A method for limiting the torque of a lifting device and the lifting device itself.

Abstract

This application relates to the field of crane safety monitoring technology, and provides a torque limiter for crane equipment and the crane equipment itself, including: a sensor network comprising multiple multimodal sensor nodes distributed at key components of the crane equipment for collecting real-time data; a backscatter energy harvesting device connected to the multimodal sensor nodes for capturing energy from a radio frequency radiation source and converting it into electrical energy to power the multimodal sensor nodes; an environmental sensing module for collecting real-time environmental data; and a processing device connected to the multimodal sensor nodes and the environmental sensing module for analyzing the real-time data to obtain torque prediction values ​​for future time points, and using risk assessment rules to assess the risk of the torque prediction values ​​and real-time environmental data, determining the risk level, and controlling the crane equipment to execute corresponding adjustment strategies. This application achieves maintenance-free continuous power supply and forward-looking and accurate risk prediction, improving system reliability and safety.

Description

Technical Field

[0001] This application relates to the field of safety monitoring technology for lifting machinery, and more specifically, to a torque limiter for lifting equipment and lifting equipment. Background Technology

[0002] Currently, in the field of crane safety monitoring, torque limiters typically employ distributed sensor networks deployed at critical components of the lifting equipment to collect data, which is then combined with a controller for risk assessment. However, existing solutions suffer from the problem of relying on batteries or external wired power supplies for sensor power, which can easily lead to battery depletion and unstable power supply, resulting in frequent sensor network interruptions or failures. This increases the maintenance burden and reduces system reliability. Furthermore, traditional solutions have weak real-time perception capabilities for dynamic environmental factors, and torque prediction relies on static models or historical data, failing to adequately respond to sudden risks and resulting in lag in safety control. Summary of the Invention

[0003] The purpose of this application is to provide a torque limiter for a lifting device and a lifting device, which uses a backscatter energy harvesting device to capture energy from radio frequency radiation to power the sensing node, and combines environmental data and torque prediction values ​​to determine risk, thereby achieving maintenance-free continuous power supply and forward-looking and accurate risk prediction.

[0004] In a first aspect, a torque limiter for a lifting device is provided, the torque limiter comprising: A sensor network, comprising multiple multimodal sensing nodes distributed at key components of the lifting equipment, is used to collect real-time data from the lifting equipment. A backscatter energy harvesting device, configured at the key components of the lifting equipment and connected to the corresponding multimodal sensing nodes, is used to capture energy from the radio frequency radiation source of the lifting equipment and convert the captured energy into electrical energy to power the multimodal sensing nodes. An environmental sensing module is used to collect real-time environmental data from the lifting equipment. A processing device, connected to the multimodal sensing nodes and the environmental sensing module, is used to analyze the received real-time data, obtain the torque prediction values ​​of each key component at each future time point, and use risk assessment rules to assess the risk of the torque prediction values ​​and the real-time environmental data, determining the risk level to control the lifting equipment to execute the adjustment strategy corresponding to the risk level.

[0005] In one possible implementation, the radio frequency radiation source is the main control cabinet of the lifting equipment; the backscatter energy harvesting device is used to capture energy from the radio frequency radiation generated when the main control cabinet is working through an antenna, and to convert the captured energy into electrical energy through an energy conversion circuit to power the multimodal sensing nodes arranged at the corresponding positions.

[0006] In one possible implementation, the energy conversion circuit includes the antenna, rectifier circuit, hybrid energy storage unit, voltage regulator circuit, and load modulation module connected in sequence; the rectifier circuit is used to convert the AC signal received by the antenna into a DC signal; the hybrid energy storage unit includes a supercapacitor and a micro battery for storing the converted DC energy; the voltage regulator circuit is used to regulate the voltage of the energy output by the hybrid energy storage unit; the load modulation module is used to modulate the collected energy according to the needs of the multimodal sensing node and to achieve backscatter communication by changing the reflection characteristics of the antenna.

[0007] In one possible implementation, the processing device is configured to: for any critical component, detect whether the predicted torque value corresponding to any future time point is greater than the rated torque, and whether the real-time environmental data meets preset obstacle conditions; if the predicted torque value corresponding to any future time point is not greater than the rated torque of a first coefficient, and the real-time environmental data meets preset safety conditions, then determine the risk level as low risk; if the predicted torque value corresponding to any future time point is between the rated torque of the first coefficient and the rated torque of the second coefficient, or the predicted torque value is not greater than the rated torque and the real-time environmental data indicates the presence of dynamic obstacles, then determine the risk level as medium risk; if the predicted torque value corresponding to any future time point is greater than the rated torque of the second coefficient, or the predicted torque value is between the rated torque of the second coefficient and the rated torque, and the real-time environmental data indicates that the ground slope is greater than the configured safety range, then determine the risk level as high risk.

[0008] In one possible implementation, the processing device is further configured to: when the risk level is determined to be low risk, control the lifting equipment to output a warning message and instruct the driver to actively adjust the operation; when the risk level is determined to be medium risk, control the lifting equipment to limit the power output of the dangerous action and allow the execution of the safe action; when the risk level is determined to be high risk, control the lifting equipment to forcibly cut off the power of the dangerous action and trigger emergency braking.

[0009] In one possible implementation, the key components of the lifting equipment include a winch, a directional pulley, a boom root, and a slewing platform; the multimodal sensing nodes include a three-pulley weight sensor fixed to the winch and the directional pulley, a multi-axis inertial measurement unit sensor installed at the boom root, and an environmental sensor installed on the slewing platform; the real-time data includes the tension of the wire rope detected by the three-pulley weight sensor, the pitch angle, roll angle, and vibration frequency of the boom measured by the multi-axis inertial measurement unit sensor, and the environmental parameters detected by the environmental sensor.

[0010] In one possible implementation, the sensor network further includes a network topology management module, which is configured to: dynamically monitor the link quality indicators and remaining energy levels of each multimodal sensor node; periodically elect a cluster head node based on the link quality indicators and remaining energy levels using a distributed consensus algorithm; the cluster head node is used to aggregate real-time data collected by other sensor nodes within the cluster, compress the data, and upload it to the processing device; when any multimodal sensor node is detected to be disconnected from the network or malfunctions, a self-healing process within the cluster is triggered, restoring network coverage by waking up neighboring dormant nodes or reallocating the affiliation of child nodes, and generating node failure records and network reconstruction logs.

[0011] In one possible implementation, the processing device further includes a data preprocessing pipeline configured to perform multi-level preprocessing operations on the received real-time data before analyzing the data. The multi-level preprocessing operations include: filtering impulse noise from the real-time data using a sliding window-based median filtering algorithm; suppressing Gaussian white noise from the real-time data using a wavelet transform-based adaptive threshold denoising algorithm; unifying real-time data from heterogeneous sampling frequencies of different multimodal sensing nodes to a target sampling frequency using a cubic spline interpolation-based time-series resampling algorithm; and reducing the dimensionality of the denoised and resampled real-time data using a principal component analysis-based feature extraction algorithm, extracting the first K principal component features as input data for subsequent analysis, where K is a preset positive integer.

[0012] In one possible implementation, the processing device is configured to: analyze the received real-time data using a configured prediction model to obtain the torque prediction values ​​of each key component at each future time point, wherein the prediction model is a time-series prediction model based on a long short-term memory network or a gated recurrent unit, and the time-series prediction model is used to predict the torque at the future time point based on the historical torque change characteristics under similar working conditions.

[0013] In a second aspect, a lifting device is provided, which may include: a torque limiter for the lifting device as described in the first aspect.

[0014] This application utilizes a backscattering energy harvesting device to capture energy from the radio frequency radiation generated by the main control cabinet of the lifting equipment and convert it into electrical energy. This enables the distributed multimodal sensor nodes to be self-powered, fundamentally eliminating the dependence on batteries or external wired power supplies. This avoids sensor network failure due to power outages, significantly reduces maintenance costs, and improves the system's operational reliability in harsh environments. Simultaneously, the application collects real-time environmental data through an environmental perception module, processes the data to predict torque values ​​at future time points, and combines this with real-time environmental data to determine risk levels. This allows the safety control strategy to move beyond static models and integrate dynamic environmental factors to achieve forward-looking predictions, effectively responding to sudden risks and improving the safety of lifting equipment operations. Attached Figure Description

[0015] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments of this application will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a schematic diagram of a torque limiter provided in an embodiment of this application; Figure 2 This is a schematic diagram of an energy conversion circuit provided in an embodiment of this application. Detailed Implementation

[0017] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application. Unless otherwise defined, the technical or scientific terms used in this application should have the ordinary meaning understood by those skilled in the art. Words such as "comprising" or "including" mean that the element or object preceding the word covers the element or object listed after the word and its equivalents, but does not exclude other elements or objects. Words such as "connection," "coupled," or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect.

[0018] The torque limiter provided in this application embodiment can be applied to various lifting equipment such as pipe-laying machines, tower cranes, crawler cranes, and truck cranes. The lifting equipment may also include winches, directional wheels, jib bases, and slewing platforms. Figure 1 As shown, the torque limiter provided in this application may include: a sensor network, multiple backscatter energy harvesting devices (e.g., 3), an environmental sensing module, and a processing device.

[0019] The sensor network includes multiple multimodal sensing nodes (e.g., 3), which are distributed at key components of the lifting equipment to collect real-time data from the lifting equipment.

[0020] Specifically, multimodal sensing nodes can be integrated sensors capable of simultaneously detecting multiple physical quantities, or they can be a combination package of multiple single-function sensors. Key components refer to locations on the lifting equipment that are subject to complex forces and are prone to fatigue or overload during operation, such as winches, sheaves, boom base, and slewing platforms. Real-time data includes, but is not limited to, parameters such as wire rope tension, boom pitch angle, roll angle, vibration frequency, and ambient temperature and humidity. Multimodal sensing nodes transmit the collected real-time data to processing equipment via wireless communication.

[0021] It should be understood that although the figure shows three multimodal sensing nodes, in actual applications, the number of nodes can be flexibly increased or decreased according to the structural complexity of the lifting equipment and the monitoring requirements, as long as the function of covering key stress points is met.

[0022] The backscatter energy harvesting device is configured at a key component of the lifting equipment and connected to a multimodal sensor node arranged at the corresponding location. It is used to capture energy from the radio frequency radiation source of the lifting equipment and convert the captured energy into electrical energy to power the multimodal sensor node.

[0023] Specifically, radio frequency (RF) radiation sources refer to electromagnetic wave radiation sources generated by the lifting equipment itself during operation. For example, the main control cabinet of a lifting equipment contains numerous motor controllers, frequency converters, and wireless communication modules. These electrical components release RF energy during operation. Backscatter energy harvesting devices capture this scattered RF energy in the environment through antennas and convert it into DC power ranging from microwatts to milliwatts, directly powering nearby multimodal sensor nodes. This power supply method eliminates the need for built-in batteries or external cables in the multimodal sensor nodes, enabling passive wireless operation.

[0024] It should be noted that the connection between the backscatter energy harvesting device and the multimodal sensing node can be a physical circuit connection or a spatial placement to achieve energy coupling, as long as electrical energy transmission can be achieved.

[0025] An environmental perception module is used to collect real-time environmental data of the lifting equipment. This module can be a Simultaneous Localization and Mapping (SLAM) module, which may include LiDAR and cameras, installed at the leading edge of the lifting equipment's slewing platform. It is used to perceive obstacle information (such as obstacle location, status, and type) and ground slope around the lifting equipment. Real-time environmental data reflects the dynamic changes in the lifting equipment's operating environment. For example, when a construction vehicle or pedestrian enters the lifting equipment's working radius, the environmental perception module can capture this dynamic obstacle information in real time.

[0026] It should be understood that the environmental perception module is not limited to the SLAM module. In other embodiments, it can also be a millimeter-wave radar or an infrared sensor, as long as it can acquire information about the environmental status around the lifting equipment.

[0027] The processing equipment is connected to multimodal sensing nodes and environmental perception modules to analyze the received real-time data, obtain the torque prediction values ​​of each key component at each future time point, and use risk judgment rules to judge the risk of the torque prediction values ​​and real-time environmental data, determine the risk level, and control the lifting equipment to execute the adjustment strategy corresponding to the risk level.

[0028] Specifically, the processing device can be a smart touch-screen all-in-one machine with a built-in edge computing chip, which interacts with sensor networks and environmental perception modules via a communication bus (such as a CAN bus) or wireless network. After receiving real-time data, the processing device analyzes the data using a built-in predictive model to deduce the torque change trend of each key component in the next few seconds (e.g., within the next 3 seconds), obtaining the torque prediction value. Subsequently, the processing device inputs the torque prediction value and real-time environmental data into risk assessment rules for comprehensive evaluation. For example, even if the torque prediction value has not yet reached the rated torque, if the real-time environmental data indicates a sudden risk such as excessive ground slope or approaching dynamic obstacles, the processing device will also increase the risk level accordingly. Finally, the processing device issues control commands to the main control system of the lifting equipment based on the determined risk level, such as outputting a warning, limiting power output, or triggering emergency braking. Through this closed-loop system of backscattering energy harvesting + environmental perception + torque prediction, the processing device can integrate mechanical parameters and dynamic environmental factors to achieve forward-looking prediction, effectively respond to sudden risks, and improve the safety and reliability of lifting equipment operation.

[0029] Based on the above embodiments, this embodiment elaborates on the energy harvesting mechanism and hardware circuit of the backscatter energy harvesting device. The radio frequency radiation source is the main control cabinet of the lifting equipment. The backscatter energy harvesting device is used to capture energy from the radio frequency radiation generated when the main control cabinet is working through an antenna, and converts the captured energy into electrical energy through an energy conversion circuit to power the multimode sensing nodes arranged at the corresponding locations.

[0030] Specifically, the main control cabinet of the lifting equipment is the core electrical cabinet that controls the operation of the entire machine. It contains numerous motor controllers, frequency converters, and wireless communication modules. These electrical components inevitably release radio frequency radiation into the surrounding space when operating at high frequencies. The antenna of the backscatter energy harvesting device captures this scattered radio frequency energy in the environment, much like collecting signals. It should be understood that, in addition to the main control cabinet 6, the radio frequency radiation source can also be other high-frequency radiating components on the lifting equipment, such as the heat dissipation radiation of the high-power frequency converter itself, or wireless communication base stations deployed at the construction site. This embodiment preferably uses the main control cabinet 6 as the energy source. Its innovative contribution lies in the fact that the main control cabinet is typically fixedly installed on the slewing platform or chassis of the lifting equipment, relatively close to and in a fixed position with the key components. This provides a stable and continuous radio frequency field strength. Compared to capturing random radio frequency signals in the environment, using the main control cabinet 6 as a directional energy harvester significantly improves the efficiency and reliability of energy harvesting, making long-term operation of passive sensing nodes possible in industrial scenarios.

[0031] Furthermore, the energy conversion circuit includes an antenna, a rectifier circuit, a hybrid energy storage unit, a voltage regulator circuit, and a load modulation module connected in sequence. The rectifier circuit converts the AC signal received by the antenna into a DC signal. The hybrid energy storage unit includes a supercapacitor and a micro battery for storing the converted DC energy. The voltage regulator circuit regulates the voltage of the energy output from the hybrid energy storage unit. The load modulation module modulates the collected energy according to the requirements of the multimodal sensing node and achieves backscatter communication by changing the antenna's reflection characteristics.

[0032] like Figure 2 As shown, the antenna receives radio waves in space and outputs a weak AC signal to the rectifier circuit. The rectifier circuit typically uses a diode bridge rectifier structure or a voltage doubler rectifier structure to convert the AC signal into a DC signal. The converted DC power is then transferred to a hybrid energy storage unit for storage. The hybrid energy storage unit contains a supercapacitor, a micro-battery, and a switching control unit. The supercapacitor has extremely high power density and fast charge / discharge characteristics to handle the transient high current demands of multimodal sensing nodes during data acquisition and wireless transmission; the micro-battery has high energy density for long-term energy storage and stable power supply.

[0033] The switching control unit dynamically switches between the supercapacitor and the micro-battery based on the operating state of the multimodal sensing node: when the node is in sleep or low-power acquisition mode, the micro-battery mainly provides microamplitude current and slowly charges the supercapacitor; when the node wakes up for RF communication, the switching control unit controls the supercapacitor to release milliamplitude transient current, thereby avoiding voltage drop in the micro-battery under instantaneous high-current discharge, extending battery life and improving transient power compensation capability. A voltage regulator circuit is connected to the output of the hybrid energy storage unit, using a linear regulator or switching regulator to clamp the fluctuating energy storage voltage to the stable operating voltage required by the multimodal sensing node. The power input of the load modulation module is connected to the output of the voltage regulator circuit, its signal input is connected to the node's microcontroller, and the antenna connection is connected back to the antenna. The load modulation module indirectly affects the load characteristics by changing the antenna's input impedance, thereby changing the antenna's reflection characteristics and achieving backscatter communication of data. This process does not require active transmission of RF signals, greatly reducing communication power consumption.

[0034] As another alternative implementation, the hybrid energy storage unit can also adopt other combinations, such as using only supercapacitor arrays, and combining multiple supercapacitors in series and parallel to simultaneously meet the requirements of energy density and power density, thereby completely eliminating the need for micro batteries in extreme temperature environments and further improving the maintenance-free nature of the system.

[0035] Based on the above embodiments, this embodiment elaborates on the joint risk assessment rules of environmental perception and torque prediction.

[0036] The processing equipment is used to: detect whether the predicted torque value at any future time point for any critical component is greater than the rated torque, and whether the real-time environmental data meets the preset obstacle conditions; If the predicted torque value at any future time point is not greater than the rated torque of the first coefficient, and the real-time environmental data meets the preset safety conditions, then the risk level is determined to be low risk. If the predicted torque value at any future time point is detected to be between the rated torque of the first coefficient and the rated torque of the second coefficient, or if the predicted torque value is not greater than the rated torque and the real-time environmental data indicates the presence of dynamic obstacles, then the risk level is determined to be medium risk. If the predicted torque value at any future time point is detected to be greater than the rated torque of the second coefficient, or if the predicted torque value is between the rated torque of the second coefficient and the rated torque, and the real-time environmental data indicates that the ground slope is greater than the configured safety range, then the risk level is determined to be high risk.

[0037] Specifically, the processing equipment has a built-in processing engine. Real-time environmental data includes obstacle information and ground slope. Preset obstacle conditions include preset safety conditions, the presence of dynamic obstacles, and ground slope exceeding the configured safety range. For any critical component of the lifting equipment (such as the base of the boom), the processing equipment detects whether the predicted torque value at a future point in time (e.g., 3 seconds later) exceeds the rated torque, and makes a comprehensive judgment based on real-time environmental data collected by the environmental perception module. If the predicted torque value is not greater than the rated torque of the first coefficient, and the real-time environmental data meets the preset safety conditions (i.e., no obstacles and ground slope within the safety range), the risk level is determined to be low risk. If the predicted torque value is between the rated torque of the first coefficient and the rated torque of the second coefficient, or if the predicted torque value is not greater than the rated torque but the real-time environmental data indicates the presence of dynamic obstacles (such as pedestrians or construction vehicles entering the working radius), the risk level is determined to be medium risk. If the predicted torque value is greater than the rated torque of the second coefficient, or if the predicted torque value is between the rated torque of the second coefficient and the rated torque, and real-time environmental data indicates that the ground slope is greater than the configured safe range (e.g., the angle with the horizontal plane is greater than 8°), then the risk level is determined to be high risk. Through this multi-dimensional joint judgment logical framework, the processing equipment can integrate mechanical parameters and dynamic environmental factors to achieve forward-looking prediction and effectively respond to sudden risks.

[0038] Furthermore, since the first coefficient is less than the second coefficient, the processing equipment is also used to: when the risk level is determined to be low risk, control the lifting equipment to output early warning information and instruct the driver to actively adjust the operation; when the risk level is determined to be medium risk, control the lifting equipment to limit the power output of the dangerous action and allow the execution of the safe action; when the risk level is determined to be high risk, control the lifting equipment to forcibly cut off the power of the dangerous action and trigger emergency braking.

[0039] Specifically, when the risk level is low, the handling equipment controls the lifting equipment to output warning messages, such as displaying a warning message box on the touch screen (e.g., "The boom torque will reach 85% of the rated value in 30 seconds, please slow down the boom extension speed"), and / or alerting the driver with a slight beep (1 time / second), and instructing the driver to actively adjust the operation (e.g., reduce the lifting speed, stop the boom extension). When the risk level is medium, the handling equipment controls the lifting equipment to restrict the power output of dangerous actions, such as cutting off the power to the boom extension cylinder, prohibiting dangerous actions such as boom extension, increasing the boom radius, and lifting heavy objects, but allowing safe actions such as boom retraction, boom lowering, load lowering, and slewing to avoid obstacles. When the risk level is high, the handling equipment controls the lifting equipment to forcibly cut off the power of dangerous actions (such as locking the hydraulic valves of hoisting, boom extension, and luffing) and trigger emergency braking (boom lock, hoisting mechanism brake). At the same time, the touch screen automatically pops up hazard avoidance suggestions (such as retracting the boom by 2 meters + turning to an area without obstacles + lowering the load), so that the driver can choose to perform automatic hazard avoidance or manual operation through the touch screen.

[0040] It should be understood that the selection of the first and second coefficients is crucial to system performance. If the first coefficient is too small (e.g., 0.7), the system will be overly sensitive to slight torque fluctuations, frequently triggering medium-risk limiting strategies and severely impacting the operating efficiency of the lifting equipment. If the first coefficient is too large (e.g., 0.95), although false alarms are reduced, the time window for system response and operator adjustment is extremely short, making it easy to directly enter a high-risk state in the event of a sudden environmental change. Similarly, if the second coefficient is too small (e.g., 0.8), emergency braking will be triggered prematurely, causing unnecessary shutdowns. If the second coefficient is too large (e.g., 0.95), the response will be severely delayed, and the torque may have already exceeded the limit before the expected time point. This embodiment, through extensive engineering verification, selects a first coefficient of 0.8 and a second coefficient of 0.9, achieving an optimal balance between safety and operating efficiency: the threshold of 0.8 can provide early warning and limit some actions, allowing the operator sufficient adjustment time, while the threshold of 0.9 serves as a final safety line, ensuring sufficient braking distance before the torque approaches its limit, thereby maximizing operating efficiency while ensuring safety.

[0041] Furthermore, based on the above embodiments, this embodiment elaborates on sensor network deployment, self-healing mechanisms, data preprocessing, and prediction algorithms in detail.

[0042] Key components of the lifting equipment include the winch, sheave, boom root, and slewing platform. Multimodal sensing nodes include a three-pulley weight sensor fixed to the winch and sheave, a multi-axis inertial measurement unit (IMU) sensor mounted at the boom root, and an environmental sensor mounted on the slewing platform. Real-time data includes the wire rope tension detected by the three-pulley weight sensor, the boom pitch angle, roll angle, and vibration frequency measured by the multi-axis IMU sensor, and environmental parameters detected by the environmental sensor.

[0043] Specifically, the three-pulley weight sensor, clamped to the wire rope, measures the rope tension using the three-point bending principle, thus indirectly reflecting the weight of the lifting load. The multi-axis inertial measurement unit (IMU) sensor is fixed to the stress concentration area at the base of the boom, used to monitor the boom's attitude changes and vibration characteristics in real time during lifting, slewing, and luffing processes. Environmental sensors are deployed on the top of the slewing platform to collect environmental parameters such as wind speed, temperature, and humidity to compensate for the mechanical sensor measurements. It should be understood that key components are not limited to the above four types; they can also include the boom head, counterweight area, etc., as long as they cover the main stress points of the lifting equipment.

[0044] The sensor network also includes a network topology management module, which is configured to: dynamically monitor the link quality indicators and remaining energy levels of each multimodal sensor node; periodically elect cluster head nodes based on the link quality indicators and remaining energy levels using a distributed consensus algorithm; aggregate real-time data collected by other sensor nodes within the cluster and upload the compressed data to the processing device; and trigger a self-healing process within the cluster when any multimodal sensor node is detected to be disconnected from the network or malfunctions, thereby restoring network coverage by waking up neighboring dormant nodes or reallocating the affiliation of child nodes, and generating node failure records and network reconstruction logs.

[0045] Specifically, the network topology management module operates as follows: First, each multimodal sensor node periodically broadcasts a beacon frame containing its own identifier, remaining energy percentage, and link quality indicators. Then, nodes exchange beacon frames using a distributed consensus algorithm, and the node with the highest remaining energy and best link quality is elected as the cluster head node. The cluster head node is responsible for collecting data from its cluster members and compressing and forwarding it to reduce energy consumption during long-distance communication with processing devices. When a node experiences a link interruption due to metal obstruction or energy depletion, the network topology management module triggers a self-healing process: waking up a physically adjacent dormant node as a relay, or reassigning the failed node's child nodes to other cluster heads, thereby restoring network coverage. Simultaneously, the system automatically generates node failure records and network reconstruction logs to provide a basis for subsequent maintenance. This embodiment overcomes the problem of data acquisition interruption caused by single-point failures in harsh construction environments with fixed network structures through dynamic topology and self-healing mechanisms.

[0046] The processing equipment also includes a data preprocessing pipeline configured to perform multi-stage preprocessing operations on the received real-time data before analysis. These multi-stage preprocessing operations include: filtering impulse noise from the real-time data using a sliding window-based median filtering algorithm; suppressing Gaussian white noise from the real-time data using a wavelet transform-based adaptive threshold denoising algorithm; unifying real-time data from heterogeneous sampling frequencies from different multimodal sensing nodes to a target sampling frequency using a cubic spline interpolation-based time-series resampling algorithm; and reducing the dimensionality of the denoised and resampled real-time data using a principal component analysis-based feature extraction algorithm, extracting the first K principal component features as input data for subsequent analysis, where K is a preset positive integer.

[0047] Specifically, the execution logic of the multi-level preprocessing operation is as follows: The first stage is median filtering. The processing device sets a sliding window of length N on the real-time data sequence, and uses the median of each data point within the window as the output of the current data point to eliminate sudden impulse noise caused by electromagnetic interference. Its core expression is:

[0048] in, Indicates the first Each filtered output data point Indicates the first One raw real-time data point, This represents the half width of the sliding window. , This indicates the median operation. Acquired directly from multimodal sensing nodes. Preset according to the sampling frequency.

[0049] The second stage is wavelet transform denoising. For Gaussian white noise, the processing equipment performs discrete wavelet decomposition on the filtered data and then uses adaptive threshold quantization to process the high-frequency wavelet coefficients before wavelet reconstruction. Its core expression is:

[0050]

[0051] in, Representing scale Peaceful relocation The wavelet coefficients below, This represents the input discrete signal sequence. Represents the discrete wavelet basis functions. This represents the wavelet coefficients after thresholding. This represents the adaptive threshold, which is calculated using a method based on Stein's unbiased risk estimation. The data is obtained from the output of the preceding median filter. The third stage involves cubic spline interpolation resampling. Because different sensors have different sampling frequencies, the processing equipment uses cubic spline interpolation to unify the heterogeneous data to the target sampling frequency. For adjacent data points... and interpolation function The expression is:

[0052] in, Representing an interval Interpolation function within, Indicates the target sampling time point. and Represents the timestamp of the original data. and Represents the original data value. This represents the spline coefficients obtained by solving for the boundary continuity condition. The target sampling time point is obtained from the output of the pre-stage denoising stage. Generated by the target sampling frequency configured by the system. Fourth level: Principal component analysis (PCA) dimensionality reduction. Before performing PCA, the processing device first standardizes the real-time data matrix X, which has undergone denoising and resampling, to eliminate the influence of different physical dimensions. The standardization formula is: Where μ is the mean of each feature column, and σ is the standard deviation of each feature column. After standardization, Each dimension is a dimensionless pure numerical value. Subsequently, the processing device calculates... The eigenvalues ​​and eigenvectors of the covariance matrix are used to construct a projection matrix by selecting the eigenvectors corresponding to the K largest eigenvalues, thus mapping the high-dimensional data to a low-dimensional space. Its core expression is:

[0053] in, This represents the feature matrix after dimensionality reduction, i.e., the extracted first K principal component features. This represents the real-time data matrix after denoising and resampling. This represents the projection matrix composed of the eigenvectors corresponding to the first K largest eigenvalues, where K is a preset positive integer and is less than or equal to the total number of original features. Preferably, the value of K is dynamically determined by the cumulative variance contribution rate (usually a threshold set to 85%~95%) to maximize the dimensionality reduction effect while preserving the integrity of the original information. Acquired from the output of the previous resampling stage. Through the The covariance matrix is ​​obtained through eigenvalue decomposition. Through multi-level preprocessing, the system effectively suppresses noise while preserving the characteristics of signal abrupt changes, and reduces the computational burden of subsequent prediction models.

[0054] The processing equipment is used to: analyze the received real-time data using a configured prediction model to obtain the torque prediction values ​​of each key component at each future time point. The prediction model is a time-series prediction model based on a long short-term memory network or a gated cyclic unit. The time-series prediction model is used to predict the torque at future time points based on the historical torque change characteristics under similar working conditions.

[0055] Specifically, the prediction model takes the preprocessed first K principal component features as input, captures long-term dependencies in the time-series data through a Long Short-Term Memory network or gated recurrent units, and outputs the torque prediction values ​​of each key component at each future time point. To achieve millisecond-level inference response on the edge computing chip built into the processing device, this embodiment adopts a lightweight deployment strategy for the prediction model. Specifically, this includes: converting network weights from 32-bit floating-point numbers to 8-bit integer representations using model quantization technology to reduce memory usage and multiplication-accumulation computation overhead; and removing redundant hidden layer channels and neuron connections in the network through structured pruning to reduce the number of floating-point operations during inference. After quantization and pruning, the model size is compressed to one-quarter of its original size, and the inference latency is controlled within 10 milliseconds, thus ensuring real-time and accurate torque prediction even with limited edge computing power.

[0056] This embodiment provides a lifting device. The lifting device includes the torque limiter of the lifting device described in any of the above embodiments.

[0057] Specifically, lifting equipment can be various types of heavy lifting machinery such as tower cranes, crawler cranes, truck cranes, or gantry cranes. The torque limiter, as a core component for safety control of the lifting equipment, has its various modules with specific assembly positions and system interaction relationships within the entire machine. Multimodal sensor nodes of the sensor network are distributed across key components of the lifting equipment, such as winches, directional wheels, the base of the boom, and the slewing platform. Backscatter energy harvesting devices are located nearby on these key components, using the radio frequency radiation generated by the main control cabinet of the lifting equipment to power the sensor nodes. An environmental perception module is installed at the leading edge of the slewing platform of the lifting equipment to obtain an unobstructed view of the surrounding environment. The processing equipment is typically located in the operator's cab of the lifting equipment and is connected to the main control cabinet and various actuators via a communication bus. At the system interaction level, the processing equipment predicts future torque based on sensor network and environmental data, assesses risks, and then directly issues control commands to the main control cabinet. The main control cabinet then drives corresponding hydraulic valve locks or electrical circuits to execute adjustment strategies such as limiting power output or emergency braking. It should be understood that the specific internal structure of the torque limiter will not be elaborated here. As the central hub for the safety monitoring of the entire machine, it fundamentally improves the safety and reliability of the lifting equipment under complex working conditions through the combination of passive wireless sensing and forward-looking risk prediction.

[0058] Although preferred embodiments have been described in this application, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the embodiments of this application.

[0059] Obviously, those skilled in the art can make various modifications and variations to the embodiments of this application without departing from the spirit and scope of the embodiments of this application. Therefore, if these modifications and variations to the embodiments of this application fall within the scope of the claims in this application and their equivalents, then this application also intends to include these modifications and variations.

Claims

1. A torque limiter for lifting equipment, characterized in that, The torque limiter includes: The sensor network includes multiple multimodal sensing nodes, which are distributed at key components of the lifting equipment to collect real-time data from the lifting equipment. A backscattering energy harvesting device is configured at the key component of the lifting equipment and connected to the multimodal sensing node arranged at the corresponding position. It is used to capture energy from the radio frequency radiation source of the lifting equipment and convert the captured energy into electrical energy to power the multimodal sensing node. The environmental sensing module is used to collect real-time environmental data of the lifting equipment; The processing device, connected to the multimodal sensing node and the environmental perception module, is used to analyze the received real-time data, obtain the torque prediction values ​​of each key component at each future time point, and use risk judgment rules to judge the risk of the torque prediction values ​​and the real-time environmental data to determine the risk level, so as to control the lifting equipment to execute the adjustment strategy corresponding to the risk level.

2. The torque limiter according to claim 1, characterized in that, The radio frequency radiation source is the main control cabinet of the lifting equipment; The backscatter energy harvesting device is used to capture energy from the radio frequency radiation generated when the main control cabinet is working through an antenna, and to convert the captured energy into electrical energy through an energy conversion circuit to power the multimodal sensing nodes arranged at the corresponding locations.

3. The torque limiter according to claim 2, characterized in that, The energy conversion circuit includes the antenna, rectifier circuit, hybrid energy storage unit, voltage regulator circuit and load modulation module connected in sequence; The rectifier circuit is used to convert the AC signal received by the antenna into a DC signal; The hybrid energy storage unit includes a supercapacitor and a micro battery for storing converted DC power. The voltage regulator circuit is used to regulate the voltage of the electrical energy output by the hybrid energy storage unit. The load modulation module is used to modulate the collected electrical energy according to the requirements of the multimodal sensing node, and to achieve backscatter communication by changing the reflection characteristics of the antenna.

4. The torque limiter according to claim 1, characterized in that, The processing device is used for: For any critical component, detect whether the predicted torque value at any future time point is greater than the rated torque, and whether the real-time environmental data meets the preset obstacle conditions; If the predicted torque value at any future time point is not greater than the rated torque of the first coefficient, and the real-time environmental data meets the preset safety conditions, then the risk level is determined to be low risk. If the predicted torque value at any future time point is detected to be between the rated torque of the first coefficient and the rated torque of the second coefficient, or if the predicted torque value is not greater than the rated torque and the real-time environmental data indicates the presence of dynamic obstacles, then the risk level is determined to be medium risk. If the predicted torque value at any future time point is detected to be greater than the rated torque of the second coefficient, or if the predicted torque value is between the rated torque of the second coefficient and the rated torque, and the real-time environmental data indicates that the ground slope is greater than the configured safety range, then the risk level is determined to be high risk.

5. The torque limiter according to claim 4, characterized in that, The processing device is also used for: When the risk level is determined to be low, the crane is controlled to output a warning message and the operator is instructed to actively adjust the operation. When the risk level is determined to be medium risk, the lifting equipment is controlled to limit the power output of dangerous actions and allow the execution of safe actions; When the risk level is determined to be high, the lifting equipment is controlled to forcibly cut off the power for the dangerous action and trigger emergency braking.

6. The torque limiter according to claim 1, characterized in that, The key components of the lifting equipment include the winch, directional rollers, boom base, and slewing platform; The multimodal sensing node includes a three-pulley weight sensor fixed to the winch and the directional wheel, a multi-axis inertial measurement unit sensor installed at the root of the boom, and an environmental sensor installed on the slewing platform; The real-time data includes the tension of the wire rope detected by the three pulley weight sensor, the pitch angle, roll angle and vibration frequency of the boom measured by the multi-axis inertial measurement unit sensor, and the environmental parameters detected by the environmental sensor.

7. The torque limiter according to claim 1, characterized in that, The sensor network also includes a network topology management module, which is configured as follows: The link quality indicators and remaining energy levels of each multimodal sensor node are dynamically monitored. Based on the link quality indicators and remaining energy levels, a distributed consensus algorithm is used to periodically elect a cluster head node. The cluster head node is used to aggregate real-time data collected by other sensor nodes in the cluster, compress the data, and then upload it to the processing device. When any multimodal sensor node is detected to be out of the network or malfunctioning, a self-healing process within the cluster is triggered. This process restores network coverage by waking up nearby dormant nodes or reallocating child node affiliations, and generates node failure records and network reconstruction logs.

8. The torque limiter according to claim 1, characterized in that, The processing equipment further includes a data preprocessing pipeline, configured as follows: Before analyzing the received real-time data, multi-level preprocessing operations are performed on the real-time data; The multi-level preprocessing operation includes: using a sliding window-based median filtering algorithm to filter out impulse noise in the real-time data; An adaptive threshold denoising algorithm based on wavelet transform is used to suppress Gaussian white noise in the real-time data; a time series resampling algorithm based on cubic spline interpolation is used to unify the real-time data from different multimodal sensing nodes with heterogeneous sampling frequencies to the target sampling frequency; and a feature extraction algorithm based on principal component analysis is used to reduce the dimensionality of the real-time data after denoising and resampling, and extract the first K principal component features as input data for subsequent analysis, where K is a preset positive integer.

9. The torque limiter according to claim 1, characterized in that, The processing device is used for: The received real-time data is analyzed using a configured prediction model to obtain the torque prediction values ​​of each key component at each future time point. The prediction model is a time-series prediction model based on a long short-term memory network or a gated cyclic unit. The time-series prediction model is used to predict the torque at the future time point based on the historical torque change characteristics under similar working conditions.

10. A lifting device, characterized in that, include: The torque limiter for the lifting equipment as described in any one of claims 1 to 9.

Images