Automobile crane high-voltage electricity approaching safety early warning method, system and device

Abstract

The invention provides a high-voltage electricity approaching safety early warning method, system and device for an automobile crane, and belongs to the technical field of engineering machinery safety protection. Comprising the steps of collecting an electric field signal radiated by a high-voltage line, reversely deducing a distance based on an electric field intensity attenuation model, and calculating an azimuth angle of a charged body through a MUSIC algorithm; controlling a laser radar to perform high-density scanning within a set angle according to the azimuth angle to obtain three-dimensional point cloud data of the high-voltage line; time synchronization and coordinate registration are carried out on the electromagnetic sensing array and laser radar data, a three-dimensional space model of the high-voltage line is established after fusion, and the shortest distance between the three-dimensional space model and a suspension arm is calculated; identifying a voltage level, calculating a basic safety distance, correcting by combining an environment factor, and multiplying by a safety coefficient to obtain a dynamic safety threshold value; and according to the real-time distance between the suspension arm and the electrified body, triggering graded early warning and control actions. According to the invention, the problem of measurement deviation of a single sensor under complex electromagnetic environment, changeable illumination and meteorological working conditions is solved.

Description

Technical Field

[0001] The present invention relates to a method, system and device for early warning of high-voltage near-electricity safety of a truck crane, belonging to the technical field of engineering machinery safety protection. Background Art

[0002] With the continued advancement of urbanization and the booming infrastructure development in my country, the number of large-scale construction projects has surged, and construction sites are becoming increasingly dense and located close to existing power facilities. Truck cranes, as key heavy-lifting equipment, play a vital role in construction activities near high-voltage transmission lines due to their flexibility and wide operating range. However, when lifting and rotating these devices in complex and changing construction environments, their metal booms, slings, and other components can easily intrude within the safe distance of live high-voltage power lines due to improper operation or spatial misjudgment, posing a serious safety threat.

[0003] When a truck crane operates near live power lines, the possibility of a near-power or near-power line collision not only damages expensive equipment and disrupts large-scale construction, but can also lead to catastrophic consequences, including but not limited to: fires caused by high-voltage arc discharges, equipment insulation breakdown leading to short circuits and tripping, and even electric shock injuries and deaths to on-site workers. These pose significant risks to the stable and reliable operation of the power grid and the safety of people's lives and property. Therefore, developing a highly reliable and precise safety warning system for truck cranes operating near live power lines, and achieving reliable near-power warnings, has become a critical technical challenge that needs to be addressed to ensure the safety of power facilities and improve construction safety.

[0004] Currently, several technologies exist for near-electrical warnings for truck cranes. A common example is an electromagnetic sensor-based warning method, which detects changes in the surrounding magnetic field strength to determine the distance to live lines. However, this method is significantly affected by the electromagnetic environment of the construction machinery itself, resulting in low ranging accuracy and a high false alarm rate in complex power grid environments, making it difficult to meet the high-precision requirements of practical projects. Ultrasonic ranging technology has also been used for near-electrical distance detection, but its results are significantly affected by factors such as ambient temperature and humidity, and the measurement angle, resulting in large measurement errors and failing to provide stable and reliable warning information.

[0005] There are also near-electrical warning methods based on LiDAR. While these methods can improve ranging accuracy and response speed to a certain extent, single LiDAR solutions are prone to missing point cloud data in adverse weather conditions such as rain and fog, resulting in inaccurate distance information output and affecting warning effectiveness. Furthermore, existing near-electrical warning systems mostly use fixed threshold warnings, failing to account for the differences in safe distances corresponding to different voltage levels. This can lead to insufficient protection or excessive warnings, making them unsuitable for complex and changing near-electrical operating conditions.

[0006] In summary, existing truck crane near-power warning technology has numerous shortcomings, making it difficult to meet the high-precision, high-reliability, and adaptability requirements of construction sites. Therefore, a more advanced and reliable safety warning system and method for near-power operations is urgently needed. This system and method has significant practical significance and application value for improving the inherent safety of truck cranes operating near live power lines, ensuring the safe and stable operation of the power grid, and protecting the lives of construction workers. Summary of the Invention

[0007] The purpose of the present invention is to provide a high-voltage near-electricity safety warning method, system and device for truck cranes, which effectively solves the measurement deviation problem of a single sensor under complex electromagnetic environments, variable lighting and meteorological conditions, and realizes the purpose of all-weather, real-time dynamic and high-precision near-electricity safety warning for high-voltage live bodies during the operation of truck cranes.

[0008] To achieve the above-mentioned purpose, the present invention is implemented through the following technical solutions: A method for early warning of high voltage electric shock safety of a truck crane comprises the following steps: The electromagnetic sensor array collects the electric field signal radiated by the high-voltage line in real time, infers the distance based on the electric field intensity attenuation model, and calculates the azimuth of the charged body using the MUSIC algorithm; According to the azimuth angle of the charged body, the laser radar is controlled to perform high-density scanning within a set angle to obtain three-dimensional point cloud data of the high-voltage line; Perform time synchronization and coordinate registration of electromagnetic sensor array and lidar data, fuse them to build a 3D spatial model of the high-voltage line, and calculate the real-time shortest distance between it and the boom; Identify the voltage level and calculate the basic safety distance. Combined with the environmental factor correction and multiplied by the safety factor, the dynamic safety threshold is obtained. Based on the dynamic safety threshold and the real-time shortest distance between the boom and the live object, graded warning and control actions are triggered.

[0009] Preferably, the electric field strength attenuation model is as follows: , , in, represents the electric field strength measured by the electromagnetic sensor array, is the dielectric constant, Indicates the identified high-voltage line voltage level, Indicates the straight-line distance between the sensor and the charged body. represents the environmental attenuation factor, is the propagation path length of the electromagnetic wave in the medium, It is the angle between the line connecting the laser radar and the charged body and the horizontal plane.

[0010] Preferably, the specific method of calculating the azimuth angle of the charged body by the MUSIC algorithm is as follows: The MUSIC algorithm is used to perform eigendecomposition on the covariance matrix of the electric field signal received by the electromagnetic sensor array. The first K large eigenvalues ​​constitute the signal subspace, and the remaining eigenvalues ​​constitute the noise subspace. The spatial spectrum function is used to search for the azimuth angle of the charged body: , in, is the array steering vector, is the azimuth of the charged body, is the spatial spectrum function in the MUSIC algorithm, Noise subspace, is the conjugate transpose of the noise subspace matrix, is the conjugate transpose of the array steering vector.

[0011] Preferably, the time synchronization and coordinate registration of the electromagnetic sensor array and the laser radar data specifically includes: Align the timestamps of electromagnetic sensor array data and multi-beam lidar data by linear interpolation; The ICP registration algorithm is used to unify the electromagnetic positioning points and the laser point cloud into a coordinate system with the boom root as the origin; The point cloud offset caused by the boom rotation is eliminated according to the correction formula, which is as follows: , , in, is the corrected point cloud coordinate, is the rotation correction matrix, is the original collected point cloud coordinate, is the linear velocity of the boom, The time interval for the laser radar to complete one frame of point cloud collection, is the angular velocity of the boom.

[0012] Preferably, the specific scheme of using the ICP registration algorithm to unify the electromagnetic positioning points and the laser point cloud into a coordinate system with the boom root as the origin is as follows: A conversion matrix is ​​used to unify the coordinates of the electromagnetic sensor array data and the lidar data into a coordinate system with the boom root as the origin; Minimize the sum of squared Euclidean distances between the electromagnetic positioning points in the electromagnetic sensor array data and the laser point cloud in the lidar data: , in, is the rotation matrix, is the translation vector, is the electromagnetic positioning point, is the laser point cloud point, is the total number of electromagnetic positioning point-laser point cloud point matching pairs involved in the registration.

[0013] Preferably, the dynamic safety threshold The calculation is as follows: , , , , , in, 、 、 and is the compensation coefficient, The dust concentration output by the environmental perception module, Indicates the identified high-voltage line voltage level, is the deviation between the ambient temperature and the standard temperature, is the relative humidity, is the environmental attenuation factor, is the dynamic safety factor.

[0014] Preferably, the voltage level recognition adopts a convolutional neural network, the input is the Mel spectrum of the electric field intensity, and the output is the voltage level classification result; the convolutional neural network includes 5 convolution layers and 2 fully connected layers, and is trained based on measured spectrum samples in the voltage range of 10kV-1000kV.

[0015] Preferably, the graded warning includes: Level 1 warning: When the distance is 1.5 times the dynamic safety threshold, an audible and visual warning is triggered; Level 2 warning: When the distance is between the dynamic safety threshold and 1.5 times the dynamic safety threshold, the boom movement speed is limited; Level 3 warning: When the distance is less than the dynamic safety threshold, emergency braking is performed.

[0016] A high-voltage near-electricity safety warning system for a truck crane, comprising: Distributed electromagnetic sensor array: It is composed of flexible three-axis electromagnetic sensors arranged in each section of the boom, used to collect electric field signals radiated by high-voltage lines; Multi-beam LiDAR: Installed on the rotating platform outside the main boom, it is used to collect point cloud data of high-voltage lines. It includes a dual-axis scanning mechanism and a dynamic focusing module. The dynamic focusing module adjusts the point cloud density according to the azimuth angle of the charged body. Environmental perception module: includes temperature, humidity, air pressure, and rain and fog sensors to monitor environmental parameters in real time; Central processing unit: built-in charged body positioning engine, space mapping module, dynamic threshold calculator and voltage level identification module, realizes charged body positioning, coordinate mapping and safety threshold calculation through multi-physics field data fusion; the charged body positioning engine reversely infers the distance of the charged body based on the electric field intensity attenuation model, and calculates the azimuth of the charged body through the MUSIC algorithm; the space mapping module calculates the azimuth of the charged body based on the charged body , control the multi-beam laser radar in Enhanced scanning of the area; aligning the timestamps of the electromagnetic sensor array data and the multi-beam laser radar data by linear interpolation; using the ICP algorithm to unify the electromagnetic positioning point and the laser point cloud to a coordinate system with the base of the boom as the origin; eliminating the point cloud offset caused by the rotation of the boom according to the correction formula; the dynamic threshold calculator calculates the safety threshold by constructing a three-dimensional safety distance model ; The voltage level identification module identifies the voltage level based on a convolutional neural network.

[0017] A novel grid-forming new energy transient reactive support strategy and setting device includes a processor and a memory storing program instructions. The processor is configured to execute the high-voltage near-electricity safety early warning method for a truck crane when running the program instructions.

[0018] The advantages of the present invention are: Through the pioneering multi-physics field spatiotemporal synchronous fusion method, the advantages of electromagnetic and laser sensors are complemented to form an efficient closed-loop detection process, which significantly improves the success rate of identifying high-voltage thin wires and fundamentally solves the problem of positioning small targets in complex environments.

[0019] A dynamic threshold calculation system based on deep learning and environmental modeling was proposed, enabling accurate and intelligent determination of safe distances. This technology performs particularly well in extreme environments such as rain and fog, keeping warning errors within industry standards and significantly improving equipment operational safety and environmental adaptability.

[0020] By introducing motion parameters and building a real-time compensation model, the measurement distortion caused by the equipment's own movement is effectively eliminated, allowing the system to maintain high-precision ranging under dynamic conditions of continuous boom movement, ensuring the accuracy and stability of measurement throughout the entire process. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention.

[0022] Figure 1 Schematic diagram of the process of the present invention.

[0023] Figure 2 This is a schematic diagram of the structure of a safety warning system for near-electrical operations of a truck crane that integrates electromagnetic sensing and laser ranging in an embodiment of the present invention.

[0024] Figure 3 This is a flowchart of the operation of the safety warning system for near-electrical operations of a truck crane that integrates electromagnetic sensing and laser ranging in an embodiment of the present invention.

[0025] Among them, 1 is the electromagnetic sensor array and 2 is the lidar. DETAILED DESCRIPTION

[0026] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0027] Example 1 like Figures 1 to 3 As shown, a high-voltage near-electricity safety warning method for a truck crane includes the following steps: S1: The electric field signal radiated by the high-voltage line is collected in real time through an electromagnetic sensor array, the distance is inferred based on the electric field intensity attenuation model, and the azimuth of the charged body is calculated using the MUSIC algorithm; S2: Controlling the laser radar to perform high-density scanning within a set angle according to the azimuth angle of the charged body to obtain three-dimensional point cloud data of the high-voltage line; S3: Time synchronization and coordinate registration of the electromagnetic sensor array and lidar data are performed. After fusion, a three-dimensional spatial model of the high-voltage line is established and the real-time shortest distance between the line and the boom is calculated. S4: Identify the voltage level and calculate the basic safety distance. Combined with the environmental factor correction, multiply it by the safety factor to obtain the dynamic safety threshold. S5: Trigger graded warning and control actions based on the dynamic safety threshold and the real-time shortest distance between the boom and the live object.

[0028] As a refinement of the above embodiment, the electric field strength attenuation model is as follows: , , in, represents the electric field strength measured by the electromagnetic sensor array, is the dielectric constant, Indicates the identified high-voltage line voltage level, Indicates the straight-line distance between the sensor and the charged body. Represents the environmental attenuation factor, which is dynamically calculated based on the temperature and humidity sensor data. is the propagation path length of the electromagnetic wave in the medium.

[0029] As a refinement of the above embodiment, the present invention uses a direction of arrival (DOA) algorithm to determine the azimuth of a charged object. It also applies the Multiple Signal Classification (MUSIC) method to crane proximity detection for the first time, leveraging the orthogonality of signal subspaces to enhance anti-interference capabilities. The specific method is as follows: S101: Use the MUSIC algorithm to calculate the covariance matrix of the electric field signal received by the electromagnetic sensor array The characteristic decomposition of is as follows: , in, is the number of electromagnetic sensor array elements, is the eigenvector matrix, is the conjugate transpose of the eigenvector matrix.

[0030] S102: Arrange the eigenvalues ​​in descending order , the first K large eigenvalues ​​constitute the signal subspace, and the remaining eigenvalues ​​constitute the noise subspace. The spatial spectrum function is used to search for the azimuth angle of the charged body: , in, is the array steering vector, is the azimuth of the charged body, is the spatial spectrum function in the MUSIC algorithm, The noise subspace matrix, is the conjugate transpose of the noise subspace matrix, is the conjugate transpose of the array steering vector, and K is the number of actual high-voltage charged bodies.

[0031] The value of K is not a fixed constant, but dynamically matches the "actual number of high-voltage charged bodies" in the current operating environment: first, the number of signal sources is preliminarily determined through the "direction and intensity characteristics" of the electric field signal, and then verified through the "amplitude mutation point" after the eigenvalues ​​are sorted, and finally the K value is determined to ensure that the signal subspace can accurately correspond to all high-voltage charged bodies to be detected, providing an accurate basis for subsequent azimuth angle calculation (searching for the peak of the spatial spectrum function).

[0032] For example, if there are two or more parallel high-voltage lines (such as double-circuit transmission lines) in the operating environment, K = 2 (or more, depending on the number of lines), and the signal subspace is composed of the first two (or more) large eigenvalues.

[0033] As a refinement of the above embodiment, the time synchronization and coordinate registration of the electromagnetic sensor array and the lidar data specifically includes: S301: aligning the timestamps of the electromagnetic sensor array data and the multi-beam lidar data by linear interpolation; Assume that the sampling time of the electromagnetic sensor array is , the laser radar scanning frame time is , time alignment is achieved by linear interpolation: .in, (LiDAR minimum scanning period), the synchronization error is controlled within ±5ms through timestamp calibration, meeting the ranging accuracy requirements (error ≤ 0.01m) for dynamic movement of the boom (maximum linear speed 1m / s), is the step size variable for time interpolation, is the upper limit of the total number of interpolations.

[0034] S302: Using the ICP registration algorithm, the electromagnetic positioning points and the laser point cloud are unified into a coordinate system with the boom root as the origin; S3021: Use a transformation matrix to unify the coordinates of the electromagnetic sensor array data and the laser radar data into a coordinate system with the boom root as the origin; Electromagnetic sensor array coordinate system : The origin of the electromagnetic sensor array coordinate system is set to the center of the array at the top of the boom; The X-axis of the electromagnetic sensor array coordinate system is set to point along the boom to the top; The Y axis of the electromagnetic sensor array coordinate system is set to be perpendicular to Horizontal side of the axis; The Z axis of the electromagnetic sensor array coordinate system is set to be vertically upward; LiDAR coordinate system : is the origin of the laser radar coordinate system, set to the radar rotation center, The X-axis of the laser radar coordinate system is set to be along the main optical axis of the radar; Y axis of the laser radar coordinate system is set to the horizontal tangent perpendicular to the main optical axis; The Z axis of the laser radar coordinate system is set to be vertically upward (same as Axis in the same direction); Unified world coordinate system : To unify the origin of the world coordinate system, set it to the hinge point at the root of the boom. To unify the world coordinate system X-axis, set it to point along the ground to the working direction; To unify the Y axis of the world coordinate system, set it to be perpendicular to The horizontal side direction of the axis (across the ground); To unify the world coordinate system Z axis, set it to vertically upward (with 、 The system unifies the “height dimension” benchmark of all sensors to ensure that electromagnetic signals and laser point clouds can be accurately integrated in three-dimensional space through coordinate transformation.

[0035] Transformation Matrix: The electromagnetic sensor array coordinates are converted to the unified world coordinate system: ,in is a 4×4 transformation matrix, including the boom length L, elevation angle , rotation angle Parameters: .

[0036] S3022: Minimize the sum of squared Euclidean distances between electromagnetic positioning points in electromagnetic sensor array data and laser point clouds in lidar data.

[0037] ICP registration algorithm: Objective function: Minimize electromagnetic positioning points With laser point cloud The sum of squared Euclidean distances: , in, is a 3×3 rotation matrix, is a 3×1 translation vector, is the electromagnetic positioning point, is the laser point cloud point, for The nearest neighbor point in the laser point cloud, is the total number of scene-level point clouds obtained from the original LiDAR scan, is the total number of electromagnetic positioning point-laser point cloud point matching pairs involved in the registration. The value of is determined by the number of electromagnetic positioning points and the number of laser point cloud points that participate in the registration simultaneously (it is necessary to ensure that n points are selected on the electromagnetic side and n corresponding points are selected on the laser side to form n matching pairs). It is the "point pair number parameter" that associates electromagnetic data with laser point cloud data and supports the registration optimization calculation.

[0038] Through iterative optimization (iteration number ≤ 20 times), the angle error of the registration error is less than or equal to 0.3° and the distance error is less than or equal to 0.1m.

[0039] S303: Eliminate the point cloud offset caused by the boom rotation according to a correction formula. The correction formula is as follows: , , in, is the corrected point cloud coordinate, is the rotation correction matrix, is the original collected point cloud coordinate, is the linear velocity of the boom, The time interval for the laser radar to complete one frame of point cloud collection, is the angular velocity of the boom. Experimental verification shows that the dynamic ranging error can be reduced from 0.5m to within 0.1m.

[0040] As a refinement of the above embodiment, the dynamic safety threshold The calculation is as follows: S401: Basic safety distance calculate; Based on the voltage reference distance of IEC61472 standard, environmental parameter compensation is introduced: , in, 、 、 and is the compensation coefficient, which is obtained by fitting 100,000 sets of experimental data using the least squares method. The dust concentration output by the environmental perception module, Indicates the identified high-voltage line voltage level, is the deviation between the ambient temperature and the standard temperature (25°C), is the relative humidity.

[0041] Voltage term : Based on the electric field attenuation characteristics, the 0.7 power fitting is consistent with the nonlinear relationship between the critical distance of air breakdown and voltage; Humidity item : For every 10% increase in humidity, the air breakdown field strength decreases by 5%, so the safety distance needs to be increased by 0.2m; Temperature item : For every 10°C increase in temperature, the insulation strength of air decreases by 3%, corresponding to a distance compensation of 0.1m.

[0042] S402: Environmental attenuation factor Correction In view of the attenuation of electric field propagation in extreme weather conditions such as rain and fog, a correction factor is introduced: , Actual safety distance after correction: .

[0043] Experimental verification: When the humidity is 95% and the dust concentration is 20mg / m³, =1.8, the safety distance of 500kV line is revised from 6.8m to 12.24m, and the error with the measured breakdown distance is ≤5%.

[0044] S403: Dynamic safety factor Adaptive adjustment according to working conditions: , The dynamic working condition adaptation unit generates the final safety threshold: At the same time, the PPO deep reinforcement learning algorithm is adopted with the goal of "no missed reports and low false alarms". Based on more than 1,000 historical accident cases and 500,000 sets of simulation data training models, real-time self-optimization of thresholds is achieved, and the error is reduced to below 2.3% in urban dense power grid scenarios.

[0045] As a refinement of the above embodiment, the voltage level recognition module uses a convolutional neural network (CNN) with electric field intensity spectrum features as input and voltage level classification results as output. The model training data consists of measured electric field samples under different voltages, distances, and environmental conditions. The CNN model uses a 5-layer convolutional + 2-layer fully connected structure, with the Mel-spectrogram of the electric field signal as input. The training data consists of 120,000 sets of measured samples from 10kV to 1000kV. At a signal-to-noise ratio (SNR) of 10dB, the recognition accuracy rate is 98.7%, which is superior to the 85.2% of traditional FFT features.

[0046] As a refinement of the above embodiment, the hierarchical warning includes: Level 1 warning: 1.5 times the distance When the sound and light prompt is triggered; Level 2 warning: distance between and 1.5 times When between, limit the boom movement speed; Level 3 warning: distance less than , perform emergency braking.

[0047] It should be noted that this embodiment has the following technical effects: A spatiotemporal synchronized multi-physics fusion method The first closed-loop fusion process of "electromagnetic orientation guidance-laser focus scanning-dynamic alignment correction" solves the positioning problem of high-voltage thin wires (diameter 15-30mm) in complex environments through the combination of MUSIC algorithm and ICP iteration. Compared with the traditional single-laser sensor solution, the recognition success rate has increased from 72% to 99.3%.

[0048] Calculation method of adaptive safety threshold A dynamic threshold system based on voltage spectrum characteristics (CNN recognition) and environmental attenuation models is constructed to implement quantitative calculation of the entire chain of "voltage level → reference distance → environmental correction → operating condition adaptation". Under extreme conditions such as rainy and foggy days (humidity 95%), the warning error is ≤5%, which is better than the industry standard (≤15%).

[0049] Real-time compensation method for motion distortion The boom motion parameters (angular velocity, linear velocity) are introduced into the point cloud correction model. Through the coupled calculation of the rotation matrix and the translation vector, the ranging deviation during dynamic operation is eliminated, and the distance measurement accuracy during the boom motion state is maintained within 0.1m.

[0050] Example 2 like Figure 2 As shown, a high-voltage near-electricity safety warning system for a truck crane includes: Distributed electromagnetic sensor array: It is composed of flexible three-axis electromagnetic sensors arranged in each section of the boom, used to collect electric field signals radiated by high-voltage lines; Multi-beam LiDAR: Installed on the rotating pan / tilt platform outside the main boom, it is used to collect point cloud data of high-voltage lines. It includes a dual-axis scanning mechanism (horizontal rotation range of ±180°, pitch angle adjustment range of -30° to +60°) and a dynamic focusing module. The dynamic focusing module adjusts the point cloud density according to the azimuth angle of the charged body. Environmental perception module: includes temperature, humidity, air pressure, and rain and fog sensors to monitor environmental parameters in real time; Central processing unit: Built-in charged body positioning engine, space mapping module, dynamic threshold calculator and voltage level identification module, realizes charged body positioning, coordinate mapping and safety threshold calculation through multi-physics field data fusion.

[0051] The charged body positioning engine infers the distance of the charged body based on the electric field intensity attenuation model and calculates the azimuth of the charged body through the MUSIC algorithm.

[0052] The spatial mapping module is based on the charged body azimuth , control the multi-beam laser radar in Enhanced scanning of the entire area; aligning the timestamps of the electromagnetic sensor array data and the multi-beam lidar data through linear interpolation; using the ICP algorithm to unify the electromagnetic positioning points and the laser point cloud into a coordinate system with the boom root as the origin, with a registration error of ≤0.3°; and eliminating point cloud offsets caused by boom rotation using a correction formula. The dynamic threshold calculator calculates the safety threshold by constructing a three-dimensional safety distance model ; The voltage level identification module identifies the voltage level based on a convolutional neural network.

[0053] Example 3 This example demonstrates the practical application of the present invention using a 35-ton truck crane performing a hoisting operation near a 220kV high-voltage line. Operating environment parameters include: temperature 30°C, relative humidity 85%, dust concentration 10mg / m³, and no rain or fog. The deployment and operation process of each system module is as follows: 1. System Deployment Details Distributed electromagnetic sensor array: A flexible triaxial electric field sensor, sampling at 1kHz, is installed at each of the second, third, and fourth boom sections (10m, 15m, and 20m from the boom base) to collect real-time spatial electric field strength and gradient data. The sensor has a measurement range of 0–5000 V / m and an accuracy of ±2%.

[0054] The multi-beam LiDAR is mounted on a rotating pan / tilt platform on the outer side of the mid-section of the main boom. Its dual-axis scanning mechanism has a horizontal rotation range of ±180° and a pitch adjustment range of -30° to +60°. The dynamic focus module has a default scanning frequency of 10Hz and a point cloud density of 50 points / m². This frequency can be increased to 40Hz in high-risk areas, increasing the point cloud density to 200 points / m².

[0055] Environmental perception module: Integrated temperature and humidity sensor (measurement accuracy ±2%RH, ±0.5℃), dust sensor (measurement range 0-50mg / m³), data output frequency 10Hz, real-time upload of environmental parameters to the central processing unit.

[0056] Central Processing Unit: An industrial-grade embedded computer (2.8 GHz main frequency, 16 GB memory) with a built-in charged body positioning engine, spatial mapping module, dynamic threshold calculator, and voltage level recognition module (the CNN model is a 5-layer convolution + 2-layer fully connected structure, with the input electric field signal Mel spectrum).

[0057] 2. Operation Process Step 1: Detect the position of the charged object The electromagnetic sensor array collects the electric field signals radiated by the high-voltage lines in real time. The measured electric field strengths E of the three sensors are 800V / m, 950V / m, and 1050V / m, respectively.

[0058] Distance inversion: Calculate the distance using the electric field intensity attenuation model, where the dielectric constant =0.85 (air), environmental attenuation factor (RH=85%, =10mg / m³).

[0059] Angle calculation: The charged body positioning engine starts the MUSIC algorithm and performs eigendecomposition on the covariance matrix of the received signals of the three array elements to obtain the signal subspace and noise subspace , by searching for the peak of the spatial spectrum function, the azimuth of the charged body is determined =45°.

[0060] Step 2: LiDAR Focus Scan The spatial mapping module converts the azimuth =45° is sent to the LiDAR dynamic focusing module, which controls it to perform high-density scanning at a frequency of 40 Hz within a cone angle of 30° to 60° (45°±15°). The point cloud density is increased to 200 points / m², and 3D point cloud data of the high-voltage lines in the area is obtained.

[0061] Step 3: Multi-data fusion modeling Time synchronization: Linear interpolation is used to align the timestamps of the electromagnetic sensor array (1kHz) and the lidar (40Hz). The synchronization error is controlled within ±4ms, meeting the ranging accuracy requirements (error ≤ 0.01m) for the dynamic motion of the boom (linear velocity 0.8m / s).

[0062] Coordinate registration: ICP algorithm is used to register the electromagnetic positioning point ( ,45°) and the laser point cloud are unified into the world coordinate system with the boom root as the origin. After 18 iterations, the registration angle error is 0.25° and the distance error is 0.07m.

[0063] Motion distortion correction: The arm rotates at 0.3 rad / s, and the scanning cycle =0.1s, and substitute the rotation matrix into the correction formula to eliminate the point cloud offset caused by the boom rotation.

[0064] 3D modeling: After fusion, the high-voltage line is determined to be a single line, and the spatial coordinates ( ) The shortest straight-line distance between the crane and the boom is preliminarily calculated to be 6.2m.

[0065] Step 4: Dynamic safety threshold calculation Voltage level recognition: The CNN inputs the Mel spectrum features of the electric field signal and outputs a recognition result of 220 kV (recognition accuracy of 99.1%).

[0066] Basic safety distance: According to the formula , The compensation coefficient , , , , , , , , calculated .

[0067] Environmental correction: Environmental attenuation factor , the actual safety distance after correction .

[0068] Safety threshold: Take the operating safety factor (Lifting operation), final safety threshold .

[0069] Step 5: Triggering of graded warnings Level 1 warning: When the distance between the boom and the charged object is 14.04m (1.5×9.36m), the level 1 warning is triggered: the yellow indicator light is on, the 1kHz buzzer starts, and the prompt "approaching the safe range of the charged object" is turned on.

[0070] Level 2 warning: When the distance drops to 10.0m ( ), a secondary warning is triggered: the red indicator light turns on, the 2kHz buzzer starts, and the system sends a PWM signal through the CAN bus to limit the boom rotation speed from 0.8rad / s to 0.3rad / s.

[0071] Level 3 warning: When the distance drops to 6.0m due to operational errors ( ), a three-level warning is triggered: the red and green indicator lights flash alternately, the buzzer sounds continuously, the central processing unit outputs instructions, controls the hydraulic system overflow valve to unload, and performs emergency braking with a braking distance of 0.4m (≤0.5m).

[0072] 3. Implementation Effect Verification In this embodiment, the system's recognition success rate for 220kV high-voltage lines is 99.3%, the dynamic ranging accuracy is maintained within 0.08m, and the warning error in extreme environments such as rainy and foggy days (humidity 95%) is ≤4.8%, all of which are better than the industry standard, verifying the effectiveness of the present invention under complex working conditions.

[0073] The disclosed embodiments also provide a high-voltage electrical proximity safety warning device for a mobile crane, comprising a processor and memory. Optionally, the device may also include a communication interface and a bus. The processor, communication interface, and memory may communicate with each other via the bus. The communication interface may be used for information transmission. The processor may invoke logic instructions stored in the memory to execute the high-voltage electrical proximity safety warning method for a mobile crane according to the aforementioned embodiments.

[0074] In addition, the logic instructions in the above-mentioned memory can be implemented in the form of software functional units and can be stored in a computer-readable storage medium when sold or used as an independent product.

[0075] Memory, as a computer-readable storage medium, can be used to store software programs and computer-executable programs, such as the program instructions / modules corresponding to the methods in the embodiments of the present disclosure. The processor executes the program instructions / modules stored in the memory to perform functional applications and data processing, thereby implementing the high-voltage proximity safety warning method for truck cranes in the above-mentioned embodiments.

[0076] The memory may include a program storage area and a data storage area. The program storage area may store an operating system and at least one application required for a function; the data storage area may store data generated based on the use of the terminal device. Furthermore, the memory may include high-speed random access memory and non-volatile memory.

[0077] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art will be able to modify the technical solutions described in the aforementioned embodiments or substitute equivalents for some of the technical features. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

Claims

1. A high voltage near-electricity safety early warning method for a truck crane, characterized in that: The following steps are involved: The electromagnetic sensor array collects the electric field signal radiated by the high-voltage line in real time, infers the distance based on the electric field intensity attenuation model, and calculates the azimuth of the charged body using the MUSIC algorithm; According to the azimuth angle of the charged body, the laser radar is controlled to perform high-density scanning within a set angle to obtain three-dimensional point cloud data of the high-voltage line; Perform time synchronization and coordinate registration of electromagnetic sensor array and lidar data, fuse them to build a 3D spatial model of the high-voltage line, and calculate the real-time shortest distance between it and the boom; Identify the voltage level and calculate the basic safety distance. Combined with the environmental factor correction and multiplied by the safety factor, the dynamic safety threshold is obtained. Based on the dynamic safety threshold and the real-time shortest distance between the boom and the live object, graded warning and control actions are triggered.

2. The high voltage near-electricity safety early warning method for a truck crane according to claim 1 is characterized in that: The electric field intensity attenuation model is as follows: ， ， in, represents the electric field strength measured by the electromagnetic sensor array, is the dielectric constant, Indicates the identified high-voltage line voltage level, Indicates the straight-line distance between the sensor and the charged body. represents the environmental attenuation factor, is the propagation path length of the electromagnetic wave in the medium, It is the angle between the line connecting the laser radar and the charged body and the horizontal plane.

3. The high voltage near-electricity safety early warning method for a truck crane according to claim 1 is characterized in that: The specific method of calculating the azimuth angle of the charged body using the MUSIC algorithm is as follows: The MUSIC algorithm is used to perform eigendecomposition on the covariance matrix of the electric field signal received by the electromagnetic sensor array. The first K large eigenvalues ​​constitute the signal subspace, and the remaining eigenvalues ​​constitute the noise subspace. The spatial spectrum function is used to search for the azimuth angle of the charged body: ， in, is the array steering vector, is the azimuth of the charged body, is the spatial spectrum function in the MUSIC algorithm, Noise subspace, is the conjugate transpose of the noise subspace matrix, is the conjugate transpose of the array steering vector.

4. The high voltage near-electricity safety early warning method for a truck crane according to claim 1 is characterized in that: The time synchronization and coordinate registration of the electromagnetic sensor array and the laser radar data specifically includes: Align the timestamps of electromagnetic sensor array data and multi-beam lidar data by linear interpolation; The ICP registration algorithm is used to unify the electromagnetic positioning points and the laser point cloud into a coordinate system with the boom root as the origin; The point cloud offset caused by the boom rotation is eliminated according to the correction formula, which is as follows: ， ， in, is the corrected point cloud coordinate, is the rotation correction matrix, is the original collected point cloud coordinate, is the linear velocity of the boom, The time interval for the laser radar to complete one frame of point cloud collection, is the angular velocity of the boom.

5. The high voltage near-electricity safety early warning method for a truck crane according to claim 4 is characterized in that: The specific scheme of using the ICP registration algorithm to unify the electromagnetic positioning points and the laser point cloud into a coordinate system with the boom root as the origin is as follows: A conversion matrix is ​​used to unify the coordinates of the electromagnetic sensor array data and the lidar data into a coordinate system with the boom root as the origin; Minimize the sum of squared Euclidean distances between the electromagnetic positioning points in the electromagnetic sensor array data and the laser point cloud in the lidar data: ， in, is the rotation matrix, is the translation vector, is the electromagnetic positioning point, is the laser point cloud point, is the total number of electromagnetic positioning point-laser point cloud point matching pairs involved in the registration.

6. The high voltage near-electricity safety early warning method for a truck crane according to claim 1 is characterized in that: The dynamic safety threshold The calculation is as follows: ， ， ， ， ， in, 、 、 and is the compensation coefficient, The dust concentration output by the environmental perception module, Indicates the identified high-voltage line voltage level, is the deviation between the ambient temperature and the standard temperature, is the relative humidity, is the environmental attenuation factor, is the dynamic safety factor.

7. The high voltage proximity safety warning method for a truck crane according to claim 6 is characterized in that: The voltage level identification adopts a convolutional neural network, the input is the electric field intensity Mel spectrum, and the output is the voltage level classification result; the convolutional neural network includes 5 convolution layers and 2 fully connected layers, and is trained based on measured spectrum samples in the voltage range of 10kV-1000kV.

8. The high voltage near-electricity safety early warning method for a truck crane according to claim 1 is characterized in that: The graded warnings include: Level 1 warning: When the distance is 1.5 times the dynamic safety threshold, an audible and visual warning is triggered; Level 2 warning: When the distance is between the dynamic safety threshold and 1.5 times the dynamic safety threshold, the boom movement speed is limited; Level 3 warning: When the distance is less than the dynamic safety threshold, emergency braking is performed.

9. A high voltage near-electricity safety warning system for a truck crane, characterized in that: The method for high-voltage proximity safety warning of a truck crane according to any one of claims 1 to 8 comprises: Distributed electromagnetic sensor array: It is composed of flexible three-axis electromagnetic sensors arranged in each section of the boom, used to collect electric field signals radiated by high-voltage lines; Multi-beam LiDAR: Installed on the rotating platform outside the main boom, it is used to collect point cloud data of high-voltage lines. It includes a dual-axis scanning mechanism and a dynamic focusing module. The dynamic focusing module adjusts the point cloud density according to the azimuth angle of the charged body. Environmental perception module: includes temperature, humidity, air pressure, and rain and fog sensors to monitor environmental parameters in real time; Central processing unit: built-in charged body positioning engine, space mapping module, dynamic threshold calculator and voltage level identification module, realizes charged body positioning, coordinate mapping and safety threshold calculation through multi-physics field data fusion; the charged body positioning engine reversely infers the distance of the charged body based on the electric field intensity attenuation model, and calculates the azimuth of the charged body through the MUSIC algorithm; the space mapping module calculates the azimuth of the charged body based on the charged body , control the multi-beam laser radar in Enhanced scanning of the area; aligning the timestamps of the electromagnetic sensor array data and the multi-beam laser radar data by linear interpolation; using the ICP algorithm to unify the electromagnetic positioning point and the laser point cloud to a coordinate system with the base of the boom as the origin; eliminating the point cloud offset caused by the rotation of the boom according to the correction formula; the dynamic threshold calculator calculates the safety threshold by constructing a three-dimensional safety distance model ; The voltage level identification module identifies the voltage level based on a convolutional neural network.

10. A high voltage electric safety warning device for a truck crane, comprising a processor and a memory storing program instructions, characterized in that: The processor is configured to execute the high-voltage proximity safety warning method for a truck crane according to any one of claims 1 to 8 when running the program instructions.

Images