An active defense method and device for a new energy distribution system considering the uncertainty of typhoon and rainstorm evolution paths

Abstract

The present invention discloses an active defense method and device for a new energy distribution system considering the uncertainty of the evolution path of typhoon and heavy rain. The method includes: obtaining vulnerable components at each moment according to the time series scenario; calculating the load loss caused by disconnecting the vulnerable components of the system at each moment through a multi-objective optimization model; considering each quantity of the number of vulnerable components at this moment, and combining all selection cases when any number of vulnerable components at each moment with vulnerable load loss are selected; calculating the combined survival probability; operating the combinations of vulnerable components in each quantity at each moment in descending order of the survival probability from the smallest quantity to the largest quantity; rescheduling the system when each combination of vulnerable components operates; judging whether there is load loss in the system when each combination of vulnerable components is put into operation according to the input operation order. If not, perform active defense at this moment according to the combination of vulnerable components and active scheduling in this operation.

Description

Technical Field

[0001] The present invention relates to the technical field of power emergency, and in particular to an active defense method and device for a new energy distribution system considering the uncertainty of the evolution path of typhoon and rainstorm. Background Art

[0002] The new energy distribution system has the characteristics of strong weather correlation and weak disaster-bearing capacity. With the intensification of climate change, extreme natural weather events such as typhoons and rainstorms occur frequently, often causing serious damage to the new energy distribution system and even triggering large-scale power outages. Therefore, the research on pre-disaster prevention measures for the new energy distribution system to cope with extreme natural events is crucial, which is also an effective way to improve the ability of the power grid to resist extreme natural disasters.

[0003] Pre-disaster prevention measures for the distribution system generally include two aspects: long-term resilience investment and short-term pre-event preparation. Long-term resilience investment is a hardening-oriented measure. Among them, the reinforcement strategy of distribution infrastructure is the most direct means to cope with typhoons. In addition, installing underground cables, raising the location of substations, and carrying out vegetation management also belong to long-term resilience investment methods. However, although long-term resilience investment can improve the system's ability to cope with typhoons to a certain extent, it still cannot completely eliminate the threat brought by typhoons and is usually costly. Short-term pre-event preparation is operation-oriented, aiming to enable the system to make active and compliant configuration adjustments when facing extreme natural disasters rather than directly collapsing. Common short-term pre-event preparations mainly include two types: 1) pre-deployment of flexible resources, such as mobile power sources, remote switches, soft switches, and maintenance personnel; 2) active scheduling management. In recent years, active scheduling management has gradually received more and more attention. As an important means of short-term pre-event preparation, active scheduling management can not only enhance the pre-disaster risk resistance ability of the distribution system but also effectively improve the post-disaster recovery efficiency, providing strong support for resisting extreme natural disasters such as typhoons and rainstorms. Some existing studies have proposed an active management method for the distribution system considering the worst-case scenario, ensuring the normal operation of important loads in the system before the storm occurs and reducing the vulnerability of the system during the disaster. However, on the one hand, the operating state of the system is significantly affected by the intensity of typhoon and rainstorm disasters, but existing studies often ignore the impact of the uncertainty of typhoon and rainstorm disaster intensity on the accuracy of the evaluation of the operating state of system components, resulting in poor scheme effects. On the other hand, existing active scheduling schemes are usually designed based on the worst-case scenario affected by the maximum wind speed of typhoons, without fully considering the sequential impact of typhoon and rainstorm disasters on the gradual damage of system components, resulting in poor flexibility and insufficient economy of the scheme.

[0004] Therefore, there is an urgent need for a new technical solution to solve the technical problem of how to design an active prevention scheme for a new energy distribution system considering the uncertainty of the evolution path of typhoon and rainstorm. Summary of the Invention

[0005] The present invention provides an active defense method and device for a new energy distribution system considering the uncertainty of typhoon and rainstorm evolution paths, aiming to solve the technical problem of how to design an active prevention scheme for a new energy distribution system considering the uncertainty of typhoon and rainstorm evolution paths.

[0006] To achieve the above object, the present invention provides an active defense method for a new energy distribution system considering the uncertainty of typhoon and rainstorm evolution paths, including:

[0007] Obtain a dual-uncertainty time-series scenario coupling the typhoon and rainstorm evolution paths and the system component states; obtain the vulnerable components at each moment during the typhoon and rainstorm duration according to the time-series scenario.

[0008] Switch the system to island operation and keep the dispatchable distributed power sources in the system using droop control; calculate the vulnerable load losses caused by disconnecting all the vulnerable components of the system at each moment through a multi-objective optimization model.

[0009] For each quantity considering the total number of vulnerable components from 0 to the current moment, combine all the selection cases when any number of vulnerable components are selected respectively at each moment with vulnerable load losses; calculate the survival probability of the combination of vulnerable components.

[0010] In the order from the smallest quantity to the largest, put the combinations of vulnerable components of each quantity at each moment with vulnerable load losses into operation in turn according to the largest survival probability to the smallest, and disconnect the vulnerable components outside the combination of vulnerable components in operation.

[0011] Reschedule the system when each combination of vulnerable components is put into operation through active dispatch and a multi-objective optimization model. The active dispatch includes power generation redispatch, pre-disaster network reconstruction, optimized setting of droop control parameters, and protective voltage regulation.

[0012] Judge in turn whether there are vulnerable load losses in the system when each combination of vulnerable components is put into operation according to the order of putting into operation. If not, perform active defense at this moment according to the combination of vulnerable components put into operation this time and the active dispatch.

[0013] Preferably, calculating the vulnerable load losses caused by disconnecting all the vulnerable components of the system at each moment through a multi-objective optimization model includes:

[0014] Case 1: If there is no load loss in the distribution system itself, directly calculate the vulnerable load losses caused by disconnecting all the vulnerable components of the system at each moment through a multi-objective optimization model.

[0015] Case 2: When there is a load loss in the distribution system itself and the vulnerability load loss caused by disconnecting all vulnerable components of the system at each moment is calculated through the multi-objective optimization model, it is necessary to first calculate the total load loss of the distribution system, and then subtract the load loss of the system itself from the total loss of the distribution system to obtain the vulnerability load loss.

[0016] The load loss of the system itself includes the load loss of the distribution system excluding the vulnerability load loss caused by disconnecting the vulnerable components; the total load loss of the distribution system includes the load loss of the system itself and the vulnerability load loss.

[0017] Preferably, obtaining the dual-uncertainty time series scenario of the coupling of the typhoon and rainstorm evolution path and the system component state includes:

[0018] Construct a typhoon and rainstorm disaster uncertainty model and a system catastrophe dynamics model under typhoon and rainstorm.

[0019] Generate a dual-uncertainty time series scenario of the coupling of the typhoon and rainstorm evolution path and the system component state according to the typhoon and rainstorm disaster uncertainty model and the system catastrophe dynamics model under typhoon and rainstorm.

[0020] Preferably, the typhoon and rainstorm disaster uncertainty model includes two parts: a typhoon and rainstorm evolution uncertainty path model and a typhoon and rainstorm wind field model:

[0021] The typhoon and rainstorm evolution uncertainty path model includes:

[0022] To simulate the possible movement path after the typhoon and rainstorm landfall, the typhoon and rainstorm empirical path mechanism model is used to calculate the moving wind speed and the direction angle :

[0023] (1)

[0024] (2)

[0025] (3)

[0026] (4)

[0027] Among them, represents the logarithmic deviation of the moving wind speed of the typhoon and rainstorm center; represents the moving direction angle of the typhoon and rainstorm center; represents the logarithm of the moving wind speed of the typhoon and rainstorm center; and respectively represent the moving wind speed and moving direction angle of the typhoon and rainstorm center at time and respectively represent The latitude and longitude of the typhoon rainstorm center at a certain moment; and are respectively the prediction errors subject to Gaussian distribution; and are the fitting coefficients obtained from historical data of different regions; is the time interval;

[0028] Based on the typhoon rainstorm empirical path mechanism model combined with the data-driven method, the prediction errors are processed through historical data, including:

[0029] Assume that the typhoon rainstorm path unfolds around the predicted path, and the prediction error follows a Gaussian distribution with a mean of zero; the uncertainty of the typhoon rainstorm path is quantified by the deviation between the empirical path and the historical true path, including the first calculation.

[0030] The first calculation includes: selecting two adjacent typhoon rainstorm positions A1 and A2 from the historical true path; calculating the empirical path from A1 to A3 through equations (1) to (4); there is a prediction error of the moving direction angle between the historical true path from A1 to A2 and the empirical path from A1 to A3 ; the predicted moving speed and moving direction angle of the empirical path from A1 to A3 can be calculated according to the coordinates of A1 and A3; the historical true moving speed and moving direction angle of the historical true path from A1 to A2 are obtained; the error between the empirical path from A1 to A3 and the historical true path from A1 to A2 can be obtained according to the predicted moving speed and moving direction angle and the historical true moving speed and moving direction angle, that is, the central moving wind speed prediction error and the prediction error of the moving direction angle .

[0031] Perform the first calculation on each historical true path to obtain the corresponding and for each historical true path, and then obtain the central moving wind speed prediction error set and the moving direction angle prediction error set .

[0032] Add and the prediction errors in to the predicted values and calculated by equations (1) and (2), and then various typhoon rainstorm path scenarios are obtained.

[0033] The typhoon rainstorm wind field model includes:

[0034] For the typhoon rainstorm path scenario moving on the ocean surface, the first model can be used to estimate the time-series wind speed at different locations , and the first model includes:

[0035] (5)

[0036] (6)

[0037] (7)

[0038] (8)

[0039] (9)

[0040] (10)

[0041] (11)

[0042] (12)

[0043] Among them, represents the circulation wind speed; and respectively represent at the moment the typhoon rainstorm wind speed and rainfall rate at the location; , and are correction factors; and are respectively the distance of the power distribution system from the typhoon rainstorm center and the maximum wind speed radius; and are the central pressure difference and the rate of change; and are respectively the coordinate positions of the power distribution system and the typhoon rainstorm center; is the duration of the typhoon rainstorm disaster; is at the moment the typhoon wind speed based on the Batts model at the location; represents the maximum circulation wind speed at the moment; represents the moving wind speed of the typhoon rainstorm center.

[0044] For the typhoon rainstorm path scenario moving on land, after the typhoon rainstorm makes landfall, affected by friction, the wind speed decreases. Therefore, an attenuation coefficient is introduced on the basis of the first model. It is assumed that the sequential wind speed of the typhoon rainstorm moving on the ocean surface calculated by the first model is , and it is assumed that the sequential wind speed of the typhoon rainstorm moving on land is , then the attenuation coefficient can be expressed as:

[0045] (13)

[0046] Among them, and represent the roughness length and the drag factor respectively.

[0047] According to the geographical location where the typhoon rainstorm is located, or is selected to calculate the time-series wind speed.

[0048] Preferably, the system catastrophe dynamics model under typhoon rainstorm includes two parts: the system component failure probability model and the photovoltaic and wind turbine output models:

[0049] The system component failure probability model includes:

[0050] Assume that the system component failures are mainly caused by wind speed. The system component failures include damage to distribution poles, faults in overhead distribution lines, and faults in photovoltaic components; assume that different components of the distribution line are independent of each other, then the failure probability of the distribution line can be calculated through a series model, which can be expressed as:

[0051] Preferably, the system catastrophe dynamics model under typhoon rainstorm includes two parts: the system component failure probability model and the photovoltaic and wind turbine output models:

[0052] (14)

[0053] (15)

[0054] (16)

[0055] Among them, and represent the failure probabilities of the distribution line, pole, and wire respectively; and represent the number of poles and the number of wire segments of the line respectively; and represent the pole and wire segment of the line respectively.

[0056] The failure probability of the photovoltaic component is modeled using the standard lognormal distribution function, which can be expressed as:

[0057] (17)

[0058] Among them, represents the standard lognormal distribution function; represents the failure probability of the photovoltaic component; and represent the standard deviation and the designed wind speed of the photovoltaic respectively.

[0059] The photovoltaic and wind turbine output models include:

[0060] Under typhoon and rainstorm conditions, the available photovoltaic output of intact equipment can be expressed as:

[0061] (18)

[0062] (19)

[0063] where represents the impact of typhoon and rainstorm disasters on light irradiance in the logarithmic space; is the light irradiance under normal conditions; and are the typhoon intensity, normalized distance, photovoltaic power generation efficiency, and area, respectively; represents the set of photovoltaic nodes.

[0064] Since wind turbines usually have the ability to resist typhoon and rainstorm, it is assumed that the wind turbines do not suffer damage or failure during typhoon and rainstorm disasters; the wind field changes caused by typhoon and rainstorm directly affect the wind power output, and the available wind power output can be calculated through the wind speed-power curve of the wind turbine , which can be expressed as:

[0065] (20)

[0066] where and represent the cut-in wind speed, rated wind speed, and cut-out wind speed of the wind turbine, with values of 3 m / s, 12 m / s, and 25 m / s, respectively; and represent the air density, power coefficient, swept area of the wind turbine blades, and rated power, respectively; represents the set of wind turbine nodes.

[0067] Preferably, the dual-uncertainty time-series scenarios of the typhoon and rainstorm evolution path and the system component state coupling generated according to the typhoon and rainstorm disaster uncertainty model and the system catastrophic dynamics model under typhoon and rainstorm include:

[0068] Obtain the coordinates of the distribution system;

[0069] Obtain the coordinates, moving wind speed, and moving direction angle of the typhoon and rainstorm center at this moment, and calculate the moving wind speed and moving direction angle of the typhoon and rainstorm center at the next moment through the typhoon and rainstorm empirical path mechanism model, namely, equations (1) to (4).

[0070] Take the prediction errors and , add the moving wind speed and moving direction angle of the typhoon rainstorm center at the next moment to the prediction error and correspondingly to obtain the coordinates of the typhoon rainstorm center at the next moment; continuously calculate the moving wind speed, moving direction angle and coordinates of the typhoon rainstorm center at subsequent moments in sequence until the maximum number of scenarios of the typhoon rainstorm path sample reaches S trackmax。

[0071] Adopt the scenario backward reduction algorithm to reduce the number of scenarios S trackmax to S reduced , and record the occurrence probability s ∈ S reduced corresponding to each typical scenario .

[0072] Perform a second calculation for each moment during the continuous period of the typhoon rainstorm. The second calculation includes:

[0073] Calculate the distance from the typhoon rainstorm center to the power distribution system in each typical scenario according to Equation (10) and the coordinates of the power distribution system , where take 1, 2… S reduced ; then the chronological expected distance considering all typical scenarios can be expressed as .

[0074] Obtain the expected predicted wind speed at the power distribution system according to Equations (7) to (9) and Equations (11) to (13) .

[0075] Calculate the expected photovoltaic output and the expected wind power output respectively according to Equations (18) and (20).

[0076] Calculate the failure probability of the photovoltaic and the failure probability of the power distribution line respectively according to Equations (14) and (17).

[0077] Use the vulnerability threshold as the standard for determining vulnerable components.

[0078] Power distribution line Vulnerability status , 0 indicates a vulnerable line and 1 indicates a normal line.

[0079] Photovoltaic component Vulnerability status , 0 represents a vulnerable photovoltaic component, and 1 represents a sound photovoltaic component.

[0080] Obtain the uncertainty vulnerability status of all components in the distribution system at this moment, as well as the uncertainty output of photovoltaic and wind power.

[0081] When the second calculation is completed for each moment during the typhoon and rainstorm duration, the time series scenario is obtained.

[0082] Preferably, considering each quantity from 0 to the total number of vulnerable components at this moment, for each moment with vulnerable load loss, the vulnerable components are respectively considered for all selection cases when any number of components are selected for combination; calculating the survival probability of the vulnerable component combination includes:

[0083] Assume that a certain moment with vulnerable load loss includes vulnerable components, being a positive integer; select components from vulnerable components, being an integer, ; obtain kinds of vulnerable component combinations; calculate kinds of survival probabilities of the vulnerable component combinations and sort them from the largest to the smallest according to the survival probability.

[0084] The calculation of the survival probability includes:

[0085] (21)

[0086] Wherein, represents the survival probability; represents the vulnerable component combination; represents the distribution line number; represents the distribution line failure probability; represents the photovoltaic equipment number; represents the photovoltaic failure probability; represents the moment; represents the typhoon and rainstorm duration.

[0087] Perform the above-mentioned all possible selection combinations for each moment with vulnerable load loss, and calculate the survival probabilities of all vulnerable component combinations.

[0088] Preferably, the multi-objective optimization model includes an objective function and constraint conditions;

[0089] The multi-objective optimization model aims to minimize the overall operating cost, including minimizing the penalty cost generated by load shedding and reducing the operating costs of dispatchable distributed power sources, photovoltaics, and wind power; the objective function of the multi-objective optimization model can be expressed as:

[0090] (22)

[0091] (23)

[0092] (24)

[0093] (25)

[0094] (26)

[0095] (27)

[0096] (28)

[0097] Among them, represents the starting time of typhoon rainstorm; represents the ending time of typhoon rainstorm; and respectively represent the load shedding penalty cost and the power generation operation cost, where the power generation operation cost includes the photovoltaic operation cost , the wind power operation cost , the dispatchable distributed power source operation cost and the mobile energy storage operation cost ; and respectively represent the unit load shedding penalty cost, the unit photovoltaic power generation operation cost, the unit wind power generation operation cost, the unit mobile energy storage operation cost and the unit dispatchable distributed power source operation cost; represents the active power of the load at time is a binary decision variable, where "1" represents load restoration, otherwise it is load shedding; and respectively represent the power generation of photovoltaic, wind power, distributed power source and mobile energy storage at time , all of which are continuous decision variables; and respectively represent the node set, the photovoltaic node set, the wind turbine node set, the distributed power source node set and the mobile energy storage node set; and respectively represent the indexes of the nodes where the load, photovoltaic, wind turbine, distributed power source and mobile energy storage are located; and are weight coefficients, and .

[0098] The constraint conditions of the multi-objective optimization model include power flow and secure operation constraints, photovoltaic and wind turbine operation constraints, a load demand model based on static voltage, operation constraints of dispatchable distributed power sources and mobile energy storage based on droop control characteristics, and radial topology constraints based on an improved spanning tree.

[0099] The power flow and secure operation constraints include:

[0100] Modeling is carried out using the optimal power flow equation based on Disflow. The AC power flow equations are given by formulas (29), (30), and (31), the line capacity constraint is represented by formula (32), and the operation security constraints are given by formulas (33) and (34):

[0101] (29)

[0102] (30)

[0103] (31)

[0104] (32)

[0105] (33)

[0106] (34)

[0107] Among them, represents the active power injected at node at time represents the reactive power injected at node at time and respectively represent the active power on lines and ; and respectively represent the reactive power on lines and ; and respectively represent the current on lines and ; and respectively represent the reactance of lines and ; and respectively represent the resistance of lines and ; represents the load reactive power at time and respectively represent the active output power of distributed power sources, photovoltaics, wind turbines, and mobile energy storage at the time node ; represents the reactive output power of the distributed power source at the time node ; is a binary variable that takes the value of 1 when the line is energized and 0 otherwise; and respectively represent the voltage amplitude and system frequency at node ; represents the voltage amplitude at node ; is a relatively large positive number; and are respectively the maximum values of node voltage, system frequency, branch current, branch active power, and reactive power; and are respectively the minimum values of node voltage, system frequency, branch current, branch active power, and reactive power; is the set of lines.

[0108] The operating constraints of photovoltaics and wind turbines include:

[0109] Photovoltaics and wind turbines are modeled as PQ control resources with a unity power factor; it is assumed that photovoltaic units and wind turbines have strong anti-typhoon and heavy rain characteristics, that is, the design wind speed is higher than the maximum wind speed of typhoons and heavy rains, and according to vulnerability theory, photovoltaic units and wind turbines will not fail during typhoons and heavy rains; at the same time, it is assumed that both photovoltaics and wind turbines are equipped with energy storage and converters, and each photovoltaic unit and wind turbine is regarded as an independent whole, ignoring the detailed models of energy storage and converters; considering the actual available output of photovoltaics and wind turbines during typhoons and heavy rains; therefore, the operating constraints of photovoltaics and wind turbines can be respectively expressed as:

[0110] (35)

[0111] (36)

[0112] Among them, and respectively represent the expected output of photovoltaics and wind power; and respectively represent the power generation of photovoltaics and wind power at time

[0113] The load demand model based on static voltage includes:

[0114] To describe the relationship between load and voltage amplitude, a polynomial is used Load model:

[0115] (37)

[0116] (38)

[0117] where and are the participation factors of the constant impedance, current, and power terms in the active power load, respectively; , and are the participation factors of the constant impedance, current, and power terms in the reactive power load, respectively; and represent the rated active power and reactive power of the load, respectively.

[0118] The operating constraints of dispatchable distributed power sources and mobile energy storage based on droop control characteristics include:

[0119] Dispatchable distributed power sources using droop control need to satisfy capacity and power constraints, as well as droop control characteristics of active - frequency and reactive - voltage, including:

[0120] (39)

[0121] (40)

[0122] (41)

[0123] (42)

[0124] (43)

[0125] (44)

[0126] (45)

[0127] (46)

[0128] (47)

[0129] where and represent the maximum active power output, maximum reactive power output, and total output capacity of the dispatchable distributed power source, respectively; and represent the node at time Decision variables for droop control of distributed power sources at the location; and represent the reference voltage amplitude and reference frequency decision variables respectively; and represent the maximum values of the reference frequency, reference voltage, active droop coefficient, and reactive droop coefficient respectively; and represent the minimum values of the reference frequency, reference voltage, active droop coefficient, and reactive droop coefficient respectively; represents the angular frequency of the system.

[0130] In addition to the controllable distributed power sources with droop control functions, mobile energy storage is also equipped with a droop controller; assuming that the mobile energy storage provides as much active power output as possible, therefore, the mobile energy storage charges or discharges according to the droop characteristics, which can be expressed as:

[0131] (48)

[0132] The charging or discharging action of the mobile energy storage is restricted by the rated capacity and can be expressed as:

[0133] (49)

[0134] Among them, is the rated capacity of the mobile energy storage; represents the power generation of the mobile energy storage at time

[0135] The radial topology constraints based on the improved spanning tree include:

[0136] During typhoon and rainstorm disasters, multiple islands may be formed in the distribution system. The improved spanning tree constraints are adopted to meet the radial topology requirements of each island, including:

[0137] (50)

[0138] (51)

[0139] Among them, and are binary variables; if node is a substation node, then takes the value of 1, otherwise 0; if node is a distributed power source node, then takes the value of 1, otherwise 0; is a binary variable, indicating whether node is the root bus in the island. If so, it takes the value of 1, otherwise 0; and are the parent node and the set of child nodes of the busbar respectively ; is a binary variable, with a value of 1 indicating that the branch is energized, and a value of 0 indicating that the branch is not energized; and are binary variables. If the node is the parent node of the node , that is , or the busbar is the parent node of the busbar , that is , then select the branch to be energized, that is, take , and cannot be 1 at the same time; represents a binary variable, indicating whether the node is the parent node of the node , when the node is the parent node of the node .

[0140] Preferably, the multi-objective optimization model further includes:

[0141] Convex optimization solution of the multi-objective optimization model:

[0142] There are non-linear equations in the multi-objective optimization model, including those in equations (29), (31), (37) and (38) , , those in equation (42) , those in equation (43) , those in equation (48) , as well as equations (32) and (41); the non-linear equations existing in the multi-objective optimization model can be divided into three types: the first type is the square of continuous variables, the second type is quadratic equations or quadratic inequalities, and the third type is the product between continuous variables; perform convex optimization on these three types of non-linear equations:

[0143] Convex optimization of the first type of non-linear equation:

[0144] Since equations (29), (31), (37) and (38) contain square terms of branch current and node voltage and , use the variables and to replace these two quadratic terms respectively, and obtain:

[0145] (52)

[0146] (53)

[0147] (54)

[0148] (55)

[0149] Among them, and are the square terms of the node voltage and the branch current, respectively; and are the square terms of the node voltage and the branch current, respectively.

[0150] Since equations (54) and (55) contain square terms , they are still non - linear expressions. Therefore, Taylor series expansion is used to linearize these equations (54) and (55), including:

[0151] Assume the initial value is , and through Taylor series expansion, we can get:

[0152] (56)

[0153] Among them, represents an infinitesimal quantity; is the initial value of the magnitude of the node voltage.

[0154] By ignoring the second - order and higher - order terms, equation (56) can be approximately expressed as:

[0155] (57)

[0156] The magnitude of the node voltage is usually about 1.0 p.u. Therefore, assuming , equation (57) can be further approximated as:

[0157] (58)

[0158] The definition of the distance function of Taylor series approximation includes:

[0159] (59)

[0160] Among them, is the distance function of Taylor series approximation.

[0161] By substituting Equation (58) into Equations (54) and (55), an approximate load demand model is obtained, which can be expressed as:

[0162] (60)

[0163] (61)

[0164] Convex optimization of the second type of nonlinear equation:

[0165] Equations (32) and (41) can be converted into second-order cone constraints and rotated second-order cone constraints respectively, which can be expressed as:

[0166] (62)

[0167] (63)

[0168] Convex optimization of the third type of nonlinear equation:

[0169] Since Equations (42), (43) and (48) all contain the product of two bounded continuous variables; the McCormick envelope approximation is used for processing, and auxiliary variables and are introduced to represent and respectively, and twelve linear constraints are used for approximation, which can be expressed as:

[0170] (64)

[0171] (65)

[0172] (66)

[0173] (67)

[0174] (68)

[0175] (69)

[0176] (70)

[0177] (71)

[0178] (72)

[0179] (73)

[0180] (74)

[0181] (75)

[0182] The present invention also provides an active defense device for a new energy distribution system considering the uncertainty of the typhoon and rainstorm evolution path, which is used for the method of the present invention. The device includes a first module, a second module, a third module, a fourth module, a fifth module and a sixth module.

[0183] The first module is used to obtain the dual-uncertainty time series scenarios of the typhoon and rainstorm evolution path and the system component status; and obtain the vulnerable components at each moment during the typhoon and rainstorm duration according to the time series scenarios.

[0184] The second module is used to switch the system to island operation and keep the dispatchable distributed power sources in the system using droop control; and calculate the vulnerable load losses caused by disconnecting all the vulnerable components of the system at each moment through a multi-objective optimization model.

[0185] The third module is used to consider each quantity from 0 to the total number of vulnerable components at this moment, and combine all the selection cases when any number of vulnerable components at each moment with vulnerable load losses are selected respectively; and calculate the survival probability of the combination of vulnerable components.

[0186] The fourth module is used to, in the order from the smallest quantity to the largest quantity, put the combinations of vulnerable components of each quantity at each moment with vulnerable load losses into operation in turn according to the largest survival probability to the smallest, and disconnect the vulnerable components outside the combination of vulnerable components in operation.

[0187] The fifth module is used to reschedule the system when each combination of vulnerable components is put into operation through active scheduling and a multi-objective optimization model. The active scheduling includes power generation rescheduling, pre-disaster network reconstruction, optimization setting of droop control parameters and protective voltage regulation.

[0188] The sixth module is used to sequentially judge whether there are vulnerable load losses in the system when each combination of vulnerable components is put into operation according to the operation order. If not, perform active defense at this moment according to the combination of vulnerable components put into operation this time and the active scheduling.

[0189] The present invention has the following beneficial effects:

[0190] The active defense method for a new energy distribution system considering the uncertainty of typhoon and rainstorm evolution paths in the present invention forms a robust scheduling strategy by integrating dual uncertainty scenarios in advance, aiming to ensure that under the condition of full-load operation of the system, the number of outage-vulnerable components is maximized, thereby effectively reducing the overall vulnerability of the system during typhoon and rainstorm disasters. Switching the system to island operation not only helps reduce the load loss of the distribution system during disasters, improve the safety of the system, but also reduces the protection coordination requirements with the main grid and alleviates other technical problems that may be caused during disasters. By using a multi-objective model, the vulnerability load loss caused by disconnecting vulnerable components at each moment is calculated. For each moment with vulnerability load loss, all possible combinations of vulnerable components are considered and combined, the survival probability of each combination is calculated, and the trial operation starts from the combination with the smallest number, and then the combination of vulnerable components with the smallest number of vulnerable components at each moment with vulnerability load loss is found and put into operation, improving the safety and robustness of the system. The method of the present invention combines mechanism analysis and data-driven models and adopts a robust strategy for optimization. Compared with the active management method that does not consider the typhoon and rainstorm evolution paths, the method of the present invention demonstrates the best comprehensive benefits. In the worst case, compared with the post-disaster scheduling results of the active management method and the passive strategy method that do not consider the typhoon and rainstorm evolution paths, the method of the present invention effectively reduces the load loss and improves the satisfaction rate of load demand. The effect, flexibility, and economy of the active defense plan obtained according to the method of the present invention all have good performances.

[0191] The active defense device for a new energy distribution system considering the uncertainty of typhoon and rainstorm evolution paths in the present invention is used for the method of the present invention and has the same beneficial effects as the method of the present invention.

[0192] In addition to the purposes, features, and advantages described above, the present invention has other purposes, features, and advantages. The following will refer to the accompanying drawings to further elaborate on the present invention in detail. Description of the Drawings

[0193] The drawings constituting a part of the present invention are used to provide a further understanding of the present invention. The schematic embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0194] Figure 1 is a schematic flowchart of a preferred embodiment of the present invention.

[0195] Figure 2 is a schematic diagram of the prediction error of the simulated typhoon and rainstorm path in a preferred embodiment of the present invention.

[0196] Figure 3 is a schematic diagram of the prediction error of the moving wind speed in a preferred embodiment of the present invention.

[0197] Figure 4 It is a schematic diagram of the prediction error of the moving direction angle in the preferred embodiment of the present invention.

[0198] Figure 5 It is a curve graph of the Taylor series approximation error in the preferred embodiment of the present invention.

[0199] Figure 6 It is a schematic diagram of the typical typhoon rainstorm movement path of the test system and simulation in the preferred embodiment of the present invention.

[0200] Figure 7 It is a schematic diagram of the modified IEEE 33-node test system in the preferred embodiment of the present invention.

[0201] Figure 8 It is a curve graph of the wind-induced vulnerability of the overhead distribution line and photovoltaic in the preferred embodiment of the present invention.

[0202] Figure 9 It is a schematic diagram of the network topology structure of the active defense of the distribution system at time t5 in the preferred embodiment of the present invention.

[0203] Figure 10 It is a schematic diagram of the network topology structure of the active defense of the distribution system at time t6 in the preferred embodiment of the present invention.

[0204] Figure 11 It is a schematic diagram of the network topology structure of the active defense of the distribution system at time t7 in the preferred embodiment of the present invention.

[0205] Figure 12 It is a schematic diagram of the active power output of the distributed power source in the preferred embodiment of the present invention.

[0206] Figure 13 It is a schematic diagram of the active power output of the fan in the preferred embodiment of the present invention.

[0207] Figure 14 It is a schematic diagram of the active power output of the photovoltaic in the preferred embodiment of the present invention.

[0208] Figure 15 It is a schematic diagram of the active power output of the mobile energy storage in the preferred embodiment of the present invention. Detailed implementation manners

[0209] The embodiments of the present invention will be described in detail below with reference to the accompanying drawings. However, the present invention can be implemented in many different ways defined and covered by the claims.

[0210] In the preferred embodiment of the present invention, during active defense, the distribution system maintains an island operation mode; since the variation range of the steady-state frequency is usually extremely small (about ±0.005 p.u.), the influence of frequency variation on the bus admittance is negligible; the dispatchable distributed power sources adopt traditional active-frequency and reactive-voltage droop control, mainly due to the influence of the coupling inductance used at the output end of the distributed power source interface converter, the network feeder impedance ratio, and the virtual inductive output impedance, to ensure the frequency and voltage stability of the island system; the droop control and damping device of the distribution system can effectively control the transient stability and dynamic oscillation of the system.

[0211] See Figure 1 , in the preferred embodiment of the present invention, an active defense method for a new energy distribution system considering the uncertainty of typhoon and rainstorm evolution paths is provided, including:

[0212] F1. Obtain the dual-uncertainty time-series scenarios of the coupling of typhoon and rainstorm evolution paths and system component states; obtain the vulnerable components at each moment during the typhoon and rainstorm duration according to the time-series scenarios.

[0213] In the preferred embodiment of the present invention, obtaining the dual-uncertainty time-series scenarios of the coupling of typhoon and rainstorm evolution paths and system component states includes:

[0214] Construct a typhoon and rainstorm disaster uncertainty model and a system catastrophe dynamics model under typhoon and rainstorm.

[0215] Generate the dual-uncertainty time-series scenarios of the coupling of typhoon and rainstorm evolution paths and system component states according to the typhoon and rainstorm disaster uncertainty model and the system catastrophe dynamics model under typhoon and rainstorm.

[0216] In the preferred embodiment of the present invention, the typhoon and rainstorm disaster uncertainty model includes two parts: a typhoon and rainstorm evolution uncertainty path model and a typhoon and rainstorm wind field model:

[0217] The typhoon and rainstorm evolution uncertainty path model includes:

[0218] In the active defense of the system, the dispatcher usually receives early warning information about typhoon and rainstorm activities, and at the same time, the meteorological department will also provide prediction data on typhoon paths. The dispatcher can use this information to evaluate the possible impact of typhoon and rainstorm disasters on the distribution system and formulate corresponding active defense plans and strategies. To simulate the possible movement paths after the typhoon and rainstorm landfall, the typhoon and rainstorm empirical path mechanism model is used to calculate the moving wind speed and the direction angle :

[0219] (1)

[0220] (2)

[0221] (3)

[0222] (4)

[0223] wherein, represents the logarithmic deviation of the moving wind speed of the typhoon rainstorm center; represents the moving direction angle of the typhoon rainstorm center; represents the logarithm of the moving wind speed of the typhoon rainstorm center; and respectively represent the moving wind speed and moving direction angle of the typhoon rainstorm center at time and respectively represent the latitude and longitude of the typhoon rainstorm center at time and are respectively the prediction errors subject to Gaussian distribution; and are the fitting coefficients obtained from historical data in different regions; is the time interval, usually taken as 1 hour.

[0224] When using the above model for typhoon rainstorm path prediction, prediction errors will inevitably occur. As time goes by, these errors affecting the moving speed and moving direction angle of the typhoon rainstorm center gradually accumulate, resulting in an increase in the uncertainty of the typhoon rainstorm path. These uncertainties cause deviations between the actual state and the predicted value of the components in the power distribution system, which may lead to the failure of existing active defense methods.

[0225] To solve the problem of the uncertainty of the typhoon rainstorm path, based on the typhoon rainstorm empirical path mechanism model combined with the data-driven method, the prediction errors are processed through historical data, including:

[0226] Assume that the typhoon rainstorm path unfolds around the predicted path, and the prediction error follows a Gaussian distribution with a mean of zero; the uncertainty of the typhoon rainstorm path is quantified by the deviation between the empirical path and the historical true path, including the first calculation.

[0227] The first calculation includes: selecting two adjacent typhoon rainstorm positions A1 and A2 from the historical true path; calculating the empirical path from A1 to A3 through equations (1) to (4); see Figure 2 that there is a prediction error in the moving direction angle between the historical true path (actual historical trajectory) from A1 to A2 and the empirical path from A1 to A3 and there is also a displacement deviation ; The predicted moving speed and moving direction angle of the empirical path from A1 to A3 can be calculated based on the coordinates of A1 and A3; the historical true moving speed and moving direction angle of the historical true path from A1 to A2 are obtained; the error between the empirical path from A1 to A3 and the historical true path from A1 to A2, that is, the central moving wind speed prediction error, can be obtained based on the predicted moving speed and moving direction angle and the historical true moving speed and moving direction angle and the prediction error of the moving direction angle .

[0228] Perform the first calculation on each historical true path to obtain the and corresponding to each historical true path, and then obtain the central moving wind speed prediction error set and the moving direction angle prediction error set , see Figure 3 and Figure 4 respectively.

[0229] Add the prediction errors in and to the predicted values and calculated by equations (1) and (2), and then obtain various typhoon rainstorm path scenarios.

[0230] The typhoon rainstorm wind field model includes:

[0231] For the typhoon rainstorm path scenario moving on the ocean surface, the first model can be used to estimate the time-series wind speed at different locations , and the first model includes:

[0232] (5)

[0233] (6)

[0234] (7)

[0235] (8)

[0236] (9)

[0237] (10)

[0238] (11)

[0239] (12)

[0240] Among them, represents the circulation wind speed; and respectively represent time the typhoon rainstorm wind speed and rainfall rate at the location; 、 and are correction factors; and are respectively the distance of the power distribution system from the typhoon rainstorm center and the maximum wind speed radius; and are the central pressure difference and change rate; and are respectively the coordinate positions of the power distribution system and the typhoon rainstorm center; is the duration of the typhoon rainstorm disaster; is time the typhoon wind speed based on the Batts model at the location; represents the maximum circulation wind speed at time; represents the moving wind speed of the typhoon rainstorm center.

[0241] For the typhoon rainstorm path scenario moving on land, after the typhoon rainstorm lands, affected by friction, the wind speed decreases. Therefore, an attenuation coefficient is introduced on the basis of the first model. Assuming that the sequential wind speed of the typhoon rainstorm moving on the ocean surface calculated by the first model is , and assuming that the sequential wind speed of the typhoon rainstorm moving on land is , then the attenuation coefficient can be expressed as:

[0242] (13)

[0243] wherein, and respectively represent the roughness length and the drag factor, and the usual values are 0.005m and 0.85 respectively, then .

[0244] According to the geographical location of the typhoon rainstorm, or is selected for the calculation of the sequential wind speed.

[0245] In the preferred embodiment of the present invention, the system catastrophe dynamics model under typhoon rainstorm includes two parts: the system component failure probability model and the photovoltaic and wind turbine output models:

[0246] The system component failure probability model includes:

[0247] Assuming that the system component failures are mainly caused by wind speed, the system component failures include damage to distribution poles, overhead distribution wire failures, and photovoltaic component failures; assuming that different components of the distribution line are independent of each other, then the failure probability of the distribution line can be calculated by a series model, which can be expressed as:

[0248] (14)

[0249] (15)

[0250] (16)

[0251] Among them, and respectively represent the failure probabilities of the distribution line, pole tower, and conductor; and respectively represent the number of pole towers and the number of conductor segments of the line; and respectively represent the pole tower of the line and the conductor segment failure probability.

[0252] Modeling the failure probability of photovoltaic components using the standard lognormal distribution function can be expressed as:

[0253] (17)

[0254] Among them, represents the standard lognormal distribution function; represents the failure probability of the photovoltaic component; and respectively represent the standard deviation and the designed wind speed of the photovoltaic.

[0255] The photovoltaic and wind turbine output models include:

[0256] Under typhoon and heavy rain, the available output of intact photovoltaics can be expressed as:

[0257] (18)

[0258] (19)

[0259] Among them, represents the impact of typhoon and heavy rain disasters on light irradiance in the logarithmic space; is the light irradiance under normal conditions; and are the typhoon intensity, normalized distance, photovoltaic power generation efficiency, and area respectively; represents the set of photovoltaic nodes.

[0260] Since wind turbines usually have the ability to resist typhoons and heavy rains, it is assumed that the wind turbines do not suffer damage or failure during typhoon and heavy rain disasters; the wind field changes caused by typhoon and heavy rain directly affect the wind power output, and the available wind power output can be specifically calculated through the wind speed-power curve of the wind turbine. , which can be expressed as:

[0261] (20)

[0262] Among them, and respectively represent the cut-in wind speed, rated wind speed and cut-out wind speed of the wind turbine, and the values are 3 m / s, 12 m / s and 25 m / s respectively; and respectively represent air density, power coefficient, swept area of the wind turbine blades and rated power; represents the set of wind turbine nodes.

[0263] In the preferred embodiment of the present invention, the dual-uncertainty time series scenarios coupling the typhoon and heavy rain evolution path and the system component state are generated according to the typhoon and heavy rain disaster uncertainty model and the system catastrophic dynamics model under typhoon and heavy rain, including:

[0264] Obtain the coordinates of the distribution system.

[0265] Obtain the coordinates, moving wind speed and moving direction angle of the typhoon and heavy rain center at this moment, and calculate the moving wind speed and moving direction angle of the typhoon and heavy rain center at the next moment through the typhoon and heavy rain empirical path mechanism model, namely formulas (1) to (4).

[0266] Take the prediction errors and , and add the moving wind speed and moving direction angle of the typhoon and heavy rain center at the next moment to the prediction errors and correspondingly to obtain the coordinates of the typhoon and heavy rain center at the next moment; continuously calculate the moving wind speed, moving direction angle and coordinates of the typhoon and heavy rain center at subsequent moments in sequence until the maximum number of scenarios of the typhoon and heavy rain path sample reaches S trackmax .

[0267] Adopt the scenario backward reduction algorithm to reduce the number of scenarios S trackmax to S reduced , and record the occurrence probability s ∈ S reduced corresponding to each typical scenario .

[0268] Perform a second calculation for each moment during the duration of typhoon rainstorms. The second calculation includes:

[0269] Calculate the distance from the typhoon rainstorm center to the power distribution system for each typical scenario according to Equation (10) and the coordinates of the power distribution system , where Take 1, 2… S reduced ; Then, considering the chronological expected distances in all typical scenarios, it can be expressed as .

[0270] Obtain the expected predicted wind speed at the power distribution system according to Equations (7) to (9) and Equations (11) to (13) .

[0271] Calculate the expected output of photovoltaic and the expected output of wind power respectively according to Equations (18) and (20).

[0272] Calculate the failure probability of photovoltaic and the failure probability of the distribution line respectively according to Equations (14) and (17).

[0273] Use the vulnerability threshold as the standard for determining vulnerable components.

[0274] Distribution line Vulnerability status , 0 represents a vulnerable line, and 1 represents a normal line.

[0275] Photovoltaic component Vulnerability status , 0 represents a vulnerable photovoltaic component, and 1 represents an intact photovoltaic component.

[0276] Obtain the uncertainty vulnerability status of all components of the power distribution system at this moment, as well as the uncertainty output situations of photovoltaic and wind power;

[0277] When the second calculation for each moment during the duration of typhoon rainstorms is completed, a chronological scenario is obtained.

[0278] F2. Switch the system to island operation and keep the dispatchable distributed power sources in the system under droop control; calculate the vulnerable load losses caused by disconnecting all vulnerable components of the system at each moment through a multi-objective optimization model.

[0279] In the preferred embodiment of the present invention, calculating the vulnerable load losses caused by disconnecting all vulnerable components of the system at each moment through a multi-objective optimization model includes:

[0280] Case 1: If there is no load loss in the distribution system itself, directly calculate the vulnerable load loss caused by disconnecting all vulnerable components of the system at each moment through the multi-objective optimization model.

[0281] Case 2: If there is load loss in the distribution system itself, when calculating the vulnerable load loss caused by disconnecting all vulnerable components of the system at each moment through the multi-objective optimization model, it is necessary to first calculate the total load loss of the distribution system, and then subtract the load loss of the system itself from the total loss of the distribution system to obtain the vulnerable load loss.

[0282] The load loss of the system itself includes the load loss of the distribution system excluding the vulnerable load loss caused by disconnecting the vulnerable components; the total load loss of the distribution system includes the load loss of the system itself and the vulnerable load loss.

[0283] In addition, when calculating the vulnerable load loss caused by disconnecting all vulnerable components of the system at each moment through the multi-objective optimization model, if there is no load loss at a certain moment, directly disconnect all vulnerable components at that moment.

[0284] If there are no vulnerable components at a certain moment, the system operates normally at that moment.

[0285] F3. Considering each quantity from 0 to the total number of vulnerable components at this moment, for each moment with vulnerable load loss, when selecting any number of vulnerable components, consider all selection cases for combination; calculate the survival probability of the combination of vulnerable components.

[0286] In the preferred embodiment of the present invention, S3 specifically includes:

[0287] Assume that a certain moment with vulnerable load loss includes vulnerable components, is a positive integer; select from vulnerable components, is an integer, ; obtain combinations of vulnerable components. It should be noted that there is only one possible corresponding component configuration, that is, all vulnerable components are out of service. Therefore, in the first subsequent iteration, apply the active scheduling to this unique configuration.

[0288] Calculate the survival probabilities of the combinations of vulnerable components and sort them from the largest to the smallest survival probability.

[0289] The calculation of the survival probability includes:

[0290] (21)

[0291] Among them, represents the survival probability; represents the combination of vulnerability components; represents the distribution line number; represents the failure probability of the distribution line; represents the photovoltaic equipment number; represents the photovoltaic failure probability; represents the time; represents the duration of typhoon and rainstorm.

[0292] For each moment with vulnerable load loss, all possible selection combinations are made, and the survival probability of all combinations of vulnerability components is calculated.

[0293] F4. In the order from the smallest quantity to the largest, for each combination of vulnerability components with each quantity at each moment with vulnerable load loss, they are put into operation in turn from the largest survival probability to the smallest, and the vulnerability components outside the combination of vulnerability components in operation are disconnected.

[0294] F5. When each combination of vulnerability components is put into operation, the system is rescheduled through an active scheduling and multi-objective optimization model. The active scheduling includes power generation rescheduling, pre-disaster network reconstruction, optimization of droop control parameters, and protective voltage regulation.

[0295] F6. Judging in turn according to the order of being put into operation whether there is vulnerable load loss in the system when each combination of vulnerability components is put into operation. If not, according to the combination of vulnerability components put into operation this time and the active scheduling, active defense at this moment is carried out.

[0296] In the preferred embodiment of the present invention, the multi-objective optimization model includes an objective function and constraint conditions.

[0297] The multi-objective optimization model aims to minimize the overall operation cost, including minimizing the penalty cost generated by load shedding and reducing the operation costs of dispatchable distributed power sources, photovoltaics, and wind power. The objective function of the multi-objective optimization model can be expressed as:

[0298] (22)

[0299] (23)

[0300] (24)

[0301] (25)

[0302] (26)

[0303] (27)

[0304] (28)

[0305] Among them, represents the start time of typhoon rainstorm; represents the end time of typhoon rainstorm; and respectively represent the load shedding penalty cost and the power generation operation cost, where the power generation operation cost includes the photovoltaic operation cost , the wind power operation cost , the dispatchable distributed power source operation cost and the mobile energy storage operation cost ; and respectively represent the unit load shedding penalty cost, the unit photovoltaic power generation operation cost, the unit wind power generation operation cost, the unit mobile energy storage operation cost and the unit dispatchable distributed power source operation cost; represents the active power of the load at time is a binary decision variable, where "1" represents load restoration, otherwise it is load shedding; and respectively represent the power generation of photovoltaic, wind power, distributed power source and mobile energy storage at time , all of which are continuous decision variables; and respectively represent the node set, the photovoltaic node set, the wind turbine node set, the distributed power source node set and the mobile energy storage node set; and respectively represent the indices of the nodes where the load, photovoltaic, wind turbine, distributed power source and mobile energy storage are located; and are weight coefficients, and .

[0306] The weight coefficients can be determined by the Pareto analysis method. In the preferred embodiment of the present invention, and are selected to preferentially reduce the load shedding rather than the power generation operation cost.

[0307] The constraint conditions of the multi-objective optimization model include power flow and safe operation constraints, photovoltaic and wind turbine operation constraints, load demand model based on static voltage, dispatchable distributed power source and mobile energy storage operation constraints based on droop control characteristics, and radial topology constraints based on improved spanning tree.

[0308] The power flow and safe operation constraints include:

[0309] Modeling is carried out using the optimal power flow equation based on Disflow. The AC power flow equation is given by formulas (29), (30), and (31), the line capacity constraint is represented by formula (32), and the operation safety constraints are given by formulas (33) and (34):

[0310] (29)

[0311] (30)

[0312] (31)

[0313] (32)

[0314] (33)

[0315] (34)

[0316] Among them, represents the active power injected at node at time represents the reactive power injected at node at time and respectively represent the active power on lines and ; and respectively represent the reactive power on lines and ; and respectively represent the currents on lines and ; and respectively represent the reactances on lines and ; and respectively represent the resistances on lines and ; represents the reactive power of the load at time and respectively represent the active output powers of distributed power sources, photovoltaics, wind turbines, and mobile energy storage at node at time represents the reactive output power of the distributed power source at node at time is a binary variable, which takes the value of 1 when the line is energized and 0 otherwise; and represent the voltage amplitude and the system frequency at node respectively; represents the voltage amplitude at node ; is a relatively large positive number; and are the maximum values of the node voltage, system frequency, branch current, branch active power, and reactive power respectively; and are the minimum values of the node voltage, system frequency, branch current, branch active power, and reactive power respectively; is the line set.

[0317] The AC power flow constraints ensure the feasibility of the distribution system operation through formulas (29)–(31), avoiding widespread load shedding or even a complete power outage due to the mismatch of active power, reactive power, and voltage constraints. Generally speaking, considerations in terms of heat and stability may limit the transmission capacity of distribution lines. The load on long transmission lines is usually limited by stability. However, distribution lines may be affected by their thermal limits, which are reflected by formulas (32) and (33), because the line power flow limit is determined by the heat capacity of the line. Formula (34) ensures that the global frequency and local voltage are maintained within a reasonable range.

[0318] The operation constraints of PV and wind turbines include:

[0319] PV and wind turbines are modeled as PQ control resources with a unity power factor; it is assumed that PV units and wind turbines have strong anti-typhoon and heavy rain characteristics, that is, the design wind speed is higher than the maximum wind speed of typhoons and heavy rains, and according to the vulnerability theory, PV units and wind turbines will not fail during typhoons and heavy rains; at the same time, it is assumed that PV and wind turbines are equipped with energy storage and converters, and each PV unit and wind turbine is regarded as an independent whole, ignoring the detailed models of energy storage and converters; considering the actual available output of PV and wind turbines during typhoons and heavy rains; therefore, the operation constraints of PV and wind turbines can be expressed as:

[0320] (35)

[0321] (36)

[0322] where and represent the expected output of PV and wind power respectively; and represent Power generation of photovoltaic and wind power at a certain moment.

[0323] The load demand model based on static voltage includes:

[0324] Protective voltage regulation controls the system voltage at the permitted lower limit to reduce peak demand, especially during extreme natural disaster events. Therefore, protective voltage regulation helps improve the resilience of the system during island operation. To describe the relationship between the load and the voltage amplitude, the polynomial ZIP load model is adopted:

[0325] (37)

[0326] (38)

[0327] Among them, and are the participation factors of the constant impedance, current, and power terms in the active power load respectively; and are the participation factors of the constant impedance, current, and power terms in the reactive power load respectively; and represent the rated active power and reactive power of the load respectively; The polynomial ZIP load model considers a relatively accurate system load prediction. If the prediction error is considered, the load demand can be expressed as the predicted value plus the error obtained according to the probability distribution of historical data.

[0328] The operation constraints of dispatchable distributed power sources and mobile energy storage based on droop control characteristics include:

[0329] Dispatchable distributed power sources such as diesel generators are usually equipped with droop controllers to provide voltage and frequency support during island operation to ensure the safe and reliable operation of the system. The dispatchable distributed power sources adopting droop control need to meet the capacity and power constraints, and the droop control characteristics of active power-frequency and reactive power-voltage, including:

[0330] (39)

[0331] (40)

[0332] (41)

[0333] (42)

[0334] (43)

[0335] (44)

[0336] (45)

[0337] (46)

[0338] (47)

[0339] Among them, and respectively represent the maximum active power output, the maximum reactive power output, and the total output capacity of the dispatchable distributed power source; and respectively represent at time node the droop control decision variables of the distributed power source; and respectively represent the reference voltage amplitude and the reference frequency decision variables; and respectively represent the maximum values of the reference frequency, the reference voltage, the active droop coefficient, and the reactive droop coefficient; and respectively represent the minimum values of the reference frequency, the reference voltage, the active droop coefficient, and the reactive droop coefficient; represents the angular frequency of the system.

[0340] It should be noted that, compared with the transmission system, the droop control dispatchable distributed power sources in the distribution system are usually smaller in volume, higher in ramp rate, and shorter in response time. Therefore, the start-up time of this type of power source is not considered in the formula, which is also a common practice in the research of the distribution system.

[0341] In addition to the controllable distributed power sources with droop control functions, mobile energy storage also is equipped with a droop controller; assuming that the mobile energy storage provides as much active power as possible, therefore, the mobile energy storage charges or discharges according to the droop characteristics, which can be expressed as:

[0342] (48)

[0343] The charging or discharging action of the mobile energy storage is restricted by the rated capacity , which can be expressed as:

[0344] (49)

[0345] Among them, is the rated capacity of the mobile energy storage; represents the power generation of the mobile energy storage at time

[0346] The radial topology constraints based on the improved spanning tree include:

[0347] During typhoon and rainstorm disasters, multiple islands may form in the power distribution system. An improved spanning tree constraint is adopted to meet the radial topology requirements of each island, including:

[0348] (50)

[0349] (51)

[0350] Among them, and are binary variables; if node is a substation node, then takes the value of 1, otherwise 0; if node is a distributed power source node, then takes the value of 1, otherwise 0; is a binary variable indicating whether node is the root bus in the island. If so, it takes the value of 1, otherwise 0; and are the sets of the parent node and child nodes of bus respectively; is a binary variable. A value of 1 indicates that branch is energized, and a value of 0 indicates that branch is not energized; and are binary variables. If node is the parent node of node , that is , or bus is the parent node of bus , that is , then branch is selected to be energized, that is, take . and cannot be 1 at the same time; represents a binary variable indicating whether node is the parent node of node . When is the parent node of node .

[0351] In the preferred embodiment of the present invention, the multi-objective optimization model further includes:

[0352] Convex optimization solution of the multi-objective optimization model:

[0353] There are non-linear equations in the multi-objective optimization model, including those in equations (29), (31), (37) and (38) , , and those in equation (42) in formula (43), in formula (48), , and formulas (32) and (41); the non - linear equations in the multi - objective optimization model can be divided into three types: the first type is the square of continuous variables, the second type is quadratic equations or quadratic inequalities, and the third type is the product of continuous variables; convex optimization is performed on these three types of non - linear equations:

[0354] Convex optimization of the first - type non - linear equations:

[0355] Since the formulas (29), (31), (37) and (38) contain square terms of branch currents and node voltages and , variables and are used to replace these two quadratic terms respectively, and we get:

[0356] (52)

[0357] (53)

[0358] (54)

[0359] (55)

[0360] where and are the square terms of the voltage of node and the current of branch respectively; and are the square terms of the voltage of node and the current of branch respectively.

[0361] Since formulas (54) and (55) contain square terms , they are still non - linear expressions. Therefore, Taylor series expansion is used to linearize these formulas (54) and (55), including:

[0362] Assume the initial value is , and through Taylor series expansion, we can get:

[0363] (56)

[0364] where represents an infinitesimal; is the initial value of the magnitude of the voltage of node .

[0365] By neglecting second - order and higher - order terms, Equation (56) can be approximately expressed as:

[0366] (57)

[0367] The magnitude of the nodal voltage is usually about 1.0 p.u., so assuming , Equation (57) can be further approximated as:

[0368] (58)

[0369] The definition of the distance function for Taylor - series approximation includes:

[0370] (59)

[0371] where is the distance function for Taylor - series approximation;

[0372] See Figure 5 , from the error curve of Taylor - series approximation, it can be seen that the calculation error is at the order of magnitude, meeting the approximation accuracy requirements within the given voltage magnitude range [0.95, 1.05]; by substituting Equation (58) into Equations (54) and (55), an approximate load - demand model can be obtained, which can be expressed as:

[0373] (60)

[0374] (61)

[0375] Convex optimization of the second - type non - linear equations:

[0376] Equations (32) and (41) can be converted into second - order cone constraints and rotated second - order cone constraints respectively, which can be expressed as:

[0377] (62)

[0378] (63)

[0379] Convex optimization of the third - type non - linear equations:

[0380] Since Equations (42), (43) and (48) all contain the product of two bounded continuous variables; using McCormick envelope approximation for processing, introducing auxiliary variables and to represent and respectively, and using twelve linear constraints for approximation, which can be expressed as:

[0381] (64)

[0382] (65)

[0383] (66)

[0384] (67)

[0385] (68)

[0386] (69)

[0387] (70)

[0388] (71)

[0389] (72)

[0390] (73)

[0391] (74)

[0392] (75)

[0393] The active defense method for a new energy distribution system considering the uncertainty of typhoon and rainstorm evolution paths in the present invention forms a robust scheduling strategy by integrating dual uncertainty scenarios in advance, aiming to maximize the number of out-of-service vulnerable components under the condition of full-load operation of the system, thereby effectively reducing the overall vulnerability of the system during typhoon and rainstorm disasters. Switching the system to island operation not only helps reduce the load loss of the distribution system during disasters, improve the safety of the system, but also reduces the protection coordination requirements with the main grid and alleviates other technical problems that may occur during disasters. The vulnerability load loss caused by disconnecting vulnerable components at each moment is calculated through a multi-objective model. All possible combinations of vulnerable components at each moment with vulnerability load loss are considered for combination, the survival probability of each combination is calculated, and the trial operation starts from the combination with the smallest number, and then the combination of vulnerable components with the smallest number of vulnerable components at each moment with vulnerability load loss is found for operation, improving the safety and robustness of the system. The method of the present invention combines mechanism analysis and data-driven models and adopts a robust strategy for optimization. Compared with the active management method that does not consider the typhoon and rainstorm evolution paths, the method of the present invention shows the best comprehensive benefits. In the worst case, compared with the post-disaster scheduling results of the active management method and the passive strategy method that do not consider the typhoon and rainstorm evolution paths, the method of the present invention effectively reduces the load loss and improves the satisfaction rate of load demand. The effect, flexibility, and economy of the active defense plan obtained according to the method of the present invention all have good performances.

[0394] In a preferred embodiment of the present invention, an active defense device for a new energy distribution system considering the uncertainty of typhoon and rainstorm evolution paths is further provided for the method of the present invention. The device includes a first module, a second module, a third module, a fourth module, a fifth module, and a sixth module.

[0395] The first module is used to obtain the dual uncertainty time series scenarios of the coupling of typhoon and rainstorm evolution paths and the states of system components; and obtain the vulnerable components at each moment during the typhoon and rainstorm duration according to the time series scenarios.

[0396] The second module is used to switch the system to island operation and keep the dispatchable distributed power sources in the system under droop control; calculate the vulnerable load loss caused by disconnecting all vulnerable components of the system at each moment through a multi-objective optimization model.

[0397] The third module is used to consider each quantity from 0 to the total number of vulnerable components at this moment, and combine all the selection situations when selecting any number of vulnerable components at each moment with vulnerable load loss; calculate the survival probability of the combination of vulnerable components.

[0398] The fourth module is used to put into operation the vulnerability component combinations of each quantity at each moment of vulnerability load loss in the order from the smallest to the largest quantity, in the order from the largest to the smallest survival probability, and disconnect the vulnerability components outside the operating vulnerability component combinations.

[0399] The fifth module is used to reschedule the system when each vulnerability component combination is put into operation through active scheduling and a multi-objective optimization model. The active scheduling includes power generation rescheduling, pre-disaster network reconstruction, optimized setting of droop control parameters, and protective voltage regulation.

[0400] The sixth module is used to sequentially determine whether there is vulnerability load loss in the system when each vulnerability component combination is put into operation in the order of being put into operation. If not, active defense at this moment is carried out according to the vulnerability component combination put into operation this time and active scheduling.

[0401] The active defense device for a new energy distribution system considering the uncertainty of the typhoon and rainstorm evolution path of the present invention, which is used for the method of the present invention, has the same beneficial effects as the method of the present invention.

[0402] Verification part:

[0403] To evaluate the effectiveness of the method of the present invention, verification is carried out from two aspects: 1) verification of the active defense plan and effectiveness; 2) verification of the superiority of the active defense method.

[0404] See Figure 6 , the modified IEEE 33-node test system is located at (110.19°E, 21.22°N) and is represented by a hexagon. The typhoon and rainstorm move northwest from the initial position (117.9°E, 17.1°N). The typhoon and rainstorm path and the component operating state introduce two layers of uncertainty, long-term and short-term. For each typhoon and rainstorm path, first, the wind field in the area where the system is located is determined, and the vulnerability state of the components and the uncertain output of photovoltaic and wind power are evaluated. In the preferred embodiment of the present invention, 50 typhoon and rainstorm path scenario samples are generated, and 5 typical path scenarios (i.e., Figure 6 the typical paths 1 to 5 in Figure 6 ) are obtained through scenario simplification, and their corresponding occurrence probabilities are 0.12, 0.37, 0.11, 0.12, and 0.28 respectively. In t 1 to t 12 ) for analysis. Table 1 shows the chronological expected distance and its corresponding chronological expected predicted wind speed under all typical scenarios.

[0405] Table 1 Time - series expected distance and its corresponding time - series expected predicted wind speed under all typical scenarios

[0406] 。

[0407] The tested IEEE 33 - node system includes 1 substation, 5 dispatchable distributed energy resources (DERs), 5 identical photovoltaic (PV) units, 2 identical wind turbines (WTs), 5 groups of mobile energy storage systems (ESSs), 5 tie - switch branches and 32 sectional - switch branches (a total of 37 switch branches, denoted as S1 - S37). Its main parameters and topological results are shown in Tables 2 to 4 and Figure 7 as shown. The voltage level of the system is 12.66 kV, the base power is 100 MVA, and the rated active load and reactive load are 3.715 MW and 2.300 Mvar respectively. In addition, it is assumed that before the typhoon and heavy rain come, all tie - switch branches are in the open state, and all mobile energy storage systems are fully charged. The test system includes commercial users, residential users and industrial users. The unit penalty cost for load shedding, the unit operation cost of PV, the unit operation cost of WT, the unit operation cost of ESS and the unit operation cost of DER are 20 $ / kW, 10 $ / kW, 20 $ / kW, 20 $ / kW and 100 $ / kW respectively. It is assumed that the power supply load demand is constant , and the maximum load in a day is regarded as the constant load of the distribution system during the typhoon and heavy rain path. This is equivalent to considering the most adverse situation under load uncertainty. Figure 8 Shows the wind - induced vulnerability curves of the distribution system components. It is assumed that the vulnerability threshold of the distribution system components is YUZ = 5%.

[0408] Table 2 Main parameters of dispatchable distributed energy resources

[0409] 。

[0410] Table 3 Droop control parameters of dispatchable distributed energy resources and mobile energy storage systems

[0411] 。

[0412] Table 4 Main parameters of wind turbines, photovoltaics and mobile energy storage systems

[0413] 。

[0414] 1) Active defense plan and effectiveness verification

[0415] According to the method of the present invention, the time - series vulnerability components under typhoon and heavy rain disasters are identified, and the specific information is shown in Table 5. It can be seen from the table that in t 1 ~ t4 and t 8 ~ t 12 During this period, since the typhoon and rainstorm center was far from the power distribution system and the wind speed at the location of the system was low, the failure probability of components was lower than the set vulnerability threshold, and no vulnerable components appeared at this time. And during t 5 ~ t 7 As the typhoon and rainstorm center gradually approached the power distribution system, the wind speed at the location of the system gradually increased, resulting in a rapid increase in the number of components whose failure probability exceeded the vulnerability threshold. For example, at t 5 the moment, the number of vulnerable branches increased to 5, namely S2, S10, S14, S24 and S30; with the further increase of the wind speed, the number of vulnerable components increased to 7 and 9 at t 6 and t 7 moments respectively. It can be seen that the impact of typhoon and rainstorm disasters on the components of the power distribution system shows obvious temporal and sequential changes. Especially within a specific time period, the impact rapidly intensifies and then rapidly weakens.

[0416] Table 5 Temporal and sequential vulnerable components under typhoon and rainstorm disasters

[0417] .

[0418] Table 6 details the active defense plan of the test case during typhoon and rainstorm t 1 ~ t 12 During this period, since no vulnerable components appeared during this period, the power distribution system was divided into 2, 2, 3 and 2 independent islands through the optimal scheduling and reconstruction of in-network resources, and each island could meet its corresponding load demand. At t 1 to t 4 time period, since no vulnerable components appeared during this period, the power distribution system was divided into 2, 2, 3 and 2 independent islands through the optimal scheduling and reconstruction of in-network resources, and each island could meet its corresponding load demand. At t 5 the moment, the vulnerable branches S2, S10, S14, S24 and S30 appeared, as Figure 9 shown. By actively shutting down all vulnerable components, the system can meet the power supply demand of all loads, and the average calculation time is 7.01 seconds. The load demand at this time is 0.9962 p.u., mainly because the protective voltage regulation reduces the node voltage, thereby reducing the voltage-related load demand.

[0419] At t 6At a certain moment, the system situation changes, such as Figure 10 shown. By adopting the proposed active defense strategy and deactivating other vulnerable components except components S5 and S31, the total load demand of the system is met, and the configuration corresponding to this combination C(7, 2) has the maximum survival probability. At this time, all tie switches are closed, and the distribution system is reconstructed into three independent islands, each of which can meet its load demand. The average time taken for the whole process is 55.51 seconds.

[0420] At t 7 a certain moment, keep the minimum vulnerability branches S5, S20, and S31 in operation, as Figure 11 shown, to make this configuration the combination with the maximum survival probability in combination C(9, 3). At the same time, deactivate S2, S7, S10, S14, S24, and S30 to meet the total power supply load demand of the system. At this time, the total load demand drops to the minimum value of 0.9884 p.u., and the system is reconstructed into three independent islands. The average time taken for the calculation process is only 330.66 seconds, meeting the requirement of the pre-disaster defense strategy for the response time.

[0421] Table 6 t 1 ~ t 12 Active defense scheduling of the distribution system during

[0422] .

[0423] Subsequently, during t 8 ~ t 12 the time period, since there are no vulnerable components, the system is divided into 3, 4, 2, 4, and 3 independent islands in sequence through optimal scheduling and reconstruction, and each island can meet the load demand inside it. At the same time, Figures 12 to 15 shows the active scheduling of distributed power sources, wind turbines, photovoltaic power, and mobile energy storage during t 1 ~ t 12 the typhoon rainstorm. The variation of the total output power of distributed power sources, wind turbines, photovoltaic power, and mobile energy storage over time is shown in Table 7.

[0424] Table 7 Total output power of distributed power sources, wind turbines, photovoltaic power, and mobile energy storage over time

[0425] .

[0426] The active defense plan obtained by this method can be applied to the distribution system before a disaster. As the typhoon rainstorm disaster develops, vulnerable components (such as those in Table 5 t6 Components such as S2 or S14 at a certain moment may fail and stop operating. As can be seen from Table 6, during the active defense scheduling t 5 ~ t 7 period, 0, 2, and 3 vulnerable components remained in operation respectively. However, in the worst case, the vulnerable components may all stop operating due to faults at t 5 , t 6 and t 7 moments. Therefore, the distribution system needs to switch to the disaster-time scheduling strategy. This strategy has been completed in the pre-disaster active defense framework and provides a scheduling plan through a pre-stored look-up table, thus achieving a rapid switch to disaster-time scheduling, as shown in Tables 8 and 9.

[0427] As can be seen from Table 8, at t 5 moment, since all 5 vulnerable components are allowed to stop operating, the system can still meet the power supply requirements of all loads. While at t 6 and t 7 moments, the total power supply loads of the system are 0.7461 p.u. and 0.6703 p.u. respectively, and the corresponding power supply rates are 74.43% and 67.1% respectively. As can be seen from Table 9, t 5 at the worst case of t 6 and t 7 moments, the outputs of distributed power sources, wind turbines, photovoltaic power, and mobile energy storage are consistent with the set values of the pre-disaster active defense scheduling strategy; while at t 6 and t 7 moments, due to the emergence of more power outage islands, some loads are forced to be cut, resulting in the outputs of distributed power sources, wind turbines, photovoltaic power, and mobile energy storage all being lower than the set values of the pre-disaster active defense scheduling strategy. At the same time, the optimal droop parameters of the distributed power sources at

[0428] Table 8 Post-disaster scheduling of the method of the present invention under the worst case

[0429] .

[0430] Table 9 Total output powers of distributed power sources, wind turbines, photovoltaic power, and mobile energy storage under the worst case

[0431] 。

[0432] Table 10 Optimal Droop Parameters of Distributed Generation

[0433] 。

[0434] 2) Verification of the Superiority of the Active Defense Method

[0435] The superiority of the method of the present invention will be verified from two aspects: ① is t 5 ~ t 7 Comparison of the overall operating costs during the period; ② is the comparison of the disaster-time scheduling in the worst case.

[0436] ① t 5 ~ t 7 Comparison of the overall operating costs during the period

[0437] The method of the present invention is compared with the active management method that does not consider the evolution path of typhoon and rainstorm. The comparison results are shown in Table 11.

[0438] Table 11 t 5 ~ t 7 Comparison of the overall operating costs during the period

[0439] 。

[0440] As can be seen from Table 11, by adopting the active management method that does not consider the evolution path of typhoon and rainstorm, the overall operating cost during t 5 ~ t 7 the period is 109.4915×10 4 $, while the overall operating cost of adopting the method of the present invention is only 108.6554×10 4 $. In contrast, the operating cost of the method of the present invention is significantly reduced, by 0.8361×10 4 $. With the increase of the unit operating cost, the advantage of the method of the present invention in the overall operating cost will become more obvious.

[0441] ② Comparison of the post-disaster scheduling in the worst case

[0442] In the worst case, a comparative analysis is carried out on the method of the present invention, the active management method that does not consider the evolution path of typhoon and rainstorm, and the passive strategy. The passive strategy refers to the operation of only relying on the protection system to shut down the components. Representative t 7Compare at the moment (i.e., the moment when the number of vulnerability components is the largest).

[0443] Table 12 Comparison of post-disaster scheduling of different methods under the worst conditions

[0444] 。

[0445] Table 12 shows the post-disaster scheduling of the distribution system using the passive strategy under the worst conditions. The results show that the distribution system is divided into 10 islands, and the number of power-off nodes is 16, mainly because some islands have no power source, or some distributed power sources with larger capacities are restricted in islands with lower loads. Under the passive strategy, only 0.4569 p.u. of the load demand is met, and the corresponding power supply rate is 45.71%. In contrast, the method of the present invention reduces the number of islands to 9 under the worst conditions, significantly reduces the number of power-off nodes, and effectively improves the load demand satisfaction rate. The power supply rate is increased by 21.39%. It should be noted that although the active management method that does not consider the evolution path of typhoon and rainstorm is the same as the method of the present invention in post-disaster scheduling, its overall operation cost is relatively high. Therefore, through the above analysis, the superiority of the method of the present invention in post-disaster scheduling and cost-benefit is verified.

[0446] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention can have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. An active defense method for a new energy distribution system considering the uncertainty of the evolution path of typhoons and rainstorms, characterized in that: include: Obtain dual uncertainty time series scenarios of typhoon and rainstorm evolution path and system component status coupling; According to the time series scenario, the vulnerable components at each moment during the typhoon and rainstorm duration are obtained; Switch the system to island operation and keep the dispatchable distributed power sources in the system in droop control; calculate the vulnerable load loss caused by disconnecting all vulnerable components of the system at each moment through a multi-objective optimization model; Considering each number of the total number of vulnerable elements from 0 to the current moment, each vulnerable element with a vulnerable load loss moment is combined by considering all selected situations when any number is selected; calculating the survival probability of the combination of vulnerable elements; In order from the smallest to the largest number, the fragile element combinations under each number at each moment of fragile load loss are put into operation in order from the largest to the smallest survival probability, and the fragile elements outside the operating fragile element combinations are disconnected; rescheduling the system for each combination of vulnerable components to be put into operation through active scheduling and multi-objective optimization models, wherein the active scheduling includes generation rescheduling, pre-disaster network reconstruction, droop control parameter optimization setting and protective voltage regulation; It is determined in turn in the order of being put into operation whether there is a vulnerable load loss in the system when each vulnerable component combination is put into operation. If not, active defense is performed at this moment based on the vulnerable component combination put into operation this time and the active scheduling.

2. The active defense method for a new energy power distribution system considering the uncertainty of the typhoon and rainstorm evolution path according to claim 1 is characterized in that: The method of calculating the vulnerable load loss caused by disconnecting all vulnerable components of the system at each moment by using the multi-objective optimization model includes: Case 1: If the distribution system does not have any load loss, the vulnerable load loss caused by disconnecting all vulnerable components of the system at each moment is directly calculated by the multi-objective optimization model. Scenario 2: The power distribution system has its own load loss. When calculating the vulnerable load loss caused by disconnecting all vulnerable components of the system at each moment through the multi-objective optimization model, it is necessary to first calculate the total load loss of the power distribution system, and then subtract the inherent load loss from the total loss of the power distribution system to obtain the vulnerable load loss; The inherent load loss includes the load loss of the power distribution system excluding the fragile load loss caused by disconnecting the fragile element; the total load loss of the power distribution system includes the inherent load loss and the fragile load loss.

3. The active defense method for a new energy power distribution system considering the uncertainty of the evolution path of typhoons and rainstorms according to claim 2 is characterized in that: The dual uncertainty time series scenario of obtaining the typhoon rainstorm evolution path and the coupling of system component states includes: Construct uncertainty model of typhoon and rainstorm disaster and system disaster dynamics model under typhoon and rainstorm; A dual uncertainty time series scenario of the coupling of the typhoon and rainstorm evolution path and the system component state is generated according to the typhoon and rainstorm disaster uncertainty model and the system disaster dynamics model under the typhoon and rainstorm.

4. The active defense method for a new energy power distribution system considering the uncertainty of the evolution path of typhoons and rainstorms according to claim 3 is characterized in that: The typhoon and rainstorm disaster uncertainty model includes two parts: the typhoon and rainstorm evolution uncertainty path model and the typhoon and rainstorm wind field model: The uncertainty path model of typhoon rainstorm evolution includes: In order to simulate the possible movement path of typhoon and rainstorm after landing, the typhoon and rainstorm empirical path mechanism model is used to calculate the moving wind speed and direction angle : (1) (2) (3) (4) in, It represents the logarithmic deviation of the moving wind speed of the typhoon rainstorm center; Indicates the moving direction angle of the typhoon rainstorm center; It represents the logarithm of the moving wind speed of the typhoon rainstorm center; and Respectively Moving wind speed and moving direction angle of the typhoon and rainstorm center at the moment; and Respectively The latitude and longitude of the typhoon and rainstorm center at the moment; and are the prediction errors that follow Gaussian distribution; and is the fitting coefficient obtained based on the historical data of different regions; is the time interval; Based on the typhoon rainstorm empirical path mechanism model combined with the data-driven method, the prediction error is processed through historical data, including: It is assumed that the typhoon rainstorm path is developed around the predicted path, and the prediction error follows a Gaussian distribution with a mean of zero; the uncertainty of the typhoon rainstorm path is quantified by the deviation between the empirical path and the historical true path, including the first calculation; The first calculation includes: selecting two adjacent typhoon rainstorm locations A1 and A2 from the historical real path; calculating the empirical path from A1 to A3 by equations (1) to (4); there is a prediction error of the moving direction angle between the historical real path A1 to A2 and the empirical path A1 to A3 ; The predicted moving speed and moving direction angle of the empirical path A1 to A3 can be calculated according to the coordinates of A1 and A3; the historical real moving speed and moving direction angle of the historical real path A1 to A2 are obtained; the error between the empirical path A1 to A3 and the historical real path A1 to A2, that is, the central moving wind speed prediction error, can be obtained according to the predicted moving speed and moving direction angle and the historical real moving speed and moving direction angle and the prediction error of the moving direction angle ; The first calculation is performed on each historical real path to obtain the corresponding and , and then get the central moving wind speed prediction error set and the moving direction angle prediction error set ; Will and The prediction error in is the same as the prediction value calculated by equations (1) and (2) and Add them together to get various typhoon and rainstorm path scenarios; The typhoon rainstorm wind field model includes: For the typhoon rainstorm path scenario moving over the ocean surface, the first model can be used to estimate the time series wind speed at different locations , the first model includes: (5) (6) (7) (8) (9) (10) (11) (12) in, Indicates the circulation wind speed; and Respectively time Typhoon rainstorm wind speed and rainfall rate; , and is the correction factor; and are the distance between the power distribution system and the typhoon and rainstorm center and the maximum wind speed radius respectively; and is the central pressure difference and its rate of change; and They are the coordinates of the power distribution system and the typhoon rainstorm center respectively; The duration of the typhoon and rainstorm disaster; for time Typhoon wind speed based on the Batts model at ; express Maximum circulation wind speed at the moment; Indicates the moving wind speed of the typhoon rainstorm center; For the typhoon rainstorm path scenario moving on land, the wind speed decreases due to friction after the typhoon rainstorm lands. Therefore, the attenuation coefficient is introduced on the basis of the first model. It is assumed that the time series wind speed of the typhoon rainstorm moving on the ocean surface calculated by the first model is , and assume that the time series wind speed of the typhoon rainstorm moving over land is , then the attenuation coefficient can be expressed as: (13) in, and represent the roughness length and retardation factor respectively; According to the geographical location of the typhoon and rainstorm, select or Calculate the time series wind speed.

5. The active defense method for a new energy power distribution system considering the uncertainty of the evolution path of typhoons and rainstorms according to claim 4 is characterized in that: The system disaster dynamics model under typhoon and rainstorm includes two parts: system component failure probability model and photovoltaic and wind turbine output model: The system component failure probability model includes: Assuming that the system component failure is mainly caused by wind speed, the system component failure includes damage to distribution towers, failure of overhead distribution wires, and failure of photovoltaic components; assuming that different components of the distribution line are independent of each other, the failure probability of the distribution line can be calculated through the series model, which can be expressed as: (14) (15) (16) in, and represent the failure probabilities of distribution lines, towers and conductors respectively; and Respectively represent the number of towers and conductor segments of the line; and Respectively represent the towers of the line and wire segments The probability of failure; The standard log-normal distribution function is used to model the failure probability of photovoltaic components, which can be expressed as: (17) in, represents the standard lognormal distribution function; represents the failure probability of the PV element; and denote the standard deviation and the design wind speed of PV, respectively; The photovoltaic and wind turbine output models include: In typhoons and rainstorms, the photovoltaic output is intact and available It can be expressed as: (18) (19) in, It represents the impact of typhoon and rainstorm disasters on light irradiance in logarithmic space; is the light irradiance under normal conditions; and They are typhoon intensity, normalized distance, photovoltaic power generation efficiency and area; represents a set of photovoltaic nodes; Since wind turbines are usually resistant to typhoons and rainstorms, it is assumed that the wind turbines are not damaged or malfunction during the typhoon and rainstorm disaster. The wind field changes caused by typhoons and rainstorms directly affect the wind power output. Specifically, the available wind power output can be calculated through the wind speed-power curve of the wind turbine. , which can be expressed as: (20) in, and They represent the cut-in wind speed, rated wind speed and cut-out wind speed of the fan, and the values ​​are 3m / s, 12m / s and 25m / s respectively; and They represent air density, power coefficient, fan blade swept area and rated power respectively; Represents a collection of wind turbine nodes.

6. The active defense method for a new energy power distribution system considering the uncertainty of the typhoon and rainstorm evolution path according to claim 5 is characterized in that: The dual uncertainty time series scenario of coupling the typhoon and rainstorm evolution path and the system component state is generated according to the typhoon and rainstorm disaster uncertainty model and the system disaster dynamics model under the typhoon and rainstorm, including: Get the coordinates of the power distribution system; Obtain the coordinates, moving wind speed and moving direction angle of the typhoon rainstorm center at this moment, and calculate the moving wind speed and moving direction angle of the typhoon rainstorm center at the next moment through the typhoon rainstorm empirical path mechanism model, i.e., equations (1) to (4); Take the prediction error of the moving wind speed and moving direction angle of the typhoon rainstorm center and , the moving wind speed and moving direction angle of the typhoon rainstorm center at the next moment are compared with the prediction error and The corresponding addition is performed to obtain the coordinates of the typhoon rainstorm center at the next moment; the moving wind speed, moving direction angle and coordinates of the typhoon rainstorm center at subsequent moments are continuously calculated in time sequence until the maximum number of scenes of the typhoon rainstorm path samples reaches S trackmax ; The scene number is reduced by using the scene backward reduction algorithm. S trackmax Reduced to S reduced , and record each typical scene s ∈ S reduced The corresponding probability of occurrence π s ; Performing a second calculation at each moment during the typhoon and rainstorm duration period, the second calculation comprising: According to formula (10) and the coordinates of the distribution system, the distance from the typhoon and rainstorm center to the distribution system in each typical scenario is calculated. ,in Take 1, 2… S reduced ; Then, considering all typical scenarios, the expected time series distance can be expressed as ; According to equations (7) to (9) and (11) to (13), the expected predicted wind speed at the distribution system is obtained: ; According to equations (18) and (20), the expected photovoltaic output is calculated as Expected output of wind power ; According to equations (14) and (17), the failure probability of photovoltaic is calculated respectively: and the failure probability of the distribution line ; Using vulnerability thresholds As a criterion for determining vulnerable elements; Distribution lines Vulnerability status , 0 represents a vulnerable line, and 1 represents a normal line; Photovoltaic components Vulnerability status , 0 represents a fragile photovoltaic element, and 1 represents a sound photovoltaic element; Obtain the uncertainty vulnerability status of all components of the distribution system at this moment, as well as the uncertainty output of photovoltaic and wind power; After completing the second calculation for each moment in the typhoon and rainstorm duration period, the time series scene is obtained.

7. The active defense method for a new energy power distribution system considering the uncertainty of the evolution path of typhoons and rainstorms according to claim 6 is characterized in that: The said considering each number of the total number of vulnerable elements from 0 to the current moment, combining each vulnerable element at the moment of vulnerable load loss when selecting any number of them, taking into account all selected situations; Calculating the survival probability of a combination of vulnerable components involves: Assume that there is a vulnerable load loss moment including A vulnerable component, is a positive integer; Choose from a number of vulnerable components components, is an integer, ;get Combinations of vulnerable components; Calculate the The survival probabilities of the combinations of vulnerable components are sorted from the highest to the lowest survival probabilities; The calculation of the survival probability includes: (21) in, represents the probability of survival; Indicates the combination of vulnerable components; Indicates the distribution line number; represents the probability of failure of the distribution line; Indicates the photovoltaic equipment number; represents the probability of PV failure; Indicates the moment; Indicates the duration of the typhoon and rainstorm; All possible selection combinations are performed at each moment when there is a fragile load loss, and the survival probabilities of all fragile element combinations are calculated.

8. The active defense method for a new energy power distribution system considering the uncertainty of the typhoon and rainstorm evolution path according to claim 7 is characterized in that: The multi-objective optimization model includes an objective function and constraints; The multi-objective optimization model aims to minimize the overall operating cost, including minimizing the penalty cost caused by load reduction and reducing the operating costs of dispatchable distributed power sources, photovoltaic and wind power. The objective function of the multi-objective optimization model can be expressed as: (22) (23) (24) (25) (26) (27) (28) in, Indicates the time when the typhoon and rainstorm begin; Indicates the time when the typhoon and rainstorm end; and They represent the load reduction penalty cost and the power generation operation cost, respectively, where the power generation operation cost includes the photovoltaic operation cost , Wind power operation costs , Dispatchy Distributed Generation Operating Cost and mobile energy storage operating costs ; and They represent the unit penalty cost of load reduction, the unit operating cost of photovoltaic power generation, the unit operating cost of wind power generation, the unit operating cost of mobile energy storage, and the unit operating cost of dispatchable distributed power generation respectively; express Active power of load at the moment; is a binary decision variable, where "1" represents load restoration, otherwise it represents load shedding; and Respectively The power generation of photovoltaic, wind power, distributed power and mobile energy storage at each moment are all continuous decision variables; and They represent the node set, photovoltaic node set, wind turbine node set, distributed power source node set and mobile energy storage node set respectively; and Respectively represent the indexes of the nodes where the load, photovoltaic, wind turbine, distributed power source and mobile energy storage are located; and is the weight coefficient, and ; The constraints of the multi-objective optimization model include power flow and safe operation constraints, photovoltaic and wind turbine operation constraints, a load demand model based on static voltage, dispatchable distributed power supply and mobile energy storage operation constraints based on droop control characteristics, and radial topology constraints based on improved spanning tree; The power flow and safe operation constraints include: The optimal power flow equation based on Disflow is used for modeling. The AC power flow equation is given by equations (29), (30) and (31), the line capacity constraint is expressed by equation (32), and the operation safety constraint is given by equations (33) and (34): (29) (30) (31) (32) (33) (34) in, express Time Node Injected active power; express Time Node Injected reactive power; and Respectively indicate lines and Active power on and Respectively indicate lines and Reactive power on and Respectively indicate lines and The current on and Respectively indicate lines and Reactance; and Respectively indicate lines and resistance; express The reactive power of the load at the moment; and Respectively Time Node Active output power of distributed power sources, photovoltaics, wind turbines and mobile energy storage; express Time Node Reactive output power of distributed power sources; is a binary variable, when the line The value is 1 when the power is on, otherwise it is 0; and Respectively represent nodes The voltage amplitude and system frequency at Representation Node The voltage amplitude at ; is a large positive number; and They are the maximum values ​​of node voltage, system frequency, branch current, branch active power and reactive power respectively; and are the minimum values ​​of node voltage, system frequency, branch current, branch active power and reactive power respectively; is a collection of lines; The photovoltaic and wind turbine operation constraints include: Photovoltaic and wind turbines are modeled as PQ control resources with unity power factor. It is assumed that the photovoltaic units and wind turbines have strong anti-typhoon and rainstorm characteristics, that is, the design wind speed is higher than the maximum wind speed of the typhoon and rainstorm, and according to the fragility theory, the photovoltaic units and wind turbines will not fail during the typhoon and rainstorm. At the same time, it is assumed that both photovoltaic and wind turbines are equipped with energy storage and converters, and each photovoltaic unit and wind turbine is regarded as an independent entity, ignoring the detailed model of energy storage and converter. The actual available output of photovoltaic and wind turbines during typhoon and rainstorm is considered. Therefore, the operation constraints of photovoltaic and wind turbines can be expressed as follows: (35) (36) in, and Represent the expected output of photovoltaic and wind power respectively; and Respectively The power generation of photovoltaic and wind power at each moment; The load demand model based on static voltage includes: To describe the relationship between load and voltage amplitude, a polynomial is used: Load Model: (37) (38) in, and are the participation factors of the constant impedance, current and power terms in the active power load, respectively; , and are the participation factors of the constant impedance, current and power terms in the reactive power load, respectively; and Respectively represent the rated active power and reactive power of the load; The dispatchable distributed power supply and mobile energy storage operation constraints based on the droop control characteristics include: The dispatchable distributed generation using droop control needs to meet the capacity and power constraints, active-frequency and reactive-voltage droop control characteristics, including: (39) (40) (41) (42) (43) (44) (45) (46) (47) in, and They represent the maximum active output power, maximum reactive output power and total output capacity of the dispatchable distributed generation respectively; and Respectively Time Node Droop control decision variables of distributed generation; and denote the reference voltage amplitude and reference frequency decision variables respectively; and Respectively represent the maximum values ​​of reference frequency, reference voltage, active power droop coefficient and reactive power droop coefficient; and Respectively represent the minimum values ​​of reference frequency, reference voltage, active power droop coefficient and reactive power droop coefficient; represents the angular frequency of the system; In addition to the controllable distributed power supply with droop control function, the mobile energy storage is also equipped with a droop controller; assuming that the mobile energy storage provides as much active output as possible, the mobile energy storage is charged or discharged according to the droop characteristics, which can be expressed as: (48) The charging or discharging action of mobile energy storage is subject to the rated capacity The restriction can be expressed as: (49) in, is the rated capacity of mobile energy storage; express The amount of electricity generated by mobile energy storage at all times; The radial topology constraint based on the improved spanning tree includes: During typhoon and rainstorm disasters, the distribution system may form multiple islands. Improved spanning tree constraints are used to meet the radial topology requirements of each island, including: (50) (51) in, and is a binary variable; if node is a substation node, then The value is 1, otherwise it is 0; if the node is a distributed power generation node, then The value is 1, otherwise it is 0; is a binary variable, indicating the node Whether it is the root bus in the island, if so, the value is 1, otherwise it is 0; and Busbar The parent node and child node set of ; It is a binary variable, and its value is 1, indicating that the branch Power on, a value of 0 indicates a branch No electricity; and is a binary variable. If the node Is a node The parent node of , or busbar It is a busbar The parent node of , then select the branch Power on, take , and Cannot be 1 at the same time; represents a binary variable, represents a node Is it a node? The parent node of Time Node For Node The parent node of .

9. The active defense method for a new energy power distribution system considering the uncertainty of the typhoon and rainstorm evolution path according to claim 8 is characterized in that: The multi-objective optimization model also includes: Convex optimization solution of the multi-objective optimization model: There are nonlinear equations in the multi-objective optimization model, including equations (29), (31), (37) and (38). , , in formula (42) , in formula (43) , in formula (48) , and equations (32) and (41); the nonlinear equations in the multi-objective optimization model can be divided into three types: the first type is the square of continuous variables, the second type is quadratic equations or quadratic inequalities, and the third type is the product of continuous variables; these three types of nonlinear equations are subjected to convex optimization processing: Convex optimization of nonlinear equations of the first type: Since equations (29), (31), (37) and (38) contain square terms of branch current and node voltage and , using the variable and To replace the two quadratic terms separately, we get: (52) (53) (54) (55) in, and Node Voltage and branches The square term of the current; and Node Voltage and branches The square term of the current; Since equations (54) and (55) contain square terms , so it is still a nonlinear expression. Therefore, Taylor series expansion is used to linearize equations (54) and (55), including: Assume that the initial value is , through Taylor series expansion we can get: (56) in, Represents an infinitesimal quantity; For Node The initial value of the voltage amplitude; By ignoring the second-order and higher-order terms, equation (56) can be approximately expressed as: (57) The node voltage amplitude is usually about 1.0 pu, so assuming , formula (57) can be further approximated as: (58) The distance function definition of the Taylor series approximation includes: (59) in, is the distance function approximated by the Taylor series; By substituting equation (58) into equations (54) and (55), an approximate load demand model is obtained, which can be expressed as: (60) (61) Convex optimization of nonlinear equations of the second type: Equations (32) and (41) can be transformed into second-order cone constraints and rotational second-order cone constraints, respectively, which can be expressed as: (62) (63) Convex optimization of the third type of nonlinear equations: Since equations (42), (43) and (48) all contain the product of two bounded continuous variables, the McCormick envelope approximation is used to process them and the auxiliary variable and Respectively and , and approximated using twelve linear constraints, can be expressed as: (64) (65) (66) (67) (68) (69) (70) (71) (72) (73) (74) (75)。 10. An active defense device for a new energy power distribution system taking into account the uncertainty of the evolution path of a typhoon or rainstorm, used in the method according to any one of claims 1 to 9, characterized in that: The device comprises a first module, a second module, a third module, a fourth module, a fifth module and a sixth module; The first module is used to obtain a dual uncertainty time series scenario of the typhoon and rainstorm evolution path and the coupling of the system component state; and obtain the vulnerable components at each moment during the typhoon and rainstorm duration period according to the time series scenario; The second module is used to switch the system to island operation and keep the dispatchable distributed power sources in the system in droop control; calculate the vulnerable load loss caused by disconnecting all vulnerable elements of the system at each moment through a multi-objective optimization model; The third module is used to consider each number of the total number of vulnerable elements from 0 to the current moment, and to combine each vulnerable element at the moment of vulnerable load loss when selecting any number of them, taking into account all the selected situations; and to calculate the survival probability of the combination of vulnerable elements; The fourth module is used to put into operation the vulnerable element combinations of each quantity at each vulnerable load loss moment in order from the smallest quantity to the largest quantity according to the survival probability from the largest to the smallest, and disconnect the vulnerable elements outside the running vulnerable element combinations; The fifth module is used to reschedule the system in which each vulnerable component combination is put into operation through active scheduling and a multi-objective optimization model, wherein the active scheduling includes power generation rescheduling, pre-disaster network reconstruction, droop control parameter optimization setting and protective voltage regulation; The sixth module is used to determine whether there is a vulnerable load loss in the system when each vulnerable component combination is put into operation in the order of operation. If not, active defense is performed at this moment based on the vulnerable component combination put into operation this time and the active scheduling.

Images