Smart City Power Distribution Network Intelligent Lightning Protection Early Warning and Adaptive Protection System

By adding an edge computing module and data interaction interface to existing smart surge arresters, multi-source data fusion and zoned adaptive surge protection are achieved, solving the problems of passive protection and low fault handling efficiency in existing surge protection systems, and improving the surge protection safety and power supply reliability of smart city power distribution networks.

CN123121441APending Publication Date: 2026-11-17SHANGHAI ZHAOBANG ELECTRIC FACILITY CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611220261.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-08-12
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

The existing passive protection mode of lightning arresters in the power distribution network is outdated, the adaptability of the unified protection strategy across the entire region is poor, the efficiency of manual handling of lightning strikes is low, and the power lightning protection system cannot be linked with the urban emergency system, which limits the improvement of power supply reliability in smart cities.

Method used

Based on the existing intelligent surge arrester equipment of Zhaobang, a lightweight upgrade is carried out, adding an edge computing module and a city-level data interaction interface to build a surge protection system with edge AI early warning, zone-adaptive dynamic protection and city-wide emergency linkage, realizing multi-source data fusion, zone-adaptive surge protection control and automatic fault linkage.

Benefits of technology

It achieves accurate lightning strike early warning within 15 to 30 minutes, differentiated lightning protection control by region, and automatic linkage handling of lightning strike faults, thereby improving the lightning protection safety and power supply stability of smart city power distribution networks and reducing fault handling time and equipment wear.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123121441A_ABST
    Figure CN123121441A_ABST
Patent Text Reader

Abstract

The Smart City Distribution Network Intelligent Lightning Protection Early Warning and Adaptive Protection System belongs to the field of intelligent lightning protection and urban power grid operation and maintenance. Based on the upgrade of existing intelligent surge arrester equipment from Zhaobang, an edge computing module and an interface for interaction with urban data are added, constructing a three-layer architecture of equipment perception, edge intelligent decision-making, and urban cloud collaboration. The system integrates multi-source data from meteorological radar, lightning monitoring, and power grid operation, and relies on an improved LSTM AI algorithm to achieve accurate prediction of lightning strike paths and intensity within 15-30 minutes, classifying multiple levels of lightning strike early warning. Addressing the differentiated lightning protection needs of urban CBDs, industrial areas, residential areas, and key load areas, the system remotely adaptively adjusts the surge arrester's action threshold, activating enhanced protection modes for critical facilities. After a lightning strike, the system can automatically locate the fault, calculate the impact of the power outage, and coordinate with the urban emergency platform to complete dispatching and repair orders. This solves the problems of passive protection, poor adaptability of the overall protection strategy, and delayed fault diagnosis inherent in traditional surge arresters, significantly improving the power supply reliability of smart city distribution networks.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to intelligent lightning protection technology and intelligent operation and maintenance technology of urban power distribution networks. Specifically, it is applicable to active lightning protection early warning, hierarchical adaptive overvoltage protection, and urban-wide emergency collaborative dispatch scenarios for 10kV / 0.4kV power distribution networks in multi-functional zoned smart cities. Background Technology

[0002] Zinc oxide surge arresters are core protective equipment for urban power distribution networks against lightning overvoltages. Existing mature applications include Zhaobang's self-developed intelligent detection device for zinc oxide surge arresters and fully insulated, separable post-type intelligent surge arresters. These devices can achieve online monitoring of surge arrester leakage current, valve body temperature, and number of operations. However, existing technical solutions still have significant technical shortcomings and cannot meet the refined, proactive, and collaborative lightning protection operation and maintenance needs of smart cities. Specific technical pain points are as follows: 1. Passive protection has strong limitations: Traditional surge arresters can only perform current discharge protection after lightning overvoltage occurs. They do not have the ability to predict lightning risk in advance, cannot identify the movement trajectory of thunderclouds in advance, and cannot estimate the intensity of lightning strikes. They are difficult to activate the protection enhancement mechanism before thunderstorms arrive, which can easily lead to damage to distribution network equipment, power outages of important loads and other faults.

[0003] 2. Poor universality and low adaptability of protection strategies: There are significant differences in load types, equipment withstand voltage levels, and lightning sensitivity across different functional zones within a city. The lightning protection requirements for high-density precision loads in the CBD, heavy-duty industrial equipment, ordinary loads in residential areas, and critical loads in transportation and healthcare vary considerably. Existing lightning protection systems use a uniform surge arrester activation threshold across the entire area, resulting in redundant protection in low-risk areas, increased equipment wear, and insufficient protection margin in high-risk, critical areas, leading to overall low protection efficiency.

[0004] 3. Low efficiency in fault handling and reliance on manual labor: After lightning tripping or surge arrester malfunction, the traditional operation and maintenance model relies on manual inspection of fault points segment by segment. Fault location, power outage range statistics, and impact assessment take a long time and cannot quickly connect with the city's emergency system. As a result, urban lifeline facilities such as transportation and medical facilities lack forward dispatch support, which seriously restricts the improvement of power supply reliability in smart cities.

[0005] 4. Severe data silos and lack of collaborative capabilities: Existing smart lightning arresters only achieve single-device status monitoring, with limited data collection dimensions. They cannot access external data such as weather radar and global lightning monitoring networks, resulting in low accuracy in lightning strike prediction and insufficient early warning. At the same time, there is a lack of standardized urban government interaction interfaces, and the power lightning protection system is isolated from the urban emergency dispatch system, making it impossible to achieve cross-industry and cross-departmental collaborative handling.

[0006] Existing technologies mostly focus on online monitoring of single devices or lightning warning from a single data source, and cannot simultaneously achieve the three core functions of high-precision lightning strike prediction, zoned adaptive dynamic protection, and city-wide emergency linkage. Furthermore, they cannot be upgraded at low cost based on existing smart surge arrester equipment, resulting in many technical defects such as low equipment reuse rate, fragmented system architecture, and lack of city-wide collaborative operation and maintenance capabilities. Summary of the Invention

[0007] 1. Purpose of the invention To address the technical shortcomings of existing power distribution network surge arresters, such as outdated passive protection modes, poor adaptability of unified protection strategies across the entire network, low efficiency of manual handling of lightning strikes, and the inability to link power lightning protection systems with urban emergency response systems, this invention provides a smart city power distribution network intelligent lightning protection early warning and adaptive protection system. This invention is based on a lightweight upgrade of existing Zhaobang intelligent surge arrester equipment, adding an edge computing module and a city-level data interaction interface to construct an integrated active lightning protection system of "edge AI advanced early warning + zoned adaptive dynamic protection + city-wide emergency linkage." This system achieves accurate lightning strike early warning within 15-30 minutes, differentiated lightning protection control by region, and automatic linkage handling of lightning strikes, effectively improving the lightning protection safety and power supply stability of smart city power distribution networks.

[0008] 2. Technical Solution To achieve the above-mentioned objectives, the present invention adopts the following technical solution: a smart city power distribution network intelligent lightning protection early warning and adaptive protection system, which adopts a hierarchical architecture design, consisting of a device perception layer, an edge intelligent decision-making layer, and a city cloud collaborative scheduling layer from top to bottom.

[0009] 1) The equipment sensing layer serves as the foundation for system data acquisition, reusing Zhaobang's mature intelligent surge arrester hardware, including intelligent detection devices for zinc oxide surge arresters and fully insulated separable pillar-type intelligent surge arresters, along with the deployment of power distribution network electrical acquisition units and atmospheric electric field monitoring terminals; the surge arresters have built-in Rogowski coil current sensing, temperature sensing, and leakage current acquisition circuits to collect valve body operating status and lightning strike pulse data in real time; the atmospheric electric field monitoring terminal collects near-ground thundercloud electric field change parameters in real time, providing auxiliary characteristic data for lightning strike prediction.

[0010] 2) The edge intelligent decision-making layer is the core decision-making unit of the system. It adds an edge computing module, a local storage unit, a surge arrester remote adjustment drive unit, and a multi-source data preprocessing unit. The edge computing module has a built-in lightweight improved LSTM lightning strike prediction model and a partitioned adaptive protection strategy library, which can independently complete data processing, risk assessment, and protection command generation. The edge computing module is connected to the surge arrester electronic control adjustment unit through the local bus to realize remote contactless dynamic control of the surge arrester action threshold. At the same time, it has the ability to operate autonomously when the network is disconnected, ensuring the continuous and stable operation of the system in the absence of network.

[0011] 3) The city cloud collaborative scheduling layer undertakes the functions of full-domain data aggregation and cross-platform linkage, including a lightning big data fusion server, a GIS power grid topology mapping module, a regional load characteristic database, and a standardized city emergency interaction interface; the cloud uniformly aggregates meteorological radar data, city lightning monitoring data, distribution network operation data, and city regional load data; the GIS module binds device geographic information and load attribute tags; the emergency interaction interface is compatible with power-specific protocols and general government protocols, and can stably connect to city emergency, transportation, and medical power supply monitoring platforms.

[0012] The complete operation flow of the system of this invention is as follows: Step 1: Synchronous Acquisition and Spatiotemporal Registration of Multi-Source Heterogeneous Data. The device sensing layer uploads the operating status of surge arresters and the voltage and current time-series data of the distribution network in real time. The city cloud collaborative scheduling layer synchronously accesses meteorological radar echoes, global lightning location data, and load tag data of urban functional zones. The spatiotemporal alignment of multi-source data is completed through NTP microsecond-level timestamps.

[0013] Step 2: Edge Data Fusion and AI-Powered Lightning Strike Early Warning. The edge computing module filters and reduces noise from heterogeneous data, normalizes features, inputs it into an improved LSTM prediction model, and outputs the cloud movement trajectory, lightning amplitude, and risk coverage area for the next 15-30 minutes. Combined with the dynamic CUSUM algorithm, multiple warning levels are defined, matching the lightning protection sensitivity characteristics of different regional loads.

[0014] Step 3: Dynamic Adjustment of Zonal Adaptive Lightning Protection Thresholds. The edge computing module, based on real-time warning levels and regional load attributes, invokes corresponding adaptive protection strategies. Standard protection strategies are applied to conventional loads in CBDs and industrial areas. For critical loads such as hospitals and rail transit, the action threshold is automatically lowered by 20%–35%, and the discharge response window is extended to enhance protection. For low-risk areas, the threshold is appropriately raised to reduce ineffective equipment losses.

[0015] Step 4: Intelligent Analysis and Collaboration of Lightning Strike Faults in Urban Emergency Response. After detecting a lightning strike fault signal, the system automatically locates the fault point based on GIS topology, simulates and calculates the power outage coverage, user scale, and duration of power outages for critical loads; and pushes fault reports, impact lists, and optimal repair routes to the urban emergency platform through standardized interfaces to support emergency dispatching of critical facilities such as urban transportation and medical facilities. At the same time, it automatically generates maintenance work orders and sends them to the repair terminals.

[0016] The method for constructing the time-series AI fusion prediction model includes: constructing a dedicated training dataset for urban lightning prediction, which includes historical thunderstorm samples, meteorological radar monitoring data for the corresponding time period, lightning fault ledgers of the power distribution network, and lightning frequency feature labels for each functional area over many years; using an improved LSTM time-series neural network to achieve deep fusion of three heterogeneous features: meteorology, lightning, and power grid operation, and outputting gridded lightning probability and lightning amplitude prediction results for the next 15 to 30 minutes; and establishing a four-level lightning warning system based on the dynamic CUSUM risk classification algorithm, namely, Level 1 low-risk warning, Level 2 medium-risk warning, Level 3 high-risk warning, and Level 4 extreme severe thunderstorm risk warning.

[0017] The aforementioned zoned adaptive protection strategy library pre-stores multiple sets of surge arrester dynamic threshold parameters adapted to different scenarios. The system divides the monitoring area into four protection zones according to urban functional zoning: CBD commercial load zone, industrial heavy load zone, ordinary residential zone, and medical and transportation first-level important load zone. Different warning levels correspond to independent surge arrester action threshold ranges. When a level 4 extreme thunderstorm risk warning is triggered, the system automatically lowers the surge arrester action voltage threshold of the corresponding area by 20% to 35% and simultaneously extends the discharge response window to achieve enhanced lightning protection for important load areas.

[0018] The aforementioned intelligent detection device for zinc oxide surge arresters and the fully insulated separable pillar-type intelligent surge arrester are based on the original acquisition circuit and have undergone intelligent upgrade and transformation. On the basis of the original hardware, an edge computing communication submodule and a remote threshold adjustment control unit are added, and a 5G / NB-IoT dual-mode communication interface is configured to realize bidirectional command interaction, status feedback and data transmission between the edge computing node and the intelligent surge arrester.

[0019] The edge computing module has the ability to operate autonomously locally when the network is interrupted. When the cloud communication link is interrupted, the edge computing module can independently complete short-term lightning risk prediction and lightning arrester adaptive protection adjustment based on locally cached historical meteorological data and lightning monitoring data. After the network communication is restored, the locally stored early warning records, equipment operation data and fault ledgers will be automatically synchronized to the city cloud collaborative scheduling layer in batches to ensure the uninterrupted and stable operation of the system.

[0020] The GIS power grid topology mapping module pre-binds the geographical coordinates, distribution network lines, and downstream load attribute tags of each smart surge arrester. While outputting lightning strike early warning information, the system can automatically generate a city-wide lightning strike risk heat map, visually presenting the lightning strike threat gradient in different areas, and providing intuitive data support for differentiated lightning protection.

[0021] The aforementioned urban emergency interaction interface supports dual-mode data output of the IEC60870-5 power-specific protocol and the JSON general government protocol. The pushed data content includes the precise location of the fault, the number of the affected distribution substation, the scale of the power outage users, the estimated power outage duration of key loads such as medical and rail transit, and the recommended optimal repair route information, meeting the two-way data interaction needs of power dispatch and urban emergency platform.

[0022] The atmospheric electric field monitoring terminal is deployed at the highest point in the city to collect real-time time-series data of near-ground atmospheric electrostatic field intensity. This data is used as an auxiliary input feature for AI prediction models, effectively improving the efficiency of identifying thundercloud anomalies and the accuracy of lightning strike warnings.

[0023] This invention also adapts to the intelligent lightning protection method of the above-mentioned systems, covering the entire business process of multi-source data preprocessing, AI lightning strike prediction, adaptive threshold control, urban emergency linkage, and model iterative optimization, forming a comprehensive protection system combining software and hardware.

[0024] 3. Beneficial effects Compared with the prior art, the present invention has the following beneficial effects: 1) Achieve proactive prediction of lightning strike risks, breaking through the limitations of traditional passive protection. This invention integrates multi-source heterogeneous big data with AI time-series prediction algorithms, which can accurately predict lightning strike risks 15 to 30 minutes in advance and activate differentiated protection enhancement mechanisms in advance, thereby reducing the probability of lightning strike failures in the power distribution network from the root.

[0025] 2) Achieve zoned adaptive dynamic protection to adapt to the differentiated load scenarios of smart cities. Based on the differences in load levels, withstand voltage characteristics, and lightning sensitivity in different functional areas of the city, dynamically adjust the surge arrester's operating threshold to avoid insufficient or redundant protection issues caused by uniform protection across the entire area, balancing power supply safety and equipment lifespan.

[0026] 3) Enables automated fault assessment and comprehensive emergency response, significantly improving fault handling efficiency. After a lightning strike, no manual on-site investigation is required; the system automatically completes location, assessment, dispatch, and cross-platform push notifications, supporting collaborative emergency dispatch of critical urban infrastructure and significantly improving the power supply reliability of smart cities.

[0027] 4) Based on existing mature equipment upgrades, the transformation cost is low and the equipment reuse rate is high. This invention is a lightweight transformation of Zhaobang's existing intelligent surge arrester hardware, only adding communication and electronic control adjustment modules, without replacing the main equipment, making it suitable for existing distribution network transformation projects and highly implementable.

[0028] 5) It has the capability to operate in both edge and cloud modes, ensuring strong system stability. Edge nodes support autonomous operation even when the network is down. They can independently perform lightning protection and control when the network is abnormal, and automatically synchronize data after reconnecting to the network, making them suitable for communication-deficient scenarios such as suburbs and tunnels.

[0029] 6) Construct a closed-loop iterative optimization system to continuously improve system adaptability. The system continuously accumulates operational sample data in the cloud, periodically iterating and optimizing the AI ​​prediction model and protection strategy library. This allows it to adapt to urban load expansion and regional functional iterations, achieving long-term adaptive optimization of the system. Attached Figure Description

[0030] Figure 1 This is a schematic diagram of the overall three-layer architecture of the system according to an embodiment of the present invention; Figure 2 A block diagram showing the internal module connections of the edge intelligent decision-making layer; Figure 3 Schematic diagram of the modified structure of the remote threshold adjustment electronic control unit for surge arresters; Figure 4 A sequence diagram of the complete lightning protection early warning and adaptive protection workflow of the system. Detailed Implementation

[0031] The technical solution of the present invention will be clearly and completely described below with reference to the accompanying drawings and specific embodiments. The following embodiments are only used to explain the technical solution of the present invention and are not intended to limit its scope of protection. This embodiment is adapted to a medium-sized smart city 10kV urban distribution network and completes the system deployment and application based on Zhaobang's existing intelligent surge arrester equipment.

[0032] System overall three-tier architecture diagram: as shown Figure 1 As shown, the overall hierarchical architecture of this system, from top to bottom, consists of the city cloud collaborative scheduling layer, the edge intelligent decision-making layer, and the equipment perception layer. The city cloud collaborative scheduling layer is used for multi-source data aggregation across the entire domain, risk topology modeling, and cross-platform emergency response; the edge intelligent decision-making layer is used for data preprocessing, AI-based lightning strike prediction, adaptive strategy decision-making, and equipment command issuance; and the equipment perception layer is used for comprehensive data collection on power grid operation, lightning protection equipment, and atmospheric environment. These three layers interact bidirectionally to form a closed-loop proactive lightning protection and control system.

[0033] Edge intelligent decision-making layer internal module connection diagram: The module topology of the system's core decision-making layer. The edge computing module is the core entity, communicating bidirectionally with the local storage unit, multi-source data preprocessing unit, and surge arrester remote adjustment drive unit. The multi-source data preprocessing unit is responsible for heterogeneous data noise reduction, registration, and normalization. The edge computing module generates control commands based on the built-in model and strategy library. The surge arrester remote adjustment drive unit is responsible for command issuance and device status feedback. The overall architecture can achieve dual-mode operation: network collaboration and offline autonomy.

[0034] Schematic diagram of the modified structure of the surge arrester remote threshold adjustment electronic control unit: (See diagram below) Figure 3As shown in the diagram, this is the intelligent upgrade structure of an existing smart surge arrester. The original sensing and acquisition circuit of the device remains unchanged, while a remote threshold adjustment control unit, a 5G / NB-IoT dual-mode communication submodule, and an edge interaction submodule are added. The control unit has a built-in dynamic gap adjustment component and threshold control chip, which can remotely adjust the surge arrester's action threshold and discharge window. The communication submodule enables bidirectional data interaction between the device and edge nodes, achieving a low-cost intelligent upgrade of existing equipment.

[0035] Complete lightning protection early warning and adaptive protection workflow sequence diagram: as follows Figure 4 As shown in the diagram, the system's full-service time sequence logic includes five major stages: multi-source data acquisition, data preprocessing and AI lightning strike early warning, zoned adaptive protection and control, lightning strike fault analysis and emergency response, and data closed-loop iterative optimization. It fully covers the entire process of lightning protection business system, including pre-event early warning, in-event protection, post-event handling, and long-term optimization.

[0036] Example 1: Layered Deployment and Implementation of System Hardware 1. Deployment of the device perception layer The existing intelligent detection devices for zinc oxide surge arresters and fully insulated separable pillar-type intelligent surge arresters in the urban distribution network are retained. Each device is equipped with a remote threshold adjustment electrical control unit and a 5G / NB-IoT dual-mode communication submodule. Electrical acquisition units are deployed at the segment locations of the distribution network lines to collect three-phase voltage and zero-sequence overvoltage time sequence data in real time. Atmospheric electric field monitoring terminals are deployed at the highest points in the city and on the roofs of substations, with a sampling period of 100ms, to collect near-ground atmospheric electric field parameters in real time.

[0037] The surge arrester's built-in sensing circuit collects resistive leakage current, valve body operating temperature, lightning pulse current amplitude, and discharge action count in real time. All raw sensing data is transmitted to the nearest edge computing node via an RS485 local bus. 2. Edge Intelligent Decision Layer Deployment Edge control units are divided according to the power distribution area of ​​the city's power distribution network. Each area is equipped with an industrial-grade edge computing gateway, with an 8G local storage unit to cache 7 days of full raw data. The gateway has a built-in lightweight improved LSTM lightning strike prediction model and a four-level early warning zone protection strategy library. It communicates bidirectionally with all smart surge arresters in the area through power line carrier or wireless Mesh network to realize command issuance and status feedback.

[0038] The edge gateway has autonomous operation capabilities during network outages. When cloud communication is interrupted, it can independently perform lightning strike risk prediction and protection control based on local cached data. After communication is restored, it automatically synchronizes historical data in batches. 3. Deployment of the city cloud collaborative scheduling layer The cloud-based system deploys a lightning big data fusion server, a GIS power grid topology server, and an urban emergency interactive gateway. The GIS system inputs the distribution network topology, equipment GPS coordinates, and transformer area load attributes, dividing the area into four major zones: CBD commercial area, industrial park, residential area, and first-level important protection load. The cloud-based system connects to the meteorological radar platform and the urban lightning positioning platform via a dedicated link to synchronize lightning monitoring data in real time.

[0039] The city emergency interactive gateway supports dual protocol output. It connects to the power dispatching master station via the IEC60870-5 protocol and connects to the city's emergency, traffic, and medical monitoring platforms via the JSON protocol, enabling standardized data push and coordinated dispatch.

[0040] Example 2: Operation Flow of AI Multi-Source Fusion Lightning Strike Early Warning Model This embodiment discloses the specific operation flow of the AI ​​multi-source fusion lightning strike early warning model: 1) Data preprocessing: The edge gateway performs spatiotemporal alignment on meteorological radar echoes, lightning location coordinates, atmospheric electric field time series, distribution network electrical parameters, and surge arrester operation data, and uses Kalman filtering algorithm to remove environmental electromagnetic noise and unify data dimensions and sampling frequency; 2) Model feature input: Five types of heterogeneous features are input into the lightweight improved LSTM neural network. The weight ratio of key features such as thundercloud movement and electric field mutation is strengthened through the attention mechanism to improve the model prediction accuracy. 3) Risk prediction output: The model outputs the gridded lightning strike probability, lightning current intensity and risk coverage for two time windows of 15 min and 30 min respectively; 4) Risk Classification: Based on historical lightning strike fault data for various functional areas of the city, a dynamic CUSUM algorithm is used to classify four warning levels: Level 1 Low Risk: Lightning strike probability <20%, standard surge arrester operation threshold maintained throughout the area; Level 2 Medium Risk: Probability of lightning strike 20%–50%, threshold lowered by 5% for industrial areas and CBD areas; Level 3 High Risk: Probability of lightning strike 50%–80%, industrial and commercial load threshold reduced by 15%, Level 1 Important Load threshold reduced by 25%; Level 4 extreme risk: probability of lightning strike > 80%, enhanced protection is activated across the entire area, and the threshold for Level 1 load areas is reduced by up to 35%.

[0041] Example 3: Dynamic Adjustment Control Flow of Zoned Adaptive Surge Arrester The edge gateway combines real-time early warning levels with GIS load zoning labels, matches them with a built-in adaptive protection strategy library, and generates differentiated threshold adjustment instructions. The specific control rules are as follows: 1) Ordinary residential areas: Level I and Level II warnings will maintain the standard threshold, Level III warning threshold will be reduced by 10%, and Level IV warning threshold will be reduced by 20%; 2) CBD commercial area and industrial park: Level II warning threshold reduced by 5%, Level III warning threshold reduced by 15%, and Level IV warning threshold reduced by 30%; 3) Hospitals, rail transit, urban data centers and other critical load areas: Level II and above warnings will directly activate the enhanced protection mode, reduce the action threshold by up to 35%, and extend the discharge conduction time to resist continuous lightning overvoltage impacts.

[0042] Control commands are sent to the surge arrester's electronic control unit via a wireless mesh network. The electronic control unit adjusts the dynamic gap parameters of the zinc oxide valve plate to achieve adaptive threshold adjustment. After adjustment, a confirmation signal is sent back to form a closed-loop control.

[0043] Example 4: Emergency Response Procedures for Lightning Strike Faults in Cities When the edge gateway identifies a surge arrester activation pulse or a line lightning overvoltage fault signal, it automatically initiates the fault linkage handling process: 1) Precise topology positioning: By combining equipment GPS coordinates, feeder topology, and transformer area load information, the fault range can be accurately located, and affected users and critical infrastructure locations can be marked; 2) Fault impact simulation: Based on the regional load database, automatically calculate the power outage coverage, estimate the recovery time, and the impact on key facilities; 3) Multi-platform standardized push: The emergency interaction gateway synchronously generates three types of messages: first, power dispatch messages, used for fault line isolation and power grid dispatch; second, urban emergency JSON briefings, including fault location, power outage scale, impact on critical loads, and optimal repair path; and third, operation and maintenance repair work orders, which are automatically sent to front-line operation and maintenance terminals. 4) Cross-domain emergency response: After receiving data, the city's emergency platform automatically triggers emergency operations such as traffic control on rail transit, switching of backup power to hospitals, and traffic management on roads to minimize the impact of lightning strikes on city operations.

[0044] Example 5: System Iterative Optimization Mechanism The city's cloud servers continuously store comprehensive data on meteorological thunderstorms, lightning activity, surge arrester operation, and distribution network faults throughout the year. The AI ​​prediction model undergoes monthly iterative training to optimize lightning prediction accuracy for different seasons and terrain scenarios. The trained, lightweight model is automatically deployed to each edge node for local updates. Simultaneously, the system iteratively optimizes the threshold parameters of the zoned protection strategy library based on annual lightning strike frequency and fault statistics, adapting to the lightning protection needs following urban new area construction, load expansion, and functional area updates, achieving long-term closed-loop self-optimization.

[0045] In summary, this invention presents an intelligent lightning protection solution with a three-layer system architecture, multi-source AI early warning, adaptive threshold control, and integrated urban emergency response technologies. Key technologies include AI prediction models, protection strategies, hardware upgrade structures, edge network outage autonomy, GIS topology mapping, and emergency interaction interfaces. It is comprehensive, stable, intelligent, and rigorous. It can be widely applied to the intelligent upgrade of existing 10kV distribution networks in cities and integrated lightning protection projects for newly built smart city distribution networks. Based on Zhaobang's existing mature surge arrester hardware system, the equipment boasts strong compatibility, low upgrade costs, and easy implementation. Simultaneously, it breaks down data barriers between power lightning protection systems and urban emergency dispatch platforms, aligning with the policy direction of new power systems and digital smart cities, and possesses extremely high engineering application value and promising prospects for industrialization.

Claims

1. A smart city power distribution network intelligent lightning protection early warning and adaptive protection system, characterized in that, include: Device perception layer, edge intelligent decision-making layer, and city cloud collaborative scheduling layer; The equipment sensing layer includes a zinc oxide surge arrester intelligent detection device, a fully insulated separable post-type intelligent surge arrester, a power distribution network electrical acquisition unit, and an atmospheric electric field monitoring terminal. Both the zinc oxide surge arrester intelligent detection device and the fully insulated separable post-type intelligent surge arrester have built-in Rogowski coil current sensors, valve plate temperature sensing modules, and leakage current acquisition circuits to collect real-time operating status parameters of the lightning protection equipment. The edge intelligent decision layer integrates an edge computing module, a local storage unit, a surge arrester remote adjustment drive unit, and a multi-source data local preprocessing unit. The edge computing module has a built-in trained AI lightning strike prediction model and a partitioned adaptive protection strategy library. The edge computing module establishes a communication connection with the surge arrester remote adjustment drive unit through a local bus to realize remote dynamic control of the surge arrester overvoltage action threshold. The city cloud collaborative scheduling layer includes a lightning big data fusion server, a GIS power grid topology mapping module, a city emergency interaction interface and a zoned load characteristic database. The city emergency interaction interface is standardized to connect to the city emergency management platform, the city traffic dispatch platform and the medical support power supply monitoring platform, so as to realize bidirectional communication between power lightning protection data and city emergency dispatch data. The system operation steps include: S1. Multi-source data synchronous acquisition: The equipment sensing layer collects the operating parameters of the surge arrester and the voltage and current time sequence operating data of the distribution network in real time. The city cloud collaborative scheduling layer synchronously accesses the meteorological radar echo data, the lightning strike coordinate and amplitude data of the city-wide lightning location monitoring network, and the city zone load label data. S2. Edge data fusion and AI lightning strike early warning: The edge computing module performs spatiotemporal registration and noise filtering on multi-source collected data, and inputs the pre-processed data into the time series AI fusion prediction model. It outputs the lightning strike prediction path, lightning current intensity and lightning strike risk coverage of the target area 15 to 30 minutes in advance, and divides the lightning strike early warning level into four levels based on the load lightning strike sensitivity characteristics of CBD, industrial area and residential area. S3, Zoned Adaptive Dynamic Lightning Protection: The edge computing module calls the zoned adaptive protection strategy library, matches the real-time warning level to generate differentiated threshold adjustment instructions, and sends them to the smart lightning arresters in the corresponding areas through the lightning arrester remote adjustment drive unit. The lightning arrester action threshold is lowered in high-risk areas, and the enhanced protection mode is activated in the first-level important load areas. S4. Comprehensive Lightning Fault Response: When a surge arrester activation signal or a line lightning overvoltage fault signal is detected, the system automatically locates the fault point, simulates and extrapolates the power outage load range, and calculates the power restoration time. It also pushes the fault information and load impact list to the city emergency management platform through the city emergency interaction interface, providing data support for the dispatch of key city facilities and power repair dispatch.

2. The intelligent lightning protection early warning and adaptive protection system for smart city power distribution networks according to claim 1, characterized in that, The method for constructing the time-series AI fusion prediction model includes: constructing a dedicated training dataset for urban lightning prediction, which includes historical thunderstorm samples, meteorological radar monitoring data for the corresponding time period, lightning fault ledgers of the power distribution network, and lightning frequency feature labels for each functional area over many years; using an improved LSTM time-series neural network to achieve deep fusion of three heterogeneous features: meteorology, lightning, and power grid operation, and outputting gridded lightning probability and lightning amplitude prediction results for the next 15 to 30 minutes; and establishing a four-level lightning warning system based on the dynamic CUSUM risk classification algorithm, namely, Level 1 low-risk warning, Level 2 medium-risk warning, Level 3 high-risk warning, and Level 4 extreme severe thunderstorm risk warning.

3. The intelligent lightning protection early warning and adaptive protection system for smart city power distribution networks according to claim 1, characterized in that: The partitioned adaptive protection strategy library pre-stores multiple sets of surge arrester dynamic threshold parameters adapted to different scenarios. The system divides the monitoring area into four protection zones according to urban functional zoning: CBD commercial load zone, industrial heavy load zone, ordinary residential zone, and medical and transportation first-level important load zone. Different warning levels correspond to independent surge arrester action threshold ranges. When a level 4 extreme thunderstorm risk warning is triggered, the system automatically lowers the surge arrester action voltage threshold of the corresponding area by 20% to 35% and simultaneously extends the discharge response window to achieve enhanced lightning protection for important load areas.

4. The intelligent lightning protection early warning and adaptive protection system for smart city power distribution networks according to claim 1, characterized in that: The zinc oxide surge arrester intelligent detection device and the fully insulated separable pillar-type intelligent surge arrester have undergone intelligent upgrades based on the original acquisition circuit. On the basis of the original hardware, an edge computing communication submodule and a remote threshold adjustment control unit have been added, and a 5G / NB-IoT dual-mode communication interface has been configured to realize bidirectional command interaction, status feedback and data transmission between the edge computing node and the intelligent surge arrester.

5. The intelligent lightning protection early warning and adaptive protection system for smart city power distribution networks according to claim 1, characterized in that: The edge computing module has the ability to operate autonomously locally when the network is interrupted. When the cloud communication link is interrupted, the edge computing module can independently complete short-term lightning risk prediction and lightning arrester adaptive protection adjustment based on locally cached historical meteorological data and lightning monitoring data. After the network communication is restored, the locally stored early warning records, equipment operation data and fault ledgers will be automatically synchronized to the city cloud collaborative scheduling layer in batches to ensure the uninterrupted and stable operation of the system.

6. The intelligent lightning protection early warning and adaptive protection system for smart city power distribution networks according to claim 1, characterized in that: The GIS power grid topology mapping module pre-binds the geographical coordinates, distribution network lines, and downstream load attribute tags of each smart surge arrester. While outputting lightning strike early warning information, the system can automatically generate a city-wide lightning strike risk heat map, visually presenting the lightning strike threat gradient in different areas, and providing intuitive data support for differentiated lightning protection.

7. The intelligent lightning protection early warning and adaptive protection system for smart city power distribution networks according to claim 1, characterized in that: The urban emergency interaction interface supports dual-mode data output of IEC60870-5 power-specific protocol and JSON general government protocol. The pushed data content includes the precise location of the fault, the number of the affected distribution substation, the scale of the power outage users, the estimated power outage duration of key loads such as medical and rail transit, and the recommended optimal repair route information, which meets the two-way data interaction needs of power dispatch and urban emergency platform.

8. The intelligent lightning protection early warning and adaptive protection system for smart city power distribution networks according to claim 1, characterized in that: The atmospheric electric field monitoring terminal is deployed at the highest point in the city to collect real-time time-series data of near-ground atmospheric electrostatic field intensity. This data is used as an auxiliary input feature for AI prediction models, effectively improving the efficiency of identifying thundercloud anomalies and the accuracy of lightning strike warnings.

9. A lightning protection method based on the intelligent lightning protection early warning and adaptive protection system for smart city power distribution networks according to any one of claims 1 to 8, characterized in that, Includes the following steps: Spatiotemporal alignment and preprocessing of multi-source heterogeneous data, edge AI-based advanced lightning strike prediction and analysis, adaptive control of threshold for zone-differentiated surge arresters, automatic assessment of lightning strike faults and linkage with urban emergency platforms, closed-loop inspection of surge arrester operating status and iterative updates of protection strategy library.