Bird risk intelligent early warning and dynamic-static coordination prevention and control system for power grid

Abstract

The present application relates to the technical field of power grid bird damage intelligent prevention and control, in particular to a power grid bird-related risk intelligent early warning and dynamic-static cooperative prevention and control system, which comprises: a panoramic sensing unit that collects power grid environment bird activity data; an intelligent decision unit that performs multi-dimensional feature analysis on the data to identify bird species and behavior and generates a hierarchical driving strategy based on risk levels; a cooperative execution unit that executes dynamic driving according to the strategy to dispatch laser and sound wave equipment and combines static physical barrier components to build a protection closed loop. In the present application, a far-middle-near three-layer defense system is built by integrating multi-modal non-harmful driving methods and physical protection facilities, effectively solving the problem of traditional single bird prevention device being easily adapted by birds and having a blind area of protection, reducing operation and maintenance costs while avoiding physical harm to birds, and realizing the coordinated development of power grid safe operation and ecological protection.

Description

Technical Field

[0001] This invention relates to the field of intelligent bird hazard prevention and control technology for power grids, and in particular to an intelligent early warning and dynamic-static coordinated prevention and control system for bird-related risks in power grids. Background Technology

[0002] Intelligent bird control in power grids refers to the use of physical, chemical, or electronic technologies to prevent birds from nesting, perching, or defecating on power lines and substation equipment, thereby ensuring the insulation safety and power supply reliability of the power grid. This field mainly involves the protection of power facilities such as transmission line towers and substation portal frames. Traditional protection technologies primarily rely on static physical devices such as fixed bird spikes and bird barriers, or simple wind-driven reflectors and single-frequency ultrasonic bird deterrents, reducing bird dwelling on tower materials by occupying bird activity space or emitting simple fright signals.

[0003] Existing technologies have significant drawbacks in practical operation. Static physical devices have limited functionality, making them highly adaptable to birds, who may even use bird spikes as "steel" supports for nests, leading to protection failure and increasing the risk of short circuits. Furthermore, traditional devices are mostly passive defenses, lacking the ability to actively detect and warn of approaching birds, making effective intervention impossible before they cause harm. Moreover, relying on single, strong stimuli to repel birds can easily harm them, violating ecological protection principles. In addition, existing equipment has poor weather resistance under complex weather conditions, requires frequent maintenance, resulting in high operation and maintenance costs, and is unable to meet the needs of large-scale, refined power grid control. Summary of the Invention

[0004] To address the technical problems existing in the prior art, embodiments of the present invention provide an intelligent early warning and dynamic-static coordinated prevention and control system for bird-related risks in power grids. The technical solution is as follows: On the one hand, an intelligent early warning and dynamic-static coordinated prevention and control system for bird-related risks in power grids is provided. This system includes: The panoramic perception unit is configured to acquire environmental data and motion trajectory data of target objects within the power grid monitoring area in real time through a radar detection module and a visual acquisition module, and generate multi-source perception signals. The intelligent decision-making unit is configured to receive the multi-source sensing signals, extract and compare the features of the target object, determine whether the target object is a bird that is involved in the crime and its specific species and behavior pattern, calculate the risk level based on the determination result, and then construct a hierarchical collaborative prevention and control strategy. The collaborative execution unit is configured to drive the dynamic drive-away module to emit photoacoustic interference signals of different intensities based on the hierarchical collaborative prevention and control strategy, and combine it with static physical barrier components to continuously occupy the key charged area, thus constructing a dynamic and static combined protection closed loop. The remote control unit is configured to connect the intelligent decision-making unit and the collaborative execution unit through a communication network, display the protection status in real time, and receive manual control instructions to correct the hierarchical collaborative prevention and control strategy.

[0005] As a further aspect of the present invention, the judgment logic of the intelligent decision-making unit for the target object specifically includes: A pre-built feature database containing the morphology, texture, and flight posture of various birds involved in harm; The collected image data is input into a feature extraction network to extract the shape contour and motion feature vector of the target object, and the similarity value between it and the samples in the feature database is calculated. When the similarity value exceeds a preset threshold, the target object is confirmed to be a bird that is causing harm, and non-target interference signals such as fallen leaves or insects are removed. By combining distance data detected by radar, the spatial location of the birds involved in the incident is determined through a coordinate mapping algorithm, generating early warning information that includes direction, distance, and species.

[0006] As a further aspect of the present invention, the dynamic removal module includes an intelligent laser emitting component, the execution logic of which is as follows: Receive bird location coordinates and species information sent by the intelligent decision-making unit; Control the gimbal mechanism to rotate horizontally and vertically, so that the laser beam is aimed at the area where the birds are located; It emits a green laser beam with a wavelength of 532nm and adjusts the beam projection direction according to the movement trajectory of the birds to generate visual stimulation signals within the field of vision of the birds involved in the incident. The power of the green laser beam is limited to a non-harmful range, and the beam projection mode supports switching between point scanning and line scanning.

[0007] As a further aspect of the present invention, the dynamic removal module further includes an acoustic wave coordination component, the execution logic of which is as follows: The integrated radar detection module has a detection range of 3 to 8 meters. Based on the hierarchical collaborative prevention and control strategy, the ultrasonic transmitter is activated to generate ultrasonic signals with a dynamic frequency between 5 kHz and 35.04 kHz. Synchronously trigger the loudspeaker player to play preset predator calls or warning sounds, with the volume intensity controlled below 120dB. By intermittently combining the ultrasonic signal with the audio from the loudspeaker, the auditory adaptation of birds is disrupted.

[0008] As a further aspect of the present invention, the static physical barrier component includes a self-opening and retractable bird spike and a rolling bird-proof wrap-around spike, wherein the structural features of the self-opening and retractable bird spike are as follows: It includes a gravity-driven mechanical connection structure and several needle units; Under normal protection conditions, the barbed wire units are distributed in an expanded manner, covering the nesting space above the insulator string; When a physical lifting action is received during maintenance, the needle unit is driven to retract inward using gravity and lever principles, and the opening angle can be adjusted between 120 and 150 degrees. After the maintenance is completed and the lifting force is released, the needle unit automatically resets to the deployed protective state; The structural features of the rolling bird-proof thorn wrapping are as follows: The rolling bird-proof barbed wire is positioned 0 to 1.5 meters from the tower, providing protection for a range of ≥0.3 meters, with a barbed length of ≥50 mm.

[0009] As a further aspect of the present invention, the static physical barrier component further includes a retractable bird-proof pin plate, the retractable bird-proof pin plate having the following structural features: It adopts a hinged telescopic structure, and its length adjustment range covers 0.5 meters to 1.2 meters; The base is equipped with both magnetic interface and U-shaped buckle for fastening, which can be used to adapt to crossarms of poles made of different materials; The needle plate is evenly covered with corrosion-resistant steel needles. The length and spacing of the steel needles are configured to block the landing points of large birds, and the ends of the steel needles have a blunt chamfered structure.

[0010] As a further aspect of the present invention, a bird-proof windmill is also included, wherein the structural features of the bird-proof windmill are as follows: Made of ABS engineering plastic, the rotating diameter is ≥ 30cm and the height is ≥ 35cm. The bird-proof windmill is equipped with windmill blades with reflective color design. It can rotate 360° under wind power. The top of the bird-proof windmill is equipped with wind blade spikes, and the wind blade spikes are equipped with bearings. After a bird lands, it will rotate due to gravity.

[0011] As a further aspect of the present invention, the dynamic and static coordination logic of the cooperative execution unit specifically includes: When the radar detects birds in the long-range warning zone, the dynamic deterrence module is activated first to provide light and sound warnings. When birds break through long-range defenses and enter the medium-range deterrence zone, increase the laser scanning frequency and sound wave intensity; When dynamic deterrence measures fail or birds attempt to stay at key locations on the pole, the static physical barrier components are used to cut off their nesting or perching paths. The integrity of the physical defense line is monitored by sensors on the static physical barrier components. If the components are found to be displaced or damaged, a maintenance request is sent to the remote control unit.

[0012] As a further aspect of the present invention, the system also includes an energy management unit; the energy management unit includes: Monocrystalline silicon solar photovoltaic panels are used to convert light energy into electrical energy; Lithium iron phosphate battery packs are used to store electrical energy and power the system; The low-temperature preheating module is configured to automatically start when the ambient temperature is lower than a preset low-temperature threshold to heat and keep the lithium iron phosphate battery pack warm, ensuring the system's continuous operation capability in low-temperature environments.

[0013] As a further embodiment of the present invention, the remote control unit also integrates an edge computing node; the edge computing node is configured to preprocess the multi-source sensing signals and perform lightweight model inference locally, and only upload the identification results and abnormal alarm data to the cloud server to reduce communication latency and bandwidth usage.

[0014] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following: By constructing a collaborative prevention and control system integrating intelligent sensing, dynamic bird removal, and static barrier, proactive defense and precise management of bird damage to the power grid have been achieved. Its core advantages lie in: utilizing AI recognition and multimodal bird removal technology to break the birds' adaptability without harming them, significantly improving the success rate of bird removal; combining retractable and adjustable static physical components to solve the problem of blind spots in protection and resolve the conflict between protection and maintenance; and achieving all-weather real-time monitoring and strategy optimization through edge computing and cloud collaboration. Overall, this solution significantly reduces power outages and maintenance costs caused by bird damage while also being eco-friendly and improving the intelligence and reliability of power grid operation. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a schematic diagram of the intelligent early warning and dynamic-static coordinated prevention and control system for bird-related risks in power grids provided in an embodiment of the present invention. Detailed Implementation

[0017] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0018] This invention provides an intelligent early warning and dynamic-static coordinated prevention and control system for bird-related risks in power grids, such as... Figure 1The diagram shown illustrates an intelligent early warning and dynamic-static coordinated prevention and control system for bird-related risks in power grids. This system includes: The panoramic perception unit is configured to acquire environmental data and motion trajectory data of target objects within the power grid monitoring area in real time through a radar detection module and a visual acquisition module, and generate multi-source perception signals. The panoramic perception unit transmits a frequency-modulated continuous wave signal (FM signal) of 77 GHz to 81 GHz to the power grid monitoring area via a millimeter-wave radar detection module. This signal uses sawtooth wave modulation and is transmitted via an antenna array covering a detection field of view with a horizontal azimuth angle of 120 degrees and an elevation angle of 30 degrees. The radar receiving antenna array receives the echo signal reflected from the target object. A mixer is used to mix the transmitted and received signals to obtain an intermediate frequency (IF) signal. Subsequently, an analog-to-digital converter (ADC) converts the IF signal into a digital signal at a sampling rate of 20 MHz. The digital signal undergoes a fast Fourier transform (FFT) to analyze the target object's distance, relative velocity, and azimuth angle data. The distance resolution is better than 0.5 meters, and the velocity resolution is better than 0.2 meters per second. Simultaneously, an industrial-grade high-definition camera in the visual acquisition module acquires visible light image data at a frame rate of 60 frames per second. The image resolution is configured as 3840 pixels multiplied by 2160 pixels. The image sensor uses a back-illuminated CMOS image sensor. During the acquisition process, the visual acquisition module reads the timestamp signal from the radar detection module and triggers the sensor via hardware. The system achieves microsecond-level time synchronization, aligning radar point cloud data and visual image data in the time dimension. It utilizes a coordinate transformation matrix to project 3D point cloud data from the radar coordinate system onto a 2D image plane in the camera coordinate system. The origin of the radar coordinate system is located at the center of the radar antenna, and the origin of the camera coordinate system is located at the optical center. Through calibration, an extrinsic parameter matrix is ​​obtained, mapping the spatial position of the target detected by the radar to image pixel coordinates, thereby generating a multi-source sensing signal containing both 3D spatial and 2D texture information of the target. Regarding environmental data acquisition, the ISP image signal processor built into the visual acquisition module automatically calculates the histogram distribution of the image and the average brightness of the current environment. When the average brightness is below 15 lux, it automatically switches to an infrared filter to enter night vision acquisition mode. Simultaneously, the radar detection module continuously outputs clutter maps of the environmental background as reference data for subsequent static interference removal. The data flow throughout the entire sensing process is uninterrupted. The generated raw sensing data is transmitted to the subsequent processing unit via a gigabit Ethernet interface, with the data packets containing uniform class fields and check bits.

[0019] The intelligent decision-making unit is configured to receive multi-source sensing signals, extract and compare the features of the target object, determine whether the target object is a bird that is involved in the crime and its specific species and behavior pattern, calculate the risk level based on the determination result, and then construct a hierarchical and collaborative prevention and control strategy. The intelligent decision-making unit's judgment logic for the target object specifically includes: A pre-built feature database containing the morphology, texture, and flight posture of various birds involved in harm; The collected image data is input into the feature extraction network to extract the shape contour and motion feature vector of the target object, and calculate the similarity value between it and the samples in the feature database. When the similarity value exceeds the preset threshold, the target object is confirmed to be a bird that is causing harm, and non-target interference signals such as fallen leaves or insects are removed. By combining distance data detected by radar, the spatial location of the birds involved in the incident is determined through a coordinate mapping algorithm, and early warning information including direction, distance and species is generated. The intelligent decision-making unit first accesses a pre-set feature database, which stores multi-dimensional feature data of 50 typical bird species involved in harm, including Oriental White Stork, Grey Heron, and Magpie. Each bird species contains no fewer than 1000 sample images from different angles, lighting conditions, and flight postures, along with corresponding wingspan ranges, flight speed ranges, and feature texture vectors. The unit receives a raw image of 3840 x 2160 pixels from the visual acquisition module, scales it to a standardized input size of 640 x 640 pixels using bilinear interpolation, and performs Z-score normalization, subtracting the mean from each pixel value and dividing by the standard deviation to ensure the data distribution is near 0. The processed image data is then input into a deep convolutional neural network for feature extraction. This network consists of an input layer, five convolutional blocks, and fully connected layers. The first convolutional block contains 64 kernels of size 7 x 7 with a stride of 2, using LeakyReLU activation to extract primary edge and texture features. The subsequent four convolutional blocks contain 128, 2... The system uses 56, 512, and 1024 convolutional kernels of size 3x3, and downsamples the feature map size to 20x20 pixels through max pooling layers between each block. The final layer outputs a 2048-dimensional feature vector through a global average pooling layer and a fully connected layer. The system then calculates the cosine similarity between this feature vector and the pre-stored standard feature vectors in the feature database. The calculation logic is as follows: divide the dot product of the real-time feature vector and the database vector by the product of their magnitudes to obtain a similarity between 0 and 1. The similarity value between them is used to determine the target. The threshold is set at 0.85, which is based on the ROC curve analysis of 100,000 historical identification records. This threshold is the optimal working point selected while ensuring a 99% recall rate. When the calculated similarity value is greater than or equal to 0.85, the target is determined to be a bird that is involved in a crime. For example, if the real-time feature vector is A and the database sample vector is B, and the dot product of the two is 0.9, the modulus of A is 1, and the modulus of B is 1, then the calculation result is 0.9 / 1=0.9, and 0.9 is greater than 0.85, the judgment result is true; for non-target interference signals, the system uses the speed data provided by the radar and the target area features extracted by vision for joint filtering. The upper limit of the area threshold for fallen leaves is set to 50 square centimeters and the upper limit of the speed threshold is 2 meters per second. The upper limit of the area threshold for insects is set to 10 square centimeters. When the parameters of the detected target meet any of the above interference conditions, the signal of the target is directly set to zero and removed; after confirming that the target is a bird that is causing harm, the system uses the Kalman filter algorithm to smooth the distance data measured by the radar. Combined with the intrinsic parameter matrix of the camera, the centroid coordinates of the bird in the image are projected backward to the world coordinate system to calculate the three-dimensional spatial coordinates (X, Y, Z) of the bird relative to the tower, where X represents the horizontal distance, Y represents the vertical height, and Z represents the vertical depth; The risk level is calculated based on three factors: bird species hazard coefficient, distance, and behavioral pattern. A weighted summation method is used, with the species hazard coefficient weighted at 0.4, distance weighted at 0.4, and behavioral pattern weighted at 0.2. The species hazard coefficient is assigned a value based on the size of the bird involved: 0.9 for large birds, 0.6 for medium-sized birds, and 0.3 for small birds. The distance factor is calculated based on the radar ranging value R: 1.0 for R less than 50 meters, 0.6 for 50 to 100 meters, and 0.2 for greater than 100 meters. In terms of behavioral patterns, circling is assigned a value of 0.5, and diving is assigned a value of 0.9. For example, if a large grey heron (coefficient 0.9) is detected diving (coefficient 0.9) at a distance of 40 meters (coefficient 1.0), the risk level is calculated as: 0.9 + 0.4 + 1.0 + 0.4 + 0.9. 0.2 = 0.94, and based on this, the system generates early warning information that includes location, distance, and extremely high risk level.

[0020] The collaborative execution unit is configured to drive the dynamic drive-away module to emit photoacoustic interference signals of different intensities based on a hierarchical collaborative prevention and control strategy, and combine with static physical barrier components to continuously occupy key charged areas, thus constructing a dynamic and static combined protection closed loop. The dynamic decoy module includes an intelligent laser emitting component, whose execution logic is as follows: Receive bird location coordinates and species information sent by the intelligent decision-making unit; Control the gimbal mechanism to rotate horizontally and vertically, so that the laser beam is aimed at the area where the birds are located; It emits a green laser beam with a wavelength of 532nm and adjusts the beam projection direction according to the movement trajectory of the birds to generate visual stimulation signals within the field of vision of the birds involved in the incident. The power of the green laser beam is limited to a non-harmful range, and the beam projection mode supports switching between point scanning and line scanning. The dynamic decoupling module also includes an acoustic wave coordination component, the execution logic of which is as follows: The integrated radar detection module has a detection range of 3 to 8 meters. Based on the hierarchical and collaborative prevention and control strategy, it activates the ultrasonic transmitter to generate ultrasonic signals with a dynamic frequency between 5 kHz and 35.04 kHz. Synchronously trigger the loudspeaker player to play preset predator calls or warning sounds, with the volume intensity controlled below 120dB. By intermittently combining ultrasonic signals with audio from a loudspeaker, the auditory adaptation of birds is disrupted. The static physical barrier components include self-opening and retractable bird spikes, the structural features of which are: It includes a gravity-driven mechanical connection structure and several needle units; Under normal protection conditions, the barbed wire units are distributed in an expanded manner, covering the nesting space above the insulator string; When a physical lifting action is received during maintenance, the needle unit is driven to retract inward using gravity and lever principles, and the opening angle can be adjusted between 120 and 150 degrees. After the maintenance is completed and the lifting force is released, the needle unit automatically resets to the deployed protective state; The static physical barrier components also include a retractable bird-proof pin plate, the structural features of which are: It adopts a hinged telescopic structure, and its length adjustment range covers 0.5 meters to 1.2 meters; The base is equipped with both magnetic interface and U-shaped buckle for fastening, which can be used to adapt to crossarms of poles made of different materials; The needle plate surface is evenly covered with anti-corrosion steel needles. The length and spacing of the steel needles are configured to block the landing points of large birds, and the ends of the steel needles have a blunt chamfered structure. The static physical barrier component also includes rolling bird-proof thorns, the structural features of which are: The rolling bird-proof spiked conductor is located 0 to 1.5 meters from the tower, with a protection range of ≥0.3m and a spike length of ≥50mm. It also includes bird-proof windmills, whose structural features are: Made of ABS engineering plastic, with a rotation diameter of ≥ 30cm and a height of ≥ 35cm, the bird-proof windmill is equipped with windmill blades. The windmill blades have a reflective color design and can rotate 360° under wind power. The top of the bird-proof windmill is equipped with wind blade spikes, and the wind blade spikes have bearings inside. After a bird lands, it will rotate due to gravity. The dynamic and static coordination logic of the coordinated execution unit specifically includes: When the radar detects birds in the long-range warning zone, the dynamic deterrence module is activated first to provide light and sound warnings. When birds break through long-range defenses and enter the medium-range deterrence zone, increase the laser scanning frequency and sound wave intensity; When dynamic deterrence methods fail or birds attempt to stay at key locations on the pole, static physical barriers are used to cut off their nesting or perching paths. The integrity of the physical defense line is monitored by sensors on the static physical barrier components. If the components are found to be displaced or damaged, a maintenance request is sent to the remote control unit. The system integrates a radar detection module, which performs real-time scanning of the target airspace with a detection range of 3 to 8 meters. This module is used for supplementary detection and distance verification of near-field flying targets, and transmits the detection results to the intelligent decision-making unit for risk level assessment. Upon receiving the warning information from the intelligent decision-making unit, the collaborative execution unit immediately analyzes the distance and risk level data. When the distance data falls within the 100-300 meter range, the target is determined to be in the long-range warning zone. The unit then controls the gimbal stepper motor of the intelligent laser emission component at a 0.5-degree angle. The unit precisely adjusts the horizontal and vertical angles, aligning a 532nm green laser beam with the airspace where the bird is located. A point scan mode is activated, with the laser power set to 30 milliwatts and the scanning frequency to 10 Hz, creating a non-harmful, strong light spot stimulus on the bird's retina. Simultaneously, the acoustic coordinating component is activated, driving the ultrasonic transmitter to emit a 20kHz fixed-frequency ultrasonic wave. When the distance data narrows to the 30-100 meter range, the target is determined to have entered the mid-range deterrence zone. The unit immediately issues an enhancement command, switching the laser scanning mode to line scan to construct a grating barrier. The laser power is increased to 45 milliwatts, and the scanning frequency is increased to 50 Hz. Simultaneously, the acoustic coordination component enters frequency conversion mode, controlling the DDS direct digital frequency synthesizer to randomly hop frequencies between 5 kHz and 35.04 kHz with a hopping interval of 0.5 seconds. This triggers a high-intensity sound player to play a peregrine falcon call at a sound pressure level of 115 dB, with audio data retrieved from a pre-stored memory chip. When the distance is less than 30 meters, or the intelligent decision-making unit determines that the bird's speed has dropped below 0.5 meters per second (i.e., attempting to perch), it is considered a high-risk situation at close range. At this point, the static physical barrier group... The components function as follows: The rolling bird-proof barbed wire is installed as follows: Within 0 to 1.5 meters of the conductor and ground wire near the tower, the barbed wire body is fixed to the conductor and ground wire using clamp / clamp connectors, so that the barbed wire forms a continuous wrap along the line; the outer surface of the barbed wire is arranged with rolling barbed wire units, the barbed wire length is not less than 50 mm, and through a rotatable structure, it generates rolling and barbed interference when birds attempt to land or grasp; the effective protective coverage length of a single set of barbed wire is not less than 0.3 meters, and if necessary, the protection range can be extended by connecting multiple sections end to end to cover key areas near the tower where birds are prone to perching.The self-opening and retractable bird spikes are installed above the insulator string. Their internal gravity-driven mechanism keeps the spike unit extended at a 135-degree angle without external force, forming a physical barrier surface with a diameter of 600 mm. The main body length of the retractable bird spike plate is adjusted and fixed at 0.8 meters, covering the key landing points of the tower crossarm. The steel needles evenly distributed on the plate are 150 mm high and spaced 50 mm apart, effectively disrupting the gripping points of birds. For maintenance, when maintenance personnel apply an upward physical pulling force to the self-opening and retractable bird spikes using the insulated operating rod, the linkage mechanism in the mechanical connection structure drives the spike unit to retract inward, reducing the overall diameter to 200 mm, clearing a maintenance passage. After maintenance, releasing the pulling force causes the gravity slider to fall, automatically resetting the spike unit to the extended state. Micro-motion on the components... The switch sensor monitors the deployment angle of the bird spike unit in real time. When the deployment angle is less than 90 degrees and there is no maintenance command, it is determined that the component is damaged or stuck due to ice accumulation, and a maintenance request signal is immediately generated. A specific example of the dynamic-static coordination logic is as follows: Assuming the radar detection range is 80 meters, the system determines that it is in a medium-range deterrence zone and automatically calculates the laser and acoustic parameters, setting the laser power to 45mW and the acoustic intensity to 115dB. If the subsequent detection range becomes 25 meters, the system determines that the dynamic deterrence has failed. At this time, although the dynamic module continues to operate, the center of gravity of protection shifts to the space occupied by the static component, and the system records the static component's in-situ status signal as "normal". If the sensor reports an abnormal bird spike angle signal, an abnormality code is reported through the communication module. Throughout the coordination process, the system refreshes the control command every 100 milliseconds.

[0021] The remote control unit is configured to connect the intelligent decision-making unit and the collaborative execution unit through a communication network, display the protection status in real time and receive manual control instructions to correct the hierarchical collaborative prevention and control strategy. The remote control unit also integrates edge computing nodes; the edge computing nodes are configured to preprocess multi-source sensing signals and perform lightweight model inference locally, and only upload the identification results and abnormal alarm data to the cloud server to reduce communication latency and bandwidth consumption. The remote control unit, through its built-in industrial-grade communication module, supports dual-link backup of 4G / 5G wireless networks and fiber optic wired networks, establishing a TCP / IP encrypted communication channel with the cloud server. Simultaneously, it connects to the intelligent decision-making unit and collaborative execution unit via an RS485 fieldbus, polling the working status registers of each unit in real time, with a polling period set to 200 milliseconds. The edge computing node integrated within the unit is equipped with an embedded AI acceleration chip, configured with a computing power of 5 TOPS, specifically designed for running lightweight deep learning models. The execution logic of the edge computing node first processes the received... Multi-source sensing signals undergo preprocessing, converting the raw YUV video stream acquired by the vision module to RGB format and adjusting the resolution to 320 x 320 pixels to suit the input requirements of the lightweight model. Subsequently, the node loads the lightweight inference model, which has undergone model pruning and INT8 quantization, compressing the model parameters from 32-bit floating-point numbers to 8-bit integers. This reduces the model size from 200MB to less than 50MB, with inference latency controlled to within 30 milliseconds. During inference, the node only extracts the confidence score and target bounding box coordinates of the decision result. When the inference result is displayed... When an image is labeled "potentially harmful bird" with a confidence level higher than 0.85, an abnormal alarm upload process is triggered. The edge node performs JPEG compression on the current frame's image data, sets the quality factor to 80, and overlays a timestamp, latitude and longitude coordinates, and bird species tags, generating an alarm data packet of approximately 50KB. The system only uploads this alarm data packet to the cloud. For background video streams without targets or detection results with a confidence level below the threshold, only the data from the last 24 hours is stored in the local circular buffer, without consuming uplink bandwidth. This mechanism reduces the average network upload speed from 5Mbps required for real-time video streams to... With speeds below 50kbps, communication latency and traffic costs are significantly reduced. The remote control unit is also responsible for receiving manual control instructions from the cloud. For example, when maintenance personnel modify the risk level parameter of a certain bird species on a remote terminal, the control unit extracts the new parameter value by parsing the instruction packet (such as adjusting the risk coefficient of the heron from 0.6 to 0.8) and writes the parameter into the configuration storage area of ​​the intelligent decision unit via the bus to complete the online correction of the strategy. If a communication link interruption is detected, the unit will automatically attempt to reconnect. The reconnection interval is executed according to the exponential backoff algorithm to ensure the reliability of the connection under network fluctuations.

[0022] The system also includes an energy management unit; the energy management unit includes: Monocrystalline silicon solar photovoltaic panels are used to convert light energy into electrical energy; Lithium iron phosphate battery packs are used to store electrical energy and power the system; The low-temperature preheating module is configured to automatically start when the ambient temperature is lower than the preset low-temperature threshold to heat and keep the lithium iron phosphate battery pack warm, ensuring the system's continuous operation capability in low-temperature environments. The energy management unit connects the monocrystalline silicon solar photovoltaic panel and the lithium iron phosphate battery pack via a maximum power point tracking (MPPT) controller, achieving efficient energy conversion and storage. The monocrystalline silicon solar photovoltaic panel has a peak power configuration of 200 watts and a photoelectric conversion efficiency of 21%. Its surface is covered with high-transmittance tempered glass, enabling it to withstand harsh outdoor environments. The lithium iron phosphate battery pack consists of 40 3.2-volt cells arranged in 4 series and 10 parallel configurations, with a rated voltage of 12.8 volts and a total capacity of 100 amp-hours. It integrates a battery management system (BMS) to monitor the voltage, current, and temperature of each cell. The low-temperature preheating module includes PTC heating elements attached around the battery pack and a high-precision temperature sensor. The temperature sensor has a measurement accuracy of ±0.5 degrees Celsius. The energy management unit's execution logic specifically includes: the system collects ambient temperature and battery pack internal temperature in real-time at 1-second intervals, setting the low-temperature preheating start threshold to -10 degrees Celsius and the stop threshold to 5 degrees Celsius; when the temperature sensor reading is below -10 degrees Celsius, the BMS system closes the heating control relay, driving the PTC heating element to heat the battery pack at a constant power of 50 watts until the temperature rises back to 5 degrees Celsius, at which point the relay is disconnected, forming a temperature hysteresis control closed loop to prevent frequent heater start-stop due to frequent temperature fluctuations; regarding the energy acquisition and consumption balance calculation logic, the system sets the light... The photovoltaic panel has an effective daily power generation duration of 5 hours and a system comprehensive efficiency coefficient (including line loss and conversion loss) of 0.8. Therefore, the daily power generation of the photovoltaic panel is calculated as follows: multiply the photovoltaic panel power of 200 watts by the effective sunshine duration of 5 hours and the efficiency coefficient of 0.8, i.e., 200 multiplied by 5 multiplied by 0.8, resulting in a daily power generation of 800 watt-hours. Regarding energy consumption, the average power consumption of the system under normal monitoring mode is set at 20 watts. The basic operating energy consumption for 24 hours is calculated as 20 multiplied by 24, equaling 480 watt-hours. If, in extremely cold weather, the low-temperature preheating module operates for a cumulative 2 hours per day with a heating power of 50 watts, then the heating energy consumption is calculated as 50 multiplied by 2, equaling 100 watt-hours. Watt-hours; the total daily energy consumption of the system is the sum of the base energy consumption and the heating energy consumption, that is, 480 plus 100 equals 580 watt-hours; by comparison, the daily power generation of 800 watt-hours is greater than the total daily energy consumption of 580 watt-hours, indicating that the system can still maintain a positive energy surplus under low temperature conditions, ensuring uninterrupted operation of the equipment; in addition, the BMS system is also set with multi-level voltage protection logic. When the battery voltage is detected to be lower than 11.0 volts, a first-level low-voltage alarm is triggered, and the energy management unit actively cuts off the power supply circuit of high-power loads such as laser emitter and loudspeaker, and only keeps the core radar detection and communication module running. At this time, the system power consumption drops to less than 5 watts to extend the standby time and wait for the light to recover.

[0023] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A smart early warning and dynamic-static coordinated prevention and control system for bird-related risks in power grids, characterized in that: The system includes: The panoramic perception unit is configured to acquire environmental data and motion trajectory data of target objects within the power grid monitoring area in real time through a radar detection module and a visual acquisition module, and generate multi-source perception signals. The intelligent decision-making unit is configured to receive the multi-source sensing signals, extract and compare the features of the target object, determine whether the target object is a bird that is involved in the crime and its specific species and behavior pattern, calculate the risk level based on the determination result, and then construct a hierarchical collaborative prevention and control strategy. The collaborative execution unit is configured to drive the dynamic drive-away module to emit photoacoustic interference signals of different intensities based on the hierarchical collaborative prevention and control strategy, and combine it with static physical barrier components to continuously occupy the key charged area, thus constructing a dynamic and static combined protection closed loop. The remote control unit is configured to connect the intelligent decision-making unit and the collaborative execution unit through a communication network, display the protection status in real time, and receive manual control instructions to correct the hierarchical collaborative prevention and control strategy.

2. The intelligent early warning and dynamic-static coordinated prevention and control system for bird-related risks in power grids according to claim 1, characterized in that, The intelligent decision-making unit's judgment logic for the target object specifically includes: A pre-built feature database containing the morphology, texture, and flight posture of various birds involved in harm; The collected image data is input into a feature extraction network to extract the shape contour and motion feature vector of the target object, and the similarity value between it and the samples in the feature database is calculated. When the similarity value exceeds a preset threshold, the target object is confirmed to be a bird that is causing harm, and non-target interference signals such as fallen leaves or insects are removed. By combining distance data detected by radar, the spatial location of the birds involved in the incident is determined through a coordinate mapping algorithm, generating early warning information that includes direction, distance, and species.

3. The intelligent early warning and dynamic-static coordinated prevention and control system for bird-related risks in power grids according to claim 1, characterized in that, The dynamic removal module includes an intelligent laser emitting component, and the execution logic of the intelligent laser emitting component is as follows: Receive bird location coordinates and species information sent by the intelligent decision-making unit; Control the gimbal mechanism to rotate horizontally and vertically, so that the laser beam is aimed at the area where the birds are located; It emits a green laser beam with a wavelength of 532nm and adjusts the beam projection direction according to the movement trajectory of the birds to generate visual stimulation signals within the field of vision of the birds involved in the incident. The power of the green laser beam is limited to a non-harmful range, and the beam projection mode supports switching between point scanning and line scanning.

4. The intelligent early warning and dynamic-static coordinated prevention and control system for bird-related risks in power grids according to claim 1, characterized in that, The dynamic removal module also includes an acoustic wave coordination component, the execution logic of which is as follows: The integrated radar detection module has a detection range of 3 to 8 meters. Based on the hierarchical collaborative prevention and control strategy, the ultrasonic transmitter is activated to generate ultrasonic signals with a dynamic frequency between 5 kHz and 35.04 kHz. Synchronously trigger the loudspeaker player to play preset predator calls or warning sounds, with the volume intensity controlled below 120dB. By intermittently combining the ultrasonic signal with the audio from the loudspeaker, the auditory adaptation of birds is disrupted.

5. The intelligent early warning and dynamic-static coordinated prevention and control system for bird-related risks in power grids according to claim 1, characterized in that, The static physical barrier component includes self-opening and retractable bird spikes and rolling bird-proof wrap-around spikes. The structural features of the self-opening and retractable bird spikes are as follows: It includes a gravity-driven mechanical connection structure and several needle units; Under normal protection conditions, the barbed wire units are distributed in an expanded manner, covering the nesting space above the insulator string; When a physical lifting action is received during maintenance, the needle unit is driven to retract inward using gravity and lever principles, and the opening angle can be adjusted between 120 and 150 degrees. After the maintenance is completed and the lifting force is released, the needle unit automatically resets to the deployed protective state; The structural features of the rolling bird-proof thorn wrapping are as follows: The rolling bird-proof barbed wire is positioned 0 to 1.5 meters from the tower, providing protection for a range of ≥0.3 meters, with a barbed length of ≥50 mm.

6. The intelligent early warning and dynamic-static coordinated prevention and control system for bird-related risks in power grids according to claim 1, characterized in that, The static physical barrier component also includes a retractable bird-proof pin plate, the structural features of which are as follows: It adopts a hinged telescopic structure, and its length adjustment range covers 0.5 meters to 1.2 meters; The base is equipped with both magnetic interface and U-shaped buckle for fastening, which can be used to adapt to crossarms of poles made of different materials; The needle plate is evenly covered with corrosion-resistant steel needles. The length and spacing of the steel needles are configured to block the landing points of large birds, and the ends of the steel needles have a blunt chamfered structure.

7. The intelligent early warning and dynamic-static coordinated prevention and control system for bird-related risks in power grids according to claim 1, characterized in that, It also includes bird-proof windmills, the structural features of which are: Made of ABS engineering plastic, the rotating diameter is ≥ 30cm and the height is ≥ 35cm. The bird-proof windmill is equipped with windmill blades with reflective color design. It can rotate 360° under wind power. The top of the bird-proof windmill is equipped with wind blade spikes, and the wind blade spikes are equipped with bearings. After a bird lands, it will rotate due to gravity.

8. The intelligent early warning and dynamic-static coordinated prevention and control system for bird-related risks in power grids according to claim 1, characterized in that, The dynamic and static coordination logic of the coordinated execution unit specifically includes: When the radar detects birds in the long-range warning zone, the dynamic deterrence module is activated first to provide light and sound warnings. When birds break through long-range defenses and enter the medium-range deterrence zone, increase the laser scanning frequency and sound wave intensity; When dynamic deterrence measures fail or birds attempt to stay at key locations on the pole, the static physical barrier components are used to cut off their nesting or perching paths. The integrity of the physical defense line is monitored by sensors on the static physical barrier components. If the components are found to be displaced or damaged, a maintenance request is sent to the remote control unit.

9. The intelligent early warning and dynamic-static coordinated prevention and control system for bird-related risks in power grids according to claim 1, characterized in that, The system also includes an energy management unit; the energy management unit includes: Monocrystalline silicon solar photovoltaic panels are used to convert light energy into electrical energy; Lithium iron phosphate battery packs are used to store electrical energy and power the system; The low-temperature preheating module is configured to automatically start when the ambient temperature is lower than a preset low-temperature threshold to heat and keep the lithium iron phosphate battery pack warm, ensuring the system's continuous operation capability in low-temperature environments.

10. The intelligent early warning and dynamic-static coordinated prevention and control system for bird-related risks in power grids according to claim 1, characterized in that, The remote control unit also integrates an edge computing node; the edge computing node is configured to preprocess the multi-source sensing signals and perform lightweight model inference locally, and only upload the identification results and abnormal alarm data to the cloud server to reduce communication latency and bandwidth consumption.

Images