A typhoon rolling prediction-based urban power grid and mobile energy storage vehicle disaster-time cooperative scheduling method

Abstract

This invention discloses a disaster-time collaborative scheduling method for urban power grids and mobile energy storage vehicles based on typhoon rolling forecasts. The method includes: calculating the current fault scenario of the urban power grid based on typhoon and power grid data; for each time when MESS (Mobile Energy Storage System) scheduling is required, using a data-physical dual-drive forecasting model to predict typhoon forward data; based on the acquired typhoon forward data, optimizing the MESS location using a MESS location optimization model that takes typhoon forecasting into account, and issuing MESS location scheduling instructions; updating the MESS access status, and optimizing the urban system's optimal load shedding based on the power grid fault scenario and MESS access status using a system optimal load shedding model that takes into account the MESS location status, to obtain the optimal MESS output; determining whether the typhoon impact has ended; if not, returning to step S1; if it has ended, ending the MESS scheduling. This invention overcomes the lag problem of current MESS real-time scheduling strategies, maximizing the utilization of MESS's disaster-time power support capabilities.

Description

Technical Field

[0001] This invention belongs to the field of power and relates to power system resilience enhancement technology, specifically to a method for coordinated dispatch of urban power grids and mobile energy storage vehicles during disasters based on typhoon rolling forecasts. Background Technology

[0002] Power system resilience refers to the ability of a power system to maintain operation and recover under extreme disaster scenarios. Extreme weather, such as typhoons, is one type of extreme disaster. With the increasing frequency of various extreme weather disasters, improving power grid resilience has gradually become a research hotspot. Compared with large transmission networks, urban power grids exhibit unique disaster characteristics when facing typhoons. Urban power grids have high equipment density and small spatial scale. Suburbs are dominated by transmission networks with high and dispersed power loads, while urban areas are dominated by distribution networks with numerous and concentrated critical loads. Their terminals are equipped with abundant flexible resources, especially Mobile Energy Storage Systems (MESS). MESS combines energy storage and mobile transportation characteristics, overcoming the geographical limitations of fixed resources. It can flexibly connect to weak points in the grid during disaster scenarios to provide emergency power to critical loads, meeting the dual requirements of flexibility and reliability for urban power grids under extreme disasters.

[0003] However, current MESS scheduling methods for extreme typhoon disaster scenarios often focus on pre-disaster configuration or post-disaster recovery. The few strategies involving dynamic scheduling during a disaster typically only use the current system fault state as a boundary condition for instantaneous optimization, lacking forward-looking predictions of typhoon trajectories and disaster evolution trends. Considering that the geographical size of urban power grids is much smaller than that of typhoon disasters, and given the complex traffic constraints of urban road networks, this passive and lagging scheduling strategy can cause MESS to miss the optimal grid connection support window, limiting its effectiveness in improving system resilience. Furthermore, the aforementioned studies primarily focus on distribution networks, and the proposed strategies are not well-suited for urban power grids with transmission and distribution coordination characteristics. Summary of the Invention

[0004] Purpose of the invention: In order to overcome the shortcomings of the existing technology, this invention provides a disaster-time collaborative scheduling method for urban power grid and mobile energy storage vehicles based on typhoon rolling forecasts. This method can overcome the lag problem of the current MESS real-time scheduling strategy and maximize the utilization of MESS power support capabilities during disasters.

[0005] Technical Solution: To achieve the above objectives, this invention provides a method for coordinated dispatching of urban power grids and mobile energy storage vehicles during disasters based on typhoon rolling forecasts, comprising the following steps:

[0006] S1: Calculate the current fault scenarios of the urban power grid based on typhoon and power grid data;

[0007] S2: For each time when MESS scheduling is required, a data-physical dual-drive prediction model is used to predict typhoon forward data;

[0008] S3: Based on the acquired typhoon forecast data, the MESS location is optimized by taking into account the typhoon forecast MESS location optimization model, and a MESS location scheduling instruction is given.

[0009] S4: Update the MESS access status. Based on the power grid fault scenario and the MESS access status, optimize the load shedding of the urban system by taking into account the MESS location status into the optimal load shedding model, and obtain the optimal output of the MESS.

[0010] S5: Determine if the typhoon's impact has ended. If it has not ended, return to step S1. If it has ended, end the scheduling of MESS.

[0011] Furthermore, the data-physical dual-drive prediction model in step S2 serves as a typhoon evolution prediction model, comprising two parts: physical characteristic prediction based on empirical formulas and big data prediction based on the MCMC method.

[0012] Furthermore, the method for predicting physical features based on empirical formulas in step S2 includes:

[0013] Dividing the typhoon development process into several equal moments, the typhoon's latitude and longitude, movement speed, and angle of movement have the following relationships with historical data:

[0014]

[0015]

[0016] In the formula, For the typhoon Speed ​​of movement at any moment For the typhoon The angle of movement at any moment , They are respectively The longitude and latitude of the typhoon center at all times. , , , , , , , The parameters are obtained by performing linear regression on historical typhoon data.

[0017] Furthermore, the big data prediction method based on the MCMC method in step S2 includes:

[0018] Typhoon The state at any given moment is defined as a six-dimensional feature vector containing the difference in latitude, difference in longitude, central air pressure, maximum wind speed, speed, and difference in travel angle. If the physical process of a typhoon is viewed as a Markov chain, then there exists a probability transition matrix.

[0019]

[0020] In the formula, The sum of both rows and columns is 1, such that:

[0021]

[0022] probability transition matrix Characterizing Markov process vectors The inherent interrelationships, their column vectors Each element in the table represents a process variable. With process vector The relationships between the other elements in the text.

[0023] Further, in step S2, solving the probability transition matrix using the Monte Carlo sampling method includes:

[0024] First, a possible transition probability matrix is ​​preset. The acceptance of this transition probability matrix is ​​calculated using the maximum likelihood method with a large amount of historical data. If the requirements are met, the matrix is ​​accepted, perturbation is added to it, and the acceptance is calculated again. If it is not accepted, the matrix is ​​discarded, perturbation is added to the matrix in the previous step, and the acceptance is calculated again.

[0025] Repeat the above process until the probability transition matrix converges. During this process, when setting the initial probability transition matrix, consider the relationships between its parameters, such as the strong correlation between latitude and longitude differences and speed, and the strong autocorrelation of travel angle differences. Simultaneously, introduce a separate weighted correction for the maximum wind speed related to the central pressure, as shown in the following formula:

[0026]

[0027] in, , The maximum wind speed value obtained using the MCMC method. The maximum wind speed value calculated based on the Batts wind field model is as follows:

[0028]

[0029]

[0030]

[0031] in, for The radius of the wind circle corresponding to the maximum wind speed at any given moment; These are empirical constants; The Coriolis coefficient; The gradient wind speed of the radio station wind; express The pressure difference between the center of the low-pressure typhoon and the surrounding environment at any given time. This represents the typhoon's maximum wind speed at that moment.

[0032] In the MCMC method, the transition matrix remains unchanged at any given time. Considering the distinct geographical characteristics of typhoon development, historical typhoon data is grouped for training to derive different MCMC models. For a typhoon to be predicted, its similarity to various typhoon models is calculated based on its weighted Euclidean distance, and the most similar model is selected for prediction.

[0033] Furthermore, the process of optimizing the location of MESS using the MESS location optimization model that takes into account typhoon prediction in step S3 includes:

[0034] For each MESS, considering its speed in the traffic network and updating its location and load shedding status in real time, the following set of optimization equations based on DC power flow is formed:

[0035]

[0036] In the formula, For the total number of MESS, For the first The output of a MESS vehicle, for Time of the first MESS vehicles at the node The access flag, and satisfies:

[0037]

[0038] In the formula, =1 indicates the first Vehicle MESS access node , =0 indicates the first Vehicles not connected to the MESS node ; Let be the set of transmission network nodes in the urban power grid, and let the MESS access flag of these nodes be always equal to 0;

[0039] Considering the characteristics of urban power grid load distribution, the node loads in suburban transmission networks are typically heavy and unsuitable for support by MESS (Mechanical Service Provider Escalator). Urban transmission networks, on the other hand, usually carry dense, important loads and should be the areas supported by MESS endpoints. The inequality constraint means that at any given time, each MESS can only connect to one node, and each node can only connect to one MESS. (This is indicated by the MESS connection flag.) Composition matrix This indicates the optimal position of MESS at the current moment;

[0040] because and The matrices are all optimization variables, so the above equation needs to be linearized by defining auxiliary variables. , Modify the power equation and add the following constraints:

[0041]

[0042] To ensure the timeliness of scheduling instructions, it is assumed that all MESS speeds are [missing information]. It is necessary to constrain the destination of MESS based on distance to ensure that all MESS can reach their destination within half an hour:

[0043]

[0044] In the formula, This represents the distance from the current node to the target node in MESS.

[0045] Furthermore, in step S3, the MESS location optimization model considering typhoon prediction aims to minimize the total journey distance of all MESS vehicles. Under the premise of meeting the arrival constraint within half an hour, a unique target node is assigned to each MESS vehicle to maximize overall scheduling efficiency. Specifically, this includes:

[0046] The calculated movement speed of MESS is ; Let be the distance traveled, where For the total number of MESS, each element Representing the MESS vehicles and The distance between nodes is km; For the MESS location indication matrix, each element Only take 0 or 1, when When =1, it represents the first... MESS vehicle destination selection to node; This is the MESS actual driving distance matrix;

[0047] For the target node, each MESS vehicle is assigned its own destination using the following set of optimization equations:

[0048] .

[0049] Furthermore, the optimal load shedding model for the system taking into account the MESS location state in step S4 is expressed as follows:

[0050]

[0051] It is important to note that This is a marker indicating the current location of MESS, and... The difference is a definite value; with similar, satisfy:

[0052]

[0053] In the formula, Let be the set of transmission network nodes in the urban power grid. The inequality constraint means that at any given time, each MESS can only connect to one node, and each node can only connect to one MESS.

[0054] Furthermore, the scheduling process for MESS in step S5 includes:

[0055] A1: Initialize time, MESS state matrix, set MESS scheduling instruction flag to 0, load initial meteorological parameters of the typhoon and basic data of the urban power grid, and calculate the system component failure rate from the current moment based on the typhoon data.

[0056] A2: Determine whether the impact of the typhoon has ended. If not, proceed to A3; if yes, proceed to A10.

[0057] A3: Update the MESS status according to the scheduling instruction; skip if the MESS scheduling instruction flag is 0.

[0058] A4: Calculate the minimum load shedding amount based on the system status;

[0059] A5: Generate typical system failure scenarios based on component failure rates and update system status;

[0060] A6: Determine whether MESS needs to be scheduled at this time. If yes, go to A7; otherwise, go to A9.

[0061] A7: Simulate the disaster situation one hour after the typhoon forecast;

[0062] A8: Calculate the optimal MESS location, and set the MESS scheduling instruction flag to 1;

[0063] A9: Advance one time step, then proceed to A2;

[0064] A10: End scheduling.

[0065] Furthermore, the process of updating the MESS state in step A3 includes:

[0066] B1: Determine whether a scheduling instruction has been received based on the MESS scheduling instruction flag. If received, proceed to B2; otherwise, proceed to B5.

[0067] B2: Determine if the MESS is going to another node. If yes, go to B3; otherwise, go to B4.

[0068] B3: MESS disconnects from the current node;

[0069] B4: Update the distance between MESS and the destination node based on vehicle speed, then proceed to B8;

[0070] B5: Determine if MESS has reached the current destination. If not, proceed to B6; if yes, proceed to B7.

[0071] B6: Update the distance between MESS and the destination based on vehicle speed, then proceed to B8;

[0072] B7: MESS connects to the destination node and participates in system scheduling in the next moment;

[0073] B8: Output MESS status.

[0074] Beneficial effects: Compared with the prior art, the present invention has the following advantages:

[0075] 1. To address the issue of lag in current MESS real-time scheduling strategies, a MESS location optimization model based on typhoon forecasting is proposed. This model enables MESS to access system fault areas in advance, reducing its travel time in the transportation network and maximizing the utilization of MESS's disaster-time power support capabilities.

[0076] 2. A MESS location optimization model is proposed to address the characteristics of transmission and distribution coordination in urban power grids. MESS is used only for distribution network node support, making the strategy more adaptable to urban power grids. Attached Figure Description

[0077] Figure 1 A flowchart of the collaborative scheduling process provided for the method of this invention;

[0078] Figure 2 This is a flowchart of updating the real-time status of MESS in this invention;

[0079] Figure 3 A topology diagram of the city's power system;

[0080] Figure 4 This is a schematic diagram of the coupling between the city's electrical and transportation networks.

[0081] Figure 5 A diagram showing the load shedding situation of the urban power grid;

[0082] Figure 6 This is a diagram showing the load shedding situation from hour 18 to 22. Detailed Implementation

[0083] The present invention will be further illustrated below with reference to the accompanying drawings and specific embodiments. It should be understood that these embodiments are for illustrative purposes only and are not intended to limit the scope of the invention. After reading this invention, any modifications of the invention in various equivalent forms by those skilled in the art will fall within the scope defined by the appended claims.

[0084] Example 1:

[0085] like Figure 1 As shown, for a specific typhoon scenario and urban power grid-transportation network coupled topology, this embodiment provides a method for coordinated scheduling of urban power grid and mobile energy storage vehicles during disasters based on typhoon rolling forecasts, including the following steps:

[0086] S1: Calculate the current fault scenarios of the urban power grid based on typhoon and power grid data;

[0087] S2: For each time when MESS scheduling is required, a data-physical dual-drive prediction model is used to predict typhoon forward data;

[0088] The data-physics dual-drive prediction model, as a typhoon evolution prediction model, includes two parts: prediction based on physical characteristics using empirical formulas and prediction based on big data using the MCMC method.

[0089] Methods for predicting physical characteristics based on empirical formulas include:

[0090] Dividing the typhoon development process into several equal moments, the typhoon's latitude and longitude, movement speed, and angle of movement have the following relationships with historical data:

[0091]

[0092]

[0093] In the formula, For the typhoon Speed ​​of movement at any moment For the typhoon The angle of movement at any moment , They are respectively The longitude and latitude of the typhoon center at all times. , , , , , , , The parameters are obtained by performing linear regression on historical typhoon data.

[0094] Big data prediction methods based on the MCMC method include:

[0095] Typhoon The state at any given moment is defined as a six-dimensional feature vector containing the difference in latitude, difference in longitude, central air pressure, maximum wind speed, speed, and difference in travel angle. If the physical process of a typhoon is viewed as a Markov chain, then there exists a probability transition matrix.

[0096]

[0097] In the formula, The sum of both rows and columns is 1, such that:

[0098]

[0099] probability transition matrix Characterizing Markov process vectors The inherent interrelationships, their column vectors Each element in the table represents a process variable. With process vector The relationships between the other elements in the text.

[0100] In this embodiment, the probability transition matrix is ​​solved using the Monte Carlo sampling method, including:

[0101] First, a possible transition probability matrix is ​​preset. The acceptance of this transition probability matrix is ​​calculated using the maximum likelihood method with a large amount of historical data. If the requirements are met, the matrix is ​​accepted, perturbation is added to it, and the acceptance is calculated again. If it is not accepted, the matrix is ​​discarded, perturbation is added to the matrix in the previous step, and the acceptance is calculated again.

[0102] Repeat the above process until the probability transition matrix converges. During this process, when setting the initial probability transition matrix, consider the relationships between its parameters, such as the strong correlation between latitude and longitude differences and speed, and the strong autocorrelation of travel angle differences. Simultaneously, introduce a separate weighted correction for the maximum wind speed related to the central pressure, as shown in the following formula:

[0103]

[0104] in, , The maximum wind speed value obtained using the MCMC method. The maximum wind speed value calculated based on the Batts wind field model is as follows:

[0105]

[0106]

[0107]

[0108] in, for The radius of the wind circle corresponding to the maximum wind speed at any given moment; These are empirical constants; The Coriolis coefficient; The gradient wind speed of the radio station wind; express The pressure difference between the center of the low-pressure typhoon and the surrounding environment at any given time. This represents the typhoon's maximum wind speed at that moment.

[0109] In the MCMC method, the transition matrix remains unchanged at any given time. Considering the distinct geographical characteristics of typhoon development, historical typhoon data is grouped for training to derive different MCMC models. For a typhoon to be predicted, its similarity to various typhoon models is calculated based on its weighted Euclidean distance, and the most similar model is selected for prediction.

[0110] S3: Based on the acquired typhoon forecast data, the MESS location is optimized by taking into account the typhoon forecast MESS location optimization model, and a MESS location scheduling instruction is given.

[0111] The process of optimizing the location of MESS using the MESS location optimization model that takes typhoon forecasting into account includes:

[0112] For each MESS, considering its speed in the traffic network and updating its location and load shedding status in real time, the following set of optimization equations based on DC power flow is formed:

[0113]

[0114] In the formula, For the total number of MESS, For the first The output of a MESS vehicle, for Time of the first MESS vehicles at the node The access flag, and satisfies:

[0115]

[0116] In the formula, =1 indicates the first Vehicle MESS access node , =0 indicates the first Vehicles not connected to the MESS node ; Let be the set of transmission network nodes in the urban power grid, and let the MESS access flag of these nodes be always equal to 0;

[0117] Considering the characteristics of urban power grid load distribution, the node loads in suburban transmission networks are typically heavy and unsuitable for support by MESS (Mechanical Service Provider Escalator). Urban transmission networks, on the other hand, usually carry dense, important loads and should be the areas supported by MESS endpoints. The inequality constraint means that at any given time, each MESS can only connect to one node, and each node can only connect to one MESS. (This is indicated by the MESS connection flag.) Composition matrix This indicates the optimal position of MESS at the current moment;

[0118] because and The matrices are all optimization variables, so the above equation needs to be linearized by defining auxiliary variables. , Modify the power equation and add the following constraints:

[0119]

[0120] To ensure the timeliness of scheduling instructions, it is assumed that all MESS speeds are [missing information]. It is necessary to constrain the destination of MESS based on distance to ensure that all MESS can reach their destination within half an hour:

[0121]

[0122] In the formula, This represents the distance from the current node to the target node in MESS.

[0123] In actual scheduling, multiple MESS vehicles are distributed in different locations within the urban transportation network, and the travel time required for each to reach different target nodes varies. After determining the globally optimal access location, it is necessary to further solve the optimal matching problem between MESS vehicles and target nodes. To address this issue, the MESS location optimization model, which takes typhoon forecasting into account, aims to minimize the total journey distance of all MESS vehicles. Under the premise of satisfying the arrival constraint within half an hour, a unique target node is assigned to each MESS vehicle to maximize overall scheduling efficiency. Specifically, this includes:

[0124] The calculated movement speed of MESS is ; Let be the distance traveled, where For the total number of MESS, each element Representing the MESS vehicles and The distance between nodes is km; For the MESS location indication matrix, each element Only take 0 or 1, when When =1, it represents the first... MESS vehicle destination selection to node; This is the MESS actual driving distance matrix;

[0125] For the target node, each MESS vehicle is assigned its own destination using the following set of optimization equations:

[0126]

[0127] The above formula ensures that each mobile MESS selects the globally optimal destination and can reach the destination and connect to the node for power support within 0.5 hours. After obtaining the optimal MESS location, the destination is reassigned based on the MESS's current location and the total distance traveled, thus obtaining the optimal MESS scheduling instruction. After receiving the scheduling instruction, the MESS begins to move towards the destination node. Subsequently, based on the MESS speed, the MESS location is updated every 15 minutes, and the MESS access flag is also updated. Considering the time it takes for the MESS to connect to the node, if the MESS reaches the destination in the current moment, the MESS will participate in the optimal power flow calculation in the next moment.

[0128] S4: Update the MESS access status. Based on the power grid fault scenario and the MESS access status, optimize the load shedding of the urban system by taking into account the MESS location status into the optimal load shedding model, and obtain the optimal output of the MESS.

[0129] The optimal load shedding model for the system, taking into account the MESS location state, is expressed as follows:

[0130]

[0131] It is important to note that This is a marker indicating the current location of MESS, and... The difference is a definite value; with similar, satisfy:

[0132]

[0133] In the formula, Let be the set of transmission network nodes in the urban power grid. The inequality constraint means that at any given time, each MESS can only connect to one node, and each node can only connect to one MESS.

[0134] S5: Determine if the typhoon's impact has ended. If it has not ended, return to step S1. If it has ended, end the scheduling of MESS.

[0135] In this embodiment, the scheduling process for MESS includes:

[0136] A1: Initialize time, MESS state matrix, set MESS scheduling instruction flag to 0, load initial meteorological parameters of the typhoon and basic data of the urban power grid, and calculate the system component failure rate from the current moment based on the typhoon data.

[0137] A2: Determine whether the impact of the typhoon has ended. If not, proceed to A3; if yes, proceed to A10.

[0138] A3: Update the MESS status according to the scheduling instruction; skip if the MESS scheduling instruction flag is 0.

[0139] A4: Calculate the minimum load shedding amount based on the system status;

[0140] A5: Generate typical system failure scenarios based on component failure rates and update system status;

[0141] A6: Determine whether MESS needs to be scheduled at this time. If yes, go to A7; otherwise, go to A9.

[0142] A7: Simulate the disaster situation one hour after the typhoon forecast;

[0143] A8: Calculate the optimal MESS location, and set the MESS scheduling instruction flag to 1;

[0144] A9: Advance one time step, then proceed to A2;

[0145] A10: End scheduling.

[0146] like Figure 2 As shown, the process of updating the MESS state in step A3 includes:

[0147] B1: Determine whether a scheduling instruction has been received based on the MESS scheduling instruction flag. If received, proceed to B2; otherwise, proceed to B5.

[0148] B2: Determine if the MESS is going to another node. If yes, go to B3; otherwise, go to B4.

[0149] B3: MESS disconnects from the current node;

[0150] B4: Update the distance between MESS and the destination node based on vehicle speed, then proceed to B8;

[0151] B5: Determine if MESS has reached the current destination. If not, proceed to B6; if yes, proceed to B7.

[0152] B6: Update the distance between MESS and the destination based on vehicle speed, then proceed to B8;

[0153] B7: MESS connects to the destination node and participates in system scheduling in the next moment;

[0154] B8: Output MESS status.

[0155] Example 2:

[0156] To verify the effectiveness of the method of the present invention, the following simulation experiments and data analysis were conducted in this embodiment:

[0157] The urban power transmission and distribution coordination test system used in the simulation consists of four IEEE standard case studies: using the IEEE-30 node system as the urban backbone power grid, and connecting three radial distribution systems (IEEE-33, IEEE-15, and IEEE-10) to its terminal nodes, forming a combined system as follows: Figure 3 As shown. The system comprises 85 electrical nodes and 101 feeder branches, with a total active load of 206.51MW. Since this simulation focuses on method verification, it is assumed that all lines are connected in a straight line in space. Because the load type of the urban power grid is unverifiable, referring to urban construction planning guidelines, the load ratio of the first, second, and third levels of the distribution network in the urban power grid is set at 4:5:1, and the transmission network at 1:2:7, with each level of load randomly allocated among the corresponding grid nodes. Based on actual urban construction conditions, it is assumed that the system's transportation network and the system's power grid geographical topology are highly consistent. In the transportation network, roads connect nodes 36 and 44, 37 and 50, 39 and 41, 43 and 53, 50 and 51, 51 and 85, and 53 and 70. The electrical and transportation coupling diagram is shown below. Figure 4 As shown. Besides the generator and energy storage station nodes, the system is equipped with six identical MESS units, each with a maximum output of 250kW. This is insufficient to support the transmission network nodes, but it can provide support for the distribution network nodes. In this embodiment, the area formed by these nodes is considered to be the urban area of ​​the city power grid. Typhoon data was obtained from meteorological monitoring data, assuming it moved from the southeast to the northwest of the city. The simulation was programmed in MATLAB 2020a, using the matpower8.0 toolkit and the Cplex solver for calculation. The city power grid has energy storage stations at eight nodes: 15, 21, 39, 33, 51, 60, 74, and 85, with maximum outputs of 50, 50, 1, 2, 2, 0.6, 0.6, and 0.6MW respectively.

[0158] When performing simulation analysis on the resilience enhancement strategy based on MESS collaborative scheduling, the starting point of MESS is fixed at nodes 31, 33, 48, 54, 61, and 68. For the urban power grid scenario under the aforementioned typhoon disaster, using the same fault chain, and following three strategies—"no MESS for load support," "using MESS but without typhoon forecasting," and "using MESS collaborative scheduling after forecasting typhoon data" (hereinafter referred to as "MESS forecasting scheduling")—the load shedding amounts under each method are as follows: Figure 5 As shown in the figure, the overall trend of load shedding in the urban power grid is similar under the three MESS dispatch strategies mentioned above, which is consistent with the characteristic that MESS is not suitable for supporting the transmission network. During periods of large load fluctuations, the damaged lines in the urban power grid are mostly transmission lines. For example, during the 5-6 hour period of a sudden increase in load shedding, transmission line 7 was damaged. This line connects two important substations in the region, and its damage resulted in a large power deficit that MESS cannot support. However, during the period when the urban distribution network was already affected, it is clear that the results of MESS dispatch are better than those of not using MESS dispatch. The analysis is as follows. Figure 6 The load shedding data from hours 18 to 22 shows that, under the same fault chain, the scheme using MESS to support the distribution network is significantly better than not using it. Furthermore, the system load shedding situation is better when using the predictive dispatch strategy than when not predicting. Analysis of the typhoon's geographical data and that of the city reveals that the typhoon's maximum wind speed circle passed precisely through part of the urban distribution network during this period, and the MESS advance dispatch strategy provided timely support for the damage in this area. In terms of total load loss, it was 685.89 MW·h without MESS, 663.13 MW·h with MESS dispatch, and 656.04 MW·h with MESS predictive dispatch.

Claims

1. A typhoon rolling prediction-based urban power grid and mobile energy storage vehicle disaster-time collaborative scheduling method, characterized in that, Includes the following steps: S1: Calculate the current fault scenarios of the urban power grid based on typhoon and power grid data; S2: For each time when MESS scheduling is required, a data-physical dual-drive prediction model is used to predict typhoon forward data; S3: Based on the acquired typhoon forecast data, the MESS location is optimized by taking into account the typhoon forecast MESS location optimization model, and a MESS location scheduling instruction is given. S4: Update the MESS access status. Based on the power grid fault scenario and the MESS access status, optimize the load shedding of the urban system by taking into account the MESS location status into the optimal load shedding model, and obtain the optimal output of the MESS. S5: Determine if the typhoon's impact has ended. If it has not ended, return to step S1. If it has ended, end the scheduling of MESS.

2. The method according to claim 1, wherein, The data-physical dual-drive prediction model in step S2 includes two parts: physical feature prediction based on empirical formulas and big data prediction based on the MCMC method.

3. The method according to claim 2, wherein, The method for predicting physical features based on empirical formulas in step S2 includes: Dividing the typhoon development process into several equal moments, the typhoon's latitude and longitude, movement speed, and angle of movement have the following relationships with historical data: ； ； wherein, is the moving speed of the typhoon at time , is the moving angle of the typhoon at time , , are the longitude and latitude of the typhoon center at time , , , , , , , , are parameters, which are obtained by linear regression on historical typhoon data.

4. The method according to claim 3, characterized in that, The big data prediction method based on the MCMC method in step S2 includes: The state of a typhoon at a time instant is defined as a six-dimensional feature vector containing the latitude difference, longitude difference, central pressure, maximum wind speed, moving speed, and moving angle difference The state of a typhoon at a time instant is defined as a six-dimensional feature vector containing the latitude difference, longitude difference, central pressure, maximum wind speed, moving speed, and moving angle difference The physical process of a typhoon is regarded as a Markov chain, and there is a probability transition matrix ； wherein The row and column are both 1, so that: ； Probability transition matrix The Markov process vector is characterized by The internal interdependencies, whose column vectors are Each element in the vector represents a process variable The interdependencies with other elements in the vector The interdependencies with other elements in the vector 5. The method according to claim 4, wherein, Step S2, which involves solving the probability transition matrix using the Monte Carlo sampling method, includes: First, a possible transition probability matrix is ​​preset. The acceptance of this transition probability matrix is ​​calculated using the maximum likelihood method with a large amount of historical data. If the requirements are met, the matrix is ​​accepted, perturbation is added to it, and the acceptance is calculated again. If it is not accepted, the matrix is ​​discarded, perturbation is added to the matrix in the previous step, and the acceptance is calculated again. Repeat the above process until the probability transition matrix converges; during this process, the relationship between its parameters is considered when setting the initial probability transition matrix; at the same time, a separate weighted correction is introduced for the maximum wind speed, as shown in the following formula: ； wherein , is the maximum wind speed value derived by the MCMC method, is the maximum wind speed value calculated according to the Batts wind farm model, in particular as follows: ； ； ； wherein, is the maximum wind speed at the time corresponding to the radius of the wind circle; is an empirical constant; is the Coriolis coefficient; is the gradient wind speed of the radio station; denotes the pressure difference between the low-pressure typhoon center and the peripheral environment at the time, is the maximum wind speed of the typhoon at the time.

6. The method according to claim 5, wherein, The process of optimizing the location of MESS in step S3, which takes into account typhoon prediction, includes: For each MESS, considering its speed in the traffic network and updating its location and load shedding status in real time, the following set of optimization equations based on DC power flow is formed: ； wherein is the total number of MESSes, is the output of the th MESS, is the th moment of the th MESS at the node and satisfies: ； wherein = 1 indicates the first vehicle MESS access node , = 0 indicates the first vehicle MESS non-access node ; is a set of transmission grid nodes in the urban grid. Accessed by MESS consisting of matrix i.e. the optimal position of the MESS at the current time instant; Due to With Both matrices are optimization variables, the above equation needs to be linearized, define auxiliary variables , , modify the power equation and add the following constraints: 。 7. The method according to claim 6, wherein, In step S3, the MESS location optimization model considering typhoon forecasting aims to minimize the total journey distance of all MESS vehicles. Under the premise of meeting the arrival constraint within half an hour, a unique target node is assigned to each MESS vehicle to maximize overall scheduling efficiency. Specifically, this includes: The calculated movement speed of MESS is ; Let be the distance traveled, where For the total number of MESS, each element Representing the MESS vehicles and The distance between nodes is km; For the MESS location indication matrix, each element Only take 0 or 1, when When =1, it represents the first... MESS vehicle destination selection to node; This is the MESS actual driving distance matrix; For the target node, each MESS vehicle is assigned its own destination using the following set of optimization equations: 。 8. The method according to claim 7, wherein, The optimal load shedding model for the system, taking into account the MESS location state, in step S4 is expressed as follows: ； flagging the current MESS location, satisfies: ； In the formula, is a set of transmission network nodes in the urban power grid, and the inequality constraint means that each MESS is connected to only one node at the same time, and each node is connected to only one MESS.

9. The method according to claim 8, wherein, The scheduling process for MESS in step S5 includes: A1: Initialize time, MESS state matrix, set MESS scheduling instruction flag to 0, load initial meteorological parameters of the typhoon and basic data of the urban power grid, and calculate the system component failure rate from the current moment based on the typhoon data. A2: Determine whether the impact of the typhoon has ended. If not, proceed to A3; if yes, proceed to A10. A3: Update the MESS status according to the scheduling instruction; skip if the MESS scheduling instruction flag is 0. A4: Calculate the minimum load shedding amount based on the system status; A5: Generate typical system failure scenarios based on component failure rates and update system status; A6: Determine whether MESS needs to be scheduled at this time. If yes, go to A7; otherwise, go to A9. A7: Simulate the disaster situation one hour after the typhoon forecast; A8: Calculate the optimal MESS location, and set the MESS scheduling instruction flag to 1; A9: Advance one time step, then proceed to A2; A10: End scheduling.

10. The method according to claim 9, wherein, The process of updating the MESS status in step A3 includes: B1: Determine whether a scheduling instruction has been received based on the MESS scheduling instruction flag. If received, proceed to B2; otherwise, proceed to B5. B2: Determine if the MESS is going to another node. If yes, go to B3; otherwise, go to B4. B3: MESS disconnects from the current node; B4: Update the distance between MESS and the destination node based on vehicle speed, then proceed to B8; B5: Determine if MESS has reached the current destination. If not, proceed to B6; if yes, proceed to B7. B6: Update the distance between MESS and the destination based on vehicle speed, then proceed to B8; B7: MESS connects to the destination node and participates in system scheduling in the next moment; B8: Output MESS status.

Images