A digital-twin-based traction wheel wear fault diagnosis method

Abstract

The application discloses a kind of based on digital twinning's traction wheel wear fault diagnosis method, it is related to elevator wear field, the method comprises: to the rigid-flexible coupling dynamics simulation of traction elevator three-dimensional model, obtains the acceleration of car in wear process, car side and the steel wire rope tension of each wheel groove on counterweight side, obtains data set;Using IPOA algorithm to optimize the hyperparameter of LSTM;IPOA algorithm is the method for generating pseudo-random number using Halton sequence to initialize population, uses the adaptive nonlinear weight factor based on sine function to update the pelican optimization algorithm of individual position;According to data set, the optimized CNN-LSTM neural network is trained and tested, and a fault diagnosis model is obtained;Diagnosis is carried out using the fault diagnosis model, the application can realize high-precision, real-time and automatic traction wheel wear fault diagnosis effect, solve elevator traction wheel wear fault diagnosis problem.

Description

Technical Field

[0001] This application relates to the field of elevator wear, and in particular to a method for diagnosing traction sheave wear faults based on digital twins. Background Technology

[0002] As a key component of elevator systems, the traction sheave directly affects the safety and operational efficiency of elevators. The traction sheave achieves the elevator's ascent and descent through friction with the steel wire rope, and its wear condition has a significant impact on the performance of the entire elevator system. How to scientifically determine whether the traction sheave has failed is an urgent problem to be solved. Studying the influence of traction sheave wear on the equivalent coefficient of friction is an effective solution for scientifically determining traction sheave failure.

[0003] Most existing inspection methods rely on manual operation, involving observation of the traction sheave surface wear. While simple, this method is low in automation and efficiency, suitable only for preliminary inspections, and prone to inconsistent results. Furthermore, it is not only time-consuming and labor-intensive but also ill-suited to the demands of modern elevator management for efficient and accurate inspections.

[0004] While some researchers use image processing methods to detect wear, camera-based inspection is highly dependent on lighting conditions. Excessive or insufficient light can affect image quality, thus impacting detection accuracy. In many cases, inspections must be conducted while the equipment is stopped to ensure clear images are captured, making real-time monitoring of traction sheave wear impossible. This not only disrupts elevator operation but also allows wear issues to worsen undetected, increasing safety hazards. Implementing real-time online inspection is challenging, especially at high speeds, where obtaining clear and stable images is difficult.

[0005] Traditional detection methods lack in-depth analysis and utilization of detection data, and cannot provide detailed wear trend analysis and prediction. This makes it difficult to achieve early prevention and planned maintenance, and can only rely on periodic inspections and post-event treatments.

[0006] Therefore, a high-precision, real-time, and automated method for diagnosing traction sheave wear faults is needed to solve the problem of diagnosing elevator traction sheave wear faults. Summary of the Invention

[0007] The purpose of this application is to provide a digital twin-based method for diagnosing traction sheave wear faults, which can achieve high-precision, real-time and automated traction sheave wear fault diagnosis, and solve the problem of diagnosing elevator traction sheave wear faults.

[0008] To achieve the above objectives, this application provides the following solution:

[0009] Firstly, this application provides a method for diagnosing traction sheave wear faults based on digital twins, including:

[0010] Under various equivalent friction coefficients between the wire ropes and traction sheaves, a rigid-flexible coupling dynamic simulation was performed on the three-dimensional model of the traction elevator to obtain sensor data corresponding to various wear amounts during normal and abnormal wear processes under different equivalent friction coefficients, thus obtaining a dataset. The sensor data during normal wear includes the car acceleration, wire rope tension in each sheave groove on the car side, and wire rope tension in each sheave groove on the counterweight side. The sensor data during abnormal wear includes single-groove wear anomaly data and synchronous wear anomaly data. The single-groove wear anomaly data includes the car acceleration, wire rope tension in each sheave groove on the car side, and wire rope tension in each sheave groove on the counterweight side during abnormal wear. The synchronous wear anomaly data includes the car acceleration, wire rope tension in the target sheave groove on the car side, and wire rope tension in the target sheave groove on the counterweight side during abnormal wear.

[0011] The IPOA algorithm is used to optimize the hyperparameters of the Long Short-Term Memory Neural Network in the CNN-LSTM neural network, resulting in an optimized CNN-LSTM neural network. The IPOA algorithm is a pelican optimization algorithm that uses the Halton sequence to generate pseudo-random numbers to initialize the population and uses an adaptive nonlinear weight factor based on a sine function to update the individual positions. The population consists of multiple individuals, and each individual includes all the hyperparameters of the Long Short-Term Memory Neural Network in the CNN-LSTM neural network.

[0012] The car acceleration, the wire rope tension in the car side wheel groove, and the wire rope tension in the counterweight side wheel groove are used as inputs, and the wear fault diagnosis results are used as outputs. The optimized CNN-LSTM neural network is trained and tested based on the dataset to obtain the fault diagnosis model. The wear fault diagnosis results are normal wear, single groove wear abnormality, or synchronous wear abnormality.

[0013] The car acceleration at the current moment, the wire rope tension in each groove on the car side, and the wire rope tension in each groove on the counterweight side are input into the fault diagnosis model to obtain the wear fault diagnosis result of the traction sheave to be diagnosed.

[0014] Optionally, under various equivalent friction coefficients between the wire rope and the traction sheave, a rigid-flexible coupling dynamic simulation is performed on the three-dimensional model of the traction elevator to obtain sensor data corresponding to various wear amounts during normal and abnormal wear processes under different equivalent friction coefficients, resulting in a dataset, specifically including:

[0015] Under any equivalent friction coefficient between the wire rope and the traction sheave, a rigid-flexible coupling dynamic simulation is performed on the three-dimensional model of the traction elevator. During the simulation, the car acceleration, wire rope tension in each wheel groove on the car side, and wire rope tension in each wheel groove on the counterweight side are obtained at a first preset wear interval as the groove angle wears from the initial angle to the first preset angle. The sensor data corresponding to each wear amount during normal wear under the equivalent friction coefficient are obtained.

[0016] A rigid-flexible coupling dynamic simulation was performed on the three-dimensional model of the traction elevator. During the simulation, the car acceleration, wire rope tension of each wheel groove on the car side, and wire rope tension of each wheel groove on the counterweight side were obtained at second preset wear intervals as the groove angle wore from the first preset angle to the second preset angle. The sensor data corresponding to each wear amount during the abnormal wear process were obtained under the equivalent friction coefficient. The first preset angle is greater than the second preset angle.

[0017] Optionally, the first preset angle is 25 degrees and the second preset angle is 0 degrees.

[0018] Optionally, the car acceleration, the wire rope tension in the car side wheel groove, and the wire rope tension in the counterweight side wheel groove are used as inputs, and the wear fault diagnosis results are used as outputs. The optimized CNN-LSTM neural network is trained and tested based on the dataset to obtain a fault diagnosis model, specifically including:

[0019] The dataset is divided into a training set and a test set;

[0020] The optimized CNN-LSTM neural network is trained based on the training set, with the car acceleration, the wire rope tension in the car side wheel groove and the wire rope tension in the counterweight side wheel groove as inputs and the wear fault diagnosis results as outputs.

[0021] The trained CNN-LSTM neural network was tested using a test set, and the test results were obtained.

[0022] If the test results do not meet the requirements, return to the steps of training the optimized CNN-LSTM neural network based on the training set, with the car acceleration, the wire rope tension in the car side wheel groove and the wire rope tension in the counterweight side wheel groove as inputs and the wear fault diagnosis results as outputs.

[0023] If the test results meet the requirements, a fault diagnosis model is obtained.

[0024] Optionally, the CNN-LSTM neural network includes a convolutional neural network and a long short-term memory neural network connected in sequence.

[0025] Optionally, an adaptive nonlinear weighting factor based on a sine function is formulated as follows:

[0026] w = w max -(w max -w min sin((t / T)) m ·(π / 2)), where w represents the adaptive nonlinear weighting factor, w max w represents the maximum value of the inertia weight. min represents the minimum inertia weight, m ​​is a real number with no practical meaning, t represents the current loop count of the IPOA algorithm, T represents the maximum number of loop counts of the IPOA algorithm, and π represents pi.

[0027] According to the specific embodiments provided in this application, this application has the following technical effects:

[0028] This application provides a digital twin-based method for traction sheave wear fault diagnosis. It initializes the population using a Halton sequence-based pseudo-random number generation method and optimizes the CNN-LSTM neural network using a pelican optimization algorithm with an adaptive nonlinear weight factor based on a sine function to update individual positions, ensuring the accuracy of the CNN-LSTM neural network's recognition. Furthermore, based on digital twin technology, a rigid-flexible coupling dynamic simulation of the traction elevator's three-dimensional model is performed under various equivalent friction coefficients between the wire rope and the traction sheave to obtain a dataset. This allows the simulation process to accurately reflect the actual operating state of the traction sheave in real time. The optimized CNN-LSTM neural network is trained using this dataset to improve the accuracy of the fault diagnosis model, achieving high-precision traction sheave wear fault diagnosis. By inputting the car acceleration, the wire rope tension in the car side sheave groove, and the wire rope tension in the counterweight side sheave groove at the current moment into the fault diagnosis model, the wear fault diagnosis result of the traction sheave is obtained, achieving real-time and automated traction sheave wear fault diagnosis. Attached Figure Description

[0029] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0030] Figure 1 This is a schematic diagram showing the contact between the wire rope and the traction sheave.

[0031] Figure 2 This is a cross-sectional view of a semi-circular shape with a notched groove.

[0032] Figure 3 This is a schematic diagram showing the angle of the wheel groove before and after wear.

[0033] Figure 4 This is a schematic diagram illustrating the analysis of changes in the incision angle.

[0034] Figure 5 A flowchart illustrating a traction sheave wear fault diagnosis method based on digital twin, provided as an embodiment of this application;

[0035] Figure 6 This is a diagram of the CNN-LSTM neural network structure.

[0036] Figure 7 This is a flowchart illustrating the process of obtaining a fault diagnosis model according to an embodiment of this application. Detailed Implementation

[0037] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0038] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0039] The root cause of wire rope slippage due to wear on the traction sheave is, for example... Figure 1 As shown.

[0040] When the elevator is loading into the car and undergoing emergency braking, the tension of the steel wire ropes on both sides of the traction sheave always satisfies the following:

[0041]

[0042] Where T1 represents the car side tension; T2 represents the counterweight side tension; e represents the natural base; f represents the equivalent friction coefficient between the traction sheave groove and the wire rope; and α represents the wrap angle of the wire rope on the traction sheave.

[0043] like Figure 2 As shown, the traction capacity of the traction sheave groove depends on the equivalent coefficient of friction and the wrap angle of the wire rope on the groove. Groove wear will reduce the wrap angle, but the reduction is negligible. The equivalent coefficient of friction of the traction sheave with its slit semicircular groove is:

[0044]

[0045] Where γ is the angle of the unworn groove, and c is the distance from the center of the groove to the upper edge of the groove when it is unworn. Considering the change in depth of the traction sheave groove after wear, the relationship between the angles of the semicircular notched grooves shows that the overall center of the wire rope will drop after wear occurs between the wire rope and the traction sheave. Figure 3 As shown, assuming that during the process of mutual running-in between the wire rope and the traction sheave, if the wear depth of the sheave groove is ε, that is, the center of the fitting circle of the arc part of the sheave groove decreases by ε, which is equivalent to the horizontal area of ​​the upper edge of the sheave groove increasing by ε height.

[0046]

[0047] AC=ε·sin0.5γ (7)

[0048] Where γ′ represents the angle after wear, and d represents the diameter of the traction sheave.

[0049] By combining equations (3) and (7), the relationship between the wheel groove angle and the wear depth of the traction sheave groove can be obtained as follows:

[0050]

[0051] exist Figure 3 In the process, when the wear of the wheel groove reaches ε0, that is, when the outermost point D of the fitted circle of the wheel groove gradually wears down to the straight section of the wheel groove, the wheel groove angle γ′ decreases to zero. After this, the wheel groove continues to wear, and the wheel groove angle γ′ no longer changes. The corresponding wear amount ε0 at this time is:

[0052]

[0053] exist Figure 4 In the diagram, β represents the initial notch angle of the slit semicircular groove, β′ represents the notch angle of the slit semicircular groove after wear, and L represents the groove width. It can be observed that the notch angle of the slit semicircular groove does not change with the amount of wear on the traction sheave.

[0054] Therefore, the equivalent coefficient of friction of the notched semi-circular groove can be expressed as:

[0055]

[0056] μ represents the friction coefficient between the wire rope sheave grooves, which is 0.1 under loading conditions.

[0057] Based on this, in an exemplary embodiment, a method for diagnosing traction sheave wear faults based on digital twins is provided. First, under the equivalent friction coefficients between each wire rope and the traction sheave, a rigid-flexible coupling dynamic simulation is performed on a three-dimensional model of the traction elevator. The digital twin method is used to enable the digital twin to accurately reflect the actual operating state of the traction sheave in real time. Second, a basic CNN-LSTM neural network structure is constructed, and the network hyperparameters are optimized using the IPOA algorithm to ensure its accuracy. Finally, real-time operating data is input into the neural network to complete the detection and fault diagnosis of the actual traction sheave's operating state, thereby solving the fault diagnosis problem during elevator operation and reducing safety hazards. Figure 5 and Figure 7 As shown, the specific steps include:

[0058] Step 201: Under the equivalent friction coefficients between the wire ropes and the traction sheaves, perform rigid-flexible coupling dynamic simulation on the three-dimensional model of the traction elevator to obtain sensor data corresponding to each wear amount during normal and abnormal wear processes under each equivalent friction coefficient, thus obtaining a dataset. The sensor data during normal wear includes the car acceleration, wire rope tension in each sheave groove on the car side, and wire rope tension in each sheave groove on the counterweight side. The sensor data during abnormal wear includes single-groove wear anomaly data and synchronous wear anomaly data. The single-groove wear anomaly data includes the car acceleration, wire rope tension in each sheave groove on the car side, and wire rope tension in each sheave groove on the counterweight side during abnormal wear. The synchronous wear anomaly data includes the car acceleration, wire rope tension in the target sheave groove on the car side, and wire rope tension in the target sheave groove on the counterweight side during abnormal wear.

[0059] Step 202: The IPOA algorithm is used to optimize the hyperparameters of the Long Short-Term Memory Network (LSTM) in the CNN-LSTM neural network, resulting in the optimized CNN-LSTM neural network. The CNN-LSTM neural network uses CNN to extract local feature information, LSTM to learn features, and Softmax for fault classification. The IPOA algorithm initializes the population using a pseudo-random number generation method based on Halton sequences and employs a pelican optimization algorithm based on an adaptive nonlinear weight factor using a sine function to update individual positions. The population consists of multiple individuals, and each individual includes all the hyperparameters of the LSTM in the CNN-LSTM neural network. The IPOA algorithm is used to optimize the number of hidden layer neurons and the learning rate of the LSTM neural network. The optimization range for the LSTM neural network hyperparameters is: the number of hidden layer neurons is [1, 150], and the initial learning rate is [0.001, 0.2]. The population size of the IPOA module is set to 30, and the maximum number of iterations is 20. Accuracy is selected as the fitness function. After optimizing the LSTM neural network using the IPOA algorithm, the optimal number of hidden layer neurons, n, is obtained. L and learning rate L r .

[0060] Step 203: Using the car acceleration, the wire rope tension in the car side wheel groove, and the wire rope tension in the counterweight side wheel groove as inputs, and the wear fault diagnosis results as outputs, train and test the optimized CNN-LST M neural network based on the dataset to obtain the fault diagnosis model; the wear fault diagnosis results are normal wear, single groove wear abnormality, or synchronous wear abnormality.

[0061] Step 204: Input the current car acceleration of the traction sheave to be diagnosed, the wire rope tension of each wheel groove (5 wheel grooves) on the car side, and the wire rope tension of each wheel groove (5 wheel grooves) on the counterweight side (obtained by sensor measurement) into the fault diagnosis model to obtain the wear fault diagnosis result of the traction sheave to be diagnosed.

[0062] In another exemplary embodiment of this application, under the equivalent friction coefficients between the wire ropes and the traction sheave, a rigid-flexible coupling dynamic simulation is performed on the three-dimensional model of the traction elevator to obtain sensor data corresponding to various wear amounts during normal and abnormal wear processes under each equivalent friction coefficient, resulting in a dataset, specifically including:

[0063] Under any equivalent friction coefficient between the wire rope and the traction sheave, a rigid-flexible coupling dynamic simulation is performed on the three-dimensional model of the traction elevator. During the simulation, the car acceleration, wire rope tension in each wheel groove on the car side, and wire rope tension in each wheel groove on the counterweight side are obtained at first preset wear intervals as the groove angle wears from the initial angle to the first preset angle. The sensor data corresponding to each wear amount during normal wear under the equivalent friction coefficient are obtained.

[0064] A rigid-flexible coupling dynamic simulation was performed on a 3D model of the traction elevator. During the simulation, the car acceleration, wire rope tension in each wheel groove on the car side, and wire rope tension in each wheel groove on the counterweight side were acquired at second preset wear intervals as the groove angle wore from a first preset angle to a second preset angle. This yielded sensor data corresponding to each wear amount during abnormal wear under the equivalent friction coefficient. The first preset angle was greater than the second preset angle. The first preset angle was 25 degrees, and the second preset angle was 0 degrees.

[0065] The specific process of obtaining the dataset through simulation is as follows:

[0066] S11: Constructing a Digital Twin of the Traction Elevator: First, a 3D model of the traction elevator is created to scale using SolidWorks software. The steel cable is modeled using SolidWorks composite curves. After modeling, it is converted to Parasolid format x_t type and imported into RecurDyn for mesh generation and flexible body modeling. Components such as the traction sheave and anti-cord sheave are treated as rigid bodies.

[0067] S12: Perform rigid-flexible coupling dynamics simulation in RecurDyn: Using RecurDyn software, continuously modify the equivalent friction coefficient between the wire rope and the traction sheave. A γ greater than 25° is considered normal wear, while a γ less than 25° and greater than 0° indicates abnormal wear. Figure 2 The figure drawn in the middle refers to the angle value of the groove, which is explained in GB7588.12020. Using the full-load downward working condition as the data collection method for each simulation, with an initial groove angle of γ of 40° as an example, when γ wears down to 25° (normal wear), the wear amount corresponding to the 10mm wire rope is 0.12mm; when γ wears down from 25° to 0° (abnormal wear), the wear amount corresponding to the 10mm wire rope is 0.88mm. Simulations were performed in 0.002mm increments of wear (0mm-0.12mm) to obtain the wire rope tension and car acceleration of the five grooves, totaling 60 sets of normal wear data. Simulations were also performed in 0.01mm increments of wear (0.126mm-0.88mm) to obtain the wire rope tension and car acceleration of the five grooves, totaling 75 sets of single-groove abnormal wear and synchronous abnormal wear data for each. A high-fidelity virtual entity was constructed by continuously adjusting the simulation parameters.

[0068] Data preprocessing involves handling missing values, outliers, data normalization, and standardization to obtain a dataset.

[0069] This application utilizes RecurDyn multibody dynamics software to construct a rigid-flexible coupling model of the traction sheave and wire rope, and continuously modifies the contact friction coefficient between the traction sheave and wire rope to obtain the elevator wire rope tension and car acceleration during the fully loaded descent process. This data serves as a dataset, and the tension and acceleration signals reflect the degree of elevator wear. Furthermore, real-time sensor data is input into a fault diagnosis model to determine the current state of the traction sheave.

[0070] In another exemplary embodiment of this application, the car acceleration, the wire rope tension in the car side wheel groove, and the wire rope tension in the counterweight side wheel groove are used as inputs, and the wear fault diagnosis results are used as outputs. The optimized CNN-LSTM neural network is trained and tested based on the dataset to obtain a fault diagnosis model, specifically including:

[0071] The dataset is divided into a training set and a test set (in a 4:1 ratio).

[0072] Using the car acceleration, the wire rope tension in the car side wheel groove, and the wire rope tension in the counterweight side wheel groove as inputs, and the wear fault diagnosis results as outputs, the optimized CNN-LSTM neural network is trained based on the training set to obtain the trained CNN-LSTM neural network. During training, training is stopped by determining whether the preset iteration number T has been reached (i).

[0073] The trained CNN-LSTM neural network is tested using a test set to obtain test results. The test set is input into the trained CNN-LSTM neural network, which then outputs prediction results. The trained CNN-LSTM neural network is evaluated using four metrics: accuracy, precision, recall, and F-measure, to obtain the test results.

[0074] If the test results do not meet the requirements, the process returns to the steps of training the optimized CNN-LSTM neural network based on the training set, with the car acceleration, the wire rope tension in the car side wheel groove and the wire rope tension in the counterweight side wheel groove as inputs and the wear fault diagnosis results as outputs.

[0075] If the test results meet the requirements, a fault diagnosis model is obtained.

[0076] In another exemplary embodiment of this application, such as Figure 6As shown, the CNN-LSTM neural network consists of a Convolutional Neural Network (CNN) and a Long Short-Term Memory (LSTM) neural network connected in sequence. The network parameters are set as follows: the CNN model has a convolutional kernel size of 64, a stride of 16, and 11 input channels; the pooling layers have a pooling kernel size of 2, a stride of 2, and use 'same' padding; the first and second LSTM layers each have 128 hidden nodes; the output layer uses the softmax activation function and has 3 output nodes; the network is trained 200 times with an initial learning rate of 0.005.

[0077] In another exemplary embodiment of this application, the IPOA algorithm is used to optimize the hyperparameters of the Long Short-Term Memory neural network in the CNN-LSTM neural network. The specific steps are as follows:

[0078] Multiple initial populations are obtained by initializing the population with pseudo-random numbers generated from the Halton sequence.

[0079] In the second phase, the pelican development phase, an adaptive nonlinear weighting factor based on a sine function was introduced to update the positions of individuals in the initial population to obtain the optimal individual. The optimal individual was determined to be the optimal hyperparameter of the Long Short-Term Memory Neural Network, i.e., the optimal number of hidden layer nodes n. L and learning rate L r This further improves the optimization performance of the Pelican algorithm.

[0080] In another exemplary embodiment of this application, the use of random initialization of the population in the Pelican Optimization Algorithm (POA) may lead to insufficient population diversity, thereby reducing the search efficiency of the algorithm and slowing down the convergence speed. This application utilizes a method of generating pseudo-random numbers using Halton sequences to initialize the population, giving the algorithm good ergodicity and allowing for a more even distribution of data during the solution process.

[0081] The specific steps for initializing the population using pseudo-random numbers generated from the Halton sequence are as follows:

[0082] For a two-dimensional Halton sequence, using two prime numbers p1 and p2 as fundamental quantities, the mathematical model for its segmentation process is as follows:

[0083]

[0084] θ(n)=b0p -1 +b1·p -2 …+b m ·p-m-1 (12)

[0085] H(n)=[θ1(n),θ2(n)] (13)

[0086] Where n is Halton's ordinal number, when calculating θ1(n) in formula (13), p in formula (11) is assigned the value p1 to obtain the value n1, and then θ(n1) is calculated according to formula (12), which is θ1(n) in formula (13). When calculating θ2(n) in formula (13), p in formula (11) is assigned the value p2 to obtain the value n2, and then θ(n2) is calculated according to formula (12), which is θ2(n) in formula (13). m is a real number with no practical meaning, p is a prime number greater than or equal to 2, and b i Let ∈{0,1,…p-1} be a constant, θ(n) be the defined sequence function, H(n) be the obtained two-dimensional Halton sequence, i.e., the initial nth population, and θ1(n) and θ2(n) be the first and second individuals in the initial nth population. In practical applications, p1 = 2 and p2 = 3.

[0087] Initialization of the pelican population:

[0088] q u,v =l v +H(n)·(d v -l v (14)

[0089] In the formula, q u,v Let l represent the value of the v-th hyperparameter in the u-th individual of the initial population. v d v These are the lower and upper bounds of the v-th hyperparameter, respectively. This application optimizes the rand function using Halton sequences.

[0090] During the optimization process, individuals can find the location of the optimal solution more quickly, thereby accelerating the convergence speed of the algorithm and improving its convergence accuracy.

[0091] In another exemplary embodiment of this application, in the second stage of the existing Pelican optimization algorithm, according to the formula... (15) Update the population, since there is a linearly decreasing term with the number of iterations. This leads to the algorithm tending towards local optima, which is inconsistent with practical applications. To address this issue, this application introduces an adaptive nonlinear weight factor based on a sine function into the sparrow search algorithm, improving the problem of the algorithm converging too quickly and getting trapped in local optima in the middle and later stages of iteration. The adaptive nonlinear weight factor based on a sine function is formulated as follows:

[0092] w = w max -(wmax -w min sin((t / T)) m ·(π / 2)) (16)

[0093] Replace w in formula (15) For population renewal, among which, After the second phase update, the value of the v-th hyperparameter in the u-th individual of the population is given by R = 0.2, where w represents the adaptive nonlinear weighting factor. max w represents the maximum value of the inertia weight. min represents the minimum inertia weight, m ​​is an adjustable parameter with no practical meaning, allowing the w weight to change non-linearly, improving the algorithm's local search capability, t represents the current iteration number of the IPOA algorithm, T represents the maximum iteration number of the IPOA algorithm, π represents pi, and w max =0.9; w min =0.2; μ=2.

[0094] The simulation model in this application realistically reflects the physical states of the wire rope and traction sheave using a rigid-flexible coupling method. This model effectively simulates the flexible deformation of the wire rope and the rigidity of the traction sheave, providing more accurate simulation results. Through simulation, the responses of the wire rope and traction sheave under different operating conditions can be analyzed, thereby effectively predicting and diagnosing potential faults in the elevator system. The accuracy and high fidelity of this model make the formulation of fault detection and maintenance strategies more scientific and reliable.

[0095] This application can accurately diagnose the wear condition of the traction sheave, providing a scientific basis for determining whether the traction sheave has failed. Furthermore, by using external acceleration and compression sensors, the operation of the elevator itself will not be affected, and the measured data can be used for in-depth analysis of the impact of wear on traction sheave failure.

[0096] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0097] In this application, all actions to acquire signals, information, or data are carried out in compliance with the relevant data protection laws and policies of the country where the location is situated, and with the authorization granted by the owner of the relevant device.

[0098] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0099] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A digital-twin-based traction sheel wear fault diagnosis method, characterized in that, The traction wheel wear fault diagnosis method based on digital twinning comprises: Under each equivalent friction coefficient of the steel wire rope and the traction wheel, perform rigid-flexible coupling dynamics simulation on the three-dimensional model of the traction elevator to obtain sensor data corresponding to each wear amount in the normal wear process and the abnormal wear process under each equivalent friction coefficient, and obtain a data set; the sensor data in the normal wear process includes the car acceleration, the steel wire rope tension of each wheel groove on the car side, and the steel wire rope tension of each wheel groove on the counterweight side in the normal wear process; the sensor data in the abnormal wear process includes single-groove wear abnormal data and synchronous wear abnormal data; the single-groove wear abnormal data includes the car acceleration, the steel wire rope tension of each wheel groove on the car side, and the steel wire rope tension of each wheel groove on the counterweight side in the abnormal wear process; the synchronous wear abnormal data includes the car acceleration, the steel wire rope tension of the target wheel groove on the car side, and the steel wire rope tension of the target wheel groove on the counterweight side in the abnormal wear process; The hyperparameters of the long short-term memory neural network in the CNN-LSTM neural network are optimized using the IPOA algorithm to obtain an optimized CNN-LSTM neural network; the IPOA algorithm is a method of initializing a population by generating pseudo-random numbers using a Halton sequence, and is an ibis optimization algorithm that updates individual positions using an adaptive nonlinear weight factor based on a sine function; the population includes multiple individuals, and each individual includes all hyperparameters of the long short-term memory neural network in the CNN-LSTM neural network; the adaptive nonlinear weight factor based on the sine function has the formula: wherein, denotes an adaptive non-linear weight factor, denotes an inertia weight maximum, denotes an inertia weight minimum, is a real number without practical meaning, denotes the number of the current loop execution of the IPOA algorithm, denotes the maximum number of loop executions of the IPOA algorithm, denotes the mathematical constant pi; The car acceleration, the steel wire rope tension of the wheel groove on the car side, and the steel wire rope tension of the wheel groove on the counterweight side are input, and the wear fault diagnosis result is output, the optimized CNN-LSTM neural network is trained and tested according to the data set to obtain a fault diagnosis model; the wear fault diagnosis result is normal wear, single-groove wear abnormality, or synchronous wear abnormality; The car acceleration, the steel wire rope tension of each wheel groove on the car side, and the steel wire rope tension of each wheel groove on the counterweight side at the current time of the traction wheel to be diagnosed are input into the fault diagnosis model to obtain the wear fault diagnosis result of the traction wheel to be diagnosed.

2. The digital-twin-based traction sheel wear fault diagnosis method according to claim 1, characterized in that, Under each equivalent friction coefficient of the steel wire rope and the traction wheel, perform rigid-flexible coupling dynamics simulation on the three-dimensional model of the traction elevator to obtain sensor data corresponding to each wear amount in the normal wear process and the abnormal wear process under each equivalent friction coefficient, and obtain a data set, specifically including: Under each equivalent friction coefficient of the steel wire rope and the traction wheel, perform rigid-flexible coupling dynamics simulation on the three-dimensional model of the traction elevator, and in the process of simulation, obtain the car acceleration, the steel wire rope tension of each wheel groove on the car side, and the steel wire rope tension of each wheel groove on the counterweight side corresponding to each wear amount in the process of wear from the initial angle to the first preset angle at a first preset wear amount interval, to obtain the sensor data corresponding to each wear amount in the normal wear process under the equivalent friction coefficient; The rigid-flexible coupling dynamics simulation is performed on the three-dimensional model of the traction elevator, and in the simulation process, the car acceleration, the steel wire rope tension of each wheel groove on the car side, and the steel wire rope tension of each wheel groove on the counterweight corresponding to each wear amount in the process of the groove angle being worn from the first preset angle to the second preset angle at a second preset wear amount interval are obtained to obtain the sensor data corresponding to each wear amount in the abnormal wear process under the equivalent friction coefficient; the first preset angle is greater than the second preset angle.

3. The digital-twin-based traction sheel wear fault diagnostic method according to claim 2, characterized in that, The first preset angle is 25 degrees, and the second preset angle is 0 degrees. 4.The digital-twin-based traction sheel wear fault diagnosis method according to claim 1, wherein, The car acceleration, the steel wire rope tension of each wheel groove on the car side, and the steel wire rope tension of each wheel groove on the counterweight are taken as inputs, and the wear fault diagnosis result is taken as output, the optimized CNN-LSTM neural network is trained and tested according to the data set to obtain a fault diagnosis model, and the method specifically comprises the following steps: The data set is divided into a training set and a test set; The car acceleration, the steel wire rope tension of each wheel groove on the car side, and the steel wire rope tension of each wheel groove on the counterweight are taken as inputs, and the wear fault diagnosis result is taken as output, the optimized CNN-LSTM neural network is trained according to the training set to obtain a trained CNN-LSTM neural network; The trained CNN-LSTM neural network is tested by using the test set to obtain a test result; If the test result does not meet the requirements, the step of taking the car acceleration, the steel wire rope tension of each wheel groove on the car side, and the steel wire rope tension of each wheel groove on the counterweight as inputs, and taking the wear fault diagnosis result as output, and training the optimized CNN-LSTM neural network according to the training set to obtain the trained CNN-LSTM neural network is returned; If the test result meets the requirements, the fault diagnosis model is obtained. 5.The digital-twin-based traction sheel wear fault diagnosis method according to claim 1, characterized in that, The CNN-LSTM neural network comprises a convolutional neural network and a long short-term memory neural network connected in sequence.

Images