Ancient building lightning protection system and method based on lightning early warning device

Abstract

The invention discloses an ancient building lightning protection system and method based on a lightning early warning device, and belongs to the technical field of lightning protection systems, and the system comprises a sensing layer, a decision layer, an execution layer and a digital twin platform. The sensing layer collects multi-mode environmental parameters through an annular atmospheric electric field monitoring array, a broadband electromagnetic pulse sensor and a meteorological sensor. The decision-making layer fuses data and outputs four levels of risk levels through the risk prediction model of the ancient building LSTM combined attention mechanism, and generates a protection instruction; and the execution layer adopts a bionic hidden lightning receiving device (hidden in a non-working state), a passive corona field lightning driving device and a multi-stage historic building SPD, and responds according to a grading strategy. The digital twin platform constructs a high-precision model, and realizes data visualization and protection strategy simulation optimization. Passive lightning protection is converted into active intelligent protection, the situation that a traditional device destroys the style and appearance of an ancient building is avoided, secondary damage caused by lightning stroke is reduced, self-learning optimization can be achieved, and the requirements for safety and cultural relic protection are met.

Description

Technical Field

[0001] This invention relates to the field of lightning protection system technology, and in particular to a lightning protection system and method for ancient buildings based on lightning early warning devices. Background Technology

[0002] Currently, the main lightning protection technology used in the field of ancient building lightning protection is the traditional lightning protection system based on the Franklin lightning rod principle. This system consists of three parts: First, external lightning protection devices, represented by metal lightning rods installed at the highest point of the building, lightning strips or lightning nets laid along the roof ridge, which are connected to the grounding device through down conductors to directly intercept lightning current; second, multi-stage surge protectors (SPDs) installed in the building's internal power distribution and low-voltage systems to protect against overvoltages generated by lightning electromagnetic pulses; in addition, some modernization renovation projects also attempt to introduce early warning systems based on atmospheric electric field monitoring, or to perform biomimetic camouflage treatment on metal lightning rods to improve safety and landscape harmony. However, these technologies have not fundamentally broken through the paradigm of passive defense.

[0003] Despite the widespread application of existing technologies, four core shortcomings have been exposed in the specific context of ancient building preservation: First, while the passive lightning interception mechanism can guide lightning current to the ground, the intense lightning discharge process inevitably involves high-temperature arcs, electromagnetic impacts, and explosive pressures, still causing irreversible secondary damage such as burning and cracking to the precious wooden structures and glazed tile roofs of ancient buildings. Second, regardless of the landscaping methods used, fixed metal components will always alter the historical appearance and spatial outline of ancient buildings, fundamentally conflicting with the principles of "authenticity" and "minimal intervention" emphasized in cultural relic preservation. Third, existing early warning systems have low levels of intelligence, typically only providing simple threshold alarms and failing to achieve accurate lightning risk prediction, and lacking the ability to coordinate with protection enforcement agencies, significantly reducing their practical value. Fourth, there is a lack of system-level coordination among various lightning protection measures; direct lightning protection, induced lightning protection, and early warning systems are independent of each other, making it difficult to form an effective protective closed loop. These limitations clearly demonstrate that the current technological system cannot simultaneously meet the dual needs of ancient buildings for safety protection and architectural preservation. Summary of the Invention

[0004] The main objective of this invention is to provide a lightning protection system and method for ancient buildings based on a lightning early warning device, which can effectively solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows: Lightning protection systems and methods for ancient buildings based on lightning early warning devices include: The sensing layer is used to collect multimodal environmental parameters related to lightning activity in real time. The sensing layer includes an atmospheric electric field monitoring array, an electromagnetic pulse sensor, and a meteorological environment sensor deployed in the area surrounding the ancient building. The decision layer, which is communicatively connected to the perception layer, is used to receive and fuse the multimodal environmental parameters, and generate lightning risk levels and protection instructions through a built-in intelligent early warning algorithm. The decision layer includes a data processing unit, a lightning risk prediction model, and a protection decision unit. The execution layer, which is communicatively connected to the decision layer, is used to receive and execute the protection command. The execution layer includes a concealed lightning arrester that can be raised and lowered according to the command, a passive corona field lightning drive device, and a multi-level surge protection device. The concealed lightning protection device is hidden outside the ancient building or inside the biomimetic decorative components when not in operation, and rises to a preset height to form a lightning protection point according to the protection command when in operation.

[0006] The core of the aforementioned system lies in transforming traditional passive lightning protection into proactive and predictable intelligent protection through a closed-loop control logic of "perception-decision-execution". This structure ensures that the system can automatically make optimal decisions based on real-time data and deploy physical lightning protection devices only when necessary, thereby providing safety guarantees while minimizing the permanent impact on the visual landscape of ancient buildings.

[0007] Preferably, the atmospheric electric field monitoring array consists of multiple three-dimensional atmospheric electric field meters arranged in a ring around the ancient building, used to monitor the spatial distribution and changing trend of the electric field intensity; the electromagnetic pulse sensor is a broadband sensor used to monitor electromagnetic signals generated by long-distance lightning activity.

[0008] The ring-shaped atmospheric electric field monitoring array can form a three-dimensional monitoring network to accurately sense the charge center, movement direction, and intensity changes of thunderclouds; while the broadband electromagnetic pulse sensor is used to detect lightning activity hundreds of kilometers away. The combination of the two provides a multi-dimensional data foundation that combines long-range and short-range data for the early warning algorithm, significantly improving the accuracy and lead time of the early warning.

[0009] Preferably, the lightning risk prediction model is a multimodal data fusion intelligent early warning algorithm, configured as follows: ① Receive time-series data streams collected by atmospheric electric field monitoring array, electromagnetic pulse sensor and meteorological environment sensor; ② Extract multidimensional features including the spatiotemporal gradient of electric field intensity (▽E), electromagnetic pulse frequency and amplitude (SF_Count, SF_Amplitude), and the trend of meteorological parameter changes (dP / dt); ③ The multidimensional features are fused with the pre-stored structural feature parameters of ancient buildings and input into a machine learning model based on Long Short-Term Memory Network (LSTM) and Attention Mechanism for calculation; ④ Output early warning signals to indicate different risk levels, which are matched with the system's graded response strategy.

[0010] The above configuration does not simply set thresholds, but rather employs an advanced machine learning model (LSTM+Attention) to perform time-series analysis on the fused complex features, dynamically focusing on key indicators. By incorporating the structural parameters of the ancient buildings themselves, the algorithm achieves personalized risk assessment, ensuring that the early warning results are highly correlated with the actual vulnerability of the specific protected object, greatly reducing false alarms and false negatives.

[0011] Preferably, the protection decision unit is configured to execute a tiered response strategy: In response to a Level 1 risk level, a monitoring and early warning instruction will be generated. In response to a level 2 risk, a command is generated to activate the passive corona field lightning arrestor. In response to a level 3 risk, a command is generated to raise the concealed lightning arrester and activate the first level of surge protection; In response to a Level 4 risk level, a comprehensive protection command is generated, which includes putting the concealed lightning protection device in maximum protection mode, activating all surge protection devices, and activating the linked fire protection system.

[0012] The above strategy clarifies a refined response process based on risk level. This tiered strategy embodies a "combination of soft and hard" protection concept: for low risk, only early warning is issued; for medium risk, passive lightning arresters are activated for "soft" intervention to avoid direct lightning strikes; only when the risk is extremely high is the lightning protection device activated for "hard" protection. This progressive response optimizes resource utilization, reduces equipment wear and tear, and minimizes unnecessary mechanical actions.

[0013] Preferably, the concealed lightning arrester includes: The lifting mechanism adopts either an electric screw or a hydraulic transmission method. Lightning arresters are designed to resemble biomimetic components that harmonize with the style of ancient architecture, including roof ornaments, finials, or pagoda finials. After receiving a lifting command, the lightning arrester moves from its hidden position to a working position at a preset height.

[0014] The lifting mechanism ensures the concealment and mobility of the device, allowing it to be completely hidden during the non-rainy season. The biomimetic design of the lightning arrester integrates it with the artistic style of ancient architecture, fundamentally solving the visual pollution problem caused by traditional metal components and achieving a unity of functionality and the aesthetic value of cultural relics.

[0015] Preferably, it also includes a digital twin platform, the platform being configured as follows: Construct a high-precision three-dimensional model of the ancient building and the lightning protection system; The real-time data collected by the perception layer is mapped onto the three-dimensional model for visualization. Virtual simulations are performed based on historical data and models to optimize the layout and protection strategies of lightning protection devices.

[0016] This platform not only provides an intuitive visual monitoring interface, but more importantly, it can perform simulations based on historical data to verify and optimize lightning protection strategies, such as the optimal installation location and height of lightning arresters. This allows for repeated trial and error in a virtual space, improving the scientific nature of the overall system design and protection efficiency.

[0017] A lightning protection method for ancient buildings based on a lightning early warning device includes: S1. Collect multimodal parameters of the environment surrounding the ancient building through multiple types of sensors in the perception layer; S2. The multimodal parameters are transmitted to the decision-making layer for data fusion and analysis, and the current and future lightning risk levels are calculated using the lightning risk prediction model; S3. Generate corresponding protection instructions based on the risk level; S4. Send the protection command to the execution layer to control the corresponding lightning protection device to perform actions. When the risk level reaches a preset threshold, control the concealed lightning protection device to rise from the hidden state to the working state.

[0018] The core contribution of this method lies in describing an automated process from data acquisition to intelligent decision-making and then to physical execution, realizing closed-loop management of lightning protection and transforming traditional decentralized and passive measures into an integrated, proactive, and intelligent organic whole.

[0019] Preferably, the calculation of risk level using a lightning risk prediction model specifically includes: ① Processing temporal multimodal parameter data based on Long Short-Term Memory (LSTM) network models; ② Employ an attention mechanism to focus on key characteristic parameters strongly correlated with lightning formation; ③ The warning threshold is dynamically adjusted based on the structural characteristics and environmental features of the ancient buildings, and the risk prediction results are output in a graded manner.

[0020] LSTM models excel at processing time-series data and can effectively learn the historical patterns of lightning formation; the attention mechanism allows the model to dynamically weight the importance of different input features, for example, paying more attention to changes in the electric field rather than humidity under specific meteorological conditions; dynamic threshold adjustment ensures customized early warnings for different ancient buildings, improving the adaptability and accuracy of the method.

[0021] Preferably, the control of the corresponding lightning protection device to perform actions adopts a graded response strategy, including: ① In the case of Level II medium risk, the passive corona field lightning suppression device should be activated first for active intervention; ②After the risk level rises to Level 3 (high risk), the concealed lightning protection device will be raised and preparations for lightning protection will be made. ③ After the lightning activity ends, control the concealed lightning arrester to descend and reset to a hidden state.

[0022] Initial intervention is achieved by prioritizing low-energy, non-damaging lightning protection techniques. If this fails, final physical protection is then activated. This sequence greatly protects the lightning arresting device and minimizes the potential impact of energy leakage on ancient buildings. The device automatically resets upon completion of the task, ensuring its concealment.

[0023] Preferably, the method further includes: ① Retrospectively analyze and process historical lightning strikes and system response through a digital twin platform; ② Continuously optimize the parameters of the lightning risk prediction model and the threshold of the graded response strategy based on the analysis results.

[0024] By reviewing each protection incident using digital twin technology, the system can accumulate experience data and continuously fine-tune the algorithm model and response strategy thresholds using this data, forming a virtuous cycle that becomes more accurate with use, demonstrating the system's evolutionary characteristics.

[0025] Compared with the prior art, the present invention has the following beneficial effects: This invention organically combines multimodal perception fusion, intelligent early warning algorithms, and a liftable concealed lightning protection device to construct an active hierarchical protection system. This achieves a fundamental shift from passive lightning protection to active early warning and protection. While significantly improving the reliability of lightning protection for ancient buildings, it completely avoids the visual damage to the historical appearance caused by traditional fixed lightning protection devices, thus achieving the dual goals of safety protection and preservation of the original appearance of cultural relics. Attached Figure Description

[0026] Figure 1 This is a schematic diagram of the system architecture and process of the present invention. Detailed Implementation

[0027] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments.

[0028] like Figure 1 The system architecture and process diagram shown below will be explained in conjunction with a detailed implementation example.

[0029] I. Specific Configuration and Deployment of Each Level of the System (a) Perception Layer: Multimodal Environmental Parameter Acquisition System The sensing layer, through a "combined long- and short-range, three-dimensional coverage" layout, collects parameters related to lightning activity in real time. The specific configuration is as follows: Atmospheric electric field monitoring array Centered on the ancient building, six DAT-500 three-dimensional atmospheric electric field meters (measurement range 0-50kV / m, accuracy ±5%, sampling frequency 1 time / second) are evenly deployed in a ring with a radius of 50m. These meters are mounted on six protective fence posts around the tower (2.5m high to avoid obstruction). This array calculates the spatiotemporal gradient (▽E) of the electric field intensity by measuring the electric field strength in the X, Y, and Z directions. For example, when thunderclouds approach from the northwest, the electric field meter in that direction first detects ▽E increasing from 0.5kV / m to 1.2kV / m, with an increase of 0.3kV / m every 10 minutes, allowing for accurate determination of the thundercloud's direction and intensity.

[0030] Principle explanation: Thunderclouds, being charged, generate an induced electric field on the ground. The three-dimensional electric field meter captures changes in the spatial electric field through miniature electric field sensors. Combined with data fusion from multiple devices in a ring layout, it can eliminate local interference from a single sensor (such as the electric field influence of trees and power lines), and achieve three-dimensional positioning of the thundercloud's charge center and movement speed.

[0031] Electromagnetic pulse sensor An EMI-3000 wideband electromagnetic pulse sensor (bandwidth 1kHz-1GHz, sensitivity 1mV / μV, sampling frequency 1 time / 0.1 seconds) was selected and deployed in the northeast and northwest directions of the tower (where thunderstorms often intrude from the northwest to northeast), 30m from the tower base, on a dedicated cement base (1.8m high, away from metal components). When a lightning strike occurs 20km away, the sensor can capture the electromagnetic pulse signal and record its frequency (SF_Count) and amplitude (SF_Amplitude). For example, a long-distance lightning strike will produce an SF_Count of 3 times / 10 minutes and an SF_Amplitude of 0.8V, while a lightning strike within 10km will cause the SF_Count to rise to 15 times / 10 minutes and the SF_Amplitude to exceed 2V.

[0032] Explanation of the principle: Lightning discharge is essentially a strong current pulse that radiates a wide-band electromagnetic signal. The signal propagation distance is inversely proportional to the amplitude (the closer the distance, the greater the amplitude), while the frequency is positively correlated with the intensity of lightning activity. These two parameters can be used to provide early warning of distant lightning 15-30 minutes in advance.

[0033] Meteorological and environmental sensors A WS-600 multi-parameter meteorological sensor is deployed at a meteorological observation station 20m south of the tower (avoiding the tower's shadow). It simultaneously collects data on temperature (-40~60℃, accuracy ±0.2℃), relative humidity (0-100%RH, accuracy ±3%), air pressure (800-1100hPa, accuracy ±0.5hPa), and wind speed (0-60m / s, accuracy ±0.3m / s), with a sampling frequency of 1 time per minute. The core monitoring parameter is the rate of change of air pressure (dP / dt). For example, one hour before a thunderstorm, the air pressure drops from 1012hPa to 1008hPa, dP / dt = -0.067hPa / min (i.e., -4hPa / h), which helps determine the development trend of thunderclouds.

[0034] (II) Decision-making level: Intelligent early warning and protection decision-making system The decision-making level is located in the cultural relics monitoring room at the base of the tower, and consists of three levels of units: "data processing - risk prediction - protection decision-making". The specific functions are as follows: Data processing unit The multimodal data received from the perception layer is processed through a three-step process: noise reduction, alignment, and fusion. Noise reduction: Kalman filtering algorithm is used to remove abnormal data, such as removing abnormal instantaneous wind speed values ​​generated by meteorological sensors due to strong winds (such as suddenly soaring to 20m / s and then immediately falling back). Alignment: Unify data from different sampling frequencies to a timestamp of 1 time / second (e.g., interpolate meteorological data from 1 time / minute to 1 time / second); Integration: Construct a data matrix, with each row containing 7 dimensions of parameters: "timestamp - ▽E - SF_Count - SF_Amplitude - dP / dt - temperature - humidity", to provide standardized input for risk prediction.

[0035] Lightning risk prediction model The machine learning model employs an "LSTM + attention mechanism," and the model training data includes: Basic data: Lightning observation data for the past 10 years provided by the local meteorological bureau (including the time, location, and intensity of 500 lightning strikes); Customized data: 3 years of historical monitoring data around the ancient building (including complete parameters of 20 thunderstorm events, 2 of which were actual lightning strikes); Model structure and parameter settings: Input layer: 6 key features (▽E, SF_Count, SF_Amplitude, dP / dt, humidity change rate, wind speed); Hidden layers: 3-layer LSTM network, 64 neurons per layer, activation function is ReLU, dropout coefficient is 0.2 (to prevent overfitting); Attention layer: An additive attention mechanism is used to dynamically allocate feature weights, such as ▽E (40%), SF_Count (30%), dP / dt (20%), and SF_Amplitude (10%) (because the electric field gradient and pulse frequency have the strongest correlation with lightning strike prediction). Output layer: 4 risk levels (Level 1: Low risk; Level 2: Medium risk; Level 3: High risk; Level 4: Extremely high risk), and output the predicted duration of the risk (e.g., "Level 3 risk will last for 40 minutes").

[0036] Example of risk level threshold:

[0037] Protection Decision Unit Based on the output of the risk prediction model, protection commands matching the graded response strategy are automatically generated. The command format adopts "target device-action-parameter-trigger condition", for example: Level 2 Risk Instruction: "Passive Corona Field Lightning Reduction Device - Start - Corona Current 10mA - Continuously Monitor ▽E Changes"; Level 4 risk instruction: "Concealed lightning protection device - maintain 1.5m elevation - activate Level 3 SPD - link fire suppression system to standby - provide status feedback every 5 seconds"; Meanwhile, the instructions are transmitted to the execution layer via a 4G industrial router with a transmission delay of ≤0.5 seconds, ensuring timely response.

[0038] (III) Execution layer: graded protection execution device The execution layer adopts a "bionic concealment + hierarchical action" design, and all devices are integrated with the tower's appearance. The specific configuration is as follows: Concealed lightning protection device Lifting mechanism: JS-100 type electric screw (rated load 50kg, lifting speed 50mm / s, stroke 1.5m, protection level IP65, suitable for outdoor thunderstorm environment) is selected and installed on the steel structure support inside the tower top spire (the support is hidden inside the copper shell of the spire and is not exposed). Lightning arrester: Designed to resemble the "dragon's kiss" at the top of the pagoda (60cm long, 40cm wide, and 20cm high), made of T2 copper (3mm thick, with excellent conductivity), with an antique bronze oxidation treatment on the surface (color consistent with the original pagoda, color difference ≤ΔE=1.5). Concealed and operational status switching: When not in operation, the lightning arrester is completely retracted into the cavity at the bottom of the tower spire (cavity dimensions: 65cm × 45cm × 25cm, inner wall lined with insulating material to prevent electrical conduction with the tower spire); upon receiving a level 3 or higher risk command, the electric screw drives the lightning arrester to rise at a speed of 50mm / s, reaching the preset height after 30 seconds (total tower height rises to 46.5m, meeting the lightning protection standard that "the lightning arrester is 1.5m higher than the protected object"), and simultaneously, through the down conductor (concealed inside the central column of the tower, using 60mm...),... 2 The copper cable is connected to the grounding device (buried 2m below the tower base, consisting of 6 2.5m long galvanized angle steels, with a grounding resistance ≤4Ω).

[0039] Principle explanation: The lightning arrester conducts lightning current to the ground through "priority lightning reception", avoiding lightning current from hitting the brick and wood structure of the tower; the biomimetic design and concealed installation completely solve the problem of traditional metal lightning arrester rods damaging the appearance of the tower.

[0040] Passive corona field lightning arrestor Four CD-200 passive corona field lightning suppression devices (requiring no external power supply and triggered by the thundercloud electric field) were selected and installed at the top of the four corner ridges of the tower (hidden and fixed under the glazed tiles, with only 12 tungsten alloy corona needles exposed). The corona needles are 15cm long, spaced 5cm apart, and have a tip curvature radius of 0.1mm (facilitating corona discharge). Upon receiving a level 2 or higher risk command, the device automatically activates. Under the influence of the thundercloud electric field, the corona needles release positive charges, neutralizing the negative charges at the bottom of the thundercloud, reducing the electric field strength around the tower by 30%-50% (e.g., ▽E from 2.5kV / m to 1.2kV / m), thereby suppressing lightning strikes.

[0041] Multi-stage surge protection device (SPD) The system is deployed in a three-tiered manner ("central-sub-design") to adapt to the protection needs of different equipment within the tower. Level 1 (Main Distribution): Install a surge protector in the main distribution box at the bottom of the tower to protect against strong induced lightning strikes from the power supply line; Second level (power distribution): Install secondary surge protectors in the lighting distribution boxes on the 3rd, 5th and 7th floors of the tower to protect the middle layer lines from induced lightning. Level 3 (Equipment End): Install Level 3 surge protectors at the power supply ends of the 12 surveillance cameras and temperature and humidity sensors inside the tower to protect sensitive electronic equipment; Each level of SPD has a status indication function (green: normal; red: fault), and fault signals are fed back to the decision-making level in real time, which facilitates operation and maintenance.

[0042] (iv) Digital Twin Platform A digital twin platform for ancient buildings and lightning protection systems was constructed using Unity3D+BIM technology. The specific functions are as follows: High-precision 3D modeling The model uses laser scanning technology (scanning accuracy ±0.1mm) to acquire complete structural data of the tower. The modeling scope includes: the brick and wood structure of the tower (the size and position of each nanmu beam and column), glazed tiles (the texture and color of each tile), and lightning protection system components (the installation position and size of sensors, lightning arresters, lightning expulsion devices, and SPDs). The model has a total of 5 million faces and can clearly display the lifting trajectory of the lightning arrester and the distribution of the corona needles of the lightning expulsion device.

[0043] Real-time data mapping Sensor data from the perception layer is displayed on the model using a "dynamic annotation + color encoding" method: Atmospheric electric field data: Real-time ▽E values ​​are marked at each electric field meter position in the ring array, and the intensity is represented by color (green: <1kV / m; yellow: 1-3kV / m; orange: 3-5kV / m; red: ≥5kV / m). Lightning arrestor status: The process of raising and lowering the lightning arrester is displayed in an animation. When it rises, the "dragon kiss" shape on the model moves upward and the height of the rise and fall is displayed in sync (e.g., "Current height 1.2m, remaining 0.3m"). SPD Status: Green / red dots indicate whether the SPD is normal or faulty at the location of the distribution box in the model. Clicking with the mouse will display parameters such as discharge current and response time.

[0044] Virtual simulation and optimization Simulation of protection effect: Simulate thunderstorms of different intensities (such as "once-in-10-year lightning strike" and "once-in-50-year lightning strike") to test the optimal raising height of the lightning arresting device (simulation verification shows that the lightning arresting efficiency at a height of 1.5m is 25% higher than that at 1m) and the number of lightning protection devices installed (4 units expand the lightning protection range by 40% compared to 2 units). Strategy optimization: Based on simulation and backtracking results, the model parameters are adjusted. For example, the dP / dt threshold for secondary risk is adjusted from -1hPa / h to -0.8hPa / h, so that the lightning protection device can be activated 10 minutes earlier, further reducing the risk of lightning strikes.

[0045] II. Complete Implementation Process of Lightning Protection Methods (Taking the Thunderstorm Process on June 15, 2024 as an Example) Step 1: Multimodal parameter acquisition At 2:00 PM one day, thunderstorms occurred in the Jiangnan region, and all sensors in the perception layer started real-time data collection: Atmospheric electric field monitoring array: The electric field meter in the northwest direction detected ▽E=1.2kV / m, and the increase was 0.3kV / m every 5 minutes; ▽E in other directions was between 0.8-1.0kV / m. Electromagnetic pulse sensor: The sensor in the northeast direction detected SF_Count = 6 times / hour, SF_Amplitude = 0.9V (indicating that lightning is 30-50km away); Meteorological environment sensor: monitored dP / dt=-0.6hPa / h, humidity=78%RH, wind speed=3.5m / s; All data is transmitted to the decision-making level in real time via the 4G network, with a stable transmission rate of 1Mbps and no data loss.

[0046] Step 2: Data Fusion and Risk Level Calculation The data processing unit at the decision-making level performs noise reduction (removing one electric field anomaly caused by a bird strike), alignment (unifying to a timestamp of 1 time / second), and fusion on the received parameters before inputting them into the lightning risk prediction model. The LSTM network processed the time series data of the first 30 minutes and found that ▽E showed a linear upward trend and SF_Count gradually increased, indicating that the thunderstorm cloud was continuously approaching. The attention mechanism assigns high weights to ▽E (weight 40%) and SF_Count (weight 30%). Combined with the structural parameters of ancient buildings, such as "brick and wood structure, susceptible to lightning strikes", the warning threshold is dynamically adjusted (the humidity threshold is reduced from 85% to 80% because the conductivity of wooden components increases after they are damp, so protection is needed in advance). The model outputs a risk level of Level 2 (medium risk) and predicts that the risk will last for 35 minutes.

[0047] Step 3: Generate and issue protection commands Based on the Level 2 risk level, the protection decision-making unit generates the following instructions: Command 1: "Passive corona field lightning dispersal device - start - continuously monitor ▽E changes - provide status feedback every 10 seconds"; Instruction 2: "Digital twin platform - update risk level to level two - mark electric field data in yellow"; The command was sent to the execution layer via a 4G industrial router. At 14:02, all four lightning protection devices were activated, and the corona needles began to release positive charges. The digital twin platform updated the model display synchronously, and the corner ridge of the tower was marked with "Lightning protection device activated".

[0048] Step 4: Perform protective actions and make dynamic adjustments At 14:30, the perception layer detected parameter changes: ▽E=3.6kV / m, SF_Count=19 times / hour, dP / dt=-1.3hPa / h, humidity=88%RH. After the data was transmitted to the decision layer, the model output risk level rose to level three (high risk), and the protection decision unit generated new instructions: Command 1: "Concealed lightning arrestor - activate electric lead screw - rise 1.5m - reach position within 30 seconds"; Command 2: "Level 1 surge protection device - activate - enter protection mode"; Instruction 3: "Digital twin platform - display animation of lightning arrester raising and lowering - mark the current risk level as Level 3"; At 14:30:05, the electric screw of the execution layer started, and the lightning arrester rose from the tower spire cavity at a speed of 50mm / s. At 14:30:35, it reached a height of 1.5m on time, and the down conductor formed a circuit with the grounding device. At the same time, the surge protector of the main distribution box at the bottom of the tower started, and the status indicator light turned green (normal).

[0049] At 14:55, the sensing layer detected extreme parameters: ▽E=6.2kV / m, SF_Count=32 times / hour, dP / dt=-2.6hPa / h, humidity=96%RH. The model output risk level rose to level four (extremely high risk), and the protection decision unit generated a comprehensive protection command. Instruction 1: "Concealed lightning arrestor - maintain a height of 1.5m - descent prohibited"; Command 2: "Level 2 surge protector and Level 3 surge protector - all activated - status feedback every 5 seconds"; Command 3: "Tower fire protection system (smoke detectors, sprinkler system) - enter standby mode - trigger threshold drops to 50℃"; At 14:55:03, all surge protectors at all levels were activated, and the smoke detectors in the fire protection system entered "high sensitivity" mode; the digital twin platform marked the electric field data in red, and the location of the lightning arrester was displayed as "maximum protection status".

[0050] Step 5: Risk Relief and Device Reset At 15:20, the perception layer detected the movement of the thunderstorm cloud: ▽E=0.7kV / m, SF_Count=2 times / hour, dP / dt=-0.3hPa / h, humidity=72%RH. The model output risk level dropped to level one (low risk), and the protection decision unit generated a reset command. Command 1: "Passive corona field lightning suppression device - shut down - stop corona discharge"; Command 2: "Concealed lightning arrestor - electric screw reverses - descends to tower brake cavity - resets and locks"; Instruction 3: "Level 2 and Level 3 surge protectors - shut down - keep Level 1 surge protector on standby - fire suppression system deactivated"; 15:20:10, lightning arrestor shuts down; 15:20:40, lightning arrester fully resets to tower brake cavity, electric lead screw locks; 15:21, secondary and tertiary surge protectors shut down, fire protection system returns to normal mode.

[0051] Step 6: System Self-Learning and Optimization After the thunderstorm ended, the digital twin platform conducted a retrospective analysis of the event: Data extraction: Extract all sensor data (43,200 records in total), system response commands (12 records in total), and device action records from 14:00 to 15:20 (lightning arrester lifting time: 30 seconds; surge protector activation response time: 8 ns). Effect evaluation: Through simulation comparison of the differences between "activating the lightning protection device" and "not activating the lightning protection device", it was found that the electric field strength around the tower body was reduced by 45% after activation, reducing 3 potential small-scale lightning strikes; Parameter optimization: Based on the analysis results, the attention weight of the lightning risk prediction model was adjusted (the weight of SF_Amplitude was increased from 10% to 15%, because this parameter is more accurate in judging the distance to lightning in this event), and the ▽E threshold of level 3 risk was adjusted from 3kV / m to 2.8kV / m, so that the lightning protection device would start 2 seconds earlier. The optimized parameters are automatically updated to the decision-making level, providing more accurate support for subsequent lightning protection work.

[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 this invention is defined by the appended claims and their equivalents.

Claims

1. A lightning protection system for ancient buildings based on a lightning early warning device, characterized in that: include: The sensing layer is used to collect multimodal environmental parameters related to lightning activity in real time. The sensing layer includes an atmospheric electric field monitoring array, an electromagnetic pulse sensor, and a meteorological environment sensor deployed in the area surrounding the ancient building. The decision layer, which is communicatively connected to the perception layer, is used to receive and fuse the multimodal environmental parameters, and generate lightning risk levels and protection instructions through a built-in intelligent early warning algorithm. The decision layer includes a data processing unit, a lightning risk prediction model, and a protection decision unit. The execution layer, which is communicatively connected to the decision layer, is used to receive and execute the protection command. The execution layer includes a concealed lightning arrester that can be raised and lowered according to the command, a passive corona field lightning drive device, and a multi-level surge protection device. The concealed lightning protection device is hidden outside the ancient building or inside the biomimetic decorative components when not in operation, and rises to a preset height to form a lightning protection point according to the protection command when in operation.

2. The ancient building lightning protection system based on a lightning early warning device according to claim 1, characterized in that: The atmospheric electric field monitoring array consists of multiple three-dimensional atmospheric electric field meters arranged in a ring around the ancient building, used to monitor the spatial distribution and variation trend of electric field intensity; the electromagnetic pulse sensor is a broadband sensor used to monitor electromagnetic signals generated by long-distance lightning activity.

3. The ancient building lightning protection system based on a lightning early warning device according to claim 1, characterized in that: The lightning risk prediction model is a multimodal data fusion intelligent early warning algorithm, configured as follows: ① Receive time-series data streams collected by atmospheric electric field monitoring array, electromagnetic pulse sensor and meteorological environment sensor; ② Extract multidimensional features including the spatiotemporal gradient of electric field intensity (▽E), electromagnetic pulse frequency and amplitude (SF_Count, SF_Amplitude), and the trend of meteorological parameter changes (dP / dt); ③ The multidimensional features are fused with the pre-stored structural feature parameters of ancient buildings and input into a machine learning model based on Long Short-Term Memory Network (LSTM) and Attention Mechanism for calculation; ④ Output early warning signals to indicate different risk levels, which are matched with the system's graded response strategy.

4. The ancient building lightning protection system based on a lightning early warning device according to claim 1, characterized in that: The protection decision unit is configured to execute a tiered response strategy: In response to a Level 1 risk level, a monitoring and early warning instruction will be generated. In response to a level 2 risk, a command is generated to activate the passive corona field lightning arrestor. In response to a level 3 risk, a command is generated to raise the concealed lightning arrester and activate the first level of surge protection; In response to a Level 4 risk level, a comprehensive protection command is generated, which includes putting the concealed lightning protection device in maximum protection mode, activating all surge protection devices, and activating the linked fire protection system.

5. The ancient building lightning protection system based on a lightning early warning device according to claim 1, characterized in that: The concealed lightning arrester includes: The lifting mechanism adopts either an electric screw or a hydraulic transmission method. Lightning arresters are designed to resemble biomimetic components that harmonize with the style of ancient architecture, including roof ornaments, finials, or pagoda finials. After receiving a lifting command, the lightning arrester moves from its hidden position to a working position at a preset height.

6. The ancient building lightning protection system based on a lightning early warning device according to claim 1, characterized in that: It also includes a digital twin platform, which is configured as follows: Construct a high-precision three-dimensional model of the ancient building and the lightning protection system; The real-time data collected by the perception layer is mapped onto the three-dimensional model for visualization. Virtual simulations are performed based on historical data and models to optimize the layout and protection strategies of lightning protection devices.

7. A lightning protection method for ancient buildings based on a lightning early warning device, characterized in that: Applied to the system as described in any one of claims 1-6, the method comprises: S1. Collect multimodal parameters of the environment surrounding the ancient building through multiple types of sensors in the perception layer; S2. The multimodal parameters are transmitted to the decision-making layer for data fusion and analysis, and the current and future lightning risk levels are calculated using the lightning risk prediction model; S3. Generate corresponding protection instructions based on the risk level; S4. Send the protection command to the execution layer to control the corresponding lightning protection device to perform actions. When the risk level reaches a preset threshold, control the concealed lightning protection device to rise from the hidden state to the working state.

8. The lightning protection method for ancient buildings based on a lightning early warning device according to claim 7, characterized in that: The calculation of risk level using a lightning risk prediction model specifically includes: ① Processing temporal multimodal parameter data based on Long Short-Term Memory (LSTM) network models; ② Employ an attention mechanism to focus on key characteristic parameters strongly correlated with lightning formation; ③ The warning threshold is dynamically adjusted based on the structural characteristics and environmental features of the ancient buildings, and the risk prediction results are output in a graded manner.

9. The lightning protection method for ancient buildings based on a lightning early warning device according to claim 7, characterized in that: The corresponding lightning protection device is controlled to perform actions using a tiered response strategy, including: ① In the case of Level II medium risk, the passive corona field lightning suppression device should be activated first for active intervention; ②After the risk level rises to Level 3 (high risk), the concealed lightning protection device will be raised and preparations for lightning protection will be made. ③ After the lightning activity ends, control the concealed lightning arrester to descend and reset to a hidden state.

10. The lightning protection method for ancient buildings based on a lightning early warning device according to claim 7, characterized in that: The method further includes: ① Retrospectively analyze and process historical lightning strikes and system response through a digital twin platform; ② Continuously optimize the parameters of the lightning risk prediction model and the threshold of the graded response strategy based on the analysis results.

Images