A method and device for identifying a pull cord risk by fusing multi-dimensional data

Abstract

The application discloses a kind of fusion multi-dimensional data's pull rope risk identification method and device, method includes: the multi-dimensional operation data of pull rope is collected, wherein, multi-dimensional operation data includes: visual data, tension data, displacement data and vibration data;Based on the multi-dimensional operation data of collection, and preliminary abnormal analysis processing is carried out to each dimensional operation data corresponding abnormal condition, and the preliminary abnormal result of each dimensional operation data is given;Based on the preliminary abnormal result of each dimensional operation data, secondary abnormal analysis processing is carried out to fusion multi-dimensional operation data, and the pull rope abnormal result is given;Based on pull rope abnormal result and the preliminary abnormal result of each dimensional operation data, the risk level of pull rope is given to carry out pull rope risk identification.Can accurately detect the operation state of pull rope under different working conditions, realize the risk identification of pull rope.

Description

Technical Field

[0001] This invention belongs to the field of power construction technology, specifically relating to a method and device for identifying rope risks by integrating multi-dimensional data. Background Technology

[0002] In the construction and maintenance of power systems, rope pulling is one of the key steps in the erection and maintenance of transmission lines. Ropes are used to pull transmission cables or construction equipment, ensuring the tension and stability of the line, which plays a crucial role in ensuring the safety and continuity of power supply. However, due to complex terrain and harsh environmental conditions, rope pulling operations face many challenges.

[0003] Controlling the tension of guy ropes is a technical challenge. Maintaining appropriate tension under varying environmental and working conditions to prevent transmission line faults or rope damage due to overload requires precise monitoring and control. Secondly, rope deviation is a common problem in guy rope operations. In complex environments, guy ropes may deviate due to terrain undulations or external forces, affecting not only construction efficiency but also threatening the safety of the transmission line or the rope itself. Furthermore, adverse weather conditions such as strong winds, heavy rain, and snow can damage guy ropes, and operational errors during construction can also cause damage. If these damages are not detected and repaired promptly, they may lead to more serious transmission line faults or construction accidents.

[0004] Currently, rope monitoring technology mainly relies on manual experience and lacks precise real-time monitoring methods, making it difficult to detect potential risks in a timely manner. To improve the safety and efficiency of rope operations, existing methods mainly use a single sensor for single detection. For example, patent CN112881947A discloses an overhead ground wire detection device based on an eddy current sensor, including an eddy current probe, an excitation signal generation module, a logarithmic detection module, a signal conditioning module, a differential amplification module, and an imaging module. The excitation signal generation module generates excitation signals of different frequencies and amplitudes. The logarithmic detection module is mainly implemented through a logarithmic amplifier and an LPF. The signal conditioning module is implemented through an AD7670 digital-to-analog converter chip and an STM chip. The differential amplification module is implemented through a differential bridge circuit. It employs eddy current testing devices and methods, solving the problems of high cost, bulky equipment, and cumbersome and time-consuming processes in current overhead ground wire testing. It boasts advantages such as not damaging the performance of the overhead ground wire, fast testing speed, and low instrument cost. It can be used to test ground wires and steel wire ropes in high-voltage transmission lines, and can also be extended to test for damage to steel wire ropes, guy wires, and OPGW stranded layers. However, it only performs single-mode testing using eddy current testing devices and does not consider abnormal parameters that may cause damage to guy wires.

[0005] Based on the above problems, how to provide a risk identification method that takes into account multiple abnormal data of the rope is a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0006] To address the shortcomings of the existing technology, this invention provides a method and apparatus for identifying risks associated with a pull rope by integrating multi-dimensional data. This method can accurately detect the working status of the pull rope under different working conditions and achieve risk identification of the pull rope.

[0007] In a first aspect, the present invention provides a method for identifying rope risks by integrating multi-dimensional data, comprising:

[0008] Collect multi-dimensional operational data of the rope pulling operation, including visual data, tension data, displacement data, and vibration data.

[0009] Based on the collected multidimensional operation data, and combined with the abnormal conditions corresponding to each dimension of operation data, preliminary anomaly analysis and processing are performed, and preliminary anomaly results for each dimension of operation data are given.

[0010] Based on the preliminary anomaly results of each dimension of operation data, the multi-dimensional operation data is integrated for secondary anomaly analysis and processing to give the rope anomaly results.

[0011] Based on the abnormal results of the rope pulling and the preliminary abnormal results of each dimension of the operation data, the risk level of the rope pulling is given to identify the rope pulling risk.

[0012] Furthermore, based on the collected multidimensional operation data and combined with the abnormal conditions corresponding to each dimension of operation data, preliminary anomaly analysis and processing are performed, and preliminary anomaly results for each dimension of operation data are given, including:

[0013] Feature extraction is performed on visual data, or visual data is compared with rope image template data to provide preliminary visual anomalies of the rope; tension data is compared with the upper and lower limits of preset tension thresholds to provide preliminary tension anomalies of the rope; the current displacement-time curve is determined based on displacement data and compared with the standard displacement-time curve of the rope under normal working conditions to provide preliminary displacement anomalies of the rope; Fourier transform is performed on vibration data to obtain frequency components and vibration amplitude, which are compared with frequency range thresholds and vibration range thresholds respectively to provide preliminary vibration anomalies of the rope.

[0014] Furthermore, a comparison is made with the standard displacement-time curve of the guy rope under normal working conditions to provide preliminary results of the guy rope displacement anomaly, including:

[0015] The similarity between the current displacement-time curve and the standard displacement-time curve is calculated, and the similarity result is compared with the preset similarity threshold. If the similarity result exceeds the preset similarity threshold, it is judged as abnormal rope displacement; otherwise, it is judged as normal rope displacement.

[0016] Fourier transform of the vibration data yields frequency components and vibration amplitude, which are then compared with frequency range thresholds and vibration range thresholds to provide preliminary results of the rope vibration anomaly, including:

[0017] The time vibration signal is determined based on the vibration data, and a Fourier transform is performed on the time vibration signal to generate a spectrum of the rope vibration. Frequency components and vibration values ​​are extracted from the generated spectrum. If the frequency components exceed the frequency range threshold and / or the vibration value exceeds the vibration range threshold, the rope vibration is judged to be abnormal; otherwise, the rope vibration is judged to be normal.

[0018] Furthermore, based on the preliminary anomaly results of each dimension of the operation data, a secondary anomaly analysis is performed by integrating multi-dimensional operation data to provide the rope anomaly results, including:

[0019] When the initial abnormal result of at least one-dimensional operation data is determined to be abnormal, a probability analysis is performed on the visual data, tension data, displacement data, and vibration data, and the abnormal result of the rope under the current multi-dimensional operation data is given based on the probability analysis results.

[0020] Furthermore, probabilistic analysis is performed on visual data, tension data, displacement data, and vibration data, and based on the results of the probabilistic analysis, abnormal results of the rope under the current multidimensional operational data are given, including:

[0021] Based on visual data, tension data, displacement data, and vibration data, the probability of the pull rope being in different abnormal states under the current multidimensional operation data is determined.

[0022] Based on the historical abnormal states of the pull rope, determine the initial probability of the pull rope being in different abnormal states;

[0023] Combining the probability of the pull rope being in different abnormal states under the current multidimensional operation data and the initial probability of the pull rope being in different abnormal states, the total probability of observing multidimensional operation data under all abnormal states of the pull rope is given.

[0024] The probability of the pull rope being in different abnormal states under the current multidimensional operation data and the initial probability of the pull rope being in different abnormal states are jointly processed, and the ratio of the joint processing result to the total probability of the pull rope being observed in multidimensional operation data under all abnormal states is calculated to give the probability of the pull rope being in different abnormal states.

[0025] Based on the probability of the rope being in different abnormal states and the probability thresholds corresponding to different abnormal states, the abnormal states that the rope exists in and the abnormal states with the highest probability are determined.

[0026] Furthermore, based on visual data, tension data, displacement data, and vibration data, the probability of the pull-up rope being in different abnormal states under the current multidimensional operational data is determined, including:

[0027] Based on a neural network classification model pre-built using historical visual data and corresponding abnormal states of the rope, and combined with current visual data, the probability of observing current visual data under different abnormal states is determined; based on a normal distribution model pre-built using historical tension data and corresponding abnormal states of the rope, and combined with current tension data, the probability of observing current tension data under different abnormal states is determined; based on the reference displacement-time curves of historical displacement data corresponding to different abnormal states of the rope, similarity calculations are performed between these curves and the current displacement-time curves corresponding to the current displacement data to determine the probability of observing current displacement data under different abnormal states; vibration data is converted into frequency domain signals using fast Fourier transform to determine the spectrum, and the distribution characteristics of the spectrum are used to determine the probability of observing current vibration data under different abnormal states;

[0028] By jointly processing the probabilities of observing current visual data, tension data, displacement data, and vibration data under the same abnormal state, the probability of the pull rope being in different abnormal states under the current multidimensional operation data is obtained.

[0029] Furthermore, based on the abnormal states observed during the historical operation of the rope, the initial probability of the rope being in different abnormal states is determined, including:

[0030] Based on the abnormal states during the historical operation of the rope, determine the number of times the rope was in each abnormal state and the total number of times all abnormal states occurred;

[0031] Based on the number of times the rope is in each abnormal state and the total number of times all abnormal states are considered, the initial probability of the rope being in different abnormal states is given.

[0032] Furthermore, combining the probability of the pull rope being in different abnormal states under the current multidimensional operation data and the initial probability of the pull rope being in different abnormal states, the total probability of observing multidimensional operation data under all abnormal states of the pull rope is given, including:

[0033] The probabilities of the pull rope being in different abnormal states under the current multidimensional operation data and the initial probabilities of the pull rope being in different abnormal states are jointly processed according to the category of abnormal state, and the results of the joint processing are superimposed to obtain the total probability of the pull rope observing multidimensional operation data under all abnormal states.

[0034] Furthermore, the probabilities of the pull rope being in different abnormal states satisfy the following relationship:

[0035]

[0036] P(Sensor data|γ i ) = P(X|γ i ) * P(Y|γ i ) * P(W|γ i ) * P(Z|γ i )P(Sensor data)

[0037] = P(Sensor data|γ1)·P(γ1) + P(Sensor data|γ2)·P(γ2)

[0038] +... + P(Sensor data|γ n )·P(γ n )

[0039] where, P(γ i |Sensor data) is the probability that the cable is in the i-th abnormal state, n is the number of abnormal states, i = 1, 2,..., n, P(γ i ) is the initial probability that the cable is in the i-th abnormal state, X, Y, W, Z are visual data, tension data, displacement data, and vibration data respectively, P(Sensor data|γ i ) is the probability that the cable is in the i-th abnormal state under the current multi-dimensional operation data, P(Sensor data) is the total probability of observing multi-dimensional operation data for the cable in all abnormal states, P(X|γ i ), P(Y|γ i ), P(W|γ i ), P(Z|γ i ) are the probabilities of observing the current visual data, tension data, displacement data, and vibration data in the i-th abnormal state respectively.

[0040] Furthermore, based on the cable abnormality result and the preliminary abnormality results of each dimension of operation data, a risk level of the cable is given for cable risk identification, including:

[0041] Based on the cable abnormality result and the preliminary abnormality results of each dimension of operation data, determine the number of cables that are abnormal;

[0042] Based on the number of cable abnormalities and the threshold range of the number of abnormalities corresponding to different risk levels, determine the risk level of the cable;

[0043] Based on the risk level of the cable and the abnormal state with the highest probability, determine the cable risk identification result.

[0044] In a second aspect, the present invention also provides a cable risk identification device that fuses multi-dimensional data, adopting the above cable risk identification method. The device includes:

[0045] The data acquisition module is used to collect multi-dimensional operational data of the rope pulling operation, including visual data, tension data, displacement data, and vibration data.

[0046] The anomaly detection module is used to perform preliminary anomaly analysis based on the collected multidimensional operation data and the anomaly conditions corresponding to each dimension of operation data, and to give the preliminary anomaly results for each dimension of operation data.

[0047] The data processing and fusion module is used to perform secondary anomaly analysis and processing by fusing multi-dimensional operation data based on the preliminary anomaly results of each dimension of operation data, and to give the rope anomaly results.

[0048] The risk warning module is used to identify rope risks by giving the risk level of the rope based on the abnormal results of the rope pulling and the preliminary abnormal results of the operation data of each dimension.

[0049] The present invention provides a method and apparatus for identifying rope risks by integrating multi-dimensional data, which has at least the following beneficial effects:

[0050] Based on preliminary and secondary anomaly analysis, this invention not only identifies the real-time status of the rope using single data points but also enables more accurate status prediction by combining multi-dimensional data. Specifically, by performing preliminary anomaly analysis on the multi-dimensional operational data, potential problems can be quickly identified. When an anomaly is detected, data from other dimensions is comprehensively considered. In other words, this invention considers not only single anomalies but also the overall operational status of the rope to take appropriate safety measures. Furthermore, it can accurately detect the operational status of the rope and perform trend prediction and anomaly identification. Attached Figure Description

[0051] Figure 1 A flowchart of a rope risk identification method that integrates multi-dimensional data provided by the present invention;

[0052] Figure 2 A flowchart illustrating abnormal results under current multidimensional operational data is provided as an embodiment of the present invention;

[0053] Figure 3 A flowchart of a rope risk identification method provided in a certain embodiment of the present invention;

[0054] Figure 4 A schematic diagram of a rope risk identification device that integrates multi-dimensional data provided by the present invention;

[0055] Figure 5 A schematic diagram of a rope risk identification device that integrates multi-dimensional data according to a certain embodiment of the present invention;

[0056] Figure 6 This is a schematic diagram of a rope risk identification device that integrates multi-dimensional data according to a certain embodiment of the present invention. Detailed Implementation

[0057] To better understand the above technical solutions, a detailed description of the solutions will be provided below in conjunction with the accompanying drawings and specific embodiments. Obviously, the described embodiments are merely some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0058] The terminology used in the embodiments of this invention is for the purpose of describing particular embodiments only and is not intended to limit the invention. The singular forms “a,” “the,” and “the” as used in the embodiments of this invention and the appended claims are also intended to include the plural forms, and “multiple” generally includes at least two unless the context clearly indicates otherwise.

[0059] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that an article or device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such an article or device. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the article or device that includes said element.

[0060] like Figure 1 As shown, the present invention provides a method for identifying rope risks by integrating multi-dimensional data, comprising:

[0061] Collect multi-dimensional operational data of the rope pulling operation, including visual data, tension data, displacement data, and vibration data.

[0062] Based on the collected multidimensional operation data, and combined with the abnormal conditions corresponding to each dimension of operation data, preliminary anomaly analysis and processing are performed, and preliminary anomaly results for each dimension of operation data are given.

[0063] Based on the preliminary anomaly results of each dimension of operation data, the multi-dimensional operation data is integrated for secondary anomaly analysis and processing to give the rope anomaly results.

[0064] Based on the abnormal results of the rope pulling and the preliminary abnormal results of each dimension of the operation data, the risk level of the rope pulling is given to identify the rope pulling risk.

[0065] This invention, based on preliminary and secondary anomaly analysis, not only identifies the real-time status of the rope using single data points but also combines multi-dimensional data for more accurate status prediction. Specifically, by performing preliminary anomaly analysis on the multi-dimensional operational data, potential problems can be quickly identified. When an anomaly is detected, other dimensions of data are comprehensively considered. In other words, this invention considers not only single anomalies but also the overall operational status of the rope to take appropriate safety measures. Furthermore, it can accurately detect the operational status of the rope and perform trend prediction and anomaly identification.

[0066] After collecting multidimensional operation data for rope pulling, preliminary anomaly analysis can be performed based on the collected multidimensional operation data and the corresponding abnormal conditions for each dimension of operation data. Preliminary anomaly results for each dimension of operation data can be provided, which may specifically include:

[0067] Feature extraction is performed on visual data, or visual data is compared with rope image template data to provide preliminary visual anomalies of the rope; tension data is compared with the upper and lower limits of preset tension thresholds to provide preliminary tension anomalies of the rope; the current displacement-time curve is determined based on displacement data and compared with the standard displacement-time curve of the rope under normal working conditions to provide preliminary displacement anomalies of the rope; Fourier transform is performed on vibration data to obtain frequency components and vibration amplitude, which are compared with frequency range thresholds and vibration range thresholds respectively to provide preliminary vibration anomalies of the rope.

[0068] Specifically, key features can be extracted from the segmented regions of the pull rope to determine preliminary visual anomalies (e.g., damage type and location). For example, methods such as Gray-Level Co-occurrence Matrix (GLCM) can be used to extract texture features from the visual data, such as roughness and contrast, to determine if there is abnormal wear or cracks on the pull rope surface. Template matching techniques can also be used to compare the current visual data with pre-established normal pull rope image templates to determine if anomalies exist and to identify preliminary visual anomalies. For example, algorithms such as Normalized Cross-Correlation (NCC) or Structural Similarity (SSIM) can be used to compare the visual data with normal pull rope image templates to determine if there is abnormal wear or cracks on the pull rope surface. The preset tension threshold includes an upper and lower limit. When the tension of the pull rope exceeds the upper limit, the rope is considered overloaded, meaning it may face the risk of breakage. When the tension is below the lower limit, there is a risk of insufficient tension, potentially leading to insufficient operational stability. In either case, a warning signal will be issued to prompt the operator to check the pull rope condition.

[0069] By comparing the displacement-time curve with that of the guy rope under normal working conditions, preliminary results of the guy rope displacement anomalies are given, including:

[0070] The similarity between the current displacement-time curve and the standard displacement-time curve is calculated, and the similarity result is compared with the preset similarity threshold. If the similarity result exceeds the preset similarity threshold, it is judged as abnormal rope displacement; otherwise, it is judged as normal rope displacement.

[0071] Fourier transform of the vibration data yields frequency components and vibration amplitude, which are then compared with frequency range thresholds and vibration range thresholds to provide preliminary results of the rope vibration anomaly, including:

[0072] The time vibration signal is determined based on the vibration data, and a Fourier transform is performed on the time vibration signal to generate a spectrum of the rope vibration. Frequency components and vibration values ​​are extracted from the generated spectrum. If the frequency components exceed the frequency range threshold and / or the vibration value exceeds the vibration range threshold, the rope vibration is judged to be abnormal; otherwise, the rope vibration is judged to be normal.

[0073] Based on the preliminary anomaly results of each dimension of the operation data, a secondary anomaly analysis is performed by integrating multi-dimensional operation data to provide the rope anomaly results, including:

[0074] like Figure 2 As shown, when the preliminary anomaly result of at least one-dimensional operational data is determined to be anomaly, a probability analysis is performed on visual data, tension data, displacement data, and vibration data. Based on the probability analysis results, the anomaly result of the rope under the current multi-dimensional operational data is given, including:

[0075] Based on visual data, tension data, displacement data, and vibration data, the probability of the pull rope being in different abnormal states under the current multidimensional operation data is determined.

[0076] Based on the historical abnormal states of the pull rope, determine the initial probability of the pull rope being in different abnormal states;

[0077] Combining the probability of the pull rope being in different abnormal states under the current multidimensional operation data and the initial probability of the pull rope being in different abnormal states, the total probability of observing multidimensional operation data under all abnormal states of the pull rope is given.

[0078] The probability of the pull rope being in different abnormal states under the current multidimensional operation data and the initial probability of the pull rope being in different abnormal states are jointly processed, and the ratio of the joint processing result to the total probability of the pull rope being observed in multidimensional operation data under all abnormal states is calculated to give the probability of the pull rope being in different abnormal states.

[0079] Based on the probability of the rope being in different abnormal states and the probability thresholds corresponding to different abnormal states, the abnormal states that the rope exists in and the abnormal states with the highest probability are determined.

[0080] After acquiring vibration data of the pull rope and performing anomaly detection, if an anomaly is detected, a subsequent secondary anomaly detection and risk level assessment will be triggered. The presence of vibration anomalies in the pull rope is determined based on the frequency components and vibration amplitude of the spectrogram; this is a preliminary anomaly detection aimed at quickly identifying potential problems. Once an anomaly is detected, a more comprehensive risk assessment is conducted by comprehensively considering other dimensions of data (such as tension information, displacement information, etc.) and the type of anomaly detected (such as tension overload, displacement deviation, vibration anomaly). By combining data from all dimensions and anomaly detection, the overall risk level of the pull rope can be assessed and classified as low, medium, or high risk. This assessment considers not only a single anomaly but also the overall state to allow for appropriate safety measures to be taken. The abnormal states in this invention can include both normal and abnormal states, or only abnormal states.

[0081] Among these, determining the probability of the pull-up rope being in different abnormal states under the current multidimensional operational data, based on visual data, tension data, displacement data, and vibration data, may include:

[0082] Based on a neural network classification model pre-built using historical visual data and corresponding abnormal states of the rope, and combined with current visual data, the probability of observing current visual data under different abnormal states is determined; based on a normal distribution model pre-built using historical tension data and corresponding abnormal states of the rope, and combined with current tension data, the probability of observing current tension data under different abnormal states is determined; based on the reference displacement-time curves of historical displacement data corresponding to different abnormal states of the rope, similarity calculations are performed between these curves and the current displacement-time curves corresponding to the current displacement data to determine the probability of observing current displacement data under different abnormal states; vibration data is converted into frequency domain signals using fast Fourier transform to determine the spectrum, and the distribution characteristics of the spectrum are used to determine the probability of observing current vibration data under different abnormal states;

[0083] By jointly processing the probabilities of observing current visual data, tension data, displacement data, and vibration data under the same abnormal state, the probability of the pull rope being in different abnormal states under the current multidimensional operation data is obtained.

[0084] In this invention, the construction processes for the neural network classification model, the normal distribution model, and the reference displacement-time curve are all conventional. Probabilities can be determined based on the constructed models. For the neural network classification model, the output is the probability of different abnormal states. When determining probabilities using the dynamic time warping method, a similarity score is determined through similarity; the smaller the similarity value, the more similar the data, and the higher the similarity score. Normalizing the similarity score yields the corresponding probability. When using the Fast Fourier Transform (FFT) to determine probabilities, distribution characteristics in the frequency graph (such as peak frequency, bandwidth, and shape of the spectrum) can be used as the basis for judging abnormal states, determining the probability of an abnormal state based on these characteristics. Furthermore, based on the FFT, historical vibration data corresponding to different abnormal states can be transformed to determine a reference spectrum. This reference spectrum is then compared with the current spectrum derived from the current vibration data to calculate the probability of observing the current vibration data under different abnormal states. The spectrum similarity can be calculated using methods such as Euclidean distance and cosine similarity, and a similarity score is determined, ultimately providing the probability of observing the current vibration data under different abnormal states. After constructing the normal distribution model based on the operational data and abnormal states in each dimension, the mean and variance of different operational data under each abnormal state can be determined by maximum likelihood estimation. Based on the corresponding mean and variance, the probability of observing the current tension data under different abnormal states can be determined, satisfying the following relationship:

[0085]

[0086] In the formula, P(η|γ) i Let γ be the probability of observing operation data η under the i-th abnormal state, where η∈{X,Y,W,Z}. i Let σ be the i-th abnormal state, i∈1,2,...,n, where n is the number of abnormal states. 2 ηi The variance, μ, of the normal distribution model constructed for the operation data η and different abnormal states. ηi The mean of the normal distribution model constructed for the operation data η and different abnormal states.

[0087] Based on the abnormal states recorded during the rope-pulling operation, the initial probability of the rope being in different abnormal states can be determined, which may include:

[0088] Based on the abnormal states during the historical operation of the rope, determine the number of times the rope was in each abnormal state and the total number of times all abnormal states occurred;

[0089] Based on the number of times the rope is in each abnormal state and the total number of times all abnormal states are considered, the initial probability of the rope being in different abnormal states is given.

[0090] Combining the probabilities of the pull rope being in different abnormal states under the current multi-dimensional operation data and the initial probabilities of the pull rope being in different abnormal states, the total probability of observing the multi-dimensional operation data for the pull rope in all abnormal states is given, including:

[0091] The probabilities of the pull rope being in different abnormal states under the current multi-dimensional operation data and the initial probabilities of the pull rope being in different abnormal states are jointly processed according to the categories of abnormal states and the results of the joint processing are superimposed to obtain the total probability of observing the multi-dimensional operation data for the pull rope in all abnormal states.

[0092] Among them, the probabilities of the pull rope being in different abnormal states satisfy the following relationship:

[0093]

[0094] P(sensor data|γ i ) = P(X|γ i ) * P(Y|γ i ) * P(W|γ i ) * P(Z|γ i )P(sensor data)

[0095] = P(sensor data|γ1)·P(γ1) + P(sensor data|γ2)·P(γ2)

[0096] +... + P(sensor data|γ n )·P(γ n )

[0097] Among them, P(γ i |sensor data) is the probability that the pull rope is in the i-th abnormal state, n is the number of abnormal states, i = 1, 2,..., n, P(γ i ) is the initial probability that the pull rope is in the i-th abnormal state, X, Y, W, Z are visual data, tension data, displacement data, and vibration data respectively, P(sensor data|γ i ) is the probability that the pull rope is in the i-th abnormal state under the current multi-dimensional operation data, P(sensor data) is the total probability of observing the multi-dimensional operation data for the pull rope in all abnormal states, P(X|γ i ), P(Y|γ i ), P(W|γ i ), P(Z|γ i ) are the probabilities of observing the current visual data, tension data, displacement data, and vibration data in the i-th abnormal state respectively.

[0098] When calculating the initial probability of the rope being in the i-th abnormal state, historical operation data of the rope can be collected within a certain time range to record the operation under different abnormal states. For example, the number of occurrences of states such as normal tension, tension overload, displacement deviation, and abnormal vibration can be recorded. The collected data is then classified according to different abnormal states, and the number of occurrences of each abnormal state is recorded: the number of occurrences for state A (normal operation), state B (tension overload), state C (displacement deviation), state D (abnormal vibration), state E (abnormal wear), and state F (surface crack) are N, respectively. A N B N C N D N E N F ; Calculate the total number of times N represents all abnormal states. total =N A +N B +N C +N D +N E +N F Based on the number of occurrences of each abnormal state, calculate the initial probability P(Ar) = N for each state. Ar / N total Ar represents state A, state B, state C, state D, state E, or state F.

[0099] When identifying rope risks based on rope anomaly results and preliminary anomaly results of each dimension of operational data, this invention may include:

[0100] Based on the abnormal results of the rope pulling and the preliminary abnormal results of each dimension of the operation data, the number of rope pullings that are abnormal is determined.

[0101] Based on the number of abnormalities in the pull rope and the threshold range of abnormality numbers corresponding to different risk levels, the risk level of the pull rope is determined.

[0102] The risk identification result of the pull rope is determined based on the risk level of the pull rope and the abnormal state with the highest probability.

[0103] The identified risk identification results for the pull rope are the risk level and the most probable abnormal state (the most important abnormal state) of the pull rope, as well as its corresponding early warning and control operations, so as to facilitate subsequent early warning and control.

[0104] Risk levels can be categorized as low, medium, and high. A zero number of abnormalities in the pull rope indicates low risk; 1-3 or 1-2 abnormalities indicate medium risk; and more than 3 or more than 2 abnormalities indicate high risk. Early warning operations are based on the pull rope's risk level and the most probable abnormal state. This can include determining the warning intensity and category based on the risk level and the most probable abnormal state. For example, in low-risk situations, normal operation of the pull rope is allowed; in medium-risk situations, the warning category is determined based on the most probable abnormal state (e.g., different abnormal states have different warning sounds / colors / methods to alert operators to the pull rope's status); in high-risk situations, safety procedures such as adjusting the pull rope tension parameters must be performed simultaneously with the corresponding warning category.

[0105] First, potential abnormal data is identified through preliminary anomaly assessment, eliminating obviously normal states. Then, secondary anomaly assessment determines the primary abnormal states. For example, if both displacement and vibration exceed thresholds, the state with the highest posterior probability is selected as the final judgment. This method balances precision and accuracy, making it particularly suitable for complex rope-pulling operation condition assessment scenarios.

[0106] After determining the risk level of the pull rope, this invention can also display other relevant information through a display interface, such as: the current tension, displacement, and vibration data of the pull rope; specific details of abnormal states and handling suggestions; and charts showing historical data trends and the current status. The display interface provides an intuitive display of risk levels and real-time data monitoring, helping operators quickly assess the current operational status. Through detailed information and user operation functions, the display interface supports operators in making timely decisions, further enhancing the safety and reliability of the system.

[0107] The multidimensional operational data in this invention is obtained through detection using corresponding sensors or devices. For example, visual data, tension data, displacement data, and vibration data can be obtained through visual sensors, pressure sensors, displacement sensors, and acceleration sensors, respectively. In practical application scenarios, such as... Figure 3 As shown, the rope risk identification method may include:

[0108] S100: Detects the working status of the rope during rope pulling operations;

[0109] S200: Acquire data from each sensor to form sensor information;

[0110] S300: Processes sensor information using a Bayesian fusion algorithm to form rope status information and rope motion trend information;

[0111] S400: Detects rope status information and rope motion trend information through threshold detection, dynamic time warping, and frequency domain analysis to identify abnormal rope states;

[0112] S500: Assess the risk level of rope pulling operations based on abnormal conditions and rope movement trends. If the risk level exceeds the safety threshold, issue a warning signal.

[0113] In this embodiment of the invention, the rope risk identification method integrating multi-dimensional data includes detecting the rope's operational status during rope pulling operations in step S100. For details, please refer to [the relevant documentation / reference]. Figure 5 , Figure 6 The corresponding descriptions are not repeated here. The method for identifying risks in a rope-pulling operation by fusing multi-dimensional data includes acquiring data from various sensors in step S200 to form sensor information. In step S300, the method includes processing the sensor information using a Bayesian fusion algorithm to form rope state information and rope motion trend information. In step S400, the method includes detecting the rope state information and rope motion trend information through threshold detection, dynamic time warping, and frequency domain analysis to identify abnormal rope states. Abnormal states may include tension overload, displacement deviation, and abnormal vibration, as well as abnormal wear and surface cracks. In step S500, the method includes assessing the risk level of the rope-pulling operation based on the abnormal state and rope motion trend information. If the risk level exceeds a safety threshold, a warning signal is issued.

[0114] like Figure 4 As shown, the present invention also provides a rope risk identification device that integrates multi-dimensional data. Employing the above-described rope risk identification method, the device includes:

[0115] The data acquisition module is used to collect multi-dimensional operational data of the rope pulling operation, including visual data, tension data, displacement data, and vibration data.

[0116] The anomaly detection module is used to perform preliminary anomaly analysis based on the collected multidimensional operation data and the anomaly conditions corresponding to each dimension of operation data, and to give the preliminary anomaly results for each dimension of operation data.

[0117] The data processing and fusion module is used to perform secondary anomaly analysis and processing by fusing multi-dimensional operation data based on the preliminary anomaly results of each dimension of operation data, and to give the rope anomaly results.

[0118] The risk warning module is used to identify rope risks by giving the risk level of the rope based on the abnormal results of the rope pulling and the preliminary abnormal results of the operation data of each dimension.

[0119] In practical applications, the device of the present invention may include: a sensor module 100, a data acquisition module 200, a data processing and fusion module 300, an anomaly detection module 400, and a risk warning module 500. The sensor module 100 includes a vision sensor 110, a pressure sensor 120, a displacement sensor 130, and an acceleration sensor 140.

[0120] In this embodiment of the invention, the sensor module 100 is used to detect the working status of the rope during rope pulling operations. The sensor module 100 includes a vision sensor 110, a pressure sensor 120, a displacement sensor 130, and an acceleration sensor 140. It can be understood that the sensor module 100 is distributed according to different parts of the rope.

[0121] Understandably, the vision sensor 110 is used to capture the surface deformation, wear, and operating posture of the pull rope in real time. Computer vision technology is used to analyze the pull rope images to identify whether the pull rope has localized wear, breakage, or external interference. The vision sensor 110 is typically installed at key nodes along the pull rope's path to ensure that the overall movement of the pull rope and changes in its condition at key locations are captured.

[0122] Pressure sensor 120 is used to detect the tension on the pull rope and is positioned at the fixed end of the rope or at a point of high stress. When the pull rope is under tension, pressure sensor 120 senses the tension change in real time and outputs tension data. Combined with the working load of the pull rope, the system can determine whether the pull rope is overloaded or under abnormal stress. Pressure sensors can be distributed at multiple locations on the pull rope to monitor the overall stress situation.

[0123] Displacement sensor 130 is used to monitor the displacement of the pull rope, especially its horizontal or vertical offset. Displacement sensor 130 can sense the relative position change of the pull rope during movement using optical, laser, or ultrasonic technology. Displacement sensors 130 are typically distributed along the movement trajectory of the pull rope. By continuously recording the position changes of the pull rope, the motion curve of the pull rope can be plotted, and its operational stability can be analyzed.

[0124] Accelerometer 140 is used to detect the vibration state of the rope, capturing minute vibrations and impacts by measuring changes in the rope's acceleration. Accelerometer 140 is typically positioned at critical points on the rope, such as locations with high tension, to identify potential abnormal vibrations (e.g., vibrations caused by external interference or internal structural damage). This data helps determine whether the rope's operation is smooth.

[0125] In this embodiment of the invention, the data acquisition module 200 is used to acquire data from each sensor in the sensor module 100 to form sensor information.

[0126] Understandably, the data acquisition module 200 is used to acquire data from each sensor in the sensor module 100 to form sensor information. The sensor module 100 includes a vision sensor 110, a pressure sensor 120, a displacement sensor 130, and an acceleration sensor 140. These sensors are distributed to detect different state parameters of the pull rope. The data acquisition module 200 collects all this data and provides it to the data processing and fusion module 300 for further analysis and processing.

[0127] In this embodiment of the invention, the data processing and fusion module 300 is used to process sensor information according to a Bayesian fusion algorithm to form rope state information and rope motion trend information.

[0128] In this embodiment of the invention, the data processing and fusion module 300 is further used to obtain the initial probability of the rope's state information, where the initial probability is the probability of the rope's initial state. Data from the vision sensor 110, pressure sensor 120, displacement sensor 130, and acceleration sensor 140 are fused to construct a likelihood function, which represents the probability of the rope's current state. The likelihood function is processed using a weighted average to obtain the evidence probability, which is the total probability of observing specific sensor data in all possible states. The posterior probability of the rope is calculated using Bayes' theorem, where the posterior probability is the probability that the rope is in a certain state given the sensor data. The rope's state estimate is updated based on the calculated posterior probability to form the rope's state.

[0129] The data processing and fusion module 300 employs a Bayesian fusion algorithm to comprehensively process data from various sensors, generating information on the rope's state and movement trends. This module can not only identify the real-time state of the rope using data from a single sensor, but also combine data from multiple sensors for more accurate state prediction.

[0130] Understandably, the data processing and fusion module 300 first acquires the initial probability of the rope's state information. The initial probability represents the probability of the rope's initial state, and it can be based on historical data, the system's initial state, or other empirical information. In the initial stage of the rope-pulling operation, the initial probability is the system's preliminary estimate of the rope's state.

[0131] The likelihood function reflects the possible states of the rope under the current conditions. The data processing and fusion module 300 extracts key information from the data of the vision sensor 110, pressure sensor 120, displacement sensor 130, and acceleration sensor 140, and constructs the corresponding likelihood function by combining it with the kinematic model of the rope. This function represents the probability of observing the current sensor data under a specific state.

[0132] The data processing and fusion module 300 performs a weighted average of the likelihood function to obtain the evidence probability. The evidence probability is the total probability of observing the current sensor data under all possible states. By calculating the evidence probability, the system can understand the overall situation of each state under uncertainty.

[0133] According to Bayes' theorem, the data processing and fusion module 300 calculates the posterior probability of the rope, that is, the probability that the rope is in a certain state given the sensor data.

[0134] In this embodiment of the invention, the anomaly detection module 400 is used to detect the state information and movement trend information of the pull rope through threshold detection, dynamic time warping, and frequency domain analysis, so as to identify the abnormal state of the pull rope, including tension overload, displacement deviation, and abnormal vibration.

[0135] The main function of the anomaly detection module 400 is to detect the state and motion trend information of the rope through threshold detection, dynamic time warping (DTW), and frequency domain analysis, thereby identifying possible abnormal states. These abnormal states include, but are not limited to, tension overload, displacement deviation, and abnormal vibration. This multi-level detection method enables comprehensive monitoring and anomaly analysis of the rope's state.

[0136] Please refer to the following: Figure 6 , Figure 5 This is a schematic diagram of a rope risk identification device that integrates multi-dimensional data, provided in another embodiment of the present invention. Figure 5 Compared to the rope risk identification device 10 that integrates multi-dimensional data shown, Figure 6 The rope risk identification device 10, which integrates multi-dimensional data, shown in the diagram, also includes a sensor module 100, a data acquisition module 200, a data processing and fusion module 300, an anomaly detection module 400, and a risk warning module 500. The sensor module 100 includes a vision sensor 110, a pressure sensor 120, a displacement sensor 130, and an acceleration sensor 140. Figure 5 The difference is that the rope risk identification device 10, which integrates multi-dimensional data, also includes: an automatic control module 600 and a display interface 700. The risk warning module 500 also includes: a threshold detection unit 510, a dynamic time warping unit 520, and a frequency domain analysis unit 530.

[0137] In this embodiment of the invention, the risk warning module 500 is used to assess the risk level of the rope pulling operation based on the abnormal state and rope movement trend information. If the risk level exceeds the safety threshold, a warning signal is issued.

[0138] In this embodiment of the invention, the risk warning module 500 includes a threshold detection unit 510. The threshold detection unit 510 is used to acquire tension and displacement information of the pull rope from the pressure sensor 120, displacement sensor 130, and acceleration sensor 140, respectively. It sets an upper and lower limit threshold for the tension information, an offset threshold for the displacement information, and a vibration frequency threshold for the vibration information. If the tension of the pull rope exceeds the set upper limit or falls below the lower limit, it is determined that the pull rope is overloaded or underloaded, and a warning signal is issued. If the displacement of the pull rope exceeds the offset threshold, it is determined that the pull rope has an abnormal displacement, and a warning signal is issued.

[0139] Pressure sensor 120 monitors the tension of the pull rope in real time. Threshold detection unit 510 sets two thresholds for the pull rope tension: an upper tension threshold and a lower tension threshold. When the tension of the pull rope exceeds the upper threshold, threshold detection unit 510 determines that the pull rope is overloaded, meaning that the pull rope may be at risk of breakage. When the tension of the pull rope is below the lower threshold, the system considers that the pull rope is at risk of insufficient tension, which may lead to insufficient operational stability. In either case, the system will issue a warning signal, prompting the operator to check the pull rope status.

[0140] The displacement sensor 130 is used to detect the displacement of the pull rope during operation, including lateral or longitudinal offset. The threshold detection unit 510 sets a displacement offset threshold; within a certain range, the displacement of the pull rope is normal, but if the displacement exceeds the set offset threshold, the system will consider that the pull rope has deviated abnormally, possibly caused by external force or operational error. Once an abnormal displacement is detected, the system will also issue a warning signal, requiring immediate investigation and handling.

[0141] Accelerometer 140 is used to detect the vibration state of the pull rope, analyzing the vibration frequency and amplitude by measuring changes in the rope's acceleration. Threshold detection unit 510 sets vibration frequency and amplitude thresholds for the vibration state. If the pull rope's vibration frequency exceeds the set frequency range, it may indicate that the rope is subjected to abnormal external force or that the equipment is unstable. Similarly, if the vibration amplitude exceeds the set threshold, it may be due to abnormal shaking of the pull rope caused by mechanical problems or environmental influences. When abnormal vibration is detected, the system will issue a warning signal, prompting the operator to check the equipment and pull rope for any problems.

[0142] The risk warning module 500 also includes a dynamic time warping unit 520. The dynamic time warping unit 520 is used to acquire the displacement time series of the pull rope under normal operating conditions to form a standard reference trajectory. The risk warning module 500 acquires the displacement time series of the pull rope collected by the displacement sensor 130 to form the displacement curve at the current moment. The risk warning module 500 compares the standard reference trajectory with the displacement curve to obtain the difference between the standard reference trajectory and the displacement curve. If the difference exceeds a similarity threshold, it determines that the pull rope has an abnormal displacement and issues a warning signal.

[0143] The risk warning module 500's function goes beyond simple threshold detection; it employs more complex algorithms to detect anomalies in displacement and vibration, improving detection accuracy and operational safety. The dynamic time warping unit 520 uses a dynamic time warping (DTW) algorithm to precisely compare the rope's displacement time series to identify anomalies. Under normal operating conditions, the system continuously records the rope's displacement time series via displacement sensor 130. After a period of recording and processing, a standard displacement reference trajectory is formed, representing the rope's displacement characteristics under normal conditions. During actual operation, the system acquires the rope's displacement data in real time, creating a displacement curve for the current moment. This curve reflects the rope's current actual state. The dynamic time warping unit 520 compares the current displacement curve with the standard reference trajectory. Through the dynamic time warping (DTW) algorithm, the system can compare the differences between two time series, achieving accurate matching even with non-linear temporal differences. After comparison, the system calculates the similarity between the standard reference trajectory and the current displacement curve. When the similarity value is lower than a preset similarity threshold, the system determines that the displacement curve of the pull rope has changed abnormally. If the similarity difference exceeds the threshold, the dynamic time warping unit 520 will consider the displacement of the pull rope to be abnormal and immediately issue a warning signal. This method can provide predictive warnings before the pull rope is about to deviate or malfunction.

[0144] In this embodiment of the invention, the risk warning module 500 further includes a frequency domain analysis unit 530. The frequency domain analysis unit 530 is used to acquire vibration data of the rope from the accelerometer and generate a time domain signal. The risk warning module 500 performs a fast Fourier transform on the acquired time domain vibration signal to obtain a frequency domain signal and generates a spectrum of the rope vibration. The risk warning module 500 obtains the frequency components and vibration amplitude based on the spectrum. If the frequency components exceed a frequency range threshold or the vibration amplitude exceeds an amplitude threshold, the module determines that the rope is vibrating abnormally and issues a warning signal.

[0145] The frequency domain analysis unit 530 is used to perform frequency domain analysis on the vibration data of the rope to identify abnormal vibration behavior. By converting the time-domain signal into a frequency-domain signal, the system can better identify the characteristics of the vibration. The system first acquires the vibration data of the rope through the accelerometer 140. This data is presented in the form of a time-domain signal, reflecting the vibration of the rope during operation. The risk warning module 500 uses Fast Fourier Transform (FFT) to convert the time-domain vibration signal into a frequency-domain signal. FFT can decompose complex time signals into different frequency components and generate corresponding spectrum diagrams. The risk warning module 500 analyzes the generated spectrum diagrams and extracts the frequency components and vibration amplitude. These two parameters reflect the vibration characteristics of the rope: Frequency components: related to the speed and frequency change of the rope vibration; Vibration amplitude: reflecting the intensity of the vibration. The system makes judgments based on preset frequency range thresholds and amplitude thresholds. If the frequency components exceed the preset frequency range, or the amplitude exceeds the set threshold, it is considered that the rope has experienced abnormal vibration. Once abnormal vibration is detected, the risk warning module 500 immediately issues a warning signal to remind the operator to check the working status of the pull rope.

[0146] In this embodiment of the invention, the rope risk identification device 10, which integrates multi-dimensional data, further includes an automatic control module 600. The risk warning module 500 is also used to assess the risk level of the rope pulling operation based on the abnormal state information and movement trend information of the rope. The risk levels include low risk, medium risk, and high risk. If the risk warning module detects a high risk, it triggers the automatic control module 600 to take safety measures. The automatic control module 600 is used to adjust the tension parameters of the rope.

[0147] The automatic control module 600 is used to take corresponding safety measures based on the risk level assessed by the risk warning module 500 to ensure the safety of the rope pulling operation. This module prevents potential accidents by adjusting the tension parameters of the rope. The automatic control module 600 receives the risk level information of the rope from the risk warning module 500. Risk levels are divided into low, medium, and high risk. The risk level assessment considers abnormal rope status information and movement trend information. Depending on the risk level, the system takes corresponding safety measures: Low risk: Normal operation, no intervention. Medium risk: Enhanced system monitoring, alerting operators to the rope status. High risk: Immediate safety measures are taken, tension parameters are adjusted to avoid accidents. If the risk level is high, the automatic control module 600 will adjust the rope tension parameters. This includes reducing tension: If tension overload is detected, the system will automatically reduce the rope tension to prevent further overload. Increasing tension: If tension is detected as too low, the system will automatically increase the rope tension to ensure the rope remains within a safe operating range. The automatic control module 600 adjusts the actual operating state of the rope through the control system. For example, the way the tension of the pull rope is applied may be adjusted or other control mechanisms may be triggered to ensure the stability of the pull rope.

[0148] In this embodiment of the invention, the rope risk identification device 10, which integrates multi-dimensional data, also includes a display interface 700. The display interface 700 is also used to display the risk level.

[0149] Understandably, Display Interface 700 serves as the system's user interface, responsible for displaying risk levels and related information to operators for real-time monitoring and decision-making. Display Interface 700 displays risk level information in real time, including low, medium, and high risk. Different risk levels are distinguished by different visual identifiers (such as colors and icons) to ensure operators can quickly identify the current risk status. In addition to risk levels, Display Interface 700 also displays other relevant information, such as: current tension, displacement, and vibration data of the rope; specific details and handling suggestions for abnormal states; and charts showing historical data trends and the current status. When the risk level reaches high risk, Display Interface 700 triggers alarm prompts, including audible alarms and flashing indicator lights, to quickly attract the operator's attention. Display Interface 700 provides an intuitive display of risk levels and real-time data monitoring, helping operators quickly assess the current operational status. Through detailed information and user operation functions, Display Interface 700 supports operators in making timely decisions, further enhancing the system's safety and reliability.

[0150] The rope risk identification device 10, which integrates multi-dimensional data in this embodiment of the invention, achieves comprehensive risk management and real-time monitoring through the cooperation of the automatic control module 600 and the display interface 700. The automatic control module 600 can automatically adjust the rope tension when a high risk is detected, reducing safety hazards; while the display interface 700 provides a clear risk level display and operating interface, supporting operators in timely monitoring and decision-making. This integrated system design effectively improves the safety and efficiency of rope pulling operations.

[0151] Although preferred embodiments of the invention have been described, 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 both the preferred embodiments and all changes and modifications falling within the scope of the invention. Clearly, those skilled in the art can make various alterations and modifications to the invention without departing from its spirit and scope. Thus, if these modifications and modifications of the invention fall within the scope of the claims and their equivalents, the invention is also intended to include these modifications and modifications.

Claims

1. A method of fusion of multi-dimensional data for pull string risk identification, characterized in that, The method comprises the following steps: Collecting multi-dimensional operation data of the pull rope, wherein the multi-dimensional operation data comprises visual data, tension data, displacement data and vibration data; Based on the collected multi-dimensional operation data, preliminary abnormal analysis processing is performed in combination with the abnormal conditions corresponding to each dimension of operation data, and preliminary abnormal results of each dimension of operation data are given; When the preliminary abnormal result of at least one dimension of operation data is determined to be abnormal, the probabilities of the pull rope being in different abnormal states under the current multi-dimensional operation data are determined based on the visual data, the tension data, the displacement data and the vibration data; wherein the abnormal states include tension overload, displacement deviation, abnormal vibration, abnormal wear and surface crack; According to the abnormal states of the pull rope in the historical operation process of the pull rope, the number of times of the pull rope being in each abnormal state and the total number of times of all abnormal states are determined; and based on the number of times of the pull rope being in each abnormal state and the total number of times of all abnormal states, initial probabilities of the pull rope being in different abnormal states are given; In combination with the probabilities of the pull rope being in different abnormal states under the current multi-dimensional operation data and the initial probabilities of the pull rope being in different abnormal states, the total probability of the pull rope being observed in all abnormal states under the multi-dimensional operation data is given; The probabilities of the pull rope being in different abnormal states under the current multi-dimensional operation data and the initial probabilities of the pull rope being in different abnormal states are jointly processed, and the ratio of the joint processing result to the total probability of the pull rope being observed in all abnormal states under the multi-dimensional operation data is calculated, so as to give the probabilities of the pull rope being in different abnormal states; Based on the probabilities of the pull rope being in different abnormal states and the probability threshold values corresponding to different abnormal states, the abnormal state of the pull rope and the abnormal state with the largest probability are determined; Based on the abnormal result of the pull rope and the preliminary abnormal result of each dimension of operation data, the risk level of the pull rope is given to identify the risk of the pull rope.

2. The pull cord risk identification method of claim 1, wherein, Based on the collected multi-dimensional operation data, preliminary abnormal analysis processing is performed in combination with the abnormal conditions corresponding to each dimension of operation data, and preliminary abnormal results of each dimension of operation data are given, which comprises: Feature extraction is performed on the visual data, or the visual data is compared with the pull rope image template data to give the preliminary visual abnormal result of the pull rope; the tension data is compared with the preset upper and lower limits of the tension threshold to give the preliminary tension abnormal result of the pull rope; based on the displacement data, the current displacement time curve is determined and compared with the standard displacement time curve of the pull rope in the normal working state to give the preliminary displacement abnormal result of the pull rope; the Fourier transform is performed on the vibration data to give the frequency component and the vibration amplitude, which are compared with the frequency range threshold and the vibration range threshold respectively to give the preliminary vibration abnormal result of the pull rope.

3. The pull cord risk identification method of claim 2, wherein, The current displacement time curve is compared with the standard displacement time curve to give the preliminary displacement abnormal result of the pull rope, which comprises: Similarity calculation is performed on the current displacement time curve and the standard displacement time curve, and the similarity result is compared with the preset similarity threshold; when the similarity result exceeds the preset similarity threshold, it is determined that the displacement of the pull rope is abnormal, otherwise, it is determined that the displacement of the pull rope is normal. The Fourier transform of the vibration data gives the frequency component and the vibration amplitude, which are compared with the frequency range threshold and the vibration range threshold respectively to give the preliminary vibration abnormality result of the pull rope, including: According to the vibration data, the time vibration signal is determined, and the Fourier transform of the time vibration signal generates the frequency spectrum diagram of the pull rope vibration; the frequency component and the vibration amplitude are extracted according to the generated frequency spectrum diagram; when the frequency component exceeds the frequency range threshold and / or the vibration amplitude exceeds the vibration range threshold, it is judged that the pull rope vibration is abnormal, otherwise, it is judged that the pull rope vibration is normal.

4. The pull cord risk identification method of claim 1, wherein, Based on the visual data, the tension data, the displacement data and the vibration data, the probability of the pull rope being in different abnormal states under the current multi-dimensional operation data is determined, including: Based on the neural network classification model constructed in advance through historical visual data and corresponding pull rope abnormal states, and combined with the current visual data, the probability of observing the current visual data under different abnormal states is determined; based on the normal distribution model constructed in advance through historical tension data and corresponding pull rope abnormal states, and combined with the current tension data, the probability of observing the current tension data under different abnormal states is determined; the reference displacement-time curve of the historical displacement data corresponding to different abnormal states of the pull rope is used to calculate the similarity of the current displacement-time curve corresponding to the current displacement data, respectively, to determine the probability of observing the current displacement data under different abnormal states; the vibration data is converted into a frequency domain signal by fast Fourier transform to determine the frequency spectrum diagram, and the distribution characteristics of the frequency spectrum diagram are used to determine the probability of observing the current vibration data under different abnormal states; The probabilities of observing the current visual data, tension data, displacement data and vibration data under the same abnormal state are jointly processed to obtain the probability of the pull rope being in different abnormal states under the current multi-dimensional operation data.

5. The pull rope risk identification method of claim 1, wherein, Combined with the probability of the pull rope being in different abnormal states under the current multi-dimensional operation data and the initial probability of the pull rope being in different abnormal states, the total probability of observing the multi-dimensional operation data under all abnormal states of the pull rope is given, including: The probability of the pull rope being in different abnormal states under the current multi-dimensional operation data and the initial probability of the pull rope being in different abnormal states are jointly processed according to the category of abnormal states and superimposed on the joint processing result to obtain the total probability of observing the multi-dimensional operation data under all abnormal states of the pull rope.

6. The pull rope risk identification method of claim 1, wherein, Based on the pull rope abnormality result and the preliminary abnormality result of each dimensional operation data, the risk level of the pull rope is given to identify the pull rope risk, including: Based on the pull rope abnormality result and the preliminary abnormality result of each dimensional operation data, the number of abnormal pull ropes is determined; Based on the abnormal number of the pull rope and the abnormal number threshold range corresponding to different risk levels, the risk level of the pull rope is determined; Based on the risk level of the pull rope and the abnormal state with the maximum probability, the pull rope risk identification result is determined.

7. A pull string risk identification apparatus that fuses multidimensional data, characterized by, The pull rope risk identification method and device according to any one of claims 1-6, including: A data acquisition module for acquiring multi-dimensional operation data of the pull rope, wherein the multi-dimensional operation data includes visual data, tension data, displacement data and vibration data; an anomaly detection module, configured to perform preliminary anomaly analysis on the collected multi-dimensional job data in combination with corresponding anomaly conditions of each dimension of job data, and to give preliminary anomaly results of each dimension of job data; a data processing and fusion module, configured to perform secondary anomaly analysis on the multi-dimensional job data based on the preliminary anomaly results of each dimension of job data, and to give a pull rope anomaly result; a risk early warning module, configured to give a risk level of the pull rope based on the pull rope anomaly result and the preliminary anomaly results of each dimension of job data, and to identify a risk of the pull rope.

Images