High-altitude operation safety alarm method and device

Abstract

The invention discloses a high-altitude operation safety alarm method and device, and belongs to the field of data processing. The high-altitude operation safety alarm method comprises the steps that S1, tension data and speed data of a safety rope are collected; calculating the second-order difference of the speed data; s2, performing similarity matching on the collected tension data and a preset tension safety range template to obtain a tension matching degree; s3, performing similarity matching on the speed change rate data and a preset speed change rate safety range template to obtain a speed change rate matching degree; s4, judging whether the safety rope is in a safe state or not according to the tension matching degree and the speed change rate matching degree; if yes, the step S5 is executed, and if not, the step S6 is executed; s5, generating a first control signal according to a preset fuzzy control rule; s6, classifying the dangerous states by using a support vector machine (SVM) algorithm to obtain danger levels; generating a second control signal according to the danger level; the high-altitude operation safety rope aims at solving the problem that in the prior art, high-altitude operation safety ropes are low in reliability, and the reliability is improved.

Description

Technical Field

[0001] The present application relates to the field of data processing, and in particular to a method and device for high-altitude operation safety alarm. Background Art

[0002] With the continuous development of modern construction projects, aerial work has become an indispensable part of the construction process. In the field of aerial work, such as building exterior wall construction, high-altitude scaffolding construction, steel structure beam and column welding, high-altitude equipment installation and maintenance, etc., safety ropes are key equipment to ensure the safety of workers. However, traditional safety ropes still have many problems in practical applications and are difficult to meet the growing demand for safety protection.

[0003] On the one hand, the length of traditional safety ropes is mostly fixed, and the few adjustable safety ropes can only be simply adjusted manually, which seriously limits the range of activities of operators in actual operations. The redundant length has a safety impact on the operation and activities of personnel, reduces work efficiency, and frequent manual operations pose great risks in high-altitude environments.

[0004] On the other hand, the anti-fall performance of traditional safety ropes is limited. Once a fall accident occurs, it is often difficult to effectively buffer the impact force, which can easily cause serious injuries to the operator. In addition, the existing safety ropes lack automatic alarm functions. When the safety rope is abnormally stretched, worn, or the connection is loose, it is impossible to send an alarm to the operator and ground monitoring personnel in time, resulting in safety hazards that cannot be discovered and handled in time. Summary of the invention

[0005] In view of the low reliability of the high-altitude work safety rope in the prior art, the present application provides a high-altitude work safety alarm device and method, which performs similarity matching between the tension data and speed data of the safety rope and a preset safety range template, and generates corresponding control parameters based on different matching results, thereby improving the reliability of the safety rope.

[0006] The purpose of this application is achieved through the following technical solutions.

[0007] One aspect of the present application provides a high-altitude work safety alarm device, comprising: an electric winder, a tension sensor, a controller and a speed gear box; the electric winder is used to control the extension and contraction of the safety rope; the tension sensor is used to collect the tension data of the safety rope and transmit the collected tension data to the controller; the speed sensor is used to collect the speed data of the safety rope and transmit the collected speed data to the controller; wherein the speed data includes the extension or contraction speed; the controller generates a control signal according to the collected tension data and speed data; the speed gear box adjusts the output speed and torque of the electric winder according to the control signal;

[0008] Another aspect of the present application provides a high-altitude operation safety alarm method, S1, collecting tension data and speed data of the safety rope; and calculating the second-order difference of the speed data to obtain speed change rate data; S2, performing similarity matching between the collected tension data and a preset tension safety range template to obtain tension matching; S3, performing similarity matching between the speed change rate data and a preset speed change rate safety range template to obtain speed change rate matching; S4, judging whether the safety rope is in a safe state according to the tension matching and the speed change rate matching; if the safety rope is in a safe state, proceeding to step S5, if not, proceeding to step S6; S5, when the safety rope is in a safe state, generating a first control signal according to a preset fuzzy control rule; adjusting the output speed and torque of the electric winder according to the first control signal; S6, for the safety rope in a dangerous state, using the support vector machine SVM algorithm to classify the dangerous state to obtain the danger level; generating a second control signal according to the danger level;

[0009] Further, S1, calculates the second-order difference of the speed data to obtain the speed change rate data, including: for each time point , get the corresponding speed data ; Calculate the difference in velocity changes at adjacent time points to obtain the second-order velocity difference value : , calculate the second-order difference of velocity Time difference with adjacent time points The ratio of the speed change acceleration is obtained. : ;

[0010] Further, S3, the speed change rate data is matched with the preset speed change rate safety range template for similarity to obtain the speed change rate matching degree, including: constructing a distance matrix between the speed change rate data sequence and the speed change rate safety range template sequence: and speed change rate safety range template sequence As the input sequence of the DTW algorithm respectively; calculate the distance between each pair of elements in the two sequences , construct the distance matrix ;in, Indicates the rate of change of speed Corresponding change rate in the safety range template distance;

[0011] Search for the optimal matching path in the distance matrix D: Use the dynamic programming algorithm to Start by recursively calculating , and get the cumulative distance matrix ; represents the minimum cumulative distance between the first i elements of sequence A and the first j elements of sequence B; Get the end point of the best matching path, start from the end point, and search backward to get the entire best matching path;

[0012] Calculate the speed change rate matching degree: sum the distance values ​​on the optimal matching path to get the total path distance ; The total path distance Divide by the path length length(path) to get the average distance , as the difference between the speed change rate data sequence and the safety range template sequence; speed change rate matching degree Dissimilarity is inversely proportional, which can be expressed as ;

[0013] The acceleration of speed change is obtained by calculating the second-order difference of speed data. The DTW algorithm not only compares the similarity of speed values ​​at different time points, but also compares the similarity of speed change trends at adjacent time points, making the similarity match more comprehensive, thereby improving the accuracy of safety rope status judgment.

[0014] Furthermore, step S2 uses an improved dynamic time warping algorithm DTW to perform similarity matching on the collected tension data and the preset tension safety range template to obtain a tension matching degree, including: constructing a distance matrix between the tension data sequence and the tension safety range template sequence: And the tension safety range template sequence As the input sequence of the improved DTW algorithm, the first-order difference of the tension data sequence F and the tension safety range template sequence G is calculated to obtain the corresponding tension change rate sequence and ; Calculate the distance matrix Elements When considering the tension value at the current position and , and also incorporates the contextual information of the k points before and after it: Calculate the mean tension of the i-th position and the k points before and after it in the tension sequence F and the mean rate of change of tension ; Calculate the mean tension of the jth position and the k points before and after it in the tension safety range template sequence G and the mean rate of change of tension ; Calculate the corrected distance based on the mean ;in and is the weight of the two items; specifically, the average or weighted average of the tension value and the rate of change of the current position and the k points before and after it is calculated, and then the distance is calculated. Introducing context information can smooth data noise to a certain extent and improve the accuracy of matching.

[0015] In the distance matrix Search for the optimal matching path in: Use a dynamic programming algorithm with local constraints, from Start by recursively calculating , and get the cumulative distance matrix ;During the recursive process, the slope and step size of the matching path are restricted to avoid excessive bending; represents the minimum cumulative distance between the first i elements of sequence F and the first j elements of sequence G under local constraints; Get the end point of the optimal matching path, start from the end point, and search backward to get the entire optimal matching path ; Specifically, the traditional DTW algorithm allows the matching path to have an arbitrary shape, which may lead to an overly curved matching path and unreasonable matching results. In order to limit the shape of the matching path, local constraints such as slope limit and step limit can be introduced into the dynamic programming recursive formula. These constraints require the matching path to meet certain shape requirements in the local range, so as to obtain more reasonable and smoother matching results.

[0016] Calculate the tension matching degree and match the path Sum the distance values ​​on the path to get the total path distance ; The total distance Divide by the path length , and get the average distance , as the difference between the optimized tension data sequence and the safety range template sequence; tension matching degree Dissimilarity with the optimized is inversely proportional, which can be expressed as .

[0017] Further, in step S5, when the safety rope is in a safe state, a first control signal is generated according to a preset fuzzy control rule; the first control signal is used to adjust the output speed and torque of the electric rope reel, including: calculating the mean value of the speed change rate data obtained in step S1 and standard deviation : Mean Reflects the average level of speed change rate data; standard deviation Reflects the fluctuation degree of speed change rate data;

[0018] According to the mean and standard deviation Determine the movement status of personnel: If and All within the preset threshold range [ E minutes , E max ] and [ s minutes , s max ] If the personnel movement status is within Mark as stable; otherwise, the personnel motion state Marked as non-stationary;

[0019] Determine whether the tension matching degree M obtained in step S2 is greater than a preset threshold :if , then the safety rope stress state Mark as uniform; otherwise, the safety rope stress state Marked as uneven;

[0020] S54, the movement status of personnel And the stress state of the safety rope As two input variables of the fuzzy controller: according to the preset personnel motion state membership function ,calculate The corresponding membership ; According to the preset safety rope force state membership function ,calculate The corresponding membership ;

[0021] S55, according to the preset fuzzy control rule library , using Mamdani inference method to calculate the membership and Perform fuzzy reasoning: Each fuzzy control rule Like "If satisfy ,and satisfy ,So and "; where v is the output speed of the electric rope reel, T is the output torque, is a fuzzy set; according to the membership degree and , calculate the trigger strength of each rule ;according to Perform weighted summation on the rule conclusions to obtain the fuzzy set of output speed v and torque T and ;

[0022] The center of gravity defuzzification algorithm is used to transform the speed fuzzy set and torque fuzzy set Converted into precise speed control and torque control , generating the first control signal: and They are and The corresponding center of gravity coordinates; the first control signal includes the speed control amount and torque control .

[0023] By introducing the statistical characteristics of the speed change rate data and using the mean and standard deviation to judge the movement state of the personnel, it is possible to effectively distinguish between smooth movement and non-smooth movement; using the threshold of the tension matching degree to judge the stress state of the safety rope, it is possible to sensitively detect abnormal stress; by fuzzifying the movement state of the personnel and the stress state of the safety rope, using them as the input of the fuzzy controller, and combining them with the preset fuzzy control rule library for reasoning, the output of the electric rope reel can be intelligently adjusted according to different working conditions. The generated speed control amount and torque control amount can dynamically adapt to various situations during high-altitude operations, effectively improving operation safety.

[0024] Further, according to the first control signal, the output speed and torque of the electric rope reel are adjusted, including: extracting the speed control amount from the first control signal and torque control : Speed ​​control amount Indicates the target speed of the electric rope reel output by the fuzzy controller; torque control value Indicates the target torque of the electric rope reel output by the fuzzy controller; obtains the current actual output speed of the electric rope reel and actual output torque :Measure the actual output speed through the electric rope reel’s own speed sensor ; The actual output torque is measured by the torque sensor of the electric rope reel itself ; Calculate the speed control amount The actual output speed The deviation Δv and the torque control amount The actual output torque Deviation ΔT: Speed ​​deviation ; Torque deviation ; According to the speed deviation and torque deviation , using PID control algorithm to calculate the speed control signal of the electric rope reel and torque control signal : Speed ​​control signal ;in, They are the proportional, integral and differential coefficients of the speed PID controller; the torque control signal ;in, They are the proportional, integral and differential coefficients of the torque PID controller respectively; the speed control signal and torque control signal Input the speed controller and torque controller of the electric rope reel respectively: the speed controller controls the signal according to the speed Adjust the motor speed of the electric rope reel to make the actual output speed Tracking target speed ; The torque controller controls the torque signal Adjust the motor torque of the electric rope reel to make the actual output torque Tracking target torque ; Repeat until the speed deviation Δv and torque deviation ΔT converge to the allowable range: the allowable range is preset according to the control accuracy of the electric rope reel; when and When , it is considered that the electric rope reel has reached the target speed and target torque, and the adjustment process ends; and The preset speed error and torque error allowable thresholds.

[0025] According to the deviation between the target speed and target torque output by the fuzzy controller and the actual output speed and actual output torque of the electric rope reel, a speed control signal and a torque control signal are generated, which can realize the precise control and rapid tracking of the output of the electric rope reel and improve the dynamic response performance of the system; by setting the speed error threshold and the torque error threshold, the electric rope reel is repeatedly adjusted until the error converges, which can ensure that the control accuracy meets the requirements;

[0026] Further, in step S6, for the safety rope in a dangerous state, the support vector machine SVM algorithm is used to classify the dangerous state to obtain the danger level; and a second control signal is generated according to the danger level, including: judging whether the safety rope is in a dangerous state: according to the mean value of the speed change rate data calculated in step S1 and standard deviation ,judge or Whether it exceeds the preset threshold range [ E minutes , E max ] or [ s minutes , s max ] ; According to the tension matching degree M calculated in step S2, determine whether M is less than the preset threshold ;if or Exceeds the preset threshold range, and M is less than , the safety rope is considered to be in a dangerous state; otherwise, the safety rope is considered to be in a safe state;

[0027] For the safety rope in a dangerous state, construct a feature vector: extract the mean, standard deviation, maximum value and minimum value of the speed change rate data obtained in step S1 as the first four components of the feature vector; extract the maximum tension and average tension of the tension data obtained in step S2 as the last two components of the feature vector; combine the extracted six components in order into a six-dimensional feature vector;

[0028] Using the pre-trained SVM model, the feature vectors are classified to obtain the danger level. Considering the non-stationary and mutational nature of the safety rope speed change rate and tension data, the RBF kernel function may be difficult to describe the true distribution of the data. This application uses the term kernel + RBF kernel: ,in, and are two eigenvectors, represents the inner product, ||*|| represents the bi-norm (Euclidean distance), d is the order of the polynomial kernel function, and γ is the shape parameter of the RBF kernel function. The composite kernel function combines the nonlinear mapping ability of the polynomial kernel and the local similarity measurement ability of the RBF kernel, which can better characterize the complex distribution of the safety rope data.

[0029] The traditional grid search has a large computational overhead and is prone to redundancy and overfitting. This application uses the Gaussian process Bayesian optimization method to optimize the hyperparameters of SVM (penalty coefficient C and kernel function parameters d, γ): define the Gaussian process prior ,in is the hyperparameter vector, K(*), is the covariance function, commonly used ones include square exponential kernel, Matérn kernel, etc.; based on the initial hyperparameter set , calculate the GP posterior distribution: ;in , ; Define the acquisition function as Expected Improvement: IE ( i ) = E [ max( 0 , f ( i ) − f ( i + )) ] ;in is the current optimal hyperparameter, is the SVM validation set accuracy corresponding to θ; find the next evaluation point : ,Will join in And update the GP posterior; repeat the iteration until the maximum number of evaluations N or Bayesian optimization can quickly approximate the global optimal parameter combination in a limited number of sampling points, avoiding the "dimensionality disaster" of grid search.

[0030] For time series data such as safety ropes, SVM classification faces the problem of sparse features and information loss. Therefore, this application introduces a sequence feature extraction module before SVM to fully explore the time dependency of the data on the speed change rate sequence. and tension sequence Perform sliding window segmentation and feature extraction: Set the window length to L, the sliding step length to s, and extract the statistical features within the kth sliding window , statistical characteristics include mean Standard Deviation Etc., feature vectors extracted from different windows Perform overlapping sampling to obtain the expanded training set {X}.

[0031] Using LSTM (Long Short-Term Memory) and other recurrent neural networks, we can perform end-to-end learning on feature sequences and automatically extract high-order temporal features. We design LSTM layers and fully connected layers, where the LSTM layer weight matrix is , the bias vector is , during forward propagation, the t-th step LSTM unit receives and , update the state h_t and memory unit , extract the last step LSTM output As a time series feature representation, it is sent to the fully connected layer and SVM classifier, where: They are forget gate, input gate and output gate respectively. is a memory unit. Use multi-scale one-dimensional CNN to extract data features of different granularities: design multiple parallel one-dimensional convolution kernels, and the convolution kernel widths are , the convolution step size is 1, the padding mode is "same", the number of convolution kernels is C, and the number of corresponding feature map channels is also C. The feature map output by the i-th convolution kernel in the l-th layer is a i ( l ) = Re LU [ w i ( l ) × a ( l − 1 ) + b i ( l ) ] ,Finally, the multi-scale feature maps are spliced ​​in the channel dimension and sent to the fully connected layer and SVM classifier.,This application introduces a deep learning method combined with SVM, which can overcome the limitations of artificial features, adaptively mine the spatiotemporal features of high-altitude safety rope data, and provide a more generalized and robust feature representation for SVM classification.

[0032] On the other hand, the standard SVM assumes that the misclassification cost of each type of sample is equal, but in the high-altitude operation scenario, the consequence of missing a high-risk state is far more serious than missing a low-risk state. Therefore, the cost-sensitive SVM is introduced to assign different penalty weights to high-risk (+1 class) and non-high-risk (-1 class) samples: Optimization objective: ; Constraints: ;in and are the penalty coefficients for high-risk (+1 category) and non-high-risk (-1 category), , ϕ(·) is the kernel function mapping, is a slack variable. For the test sample x, according to its relationship with the support vector The kernel function value of And the dual coefficient , calculate the classification decision function: , N is the total number of support vectors, for , b is the bias term. If f(x)>0, x is predicted to be high-risk (+1); if f(x)<0, x is predicted to be non-high-risk (-1); if f(x)=0, x is on the decision boundary.

[0033] According to the danger level, a second control signal is generated to control the operation of the electric rope reel. When the danger level is high (+1), an emergency braking command is generated to stop and lock the electric rope reel in the shortest time; when the danger level is not high (-1), a deceleration or speed limit control command is generated according to the specific speed and tension level to adjust the operation state of the electric rope reel until the safety rope returns to a safe range.

[0034] Further, constructing the feature vector includes: Perform feature extraction: Calculate the mean of the speed change rate data ; Calculate the standard deviation of the velocity change rate data ; Calculate the maximum value of the speed change rate data and minimum value ; , , , As the first 4 components of the eigenvector;

[0035] The tension data obtained in step S2 Perform feature extraction: calculate the maximum value of the tension data max (F); calculate the average value of the tension data ; , As the fifth and sixth components of the feature vector; the extracted four velocity change rate features and the extracted two tension features are combined in sequence to obtain a 6-dimensional feature vector .

[0036] Further, a second control signal is generated according to the danger level to control the operation of the electric rope reel, including: when the danger level is high: generating an emergency braking command to reduce the target speed of the electric rope reel to Set to 0; the braking torque Set to the preset maximum braking torque The generated emergency braking command is used as the second control signal to control the emergency braking of the electric rope reel;

[0037] When the danger level is medium, the generation process of the second control signal is similar to step S5, but different input variables and fuzzy control rule base are used; the current actual speed of the electric rope reel is obtained. ;according to and the preset fuzzy control rule base under medium-risk conditions, and use fuzzy reasoning to calculate the target speed and braking torque ;Will and As the second control signal, control the electric rope reel to decelerate and run;

[0038] When the danger level is low, the generation process of the second control signal is similar to step S5, but different input variables and fuzzy control rule base are used; the current actual speed of the electric rope reel is obtained. ;according to and the preset fuzzy control rule base under low-risk conditions, and use fuzzy reasoning to calculate the target speed ;Will As the second control signal, the electric rope winder is controlled to resume normal speed operation.

[0039] Compared with the prior art, the advantages of this application are:

[0040] The traditional DTW algorithm uses the original time series data to directly perform similarity matching, but there are limitations in the analysis of the speed data of the safety rope. It is difficult to accurately characterize the dynamic change characteristics of the speed, and it is easy to ignore the instantaneous characteristics of the speed change, resulting in lag and inaccuracy in the safety state judgment. Therefore, this application calculates the second-order difference of the speed data to obtain the speed change rate data, which reflects the dynamic change characteristics of the speed; the DTW algorithm is used to perform similarity matching on the speed change rate data and the tension data with the preset safety range template, which can effectively judge the similarity of the data sequence, obtain the speed change rate matching degree and the tension matching degree, and then comprehensively evaluate whether the safety rope is in a safe state. This improved method not only takes into account the absolute value of the speed, but also fully explores the dynamic characteristics of the speed change, improves the timeliness and accuracy of the safety state judgment, and makes up for the shortcomings of the traditional DTW algorithm.

[0041] For the safety rope in a safe state, most traditional control strategies use fixed thresholds or single variable feedback control, lacking comprehensive consideration of the movement state of the personnel and the stress state of the safety rope. The control accuracy and adaptability are not high, and it is difficult to cope with complex and changeable working conditions. This application has made improvements from the perspective of comprehensive decision-making. By integrating multi-source information such as speed change rate and tension matching, a multivariable fuzzy controller with the movement state of the personnel and the stress state of the safety rope as input is constructed, which solves the limitation of the traditional control method. Therefore, for the safety rope in a safe state, an adaptive fuzzy control strategy is designed. According to the mean and standard deviation of the speed change rate data, it is judged whether the movement state of the personnel is stable; combined with the tension matching degree, it is judged whether the force of the safety rope is uniform; the movement state of the personnel and the stress state of the safety rope are used as two inputs of the fuzzy controller, and the speed control quantity and torque control quantity are generated through membership calculation, fuzzy reasoning and defuzzification steps, and the output of the electric winder is adjusted to realize the intelligent control of the safety rope extension process and ensure the stability and safety of the movement. This strategy can adaptively adjust the control output according to the actual movement of personnel and the stress state of the safety rope, and has higher control accuracy and real-time response capability.

[0042] According to the danger classification results of the SVM algorithm, a graded control strategy is adopted. When the danger level is high, an emergency braking response is made quickly, the target speed of the electric winder is set to zero, and the maximum braking torque is applied to prevent the deterioration of the dangerous situation in time; when the danger level is medium or low, the corresponding control instructions are generated according to the current speed state using fuzzy control rules to control the electric winder to slow down or restore normal speed operation, avoiding excessive or insufficient control problems caused by deviations in danger level judgment and improving the accuracy and reliability of control. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] The present application will be further described in the form of exemplary embodiments, which will be described in detail by the accompanying drawings. These embodiments are not restrictive, and in these embodiments, the same number represents the same structure, wherein:

[0044] Figure 1 is an exemplary module diagram of a high-altitude operation safety alarm device according to some embodiments of the present application;

[0045] Figure 2 is an exemplary flow chart of a high-altitude work safety alarm method according to some embodiments of the present application;

[0046] Figure 3 is an exemplary flow chart of calculating tension matching degree according to some embodiments of the present application;

[0047] Figure 4is an exemplary flow chart of calculating the speed change rate matching degree according to some embodiments of the present application;

[0048] Figure 5 This is an exemplary flow chart of generating a first control signal according to some embodiments of the present application. DETAILED DESCRIPTION

[0049] The method and system provided in the embodiments of the present application are described in detail below with reference to the accompanying drawings.

[0050] like Figure 1 As shown, a high-altitude work safety alarm device includes: an electric winder, a tension sensor, a speed sensor, a controller and a speed gearbox. The electric winder is used to control the extension and contraction of the safety rope, and its output shaft is connected to the input shaft of the speed gearbox. The tension sensor is arranged on the safety rope and is used to collect the tension data of the safety rope in real time. The sampling frequency is 100Hz, the range is 0-1000N, and the accuracy is ±0.5%FS. The speed sensor is connected to the output shaft of the electric winder and is used to collect the extension or contraction speed of the safety rope in real time. The sampling frequency is 100Hz, the range is ±10m / s, and the accuracy is ±0.1%FS.

[0051] The data collected by the tension sensor and speed sensor are transmitted to the controller in real time via the RS485 bus. The controller uses the STM32H743 microcontroller based on the ARM Cortex-M7 core, with an operating frequency of 480MHz, built-in 1MB Flash and 512KB SRAM, and an external 8GB TF card for data storage. The controller communicates with the speed change gearbox via the CAN bus.

[0052] After receiving the tension data and speed data, the controller extracts the data features, calculates the second-order difference of the speed data, and obtains the speed change rate data. Then, the DTW algorithm is used to perform similarity matching between the speed change rate data and the tension data and the preset safety range template, with 50 data points as a group, and the matching distances in both directions are calculated. If the speed change rate matching distance and the tension matching distance are both within the threshold range, the safety rope is judged to be in a safe state; otherwise, the safety rope is judged to be in a dangerous state.

[0053] When the safety rope is in a dangerous state, the controller uses the SVM algorithm to classify the extracted speed change rate characteristics and tension characteristics to identify the current danger level. If the danger level is low, a deceleration control signal is generated to set the target speed of the electric winder to 80% of the current speed until the speed and tension return to a safe range; if the danger level is medium, a deceleration control signal is generated to set the target speed to 50% of the current speed; if the danger level is high, an emergency braking signal is generated to set the target speed to 0 and apply the maximum braking torque.

[0054] The speed change gearbox uses the PID control algorithm to adjust the output speed and torque of the electric winder according to the control signal received, so as to realize the automatic extension and retraction control of the safety rope. The speed change gearbox adopts planetary gear transmission, with a transmission ratio range of 1:11:5, a rated output speed of 01000rpm, a rated output torque of 20N·m, a fully sealed structure, and a protection level of IP67.

[0055] like Figure 2 As shown, S1, collecting the tension data and speed data of the safety rope; and calculating the second-order difference of the speed data to obtain the speed change rate data; S2, performing similarity matching between the collected tension data and the preset tension safety range template to obtain the tension matching degree; S3, performing similarity matching between the speed change rate data and the preset speed change rate safety range template to obtain the speed change rate matching degree; S4, judging whether the safety rope is in a safe state according to the tension matching degree and the speed change rate matching degree; if the safety rope is in a safe state, proceeding to step S5, if not, proceeding to step S6; S5, when the safety rope is in a safe state, generating a first control signal according to a preset fuzzy control rule; adjusting the output speed and torque of the electric winder according to the first control signal; S6, for the safety rope in a dangerous state, classifying the dangerous state using the support vector machine SVM algorithm to obtain the danger level; generating a second control signal according to the danger level.

[0056] Specifically, S1 collects the tension data and speed data of the safety rope. The tension sensor and speed sensor set on the safety rope collect the tension and speed information of the safety rope in real time. The tension sensor uses a piezoresistive sensor with a range of 0-1000N, an accuracy of ±0.5%FS, and a sampling frequency of 100Hz; the speed sensor uses an incremental encoder with a range of ±10m / s, an accuracy of ±0.1%FS, and a sampling frequency of 100Hz. The sensor transmits the collected data to the controller in real time via the RS485 bus.

[0057] After receiving the speed data, the controller performs a second-order difference operation on it to obtain the speed change rate data. For each time point , get the corresponding speed data Among them, the time point The interval is the sampling period of the speed sensor, that is, Δt=0.01s. Calculate the difference in speed change at adjacent time points to obtain the second-order speed difference value , the calculation formula is: ; For example, at time point The speed data collected at , , , , then the second-order difference of velocity and They are: ; .

[0058] Calculate the second-order difference of velocity Time difference with adjacent time points The ratio of the speed change acceleration is obtained. , the calculation formula is: ; Among them, the time difference between adjacent time points In this solution, the fixed value is 0.01s. Continuing with the above data as an example, the speed change acceleration and They are: ; .

[0059] like Figure 3 As shown, step S2 uses the improved dynamic time warping algorithm DTW to perform similarity matching on the collected tension data and the preset tension safety range template to obtain the tension matching degree, including: constructing a distance matrix between the tension data sequence and the tension safety range template sequence. and preset tensile safety range template sequence They are used as the input sequences of the improved DTW algorithm. Among them, the length of the tension data sequence F is n=1000, and the sampling frequency is 100Hz, that is, a tension data is collected every 0.01s; the length of the tension safety range template sequence G is m=100, which is set by experts based on experience and represents the standard range of the safety rope force. The first-order difference of the tension data sequence F and the tension safety range template sequence G is calculated respectively to obtain the corresponding tension change rate sequence and ,in , .

[0060] Calculate the distance matrix Elements When considering the tension value at the current position and , and also incorporates the context information of the k=5 points before and after it. Specifically, calculate the mean tension of the i-th position and the 5 points before and after it (a total of 11 points) in the tension sequence F and the mean rate of change of tension ; Similarly, calculate the average tension of the jth position and the five points before and after it in the tension safety range template sequence G and the mean rate of change of tension . Calculate the corrected distance based on the mean , where the weight of the mean difference of tension is , the weight of the mean difference of tension change rate .

[0061] Use a dynamic programming algorithm with slope and step size constraints to search for the optimal matching path in the distance matrix D'. Specifically, initialize the cumulative distance matrix ,Will Set to ,Right now .from Start by recursively calculating the cumulative distance matrix Other elements until The recursive formula is: ; Constraints: , , ;in, It represents the minimum cumulative distance between the first i elements of sequence F and the first j elements of sequence G under local constraints; is the coordinate of the matching point in the previous step. Specifically, for the current position (i, j), first calculate its coordinate with the matching point in the previous step The coordinate difference is used to determine whether the constraint condition is met: |ij|≤5 limits the slope of the matching path to approximately 45°; and The maximum step lengths in the vertical and horizontal directions are limited respectively to make the matching path smoother. If the constraints are met, the recursive formula is used to calculate . Indicates that there are three possible paths from the previous state to the current state (i, j): diagonal (i-1, j-1), vertical (i-1, j) and horizontal (i, j-1). The path with the smallest cumulative distance is selected as the optimal subpath, plus the distance of the current position. ,get If the constraints are not met, Set to a large value (such as infinity) to indicate that the current location cannot be part of the optimal matching path. Repeat until the cumulative distance matrix is ​​calculated. All elements.

[0062] For example, the distance matrix Some of the elements are as follows: D ' = [ 0 . 2 0 . 5 0 . 8 0 . 3 0 . 1 0 . 6 0 . 7 0 . 4 0 . 2 ] , according to the recursive formula, calculate the cumulative distance matrix The corresponding element: S ' = [ 0 . 2 0 . 7 1 . 5 0 . 5 0 . 3 1 . 1 1 . 2 0 . 8 0 . 5 ] .in, ; ; ; .

[0063] exist Get the end point of the best matching path, and then start from the end point and search backward to get the entire best matching path Specifically, from First, compare its three adjacent elements , and , select the element with the smallest value as the previous matching point, and add its coordinates to the optimal matching path Repeat this process until you reach , get the complete optimal matching path Taking the above matrix as an example, the optimal matching path is: , the corresponding matching point is .

[0064] Calculate the tension matching degree. For the optimal matching path Sum the distance values ​​on the path to get the total path distance . The total distance Divide by the path length length(path') = 723 to get the average distance , as the difference between the optimized tension data sequence and the safety range template sequence. Finally, the tension matching degree The closer the value is to 1, the more the collected tension data matches the safety range template, and the safer the state of the safety rope is; otherwise, it means that the current tension state may have safety hazards and further analysis is needed.

[0065] like Figure 4 As shown, S3, the speed change rate data is matched with the preset speed change rate safety range template for similarity to obtain the speed change rate matching degree, including: constructing a distance matrix between the speed change rate data sequence and the speed change rate safety range template sequence. and preset speed change rate safety range template sequence They are used as the input sequences of the DTW algorithm respectively. Among them, the length of the speed change rate data sequence A is n-2=998, and the length of the speed change rate safety range template sequence B is m=100, which is set by experts based on experience and represents the standard range of safety rope speed change.

[0066] Calculates the distance between each pair of elements in two sequences , construct the distance matrix .in, Indicates the rate of change of speed Corresponding change rate in the safety range template The distance is calculated as: , for example, if , ,but .

[0067] Search for the best matching path in the distance matrix D. Use the dynamic programming algorithm, starting from D(1, 1), and recursively calculate the cumulative distance matrix , until The recursive formula is: ,in, Represents the minimum cumulative distance between the first i elements of sequence A and the first j elements of sequence B.

[0068] exist The end point of the optimal matching path is obtained, and then starting from the end point, the entire optimal matching path is searched in reverse. First, compare its three adjacent elements , and , select the element with the smallest value as the previous matching point, and add its coordinates to the optimal matching path. Repeat this process until you backtrack to , get the complete optimal matching path path.

[0069] Calculate the speed change rate matching degree. Sum the distance values ​​on the optimal matching path to get the total path distance . Then the total path distance Divide by the path length length(path) to get the average distance , as the difference between the speed change rate data sequence and the safety range template sequence. Finally, use the formula Calculate the speed change rate matching degree. After calculation, , , length(path)=998, then: , . Speed ​​change rate matching The closer it is to 1, the more the speed change rate data sequence matches the safety range template sequence, and the safer the speed change state of the safety rope is; conversely, it means that the current speed change state may have safety hazards and requires timely warning. Compared with directly comparing speed values, by calculating the second-order difference of the speed to obtain the speed change acceleration, and then performing DTW matching, not only the similarity of speed values ​​at different time points is compared, but also the similarity of speed change trends at adjacent time points is compared, making the similarity matching more comprehensive, thereby improving the accuracy of safety rope status judgment.

[0070] like Figure 5As shown, S4, according to the tension matching degree and the speed change rate matching degree, determine whether the safety rope is in a safe state; if the safety rope is in a safe state, enter step S5, if not, enter step S6; in step S5, when the safety rope is in a safe state, a first control signal is generated according to a preset fuzzy control rule; the first control signal is used to adjust the output speed and torque of the electric rope reel, including: calculating the mean of the speed change rate data obtained in step S1 and standard deviation Mean Reflects the average level and standard deviation of the speed change rate data Reflects the fluctuation degree of speed change rate data. Assume that the speed change rate data sequence is ,but: ; ; For example, when n=1000, if , ,but: ; .

[0071] According to the mean and standard deviation Determine the movement status of the person. E ( v ') ∈ [ 1600 , 2000 ] and s ( v ') ∈ [ 400 , 600 ] , then the personnel motion state Mark as stable; otherwise, Marked as unstable. The threshold here is [ E minutes , E max ] and [ s minutes , s max ] It can be set according to actual situation.

[0072] Determine whether the tension matching degree M obtained in step S2 is greater than a preset threshold If M>0.85, the safety rope stress state Mark as uniform; otherwise, Marked as non-uniform.

[0073] The movement status of personnel And the stress state of the safety rope As two input variables of the fuzzy controller respectively. According to the preset membership function of the personnel motion state ,calculate The corresponding membership ; According to the preset safety rope force state membership function ,calculate The corresponding membership The membership function is as follows: If So ;if So .if So ;if So .

[0074] According to the preset fuzzy control rule library , using Mamdani inference method to calculate the membership and Perform fuzzy reasoning. Known personnel movement status And the stress state of the safety rope The membership degrees are: ; .

[0075] Then the membership degrees of the premise parts of the four fuzzy control rules are: : ; : ; : ; : . Among them, ∧ represents the smaller operation. The rule The trigger strength .

[0076] According to how Mamdani reasoning works, using trigger strength Rules The conclusion part is intercepted to obtain the membership function of the output variable under this rule: in conclusion: m v 1 ( v ) = m VS , 3 ( v ) ∧ w 1 = trimf ( v , [ 400 , 600 , 800 ] ) ∧ 0 . 4 , m T 1 ( T ) = m VT , 3 ( T ) ∧ w 1 = trimf ( T , [ 40 , 60 , 80 ] ) ∧ 0 . 4 ; in conclusion: m v 2 ( v ) = m VS , 2 ( v ) ∧ w 2 = trimf ( v , [ 200 , 400 , 600 ] ) ∧ 0 . 6 , m T 2 ( T ) = m VT , 2 ( T ) ∧ w 2 = trimf ( T , [ 20 , 40 , 60 ] ) ∧ 0 . 6 ; in conclusion: m v 3 ( v ) = m VS , 1 ( v ) ∧ w 3 = trimf ( v , [ 0 , 200 , 400 ] ) ∧ 0 . 2 , m T 3 ( T ) = m VT , 1 ( T ) ∧ w 3 = trimf ( T , [ 0 , 20 , 40 ] ) ∧ 0 . 2 ; in conclusion: m v 4 ( v ) = m VS , 0 ( v ) ∧ w 4 = trimf ( v , [ − 200 , 0 , 200 ] ) ∧ 0 . 2 , m T 4 ( T ) = m VT , 0 ( T ) ∧ w 4 = trimf ( T , [ − 20 , 0 , 20 ] ) ∧ 0 . 2 .

[0077] Aggregate the output membership functions in the four rule conclusions to obtain the fuzzy sets of output speed v and torque T and : ; . Among them, ∨ represents the maximum operation. Substituting the membership function of each rule conclusion into: m v ( v ) = [ trimf ( v , [ 400 , 600 , 800 ] ) ∧ 0 . 4 ] ∨ [ trimf ( v , [ 200 , 400 , 600 ] ) ∧ 0 . 6 ] ∨ [ trimf ( v , [ 0 , 200 , 400 ] ) ∧ 0 . 2 ] ∨ [ trimf ( v , [ − 200 , 0 , 200 ] ) ∧ 0 . 2 ] ; For the above and Discretize points in the definition domain, use the maximum operation to aggregate, and finally get the output speed fuzzy set and output torque fuzzy set , which integrate the reasoning results of four fuzzy control rules.

[0078] current , , , , the triggered rule is and , the corresponding trigger strength is , , from which we get: , .

[0079] S5.6 uses the centroid defuzzification algorithm to convert the speed fuzzy set and torque fuzzy set Converted into precise speed control and torque control , generate the first control signal, and the speed fuzzy set obtained by fuzzy reasoning and torque fuzzy set Several fuzzy subsets and (i=0, 1, 2, 3), each subset corresponds to a membership function. and The membership functions are trigonometric functions and , whose analytical formula is: m VS 0 ( v ) = trimf ( v , [ − 200 , 0 , 200 ] ) = { 0 , v ≤− 200 , or , v ≥ 200 v + 200 200 , − 200 ≤ v < 0 200 − v 200 , 0 ≤ v < 200 ; m VS 1 ( v ) = trimf ( v , [ 0 , 200 , 400 ] ) = { 0 , v ≤ 0 , or , v ≥ 400 v 200 , 0 ≤ v < 200 400 − v 200 , 200 ≤ v < 400 ; m VS 2 ( v ) = trimf ( v , [ 200 , 400 , 600 ] ) = { 0 , v ≤ 200 , or , v ≥ 600 v − 200 200 , 200 ≤ v < 400 600 − v 200 , 400 ≤ v < 600 ; m VS 3 ( v ) = trimf ( v , [ 400 , 600 , 800 ] ) = { 0 , v ≤ 400 , or , v ≥ 800 v − 400 200 , 400 ≤ v < 600 800 − v 200 , 600 ≤ v < 800 ; m VT 0 ( T ) = trimf ( T , [ − 20 , 0 , 20 ] ) = { 0 , T ≤− 20 , or , T ≥ 20 T + 20 20 , − 20 ≤ T < 0 20 − T 20 , 0 ≤ T < 20 ; m VT 1 ( T ) = trimf ( T , [ 0 , 20 , 40 ] ) = { 0 , T ≤ 0 , or , T ≥ 40 T 20 , 0 ≤ T < 20 40 − T 20 , 20 ≤ T < 40 ; m VT 2 ( T ) = trimf ( T , [ 20 , 40 , 60 ] ) = { 0 , T ≤ 20 , or , T ≥ 60 T − 20 20 , 20 ≤ T < 40 60 − T 20 , 40 ≤ T < 60 ; m VT 3 ( T ) = trimf ( T , [ 40 , 60 , 80 ] ) = { 0 , T ≤ 40 , or , T ≥ 80 T − 40 20 , 40 ≤ T < 60 80 − T 20 , 60 ≤ T < 80 Among them, trimf(x, [abc]) represents the triangular membership function with a, b, and c as parameters, the domain of v is [-200, 800], and the domain of T is [-20, 80].

[0080] According to the center of gravity defuzzification method, the speed control quantity and torque control The calculation formula is: , ;in, and They are the composite membership functions of the speed fuzzy set and the torque fuzzy set respectively: ; ;Will , Substitution and The calculation formula of can be obtained by numerical integration: ; ; In the formula, and are the discrete values ​​of the speed and torque domains, and is the discrete step length, n and m are the discrete points. , , n=1000, m=1000, substitute into the above formula to get the precise speed control value and torque control .

[0081] Obtained through fuzzy reasoning m v = 0 . 4 × VS 3 + 0 . 6 × VS 2 = { 0 . 4 / [ 400 , 600 , 800 ] , 0 . 6 / [ 200 , 400 , 600 ] } , m T = 0 . 4 × VT 3 + 0 . 6 × VT 2 = { 0 . 4 / [ 40 , 60 , 80 ] , 0 . 6 / [ 20 , 40 , 60 ] } ,but: m v ( v ) = max { 0 . 4 × trimf ( v , [ 400 , 600 , 800 ] ) , 0 . 6 × trimf ( v , [ 200 , 400 , 600 ] ) } ; m T ( T ) = max { 0 . 4 × trimf ( T , [ 40 , 60 , 80 ] ) , 0 . 6 × trimf ( T , [ 20 , 40 , 60 ] ) } ; Substituting into the discrete integral formula, we get , The present application judges the motion state of personnel and the stress state of safety rope by introducing the mean and standard deviation of speed change rate, and then uses them as the input of fuzzy controller after fuzzy processing, uses fuzzy rule base for reasoning, and uses the center of gravity defuzzification algorithm to convert the reasoning result into speed control quantity. and torque control Generate a first control signal, and finally control the electric rope reel output to track the control signal. The entire control strategy can dynamically adapt to various working conditions such as smooth / non-smooth movement and uniform / non-uniform force during aerial operations. The safety rope is always in a safe state through automatic adjustment of speed and torque, effectively improving operation safety.

[0082] According to the speed deviation and torque deviation , using PID control algorithm to calculate the speed control signal of the electric rope reel and torque control signal The PID control algorithm is as follows: u v ( k ) = K p 1 × D v ( k ) + K i 1 × ∑ j = 0 k D v ( j ) + K d 1 × [ D v ( k ) − D v ( k − 1 ) ] ; u T ( k ) = K p 2 × D T ( k ) + K i 2 × ∑ j = 0 k D T ( j ) + K d 2 × [ D T ( k ) − D T ( k − 1 ) ] ;in, and are the speed control signal and torque control signal at the kth sampling moment, respectively. , , , , , , the sampling period is 10ms.

[0083] The speed control signal and torque control signal The speed controller uses SVPWM control algorithm to drive the motor to adjust the output speed. track ; The torque controller uses a constant torque control algorithm to adjust the motor excitation current. track The rated power of the motor is 3kW, the rated speed is 1500r / min, and the rated torque is 20N·m. Repeat until the speed deviation and torque deviation All converge to the allowable range. Set the speed error allowable threshold , torque error allowable threshold ,when and When 20 sampling cycles (i.e. 0.2s) are maintained continuously, it is considered that the electric rope reel has reached the target speed and target torque, and the adjustment process is ended. This application uses the PID control algorithm to achieve precise tracking and control of the speed and torque of the electric rope reel respectively. The speed loop and the torque loop are adjusted separately, and good dynamic and steady-state characteristics can be obtained through parameter setting. At the same time, by setting the speed error and torque error allowable thresholds and detecting their continuous holding time, it is ensured that the electric rope reel can reliably reach the target value. The entire control process has a high degree of automation and fast dynamic response, which is conducive to improving the real-time adjustment capability of the intelligent safety rope.

[0084] In step S6, for the safety rope in a dangerous state, the support vector machine (SVM) algorithm is used to classify the dangerous state to obtain the danger level; and a second control signal is generated according to the danger level, including: judging whether the safety rope is in a dangerous state. The mean value of the speed change rate data calculated in step S1 and standard deviation ,judge Whether it exceeds the preset threshold range [1000, 2500], Whether it exceeds the preset threshold range [400, 800]; According to the tension matching degree M calculated in step S2, determine whether M is less than the preset threshold M_0=0.6. or or or , and M<0.6, the safety rope is considered to be in a dangerous state; otherwise, the safety rope is considered to be in a safe state.

[0085] For the safety rope in a dangerous state, construct a feature vector. For the speed change rate data obtained in step S1 , n=1000, feature extraction: calculate the mean value of speed change rate ; Calculate the standard deviation of the velocity change rate ; Calculate the maximum speed change rate , minimum ; , , , as the first 4 components of the eigenvector.

[0086] The tension data obtained in step S2 , m=1000 for feature extraction: Calculate the maximum tension ; Calculate the average pulling force ; , As the 5th and 6th components of the eigenvector, we finally get a 6-dimensional eigenvector (1423.5, 562.8, 2256.2, 463.1, 1538.6, 962.7).

[0087] Using the pre-trained SVM model, the feature vector is classified to obtain the danger level. Construct SVM, where γ is the kernel function parameter, Represents the bi-norm. Through grid search and 5-fold cross validation, the optimal parameters C=100, γ=0.1 are obtained. At this time, the classification accuracy of the SVM model reaches more than 95%. Using 500 sets of feature vector samples of historical dangerous working conditions, 250 of which are high-risk samples (label y=-1), 150 are medium-risk samples (label y=0), and 100 are low-risk samples (label y=1), the optimized SVM model is trained to obtain the optimal decision hyperplane and 361 support vectors.

[0088] The constructed feature vector (1423.5, 562.8, 2256.2, 463.1, 1538.6, 962.7) is input into the trained SVM model. After calculation, the feature vector falls on the decision hyperplane. If the negative half space is too large, it is classified as high risk.

[0089] Generate a second control signal based on the danger level to control the operation of the electric rope reel. When the danger level is high: Generate an emergency braking command to reduce the target speed Set to 0, braking torque Set to maximum braking torque , the emergency braking command is output as the second control signal. When the danger level is medium: Get the current actual speed ,according to and the preset medium-risk rule base {if So , }, calculate the target speed , braking torque , as the second control signal output. When the danger level is low: Get the current actual speed ,according to and the preset low-risk rule base {if So }, calculate the target speed , as the second control signal output. This application introduces the SVM classification algorithm to effectively identify and classify the dangerous state of the safety rope according to the speed change rate and the tension feature vector, and then generates corresponding speed or torque control instructions for different danger levels to form a complete second control signal, which can quickly and reliably control the emergency braking or deceleration operation of the electric rope reel under dangerous conditions, further improving the safety protection effect.

Claims

1. A high-altitude operation safety alarm method, characterized in that: include: S1, collects the tension data and speed data of the safety rope; And calculate the second-order difference of the speed data to obtain the speed change rate data; S2, performing similarity matching between the collected tension data and the preset tension safety range template to obtain a tension matching degree; S3, performing similarity matching between the speed change rate data and a preset speed change rate safety range template to obtain a speed change rate matching degree; S4, judging whether the safety rope is in a safe state according to the tension matching degree and the speed change rate matching degree; If the safety rope is in a safe state, proceed to step S5, if not, proceed to step S6; S5, when the safety rope is in a safe state, generating a first control signal according to a preset fuzzy control rule; adjusting the output speed and torque of the electric winder according to the first control signal; S6, for the safety rope in a dangerous state, the support vector machine (SVM) algorithm is used to classify the dangerous state to obtain the danger level; A second control signal is generated according to the risk level.

2. The high-altitude operation safety alarm method according to claim 1 is characterized in that: S1, calculate the second-order difference of the speed data to obtain the speed change rate data, including: For each time point , get the corresponding speed data ; Calculate the difference in velocity changes at adjacent time points to obtain the second-order velocity difference value : ； Calculate the second-order difference of velocity Time difference with adjacent time points The ratio of the speed change acceleration is obtained. : 。 3. The high-altitude operation safety alarm method according to claim 2 is characterized in that: S3, performing similarity matching between the speed change rate data and a preset speed change rate safety range template to obtain a speed change rate matching degree, including: Calculate the distance between the speed change rate data and each pair of data in the preset speed change rate safety range template to generate a distance matrix; The dynamic programming algorithm DTW is used to search for the matching path with the smallest cumulative distance in the distance matrix as the optimal matching path; Calculate the average distance between the speed change rate data and the preset speed change rate safety range template according to the optimal matching path as the difference; The speed change rate matching degree is calculated based on the dissimilarity.

4. The high-altitude operation safety alarm method according to claim 1 is characterized in that: S2, performing similarity matching between the collected tension data and the preset tension safety range template to obtain the tension matching degree, including: Calculate the first-order difference of the collected tension data sequence {F} and the preset tension safety range template sequence {G} respectively, and get the corresponding and ; According to the tension data sequence {F} and the preset tension safety range template sequence {G}, and and , construct the distance matrix ; Among them, the distance matrix The element d(i, j) of represents the corrected distance between the i-th element in {F} and the j-th element in {G}; Using dynamic programming algorithm DTW to calculate the distance matrix Perform recursive calculation to obtain the cumulative distance matrix and the best matching path ; Among them, the cumulative distance matrix Elements It represents the minimum cumulative distance between the first i elements of {F} and the first j elements of {G} under the local constraints; the optimal matching path Indicated in The matching end point is obtained at , and according to the recursive direction of the DTW algorithm, the path obtained by reverse search from the matching end point; By calculating the optimal matching path The average distance on , as the dissimilarity; According to the dissimilarity, the matching degree between the tension data sequence {F} and the preset tension safety range template sequence {G} is calculated through an inverse proportional function relationship as the tension matching degree.

5. The high-altitude operation safety alarm method according to any one of claims 1 to 4, characterized in that: S5, when the safety rope is in a safe state, a first control signal is generated according to a preset fuzzy control rule; the first control signal is used to adjust the output speed and torque of the electric winder, including: Calculate the mean and standard deviation of the speed change rate data obtained in step S1; If the mean and standard deviation are within the preset threshold range, the person's motion state is marked as stable, otherwise it is marked as unstable; Determine whether the tension matching degree obtained in step S2 is greater than a preset threshold value. If so, mark the stress state of the safety rope as uniform, otherwise mark it as uneven; The motion state of the personnel and the stress state of the safety rope are used as two input variables of fuzzy control respectively; Calculate the personnel motion state membership according to a preset personnel motion state membership function; calculate the safety rope force state membership according to a preset safety rope force state membership function; According to the preset fuzzy control rule base, the Mamdani reasoning method is used to perform fuzzy reasoning on the membership of the personnel's motion state and the safety rope's force state, and the fuzzy sets of speed and torque are obtained. The center of gravity defuzzification algorithm is used to convert the fuzzy sets of speed and torque into speed control quantities. and torque control , as the first control signal.

6. The high-altitude operation safety alarm method according to claim 5 is characterized in that: According to the first control signal, the output speed and torque of the electric winder are adjusted, including: Extract the speed control amount from the first control signal and torque control ; Get the current speed of the electric winder and current torque ; According to the speed control and torque control , and the current speed and current torque , using PID control algorithm to calculate the speed control signal of the electric winder and torque control signal .

7. The high-altitude operation safety alarm method according to claim 5 is characterized in that: S6, generating a second control signal according to the danger level, including: When the mean or standard deviation of the speed change rate data exceeds the preset threshold range, and the tension matching degree is lower than the preset threshold, the safety rope is considered to be in a dangerous state; For the safety rope in dangerous state, construct the feature vector; Using the pre-trained SVM model, the feature vectors are classified to obtain the risk level, where the risk level includes low risk, medium risk and high risk; According to the danger level, a second control signal is generated to control the operation of the electric winder.

8. The high-altitude operation safety alarm method according to claim 7 is characterized in that: Construct a feature vector, including: Extract the mean, standard deviation, maximum and minimum values ​​of the speed change rate data; Extract the maximum tension and average tension of the tension data; Combine the extracted data into a feature vector.

9. The high-altitude operation safety alarm method according to claim 7, characterized in that: According to the danger level, a second control signal is generated to control the operation of the electric winder, including: When the danger level is high, the target speed of the electric winder is set to zero, and the braking torque is set to a preset maximum value; When the danger level is medium or low, a second control signal is generated according to the current speed data using a preset fuzzy control rule.

10. A safety alarm device for high-altitude operations, characterized in that: include: Electric winder, tension sensor, controller and speed change gearbox; Electric winder, used to control the extension and retraction of the safety rope; A tension sensor is used to collect the tension data of the safety rope and transmit the collected tension data to the controller; A speed sensor is used to collect speed data of the safety rope and transmit the collected speed data to the controller; wherein the speed data includes an extension or contraction speed; A controller generates a control signal according to the collected tension data and speed data; The speed change gearbox adjusts the output speed and torque of the electric winder according to the control signal.

Images