Power system planning typical scene generation method based on data driving

Abstract

The invention discloses a power system planning typical scene generation method based on data driving, and relates to the technical field of typical scene generation in a power system, and the method comprises the steps: carrying out the clustering of a scene through employing a K-medios clustering algorithm based on the wind and light output and load, and obtaining a typical scene; then, constructing a power system capacity expansion planning model based on the typical scene, and obtaining an optimal planning result of a unit and a line in the typical scene; and finally, based on a planning result, constructing a full-time optimization operation model, identifying a scene with extremely unbalanced supply and demand, if the scene with unbalanced supply and demand exists, extracting the scene with unbalanced supply and demand, putting the scene with unbalanced supply and demand into a typical scene, and repeating the previous step. And if the supply-demand imbalance scene does not exist, ending the circulation, and outputting a supply-demand imbalance typical scene. According to the method provided by the invention, the planning efficiency is improved while a typical scene of a supply-demand extreme imbalance scene is considered, and the planning precision is not sacrificed.

Description

Technical Field

[0001] The present invention relates to the technical field of typical scenario generation in power systems, and in particular to a typical scenario generation method for power system planning based on data-driven. Background Art

[0002] In recent years, with the development of new energy power generation, the planning of wind, solar and storage installed capacity has played a significant role in promoting the application of renewable energy, optimizing the energy structure, responding to electricity demand, and improving the stability of the power grid. Reasonable planning not only helps to cope with the challenges of climate change and energy transformation, but also improves the economy and sustainability of the power system, providing important support for achieving the goal of carbon neutrality.

[0003] In order to accurately and comprehensively plan and configure the power system that includes wind power and photovoltaic resources, it is necessary to comprehensively consider the time-varying characteristics of loads and wind and photovoltaic resources, and analyze a large number of wind and photovoltaic power generation output and load scenarios. However, the large number of such scenarios increases the complexity of computational analysis and prolongs the calculation time accordingly. Therefore, how to extract a small number of representative scenarios and improve computational efficiency while ensuring planning robustness has become an urgent problem to be solved. In existing methods, typical scenario generation can use a small number of representative scenarios to describe the characteristics of a large number of complex scenarios, reduce the number of similar scenarios within the research cycle, and achieve the effect of reducing computational complexity. However, it is difficult to take into account extreme scenarios in the current typical scenario generation method, resulting in planning results that are inconsistent with actual requirements, and a large number of load shedding and power abandonment phenomena occur in extremely unbalanced scenarios.

[0004] In view of this, a data-driven method for generating typical scenarios for power system planning is needed. Summary of the invention

[0005] In view of the above-mentioned defects of the prior art, the purpose of the present invention is to find a typical scenario generation method that considers the extreme imbalance between supply and demand from a planning perspective, so as to improve the planning efficiency without sacrificing the planning accuracy.

[0006] To achieve this purpose, the present invention proposes a data-driven method for generating typical scenarios for power system planning, the method comprising the following steps:

[0007] S1: Based on wind and solar power output and load, the K-mediods clustering algorithm is used to cluster the scenarios to obtain typical scenarios;

[0008] S2: Construct a power system capacity expansion planning model based on typical scenarios to obtain the optimal planning results of units and lines under typical scenarios;

[0009] S3: Based on the planning results, a full-time optimization operation model is constructed to identify scenarios with extreme imbalance between supply and demand. If there is a supply and demand imbalance scenario, the supply and demand imbalance scenario is extracted and included in the typical scenario, and step S2 is repeated; if there is no supply and demand imbalance scenario, the loop is terminated and the typical supply and demand imbalance scenario is output.

[0010] The clustering method of step S1 based on scene scenery processing and load is as follows:

[0011] S11: K-mediods clustering algorithm is a commonly used clustering algorithm, which is mainly used to divide the data set into k clusters;

[0012] 1) Initialization: Randomly select k initial cluster centers.

[0013] 2) Assign samples to the nearest cluster center: For each scene, calculate its Euclidean distance to each cluster center:

[0014]

[0015] Where: x i is the i-th scene; c k is the kth cluster center; n is the dimension of the time within the day; x ij is the value of the i-th scene at time j; c kj is the value of the kth cluster center at time j.

[0016] And assign the scene to the cluster with the closest distance.

[0017] 3) Reselect the center point: For each cluster, select a new center point. The new center point is the data point in the cluster that minimizes the total distance. The distance calculation method within the cluster is:

[0018]

[0019] Where: x i For cluster S k Any point in .

[0020]

[0021] For cluster S k Each point c k , calculate the sum of the distances for it to become the center point, and if its total distance is the smallest, replace it with the new center point.

[0022] 4) Iterative calculation: Repeat the allocation and update process until the changes in cluster centers become stable.

[0023] The goal of the K-medoids algorithm is to minimize the total distance of all clusters:

[0024]

[0025] Where: K is the number of clusters.

[0026] Different scenarios have different characteristics in terms of wind and solar power output and load. Based on this, the scenarios can be clustered based on the K-mediods clustering algorithm to obtain k typical scenarios and their occurrence probabilities.

[0027] The step S2 constructs a power system capacity expansion planning model, and the method steps for obtaining the optimal planning result are as follows:

[0028] S21: Construct the power system capacity expansion planning problem: the planning variables are the new capacity of units and lines. The objective function is to minimize the total cost, including investment cost, maintenance cost, and operation cost:

[0029] C tot =C INV +C OM +C OP

[0030] Where: C tot is the total cost; C INV is the investment cost; C OM is the maintenance cost; C OP For running costs.

[0031] The investment cost consists of the generator investment cost and the line investment cost:

[0032] C INV =C INV,gen +C INV,line

[0033]

[0034] Where: C INV,gen is the investment cost of the generator set; C INV,line is the investment cost of transmission line; N g is the number of generator sets; gen is the capital recovery factor of the generator set, i r is the investment discount rate; gen,g is the investment payback period of unit g; G g Investment capacity for generator sets; c inv,g is the construction cost per unit capacity of unit g; line is the capital recovery coefficient of the line, y line,l F is the investment payback period of transmission line l; l is a variable representing the investment status of transmission lines; c inv,l is the construction cost per unit length of the transmission line; Dl is the construction length of the transmission line.

[0035] The maintenance cost is expressed as follows:

[0036]

[0037] Where: r om,g is the maintenance cost per unit capacity of the generator set g; r om,l is the maintenance cost per unit length of transmission line l; F l is a variable representing the investment status of transmission lines; G g Investment capacity for generator sets; c inv,l is the construction cost per unit length of the transmission line; D l The construction length of the transmission line.

[0038] The operating cost is composed of fuel cost, start-stop cost and load shedding cost, which can be expressed as follows:

[0039] C OP =C F +C UD +C LS

[0040]

[0041]

[0042] Where: C F is the generator fuel cost; C UD is the cost of starting and stopping the generator; C LS is the system load shedding cost; ψ T is a collection of thermal power generators; N s is the number of typical scenes; N d is the number of days included in the planning cycle; ω s is the probability of the typical scene s appearing; N t is the number of hours in a typical scenario; c a,g 、c b,g With c c,g is the power generation cost coefficient of the unit. In order to improve the solution efficiency, it is approximated by using a piecewise linear function; P g,t is the power generation of unit g at time t; S g,t is the binary variable of the running state of unit g at time t, S g,t =1 means the unit is in the on state, otherwise it is in the off state; c on,g and c off,g are the startup and shutdown costs of unit g respectively; O g,t and C g,tare whether the unit g performs startup and shutdown actions at time t, O g,t =1 means that the unit g starts up at time t, C g,t =1 means unit g stops at time t, O g,t =0 while C g,t =0 means the state of unit g has not changed at time t; L cur,t is the load shedding amount of the system at time t.

[0043] The planning problem needs to satisfy the following constraints:

[0044] 1) Power generation technology constraints

[0045] Thermal power units Output constraints:

[0046] α g S g,t G g ≤P g,t ≤S g,t G g

[0047] Where: α g It is the minimum output technical coefficient of thermal power unit g.

[0048] Conventional unit Start-Stop Constraints:

[0049] S g,t -S g,t-1 =O g,t -C g,t

[0050]

[0051] Where: They are the actual startup time at minimum output and the minimum shutdown time of the unit respectively.

[0052] Conventional units Hill climbing constraints:

[0053]

[0054] Where: They are the upward and downward climbing capabilities of the unit respectively.

[0055] Wind and solar units Output constraints:

[0056] 0≤P g,t ≤β g,t G g

[0057] Where: β g,tis the capacity factor of the unit at time t, and its value varies between [0,1] according to different wind and solar output conditions. g Invest in capacity for generator sets.

[0058] 2) Power line transmission constraints:

[0059]

[0060] Where: Γ l,g is the transfer distribution factor of unit g for line l, H l,d is the transfer distribution factor of load d for line l.

[0061] 3) Power system power balance constraints:

[0062]

[0063] Where: P cha,t , P dis,t They are the charging and discharging power of long-term energy storage respectively.

[0064] 4) Long-term energy storage constraints:

[0065]

[0066] SOC min ≤SOC t ≤SOC max

[0067] SOC1=SOC end

[0068] Where: SOC t The storage state of long-term energy storage, SOC max , SOC min They are the upper and lower limits of long-term energy storage capacity respectively.

[0069] S22: The typical scenario obtained in S1 and its probability of occurrence are placed into the planning model to obtain the optimal expansion capacity under the conditions of this typical scenario.

[0070] The step S3 constructs a full-time optimization operation model to identify scenarios with extreme imbalance between supply and demand. The method for reconstructing a typical scenario set based on the extreme imbalance scenario is as follows:

[0071] S31: Based on the capacity configuration obtained in step S2, a power system optimization operation model is constructed, the optimization variables are the unit output and the load shedding amount, and the optimization goal is to minimize the operating cost:

[0072] C OP =C F +C UD +CLS

[0073] The optimization problem also needs to meet the unit power generation technology constraints, line transmission constraints, power balance constraints and long-term energy storage constraints.

[0074] The unit capacity G in the model g and line capacity F l , adopt the configuration capacity in the optimal planning scheme in step S2.

[0075] S32: Put the full-time scenario into the optimization model to obtain the optimal operation plan. According to the operation optimization results, find the scenario with the maximum load shedding, regard it as an extremely unbalanced scenario, and extract the wind and solar power output and load data corresponding to this scenario.

[0076] S33: Re-cluster the scene set with the extremely unbalanced scene removed by K-mediods to obtain k typical scenes and their weights. The extremely unbalanced scene is taken as the typical scene, and its weight is 1. Then the probability of each typical scene appearing is:

[0077]

[0078] Where: S s is the weight of the typical scene s, S total is the total weight of all typical scenarios;

[0079] A set of typical scenes are obtained for reconstruction.

[0080] S34: Enter the loop, reapply the reconstructed typical scenarios and their probabilities to the power system planning expansion model, obtain the optimal planning capacity (step S2), put the optimal planning capacity into the power system full-time operation model (step S31), identify extreme unbalanced scenarios (step S32), continue the loop if there are extreme unbalanced scenarios, and reconstruct the typical scenario set (step S33); if there are no extreme unbalanced scenarios, exit the loop.

[0081] Compared with the prior art, the present invention has the following beneficial effects:

[0082] The present invention uses K-mediods clustering algorithm based on wind and solar power output and load to cluster scenarios to obtain typical scenarios; then, a power system capacity expansion planning model based on typical scenarios is constructed to obtain the optimal planning results of units and lines under typical scenarios; finally, based on the planning results, a full-time optimization operation model is constructed to identify scenarios with extreme imbalance in supply and demand. If there is a scenario with imbalance in supply and demand, the scenario with imbalance in supply and demand is extracted and included in the typical scenario, and step S2 is repeated; if there is no scenario with imbalance in supply and demand, the cycle is terminated and a typical scenario with imbalance in supply and demand is output. The method of the present invention improves planning efficiency while considering typical scenarios of extreme imbalance in supply and demand without sacrificing planning accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0083] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following is a brief introduction to the drawings required for the specific embodiments or the description of the prior art. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn according to the actual scale.

[0084] Figure 1 Schematic flow chart of the method for generating typical scenarios for data-driven power system planning.

[0085] Figure 2 This is a schematic diagram of the IEEE 24-node system structure.

[0086] Figure 3 is the load per unit value curve of the example.

[0087] Figure 4 This is the per-unit value curve of wind power output for the example.

[0088] Figure 5 This is the per-unit photovoltaic output curve for the example.

[0089] Figure 6 The load shedding result graph is shown for the planned capacity obtained by the existing clustering method. DETAILED DESCRIPTION

[0090] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments 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.

[0091] It should be understood that when used in this specification and the appended claims, the terms "include" and "comprises" indicate the presence of described features, integers, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or combinations thereof.

[0092] It should also be understood that the terms used in the present specification are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the present specification and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0093] It should be further understood that the term "and / or" used in the present description and the appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0094] In one embodiment of the present invention, a method for generating typical scenarios for power system planning based on data-driven is provided, which comprises:

[0095] S1: Based on wind and solar power output and load, the K-mediods clustering algorithm is used to cluster the scenarios to obtain typical scenarios;

[0096] S2: Construct a power system capacity expansion planning model based on typical scenarios to obtain the optimal planning results of units and lines under typical scenarios;

[0097] S3: Based on the planning results, a full-time optimization operation model is constructed to identify scenarios with extreme imbalance between supply and demand. If there is a supply and demand imbalance scenario, the supply and demand imbalance scenario is extracted and included in the typical scenario, and step S2 is repeated; if there is no supply and demand imbalance scenario, the loop is terminated and the typical supply and demand imbalance scenario is output.

[0098] The clustering method of step S1 based on scene scenery processing and load is as follows:

[0099] S11: K-mediods clustering algorithm is a commonly used clustering algorithm, which is mainly used to divide the data set into k clusters;

[0100] 5) Initialization: Randomly select k initial cluster centers.

[0101] 6) Assign samples to the nearest cluster center: For each scene, calculate its Euclidean distance to each cluster center:

[0102]

[0103] Where: x iis the i-th scene; c k is the kth cluster center; n is the dimension of the time within the day; x ij is the value of the i-th scene at time j; c kj is the value of the kth cluster center at time j.

[0104] And assign the scene to the cluster with the closest distance.

[0105] 7) Reselect the center point: For each cluster, select a new center point. The new center point is the data point in the cluster that minimizes the total distance. The distance calculation method within the cluster is:

[0106]

[0107] Where: x i For cluster S k Any point in .

[0108]

[0109] For cluster S k Each point c k , calculate the sum of the distances for it to become the center point, and if its total distance is the smallest, replace it with the new center point.

[0110] 8) Iterative calculation: Repeat the process of allocation and updating until the changes in cluster centers become stable.

[0111] The goal of the K-medoids algorithm is to minimize the total distance of all clusters:

[0112]

[0113] Where: K is the number of clusters.

[0114] Different scenarios have different characteristics in terms of wind and solar power output and load. Based on this, the scenarios can be clustered based on the K-mediods clustering algorithm to obtain k typical scenarios and their occurrence probabilities.

[0115] The step S2 constructs a power system capacity expansion planning model, and the method steps for obtaining the optimal planning result are as follows:

[0116] S21: Construct the power system capacity expansion planning problem: the planning variables are the new capacity of units and lines. The objective function is to minimize the total cost, including investment cost, maintenance cost, and operation cost:

[0117] C tot =C INV +C OM +C OP

[0118] Where: Ctot is the total cost; C INV is the investment cost; C OM is the maintenance cost; C OP For running costs.

[0119] The investment cost consists of the generator investment cost and the line investment cost:

[0120] C INV =C INV,gen +C INV,line

[0121]

[0122] Where: C INV,gen is the investment cost of the generator set; C INV,line is the investment cost of transmission line; N g is the number of generator sets; gen is the capital recovery factor of the generator set, i r is the investment discount rate; gen,g is the investment payback period of unit g; G g Investment capacity for generator sets; c inv,g is the construction cost per unit capacity of unit g; line is the capital recovery coefficient of the line, y line,l F is the investment payback period of transmission line l; l is a variable representing the investment status of transmission lines; c inv,l is the construction cost per unit length of the transmission line; D l The construction length of the transmission line.

[0123] The maintenance cost is expressed as follows:

[0124]

[0125] Where: r om,g is the maintenance cost per unit capacity of the generator set g; r om,l is the maintenance cost per unit length of transmission line l; F l is a variable representing the investment status of transmission lines; G g Investment capacity for generator sets; c inv,l is the construction cost per unit length of the transmission line; D l The construction length of the transmission line.

[0126] The operating cost is composed of fuel cost, start-stop cost and load shedding cost, which can be expressed as follows:

[0127] C OP =C F +C UD +CLS

[0128]

[0129]

[0130] Where: C F is the generator fuel cost; C UD is the cost of starting and stopping the generator; C LS is the system load shedding cost; ψ T is a collection of thermal power generators; N s is the number of typical scenes; N d is the number of days included in the planning cycle; ω s is the probability of the typical scene s appearing; N t is the number of hours in a typical scenario; c a,g 、c b,g With c c,g is the power generation cost coefficient of the unit. In order to improve the solution efficiency, it is approximated by using a piecewise linear function; P g,t is the power generation of unit g at time t; S g,t is the binary variable of the running state of unit g at time t, S g,t =1 means the unit is in the on state, otherwise it is in the off state; c on,g and c off,g are the startup and shutdown costs of unit g respectively; O g,t and C g,t are whether the unit g performs startup and shutdown actions at time t, O g,t =1 means that the unit g starts up at time t, C g,t =1 means unit g stops at time t, O g,t =0 while C g,t =0 means the state of unit g has not changed at time t; L cur,t is the load shedding amount of the system at time t.

[0131] The planning problem needs to satisfy the following constraints:

[0132] 1) Power generation technology constraints

[0133] Thermal power units Output constraints:

[0134] α g S g,t G g ≤P g,t ≤S g,t G g

[0135] Where: α gIt is the minimum output technical coefficient of thermal power unit g.

[0136] Conventional units Start-Stop Constraints:

[0137] S g,t -S g,t-1 =O g,t -C g,t

[0138]

[0139] Where: They are the actual startup time at minimum output and the minimum shutdown time of the unit respectively.

[0140] Conventional units Hill climbing constraints:

[0141]

[0142] Where: They are the upward and downward climbing capabilities of the unit respectively.

[0143] Wind and solar units Output constraints:

[0144] 0≤P g,t ≤β g,t G g

[0145] Where: β g,t is the capacity factor of the unit at time t, and its value varies between [0,1] according to different wind and solar output conditions. g Invest in capacity for generator sets.

[0146] 5) Power line transmission constraints:

[0147]

[0148] Where: Γ l,g is the transfer distribution factor of unit g for line l, H l,d is the transfer distribution factor of load d for line l.

[0149] 6) Power system power balance constraints:

[0150]

[0151] Where: P cha,t , P dis,t They are the charging and discharging power of long-term energy storage respectively.

[0152] 7) Long-term energy storage constraints:

[0153]

[0154] SOC min ≤SOC t ≤SOC max

[0155] SOC1=SOC end

[0156] Where: SOC t The storage state of long-term energy storage, SOC max , SOC min They are the upper and lower limits of long-term energy storage capacity respectively.

[0157] S22: The typical scenario obtained in S1 and its probability of occurrence are placed into the planning model to obtain the optimal expansion capacity under the conditions of this typical scenario.

[0158] The step S3 constructs a full-time optimization operation model to identify scenarios with extreme imbalance between supply and demand. The method for reconstructing a typical scenario set based on the extreme imbalance scenario is as follows:

[0159] S31: Based on the capacity configuration obtained in step S2, a power system optimization operation model is constructed, the optimization variables are the unit output and the load shedding amount, and the optimization goal is to minimize the operating cost:

[0160] C OP =C F +C UD +C LS

[0161] The optimization problem also needs to meet the unit power generation technology constraints, line transmission constraints, power balance constraints and long-term energy storage constraints.

[0162] The unit capacity G in the model g and line capacity F l , adopt the configuration capacity in the optimal planning scheme in step S2.

[0163] S32: Put the full-time scenario into the optimization model to obtain the optimal operation plan. According to the operation optimization results, find the scenario with the maximum load shedding, regard it as an extremely unbalanced scenario, and extract the wind and solar power output and load data corresponding to this scenario.

[0164] S33: Re-cluster the scene set with the extremely unbalanced scene removed by K-mediods to obtain k typical scenes and their weights. The extremely unbalanced scene is taken as the typical scene, and its weight is 1. Then the probability of each typical scene appearing is:

[0165]

[0166] Where: Ss is the weight of the typical scene s, S total is the total weight of all typical scenarios;

[0167] A set of typical scenes are obtained for reconstruction.

[0168] S34: Enter the loop, reapply the reconstructed typical scenarios and their probabilities to the power system planning expansion model, obtain the optimal planning capacity (step S2), put the optimal planning capacity into the power system full-time operation model (step S31), identify extreme unbalanced scenarios (step S32), continue the loop if there are extreme unbalanced scenarios, and reconstruct the typical scenario set (step S33); if there are no extreme unbalanced scenarios, exit the loop. The algorithm flow is as follows Figure 1 shown.

[0169] In order to show the advantages of the method of the present invention, the typical scene obtained by clustering based on the existing K-medoids method is compared with the typical scene obtained by the method proposed in the present invention, and the IEEE24 standard example is used as a test example. Figure 2 As shown in the figure, the system has a total of 24 nodes, 38 transmission lines and 16 generator sets, including 12 thermal power generator sets (connected to nodes 1, 3, 7, 13, 14, 15, 16, 18, 21, 22, 23, and 23 respectively), 1 wind power generator set (connected to nodes 3 and 5 respectively), 2 photovoltaic units (connected to nodes 13 and 21 respectively), and 1 long-term energy storage device (connected to node 13). The load output curve and the wind and light processing curve both include data at 1440, as shown in the figure. Figure 3 , Figure 4 , Figure 5 shown.

[0170] The planned capacities obtained by the two methods are respectively put into the operation model to calculate the load shedding amount. S1 is the existing K-medoids method, and S2 is the method proposed in this paper. The calculation results are:

[0171] Table 1 Best reconstruction results of the comparison methods

[0172]

[0173]

[0174] It can be seen from Table 1 that based on the method proposed in this paper, after introducing the extreme unbalanced scenario as the planning scenario, the typical scenario set considering the extreme unbalanced scenario is constructed, and the planned capacity configuration obtained can effectively reduce the load shedding phenomenon and ensure power supply reliability.

[0175] The load shedding amount at each moment in the existing method is as follows Figure 6 shown.

[0176] Those of ordinary skill in the art will appreciate that the units of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition of each example has been generally described in terms of function in the above description. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0177] In the embodiments provided by the present invention, it should be understood that the division of units is only a logical function division, and there may be other division methods in actual implementation, for example, multiple units can be combined into one unit, one unit can be split into multiple units, or some features can be ignored, etc.

[0178] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0179] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or 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, read-only memory (ROM, Read-0nlyMemory), random access memory (RAM, RandomAccessMemory), mobile hard disk, magnetic disk or optical disk, etc., which can store program code.

[0180] 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 replace some or all of the technical features therein by equivalents. These modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be included in the scope of the claims and specification of the present invention.

Claims

1. A data-driven method for generating typical scenarios for power system planning, characterized in that: The following steps are involved: S1: Based on wind and solar power output and load, a clustering algorithm is used to cluster the scenarios to obtain typical scenarios; S2: Taking the minimum total cost as the objective function, a power system capacity expansion planning model based on typical scenarios is constructed. The typical scenarios obtained in S1 and their probability of occurrence are placed in the planning model to obtain the optimal expansion capacity under the typical scenario conditions; S3: Based on the planning results, identify scenarios of supply and demand imbalance. If there are supply and demand imbalance scenarios, extract the supply and demand imbalance scenarios and include them in the typical scenarios, and repeat step S2; if there are no supply and demand imbalance scenarios, end the loop and output the typical supply and demand imbalance scenarios.

2. According to claim 1, a method for generating typical scenarios for power system planning based on data drive, characterized in that: Step S1 is specifically as follows: Initialization: Randomly select k initial cluster centers; Assign samples to the nearest cluster center: For each scene, calculate its Euclidean distance to each cluster center: Where: x i is the i-th scene; c k is the kth cluster center; n is the dimension of the time within the day; x ij is the value of the i-th scene at time j; c kj is the value of the k-th cluster center at time j; Assign the scene to the cluster with the closest distance; Reselect the center point: For each cluster, select a new center point. The new center point is the data point in the cluster that minimizes the total distance. The distance within the cluster is calculated as follows: Where: x i For cluster S k Any point in ; For cluster S k Each point c k , calculate the sum of the distances to become the center point, if its total distance is the smallest, replace it with the new center point; Iterative calculation: Repeat the allocation and update process until the changes in cluster centers tend to be stable; The goal of the K-medoids algorithm is to minimize the total distance of all clusters: Where: K is the number of clusters.

3. The method for generating typical scenarios for power system planning based on data drive according to claim 2, characterized in that: The total cost includes investment cost, maintenance cost, and operating cost. The objective function is as follows: C tot =C INV +C OM +C OP Where: C tot is the total cost; C INV is the investment cost; C OM is the maintenance cost; C OP For running costs.

4. The method for generating typical scenarios for power system planning based on data-driven according to claim 3, characterized in that: The investment cost consists of the generator investment cost and the line investment cost: C INV =C INV,gen +C INV,line Where: C INV,gen is the investment cost of the generator set; C INV,line is the investment cost of transmission line; N g is the number of generator sets; gen is the capital recovery factor of the generator set, i r is the investment discount rate; gen,g is the investment payback period of unit g; G g Investment capacity for generator sets; c inv,g is the construction cost per unit capacity of unit g; line is the capital recovery coefficient of the line, y line,l F is the investment payback period of transmission line l; l is a variable representing the investment status of transmission lines; c inv,l is the construction cost per unit length of the transmission line; D l is the construction length of the transmission line.

5. The method for generating typical scenarios for power system planning based on data-driven according to claim 3, characterized in that: The maintenance cost is expressed as follows: Where: r om,g is the maintenance cost per unit capacity of the generator set g; r om,l is the maintenance cost per unit length of transmission line l; F l is a variable representing the investment status of transmission lines; G g Investment capacity for generator sets; c inv,l is the construction cost per unit length of the transmission line; D l is the construction length of the transmission line.

6. The method for generating typical scenarios for power system planning based on data drive according to claim 3 is characterized in that: The operating cost is composed of fuel cost, start-stop cost and load shedding cost, which can be expressed as follows: C OP =C F +C UD +C LS Where: C F is the generator fuel cost; C UD is the cost of starting and stopping the generator; C LS is the system load shedding cost; ψ T is a collection of thermal power generators; N s is the number of typical scenes; N d is the number of days included in the planning cycle; ω s is the probability of the typical scene s appearing; N t is the number of hours in a typical scenario; c a,g 、c b,g With c c,g is the power generation cost coefficient of the unit. In order to improve the solution efficiency, it is approximated by using a piecewise linear function; P g,t is the power generation of unit g at time t; S g,t is the binary variable of the running state of unit g at time t, S g,t =1 means the unit is in the on state, otherwise it is in the off state; c on,g and c off,g are the startup and shutdown costs of unit g respectively; O g,t and C g,t are whether the unit g performs startup and shutdown actions at time t, O g,t =1 means that the unit g starts up at time t, C g,t =1 means unit g stops at time t, O g,t =0 while C g,t =0 means the state of unit g has not changed at time t; L cur,t is the load shedding amount of the system at time t.

7. The method for generating typical scenarios for power system planning based on data drive according to claim 3 is characterized in that: The planning problem needs to satisfy the following constraints: Output constraints of thermal power units: a g S g,t G g ≤P g,t ≤S g,t G g Where: α g is the minimum output technical coefficient of thermal power unit g; G g P is the investment capacity of the generator set; g,t is the power generation of unit g at time t; S g,t is the binary variable of the running state of unit g at time t, S g,t =1 means the unit is in the on state, otherwise it is in the off state; Conventional unit start and stop constraints: S g,t -S g,t-1 =O g,t -C g,t Where: They are the actual startup time and minimum shutdown time of the unit at minimum output respectively; g,t and C g,t are whether the unit g performs startup and shutdown actions at time t, O g,t =1 means that the unit g starts up at time t, C g,t =1 means unit g stops at time t, O g,t =0 while C g,t =0 means the state of unit g does not change at time t; Conventional unit climbing constraints: Where: They are the upward and downward climbing capabilities of the unit respectively; Wind and solar power unit output constraints: 0≤P g,t ≤β g,t G g Where: β g,t is the capacity factor of the unit at time t, and its value varies between [0,1] according to different wind and solar output conditions. g Invest in capacity for generator sets.

8. The method for generating typical scenarios for power system planning based on data drive according to claim 3 is characterized in that: Step S3 is as follows: S31: Based on the capacity configuration obtained in step S2, a power system optimization operation model is constructed, the optimization variables are the unit output and the load shedding amount, and the optimization goal is to minimize the operation cost; S32: Put the full-time scenario into the optimization model to obtain the optimal operation plan. According to the operation optimization results, find the scenario with the maximum load shedding, regard it as an extremely unbalanced scenario, and extract the wind and solar power output and load data corresponding to this scenario; S33: re-cluster the scene set with the extremely unbalanced scene removed by K-mediods to obtain k typical scenes and their weights, and take the extremely unbalanced scene as the typical scene to obtain a reconstructed typical scene set; S34: Enter the loop, reapply the reconstructed typical scenarios and their probabilities to the power system planning expansion model to obtain the optimal planning capacity, and put the optimal planning capacity into the full-time operation model of the power system to identify extreme unbalanced scenarios. If there are extreme unbalanced scenarios, continue the loop and reconstruct the typical scenario set; if there are no extreme unbalanced scenarios, exit the loop.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored program, wherein when the program is running, the device where the computer-readable storage medium is located is controlled to execute the data-driven typical scenario generation method for power system planning as described in any one of claims 1 to 8.

10. A processor, characterized in that: The processor is used to run a program, wherein when the program is run, the data-driven method for generating typical scenarios for power system planning described in any one of claims 1 to 8 is executed.

Images