A road network power emergency guarantee and automatic recovery system

Abstract

The application provides a road network power emergency guarantee and automatic recovery system, relates to the technical field of power distribution network emergency guarantee and self-healing control, and comprises a multi-modal fusion sensing module, a multi-source emergency power source cooperative scheduling module, a power distribution network flexible reconstruction recovery module, an extreme weather active protection module and a cross-department closed-loop management and control platform. The multi-modal fusion sensing module accesses electrical quantity, temperature, partial discharge sensing data and roadside AI vision data of a linear power distribution line of a road network, and is used for fault section positioning and implicit fault prediction. The multi-source emergency power source cooperative scheduling module is based on a road network traffic load grading system. Through the multi-modal fusion linear power distribution network fault accurate tracing and implicit fault prediction module, the application realizes full-dimensional sensing upgrading of the road network power distribution scene, greatly improves the fault identification efficiency and positioning accuracy of the road network power distribution network, and solves the core pain point of long-distance linear line fault troubleshooting difficulty.

Description

Technical Field

[0001] This invention relates to the field of power distribution network emergency protection and self-healing control technology, and in particular to a power network emergency protection and automatic recovery system. Background Technology

[0002] With the rapid development of transportation networks, linear distribution networks along roads are the power supply foundation for core transportation facilities such as tunnel lighting, traffic lights, and roadside monitoring. Their power supply reliability directly affects the safety and efficiency of road network traffic. Existing distribution network operation and maintenance systems generally rely on monitoring electrical quantities such as current and voltage to detect faults, and are equipped with uninterruptible power supplies (UPS) and energy storage devices to ensure power supply to critical loads. Simultaneously, they utilize distribution network self-healing and reconfiguration technologies to restore power to non-faulty areas. Some intelligent solutions introduce distributed sensor networks to achieve line status awareness and are complemented by operation and maintenance management platforms to dispatch fault work orders and track processes, providing mature technical support for the stable operation of conventional distribution networks.

[0003] However, the existing power distribution network protection system has the following shortcomings in adaptability to the specific scenarios of linear road network distribution: First, fault perception mainly relies on the collection of pure electrical parameters, which cannot identify external fault causes that frequently occur in road network power distribution, such as fallen trees and damage from external construction. Furthermore, it lacks the ability to predict latent faults such as insulation aging and poor contact, and the fault location accuracy is insufficient to meet the operation and maintenance needs of long-distance linear lines. Second, emergency power dispatching does not combine the traffic safety value of the road network load for hierarchical protection, the power supply priority of core transportation facilities is unclear, and the coordinated dispatching efficiency of multiple types of emergency power supplies is low. Third, conventional self-healing reconfiguration schemes for distribution networks are mainly adapted to radial and ring-shaped urban distribution networks. The existing power grids suffer from several problems. First, they lack adaptability to long-distance, multi-segmented linear topologies, and their fault isolation and power restoration speeds cannot meet the requirements for safe operation. Second, they lack proactive protection mechanisms against extreme weather conditions for exposed distribution lines, relying mainly on reactive repairs after faults occur, resulting in weak system resilience under extreme disaster scenarios. Third, there are significant data barriers between power distribution maintenance and traffic control, road administration, and emergency command departments, leading to lagging cross-departmental collaboration in fault handling and the absence of a closed-loop management system covering the entire chain from fault warning to repair acceptance. Therefore, this invention proposes a power emergency protection and automatic recovery system for road networks to address these issues. Summary of the Invention

[0004] To address the aforementioned issues, this invention proposes a road network power emergency protection and automatic recovery system. This system achieves a comprehensive upgrade in perception of road network power distribution scenarios through a multi-modal fusion module for precise fault tracing and latent fault prediction in linear distribution networks. On one hand, it breaks through the limitations of traditional pure electrical perception by integrating roadside monitoring AI visual data, enabling accurate identification of externally induced faults that frequently occur in road network power distribution, such as fallen trees and external damage, perfectly adapting to the operating environment of open-air lines linearly distributed along the road network. On the other hand, it employs a two-level algorithm of dual-end traveling wave initial positioning + graph neural network fine positioning to achieve meter-level fault location accuracy, while also possessing a 7-15 day latent fault prediction capability, covering the entire process of pre-warning and in-process location. This provides high-precision perception input for subsequent emergency dispatch and fault recovery, significantly improving the efficiency and accuracy of fault identification in road network power distribution networks, and solving the core pain point of difficult fault investigation in long-distance linear lines.

[0005] To achieve the objectives of this invention, the following technical solution is provided: a road network power emergency protection and automatic recovery system, comprising a multimodal fusion sensing module, a multi-source emergency power supply collaborative scheduling module, a distribution network flexible reconfiguration and recovery module, an extreme weather active protection module, and a cross-departmental closed-loop management platform. The multimodal fusion sensing module accesses electrical quantity, temperature, and partial discharge sensor data of the linear distribution lines of the road network, along with roadside AI visual data, for fault section location and latent fault prediction. The multi-source emergency power supply collaborative scheduling module, based on the road network traffic load classification system, coordinates multiple types of emergency power supplies to achieve classified power supply protection.

[0006] The flexible reconfiguration and recovery module of the power distribution network adapts to the linear power distribution topology of the road network and realizes the isolation of faulty sections and the restoration of power supply to non-faulty sections through flexible interconnection devices; the extreme weather active protection module accesses meteorological early warning and road network GIS data for risk section identification, differentiated active protection and flexible operation and management during disasters; the cross-departmental closed-loop management and control platform communicates with each of the above modules and connects to the data interfaces of multiple departmental systems to realize closed-loop management and control of the entire emergency process.

[0007] A further improvement lies in the fact that the multimodal fusion sensing module adopts a two-level fault location mechanism combining dual-end traveling wave initial localization with graph neural network fine localization. The fault distance calculation formula for dual-end traveling wave initial localization is as follows:

[0008]

[0009] In the formula, L is the distance from the fault point to the beginning of the distribution line, in meters; v is the propagation speed of the transient traveling wave in the distribution line, in meters per second; ts is the time it takes for the transient traveling wave to reach the detection terminal at the beginning of the line, in seconds; and tm is the time it takes for the transient traveling wave to reach the detection terminal at the end of the line, in seconds.

[0010] A further improvement lies in the fact that the graph neural network precise localization calculates the fault matching degree of each segment by fusing multimodal features. The formula for calculating the fault matching degree is:

[0011]

[0012] In the formula, Si is the fault matching degree of the i-th distribution section, with a value range of 0-1; Ei is the electrical quantity deviation characteristic value of the i-th section, which is calculated by normalizing the deviation of current and voltage amplitude; Ti is the temperature anomaly characteristic value of the i-th section, which is calculated by normalizing the deviation of line temperature from historical benchmark value; Vi is the visual anomaly characteristic value of the i-th section, which is calculated by normalizing the tree obstruction and external force damage characteristics of roadside AI visual recognition; α, β, and γ are the weighting coefficients of electrical quantity, temperature, and visual characteristics, respectively, satisfying α+β+γ=1.

[0013] A further improvement lies in the fact that the multi-source emergency power supply collaborative scheduling module constructs a spatiotemporal supply and demand matching scheduling model, the objective function of which is:

[0014]

[0015] In the formula, F is the comprehensive scheduling cost; T2,j is the power outage duration of the j-th secondary load, in minutes; η is the comprehensive utilization rate of all emergency power supplies, which is the ratio of actual power supply capacity to rated total capacity; ω1 and ω2 are the weighting coefficients of the power outage duration and power utilization rate of the secondary load, respectively; the constraint condition of the scheduling model is that the cumulative power outage duration T1 of all primary loads is 0.

[0016] Further improvements include: the road network traffic load classification system is divided into three levels according to the degree of impact on traffic safety, where level one load includes tunnel ventilation and lighting facilities, ramp traffic lights, and monitoring equipment for accident-prone sections; level two load includes main road lighting facilities and monitoring equipment for ordinary road sections; and level three load includes landscape lighting facilities. The multi-source emergency power supply includes four types: roadside distributed photovoltaic, fixed energy storage devices, segmented UPS equipment, and mobile power generation vehicles, and is equipped with a millisecond-level UPS seamless switching mechanism and a linkage energy replenishment mechanism between energy storage and mobile power generation vehicles.

[0017] A further improvement is made in that the flexible reconfiguration and recovery module for the distribution network uses an improved discrete particle swarm optimization algorithm to solve for the optimal topology reconfiguration scheme, and its fitness calculation formula is as follows:

[0018]

[0019] In the formula, fit is the fitness value of the topology scheme, and the smaller the value, the better the scheme; Ploss,k is the active power loss of the kth distribution branch, in kW; Udev,k is the voltage deviation rate of the kth distribution node, which is the ratio of the difference between the actual voltage and the rated voltage to the rated voltage; n is the total number of distribution branches, and m is the total number of distribution nodes; λ1 and λ2 are the weighting coefficients of network loss and voltage deviation, respectively.

[0020] Further improvements include: the flexible interconnection device is composed of a combination of solid-state circuit breakers and soft switching equipment, deployed at each segment node of the linear distribution network, achieving precise isolation of fault sections within ≤20ms when a fault occurs; the flexible reconfiguration and recovery module of the distribution network has a built-in adaptive reconfiguration and automatic recovery mechanism, which covers the entire process of fault isolation, topology reconfiguration, and safety verification. After the topology reconfiguration is completed, an automatic overload control mechanism is provided to avoid secondary faults caused by line overload by adjusting the power output and load switching.

[0021] A further improvement is made in that: the extreme weather active protection module divides disaster risk zones using a risk assessment model, and the risk value calculation formula is as follows:

[0022]

[0023] In the formula, Rk is the meteorological disaster risk value of the kth power distribution section; Mk is the meteorological warning intensity coefficient of the corresponding section, which is taken as 0.2-1.0 according to the warning level of rainstorm, snow, strong wind, and lightning; Gk is the terrain vulnerability coefficient of the corresponding section, which is taken as 0.5-1.5 according to the historical data of road section altitude, slope, water accumulation / ice accumulation; μ is the scene correction coefficient, which is taken according to the type of overhead / cable laying of the line.

[0024] Further improvements include: the extreme weather active protection module is configured with differentiated active protection strategies and a disaster resilience mechanism. The differentiated active protection strategies are configured according to the risk level of the affected area, including DC de-icing control for high-risk icing areas, pre-switching energy storage power supply for low-lying waterlogged areas, and encrypted monitoring and early warning for areas prone to tree obstruction. The disaster resilience mechanism includes low-power endurance mode for edge nodes, data linkage access for UAV inspections, and dynamic adjustment of emergency dispatch and topology reconstruction schemes to ensure continuous power supply to core primary loads under extreme disasters.

[0025] Further improvements include: the cross-departmental closed-loop management platform connects the power operation and maintenance system, traffic control system, road administration management system, and emergency command system with standardized data interfaces; after a fault is triggered, it automatically links with the traffic management system to adjust traffic light timings and issue detour prompts for road sections, while automatically generating emergency repair work orders and planning the optimal inspection and repair route; after the repair is completed, it automatically verifies the power supply status and synchronizes information across all departments, constructing a fully visualized management and control system with a closed loop for fault early warning, location isolation, emergency power supply, network restoration, operation and maintenance repair, and acceptance.

[0026] The beneficial effects of this invention are as follows:

[0027] 1. This invention achieves a comprehensive upgrade of perception in road network power distribution scenarios through a multi-modal fusion module for precise fault tracing and latent fault prediction in linear power distribution networks. On the one hand, it breaks through the limitations of traditional pure electrical perception by integrating roadside monitoring AI visual data, which can accurately identify externally induced faults that frequently occur in road network power distribution, such as fallen trees and external damage, perfectly adapting to the operating environment of open-air lines linearly distributed along the road network. On the other hand, it adopts a two-level algorithm of dual-end traveling wave initial positioning + graph neural network fine positioning to achieve meter-level fault location accuracy, while also having a 7-15 day latent fault prediction capability, covering the entire process of pre-warning and in-process location, providing high-precision perception input for subsequent emergency dispatch and fault recovery, significantly improving the fault identification efficiency and location accuracy of road network power distribution networks, and solving the core pain point of difficult fault investigation in long-distance linear lines.

[0028] 2. The multi-source emergency power supply collaborative scheduling system for road network load classification and matching of the present invention constructs a dedicated power supply guarantee system adapted to road network scenarios. Based on the impact of traffic safety, a three-level load classification system is established to accurately match the safety value of road network-specific loads such as tunnel ventilation lighting, ramp traffic lights, and accident-prone section monitoring, and clarifies the priority boundaries of graded protection. By integrating four types of multi-source emergency power supplies—roadside photovoltaics, fixed energy storage, segmented UPS, and mobile generators—and constructing a spatiotemporal supply and demand matching scheduling model, the system optimizes for multiple objectives: "zero power outages for primary loads, shortest power outage duration for secondary loads, and highest power utilization." It is equipped with a millisecond-level seamless switching and energy storage-generator linkage replenishment mechanism, achieving full-scenario coverage of short-term seamless power supply and long-term continuous protection, effectively improving the power supply reliability of core road network transportation facilities and the resource utilization efficiency of emergency power supplies.

[0029] 3. The adaptive reconfiguration and automatic recovery mechanism for distribution networks based on flexible interconnection of this invention realizes rapid self-healing of linear distribution networks. Through a flexible interconnection device composed of solid-state circuit breakers and soft switches, it can achieve precise isolation of fault sections within 20ms, avoiding the expansion of the fault range. With the optimal topology reconfiguration scheme solved by the improved discrete particle swarm optimization algorithm, it can achieve second-level power restoration of non-faulty sections. At the same time, it is equipped with an automatic overload control mechanism to carry out safety verification after reconfiguration, avoiding secondary faults caused by line overload. For the linear topology characteristics of distribution networks with many segments and long paths, flexible reconfiguration of linear networks is achieved through flexible transformation of segmented switches, which fully covers the entire self-healing process of fault isolation-topology reconfiguration-safety verification, significantly shortening the power outage time in non-faulty areas and reducing the impact of power outages on road network traffic.

[0030] 4. The active protection and disaster resilience mechanism linked to extreme weather of this invention upgrades the road network power distribution system from passive emergency response to active defense. By connecting to the meteorological early warning and road network GIS system, it identifies risk zones based on four core extreme weather events affecting the road network: rainstorms, snow, strong winds, and lightning. It also provides differentiated active protection strategies to reduce the failure rate under extreme weather conditions. At the same time, it has the ability to operate resiliently during extreme disasters. Through low-power endurance of edge nodes, linkage with drone inspection data, and dynamic adjustment and reconfiguration schemes, it ensures continuous power supply to core loads, making up for the shortcomings of conventional self-healing systems in extreme disaster scenarios and significantly improving the disaster resilience of the road network power distribution system.

[0031] 5. The cross-departmental data interoperability full-link emergency closed-loop management platform of this invention breaks down the cross-departmental collaboration barriers of the road network power system. It is the central brain of the entire system. By connecting the standardized data interfaces of the four major systems of power operation and maintenance, traffic control, road administration, and emergency command, it can coordinate multiple departments to perform actions such as traffic signal timing adjustment, detour notice issuance, automatic generation of repair work orders, and inspection route planning after a fault occurs. After the repair is completed, it automatically verifies and synchronizes information from all departments, perfectly adapting to the core attributes of road network power supply services for traffic and emergency linkage with multiple departments. At the same time, it constructs a full-process visualized management and control system, integrating the technical capabilities of the aforementioned four modules to achieve full-link implementation from technical actions to business processes, comprehensively improving the collaborative efficiency and refined management level of road network power emergency response. Attached Figure Description

[0032] Figure 1 This is a diagram illustrating the composition of the present invention. Detailed Implementation

[0033] To enhance understanding of the present invention, the present invention will be further described in detail below with reference to embodiments. These embodiments are only used to explain the present invention and do not constitute a limitation on the scope of protection of the present invention.

[0034] Example 1

[0035] according to Figure 1 As shown, this embodiment proposes a road network power emergency protection and automatic recovery system. The application scenario is a 12km long 10kV overhead linear distribution line along a city's ring expressway. This line is laid along the outer side of the road, passing through suburban woodlands, construction sections, and other areas prone to tree obstruction and external damage, providing power support for all traffic facilities along the route. High-precision transient traveling wave acquisition terminals, fiber optic distributed temperature sensors, and ultrasonic partial discharge detection devices are simultaneously deployed at the beginning and end of the line and at every 2km segment node. The sampling frequency is set to 10MHz to achieve synchronous acquisition and timestamp alignment of electrical quantities, temperature, and partial discharge data. Simultaneously, 24 high-definition roadside surveillance cameras already deployed along the line are connected. A lightweight YOLOv5 visual recognition model is deployed at the edge gateway, which can identify three types of high-incidence external anomalies in the road network in real time: fallen trees, vehicles hitting poles, and construction machinery crossing boundaries. The target recognition accuracy can reach over 96%. When a line fault occurs, the system automatically initiates a two-stage positioning process: the first step performs dual-end traveling wave initial positioning, capturing the fault transient current traveling wave fronts through the traveling wave terminals at both ends of the line, and accurately recording the arrival times ts and tm of the wave fronts. The transient traveling wave propagation velocity v in the overhead line is taken as 2.94 × 10⁻⁶. 8 The first step involves substituting the speed (m / s) into the traveling wave ranging formula to calculate the approximate range of the fault point. The initial positioning error can be controlled within 500m. The second step executes graph neural network fine positioning, extracting the electrical quantity deviation features Ei, temperature anomaly features Ti, and visual anomaly features Vi from three subdivided sections within the initial positioning coverage area. The weight coefficients for regular road sections are set to α=0.6, β=0.2, and γ=0.2. For sections with a high incidence of external force damage, the visual feature weight γ is increased to 0.35. The fault matching degree formula is then used to calculate the fault probability of each section. The section with the highest matching degree is the accurate fault section. The final positioning accuracy can reach within 5 meters. At the same time, it can automatically distinguish between three types of faults: short circuit faults, grounding faults, and external force damage faults. To predict latent faults, the system continuously collects 90 days of time-series data on line joint temperature, partial discharge amplitude, and contact resistance. It extracts degradation features using the ARIMA time-series trend fitting algorithm. When insulation aging-related features exceed the warning threshold for 7 consecutive days, the system automatically issues a hidden danger warning. It can identify latent defects such as insulation aging and poor terminal contact 7-15 days in advance, enabling proactive investigation and preventing sudden faults from affecting the operation of the road network.

[0036] Example 2

[0037] according to Figure 1As shown, this embodiment proposes a road network power emergency protection and automatic restoration system, taking a 10kV power distribution system along a mountainous highway as the application scenario. The line is 28km long, passing through 5 long tunnels and 12 interchange ramps, and all loads along the line are traffic-specific facilities. The system establishes a three-level load classification system based on the degree of impact on traffic safety: Level 1 loads include ventilation and lighting facilities in 5 tunnels, traffic lights at 12 entrance ramps, and high-definition monitoring equipment at 8 accident-prone sections, with a total capacity of 860kW. This has the highest priority and power supply cannot be interrupted under any circumstances. Level 2 loads include lighting facilities on the main road and monitoring equipment at 22 ordinary road sections, with a total capacity of 1200kW. This has the second highest priority and short-term interruptions are allowed with priority restoration. Level 3 loads include 15 roadside landscape lighting and advertising screens, with a total capacity of 420kW. These are interruptible loads and can be directly disconnected when there is a large power shortage. The emergency power supply integrates 12 roadside distributed photovoltaic systems (total installed capacity 2.4MW), 5 fixed energy storage stations at tunnel entrances (total capacity 6MWh), and segmented UPS devices in key sections (total capacity 1.2MWh). It also incorporates 4 500kW mobile generators on standby at 3 highway service areas as regional backup capacity. When a line fault causes a power outage in a section, the multi-source emergency power supply collaborative scheduling module immediately activates the spatiotemporal supply and demand matching scheduling model. With the optimization objective function as the core, weighting coefficients ω1=0.7 and ω2=0.3 are set to prioritize the reliability of power supply to primary loads while also considering power utilization efficiency. The model strictly meets the constraint that "the cumulative outage time of primary loads T1=0". The dispatch execution is divided into three levels of linkage: The first level is millisecond-level seamless switching, where segmented UPS automatically engages within 10ms after a fault is triggered, ensuring uninterrupted power supply to primary loads and avoiding traffic risks caused by power outages in transportation facilities; the second level is continuous power supply from fixed energy storage, dynamically allocating output from nearby energy storage stations based on the total load capacity and SOC status of the faulty section to maintain normal power supply to secondary loads; the third level is mobile generator dispatch for supplementary power, where if the predicted fault repair time exceeds the remaining battery life of the energy storage (energy storage SOC below 20%), the system automatically matches the nearest mobile generator, plans the optimal route based on real-time highway traffic conditions, and coordinates with road administration departments in advance to prepare access routes, achieving long-term continuous power supply assurance. If a significant power shortage occurs during dispatch, the system can automatically disconnect all tertiary loads and reduce the supply to secondary loads in stages as needed, achieving the optimization goal of minimizing the power outage time of secondary loads and maximizing the overall utilization rate of power supply while ensuring power supply to core transportation facilities.

[0038] Example 3

[0039] according to Figure 1As shown, this embodiment proposes a road network power emergency protection and automatic restoration system. It focuses on the flexible transformation of a 10kV single-radial linear distribution line on a city's outer ring expressway. The line is 16km long and has 9 segment nodes. During normal operation, it uses a single-end power supply mode, with intermediate segments serving as normally open connection points, connecting to adjacent distribution branches as backup power sources. At each 2km segment node, a flexible interconnection device consisting of fast solid-state circuit breakers and bidirectional soft switches is deployed. The solid-state circuit breakers have a rated opening time of ≤12ms, enabling rapid interruption of fault current; the soft switches have a rated capacity of 1MW, allowing for bidirectional power smooth adjustment and supporting flexible power transfer. A total of 8 sets of flexible interconnection devices are deployed along the entire line, upgrading traditional mechanical segment switches to fully controllable flexible nodes. When a permanent short-circuit fault occurs in a section of the line, the distribution network flexible reconfiguration and recovery module first triggers rapid fault isolation: after the flexible interconnection devices at both ends of the faulty section detect the overcurrent fault characteristics, the solid-state circuit breaker immediately triggers tripping, completing electrical isolation between the two ends of the faulty section within 18ms, effectively preventing the fault current from spreading to upstream and downstream normal sections and reducing the scope of the power outage. Subsequently, the adaptive topology reconfiguration process is initiated, using an improved discrete particle swarm optimization algorithm to solve for the optimal reconfiguration scheme. The algorithm uses the on / off state of each flexible switch as a particle code (0 represents open, 1 represents closed), with an initial population size of 30, a maximum number of iterations of 50, and a convergence threshold of 1×10⁻⁶. −5 Using the fitness function as the optimization objective, weighting coefficients λ1=0.4 and λ2=0.6 are set to prioritize node voltage quality while minimizing line network losses. Combining the current line topology, backup power capacity, and real-time load conditions, all feasible solutions are traversed to ultimately output the optimal switch combination scheme that satisfies safety constraints. After the scheme is implemented, non-faulty sections can be resupplied to adjacent normal power supply branches via tie points, achieving power restoration within 3 seconds for non-faulty sections. After topology reconfiguration, the system automatically performs a safety verification process, monitoring the load rate of each branch and the voltage deviation of each node in real time. Node voltage deviation is required to not exceed ±7% of the rated value, and branch load rate not to exceed 85% of the rated capacity. If overload or voltage exceeding limits is detected, the automatic overload control mechanism is immediately activated. This eliminates the risk of exceeding limits by adjusting the charging and discharging output of nearby energy storage devices and disconnecting tertiary loads as needed, preventing secondary faults caused by line overload, and fully covering the self-healing process of "fault isolation - topology reconfiguration - safety verification".

[0040] Example 4

[0041] according to Figure 1As shown, this embodiment proposes a road network power emergency protection and automatic recovery system, which accesses the four types of extreme weather warning data of rainstorm, snow, strong wind and lightning from the local meteorological department, and simultaneously imports road network GIS topographic data and the historical disaster ledger of the line in the past 5 years. The system takes the 32 distribution sections of the entire line as independent assessment units and carries out disaster risk classification through risk value calculation formula. The meteorological warning intensity coefficient Mk is determined based on the official warning level: 0.2 for blue warning, 0.4 for yellow warning, 0.7 for orange warning, and 1.0 for red warning; the terrain vulnerability coefficient Gk is determined based on the geographical attributes of the section: 1.5 for low-lying areas prone to water accumulation, 1.4 for high-altitude areas prone to icing, and 0.6 for ordinary plain areas; the scene correction coefficient μ is determined based on the line laying method: 1.0 for overhead lines and 0.5 for buried cable sections; finally, the sections are divided into three levels according to the risk value Rk: high risk (Rk≥0.7), medium risk (0.3≤Rk<0.7), and low risk (Rk<0.3). The system implements differentiated proactive protection strategies for sections with different risk levels: When an orange alert for snow and ice is issued in winter and the predicted ice thickness on the lines exceeds 5mm, the DC de-icing device in high-risk overhead sections is automatically activated to eliminate the risk of ice-induced power outages in advance; when a red alert for heavy rain is issued, the power supply mode of low-lying waterlogged sections is switched to independent energy storage power supply one hour in advance, and the main grid connection is disconnected to prevent water from entering the lines and causing short-circuit faults to escalate; when a blue or higher alert for strong winds is issued, the frequency of visual monitoring in sections with a high incidence of tree obstacles is increased from once every 15 minutes to once every minute to identify the risk of tree falling in advance and push maintenance warnings. When extreme disasters cause public network communication to be interrupted in some sections, the system automatically activates the disaster resilience protection mechanism: edge computing nodes switch to low-power endurance mode, turn off unnecessary monitoring and data upload functions, and only retain core load management and local wireless communication capabilities, which can extend the device's endurance to more than 72 hours; at the same time, it accesses the line image data transmitted back by maintenance drones, automatically identifies fault points and updates line status, and dynamically adjusts emergency power scheduling and topology reconstruction schemes, so as to ensure the continuous power supply of primary core loads even in communication-constrained scenarios, making up for the shortcomings of conventional self-healing systems in extreme disaster scenarios.

[0042] Example 5

[0043] according to Figure 1As shown, this embodiment proposes a road network power emergency protection and automatic recovery system, establishing a city-level road network power emergency closed-loop management and control platform as the central dispatch core of the entire system. The platform uses standardized data interfaces to connect four major business systems: it connects to the power operation and maintenance system via the IEC61850 power communication protocol to obtain real-time distribution network operation status, fault information, and operation and maintenance team resource data; it connects to the traffic control platform via the NTCIP traffic signal standard protocol to achieve linkage between traffic light timing adjustment and traffic information release; it connects to the road administration management system via the government data sharing interface to obtain road construction, control information, and real-time traffic data; and it connects to the emergency command platform via a dedicated emergency network to achieve unified dispatch and information synchronization for emergencies. The platform's end-to-end data synchronization frequency is once every 5 seconds, ensuring the real-time performance and consistency of data from all departments.

[0044] When the system detects a fault or issues a potential hazard warning, the platform automatically triggers a full-link closed-loop management process: The first step is coordinated warning and traffic management. Within 10 seconds of fault assessment, the location of the faulty road segment, the expected impact range, and the recovery time are pushed to the traffic control platform. This involves coordinating adjustments to the signal timings at upstream and downstream intersections of the faulty road segment, extending the green light duration on the main line, controlling ramp entrance traffic, and simultaneously publishing detour information via roadside information screens and navigation apps to reduce the impact of power outages on road network efficiency. The second step is emergency repair resource scheduling. The system automatically matches the nearest maintenance team and repair materials, and automatically generates a temporary road occupancy emergency repair application based on road control information from the road administration system. The system first plans the optimal repair route, and then automatically dispatches work orders to the handheld terminals of maintenance personnel. The second step is process tracking and multi-department collaboration. The real-time location of the repair personnel and the progress of on-site handling are simultaneously pushed to the emergency command platform and the road administration system, allowing multiple departments to monitor the progress of the handling in real time and coordinate their efforts. The third step is acceptance closure. After the repair is completed, the system automatically verifies the power supply operation data of the faulty section. If there are no abnormalities for 30 consecutive minutes, the case is automatically closed, and the handling results, fault causes, and rectification suggestions are synchronized to all related systems. This forms a fully visualized management and control system that includes "fault warning - location isolation - emergency power supply - network restoration - maintenance and repair - acceptance closure", realizing the full-link implementation from technical handling to business collaboration.

[0045] Comparative example:

[0046] This comparative example adopts a traditional general power distribution network protection scheme, which only monitors electrical quantities such as current and voltage to detect faults and uses the traditional impedance method for fault location. The emergency power supply uses a uniformly configured UPS and energy storage device, without classifying the road network load, and provides power supply according to the principle of proximity after a fault. The power distribution network uses traditional mechanical sectionalizing switches, combined with conventional self-healing reconfiguration algorithms to achieve fault recovery. There is no active protection mechanism for extreme weather, and manual inspection and repair are carried out after a fault occurs. Operation and maintenance management relies solely on the internal management platform of the power department and does not carry out data linkage with transportation, road administration and other departments.

[0047] Validation data:

[0048] Selecting the same 10kV power distribution network scenario and setting the same permanent fault condition in the intermediate section, the core performance indicators of the system of this invention are compared with those of the comparative scheme. The verification results are shown in the table below:

[0049] Fault location accuracy ≤5m ≥210m Fault section isolation time ≤18ms ≥230ms Cumulative power outage duration of Level 1 load 0min 6.2min Average power outage duration for secondary loads 1.7min 18.5min Early warning days for hidden faults 7-15 days Lack of predictive ability Decrease in failure rate under extreme weather conditions 68% No active protection Power restoration time in non-faulty sections 2.6s 36s Cross-departmental collaborative response time 8s 37min

[0050] The verification data shows that the present invention is significantly superior to traditional solutions in terms of fault location accuracy, fault isolation and recovery speed, core load protection capability, disaster protection capability, and cross-departmental collaboration efficiency. It is fully adapted to the emergency protection needs of linear power distribution scenarios in road networks, and the technical solution has sufficient rationality and engineering feasibility.

[0051] This road network power emergency protection and automatic recovery system achieves a full-dimensional perception upgrade of the road network power distribution scenario through a multi-modal fusion module for precise fault tracing and latent fault prediction in linear distribution networks. On the one hand, it breaks through the limitations of traditional pure electrical perception by accessing roadside monitoring AI visual data, which can accurately identify external causes of faults that frequently occur in road network power distribution, such as fallen trees and external damage, perfectly adapting to the operating environment of open-air lines linearly distributed along the road network. On the other hand, it adopts a two-level algorithm of dual-end traveling wave initial positioning + graph neural network fine positioning to achieve meter-level fault location accuracy, while also having a 7-15 day latent fault prediction capability, covering the entire process of pre-warning and in-process location, providing high-precision perception input for subsequent emergency dispatch and fault recovery, significantly improving the fault identification efficiency and location accuracy of the road network power distribution network, and solving the core pain point of difficult fault investigation in long-distance linear lines. The multi-source emergency power supply collaborative scheduling system for road network load classification and matching of this invention constructs a dedicated power supply guarantee system adapted to road network scenarios. Based on the impact of traffic safety, a three-level load classification system is established to accurately match the safety value of road network-specific loads such as tunnel ventilation lighting, ramp traffic lights, and accident-prone section monitoring, clarifying the priority boundaries of graded protection. By integrating four types of multi-source emergency power supplies—roadside photovoltaics, fixed energy storage, segmented UPS, and mobile generators—and constructing a spatiotemporal supply and demand matching scheduling model, the system optimizes for multiple objectives: "zero power outages for primary loads, shortest power outage duration for secondary loads, and highest power utilization." It is equipped with a millisecond-level seamless switching and energy storage-generator linkage replenishment mechanism, achieving full-scenario coverage of short-term seamless power supply and long-term continuous protection, effectively improving the power supply reliability of core road network transportation facilities and the resource utilization efficiency of emergency power supplies. The adaptive reconfiguration and automatic recovery mechanism for distribution networks based on flexible interconnection of this invention realizes rapid self-healing of linear distribution networks. Through a flexible interconnection device composed of solid-state circuit breakers and soft switches, it can accurately isolate faulty sections within 20ms, preventing the fault range from expanding. With the optimal topology reconfiguration scheme solved by the improved discrete particle swarm optimization algorithm, it can achieve second-level power restoration of non-faulty sections. At the same time, it is equipped with an automatic overload control mechanism to carry out safety verification after reconfiguration, avoiding secondary faults caused by line overload. For the linear topology characteristics of distribution networks with many segments and long paths, flexible reconfiguration of linear networks is achieved through flexible modification of segmented switches, which fully covers the entire self-healing process of fault isolation-topology reconfiguration-safety verification, significantly shortening the power outage time in non-faulty areas and reducing the impact of power outages on network traffic.The active protection and disaster resilience mechanism linked to extreme weather of this invention upgrades the road network power distribution system from passive emergency response to active defense. By connecting to the meteorological early warning and road network GIS system, it identifies risk zones based on four core extreme weather events affecting the road network: rainstorms, snow, strong winds, and lightning. It also provides differentiated active protection strategies to reduce the failure rate under extreme weather conditions by combining data on road elevation, terrain, and historical water / ice accumulation. Simultaneously, it possesses the ability to operate resiliently during extreme disasters. Through low-power endurance of edge nodes, data linkage with UAV inspections, and dynamic adjustment and reconfiguration schemes, it ensures continuous power supply to core loads, making up for the shortcomings of conventional self-healing systems in extreme disaster scenarios and significantly improving the disaster resilience of the road network power distribution system. This invention's cross-departmental data interoperability end-to-end emergency closed-loop management platform breaks down cross-departmental collaboration barriers in the road network power system. It serves as the central brain of the entire system, connecting standardized data interfaces across four major systems: power operation and maintenance, traffic control, road administration, and emergency command. Upon a fault, it can trigger coordinated actions from multiple departments, including adjusting traffic signal timings, issuing detour notices, automatically generating repair work orders, and planning inspection routes. After repairs are completed, it automatically verifies and synchronizes information across all departments, perfectly adapting to the core attributes of road network power supply services for traffic and emergency coordination among multiple departments. Simultaneously, it constructs a full-process visualized management system, integrating the technical capabilities of the aforementioned four modules to achieve end-to-end implementation from technical actions to business processes, comprehensively improving the collaborative efficiency and refined management level of road network power emergency response.

[0052] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A road network power emergency protection and automatic recovery system, comprising a multi-modal fusion sensing module, a multi-source emergency power supply collaborative scheduling module, a distribution network flexible reconfiguration and recovery module, an extreme weather active protection module, and a cross-departmental closed-loop management and control platform, characterized in that: The multimodal fusion sensing module accesses electrical quantity, temperature, and partial discharge sensor data of the linear power distribution lines of the road network, as well as roadside AI visual data, for fault section location and latent fault prediction; the multi-source emergency power supply collaborative scheduling module is based on the road network traffic load classification system, and coordinates multiple types of emergency power supplies to achieve classified power supply protection. The flexible reconfiguration and recovery module of the power distribution network adapts to the linear power distribution topology of the road network and realizes the isolation of faulty sections and the restoration of power supply to non-faulty sections through flexible interconnection devices; the extreme weather active protection module accesses meteorological early warning and road network GIS data for risk section identification, differentiated active protection and flexible operation and management during disasters; the cross-departmental closed-loop management and control platform communicates with each of the above modules and connects to the data interfaces of multiple departmental systems to realize closed-loop management and control of the entire emergency process.

2. The road network power emergency protection and automatic restoration system according to claim 1, characterized in that: The multimodal fusion sensing module adopts a two-level fault location mechanism combining dual-end traveling wave initial localization and graph neural network fine localization. The fault distance calculation formula for dual-end traveling wave initial localization is as follows: In the formula, L is the distance from the fault point to the beginning of the power distribution line, in meters; v is the propagation speed of the transient traveling wave in the power distribution line, with the unit being m / s; ts is the time it takes for the fault transient traveling wave to reach the detection terminal at the beginning of the line, in seconds; tm is the time it takes for the fault transient traveling wave to reach the detection terminal at the end of the line, in seconds.

3. The road network power emergency protection and automatic restoration system according to claim 2, characterized in that: The graph neural network-based precise localization calculates the fault matching degree of each segment by fusing multimodal features. The formula for calculating the fault matching degree is as follows: In the formula, Si is the fault matching degree of the i-th distribution section, with a value range of 0-1; Ei is the characteristic value of the electrical quantity deviation of the i-th section, which is calculated by normalizing the current and voltage amplitude deviations. Ti is the temperature anomaly characteristic value of the i-th section, which is calculated by normalizing the deviation between the line temperature and the historical benchmark value. Vi is the visual anomaly feature value of the i-th segment, which is obtained by normalizing the tree obstruction and external force damage features of roadside AI visual recognition; α, β, and γ are the weight coefficients of electrical quantity, temperature, and visual feature, respectively, satisfying α+β+γ=1.

4. The road network power emergency protection and automatic restoration system according to claim 1, characterized in that: The multi-source emergency power supply collaborative scheduling module constructs a spatiotemporal supply and demand matching scheduling model, and its optimization objective function is: In the formula, F is the comprehensive scheduling cost; T2,j is the power outage duration of the j-th secondary load, in minutes; η is the comprehensive utilization rate of all emergency power supplies, which is the ratio of actual power supply capacity to rated total capacity; ω1 and ω2 are the weighting coefficients of the power outage duration and power utilization rate of the secondary load, respectively; the constraint condition of the scheduling model is that the cumulative power outage duration T1 of all primary loads is 0.

5. A road network power emergency protection and automatic restoration system according to claim 4, characterized in that: The road network traffic load classification system is divided into three levels according to the degree of impact on traffic safety. The first level of load includes tunnel ventilation and lighting facilities, ramp traffic lights, and monitoring equipment for accident-prone sections. The second level of load includes main road lighting facilities and monitoring equipment for ordinary road sections. The third level of load includes landscape lighting facilities. The multi-source emergency power supply includes four types: roadside distributed photovoltaic, fixed energy storage devices, segmented UPS equipment, and mobile power generation vehicles. It is equipped with a millisecond-level UPS seamless switching mechanism and a linkage energy replenishment mechanism between energy storage and mobile power generation vehicles.

6. The road network power emergency protection and automatic restoration system according to claim 1, characterized in that: The flexible reconfiguration and recovery module for the distribution network uses an improved discrete particle swarm optimization algorithm to solve for the optimal topology reconfiguration scheme. Its fitness calculation formula is as follows: In the formula, fit is the fitness value of the topology scheme, and the smaller the value, the better the scheme; Ploss,k is the active power loss of the kth distribution branch, in kW; Udev,k is the voltage deviation rate of the kth distribution node, which is the ratio of the difference between the actual voltage and the rated voltage to the rated voltage; n is the total number of distribution branches, and m is the total number of distribution nodes; λ1 and λ2 are the weighting coefficients of network loss and voltage deviation, respectively.

7. A road network power emergency protection and automatic restoration system according to claim 6, characterized in that: The flexible interconnection device is composed of a combination of solid-state circuit breakers and soft switching equipment, and is deployed at each segment node of the linear distribution network. When a fault occurs, it can accurately isolate the fault segment within ≤20ms. The flexible reconfiguration and recovery module of the distribution network has a built-in adaptive reconfiguration and automatic recovery mechanism. The adaptive reconfiguration and automatic recovery mechanism covers the entire process of fault isolation, topology reconfiguration, and safety verification. After the topology reconfiguration is completed, an automatic overload control mechanism is provided to avoid secondary faults caused by line overload by adjusting the power output and load switching.

8. The road network power emergency protection and automatic restoration system according to claim 1, characterized in that: The extreme weather active protection module divides disaster risk zones using a risk assessment model, and the risk value is calculated using the following formula: In the formula, Rk is the meteorological disaster risk value of the kth power distribution section; Mk is the meteorological warning intensity coefficient of the corresponding section, which is taken as 0.2-1.0 according to the warning level of rainstorm, snow, strong wind, and lightning; Gk is the terrain vulnerability coefficient of the corresponding section, which is taken as 0.5-1.5 according to the historical data of road section altitude, slope, water accumulation / ice accumulation; μ is the scene correction coefficient, which is taken according to the type of overhead / cable laying of the line.

9. A road network power emergency protection and automatic restoration system according to claim 8, characterized in that: The extreme weather active protection module is configured with differentiated active protection strategies and a disaster resilience mechanism. The differentiated active protection strategies are configured according to the risk level of the section, including DC de-icing control for high-risk icing sections, pre-switching energy storage power supply for low-lying waterlogged sections, and encrypted monitoring and early warning for sections prone to tree obstruction. The disaster resilience mechanism includes low-power endurance mode for edge nodes, data linkage access for UAV inspections, and dynamic adjustment of emergency dispatch and topology reconstruction schemes to ensure continuous power supply to core primary loads under extreme disasters.

10. A road network power emergency protection and automatic restoration system according to claim 1, characterized in that: The cross-departmental closed-loop management platform connects the power operation and maintenance system, traffic control system, road administration management system, and emergency command system through standardized data interfaces. After a fault is triggered, it automatically links with the traffic management system to adjust traffic light timings and issue detour notices for the affected road sections. At the same time, it automatically generates emergency repair work orders and plans the optimal inspection and repair routes. After the repair is completed, it automatically verifies the power supply status and synchronizes information across all departments, thus constructing a fully visualized management and control system that includes fault early warning, location isolation, emergency power supply, network restoration, operation and maintenance repair, and acceptance.

Images