A power distribution network repair strategy updating method and system considering road network traffic state

Abstract

The application discloses a power distribution network repair strategy updating method and system considering road network traffic state, calculates dynamic traffic time of a repair team to a fault point and total disposal time of the fault point, obtains baseline recovery results in a preset rolling time domain through online recovery optimization model output of a power distribution network, determines a priority repair point from the fault point according to the baseline recovery results, the dynamic traffic time and the total disposal time, and in the process of executing a repair task corresponding to the priority repair point in the preset rolling time domain, re-determines the priority repair point which has not executed the repair according to the latest post-disaster road edge state until post-disaster repair and recovery are completed. The application improves the speed of post-disaster repair scheduling, improves the post-disaster recovery capability of the power distribution network, and meets the actual demand of fast recovery of power supply.

Description

Technical Field

[0001] This invention relates to the field of power system dispatching technology, and in particular to a method and system for updating distribution network emergency repair strategies that takes into account the road network traffic status. Background Technology

[0002] With global climate change, the intensity and frequency of extreme events are gradually increasing, easily triggering large-scale power outages and slow recovery in power distribution networks, necessitating post-disaster emergency repairs. During post-disaster repairs, the dispatch sequence of maintenance teams, fault repair time, and network reconstruction and recovery strategies are interdependent, jointly affecting power restoration efficiency and the duration of power outages for users. As the scale of power distribution networks expands and their structure becomes increasingly complex, relying solely on experience for post-disaster repair scheduling is no longer sufficient to meet the actual needs for rapid power restoration.

[0003] In existing technologies, the problem of post-disaster power distribution network restoration is usually addressed by two main approaches: one is to first generate a repair sequence based on the fault location, load size, or equipment importance, and then organize maintenance teams to carry out repair tasks in that sequence; the other is to establish a power distribution network restoration optimization model with the goal of network reconstruction and load restoration, and solve for fault isolation, tie switch switching, and load restoration schemes.

[0004] However, while the above process can improve the speed of post-disaster emergency repair dispatch, the total handling time and travel time are usually simplified to fixed constants, leading to discrepancies between work hour estimates and the actual repair process, affecting the accuracy of dispatch decisions. Furthermore, in actual repair operations, the ability to update decisions promptly after events such as task completion, new faults occurring, or changes in roadside conditions makes it difficult to achieve dynamic coordination between emergency repair dispatch and the power grid restoration process. This results in low post-disaster recovery capabilities of the power grid and low power restoration efficiency, failing to meet the actual demand for rapid power restoration. Summary of the Invention

[0005] The technical problem to be solved by this invention is: how to improve the post-disaster recovery capability of the power distribution network while increasing the speed of post-disaster emergency repair and dispatch, so as to meet the actual needs of rapid power restoration.

[0006] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows: A method for updating emergency repair strategies for a power distribution network that takes into account the road network traffic status includes the following steps: Construct a time-varying traffic network containing fault points based on the post-disaster roadside conditions, and calculate the dynamic travel time of the repair team to the fault points. Obtain the fault baseline working time, and combine it with the dynamic passage time, the fault baseline working time, real-time weather conditions, real-time terrain conditions, and real-time material status to calculate the total handling time of the fault point in real time. A power distribution network online recovery optimization model is constructed, and the baseline recovery result within a preset rolling time domain is obtained by combining the post-disaster roadside conditions through the output of the power distribution network online recovery optimization model. Based on the baseline recovery results, the dynamic passage time, and the total handling time, priority repair points are determined from the fault points; During the execution of emergency repair tasks corresponding to the priority repair points within the preset rolling time domain, the priority repair points that have not yet been repaired are re-determined based on the latest post-disaster roadside status, until the post-disaster emergency repair and restoration are completed.

[0007] To solve the above-mentioned technical problems, another technical solution adopted by the present invention is as follows: A power distribution network emergency repair strategy update system that considers road network traffic conditions includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the aforementioned power distribution network emergency repair strategy update method that considers road network traffic conditions.

[0008] The beneficial effects of this invention are as follows: It provides a method and system for updating a power distribution network emergency repair strategy that considers road network traffic conditions. Based on the post-disaster roadside conditions, a time-varying traffic network containing fault points is constructed. The dynamic travel time from the repair team to the fault point is calculated. The time-varying traffic network accurately reflects the impact of post-disaster road traffic conditions on the arrival time of the repair team, determines the dynamic travel time, and improves the accuracy of emergency repair travel time estimation. By combining the fault baseline working time with weather conditions, terrain conditions, and material conditions, the fault repair time is dynamically estimated, improving the realism and rationality of post-disaster repair process modeling. An online power distribution network recovery optimization model is constructed. Combined with the post-disaster roadside conditions, the output of the online power distribution network recovery optimization model yields the results within a preset rolling timeframe. The baseline restoration results within the time domain; based on the baseline restoration results, dynamic passage time, and total handling time, priority repair points are determined from the fault points; thus avoiding the limitations of pre-determining a fixed repair sequence, improving the method's adaptability to changes in the dynamic post-disaster environment, and combining the selection of repair targets with online restoration optimization of the distribution network, during the execution of repair tasks for corresponding priority repair points within the preset rolling time domain, the latest post-disaster roadside status is obtained and used to re-determine priority repair points that have not yet been repaired, until post-disaster repair and restoration are completed, enabling repair decisions at each moment to be adjusted based on the current system status, improving the speed of post-disaster repair scheduling while enhancing the post-disaster recovery capability of the distribution network, and meeting the actual needs of rapid power restoration. Attached Figure Description

[0009] Figure 1 This is a schematic diagram illustrating the steps of a distribution network emergency repair strategy update method considering road network traffic conditions according to an embodiment of the present invention. Figure 2 This is a statistical diagram of time-varying roadside status for different schemes in a power distribution network emergency repair strategy update method that considers road network traffic conditions, according to an embodiment of the present invention. Figure 3 This is a repair timing diagram in a power distribution network repair strategy update method that considers road network traffic status according to an embodiment of the present invention. Figure 4 This is a statistical chart of load loss under different schemes in a power distribution network emergency repair strategy update method that considers road network traffic conditions, according to an embodiment of the present invention. Figure 5 This is a system block diagram of a power distribution network emergency repair strategy update system that considers the road network traffic status, according to an embodiment of the present invention. Detailed Implementation

[0010] To explain in detail the technical content, objectives, and effects of the present invention, the following description is provided in conjunction with the embodiments and accompanying drawings.

[0011] Existing post-disaster emergency repair strategies for power distribution networks have at least several shortcomings: First, repair time is usually simplified to a fixed constant, failing to fully consider the impact of different fault types, weather conditions, terrain conditions, and material status on repair time, leading to discrepancies between estimated work hours and the actual repair process. Second, the travel time for repair teams to reach the fault point is usually estimated using static distance or a fixed congestion coefficient, failing to effectively describe the dynamic traffic conditions such as road closures, traffic restrictions, and gradual restoration after a disaster, thus affecting the accuracy of dispatch decisions. Third, existing methods mostly generate the repair sequence all at once at the initial moment, lacking the ability to update decisions in a timely manner after events such as task completion, new faults taking effect, or changes in roadside conditions occur, making it difficult to achieve dynamic coordination between emergency repair dispatch and the power distribution network restoration process.

[0012] Therefore, there is an urgent need to propose a post-disaster power distribution network emergency repair and restoration method that can comprehensively consider the dynamic changes in roadside conditions, the dynamic changes in fault repair time, and the power distribution network restoration and reconstruction process, and can be updated in a rolling manner when triggered by events, so as to improve power supply restoration efficiency and enhance post-disaster recovery capabilities.

[0013] To address at least the aforementioned problems, this invention calculates the dynamic travel time from the maintenance team to the fault point and the total handling time of the fault point. It then outputs the baseline restoration results within a preset rolling time domain through an online power distribution network recovery optimization model. Based on the baseline restoration results, dynamic travel time, and total handling time, priority repair points are determined from the fault points. During the execution of repair tasks for the corresponding priority repair points within the preset rolling time domain, the latest post-disaster roadside conditions are acquired and used to re-determine priority repair points that have not yet been repaired, until post-disaster repair and restoration are completed. This invention improves the speed of post-disaster repair scheduling while enhancing the post-disaster recovery capability of the power distribution network, meeting the practical needs for rapid power restoration.

[0014] Before detailing the embodiments of this application, some related concepts will first be explained: Shortest path algorithm: Given the network structure and road edge weights, a method to search for the path with the minimum total travel cost between the current position and the corresponding position of the target fault point.

[0015] Single-item flow method: In distribution network optimization, this method simulates unidirectional flow from source nodes to load nodes by introducing virtual flow variables and flow conservation constraints. This method can rigorously describe the radial structure of the distribution network (connected and acyclic) in the form of linear constraints and is commonly used in network reconfiguration and fault recovery models.

[0016] Please refer to Figure 1 and Figure 2 This invention discloses a method 100 for updating a power distribution network emergency repair strategy that considers the road network traffic status. The method 100 includes the following steps: In step S102, a time-varying traffic network containing fault points is constructed based on the post-disaster roadside conditions, and the dynamic travel time from the repair team to the fault points is calculated.

[0017] In this invention, the travel time of the repair team to the fault point is directly affected by the dynamic changes in the post-disaster road traffic conditions. To accurately reflect the travel costs required for the repair team to reach the fault point at different times, this invention classifies roadside states into three categories: closed, restricted, and open. A closed state indicates that the corresponding roadside is not passable at the current time; a restricted state indicates that the corresponding roadside is passable at the current time, but the traffic efficiency is lower than normal; and an open state indicates that the corresponding roadside has restored normal traffic capacity at the current time. Therefore, let the set of roadsides in the time-varying traffic network be... For any roadside Record its absolute moment The roadside condition is Then the roadside state satisfies: .

[0018] In the formula, Indicates the side of the road At any moment The state.

[0019] The evolution of the roadside state over time can be represented as: .

[0020] In the formula, This indicates the moment when road edge e changes from a closed state to a restricted state. Let e ​​represent the moment when road edge e changes from a restricted state to an open state, and satisfy the following condition: .

[0021] The above expression indicates that the roadside can change from a closed state to a restricted state over time, and then from a restricted state to an open state.

[0022] Furthermore, for some roadsides that were under traffic restrictions immediately after a disaster, the following can be ordered: .

[0023] At this point, road edge e can enter the restricted passage state at the initial moment, and in subsequent moments... It was later switched to an open state.

[0024] In the program implementation, basic traffic attributes, state change times, and traffic restriction amplification coefficients can be preset for each roadside, and the roadside state can be determined at the current time accordingly.

[0025] After classifying the roadside states, in order to further map the closed, restricted, and open states of the roadsides to the passage costs in the maintenance team's path calculation, it is necessary to determine the roadside weights based on the current roadside states. These roadside weights characterize the time cost required for the maintenance team to traverse the corresponding roadside and serve as the basis for subsequently constructing the time-varying traffic network and solving for the shortest path.

[0026] The roadside weights are determined based on the roadside conditions, and the expression is as follows: .

[0027] in, Indicates the weight of the roadside. This indicates the basic travel time at roadside e. Let e ​​represent the amplification factor of travel time at roadside e under traffic restriction conditions, and it usually satisfies: .

[0028] When a roadside is closed, its passage weight is positive infinity, indicating that the roadside is not passable at the current moment; when a roadside is restricted, its passage weight is the product of the basic passage time and the restriction amplification factor; when a roadside is open, its passage weight is equal to the basic passage time.

[0029] Furthermore, when roadside e satisfies: .

[0030] Since its weight is positive infinity, when constructing the traffic network at the current moment, this road edge can be considered an unusable edge and not included in the edge set construction of the current traffic graph; when road edge e satisfies one of the following: .

[0031] or .

[0032] Then, it is added to the time-varying traffic network with the corresponding weight. Through this method, both the edge set and edge weight in the time-varying traffic network can change with time. It updates dynamically according to changes.

[0033] Furthermore, in the program implementation of this invention, the determination of roadside weights is performed in real time based on the current state of the roadside. For each roadside, the weights are first determined according to the current time... Determine its status Then, its passage weight is determined according to the above weight calculation rules. .

[0034] In this invention, the travel time from the maintenance team to the fault point is not fixed in advance, but is calculated in real time based on the time-varying traffic network constructed at the current moment.

[0035] Set time The time-varying traffic network is represented as follows: .

[0036] A time-varying traffic network is constructed based on the current roadside state. The dynamic travel time from the repair team to the fault point is calculated using the shortest path method, as shown in the following expression: .

[0037] in, ; In the formula, This indicates that the maintenance team is always Location, This indicates the location corresponding to fault point i. Indicates time-varying transportation networks Up by arrive The set of feasible paths Representing a path The set of road edges included.

[0038] Furthermore, the set of road edges contained in path π is denoted as... The total travel time for this path can then be expressed as: .

[0039] In the formula, Indicates path π at time 10:00 The corresponding total travel time, We Indicates the time e at roadside The passage weights are then used. Therefore, the dynamic passage time from the maintenance team to the fault location i can be further expressed as: .

[0040] In the formula, ) indicates at time Below, from node to node n i The set of all feasible paths.

[0041] The above expression means that the dynamic travel time from the maintenance team to the fault point is the shortest path travel time from the current location node to the corresponding fault node in the traffic map at the current moment.

[0042] Furthermore, in the program implementation of this invention, the dynamic travel time from the repair team to each fault point can be calculated based on the time-varying traffic network constructed at the current moment using the shortest path search algorithm; when the location of the repair team or the state of the roadside changes, the traffic map is reconstructed and the dynamic travel time is updated.

[0043] In step S104, the fault baseline working time is obtained, and the total handling time of the fault point is calculated in real time by combining the dynamic passage time, fault baseline working time, real-time weather conditions, real-time terrain conditions and real-time material status.

[0044] In this invention, the total handling time for the fault to be repaired is not a fixed constant, but is determined by the fault baseline working time corresponding to the fault type, combined with weather conditions, terrain conditions, and material status, in order to reflect the dynamic impact of the post-disaster environment on maintenance efficiency, including the following process: First, set the corresponding fault baseline working hours according to the fault type of the fault point.

[0045] Let the current set of faults to be repaired be... For any fault point Record its fault type as Then, the corresponding fault baseline working hours can be preset according to the fault type, denoted as: .

[0046] In the formula, The baseline working time represents the time required to resolve the fault at point i. Indicates the fault type at fault point i; This represents the mapping relationship from fault type to fault baseline working time.

[0047] Secondly, weather influence factors, terrain influence factors, and material influence factors are established based on real-time weather conditions, real-time terrain conditions, and real-time material status. Considering the impact of post-disaster external weather conditions on maintenance efficiency, a weather influence factor is also established. The weather impact factor is used to characterize the effect of the weather conditions faced by fault point i at the start of maintenance on the baseline working hours. Let the start time of maintenance for fault point i be... Then the weather influencing factor can be expressed as: .

[0048] In the formula, The weather influence factor represents the fault point i; This indicates the time when the maintenance team arrives at the location corresponding to fault point i; This represents a weather-related function.

[0049] Furthermore, in one embodiment of the present invention, the weather impact function can be designed to gradually decrease over time after a disaster or change with the gradual improvement of environmental conditions. That is, in the early post-disaster period, the weather impact factor is relatively large; as time goes on, the weather impact factor gradually decreases and tends to stabilize. Typically, it can be set as follows: .

[0050] The above inequality indicates that the impact of weather conditions on repair time is mainly reflected in the amplification or maintenance of the baseline working time, and will not cause the repair time to be lower than the baseline working time for this type of fault.

[0051] Furthermore, considering the impact of terrain conditions in the area where the fault point is located on maintenance efficiency, a terrain influence factor is established. The terrain influence factor is used to characterize the effect of differences in road complexity, terrain conditions, or on-site working environment at the fault location on repair time. Let n be the location identifier corresponding to fault point i. i Then the topographic influence factor can be expressed as: .

[0052] In the formula, The terrain influence factor represents the fault point i; n i Indicates the location corresponding to fault point i; This represents the mapping relationship from the fault location to the terrain influence factor.

[0053] Furthermore, different locations may have different terrain conditions, and therefore different terrain influence factors. Typically, we can let: .

[0054] Furthermore, considering the impact of material reserves and supply on repair efficiency during emergency repairs, a material impact factor is established. Let the current material state variable be... Then the material impact factor at fault point i can be expressed as: .

[0055] In the formula, The material impact factor represents the failure point i. This indicates the status of the materials at the moment the maintenance team arrives at the fault point i. This represents the mapping relationship between the state of materials and their influencing factors.

[0056] Furthermore, when material reserves are sufficient, the material impact factor can be taken to be close to... The value of the material impact factor increases accordingly when there is a shortage of materials or critical materials. Commonly included are: .

[0057] After obtaining the baseline working hours for the fault type, as well as the weather, terrain, and material impact factors, the dynamic maintenance time for fault point i can be further obtained.

[0058] Next, based on the baseline working time of the fault, weather impact factors, terrain impact factors, and material impact factors, the dynamic maintenance time of the fault repair point is calculated. Its expression is as follows: .

[0059] The above expression indicates that the dynamic repair time for a fault is obtained by adjusting the baseline working time for the fault type based on weather, terrain, and material factors, and is dynamically updated as the current scenario changes.

[0060] The total processing time is obtained by combining dynamic maintenance time and dynamic traffic time. Its expression is as follows: .

[0061] In step S106, an online restoration optimization model for the power distribution network is constructed. Based on the post-disaster roadside conditions, the baseline restoration results within a preset rolling time domain are obtained through the output of the online restoration optimization model for the power distribution network.

[0062] In this invention, the distribution network online recovery optimization model is used to characterize the system's recovery effect in the subsequent rolling time domain based on the current known fault set, power outage event information, and line availability status. To reflect the overall level of power supply recovery in the rolling time domain, an objective function is established with the goal of maximizing the cumulative restored load in the rolling time domain.

[0063] Furthermore, suppose the rolling recovery time domain includes There are discrete time periods, denoted by h, where The duration of each time period is Let the set of load nodes be... For any load node Record its active load as , record it in the first The load recovery state variables for each time period are: Specifically, when load node i restores power in the h-th time period, When load node i fails to restore power during the h-th time period, Then the first The recovery load for each time period can be expressed as: .

[0064] In the formula, This represents the total active load that the system has recovered during the h-th time period.

[0065] Based on this, the objective function established in this invention can be expressed as: .

[0066] In the formula, M represents the recovery index in the rolling time domain; H represents the total number of discrete time periods included in the rolling recovery time domain; Indicates the duration of a single discrete time period; This represents the total active load that the system has recovered during the h-th time period.

[0067] Furthermore, Substituting into the above objective function, we get: .

[0068] The above expression indicates that the objective function measures the cumulative recovery effect of the system within the entire rolling window by accumulating the recovery load of each time period within the rolling time domain.

[0069] In subsequent evaluations of candidate fault hypothesis repair scenarios, the present invention uses the aforementioned recovery index M as the evaluation standard.

[0070] In this invention, the restoration network corresponding to the online restoration optimization model of the distribution network should meet the radial operation requirements, that is, the energized network should remain connected and not form loops after restoration. To this end, this invention uses the single-commodity flow method to characterize the connectivity and radiality of the restoration network, so as to ensure that the restoration scheme obtained by the online restoration optimization model meets the topology operation constraints of the distribution network.

[0071] Furthermore, let the set of distribution network nodes be... The route set is For any line Let its operational state variable in the h-th time period be denoted as . Specifically, when line e is put into operation in the h-th time period, =1; when line e is disconnected in the h-th time period. =0. Simultaneously, record the node. The energized state variable in the h-th time period is Specifically, when node i is in the energized state during the h-th time period, =1; when node i is not powered on in the h-th time period. =0.

[0072] To characterize the connectivity of the restored network, this invention uses a set of lines... Virtual flow variables are introduced onto the induced set of directed arcs. Furthermore, let the set of lines... The set of directed arcs obtained by expansion is For any directed arc Introduce virtual flow variables for the h-th time period. This is used to characterize the virtual tracer flow from node i to node j in the recovery network. The virtual flow does not correspond to the actual physical power flow, but is used to describe the connectivity between each energized node and the power supply root node in the recovery network.

[0073] Furthermore, in one embodiment of the present invention, a power supply root node may be selected. As a virtual source node. For any non-root node The following virtual flow conservation constraints are established: .

[0074] In the formula, This represents the virtual traffic flowing from node j to node i in the h-th time period. Let represent the virtual flow from node i to node k in the h-th time period. The above constraint means that for each non-root node that is powered on, its net inflow virtual flow should be 1; for nodes that are not powered on, since... It does not need to obtain virtual connected flow from the root node.

[0075] Furthermore, for the root node r, the following virtual flow balance constraint is established: .

[0076] The above constraint means that the root node provides all other charged nodes with a virtual flow that matches the total number of their charged states in the h-th time period, thereby ensuring that all charged nodes can be reached by the root node through the recovery network.

[0077] Furthermore, in addition to the aforementioned virtual flow conservation constraints, to ensure that the restored network satisfies the radial structure, a correspondence constraint between the number of lines and the number of charged nodes is also established: .

[0078] The above constraint means that, combined with the aforementioned virtual flow conservation constraint, it can ensure that the restored network is both connected and loop-free, thereby satisfying the radial topology requirement.

[0079] The radial topology constraints based on the single-commodity flow method are established separately for each discrete time period in the rolling time domain to ensure that the recovered topology in each time period meets the requirements of the radial structure.

[0080] In this invention, after establishing the objective function and the radial topology constraints based on the single commodity flow method, it is also necessary to further establish line availability constraints, load recovery constraints, node power balance constraints, and voltage and power flow operation constraints to ensure that the recovery scheme obtained by the online recovery optimization model not only meets the topology requirements, but also meets the fault state constraints and distribution network safe operation constraints.

[0081] Furthermore, let the set of load nodes be... For any line Let its operational status in the h-th time period be denoted as . The line is available. Specifically, when line e is ready for operation in the h-th time period, When line e is unavailable for operation in the h-th time period due to an unrepaired fault, The available state constraints of the line can then be expressed as: .

[0082] The above constraint means that the line is only allowed to be put into operation when it is in an available state, so that the progress of fault repair is explicitly fed back into the online recovery optimization model.

[0083] Furthermore, for any load node Let its load recovery state variable in the h-th time period be denoted as . The node's energized state variable is The load recovery constraint can then be expressed as: .

[0084] The above constraint means that the load on node i is allowed to be restored only if node i is energized in the h-th time period; if the node is not energized, the corresponding load cannot be mistakenly judged as restored. This ensures that the calculation of restored load is consistent with the node's energized state.

[0085] Furthermore, to describe the active and reactive power balance relationship of nodes during the recovery process, let... Let i represent the set of routes that terminate at node i. Let represent the set of lines starting from node i. For any line e, let its active power in the h-th time period be denoted as . reactive power is The square of the current is The line resistance is The line reactance is For any node Let the active power injection of the distributed power source in the h-th time period be denoted as . Reactive power injection is The active load of the node is Reactive load is The active power balance constraint at the nodes can then be expressed as: .

[0086] The corresponding reactive power balance constraint can be expressed as: .

[0087] The aforementioned active and reactive power balance constraints are used to ensure that the power conservation relationship holds true at each node in each time period. Furthermore, to ensure that the distribution network operates within a safe range during the restoration process, voltage and power flow constraints also need to be established. Let the square of the voltage at node i in the h-th time period be... The upper and lower limits of the voltage are respectively and Then the node voltage constraint can be expressed as: .

[0088] The above constraint means that when node i is energized in the h-th time period, its voltage must be within the allowable range.

[0089] Furthermore, the route was designed. Connecting node i and node j, the voltage drop relationship of the line can be expressed as: .

[0090] Meanwhile, to ensure consistency between line power flow variables and current and voltage variables, the following second-order cone constraint is established: .

[0091] The above constraints indicate that the active power, reactive power, current square, and transmitting voltage square of the line satisfy the power flow operation relationship under the distribution network model.

[0092] After establishing the objective function and various topology and operational constraints, they are combined to form the distribution network online recovery optimization model corresponding to the current moment, and the baseline recovery result under the current state is obtained based on the model.

[0093] The known fault set, power outage event information, and line availability status are obtained from the post-disaster roadside conditions. These are then used as input to the online distribution network recovery optimization model to obtain the baseline recovery results, as detailed below: Let the current absolute time be Given the current known fault set, power outage event information, and line availability status, solving the online recovery optimization model yields the current time value. The optimal recovery result is denoted as ,Right now: .

[0094] In the formula, Indicates the current time The baseline recovery index is defined below; H represents the total number of discrete time periods included in the rolling recovery time domain. Indicates the duration of a single discrete time period; This represents the total active load that the system has recovered during the h-th time period.

[0095] In step S108, based on the baseline recovery results, dynamic passage time, and total handling time, priority repair points are determined from the fault points, including the following process: First, based on the total handling time and the condition of the roadside after the disaster, a hypothetical repair scenario is constructed.

[0096] Corresponding to the current state, without additional priority assumptions about repairing a candidate fault, the optimal recovery level that the system can achieve through network reconstruction and load recovery is, and therefore can be used as a benchmark characterization of the system's recovery capability at the current moment.

[0097] During the priority repair point assessment process, for any candidate fault point i, a hypothetical priority repair scenario is constructed, and after resolving the online recovery optimization model, the corresponding expected repair result can be obtained. By saying and By comparing these factors, we can further obtain the incremental contribution of candidate fault point i to the system recovery effect.

[0098] In this invention, not all known faults requiring repair need to have hypothetical repair scenarios constructed and recovery effects evaluated at any given moment. To reduce the computational overhead of solving the online recovery optimization model for each fault individually, it is preferable to first form a current set of candidate faults based on fault reachability and time cost, and then evaluate the candidate faults for subsequent scenarios.

[0099] Furthermore, based on the dynamic travel time, the accessibility of the fault point is determined; based on the accessibility and total handling time, a candidate fault set is formed from the fault points. Since a fault only has practical significance for immediate dispatch and execution if the maintenance team can reach the corresponding fault point at the current moment, it is preferable to first screen the known faults to be repaired based on the dynamic travel time to form a candidate fault set. It can be represented as: .

[0100] In the formula, This represents the set of candidate faults at the current moment. The above expression means that a fault is considered reachable at the current moment only if the dynamic passage time corresponding to fault point i is a finite value, thus entering the subsequent candidate fault screening process.

[0101] Furthermore, based on the aforementioned candidate fault set, the faults can be sorted in ascending order of total fault handling time, and the top K faults can be selected to form the current candidate fault set, where K is a preset parameter for the number of candidate faults. If the number of currently evaluable faults is less than or equal to K, then we can directly set: .

[0102] If the number of currently assessable faults is greater than K, then from The K faults with the shortest total processing time are selected to form the core. Therefore, the current candidate fault set can be understood as a group of faults that, at the current moment, both meet the arrival conditions and have a high priority evaluation value in terms of time cost.

[0103] Furthermore, the formation process of the aforementioned candidate fault set does not directly determine the final repair target, but is used to narrow down the scope of subsequent hypothetical repair scenario evaluation.

[0104] Furthermore, the following conditions are met at the current moment: .

[0105] This indicates that there are currently no candidate faults available for subsequent evaluation. In this case, it is preferable not to generate a new repair task, but to maintain the current state and wait for the next event to trigger a re-evaluation. The event includes at least one of the following: repair task completion event, roadside state change event, and new fault activation event.

[0106] After obtaining the baseline recovery index, a hypothetical repair scenario is constructed for each candidate fault in the current candidate fault set to characterize the impact of prioritizing the repair of the candidate fault at the current moment on the subsequent rolling recovery process.

[0107] Based on the total handling time, determine the estimated completion time of the corresponding pending fault points; update the availability status of the lines corresponding to the pending fault points based on the estimated completion time, and construct the corresponding hypothetical repair scenarios.

[0108] For any candidate fault Its total processing time at the current moment is: .

[0109] In the formula, This indicates that the maintenance team is always The dynamic travel time required to reach candidate fault point i. This represents the dynamic repair time for candidate fault point i. Indicates that candidate fault point i is at time i The total processing time.

[0110] Therefore, the expected repair completion time corresponding to candidate fault point i can be obtained. ,Right now: , In the formula, Indicates if at the current moment If the maintenance team is given priority to handle candidate fault point i, then the estimated completion time for the repair of that candidate fault is as follows.

[0111] Furthermore, when constructing the hypothetical repair scenario for candidate fault point i, the repair completion time of the line corresponding to the candidate fault is updated to... The availability status of the corresponding line in the rolling recovery time domain is updated accordingly. That is, under the hypothetical repair scenario, the line corresponding to candidate fault point i remains unavailable before the expected repair is completed, and becomes available after the expected repair is completed.

[0112] Secondly, after constructing the hypothetical repair scenarios for each candidate fault, the aforementioned online recovery optimization model is called to solve each hypothetical repair scenario to obtain the expected repair results when the corresponding candidate fault is repaired first.

[0113] For any candidate fault By solving the online recovery optimization model under its assumed repair scenario, the corresponding scenario recovery index can be obtained. ,Right now: .

[0114] In the formula, Indicates the current time of candidate fault point i. The recovery metrics corresponding to the hypothetical repair scenario; This represents the total active load that the system has restored in the h-th time period under the hypothetical repair scenario at candidate fault point i.

[0115] Therefore, for the current set of candidate faults From all candidate faults in the dataset, we can obtain the corresponding set of scenario recovery metrics: .

[0116] In the formula, Indicates the current time The set of scenario recovery indicators corresponding to all candidate faults.

[0117] Then, based on the difference between the expected repair results and the baseline recovery results, the candidate fault comprehensive score of the corresponding fault point is calculated. Based on the candidate fault comprehensive score, dynamic passage time and total handling time, the priority repair point is determined from the fault points.

[0118] Obtaining baseline recovery indicators and the scenario recovery metrics corresponding to each candidate fault. Then, the incremental recovery metrics resulting from prioritizing the repair of each candidate fault can be calculated. For any candidate fault... Its recovery indicator increment It can be represented as: .

[0119] In the formula, Indicates the current time of candidate fault point i. The increase in recovery indicators.

[0120] The above expression indicates that the recovery index increment is used to characterize the additional recovery benefits brought about by prioritizing the repair of candidate faults in the current state relative to the baseline recovery results.

[0121] Furthermore, to comprehensively consider the recovery benefits of prioritizing the repair of candidate faults and the time cost required to complete the repair, this invention constructs a comprehensive score for candidate faults based on the increment of the recovery index and the total handling time. For any candidate fault... Its candidate fault comprehensive score It can be represented as: .

[0122] In the formula, Indicates the current time of candidate fault point i. The comprehensive score of candidate faults; This indicates a preset positive number, used to avoid the denominator being zero.

[0123] The above expression indicates that the comprehensive score reflects the level of recovery benefits that can be achieved per unit of total treatment time.

[0124] After obtaining the comprehensive score for each candidate fault, the current repair target can be determined based on the comprehensive score. Further, let the current time... The following is a comprehensive score set for candidate faults: .

[0125] The current repair target It can be represented as: .

[0126] In the formula, Indicates the current time The determined emergency repair targets are then identified. The above expression means: from the current set of candidate faults, the candidate fault with the highest comprehensive score is selected as the fault to be prioritized for the repair team to handle.

[0127] Furthermore, if the scenario recovery result corresponding to the candidate fault cannot be effectively obtained in certain situations, the principle of minimizing the total processing time can be used as a supplementary selection criterion.

[0128] In step S110, during the process of performing emergency repair tasks for the corresponding priority repair points within the preset rolling time domain, the priority repair points that have not yet been repaired are re-determined based on the latest post-disaster roadside status, until the post-disaster emergency repair and restoration are completed.

[0129] In this invention, the post-disaster emergency repair and recovery process adopts an event-driven rolling update approach, rather than generating a fixed repair sequence all at once at the initial moment. During the execution of the repair task, when the repair task is completed, the roadside status changes, or a new fault takes effect, the current fault information, power outage event information, line availability status, and traffic network status are updated, and the subsequent decision-making process is repeated in a rolling manner.

[0130] In this invention, the post-disaster repair and recovery process adopts an event-driven rolling update method. The events mainly include repair task completion events, roadside status change events, and new fault activation events. After any one of these three events occurs, the post-disaster roadside status is updated. Specifically, the repair task completion event refers to the current repair target being repaired at the expected completion time, the corresponding faulty line changing from an unavailable state to an available state, and the repair team's current location being updated to the location of the fault point.

[0131] If the current repair target is a fault The estimated completion time for the repair is Then when .

[0132] The completion of the emergency repair task is determined by the occurrence of an event. A roadside status change event refers to an event in which a roadside changes from a closed state to a restricted state, or from a restricted state to an open state, during operation, resulting in a change in the current traffic network structure or roadside rights; if the roadside... The state change times are respectively and Then, if one of the following conditions is met: ; or .

[0133] The occurrence of a roadside state change event is determined. A new fault activation event refers to an event in which a future fault, after reaching its predetermined fault time, is formally added to the current active fault set and affects power outage event information and line availability; if the predetermined activation time of future fault point j is... Then: .

[0134] A new fault event has been determined to have occurred.

[0135] Furthermore, after any event occurs, it is necessary to update the current fault information, power outage event information, line availability status, and traffic network status, and re-execute the online recovery optimization and subsequent fault assessment process based on the updated system status.

[0136] Furthermore, when a repair task has already been issued and is in the execution phase, it is preferable not to preempt the ongoing repair task. That is, if the repair team has already gone to a fault point and is in the process of execution, changes in the roadside status or the occurrence of a new fault will only trigger system status updates and recalculation of the recovery plan, without directly changing the currently executing repair target.

[0137] Furthermore, to provide a unified description of the triggering times of various events, they can be categorized into a set of event moments. Let the current time be... Then the nearest event trigger time after the current time can be represented as .

[0138] In the formula, This represents the most recent event moment after the current moment. The system advances along the event time axis and, at each event point, reconstructs the traffic network based on the updated system state, updates the availability of routes, invokes the online recovery optimization model, calculates baseline recovery indicators, evaluates candidate fault scenarios, and determines subsequent repair targets, thus forming a rolling closed-loop repair and recovery process of "event triggering - state update - recovery solution - candidate evaluation - continued execution".

[0139] After any of the aforementioned events occur, the current system state must be updated first to ensure that subsequent recovery solutions and emergency repair decisions are based on the latest state.

[0140] The system state after the event update can be abstractly represented as: .

[0141] In the formula, Indicates time The system status under the following conditions; Represents the current active fault set; Indicates information about a power outage event; Represents the set of available line states; Represents a time-varying transportation network at the current moment; This indicates the current location of the maintenance team. By uniformly updating the above status variables, it can be ensured that the subsequent rolling recovery and emergency repair assessment processes use consistent and up-to-date system information.

[0142] After completing the system status update, based on the updated... The subsequent assessment and decision-making process will be re-executed. Specifically, this will begin with the updated time-varying traffic network. The dynamic travel time from the maintenance team to each fault location is recalculated, and the total handling time for each fault is obtained by combining this with the aforementioned dynamic maintenance time model. Next, the aforementioned online recovery optimization model is invoked to obtain the baseline recovery index corresponding to the current moment. Then, based on the currently reachable faults, a set of candidate faults is formed. For each candidate fault, a hypothetical repair scenario is constructed, and the corresponding scenario recovery index is obtained. Finally, a comprehensive score is constructed based on the increase in recovery indicators and the total handling time, and the next repair targets at the current moment are determined.

[0143] Furthermore, if the updated set of candidate faults at the current moment is denoted as... After reassessment, the subsequent repair targets can still be determined by the following formula:

[0144] And it satisfies: .

[0145] In the formula, Indicates the current time The newly identified repair targets. Indicates that candidate fault point i is at time i The overall score is as follows.

[0146] Based on the aforementioned event triggering, state updating, recovery solving, and emergency repair assessment, this invention forms an event-driven rolling closed-loop emergency repair and recovery mechanism. This mechanism does not determine the complete repair sequence all at once at the initial moment, but rather progresses step-by-step along the event timeline, and after each event trigger, it re-determines the subsequent repair targets based on the updated system state.

[0147] During the execution of emergency repair tasks at the corresponding priority repair points, the system monitors emergency repair task completion events, roadside status change events, and new fault activation events. After the emergency repair task completion event is completed, the fault points that have been repaired by the emergency repair task completion event are removed from the fault set, and the post-disaster roadside status is updated. If the fault set is empty and no new fault activation event occurs, the emergency repair and recovery process ends.

[0148] To verify the effectiveness of the proposed rolling update method for post-disaster emergency repair strategies of distribution networks that considers road network traffic conditions, an IEEE 33-node distribution network was selected as the test system, and a 5-node distribution network coupled with its geographical location was constructed. A 14-grid traffic network is used to describe the road accessibility, traffic status evolution, and travel time to the fault point for the repair team in a post-disaster scenario. In this embodiment, the initial fault route numbers are 3, 7, 28, 6, 36, 21, and 31. Furthermore, to simulate the continuous emergence of new faults during post-disaster recovery, subsequent sudden fault routes 18 and 24 are also set, with corresponding effective times of 2 hours and 4 hours, respectively. Through these settings, a post-disaster recovery scenario where initial concentrated faults and subsequent newly added faults coexist can be reflected in the same calculation.

[0149] To demonstrate the impact of each key technical element on the recovery outcome, the following three comparison schemes were set up: Option 1 is a rolling emergency repair scheme with no recovery-emergency repair coupling benefit assessment. This scheme reorders faults based on static fault priority and prioritizes the fault with the smaller sum of passage time and repair time among the currently available faults. Option 2 is a rolling emergency repair scheme without time-varying road networks and dynamic maintenance time. That is, under the rolling framework, the evolution of roadside conditions over time is not considered, nor is the maintenance time affected by changes in environmental conditions. Scheme 3 is the scheme of this invention, which comprehensively considers the time-varying status of the roadside after the disaster, the dynamic changes in the repair time, and the evaluation of the benefits of the recovery-emergency repair coupling, and adopts an event-driven approach to continuously update the current emergency repair target.

[0150] Table 1 Comparison of key indicators under different schemes

[0151] As shown in Table 1, Scheme 3 outperforms the other two schemes in both the key indicators of "90% load recovery time" and "load reduction amount". Specifically, the 90% load recovery time for Scheme 3 is... The load reduction in Scheme 3 is 17.365 MWh, significantly earlier than the 15.1 h in Scheme 1 and 12.6 h in Scheme 2. This indicates that the method proposed in this invention can improve the load recovery level more quickly in the early stage of post-disaster recovery and effectively reduce the scale of system load loss, thereby improving the emergency recovery efficiency of the distribution network.

[0152] Combination Figure 2 It is evident that the condition of roadsides continuously improves over time after a disaster, with the number of closed and restricted road sections decreasing rapidly and the number of unobstructed road sections gradually increasing. This indicates that road accessibility exhibits a significant temporal evolution during post-disaster repair. Ignoring this time-varying process and treating roadside conditions as fixed would fail to accurately reflect the changes in accessibility of fault points at different recovery stages, and would also make it difficult to truly characterize the differences in travel costs for repair teams in the post-disaster environment. Therefore, introducing a time-varying road network model into post-disaster recovery decision-making helps to more accurately describe the actual accessibility of repair resources at different stages. Figure 3 It can be seen that the condition of roadsides after a disaster continues to improve over time, with the number of closed and restricted road sections gradually decreasing and the number of unobstructed road sections gradually increasing. This indicates that introducing a time-varying road network model into post-disaster recovery decision-making helps to more accurately describe the accessibility of faults and the cost of emergency repairs at different stages.

[0153] Figure 3 This indicates that Scheme 3 does not simply sort the tasks by proximity or shortest repair time, but dynamically rearranges the emergency repair tasks after each event is triggered, taking into account the current roadside status, fault repair time, and recovery benefits after candidate fault repair. Figure 4 Further analysis shows that Scheme 3 exhibits a faster decrease in load loss and enters a low load loss state earlier, verifying that the present invention has superior recovery performance in the critical stage of post-disaster recovery. Further analysis of Schemes 1 and 3 reveals that although both employ a rolling update mechanism, their decision-making bases differ. Scheme 1's rolling mechanism primarily involves reordering unprocessed faults based on static priorities and the current repair time cost after system state changes; its essence remains a rolling reordering based on static priorities. Scheme 3, however, further considers the time-varying nature of roadside conditions, the dynamic nature of repair time, and the recovery benefits after candidate fault repair after each event trigger, re-evaluating and selecting the current repair target. Therefore, it belongs to an event-driven rolling decision-making process coupled with recovery and repair.

[0154] It should be noted that although the fault repair completion time of Scheme 3 is slightly later than that of Scheme 1 and Scheme 2, this does not mean that the overall performance is worse. Rather, it shows that the optimization goal of this invention is not simply to complete all faults as early as possible, but to prioritize the repair of faults that contribute more to load recovery, so as to improve the recovery rate more quickly and significantly reduce the load loss under the condition of limited repair resources.

[0155] In summary, while Scheme 1 possesses rolling repair capabilities, its rolling updates are primarily based on static priorities and current repair time costs, failing to incorporate real-time recovery benefits after candidate fault repairs into unified decision-making. Scheme 2 does not consider time-varying road networks and dynamic maintenance durations, making it difficult to accurately reflect post-disaster traffic conditions and actual repair costs. Scheme 3, by introducing time-varying road networks, dynamic maintenance durations, and event-driven rolling decision-making that couples recovery and repair, achieves coordinated optimization of traffic accessibility, repair time, and recovery benefits, thus demonstrating superior performance in both load recovery speed and load loss control.

[0156] Please refer to Figure 5 A power distribution network emergency repair strategy update system 200 that takes into account the road network traffic status includes a memory 202, a processor 204, and a computer program stored in the memory 202 and executable on the processor 204. When the processor 204 executes the computer program, it implements the aforementioned power distribution network emergency repair strategy update method that takes into account the road network traffic status.

[0157] The above description is merely an embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent modifications made based on the content of the present invention specification and drawings, or direct or indirect applications in related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A method for updating emergency repair strategies for a distribution network considering road network traffic conditions, characterized in that, Includes the following steps: Construct a time-varying traffic network containing fault points based on the post-disaster roadside conditions, and calculate the dynamic travel time of the repair team to the fault points. Obtain the fault baseline working time, and combine it with the dynamic passage time, the fault baseline working time, real-time weather conditions, real-time terrain conditions, and real-time material status to calculate the total handling time of the fault point in real time. A power distribution network online recovery optimization model is constructed, and the baseline recovery result within a preset rolling time domain is obtained by combining the post-disaster roadside conditions through the output of the power distribution network online recovery optimization model. Based on the baseline recovery results, the dynamic passage time, and the total handling time, priority repair points are determined from the fault points; During the execution of emergency repair tasks corresponding to the priority repair points within the preset rolling time domain, the priority repair points that have not yet been repaired are re-determined based on the latest post-disaster roadside status, until the post-disaster emergency repair and restoration are completed. 2.The method of claim 1, wherein, The baseline recovery result obtained by the online recovery optimization model of the distribution network within the preset rolling time domain is as follows: Obtain the current known fault set, power outage event information, and line availability status from the post-disaster roadside conditions; The known fault set, power outage event information, and line availability status are input into the distribution network online recovery optimization model to obtain the baseline recovery results. 3.The method of claim 1, wherein, The specific steps for constructing the online recovery optimization model for the power distribution network are as follows: A target function is established with the objective of maximizing the number of load nodes in the load node set that are cumulatively restored within the preset rolling time domain, based on a preset set of load nodes corresponding to the time-varying traffic network. Establish constraints on line availability, load recovery, node power balance, voltage and power flow operation, line switch operation, and radial topology constraints based on the single commodity flow method. By combining the objective function, the line availability constraints, the load recovery constraints, the node power balance constraints, the voltage and power flow operation constraints, the line switch action constraints, and the radial topology constraints, the online recovery optimization model of the distribution network is constructed.

4. The method of claim 1, wherein, The construction of a time-varying traffic network containing fault points, and the calculation of the dynamic travel time from the maintenance team to the fault points, include: The roadside status is divided into closed status, restricted traffic status, and open status; The roadside weights are determined based on the roadside conditions, and the expression is as follows: ； in, Indicates the weight of the roadside. Indicates the time e at roadside state, This indicates the basic travel time at roadside e. This represents the amplification factor of travel time at roadside e under traffic restriction conditions; Construct a time-varying traffic network based on the current roadside state; The dynamic travel time from the maintenance team to the fault point is calculated based on the shortest path method: ； in, ； In the formula, This indicates that the maintenance team is always Location, This indicates the location corresponding to fault point k. Indicates time-varying transportation networks Up by arrive The set of feasible paths Representing a path The set of road edges included.

5. The method of claim 1, wherein, The specific method for calculating the total handling time of the fault point in real time is as follows: Set the corresponding fault baseline working hours according to the fault type of the fault point: Based on real-time weather conditions, real-time terrain conditions, and real-time material status, weather influence factors, terrain influence factors, and material influence factors are established respectively. The dynamic maintenance time of the fault repair point is calculated based on the fault baseline working time, the weather impact factor, the terrain impact factor, and the material impact factor. The total processing time is obtained by combining the dynamic maintenance time and the dynamic passage time.

6. The method of claim 1, wherein, The specific steps for determining the priority repair point from the fault point are as follows: Based on the total handling time and the post-disaster roadside condition, a hypothetical repair scenario is constructed; The online recovery optimization model for the power distribution network outputs the expected repair results of the fault point under the hypothetical repair scenario. Based on the difference between the expected repair result and the baseline recovery result, calculate the comprehensive score of the candidate fault for the corresponding fault point; Based on the comprehensive score of the candidate faults, the dynamic passage time, and the total handling time, priority repair points are determined from the fault points.

7. The method for updating the emergency repair strategy of a distribution network considering the road network traffic status according to claim 6, characterized in that, The specific scenario for constructing the hypothesis repair is as follows: The accessibility of the fault point is determined based on the dynamic passage time. Based on the fault reachability and the total handling time, a candidate fault set is formed from the fault points; Based on the total processing time, determine the estimated completion time of the repair for the corresponding pending fault points; Based on the expected completion time of the repair, update the availability status of the lines corresponding to the pending fault points, and construct the corresponding hypothetical repair scenarios. 8.The method of claim 1, wherein, The specific steps for obtaining and basing the latest post-disaster roadside conditions are as follows: During the execution of emergency repair tasks at the corresponding priority repair points, monitor emergency repair task completion events, roadside status change events, and new fault activation events; The post-disaster roadside status is updated after any one of the following events occurs: completion of emergency repair, change of roadside status, or new fault activation.

9. The method of claim 8, wherein the method further comprises: Also includes: Establish a fault set that includes all fault points; After the emergency repair task is completed, remove the fault points that have been repaired by the emergency repair task completion event from the fault set, and update the post-disaster roadside status. If the fault set is empty and no new fault activation event occurs, then the emergency repair and post-disaster recovery are completed. 10.A system for updating a power distribution network repair strategy considering road network traffic state, characterized in that, The device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the method for updating the emergency repair strategy of a power distribution network that takes into account the road network traffic status, as described in any one of claims 1 to 9.

Images