to form a representative daily scenario dataset; constructing a node RoCoF calculation model that takes into account the probability distribution of disturbance sources, determining the node RoCoF over-limit probability and expectation, and constructing a node inertia vulnerability assessment model; determining the participating resources on each side of the source-grid-load and their basic planning constraints, and constructing a source-grid-load coordinated expansion planning model; using the representative daily scenario dataset as input to the source-grid-load coordinated expansion planning model, determining inertia resource construction constraints and line construction constraints, performing an assessment using the node inertia vulnerability assessment model, and iteratively optimizing and solving a planning solution based on the assessment results until an optimal source-grid-load coordinated expansion planning solution is output. Compared with existing technologies, the present invention has the advantages of comprehensive considerations and meeting the needs of power system development.

Description

Technical Field

[0001] The present invention relates to the field of power system planning, and in particular to a source-grid-load coordinated expansion planning method considering node inertia weakness evaluation. Background Art

[0002] In recent years, the frequency security of power systems has become increasingly prominent under the high proportion of renewable energy access. The rapid frequency drop in the early stages of a frequency event due to insufficient or irrational distribution of inertia resources is a major factor affecting frequency security. Existing planning and optimization methods that take frequency security into account mostly use centralized modeling, ignoring the spatial differences in system inertia response. When a large disturbance occurs in the power system, the rate of frequency drop in the early stages of the disturbance is closely related to the system inertia distribution. The distribution of inertia resources and the network line structure both affect the distributed inertia response of the power system. Therefore, it is necessary to construct a node inertia vulnerability assessment method to describe the distributed inertia support capacity of the power system.

[0003] The inertia response of a power system is not only limited by the power source structure and unit capacity, but also by transmission capacity and network topology, which restrict power source startup methods and indirectly affect the system's inertia support capabilities. Load-side rotating devices and virtual inertia devices directly affect the system's inertia level. Existing planning models fail to consider the collaborative contribution of source, grid, and load resources to improving the system's inertia response capability, nor do they consider the probability distribution of disturbance sources. They often use the N-1 principle to determine the magnitude of disturbances, which is incomplete. Summary of the Invention

[0004] The purpose of the present invention is to provide a source-grid-load coordinated expansion planning method that considers the evaluation of node inertia weakness, which is more comprehensive and the planning scheme is more economical and more practical.

[0005] The purpose of the present invention can be achieved by the following technical solutions:

[0006] A source-grid-load coordinated expansion planning method considering node inertia weakness assessment includes the following steps:

[0007] Step 1) Obtain annual load and renewable energy output data, and cluster the data using an improved load-renewable energy scenario reduction ordered clustering model based on comprehensive class diameter to form a representative daily scenario dataset;

[0008] Step 2) Considering the uncertainty of the disturbance source, a node RoCoF calculation model is constructed that takes into account the probability distribution of the disturbance source;

[0009] Step 3) Determine the node RoCoF exceeding probability and node RoCoF expectation based on the node RoCoF calculation model, and construct a node inertia weakness assessment model;

[0010] Step 4) Determine the participating resources on each side of the source, grid, and load and their basic planning constraints, and build a source, grid, and load coordinated expansion planning model;

[0011] Step 5) Using the representative daily scenario dataset as input to the source-grid-load coordinated expansion planning model, the inertia resource construction constraints and line construction constraints are determined. The nodes of the planning scheme are evaluated using the node inertia vulnerability assessment model. Based on the evaluation results, the planning scheme is iteratively optimized and solved until the optimal source-grid-load coordinated expansion planning scheme is output.

[0012] The step 1) comprises the following steps:

[0013] Step 1-1) Obtain annual load and renewable energy output data and set the cut point range;

[0014] Step 1-2) Generate the optimal load shedding scheme using an improved load-new energy scenario reduction ordered clustering model based on comprehensive class diameter;

[0015] Steps 1-3) Generate a preliminary set of new energy scenarios based on the optimal load shedding solution;

[0016] Steps 1-4) Calculate the distance between the new energy output in the scene set and the original centroid at different segmentation points;

[0017] Step 1-5) Group the output data according to the distance from the centroid and recalculate the centroid based on the grouped data;

[0018] Step 1-6) determines whether the centroid distance is less than the pre-configured threshold. If so, the final clustering result of the new energy is determined based on the grouped output data and load data to form a representative daily scene data set. If not, the centroid is updated and returns to step 1-5).

[0019] The improved load-new energy scenario reduction ordered clustering model based on comprehensive class diameter is:

[0020]

[0021] In the formula, n ordered samples are divided into k categories, and the G-th sample group includes samples {X i ,X i+1 ,…,X t ,…X j}; Represents the sample mean corresponding to the G-th sample group; represents the sample mean trend corresponding to the G-th sample group; D(i,j) represents the comprehensive class diameter corresponding to the G-th sample group; L[(b(n,k))] represents a method of dividing n sample data into k categories; p(n,k) represents an optimal solution that minimizes L[(b(n,k))] under the aforementioned classification method.

[0022] The probability distribution of the disturbance source is expressed as:

[0023]

[0024] Where: Refers to the working condition of component k1; 0 represents normal operation; 1 represents a fault; Indicates the probability of failure; R k1 is the random number drawn for element k1.

[0025] The node RoCoF calculation model taking into account the probability distribution of disturbance sources is:

[0026]

[0027] Where ΔP i Indicates the disturbance power allocated to inertia node i; M ik is the synchronous power coefficient between inertia node i and disturbance node k; E i is the voltage of inertia node i; U k is the voltage at the disturbance point k; B ik and δ ik are the susceptance and initial angle difference between inertia node i and disturbance node k respectively; Represents the maximum value of discrete distribution fault disturbance source k1; represents the maximum value of the continuous distribution fault disturbance source k2; m1 represents the number of discrete distribution fault disturbance sources; m2 represents the number of continuous distribution fault disturbance sources; K ik represents the power distribution coefficient between inertia node i and disturbance node k.

[0028] The node inertia weakness assessment model is:

[0029]

[0030] Where, γ i represents the RoCof exceeding limit probability of node i; δ i represents the expected value of RoCoF of node i; w γ 、w δ , represent the weight coefficients of RoCoF crossing probability and RoCoF expected value respectively; Indicates the number of times the node iRoCoF exceeds the critical value; represents the total number of disturbance simulations; J i and B j Indicates the number and length of intervals for classifying RoCoF; R mj Represents the median RoCoF value in interval j.

[0031] The optimization objective function of the source-grid-load coordinated expansion planning model is:

[0032]

[0033] Where t, k, and h are the indexes of the planning year, scenario day, and hour within a day, respectively; g, r, e, and l are the indexes of the thermal power units to be built, the small thermal power units to be transformed, the energy storage devices to be built, and the lines to be built, respectively; Ω represents the set corresponding to each resource; C inv and is the overall cost and equivalent annual cost of the components to be constructed or renovated; t represents the current market value coefficient; d rate represents the discount rate; p k d k Indicates the probability of scenario k occurring and the corresponding number of days; Respectively represent the hourly output of thermal power units, energy storage devices, and wind turbines; Represents the predicted output of wind turbine; y t A 0-1 variable indicating whether the component is under construction; u tkh A 0-1 variable indicating the start and stop status of the device every hour.

[0034] The basic constraints of the source-grid-load coordinated expansion planning model are:

[0035]

[0036] Where m and s represent the indexes of the sending node and the receiving node respectively; d represents the index of the original device of the node; X l is the reactance of line l; M is a large constant value; P l max is the upper limit of the power flow of line l; EL represents the set of existing lines; CL represents the set of candidate lines; and are the phase angles of the sending and receiving nodes of line l respectively; and Represent the phase angle of node b and its upper and lower limits respectively; They represent the upper and lower output limits of unit g respectively; EG and CG represent the collection of existing generator sets and newly built generator sets respectively; Represents the predicted output of wind farm w; M ceh is the charging power multiple of the energy storage system e in time period h; and M ceh Upper and lower limits of M deh is the discharge power multiple of the energy storage device system e in time period h; and Mdeh Upper and lower limits of E eh is the amount of electricity of the energy storage system e in time period h; and E eh Upper and lower limits of E e0 and E eT are the electricity consumption of the energy storage system e at the beginning and end of the scene respectively.

[0037] The inertia resource construction constraints are:

[0038]

[0039] Where, Ω i represents the set of inertia resources to be built or modified at node i; Ω g Represents the set of inertia resources to be built or modified for all nodes; They are the inertia time constants of the power generating unit to be built, the small thermal power unit to be transformed, and the energy storage device to be built;

[0040] The line construction constraints are:

[0041]

[0042] Where, ξ t,max , χ t,max They represent the maximum electrical coupling degree and the maximum electrical distance from the disturbance source of all nodes in the system in the tth planning year; Z ii , Z jj Represent the self-impedance of node i and node j respectively; Z ij represents the mutual impedance between node j and node j.

[0043] The step 5) comprises the following steps:

[0044] Step 5-1) Use the representative daily scenario dataset as input to the source-grid-load coordinated expansion planning model, along with other initial information;

[0045] Step 5-2) Determine the inertia resource construction constraints and line construction constraints,

[0046] Step 5-3) Based on the initial information and the control factors in the node inertia weakness assessment model, solve the system expansion planning scheme while satisfying the constraints of the source-grid-load coordinated expansion planning model, and update the system network information and inertia resource information;

[0047] Step 5-4) Evaluate the nodes of the planning scheme using the node inertia weakness assessment model;

[0048] Step 5-5) determines whether all nodes meet the evaluation requirements. If so, proceed to step 5-6). Otherwise, increase the control factor for the nodes that do not meet the evaluation requirements and return to step 5-3) to re-solve the system expansion plan.

[0049] Step 5-6) determines whether the number of iterations reaches the pre-configured value. If so, go to step 5-7). Otherwise, record the solution result to the feasible solution database, increase the control factor, and return to step 5-3) to re-solve the system expansion plan.

[0050] Steps 5-7) Extract the optimal solution from the feasible solution library and output the optimal source-grid-load coordinated expansion planning solution.

[0051] Compared with the prior art, the present invention has the following beneficial effects:

[0052] (1) The present invention adopts a load-new energy scenario reduction method based on comprehensive class diameter, which can reduce the number of clustering scenarios while ensuring the clustering effect and reduce the calculation scale in subsequent planning;

[0053] (2) The present invention considers the uncertainty of the disturbance source and can account for the random factors in the occurrence of disturbances, making the simulation results more realistic and more suitable for application scenarios with a high proportion of new energy access;

[0054] (3) The present invention evaluates the weakness of node inertia, can fully consider the uncertainty disturbance source, the influence of network topology on the node inertia support capacity, can quantitatively characterize the distributed inertia response level of the system, and can be applied to the distributed inertia support capacity evaluation of different planning schemes.

[0055] (4) The present invention embeds the node inertia vulnerability assessment into the source-grid-load collaborative planning model, which can collaboratively optimize the inertia resources on each side of the source-grid-load, improve the distributed inertia response capability of the planning scheme, and reduce the RoCoF over-limit risk under large disturbances. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 This is the overall architecture diagram of the present invention;

[0057] Figure 2 Reduce the flow chart for the load-renewable energy scenario;

[0058] Figure 3 This is a flow chart for source-grid-load collaborative planning based on node inertia weakness assessment;

[0059] Figure 4 is the load clustering result of Example 1.3;

[0060] Figure 5 is the RoCoF probability distribution of the system as a whole and nodes;

[0061] Figure 6 Fitting a heat map for the inertia weakness distribution of system nodes;

[0062] Figure 7 Comparison of the node inertia weakness assessment results for Examples 3.1 and 3.2;

[0063] Figure 8 Comparison of planning results for Examples 3.1, 3.2, and 3.3;

[0064] Figure 9 This is a comparison of the investment and construction costs and penalty cost results for Examples 4.1, 4.2, and 4.3. DETAILED DESCRIPTION

[0065] The present invention is described in detail below with reference to the accompanying drawings and specific embodiments. This embodiment is implemented based on the technical solution of the present invention, and provides a detailed implementation method and specific operation process, but the protection scope of the present invention is not limited to the following embodiments.

[0066] This embodiment provides a source-grid-load coordinated expansion planning method considering node inertia weakness assessment, such as Figure 1 As shown, the following steps are included:

[0067] Step 1) Obtain annual load and renewable energy output data, and cluster the data using the improved load-renewable energy scenario reduction ordered clustering model based on comprehensive class diameter to form a representative daily scenario dataset. The process is as follows: Figure 2 As shown, the specific steps include:

[0068] Step 1-1) Obtain annual load and renewable energy output data and set the cut point range.

[0069] Step 1-2) uses an improved load-new energy scenario reduction ordered clustering model based on comprehensive class diameter to generate the optimal load cutting scheme.

[0070] Step 1-2-1) Taking the minimum comprehensive diameter as the optimization goal, calculate the optimal cutting position under the current cutting point;

[0071] Step 1-2-2) Determine whether the number of cut points has reached the maximum. If so, calculate the corresponding silhouette coefficient value of each scheme, and compare and determine the optimal load cutting scheme under different cut points. Otherwise, add 1 to the number of cut points and return to step 1-2-1).

[0072] Specifically, the improved load-new energy scenario reduction ordered clustering model based on comprehensive class diameter is:

[0073]

[0074] In the formula, n ordered samples are divided into k categories, and the G-th sample group includes samples {X i ,X i+1 ,…,X t ,…X j}; Represents the sample mean corresponding to the G-th sample group; represents the sample mean trend corresponding to the G-th sample group; D(i,j) represents the comprehensive class diameter corresponding to the G-th sample group; L[(b(n,k))] represents a method of dividing n sample data into k categories; p(n,k) represents an optimal solution that minimizes L[(b(n,k))] under the aforementioned classification method.

[0075] Steps 1-3) Generate a preliminary set of new energy scenarios based on the optimal load shedding solution.

[0076] Steps 1-4) Calculate the distance between the new energy output in the scene set and the original center of mass at different segmentation points.

[0077] Step 1-5) Group the output data according to the distance from the centroid, and recalculate the centroid based on the grouped data.

[0078] Step 1-6) determines whether the centroid distance is less than the pre-configured threshold. If so, the final clustering result of the new energy is determined based on the grouped output data and load data to form a representative daily scene data set. If not, the centroid is updated and returns to step 1-5).

[0079] Step 2) Considering the uncertainty of the disturbance source, a node RoCoF calculation model is constructed that takes into account the probability distribution of the disturbance source.

[0080] The probability distribution of the disturbance source is expressed as:

[0081]

[0082] Where: Refers to the working condition of component k1; 0 represents normal operation; 1 represents a fault; Indicates the probability of failure; is the random number drawn for element k1.

[0083] The node RoCoF calculation model taking into account the probability distribution of disturbance sources is:

[0084]

[0085] Where ΔP i Indicates the disturbance power allocated to inertia node i; M ik is the synchronous power coefficient between inertia node i and disturbance node k; E i is the voltage of inertia node i; Uk is the voltage at the disturbance point k; B ik and δ ik are the susceptance and initial angle difference between inertia node i and disturbance node k respectively; Represents the maximum value of discrete distribution fault disturbance source k1; represents the maximum value of the continuous distribution fault disturbance source k2; m1 represents the number of discrete distribution fault disturbance sources; m2 represents the number of continuous distribution fault disturbance sources; K ik represents the power distribution coefficient between inertia node i and disturbance node k.

[0086] Step 3) Based on the node RoCoF calculation model, determine the node RoCoF exceeding probability and the node RoCoF expectation, and build a node inertia weakness assessment model.

[0087] The node inertia weakness assessment model is:

[0088]

[0089] Where, γ i represents the RoCof exceeding limit probability of node i; δ i represents the expected value of RoCoF of node i; w γ 、w δ , represent the weight coefficients of RoCoF crossing probability and RoCoF expected value respectively; Indicates the number of times the node iRoCoF exceeds the critical value; represents the total number of disturbance simulations; J i and B j Indicates the number and length of intervals for classifying RoCoF; R mj Indicates the median value of RoCoF in interval j. i Represents the evaluation index, which is used to evaluate whether the node meets the requirements, γ i , δ i Determined based on the node RoCoF calculation model, in the present invention, it is also called the node control factor.

[0090] Step 4) Determine the participating resources on each side of the source, grid and load and their basic planning constraints, and build a source, grid and load coordinated expansion planning model.

[0091] The optimization objective function of the source-grid-load coordinated expansion planning model is:

[0092]

[0093] Where t, k, and h are the indexes of the planning year, scenario day, and hour within a day, respectively; g, r, e, and l are the indexes of the thermal power units to be built, the small thermal power units to be transformed, the energy storage devices to be built, and the lines to be built, respectively; Ω represents the set corresponding to each resource; Cinv and is the overall cost and equivalent annual cost of the components to be constructed or renovated; t represents the current market value coefficient; d rate represents the discount rate; p k d k Indicates the probability of scenario k occurring and the corresponding number of days; Respectively represent the hourly output of thermal power units, energy storage devices, and wind turbines; Represents the predicted output of wind turbine; y t A 0-1 variable indicating whether the component is under construction; u tkh A 0-1 variable indicating the start and stop status of the device every hour.

[0094] The basic constraints of the source-grid-load coordinated expansion planning model are:

[0095]

[0096] Where m and s represent the indexes of the sending node and the receiving node respectively; d represents the index of the original device of the node; X l is the reactance of line l; M is a large constant value; P l max is the upper limit of the power flow of line l; EL represents the set of existing lines; CL represents the set of candidate lines; and are the phase angles of the sending and receiving nodes of line l respectively; and Represent the phase angle of node b and its upper and lower limits respectively; They represent the upper and lower output limits of unit g respectively; EG and CG represent the collection of existing generator sets and newly built generator sets respectively; Represents the predicted output of wind farm w; M ceh is the charging power multiple of the energy storage system e in time period h; and M ceh Upper and lower limits of M deh is the discharge power multiple of the energy storage device system e in time period h; and M deh Upper and lower limits of E eh is the amount of electricity of the energy storage system e in time period h; and E eh Upper and lower limits of E e0 and E eT are the electricity consumption of the energy storage system e at the beginning and end of the scene respectively.

[0097] Step 5) Using the representative daily scenario dataset as input to the source-grid-load coordinated expansion planning model, the inertia resource construction constraints and line construction constraints are determined. The nodes of the planning scheme are evaluated using the node inertia vulnerability assessment model. Based on the evaluation results, the planning scheme is iteratively optimized and solved until the optimal source-grid-load coordinated expansion planning scheme is output.

[0098] In step 5-1), the representative daily scenario dataset is used as the input of the source-grid-load coordinated expansion planning model, along with other initial information such as unit capacity, network line impedance, and location type of the components to be built.

[0099] Step 5-2) Determine inertia resource construction constraints and line construction constraints.

[0100] The inertia resource construction constraints are:

[0101]

[0102] Where, Ω i represents the set of inertia resources to be built or modified at node i; Ω g Represents the set of inertia resources to be built or modified for all nodes; They are the inertia time constants of the power generating unit to be built, the small thermal power unit to be transformed, and the energy storage device to be built;

[0103] The line construction constraints are:

[0104]

[0105] Where, ξ t,max , χ t,max They represent the maximum electrical coupling degree and the maximum electrical distance from the disturbance source of all nodes in the system in the tth planning year; Z ii , Z jj Represent the self-impedance of node i and node j respectively; Z ij represents the mutual impedance between node j and node j.

[0106] Step 5-3) Based on the initial information and the control factors in the node inertia weakness assessment model, the system expansion planning scheme is solved under the constraints of the source-grid-load coordinated expansion planning model, and the system network information and inertia resource information are updated.

[0107] Step 5-4) Use the node inertia weakness assessment model to evaluate the nodes of the planning scheme.

[0108] In step 5-5), it is determined whether all nodes meet the evaluation requirements. If so, the process goes to step 5-6). Otherwise, the control factors of the nodes that do not meet the evaluation requirements are increased, and the process returns to step 5-3 to re-solve the system expansion plan.

[0109] Step 5-6) determines whether the number of iterations reaches the preconfigured value. If so, go to step 5-7). Otherwise, record the solution result to the feasible solution library, increase the control factor, and return to step 5-3) to re-solve the system expansion plan.

[0110] Steps 5-7) Extract the optimal solution from the feasible solution library and output the optimal source-grid-load coordinated expansion planning solution.

[0111] In this embodiment, MATLAB+CPLEX software is used for optimization and solution.

[0112] This example uses a modified IEEE 24-node system for case analysis. All examples are solved using the commercial CPLEX solver on a 64-bit Windows PC with 16GB of memory and an i7-4790 CPU, running in the MATLAB R2018a environment.

[0113] The revised IEEE RTS-24 system includes 24 nodes, 38 lines, 12 thermal power units, one wind farm, one DC feed-in node, 19 candidate units, 48 ​​candidate lines, and 24 candidate energy storage devices. The wind farm has 19 nodes and the DC feed-in node has 17, with a baseline capacity of 500MW.

[0114] The total planning period is set to 5 years, and the annual growth rates of load, wind power and DC feed-in are 10%, 15% and 10% respectively. The network structure is as follows Figure 4 The maximum single-unit outage is simulated using a DC blocking fault. The DC blocking fault, wind power prediction error, and other main parameters are shown in Table 1.

[0115] Table 1 Description of main parameters

[0116]

[0117] Example 1.1: Using the traditional ordered clustering method

[0118] Example 1.2: Improving the Method Using Trend Metrics

[0119] Example 1.3: Using an Improved Load-New Energy Scenario Reduced Ordered Clustering Method Based on Comprehensive Class Diameter

[0120] Table 2 Comparison of clustering results of three improved methods

[0121]

[0122] As can be seen from Table 2, Example 1.2 obtains the optimal silhouette coefficient value of 0.816 when the optimal cutting point is 17. Compared with Example 1.1, the optimal silhouette coefficient is significantly improved. However, the number of cutting points is increased by 11, resulting in the number of load clustering scenarios increasing by 1.83 times compared with Example 1.1, and the amount of calculation in subsequent application scenarios increases significantly. The number of cutting points for the optimal silhouette coefficient of Example 1.3 increases by 0.67 times compared with Example 1.1, but the optimal silhouette coefficient value can reach 96.9% of that of Example 1.2. In other words, the improved method proposed by the present invention can achieve a relatively ideal clustering effect while increasing the number of cutting points relatively little.

[0123] Based on the initial data from the IEEE RTS-24 system, the inertia time constant of each node's generator was reduced by 20%, and the actual RoCoF of each node was calculated and analyzed. A sampling frequency of 50,000 was selected for the evaluation. Wind power prediction errors followed a normal distribution, while HVDC lockout faults followed a 0-1 distribution.

[0124] like Figure 5 As shown in the figure, under the combined effects of HVDC locking fault and wind power error power disturbance, the actual RoCoF probability distribution of different nodes in the system has significant differences. When the RoCoF limit is set to 0.25Hz / s, there is no over-limit situation in the RoCoF probability distribution of the system as a whole and node 2, and the expected value of the distribution is relatively small; while there is an over-limit probability at nodes 16 and 18, and the absolute value of the over-limit probability is large. It can be seen that the inertia response of the system nodes under large disturbances is different, and the spatial difference is significant. The method in this paper quantitatively describes this difference. The overall distribution of the system inertia weakness degree is shown in the figure. Figure 6 As shown, from dark to light, the weakness of the system's distributed node inertia gradually deepens, and the rating gradually decreases from AD.

[0125] In order to analyze the actual effect of the distributed inertia constraint of the method of the present invention, based on the planning of traditional units, lines, and energy storage devices, the following three schemes are set for specific analysis:

[0126] Example 3.1: Basic planning model without considering inertia constraints

[0127] Example 3.2: Considering centralized inertial constraints

[0128] Example 3.3: Considering Distributed Inertial Constraints

[0129] Figure 7 The results of the node inertia weakness assessment of the planning schemes of Examples 3.1 and 3.2 are given. Table 2 shows the comparison results of the units and energy storage devices invested and built in the three example planning schemes, among which G 5,1 Indicates that the fifth generator among the candidate generators will be put into construction in the first planning year. 17,2This indicates that the 17th candidate energy storage device will be put into construction in the second planning year.

[0130] Depend on Figure 7 Comparing the evaluation results of the alternatives shows that Example 3.1, which ignores the system inertia constraint, has nodes failing the evaluation in planning years 1, 2, and 4, resulting in a poor overall evaluation result. Example 3.2 considers the overall system inertia level requirement, improving the overall inertia response capability to a certain extent. However, nodes 15, 16, and 17 still fail the evaluation, lacking consideration for weak nodes. The evaluation results of Example 3.3 strictly meet the requirements throughout the entire planning cycle, demonstrating that the proposed long-term expansion planning method for considering distributed node inertia vulnerability assessment can effectively account for the distribution characteristics of node inertia vulnerability.

[0131] Table 3 Comparison of planning results and costs of Examples 3.1, 3.2 and 3.3

[0132]

[0133] Combine Figure 7 、 Figure 8 From Table 3, we can see that in Example 3.2 and Example 3.3, compared to Example 3.1, the number of newly built energy storage units increased significantly to meet the system inertia requirements. This, in turn, promoted an increase in the wind power absorption rate, reduced the peak-valley fluctuations on representative days, and reduced the total system power generation capacity demand, ultimately leading to a decrease in the planned capacity of traditional units. Furthermore, compared to Example 3.2, in Example 3.3, while the construction capacity was similar, the construction locations of the units and energy storage units were tilted towards nodes 15, 16, and 18. This also shows that Example 3.3 can improve the system's distributed inertia response capability by adjusting the device construction location at a similar capacity level. Furthermore, in terms of line construction, Example 3.3 has more lines than Example 3.1 and Example 3.2, and the construction locations are more inclined towards inertia-weak areas, indicating that Example 3.3 can improve the inertia response capability of weak inertia nodes by increasing the location and number of newly built lines. From a cost perspective, compared to Example 3.1, Examples 3.2 and 3.3 have higher requirements for the system inertia support capability, resulting in an increase in overall planning costs. However, the cost of Example 3.3 is similar to that of Example 3.2, indicating that considering distributed inertia constraints does not result in a significant cost increase.

[0134] To analyze the impact of different resource participation on the overall planning cost under distributed inertia constraints, the present invention proposes the following three schemes for analysis. Among them, unit transformation refers to the transformation of small thermal power units in the actual system that have to be retired due to excessive unit energy consumption or low-carbon emission reduction requirements, so that they can operate online but do not provide active power, but only provide inertial response.

[0135] Example 4.1: Coordinated Planning of Thermal Power Units and Transmission Lines

[0136] Example 4.2: Adding energy storage to Example 4.1

[0137] Example 4.3: Adding a new unit modification based on Example 4.2

[0138] Figure 9 The comparison results of the investment, construction and penalty costs of the three options are given. Table 3 gives the specific planning schemes and costs of the three options.

[0139] Table 4 Comparison of planning results and costs of Examples 4.1, 4.2 and 4.3

[0140]

[0141] Combine Figure 9 Compared to Table 3, we can see that both Cases 4.3 and 4.2 incorporate energy storage, significantly reducing the wind curtailment penalty costs compared to Case 4.1. However, Case 4.3 incorporates a unit retrofit project, requiring less newly built energy storage capacity and resulting in relatively lower overall investment and construction costs. Furthermore, the increase in wind curtailment penalty costs in Case 4.3 is minimal compared to Case 4.2. This is because a portion of the energy storage in Case 4.2 primarily serves as inertia support. This creates a certain amount of energy storage redundancy to increase wind power absorption. This redundancy is released through the introduction of unit retrofit projects, reducing overall operating and investment costs. Overall, collaborative planning with multiple resources offers greater economic benefits. It is essential to fully account for resources such as load-side energy storage and unit retrofit projects that can provide peak load regulation or inertia support in planning.

[0142] The above describes in detail the preferred embodiments of the present invention. It should be understood that those skilled in the art can make numerous modifications and variations based on the concepts of the present invention without inventive effort. Therefore, any technical solutions that can be derived by those skilled in the art through logical analysis, reasoning, or limited experimentation based on the concepts of the present invention and the prior art should be within the scope of protection defined by the claims.

Claims

1. A source-grid-load coordinated expansion planning method considering node inertia weakness assessment, characterized by: The following steps are involved: Step 1) Obtain annual load and renewable energy output data, and cluster the data using an improved load-renewable energy scenario reduction ordered clustering model based on comprehensive class diameter to form a representative daily scenario dataset; Step 2) Considering the uncertainty of the disturbance source, a node RoCoF calculation model is constructed that takes into account the probability distribution of the disturbance source; Step 3) Determine the node RoCoF limit violation probability and node RoCoF expectation based on the node RoCoF calculation model, and construct a node inertia weakness assessment model; the node inertia weakness assessment model is: Where, γ i represents the RoCof exceeding limit probability of node i; δ i represents the expected value of RoCoF of node i; w γ 、w δ , represent the weight coefficients of RoCoF crossing probability and RoCoF expected value respectively; Indicates the number of times the node iRoCoF exceeds the critical value; represents the total number of disturbance simulations; J i and B j Indicates the number and length of intervals for classifying RoCoF; R mj represents the median value of RoCoF in interval j; Step 4) Determine the participating resources on each side of the source, grid, and load and their basic planning constraints, and construct a source, grid, and load coordinated expansion planning model; the optimization objective function of the source, grid, and load coordinated expansion planning model is: Where t, k, and h are the indexes of the planning year, scenario day, and hour within a day, respectively; g, r, e, and l are the indexes of the thermal power units to be built, the small thermal power units to be transformed, the energy storage devices to be built, and the lines to be built, respectively; Ω represents the set corresponding to each resource; C inv and is the overall cost and equivalent annual cost of the components to be constructed or renovated; t represents the current market value coefficient; d rate represents the discount rate; p k d k Indicates the probability of scenario k occurring and the corresponding number of days; Respectively represent the hourly output of thermal power units, energy storage devices, and wind turbines; Represents the predicted output of wind turbine; y t A 0-1 variable indicating whether the component is under construction; u tkh A 0-1 variable indicating the start and stop status of the device every hour; Step 5) Using the representative daily scenario dataset as input to the source-grid-load coordinated expansion planning model, the inertia resource construction constraints and line construction constraints are determined. The nodes of the planning scheme are evaluated using the node inertia vulnerability assessment model. Based on the evaluation results, the planning scheme is iteratively optimized and solved until the optimal source-grid-load coordinated expansion planning scheme is output.

2. A source-grid-load coordinated expansion planning method considering node inertia weakness assessment according to claim 1, characterized in that: The step 1) comprises the following steps: Step 1-1) Obtain annual load and renewable energy output data and set the cut point range; Step 1-2) Generate the optimal load shedding scheme using an improved load-new energy scenario reduction ordered clustering model based on comprehensive class diameter; Steps 1-3) Generate a preliminary set of new energy scenarios based on the optimal load shedding solution; Steps 1-4) Calculate the distance between the new energy output in the scene set and the original centroid at different segmentation points; Step 1-5) Group the output data according to the distance from the centroid and recalculate the centroid based on the grouped data; Step 1-6) determines whether the centroid distance is less than the pre-configured threshold. If so, the final clustering result of the new energy is determined based on the grouped output data and load data to form a representative daily scene data set. If not, the centroid is updated and returns to step 1-5).

3. The source-grid-load coordinated expansion planning method considering node inertia weakness assessment according to claim 2 is characterized in that: The improved load-new energy scenario reduction ordered clustering model based on comprehensive class diameter is: In the formula, n ordered samples are divided into k categories, and the G-th sample group includes samples {X i ,X i+1 ,…,X t ,…X j }; Represents the sample mean corresponding to the G-th sample group; represents the sample mean trend corresponding to the Gth sample group; D(i,j) represents the comprehensive class diameter corresponding to the Gth sample group; L[(b(n,k))] represents a method of dividing n sample data into k categories; p(n,k) represents an optimal solution that minimizes L[(b(n,k))] under the aforementioned method.

4. The source-grid-load coordinated expansion planning method considering node inertia weakness assessment according to claim 1 is characterized in that: The probability distribution of the disturbance source is expressed as: Where: Refers to the working condition of component k1; 0 represents normal operation; 1 indicates a fault has occurred; Indicates the probability of failure; is the random number drawn for element k1.

5. The source-grid-load coordinated expansion planning method considering node inertia weakness assessment according to claim 4 is characterized in that: The node RoCoF calculation model taking into account the probability distribution of disturbance sources is: Where ΔP i Indicates the disturbance power allocated to inertia node i; M ik is the synchronous power coefficient between inertia node i and disturbance node k; E i is the voltage of inertia node i; U k is the voltage at the disturbance point k; B ik and δ ik are the susceptance and initial angle difference between inertia node i and disturbance node k respectively; Represents the maximum value of discrete distribution fault disturbance source k1; represents the maximum value of the continuous distribution fault disturbance source k2; m1 represents the number of discrete distribution fault disturbance sources; m2 represents the number of continuous distribution fault disturbance sources; K ik represents the power distribution coefficient between inertia node i and disturbance node k.

6. The source-grid-load coordinated expansion planning method considering node inertia weakness assessment according to claim 1 is characterized in that: The basic constraints of the source-grid-load coordinated expansion planning model are: Where m and s represent the indexes of the sending node and the receiving node respectively; d represents the index of the original device of the node; X l is the reactance of line l; M is a large constant value; P l max is the upper limit of the power flow of line l; EL represents the existing line set; CL represents the line set to be selected; and are the phase angles of the sending and receiving nodes of line l respectively; and Represent the phase angle of node b and its upper and lower limits respectively; They represent the upper and lower output limits of unit g respectively; EG and CG represent the collection of existing generator sets and newly built generator sets respectively; Represents the predicted output of wind farm w; M ceh is the charging power multiple of the energy storage system e in time period h; and M ceh Upper and lower limits of M deh is the discharge power multiple of the energy storage device system e in time period h; and M deh Upper and lower limits of E eh is the amount of electricity of the energy storage system e in time period h; and E eh Upper and lower limits of E e0 and E eT are the electricity consumption of the energy storage system e at the beginning and end of the scene respectively.

7. The source-grid-load coordinated expansion planning method considering node inertia weakness assessment according to claim 6 is characterized in that: The inertia resource construction constraints are: Where, Ω i represents the set of inertia resources to be built or modified at node i; Ω g Represents the set of inertia resources to be built or modified for all nodes; They are the inertia time constants of the power generating unit to be built, the small thermal power unit to be transformed, and the energy storage device to be built; The line construction constraints are: Where, ξ t,max , χ t,max They represent the maximum electrical coupling degree and the maximum electrical distance from the disturbance source of all nodes in the system in the tth planning year; Z ii , Z jj Represent the self-impedance of node i and node j respectively; Z ij represents the mutual impedance between node j and node j.

8. The source-grid-load coordinated expansion planning method considering node inertia weakness assessment according to claim 1 is characterized in that: The step 5) comprises the following steps: Step 5-1) Use the representative daily scenario dataset as input to the source-grid-load coordinated expansion planning model, along with other initial information; Step 5-2) Determine the inertia resource construction constraints and line construction constraints, Step 5-3) Based on the initial information and the control factors in the node inertia weakness assessment model, solve the system expansion planning scheme while satisfying the constraints of the source-grid-load coordinated expansion planning model, and update the system network information and inertia resource information; Step 5-4) Evaluate the nodes of the planning scheme using the node inertia weakness assessment model; Step 5-5) determines whether all nodes meet the evaluation requirements. If so, proceed to step 5-6). Otherwise, increase the control factor for the nodes that do not meet the evaluation requirements and return to step 5-3) to re-solve the system expansion plan. Step 5-6) determines whether the number of iterations reaches the pre-configured value. If so, go to step 5-7). Otherwise, record the solution result to the feasible solution database, increase the control factor, and return to step 5-3) to re-solve the system expansion plan. Steps 5-7) Extract the optimal solution from the feasible solution library and output the optimal source-grid-load coordinated expansion planning solution.

Images