Vehicle-pile-network cooperative recovery method and device considering multiple types of mobile energy storage vehicles under emergency condition

Abstract

The invention relates to a vehicle-pile-network cooperative recovery method and device considering multiple types of mobile energy storage vehicles under an emergency condition, and the method comprises the following steps: constructing a power-loss load echelon decomposition recovery model of a power distribution network through calculating four index values of power utilization reliability, static voltage stability, load importance and load power-loss economic loss of various loads; based on the constructed power-loss load echelon decomposition recovery model of the power distribution network, establishing an emergency power supply recovery optimization scheduling model integrating three types of flexible scheduling resources of a mobile energy storage vehicle, an electric vehicle cluster and a diesel power generation vehicle; and solving the emergency power supply recovery optimization scheduling model, solving a mixed integer nonlinear optimization problem of the emergency power supply recovery optimization scheduling model, and obtaining a power loss load power supply recovery priority and a multi-type mobile power supply scheduling scheme.

Description

Technical Field

[0001] This invention belongs to the fields of energy distribution technology for multiple types of mobile energy storage vehicles and coordinated scheduling technology of electric vehicles-charging piles-distribution networks under emergency conditions. It relates to a method and device for coordinated recovery of vehicles-charging piles-networks under emergency conditions, and in particular, a method for coordinated recovery of vehicles-charging piles-networks under emergency conditions that takes into account multiple types of mobile energy storage vehicles. Background Technology

[0002] In recent years, frequent natural disasters have triggered severe power outages in distribution networks, causing significant economic losses. Sudden faults in distribution networks lead to changes in system topology, resulting in poorer power restoration of lost loads. The lack of collaborative restoration methods involving multiple types of mobile power sources and distribution network system reconfiguration leads to wasted dispatchable resources and excessively high restoration costs. Therefore, collaborative dispatch of multiple types of mobile power sources is a crucial strategy for ensuring efficient power restoration of distribution networks in emergency situations.

[0003] Current research primarily focuses on the application of mobile energy storage vehicles in day-ahead and real-time optimization and scheduling of distribution systems, promoting the consumption of distributed energy and improving the reliability of distribution networks. While mobile energy storage vehicles can achieve optimal load restoration in reconfigurable distribution networks, they lack consideration for the stability of the distribution network during load restoration. For large-scale power outages caused by extreme weather, some studies have proposed a post-disaster multi-source collaborative islanding operation strategy for distribution networks that considers mobile energy storage scheduling. Some technologies have quantitatively analyzed the control capabilities of mobile energy storage vehicles and proposed economic scheduling strategies to ensure adequate voltage in distribution networks under extreme scenarios, but they lack consideration for the coordinated optimization of other flexible scheduling resources in active distribution networks.

[0004] In summary, existing research neglects the impact of coordinated scheduling of multiple types of mobile power sources, such as mobile energy storage vehicles, electric vehicle clusters, and diesel generators, on the power restoration of the distribution network. Furthermore, the definition of the order of load power restoration is vague, and most studies only consider one type of dispatchable resource to participate in power restoration, resulting in a lack of flexibility in power restoration and high power restoration costs.

[0005] To address the aforementioned issues, this invention proposes a vehicle-pile-grid collaborative recovery method and device that considers multiple types of mobile energy storage vehicles under emergency conditions.

[0006] A search revealed no publicly available literature of the same or similar prior art as this invention. Summary of the Invention

[0007] To address the shortcomings of existing technologies, this invention proposes a vehicle-pile-grid collaborative restoration method and device under emergency conditions that considers multiple types of mobile energy storage vehicles. This method can obtain the priority order for power restoration of power outage loads and the scheduling scheme for multiple types of mobile power sources.

[0008] The above-mentioned objective of this invention is achieved through the following technical solution: A vehicle-pile-grid collaborative recovery method considering multiple types of mobile energy storage vehicles under emergency conditions includes the following steps: By calculating four types of indicators for various loads—power reliability, static voltage stability, load importance, and economic loss due to load power failure—a tiered decomposition and recovery model for power failure loads in the distribution network is constructed. Based on the constructed distribution network load loss tiered decomposition recovery model, an emergency power supply recovery optimization scheduling model is established that takes into account three types of flexible scheduling resources: mobile energy storage vehicles, electric vehicle clusters, and diesel generator vehicles. By inputting the power and state of charge of the flexible scheduling resources, as well as four types of indicators such as the load's power reliability, static voltage stability, load importance, and economic loss due to load power failure, the emergency power restoration optimization scheduling model is solved to obtain the priority order of power restoration for power-loss loads and multiple types of mobile power dispatch schemes, so as to achieve coordinated restoration of vehicles, charging piles, and networks under emergency conditions.

[0009] Furthermore, the specific steps for constructing a tiered decomposition and recovery model for power outage loads in the distribution network by calculating four types of indicators—power reliability, static voltage stability, load importance, and economic loss due to load power outage—include the following: Power reliability assessment under emergency conditions Average short-term power outage duration. The total average duration of short-term power outages experienced by all power-deprived loads during the entire power restoration period is denoted as... (h / household), the calculation method is as follows: (1) In the formula: Indicates the first in the distribution network n The load in the first j The duration of the short-term power outage; N load This indicates the total number of electrical loads in the power distribution network.

[0010] Average duration of power outage. The total average duration of continuous power outages experienced by all electrical loads during the entire power restoration period is denoted as . (h / household), the calculation method is as follows: (2) In the formula: Indicating the first in the distribution network n The electrical load is at the first j Duration of a single, continuous power outage.

[0011] Voltage compliance rate. The ratio of the duration of voltage compliance with the electrical load to the total duration of power restoration during the entire power restoration period, denoted as voltage compliance rate. l U-qualifiedThe calculation method is as follows: (3) In the formula: Indicates the distribution network number n The duration of voltage compliance for each electrical load during power restoration. T rco Indicates the time required for power restoration.

[0012] Static voltage stability assessment of power distribution system under emergency conditions i For the sending node, j As the receiving node, Z ji For line impedance, I ji For line current, For the apparent power at the receiving end, i ji Indicates the phase angle of the line. and These represent active power and reactive power, respectively. Z ji = R ji -jX ji The voltage stability index is calculated for any line in the system as follows: (4) In the formula: This is a voltage stability indicator. For each line in a real distribution network, It is a positive real number. If all lines If all values ​​are less than 1, it indicates that the distribution network system is statically stable; if at least one line... If the value is equal to 1, then the entire distribution network system is located on the stability boundary; if at least one line... If the value is greater than 1, the voltage of the distribution network system will become unstable.

[0013] Power loss load importance assessment System Expected Power Loss: The system expected power loss is used as an indicator to assess the importance of power outage loads. SEMP refers to the amount of power outage load that can be restored after repairs to a fault point in the distribution network following a fault, reflecting the extent of damage caused by the fault. The calculation method is as follows: (5) In the formula: T j To repair faulty load points j Duration used. N For the fault load point j The resulting power outage load collection. The SEMP weight value includes the impact coefficient of failing to complete the emergency repair task within the expected timeframe on subsequent repair work. α j The coefficient set for the travel time from the power bank to the fault location. β j and load weighting level coefficient c j ,Right now . P j To repair the fault point j The total power of the lost load that has been restored.

[0014] Importance factor of fault load point: Defined as the ratio of uncontrollable primary and secondary loads to the total power loss caused by the fault load point in the power loss load caused by the fault load point. .

[0015] (6) In the formula: The sum of primary loads among the power loss loads caused by the fault load point; This refers to the sum of uncontrollable secondary loads among the power loss loads caused by the fault point; P sum This represents the total load power.

[0016] Economic loss assessment of load power failure Combining load importance and load capacity, a calculation method for assessing the economic loss of load power failure can be derived, as shown in equation (7) below: (7) In the formula: E κ For load k The economic losses from power outages; P κ The capacity of the power outage load; v τ In the first t The rate of increase of load power loss during a power outage period, i.e., the rate of increase of load power loss; e For economic parameters; Δ t τ For the first t The duration of each power outage period; Γ is the total number of power outage periods.

[0017] By calculating four indicators—power reliability, static voltage stability, load importance, and economic loss due to load power outage—a tiered decomposition and recovery model for power outage loads in a distribution network was constructed. Furthermore, the specific steps for establishing an emergency power supply restoration optimization scheduling model that takes into account three types of flexible dispatch resources—mobile energy storage vehicles, electric vehicle clusters, and diesel generator vehicles—based on the constructed distribution network power outage load tiered decomposition and recovery model include: Construct an optimized scheduling model for emergency power restoration After a fault occurs, prioritizing the restoration of critical power-loss loads while ensuring the reliability and stability of the distribution network during power restoration, and also ensuring the economic efficiency of mobile power dispatch, are crucial considerations. Therefore, four types of objective functions are established to comprehensively evaluate the importance of power-loss loads and achieve economical and efficient restoration.

[0018] Establish an objective function for the reliability of load power consumption. As shown in formula (8), the optimization objective is to minimize the average short-term power outage duration and the average continuous power outage duration, and maximize the voltage qualification time of the power supply load.

[0019] (8) Where: the entire scheduling cycle is T Each scheduling time period is t , B It is a set of distribution network nodes.

[0020] Establish an objective function for the static voltage stability of the distribution network. As shown in formula (9), the optimization objective is to optimize all lines in the system. The maximum value of the indicator should be as small as possible.

[0021] (9) Prioritize the restoration of critical power-loss loads and establish an objective function. The formula is as follows: (10) To minimize the economic losses from load power outages, an objective function is established. The formula is as follows: (11) Based on the objective function , , and After determining the priority order for restoring power outage loads, the objective function is set to minimize the mobile power dispatch cost. for: (12) In the formula: M The set representing the total number of MPS; GThis represents the set of all mobile emergency generators (MEGs). S This represents the set of all mobile energy storage vehicles (MESVs). V Represents the set of all schedulable EV clusters; Indicates the first m The transportation cost coefficient of MPS vehicles; Indicates the first m MPS in time t The travel status (value is 1 when in a travel status, otherwise value is 0); k m Indicates the first m Battery charge / discharge degradation slope of a MESV or EV cluster; Indicates the first m Rated electricity price (RMB / kWh) for MESV or EV clusters; and Indicates the first m MESV or EV clusters in time t Charging and discharging power (kW); , d m Indicates the first m The electricity generation cost coefficient of a MEG generator; Indicates the first m MEG vehicles in time t Active power output (kW).

[0022] Constructing optimized scheduling security constraints set up B m Indicates that it can be connected to the first m The set of busbar nodes in the distribution network of a vehicle MPS. t This represents the travel time of the mobile power source upon receiving the dispatch instruction and arriving at the target location. When multiple types of mobile power sources collaborate, the power distribution network fault recovery problem needs to consider the following constraints.

[0023] Power bank connection constraints The mobile power supply can connect to at most one predetermined grid access point at any given time to provide power when needed, as shown in Equation (13).

[0024] (13) In the formula: Indicates the first m MPS in time t Time and Node i The connection status is set to 1 when connected and 0 otherwise.

[0025] Constraint (14) indicates that the number of mobile power sources allowed to be connected to a node is limited by the capacity of each candidate node.

[0026] (14) In the formula: Represents a node i The number of MPS allowed to access the site; M i Indicates that it can be connected to a node. i The MPS set.

[0027] Constraint (15) states that once a mobile power source is connected to a candidate node, it cannot be moved to another node.

[0028] (15) Mobile power path constraints Constraint (16) indicates that MPS transportation between different distribution network nodes meets the required travel time.

[0029] (16) In the formula: Indicates the first m MPS from node i To the node j Travel time, .

[0030] Assume that during the recovery process, MEGs equipped with fuel tanks can complete the dispatch by refueling from the tanks, while MESVs and EVs consume electrical energy during the dispatch process. The change of MESV's SOC over time is determined by its charging and discharging behavior, as shown in Equation (17): (17) In the formula: and Indicates MESV or EV cluster m The charging and discharging efficiency; Indicates the first m MESV in time t SOC. Indicates the first m The battery capacity of a MESV.

[0031] The SOC of an EV is determined by its charging and discharging and its travel behavior, as shown in Equation (18).

[0032] (18) In the formula: Indicates the first m Energy consumption (kW) of a group of EV clusters during travel; Indicates the firstm Group EV cluster in time t SOC. Indicates the first m Battery capacity of the EV cluster.

[0033] Constraint (19) represents the SOC range of MESV and EV over all time periods.

[0034] (19) In the formula: and Indicates MESV or EV cluster SOC m,t The minimum and maximum values.

[0035] Constraints (20) and (21) impose charging and discharging power constraints on the MESV and EV respectively based on their respective rated power.

[0036] (20) (twenty one) In the formula: , and Indicates the first m MESV or EV clusters in time t The charging and discharging state (the value is 1 when it is in the charging or discharging state, and 0 otherwise). and This indicates the maximum charging and discharging power (kW) of the MESV or EV cluster.

[0037] When MESV and EV are not connected to the distribution network, both charging and discharging power are forced to 0. The charging and discharging of MESV and EV are mutually exclusive in all time periods, as shown in Equation (22), which means that MPS disconnected from the distribution network can neither charge nor discharge.

[0038] (twenty two) Constraints (23) and (24) set the range of active and reactive power output of the MEG according to its rated power, and force the MEG to have zero active and reactive power output when disconnected from the distribution network.

[0039] (twenty three) (twenty four) In the formula: and Indicates the first m Maximum active and reactive power output of a MEG vehicle (kW, kVar). and Indicates the first m MEG vehicles in time t The active and reactive power output (kW, kVar).

[0040] Radial constraints of power distribution system Constraint (25) ensures that the remaining radiation network of the power distribution system can cover all time periods.

[0041] (25) In the formula: Indicates branch ( i , j In time t The connection status (1 if the branch is connected, 0 otherwise); D B Indicates the total number of nodes in the distribution network; Indicates due to time t The number of isolated islands formed by damaged and unrepaired branch roads.

[0042] Constraints (26) and (27) ensure power flow balance between load nodes and power nodes in each island.

[0043] (26) (27) In the formula: Indicates time t Time branch ( i , j The trend on ) Represents a node i In time t The load; Indicates time t node i The power supply.

[0044] Constraint (28) is a mandatory constraint, which ensures that the power flow is 0 in the open-circuit state.

[0045] (28) In the formula: N m It represents a sufficiently large positive number.

[0046] Branch state constraints Constraint (29) indicates that during the time period t The damaged branch must remain disconnected until it is repaired.

[0047] (29) In the formula: Indicates branch ( i , j In time t The damage status is indicated by a 0-1 symbol (1 if the branch is not damaged or has been repaired, otherwise 0).

[0048] Constraint (30) means that branches without remote control switches and without damage remain in their initial state at all times.

[0049] (30) In the formula: Indicates branch ( i , j The initial state is indicated by 0-1 symbols (1 if the branch is connected, 0 otherwise). , Indicates time t A group of branches that were damaged and have not yet been repaired; L swch This indicates a group of branches equipped with a remote control switch; L This represents the set of branches in the distribution network.

[0050] Power supply output power constraint Constraints (31) and (32) state that at a candidate node of an MPS, the active / reactive power injected or outflowed is equal to the sum of the active or reactive power outputs of all MPS.

[0051] (31) (32) In the formula: and Representing nodes respectively i The active and reactive power outputs (kW, kVar) of the MPS.

[0052] Constraint (22) indicates that the active or reactive power from the MPS to the unconnected MPS node is 0.

[0053] (33) Furthermore, the input of the flexible scheduling resources includes power and state of charge, as well as four types of indicators: load power reliability, static voltage stability, load importance, and economic loss due to load power outage. The specific steps for solving the emergency power supply restoration optimization scheduling model, obtaining the priority order for power supply restoration of power-outage loads and multiple types of mobile power supply scheduling schemes, thereby achieving coordinated restoration of vehicles, charging piles, and the grid under emergency conditions, include: The main steps for obtaining the Pareto optimal solution of a multi-objective optimization model using the adaptive weighted sum method are as follows.

[0054] objective function If ≥0, then the objective function , , and The priority Ф for restoring power-loss loads can be written in the following compact form: (34) (35) (36) In the formula: f ( x )=0 indicates equality constraints in the model. g ( x )≤0 indicates an inequality constraint in the model.

[0055] Construct the minimum objective function The single-objective optimization problem is solved by calculation. minimum value At the same time, the corresponding results can be obtained. maximum value , minimum value and maximum value Similarly, construct the objective function respectively. , , For a single-objective optimization problem, the objective function is obtained. , , and Maximum / Minimum Values , , and .

[0056] By calculating the maximum and minimum values ​​of the four objective functions, the objective function expression is normalized, i.e. (37) In the formula: , , and They are respectively , , and The per-unit value.

[0057] Using the weighted sum method with a smaller number of segments n initialSolving the normalized multi-objective model, the weighting factors of the objective function. l i ( i The uniform step size of (=1, 2, 3) is determined by the number of segments. n initial Decision, that is (38) Calculate the Euclidean distance between all adjacent solutions, according to the set precision. m Remove nearly overlapping solutions, i.e., solutions whose Euclidean distance is less than 1. m When the time comes, keep one solution and delete the rest.

[0058] The selected solutions are organized, with each pair of adjacent solutions forming a region. a Then determine the number of calculations required for each region. a The longer the Euclidean distance between the two endpoints of the region, the better. a The more times the calculations are refined, the more detailed the result. The degree of refinement for each region is determined based on the relative length of the Euclidean distance between the two endpoints of the region, as shown in equation (39): (39) In the formula: n a For the region a The number of calculations that need to be refined; l a For the region a Euclidean distance between the two endpoints; l avg This is the average length of the Euclidean distance between the two endpoints of all regions; C To pre-determine constants.

[0059] like n a ≤1, then the region a No further detailed calculations are needed; if n a If the value is greater than 1, then proceed to step (8).

[0060] Define area a The offset distance between the two endpoints. Connecting region. a endpoints and Create a piecewise linearized secant, and then select the Pareto front offset distance along the piecewise linearized secant direction. ( k =2, 3, 4; j =1, 2, 3, 4).

[0061] In the region aAdd additional inequality constraints and apply a weighted sum method to the region. a The optimization is performed, and the optimization problem is shown in equations (40) to (43): (40) (41) (42) · (43) In the formula: For the region a Weighting factors of the feasible region , i =1, 2, 3. Their uniform step size The number of refinement calculations obtained from step (6) n a The decision is as shown in equation (44): (44) The regions that cannot converge to the optimal solution during the refinement calculation are deleted because, in this case, the deleted regions are non-convex and do not contain the Pareto optimal solution.

[0062] Calculate the Lagrange distance between all adjacent solutions, and remove overlapping solutions. If the Lagrange distance between the endpoints of all regions is less than 1, then the solution is considered complete. m If the Lagrange distance between any two endpoints is greater than 1, then the optimization process ends. m If the selected solution is found in a certain region, the process jumps to the step of organizing the selected solutions for iterative calculation.

[0063] After obtaining the priority of load recovery using the adaptive weighted sum method, a genetic algorithm is used to solve the optimal economic scheduling scheme for various types of mobile power supplies, so as to achieve coordinated recovery of vehicle-pile-network under emergency conditions.

[0064] A vehicle-pile-grid collaborative recovery device for emergency conditions considering multiple types of mobile energy storage vehicles, comprising: The module for constructing a tiered decomposition and recovery model for power outage loads in a distribution network constructs a tiered decomposition and recovery model for power outage loads by calculating four types of indicators: power reliability, static voltage stability, load importance, and economic loss of loads when they lose power. The emergency power supply restoration optimization scheduling model construction module, based on the constructed distribution network power loss load tiered decomposition restoration model, establishes an emergency power supply restoration optimization scheduling model that takes into account three types of flexible scheduling resources: mobile energy storage vehicles, electric vehicle clusters, and diesel generator vehicles. The collaborative recovery module takes into account the power and charge status of the flexibly dispatchable resources, as well as four types of indicators: load power reliability, static voltage stability, load importance, and economic loss of load power failure. It solves the emergency power supply recovery optimization scheduling model to obtain the priority order of power supply recovery for power failure loads and multiple types of mobile power supply scheduling schemes, thereby realizing the collaborative recovery of vehicle-pile-grid under emergency conditions.

[0065] Furthermore, the power distribution network load loss cascade decomposition and recovery model construction module also includes: The emergency load power reliability assessment module is used for: The average duration of short-term power outages is the total average duration of short-term power outages experienced by all power-depleted loads during the entire power restoration period. (h / household), the calculation method is as follows: (1) In the formula: Indicating the first in the distribution network n The load in the first j The duration of the short-term power outage; N load This indicates the total number of electrical loads in the distribution network; Average duration of continuous power outage: The total average duration of continuous power outages experienced by all electrical loads during the entire power restoration period is denoted as . (h / household), the calculation method is as follows: (2) In the formula: Indicating the first in the distribution network n The electrical load is at the first j Duration of a single continuous power outage; Voltage qualification rate, denoted as the ratio of the duration of voltage qualification for the electrical load to the duration of power restoration during the entire power restoration period, is denoted as... l U-qualified The calculation method is as follows: (3) In the formula: Indicates the distribution network number n The duration of voltage compliance for each electrical load during power restoration. T rco Indicates the power restoration time; The static voltage stability assessment module for power distribution systems in emergency situations is used for: i For the sending node, j As the receiving node, Z ji For line impedance, I jiFor line current, For the apparent power at the receiving end, i ji Indicates the phase angle of the line. and These represent active power and reactive power, respectively. Z ji = R ji -jX ji The voltage stability index is calculated for any line in the system as follows: (4) In the formula: For voltage stability indicators; for each line in an actual distribution network, It is a positive real number; if all lines If all values ​​are less than 1, it indicates that the distribution network system is statically stable; if at least one line... If the value is equal to 1, then the entire distribution network system is located on the stability boundary; if at least one line... If the value is greater than 1, the voltage of the distribution network system will become unstable; The power failure load importance assessment module is used for: System Expected Power Loss: The system expected power loss is used as an indicator to assess the importance of power outage loads; SEMP refers to the amount of power outage load that can be restored after repairing a fault point in the distribution network, reflecting the degree of damage caused by the fault. The calculation method is as follows: (5) In the formula: T j To repair faulty load points j Duration used; N For the fault load point j The resulting collection of power loss loads; The SEMP weight value includes the impact coefficient of failing to complete the emergency repair task within the expected timeframe on subsequent repair work. α j The coefficient set for the travel time from the power bank to the fault location. β j and load weighting level coefficient c j ,Right now ; P j To repair the fault point j Total power of the lost load restored; Importance factor of fault load point: Defined as the ratio of uncontrollable primary and secondary loads to the total power loss caused by the fault load point in the power loss load caused by the fault load point. ; (6) In the formula: The sum of primary loads among the power loss loads caused by the fault load point; This refers to the sum of uncontrollable secondary loads among the power loss loads caused by the fault point; This represents the total load power. The load power failure economic loss assessment module is used for: Combining load importance and load capacity, a calculation method for assessing the economic loss of load power failure is derived, as shown in equation (7) below: (7) In the formula: E κ For load k The economic losses from power outages; P κ The capacity of the power outage load; v τ In the first t The rate of increase of load power loss during a power outage period, i.e., the rate of increase of load power loss; e For economic parameters; Δ t τ For the first t The duration of each power outage period; Γ represents the total number of power outage periods; By calculating four types of indicators for various loads—power reliability, static voltage stability, load importance, and economic loss due to load power failure—a tiered decomposition and recovery model for power loss loads in the distribution network is constructed.

[0066] A computer-readable storage medium having program instructions stored thereon, which, when executed by a processor, implement the vehicle-pile-grid collaborative recovery method under emergency conditions considering multiple types of mobile energy storage vehicles, as described in any one of claims 1 to 4.

[0067] The advantages and positive effects of this invention are as follows: 1. This invention proposes a vehicle-pile-grid collaborative restoration method and device under emergency conditions that considers multiple types of mobile energy storage vehicles. It can comprehensively evaluate the priority of power grid load restoration under emergency conditions by considering four types of indicators: load power reliability, distribution network static voltage stability, load importance, and economic loss of load power failure.

[0068] 2. It can fully utilize various types of mobile power sources, including mobile energy storage vehicles, electric vehicle clusters, and diesel generators, to minimize the cost of mobile power source dispatch while prioritizing the restoration of critical loads.

[0069] 3. This invention can obtain a uniformly distributed Pareto front solution, thereby obtaining the priority order for power restoration of power-outage loads and multiple types of mobile power dispatching schemes. The emergency power restoration strategy for distribution networks considering mobile energy storage vehicles proposed in this invention reduces fault recovery time by 6.44% and 8.02% respectively, and reduces dispatching costs by 18.14% and 13.23% respectively compared with the comparative strategies, improving the priority restoration capability for important power-outage loads and ensuring the efficiency and flexibility of emergency power restoration of distribution networks. Attached Figure Description

[0070] Figure 1 This is a single-line diagram of the power distribution system cable line of the present invention; Figure 2 This is a diagram illustrating the islanding of the distribution network after a fault, as presented in this invention. Figure 3 This is a comparison chart of load recovery under different scenarios according to the present invention. Detailed Implementation

[0071] The structure of the present invention will be further described below with reference to the accompanying drawings and embodiments. It should be noted that these embodiments are descriptive and not limiting.

[0072] A vehicle-pile-grid collaborative recovery method considering multiple types of mobile energy storage vehicles under emergency conditions includes the following steps: By calculating four types of indicators for various loads—power reliability, static voltage stability, load importance, and economic loss due to load power failure—a tiered decomposition and recovery model for power failure loads in the distribution network is constructed. The specific steps for constructing a tiered decomposition and recovery model for power outage loads in the distribution network by calculating four types of indicators—power reliability, static voltage stability, load importance, and economic loss due to load power outage—include the following: 1) Power supply reliability assessment under emergency conditions (1) Average duration of short-term power outage. The total average duration of short-term power outages experienced by all power-loss loads during the entire power restoration period is denoted as . (h / household), the calculation method is as follows: (1) In the formula: Indicating the first in the distribution network n The load in the first j The duration of the short-term power outage; N load This indicates the total number of electrical loads in the power distribution network.

[0073] (2) Average duration of continuous power outage. The total average duration of continuous power outages experienced by all electrical loads during the entire power restoration period is denoted as . (h / household), the calculation method is as follows: (2) In the formula: Indicating the first in the distribution network n The electrical load is at the first j Duration of a single, continuous power outage.

[0074] (3) Voltage Qualification Rate. The ratio of the duration of voltage qualification for the electrical load to the duration of power restoration during the entire power supply restoration period is denoted as: l U-qualified The calculation method is as follows: (3) In the formula: Indicates the distribution network number n The duration of voltage compliance for each electrical load during power restoration. T rco Indicates the time required for power restoration.

[0075] 2) Static voltage stability assessment of power distribution system under emergency conditions like Figure 1 As shown, i For the sending node, j As the receiving node, Z ji For line impedance, I ji For line current, For the apparent power at the receiving end, i ji Indicates the phase angle of the line. and These represent active power and reactive power, respectively. Z ji = R ji -jX ji The voltage stability index is calculated for any line in the system as follows: (4) In the formula: This is a voltage stability indicator. For each line in a real distribution network, It is a positive real number. If all lines If all values ​​are less than 1, it indicates that the distribution network system is statically stable; if at least one line... If the value is equal to 1, then the entire distribution network system is located on the stability boundary; if at least one line... If the value is greater than 1, the voltage of the distribution network system will become unstable.

[0076] 3) Assessment of the importance of power loss load (1) System Expects Missing Power (SEMP): The system expected missing power is used as an indicator to assess the importance of power loss load. SEMP refers to the amount of power loss load that can be restored after emergency repairs to a fault point in the distribution network, reflecting the degree of damage caused by the fault. The calculation method is as follows: (5) In the formula: T j To repair faulty load points j Duration used. N For the fault load point j The resulting power outage load collection. The SEMP weight value includes the impact coefficient of failing to complete the emergency repair task within the expected timeframe on subsequent repair work. α j The coefficient set for the travel time from the power bank to the fault location. β j and load weighting level coefficient c j ,Right now . P j To repair the fault point j The total power of the lost load that has been restored.

[0077] (2) Failure Point Importance Coefficient (FPIC): The ratio of uncontrollable primary and secondary loads to the total power loss caused by the fault point in the power loss load caused by the fault point is defined as the FPIC. .

[0078] (6) In the formula: The sum of primary loads among the power loss loads caused by the fault load point; This refers to the sum of uncontrollable secondary loads among the power loss loads caused by the fault point; This represents the total load power.

[0079] 4) Assessment of economic losses from load power failure Combining load importance and load capacity, a calculation method for assessing the economic loss of load power failure can be derived, as shown in equation (7) below: (7) In the formula: E κ For load k The economic losses from power outages; P κ The capacity of the power outage load; v τ In the first t The rate of increase of load power loss during a power outage period, i.e., the rate of increase of load power loss; e For economic parameters; Δ t τ For the first t The duration of each power outage period; Γ is the total number of power outage periods.

[0080] By calculating four indicators—power reliability, static voltage stability, load importance, and economic loss due to load power outage—a tiered decomposition and recovery model for power outage loads in a distribution network was constructed. Based on the constructed distribution network power loss load tiered decomposition recovery model, an emergency power supply recovery optimization scheduling model is established for three types of flexible scheduling resources: integrated mobile energy storage vehicles, electric vehicle clusters, and diesel generator vehicles. The specific steps for establishing an optimized scheduling model for emergency power supply restoration based on the constructed distribution network load loss tiered decomposition recovery model, which integrates three types of flexible dispatch resources—mobile energy storage vehicles, electric vehicle clusters, and diesel generators—include the following: 1) Construct an optimized scheduling model for emergency power supply restoration After a fault occurs, prioritizing the restoration of critical power-loss loads while ensuring the reliability and stability of the distribution network during power restoration, and also ensuring the economic efficiency of mobile power dispatch, are crucial considerations. Therefore, four types of objective functions are established to comprehensively evaluate the importance of power-loss loads and achieve economical and efficient restoration.

[0081] (1) Establish an objective function for the reliability of load power consumption. As shown in formula (8), the optimization objective is to minimize the average short-term power outage duration and the average continuous power outage duration, and maximize the voltage qualification time of the power supply load.

[0082] (8) Where: the entire scheduling cycle is T Each scheduling time period is t , B It is a set of distribution network nodes.

[0083] (2) Establish an objective function for the static voltage stability of the distribution network. As shown in formula (9), the optimization objective is to optimize all lines in the system. The maximum value of the indicator should be as small as possible.

[0084] (9) (3) Prioritize the restoration of critical power-loss loads and establish an objective function. The formula is as follows: (10) (4) Considering minimizing the economic losses from load power outages, establish the objective function. The formula is as follows: (11) (5) Based on the objective function , , and After determining the priority order for restoring power outage loads, the objective function is set to minimize the mobile power dispatch cost. for: (12) In the formula: M The set representing the total number of MPS; G This represents the set of all mobile emergency generators (MEGs). S This represents the set of all mobile energy storage vehicles (MESVs). V Represents the set of all schedulable EV clusters; Indicates the first m The transportation cost coefficient of MPS vehicles; Indicates the first m MPS in time t The travel status (value is 1 when in a travel status, otherwise value is 0); k m Indicates the first m Battery charge / discharge degradation slope of a MESV or EV cluster; Indicates the first m Rated electricity price (RMB / kWh) for MESV or EV clusters; and Indicates the first m MESV or EV clusters in time t Charging and discharging power (kW); , d m Indicates the first m The electricity generation cost coefficient of a MEG generator; Indicates the first m MEG vehicles in time t Active power output (kW).

[0085] 2) Construct optimized scheduling security constraints set up B m Indicates that it can be connected to the firstm The set of busbar nodes in the distribution network of a vehicle MPS. t This represents the travel time of the mobile power source upon receiving the dispatch instruction and arriving at the target location. When multiple types of mobile power sources collaborate, the power distribution network fault recovery problem needs to consider the following constraints.

[0086] (1) Power supply connection constraints The mobile power supply can connect to at most one predetermined grid access point at any given time to provide power when needed, as shown in Equation (13).

[0087] (13) In the formula: Indicates the first m MPS in time t Time and Node i The connection status is set to 1 when connected and 0 otherwise.

[0088] Constraint (14) indicates that the number of mobile power sources allowed to be connected to a node is limited by the capacity of each candidate node.

[0089] (14) In the formula: Represents a node i The number of MPS allowed to access the site; M i Indicates that it can be connected to a node. i The MPS set.

[0090] Constraint (15) states that once a mobile power source is connected to a candidate node, it cannot be moved to another node.

[0091] (15) (2) Mobile power supply path constraints Constraint (16) indicates that MPS transportation between different distribution network nodes meets the required travel time.

[0092] (16) In the formula: Indicates the first m MPS from node i To the node j Travel time, .

[0093] Assume that during the recovery process, MEGs equipped with fuel tanks can complete the dispatch by refueling from the tanks, while MESVs and EVs consume electrical energy during the dispatch process. The change of MESV's SOC over time is determined by its charging and discharging behavior, as shown in Equation (17): (17) In the formula: and Indicates MESV or EV cluster m The charging and discharging efficiency; Indicates the first m MESV in time t SOC. Indicates the first m The battery capacity of a MESV.

[0094] The SOC of an EV is determined by its charging and discharging and its travel behavior, as shown in Equation (18).

[0095] (18) In the formula: Indicates the first m Energy consumption (kW) of a group of EV clusters during travel; Indicates the first m Group EV cluster in time t SOC. Indicates the first m Battery capacity of the EV cluster.

[0096] Constraint (19) represents the SOC range of MESV and EV over all time periods.

[0097] (19) In the formula: and Indicates MESV or EV cluster SOC m,t The minimum and maximum values.

[0098] Constraints (20) and (21) impose charging and discharging power constraints on the MESV and EV respectively based on their respective rated power.

[0099] (20) (twenty one) In the formula: , and Indicates the first m MESV or EV clusters in time t The charging and discharging state (the value is 1 when it is in the charging or discharging state, and 0 otherwise). and This indicates the maximum charging and discharging power (kW) of the MESV or EV cluster.

[0100] When MESV and EV are not connected to the distribution network, both charging and discharging power are forced to 0. The charging and discharging of MESV and EV are mutually exclusive in all time periods, as shown in Equation (22), which means that MPS disconnected from the distribution network can neither charge nor discharge.

[0101] (twenty two) Constraints (23) and (24) set the range of active and reactive power output of the MEG according to its rated power, and force the MEG to have zero active and reactive power output when disconnected from the distribution network.

[0102] (twenty three) (twenty four) In the formula: and Indicates the first m Maximum active and reactive power output of a MEG vehicle (kW, kVar). and Indicates the first m MEG vehicles in time t The active and reactive power output (kW, kVar).

[0103] (3) Radial constraints of the power distribution system Constraint (25) ensures that the remaining radiation network of the power distribution system can cover all time periods.

[0104] (25) In the formula: Indicates branch ( i , j In time t The connection status (1 if the branch is connected, 0 otherwise); D B Indicates the total number of nodes in the distribution network; Indicates due to time t The number of isolated islands formed by damaged and unrepaired branch roads.

[0105] Constraints (26) and (27) ensure power flow balance between load nodes and power nodes in each island.

[0106] (26) (27) In the formula: Indicates time t Time branch ( i , j The trend on ) Represents a nodei In time t The load; Indicates time t node i The power supply.

[0107] Constraint (28) is a mandatory constraint, which ensures that the power flow is 0 in the open-circuit state.

[0108] (28) In the formula: N m It represents a sufficiently large positive number.

[0109] (4) Branch state constraints Constraint (29) indicates that during the time period t The damaged branch must remain disconnected until it is repaired.

[0110] (29) In the formula: Indicates branch ( i , j In time t The damage status is indicated by a 0-1 symbol (1 if the branch is not damaged or has been repaired, otherwise 0).

[0111] Constraint (30) means that branches without remote control switches and without damage remain in their initial state at all times.

[0112] (30) In the formula: Indicates branch ( i , j The initial state is indicated by 0-1 symbols (1 if the branch is connected, 0 otherwise). , Indicates time t A group of branches that were damaged and have not yet been repaired; L swch This indicates a group of branches equipped with a remote control switch; L This represents the set of branches in the distribution network.

[0113] (5) Output power constraint of mobile power supply Constraints (31) and (32) state that at a candidate node of an MPS, the active / reactive power injected or outflowed is equal to the sum of the active or reactive power outputs of all MPS.

[0114] (31) (32) In the formula: and Representing nodes respectively i The active and reactive power outputs (kW, kVar) of the MPS.

[0115] Constraint (22) indicates that the active or reactive power from the MPS to the unconnected MPS node is 0.

[0116] (33) To address the difficulty of obtaining uniformly distributed Pareto front solutions directly from emergency power restoration optimization scheduling models using traditional algorithms, an adaptive weighted genetic algorithm is proposed to solve the emergency power restoration optimization scheduling model. This algorithm solves the mixed-integer nonlinear optimization problem of the emergency power restoration optimization scheduling model, obtaining the priority order for power restoration of power-outage loads and multiple types of mobile power supply scheduling schemes.

[0117] The specific steps for solving the emergency power supply restoration optimization scheduling model, which involves solving a mixed-integer nonlinear optimization problem to obtain the priority order for power supply restoration of power-outage loads and multiple types of mobile power supply scheduling schemes, include: The main steps for obtaining the Pareto optimal solution of a multi-objective optimization model using the adaptive weighted sum method are as follows.

[0118] Step (1): Objective function Then the objective function , , and The priority Ф for restoring power-loss loads can be written in the following compact form: (34) (35) (36) In the formula: f ( x )=0 indicates equality constraints in the model. g ( x )≤0 indicates an inequality constraint in the model.

[0119] Step (2): Construct the minimum objective function The single-objective optimization problem is solved by calculation. minimum value At the same time, the corresponding results can be obtained. maximum value , minimum value and maximum value Similarly, construct the objective function respectively. , , For a single-objective optimization problem, the objective function is obtained. , , and Maximum / Minimum Values , , and .

[0120] Step (3): Using the maximum and minimum values ​​of the four objective functions obtained in steps (1) and (2), normalize the objective function expression, i.e. (37) In the formula: , , and They are respectively , , and The per-unit value.

[0121] Step (4): Use the weighted sum method with a smaller number of segments. n initial Solving the normalized multi-objective model, the weighting factors of the objective function. l i ( i The uniform step size of (=1, 2, 3) is determined by the number of segments. n initial Decision, that is (38) Because the step size of the weighting factor in step (4) is relatively large, the number of solutions obtained in this step is relatively small.

[0122] Step (5): Calculate the Euclidean distance between all adjacent solutions obtained in step (4), according to the set precision. m Remove nearly overlapping solutions, i.e., solutions whose Euclidean distance is less than 1. m When the time comes, keep one solution and delete the rest.

[0123] Step (6): Organize the solutions selected in step (5), with each pair of adjacent solutions forming a region. a Then determine the number of calculations required for each region. a The longer the Euclidean distance between the two endpoints of the region, the better. aThe more times the calculations are refined, the more detailed the result. The degree of refinement for each region is determined based on the relative length of the Euclidean distance between the two endpoints of the region, as shown in equation (39): (39) In the formula: n a For the region a The number of calculations that need to be refined; l a For the region a Euclidean distance between the two endpoints; l avg This is the average length of the Euclidean distance between the two endpoints of all regions; C To pre-determine constants.

[0124] Step (7): If n a ≤1, then the region a No further detailed calculations are needed; if n a If the value is greater than 1, then proceed to step (8).

[0125] Step (8): Determine the area a The offset distance between the two endpoints. Connecting region. a endpoints and Create a piecewise linearized secant, and then select the Pareto front offset distance along the piecewise linearized secant direction. ( k =2, 3, 4; j =1,2,3,4).

[0126] Step (9): In the area a Add additional inequality constraints and apply a weighted sum method to the region. a The optimization is performed, and the optimization problem is shown in equations (40) to (43): (40) (41) (42) · (43) In the formula: For the region a Weighting factors of the feasible region I am ∈(0.1), i =1, 2, 3. Their uniform step size The number of refinement calculations obtained from step (6) n a The decision is as shown in equation (44): (44) The regions that cannot converge to the optimal solution during the refinement calculation are deleted because, in this case, the deleted regions are non-convex and do not contain the Pareto optimal solution.

[0127] Step (10): Calculate the Lagrange distance between all adjacent solutions and delete overlapping solutions. If the Lagrange distance between the two endpoints of all regions is less than 10, then the solution is considered to be incomplete. m If the Lagrange distance between any two endpoints is greater than 1, then the optimization process ends. m If the region is not specified, then proceed to step (6) for iterative calculation.

[0128] Step (11): After obtaining the priority of load recovery using the adaptive weighted sum method, the genetic algorithm is used to solve the optimal economic scheduling scheme for various types of mobile power supplies.

[0129] The innovation of this invention lies in: (1) Based on the constructed distribution network power loss load tiered decomposition recovery model, an emergency power supply recovery optimization scheduling model is established for three types of flexible scheduling resources: integrated mobile energy storage vehicles, electric vehicle clusters and diesel generator vehicles.

[0130] (2) To address the problem that the Pareto front solution with uniform distribution is difficult to obtain directly through traditional algorithms in the emergency power restoration optimization scheduling model, an algorithm for solving the emergency power restoration optimization scheduling model based on adaptive weighted genetic algorithm is proposed. This algorithm solves the mixed integer nonlinear optimization problem of the emergency power restoration optimization scheduling model and obtains the priority order of power restoration of power outage loads and multiple types of mobile power scheduling schemes.

[0131] It should be emphasized that the embodiments described in this invention are illustrative rather than limiting. Therefore, this invention includes, but is not limited to, the embodiments described in the specific implementation. Any other implementations derived by those skilled in the art based on the technical solutions of this invention are also within the scope of protection of this invention.

Claims

1. A vehicle-pile-grid collaborative recovery method under emergency conditions considering multiple types of mobile energy storage vehicles, characterized in that: Includes the following steps: By calculating four types of indicators for various loads—power reliability, static voltage stability, load importance, and economic loss due to load power failure—a tiered decomposition and recovery model for power failure loads in the distribution network is constructed. Based on the constructed distribution network load loss tiered decomposition recovery model, an emergency power supply recovery optimization scheduling model is established that takes into account three types of flexible scheduling resources: mobile energy storage vehicles, electric vehicle clusters, and diesel generator vehicles. By inputting the power and state of charge of the flexible scheduling resources, as well as four types of indicators such as the load's power reliability, static voltage stability, load importance, and economic loss due to load power failure, the emergency power restoration optimization scheduling model is solved to obtain the priority order of power restoration for power-loss loads and multiple types of mobile power dispatch schemes, thereby achieving coordinated restoration of vehicles, charging piles, and the network under emergency conditions.

2. The vehicle-pile-grid collaborative recovery method under emergency conditions considering multiple types of mobile energy storage vehicles as described in claim 1, characterized in that: The specific steps for constructing a tiered decomposition and recovery model for power outage loads in the distribution network by calculating four types of indicators—power reliability, static voltage stability, load importance, and economic loss due to load power outage—include the following: Emergency load power reliability assessment: Average short-term power outage duration, which is the total average duration of all power outages experienced by all power-loss loads during the entire power restoration period. (h / household), the calculation method is as follows: In the formula: N represents the duration of the power outage of the nth load in the distribution network during the j-th short-time power outage; load This indicates the total number of electrical loads in the distribution network; Average duration of continuous power outage: The total average duration of continuous power outages experienced by all electrical loads during the entire power restoration period is denoted as . (h / household), the calculation method is as follows: In the formula: This represents the duration of the power outage for the nth electrical load in the distribution network during the jth continuous power outage. Voltage qualification rate, denoted as λ, is the ratio of the duration of voltage qualification for the electrical load to the duration of power restoration during the entire power restoration period. U-qualified The calculation method is as follows: In the formula: t qn T represents the duration of voltage compliance for the nth electrical load in the distribution network during the power restoration period. rco Indicates the time required for power restoration; Static voltage stability assessment of power distribution system under emergency conditions: i is the sending-end node, j is the receiving-end node, Z ji I is the line impedance. ji S is the line current. SVSI ji For the apparent power at the receiving end, θ ji P represents the phase angle of the line. SVSI ji and Q SVSI ji Z represents active power and reactive power respectively; ji =R ji -jX ji The voltage stability index is calculated for any line in the system as follows: In the formula: γ SVSI ji γ is a voltage stability indicator; for each line in a real distribution network, γ SVSI ji It is a positive real number; if γ of all lines SVSI ji If all values ​​are less than 1, it indicates that the distribution network system is statically stable; if at least one line has a γ value less than 1, it indicates that the distribution network system is statically stable. SVSI ji If the value is equal to 1, the entire distribution network system is located on the stability boundary; if at least one line's γ value is equal to 1, then the entire distribution network system is located on the stability boundary. SVSI ji If the value is greater than 1, the voltage of the distribution network system will become unstable; Power Loss Load Importance Assessment: System Expected Power Loss: The system expected power loss is used as an indicator for assessing the importance of power loss loads; SEMP refers to the amount of power loss load that can be restored after repairing a fault point in the distribution network, reflecting the degree of damage caused by the fault. The calculation method is as follows: In the formula: T j The time required to repair faulty load point j; N is the set of power loss loads caused by faulty load point j; σ SEMP j The SEMP weight value includes the impact coefficient α of failing to complete the emergency repair task within the expected timeframe on subsequent repair work. j The coefficient β set for the travel time from the power bank to the fault location. j and load weighting level coefficient γ j , i.e. σ SEMP j =α j +β j +γ j ;P j The total power of the lost load restored to the fault point j during emergency repair; Importance coefficient of fault load point: Defined as the ratio of uncontrollable primary and secondary loads to the total power loss caused by the fault load point in the power loss load caused by the fault load point. FPIC i ; In the formula: P sum1 The sum of primary loads among the power loss loads caused by the fault load point; P uncont 1 P is the sum of uncontrollable secondary loads in the power loss load caused by the fault point; sum This represents the total load power. Economic loss assessment of load power failure: Combining load importance and load capacity, the calculation method for assessing the economic loss of load power failure is derived as shown in the following formula (7): In the formula: E κ The economic loss due to power failure of load κ; P κ The capacity of the power-off load; v τ Let ε be the growth rate of load power loss during the τ-th power outage period, i.e., the load power loss growth rate; ε is an economic parameter; Δt τ Γ represents the duration of the τth power outage period; Γ represents the total number of power outage periods. By calculating four types of indicators for various loads—power reliability, static voltage stability, load importance, and economic loss due to load power failure—a tiered decomposition and recovery model for power loss loads in the distribution network is constructed.

3. The vehicle-pile-grid collaborative recovery method under emergency conditions considering multiple types of mobile energy storage vehicles as described in claim 1, characterized in that: The specific steps for establishing an emergency power supply restoration optimization scheduling model that takes into account three types of flexible dispatch resources—mobile energy storage vehicles, electric vehicle clusters, and diesel generators—based on the constructed distribution network power outage load tiered decomposition and recovery model include: Construct an optimized scheduling model for emergency power restoration: Establish an objective function for the reliability of load power consumption. As shown in formula (8), the optimization objective is to minimize the average short-term power outage duration and the average continuous power outage duration, and maximize the voltage qualification time of the power supply load. In the formula: the entire scheduling cycle is T, each scheduling time period is t, and B is the set of distribution network nodes; Establish an objective function for the static voltage stability of the distribution network. As shown in formula (9), the optimization objective is to optimize all lines γ of the system. SVSI ji The maximum value of the indicator should be as small as possible; Prioritize the restoration of critical power-loss loads and establish an objective function. The formula is as follows: To minimize the economic losses from load power outages, an objective function is established. The formula is as follows: Based on the objective function and After determining the priority order for restoring power outage loads, the objective function is set to minimize the mobile power dispatch cost. for: In the formula: M represents the total set of MPS; G represents the set of all mobile emergency generators (MEG); S represents the set of all mobile energy storage vehicles (MESV); V represents the set of all dispatchable EV clusters; κ represents the transportation cost coefficient of the m-th MPS; tstat m,t This represents the travel status of the m-th MPS at time t (value 1 when in a travel state, 0 otherwise); k m This represents the battery charge / discharge degradation slope of the m-th MESV or EV cluster. p represents the rated electricity price (yuan / kWh) for the m-th MESV or EV cluster; chg m,t and p dchg m,t This represents the charging and discharging power (kW) of the m-th MESV or EV cluster at time t; δ m p represents the generation cost coefficient of the m-th MEG; MEG m,t This represents the active power output (kW) of the m-th MEG at time t; Constructing optimized scheduling security constraints Setting B m Let represent the set of distribution network bus nodes that can be connected to the m-th MPS; τ represents the travel time of the mobile power source to reach the target location after receiving the dispatch instruction; under the collaborative participation of multiple types of mobile power sources, the distribution network fault recovery problem needs to consider the following constraints; Power bank connection constraints The mobile power supply can connect to at most one predetermined grid access point at any given time to provide power when needed, as shown in Equation (13); In the formula: L stat m,i,t This indicates the connection status of the mth MPS with node i at time t. The value is 1 when connected and 0 otherwise. Constraint (14) indicates that the number of mobile power sources allowed to be connected to a node is limited by the capacity of each candidate node; In the formula: M represents the number of MPS allowed to access at node i; i This represents the set of MPS that can be connected to node i; Constraint (15) states that once a mobile power source is connected to a candidate node, it cannot be moved to another node; Mobile power path constraints Constraint (16) indicates that MPS transportation between different distribution network nodes meets the required travel time; In the formula: Let represent the travel time of the m-th MPS from node i to node j. Assuming that during the recovery process, MEGs equipped with fuel tanks can be refueled to complete the dispatch, while MESVs and EVs consume electrical energy during the dispatch process; the change of MESV's SOC over time is determined by its charging and discharging behavior, as shown in Equation (17): In the formula: and Indicates the charging and discharging efficiency of the MESV or EV cluster m; SOC MESV m,t E represents the SOC of the m-th MESV at time t; MESV m This represents the battery capacity of the m-th MESV; The SOC of an EV is determined by its charging and discharging and its travel behavior, as shown in equation (18). In the formula: This represents the energy consumption (kW) of the m-th EV cluster during travel; SOC EV m,t E represents the SOC of the m-th EV cluster at time t; EV m This represents the battery capacity of the m-th EV cluster; Constraint (19) represents the SOC range of MESV and EV over all time periods; Where: SOC min m and SOC max m Indicates MESV or EV cluster SOC m,t The minimum and maximum values; Constraints (20) and (21) impose charging and discharging power constraints on MESV and EV respectively based on their respective rated power; In the formula: S chg m,t and S dchg m,t This represents the charging and discharging state of the m-th MESV or EV cluster at time t (value 1 when charging or discharging, otherwise 0); P chg,max m,t Pdchg,maxm,t represents the maximum charging and discharging power (kW) of the MESV or EV cluster; When MESV and EV are not connected to the distribution network, the charging and discharging power is forced to 0; the charging and discharging of MESV and EV are mutually exclusive in all time periods, as shown in Equation (22), which means that the MPS disconnected from the distribution network can neither charge nor discharge. Constraints (23) and (24) set the range of active and reactive power output of MEG according to its rated power, and force MEG to have zero active and reactive power output when disconnected from the distribution network. In the formula: P MEG,max m and Q MEG,max m p represents the maximum active and reactive power output (kW, kVar) of the m-th MEG; MEG m,t and q MEG m,t Let represent the active and reactive power output (kW, kVar) of the m-th MEG at time t; Radial constraints of power distribution system Constraint (25) ensures that the remaining radiation network of the power distribution system can cover all time periods; In the formula: S linkij,t This represents the connection state of branch (i,j) at time t (1 if the branch is connected, 0 otherwise); D B Indicates the total number of nodes in the distribution network; This represents the number of isolated islands formed due to branches that are damaged and not repaired at time t; Constraints (26) and (27) ensure power flow balance between load nodes and power nodes in each island; In the formula: y flow ij,t R represents the power flow on branch (i,j) at time t; fict i,t G represents the load of node i at time t; fict i,t Indicates the power supply at node i at time t; Constraint (28) is a mandatory constraint, which ensures that the power flow is 0 in the open-circuit state; Where: N m Represents a sufficiently large positive number; Branch state constraints Constraint (29) means that the damaged branch must be disconnected before it is repaired in time period t; In the formula: S dmg ij,t The 0-1 flag represents the damage status of branch (i,j) at time t (1 if the branch is not damaged or has been repaired, 0 otherwise). Constraint (30) means that branches without remote control switches and without damage remain in their initial state at all times; In the formula: A 0-1 flag representing the initial state of branch (i,j) (1 if the branch is connected, 0 otherwise); L represents a group of branches that are damaged and have not yet been repaired at time t; swch This indicates a group of branches equipped with remote control switches; L represents a set of distribution network branches. Power bank output power constraints Constraints (31) and (32) state that at a candidate node of an MPS, the active / reactive power injected or outflowed is equal to the sum of the active or reactive power outputs of all MPS. In the formula: p MPSi,t and q MPSi,t These represent the active power and reactive power output (kW, kVar) of the MPS at node i, respectively. Constraint (22) indicates that the active or reactive power from the MPS to the unconnected MPS node is 0; 4. The vehicle-pile-grid collaborative recovery method under emergency conditions considering multiple types of mobile energy storage vehicles, as described in claim 1, is characterized in that: The input includes the power and state of charge of the flexible scheduling resources, as well as four types of indicators: load power reliability, static voltage stability, load importance, and economic loss due to load power failure. The emergency power restoration optimization scheduling model is then solved to obtain the priority order for power restoration of power-loss loads and multiple types of mobile power supply scheduling schemes. The specific steps to achieve coordinated restoration of vehicle-pile-grid under emergency conditions include: objective function Then the objective function and The priority Ф for restoring power-loss loads can be written in the following compact form: stf(x)=0 (35) g(x)≤0 (36) In the formula: f(x)=0 represents the equality constraint in the model, and g(x)≤0 represents the inequality constraint in the model; Construct the minimum objective function The single-objective optimization problem is solved by calculation. The minimum value F Simultaneously, the corresponding result can be obtained. maximum value minimum value and The maximum value F Similarly, construct the objective function respectively. For a single-objective optimization problem, the objective function is obtained. and Maximum / Minimum Values and By calculating the maximum and minimum values ​​of the four objective functions, the objective function expression is normalized. In the formula: and They are respectively and The per-unit value; Using the weighted sum method with a smaller number of segments n initial Solve the normalized multi-objective model, the weight factor λ of the objective function. i The uniform step size for (i = 1, 2, 3) is determined by the number of segments n. initial Decision, that is Calculate the Euclidean distance between all adjacent solutions, and delete almost overlapping solutions according to the set precision μ. That is, when the Euclidean distance between adjacent solutions is less than μ, keep one solution and delete the rest. The selected solutions are organized, with each pair of adjacent solutions forming a region a. The number of refinement calculations required for each region is then determined. The longer the Euclidean distance between the two endpoints of region a, the more refinement calculations are required for region a. The degree of refinement for each region is determined based on the relative length of the Euclidean distance between the two endpoints of the region, as shown in equation (39). Where: n a The number of calculations required for region a; l a Let l be the Euclidean distance between the two endpoints of region a; avg is the average Euclidean distance between the two endpoints of all regions; C is a pre-defined constant. If n a If n ≤ 1, then region a does not require further detailed calculation; if n a If the value is greater than 1, proceed to the next step. Determine the offset distance between the two endpoints of region a; connect the endpoints of region a. and Create a piecewise linearized secant, and then select the Pareto front offset distance along the piecewise linearized secant direction. Add additional inequality constraints to region a, and use the weighted sum method to optimize region a. The optimization problem is shown in equations (40) to (43): stf(x)=0 (42) ·g(x)≤0 (43) In the formula: Let be the weighting factor for the feasible region of region a. i = 1, 2, 3; its uniform step size Δλ i a The number of refinement calculations n obtained from step (6) a The decision is as shown in equation (44): Delete regions that cannot converge to the optimal solution during the refinement calculation, because in this case the deleted regions are non-convex and do not contain Pareto optimal solutions; Calculate the Lagrange distance between all adjacent solutions and delete overlapping solutions; if the Lagrange distance between the two endpoints of all regions is less than μ, then end the optimization process; if there are regions where the Lagrange distance between the two endpoints is greater than μ, then jump to the step of sorting out the filtered solutions for iterative calculation. After obtaining the priority of load recovery using the adaptive weighted sum method, a genetic algorithm is used to solve the optimal economic scheduling scheme for various types of mobile power supplies, so as to achieve coordinated recovery of vehicle-pile-network under emergency conditions.

5. A vehicle-pile-grid collaborative recovery device under emergency conditions considering multiple types of mobile energy storage vehicles, characterized in that: include: The module for constructing a tiered decomposition and recovery model for power outage loads in a distribution network constructs a tiered decomposition and recovery model for power outage loads by calculating four types of indicators: power reliability, static voltage stability, load importance, and economic loss of loads when they lose power. The emergency power supply restoration optimization scheduling model construction module, based on the constructed distribution network power loss load tiered decomposition restoration model, establishes an emergency power supply restoration optimization scheduling model that takes into account three types of flexible scheduling resources: mobile energy storage vehicles, electric vehicle clusters, and diesel generator vehicles. The collaborative recovery module takes into account the power and charge status of the flexibly dispatchable resources, as well as four types of indicators: load power reliability, static voltage stability, load importance, and economic loss of load power failure. It solves the emergency power supply recovery optimization scheduling model to obtain the priority order of power supply recovery for power failure loads and multiple types of mobile power supply scheduling schemes, thereby realizing the collaborative recovery of vehicle-pile-grid under emergency conditions.

6. A vehicle-pile-grid collaborative recovery device under emergency conditions considering multiple types of mobile energy storage vehicles, as described in claim 5, is characterized in that: The power distribution network load cascade decomposition and recovery model construction module also includes: The emergency load power reliability assessment module is used for: The average duration of short-term power outages is the total average duration of short-term power outages experienced by all power-depleted loads during the entire power restoration period. (h / household), the calculation method is as follows: In the formula: N represents the duration of the power outage of the nth load in the distribution network during the j-th short-time power outage; load This indicates the total number of electrical loads in the distribution network; Average duration of continuous power outage: The total average duration of continuous power outages experienced by all electrical loads during the entire power restoration period is denoted as . (h / household), the calculation method is as follows: In the formula: This represents the duration of the power outage for the nth electrical load in the distribution network during the jth continuous power outage. Voltage qualification rate, denoted as λ, is the ratio of the duration of voltage qualification for the electrical load to the duration of power restoration during the entire power restoration period. U-qualified The calculation method is as follows: In the formula: t q n T represents the duration of voltage compliance for the nth electrical load in the distribution network during the power restoration period. rco Indicates the time required for power restoration; The static voltage stability assessment module for power distribution systems in emergency situations is used for: i is the sending node, j is the receiving node, and Z is the receiving node. ji I is the line impedance. ji S is the line current. SVSI ji For the apparent power at the receiving end, θ ji P represents the phase angle of the line. SVSI ji and Q SVSI ji Z represents active power and reactive power respectively; ji =R ji -jX ji The voltage stability index is calculated for any line in the system as follows: In the formula: γ SVSI ji γ is a voltage stability indicator; for each line in a real distribution network, γ SVSI ji It is a positive real number; if γ of all lines SVSI ji If all values ​​are less than 1, it indicates that the distribution network system is statically stable; if at least one line has a γ value less than 1, it indicates that the distribution network system is statically stable. SVSI ji If the value is equal to 1, the entire distribution network system is located on the stability boundary; if at least one line's γ value is equal to 1, then the entire distribution network system is located on the stability boundary. SVSI ji If the value is greater than 1, the voltage of the distribution network system will become unstable; The power failure load importance assessment module is used for: System Expected Power Loss: The system expected power loss is used as an indicator to assess the importance of power outage loads; SEMP refers to the amount of power outage load that can be restored after repairing a fault point in the distribution network, reflecting the degree of damage caused by the fault. The calculation method is as follows: In the formula: T j The time required to repair faulty load point j; N is the set of power loss loads caused by faulty load point j; σ SEMP j The SEMP weight value includes the impact coefficient α of failing to complete the emergency repair task within the expected timeframe on subsequent repair work. j The coefficient β set for the travel time from the power bank to the fault location. j and load weighting level coefficient γ j , i.e. σ SEMP j =α j +β j +γ j ;P j The total power of the lost load restored to the fault point j during emergency repair; Importance coefficient of fault load point: Defined as the ratio of uncontrollable primary and secondary loads to the total power loss caused by the fault load point in the power loss load caused by the fault load point. FPIC i ; In the formula: P sum 1 The sum of primary loads among the power loss loads caused by the fault load point; P uncont 1 P is the sum of uncontrollable secondary loads in the power loss load caused by the fault point; sum This represents the total load power. The load power failure economic loss assessment module is used for: Combining load importance and load capacity, a calculation method for assessing the economic loss of load power failure is derived, as shown in equation (7) below: In the formula: E κ The economic loss due to power failure of load κ; P κ The capacity of the power-off load; v τ Let ε be the growth rate of load power loss during the τ-th power outage period, i.e., the load power loss growth rate; ε is an economic parameter; Δt τ Γ represents the duration of the τth power outage period; Γ represents the total number of power outage periods. By calculating four types of indicators for various loads—power reliability, static voltage stability, load importance, and economic loss due to load power failure—a tiered decomposition and recovery model for power loss loads in the distribution network is constructed.

7. A computer-readable storage medium storing program instructions thereon, characterized in that: When the program instructions are executed by the processor, they implement the vehicle-pile-grid collaborative recovery method under emergency conditions that considers multiple types of mobile energy storage vehicles, as described in any one of claims 1 to 4.

Images