Intelligent monitoring method and system for roundabout cableway

Abstract

The invention discloses an intelligent monitoring method and system for a roundabout cableway, and relates to the technical field of cableway engineering, and the method comprises the steps: configuring a regional sensing network, and generating cableway digital mapping data; executing space-time sequence characteristic analysis on the cableway digital mapping data, extracting key component state parameters, and outputting cableway system operation characteristics; integrating the expert rule base and the depth Q network in a distributed calculation mode, and generating a cableway key component health degree matrix; candidate circuitous paths are calculated, and a fault-tolerant circuitous scheme is output in combination with dynamic constraints and a path optimization strategy; reading a fault-tolerant roundabout scheme, configuring parameters of a fuzzy controller, adjusting operation parameters of a cableway system, and updating a digital mapping state based on edge nodes. According to the invention, the operation state of the cableway system can be accurately monitored in real time by arranging a regional sensing network constructed by the optical fiber strain sensor, the inertial measurement unit and the acoustic emission sensor and combining temperature-strain compensation and edge computing technologies.

Description

Technical Field

[0001] The present invention relates to the technical field of cableway engineering, and in particular to an intelligent monitoring method and system for a circuitous cableway. Background Art

[0002] Cableway systems are important transportation facilities, widely used in mountain tourism, mineral mining, engineering construction, and other fields. However, due to factors such as complex terrain and changing climates, some sections of cableway systems may present safety hazards or require maintenance. Traditional cableway maintenance often requires downtime, which not only impacts operational efficiency but can also result in financial losses.

[0003] At present, the monitoring of cableway systems mainly relies on manual inspections and simple sensor monitoring, which has the following problems: first, the monitoring method is single, making it difficult to fully grasp the operating status of the cableway system; second, the data analysis method is backward and cannot detect potential faults in time; finally, there is a lack of intelligent decision-making support, which makes it impossible to achieve rapid response and scientific decision-making when a fault occurs. Summary of the Invention

[0004] In view of the problems of the existing cableway monitoring system, such as single monitoring means, insufficient analysis ability, and lack of decision support, the present invention is proposed.

[0005] Therefore, the problem to be solved by the present invention is to achieve comprehensive perception, intelligent analysis and adaptive control of the cableway system, and realize roundabout operation while ensuring safety.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: In the first aspect, an embodiment of the present invention provides an intelligent monitoring method for a detour cableway, which includes deploying a monitoring sensor array, configuring a regional sensor network of optical fiber strain sensors, inertial measurement units and acoustic emission sensors, performing temperature-strain compensation operations on sensor data through edge computing nodes, and generating cableway digital mapping data; constructing a deep neural network architecture, performing spatial-temporal feature analysis on the cableway digital mapping data, extracting key component state parameters through an adaptive pooling layer and a cross-attention network, and outputting cableway system operation characteristics; constructing a hierarchical intelligent agent decision architecture, receiving cableway system operation characteristics, integrating an expert rule base and a deep Q network in a distributed computing manner, and generating a cableway key component health matrix; constructing a cableway network hierarchical graph model, importing the cableway key component health matrix, applying a hierarchical path search algorithm to calculate candidate detour paths, and outputting a fault-tolerant detour plan in combination with dynamic constraints and path optimization strategies; reading the fault-tolerant detour plan, configuring fuzzy controller parameters, adjusting the cableway system operation parameters, and updating the digital mapping state based on edge nodes.

[0007] As a preferred solution of the intelligent monitoring method for a circuitous cableway described in the present invention, generating cableway digital mapping data includes the following steps: arranging a fiber optic strain sensor grid, a multi-axis inertial measurement unit group and an acoustic emission sensor array in the support interval to collect support structure strain data, cableway motion parameters and mechanical vibration signals; wherein the cableway motion parameters include acceleration and angular velocity; performing temperature-strain compensation calculation to compensate the support structure strain data to obtain corrected strain data, processing the cableway motion parameters to obtain motion trajectory data, and analyzing the mechanical vibration signal to obtain fault characteristic data; establishing a parameterized cableway model based on the corrected strain data, motion trajectory data and fault characteristic data, and calculating the dynamic stress distribution, deformation field distribution and damage state distribution of the cableway; using the tensor fusion method to combine the dynamic stress distribution, deformation field distribution and damage state distribution, constructing a digital twin model of the cableway system, and generating cableway digital mapping data.

[0008] As a preferred solution of the intelligent monitoring method for a circuitous cableway described in the present invention, the following steps are included: constructing a deep neural network architecture: parsing the cableway digital mapping data into a bracket vibration feature layer, a steering motion feature layer, a driving force feature layer, and a carrying state feature layer to generate a multidimensional feature space; designing an enhanced feature extraction unit, adopting a space-time dual-stream network structure to process the multidimensional feature space, performing frequency domain decomposition and time domain analysis, and outputting a spatiotemporal feature matrix; configuring an adaptive pooling layer, adjusting the pooling parameters based on a dynamic threshold mechanism, adaptively selecting feature importance for different operating conditions, and generating a core feature matrix; constructing a multi-scale cross-attention network, performing local-global feature fusion on the core feature matrix, establishing a feature correlation graph, and extracting key component state parameters; constructing a cableway system state assessment model based on the key component state parameters, and outputting the cableway system operation characteristics including equipment state, operating parameters and safety risks.

[0009] As a preferred solution of the intelligent monitoring method for a roundabout cableway described in the present invention, generating a health matrix of key cableway components includes the following steps: constructing a hierarchical intelligent agent decision architecture, dividing the diagnosis area according to the operation characteristics of the cableway system, and configuring a structural intelligent agent, a steering intelligent agent, a drive intelligent agent, and a carrying intelligent agent in the diagnosis area, wherein the intelligent agent adopts a modular heterogeneous structure; deploying an intelligent agent diagnosis model, wherein the structural intelligent agent extracts structural stress characteristic indicators, the steering intelligent agent analyzes steering performance characteristics, the drive intelligent agent calculates dynamic response characteristics, and the carrying intelligent agent monitors operating status parameters to generate intelligent agent diagnosis data; installing an intelligent agent communication module, establishing an intelligent agent shared knowledge base based on the intelligent agent diagnosis data, adopting a two-way verification mechanism to integrate fault mode data in the expert rule base, and outputting multi-agent diagnosis opinions; setting up a reinforcement learning training environment, inputting multi-agent diagnosis opinions, executing a deep Q network algorithm based on the Actor-Critic framework, and generating an intelligent agent collaboration strategy using an experience-first replay mechanism; applying the intelligent agent collaboration strategy, respectively calculating the support system structure score, steering system performance score, drive system dynamic score, and carrying system operation score to generate a health matrix of key cableway components.

[0010] As a preferred solution of the intelligent monitoring method for a detour cableway described in the present invention, the method comprises the following steps: constructing a cableway network hierarchical graph model, marking node weights based on the health matrix of key cableway components, setting a support node set, a steering node set, a drive node set, and a carrying node set, using an adjacency matrix to represent the connection relationship between nodes, and calculating the maximum capacity between nodes; executing a hierarchical path search algorithm, normalizing the node weights, constructing a path accessibility criterion, searching for node combinations that meet the capacity constraints based on a depth-first strategy, and generating a set of candidate detour paths; executing preliminary path optimization, mapping the set of candidate detour paths to an evaluation index distribution, constructing a path evaluation function, designing a parameter update rule, and iteratively calculating an initial optimal detour path; performing dynamic calculations on the initial optimal detour path, extracting the force state, motion parameters, and control boundaries of key nodes, evaluating the safety margin of the detour path, and outputting local optimization parameters; updating the path evaluation function based on the local optimization parameters, executing path refinement optimization, performing a local search on the initial optimal detour path, and generating a fault-tolerant detour solution that meets the dynamic constraints.

[0011] As a preferred solution of the intelligent monitoring method for a detour cableway described in the present invention, wherein: adjusting the operating parameters of the cableway system includes the following steps: establishing a control priority queue based on the node force state, motion parameters and control boundaries in the fault-tolerant detour scheme, and generating a segmented control timing table; configuring the system operating parameters according to the segmented control timing table, establishing a state buffer for the drive system, steering system, support system, and carrying system, loading a multi-stage control strategy matrix, and setting a parameter smooth transition threshold; according to the control priority queue order, based on the control strategy matrix of the state buffer, issuing driving force adjustment instructions, steering angle adjustment instructions, support compensation instructions, and carrying speed adjustment instructions in sequence, and collecting working condition data during the adjustment process; using edge computing nodes to perform segmented verification on the working condition data, evaluating the control effect through strain-vibration joint analysis, and generating a parameter deviation matrix; updating the dynamic stress distribution, deformation field distribution, and damage state distribution in the cableway digital mapping data based on the parameter deviation matrix to form a closed-loop feedback channel.

[0012] As a preferred solution of the intelligent monitoring method for a circuitous ropeway according to the present invention, the specific formula of the path evaluation function is as follows: in, is the path health component, is the traffic capacity component, is the resource consumption component, is the weight coefficient, and p represents the candidate detour path.

[0013] In a second aspect, an embodiment of the present invention provides an intelligent monitoring system for a detour cableway, which includes a data acquisition module for deploying a monitoring sensor array, configuring a regional sensor network of optical fiber strain sensors, inertial measurement units and acoustic emission sensors, performing temperature-strain compensation operations on sensor data through edge computing nodes, and generating cableway digital mapping data; a feature extraction module for constructing a deep neural network architecture, performing spatial-temporal feature analysis on the cableway digital mapping data, extracting key component state parameters through an adaptive pooling layer and a cross-attention network, and outputting the cableway system operation characteristics; a health assessment module for constructing a hierarchical intelligent agent decision architecture, receiving the cableway system operation characteristics, integrating the expert rule base and the deep Q network in a distributed computing manner, and generating a health matrix of key cableway components; a path planning module for constructing a cableway network hierarchical graph model, importing the health matrix of key cableway components, applying a hierarchical path search algorithm to calculate candidate detour paths, and outputting a fault-tolerant detour plan in combination with dynamic constraints and path optimization strategies; an execution control module for reading the fault-tolerant detour plan, configuring fuzzy controller parameters, adjusting the cableway system operation parameters, and updating the digital mapping state based on edge nodes.

[0014] In a third aspect, an embodiment of the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein when the processor executes the computer program, any step of the above-mentioned intelligent monitoring method for a circuitous cableway is implemented.

[0015] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, any step of the above-mentioned intelligent monitoring method for a circuitous cableway is implemented.

[0016] The beneficial effects of the present invention are as follows: the present invention can realize real-time and accurate monitoring of the operating status of the cableway system by deploying a regional sensing network constructed by optical fiber strain sensors, inertial measurement units and acoustic emission sensors, combined with temperature-strain compensation and edge computing technology. Based on the spatial-temporal feature analysis and adaptive pooling layer design of deep neural networks, it can automatically extract the state parameters of key components and improve the accuracy and robustness of feature extraction. By combining a hierarchical intelligent agent decision-making architecture with a deep Q network, the health status of key cableway components can be accurately evaluated through multi-agent collaborative diagnosis and reinforcement learning training. By constructing a hierarchical graph model of the cableway network and combining it with dynamic constraints, applying a hierarchical path search algorithm and a multi-stage optimization strategy, a safe and reliable fault-tolerant detour plan can be quickly generated to ensure the continuous operation of the cableway system. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0018] Figure 1 The framework flow chart of the intelligent monitoring method for a circuitous cableway is shown in FIG.

[0019] Figure 2 Design a flowchart for the deep neural network architecture of the intelligent monitoring method for the circuitous ropeway.

[0020] Figure 3 A flow chart is constructed for the health matrix of key ropeway components for the intelligent monitoring method of the detour ropeway. DETAILED DESCRIPTION

[0021] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0022] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0023] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0024] Example 1, reference Figures 1 to 3 , which is the first embodiment of the present invention, provides an intelligent monitoring method for a circuitous cableway, and the framework flow chart is as follows Figure 1 Shown, including, S1: Deploy a monitoring sensor array and configure a regional sensor network consisting of fiber optic strain sensors, inertial measurement units, and acoustic emission sensors. Edge computing nodes are used to perform temperature-strain compensation operations to process sensor data and generate digital mapping data for the cableway.

[0025] Specifically, the method includes the following steps: S1.1: Arrange a fiber optic strain sensor grid, a multi-axis inertial measurement unit group, and an acoustic emission sensor array in the support area to collect support structure strain data, cableway motion parameters, and mechanical vibration signals.

[0026] Among them, the cableway motion parameters include acceleration and angular velocity.

[0027] S1.2: Perform temperature-strain compensation calculations to compensate the support structure strain data to obtain corrected strain data, process the cableway motion parameters to obtain motion trajectory data, and analyze the mechanical vibration signal to obtain fault characteristic data.

[0028] It should be noted that temperature changes will affect the strain measurement results. Temperature-strain compensation is a basic processing method in fiber optic sensing technology. It can monitor ambient temperature changes by deploying temperature sensors and use the compensation coefficient method to eliminate the temperature effect.

[0029] S1.3: Establish a parameterized cableway model based on the corrected strain data, motion trajectory data, and fault characteristic data, and calculate the dynamic stress distribution, deformation field distribution, and damage state distribution of the cableway.

[0030] Accordingly, the specific implementation of this step includes: processing the corrected strain data, calculating the principal stress value and principal stress direction at the key point position of the bracket, and obtaining the bracket node stress data; calculating the horizontal deflection angle and vertical deflection angle of the wire rope in each bracket span interval based on the motion trajectory data, and obtaining the tension distribution data in combination with the cableway running speed; extracting the vibration acceleration amplitude and frequency in the fault characteristic data, marking the abnormal vibration points that exceed the preset threshold, and generating vibration characteristic data; performing regional cumulative calculation on the bracket node stress data to obtain the dynamic stress distribution, and performing interpolation operation on the wire rope tension distribution data and the bracket displacement data to obtain the deformation field distribution; determining the stress exceeding limit area according to the dynamic stress distribution, identifying the abnormal deformation section based on the deformation field distribution, locating the high-frequency vibration part in combination with the vibration characteristic data, and forming the cableway damage state distribution.

[0031] In terms of technical implementation, the method of establishing a parametric cableway model processes and maps the corrected strain data, motion trajectory data and fault feature data in a hierarchical manner. Compared with the simple superposition or linear combination method in the existing technology, it can more accurately reflect the coupling relationship between various types of data. It not only improves the accuracy of stress distribution calculation, but also can effectively identify abnormal deformation sections and realize the precise positioning of the cableway damage status.

[0032] S1.4: Use the tensor fusion method to combine the dynamic stress distribution, deformation field distribution, and damage state distribution to construct a digital twin model of the cableway system and generate digital mapping data for the cableway.

[0033] Among them, the detailed implementation steps can be described as follows: construct a third-order tensor structure, map the dynamic stress distribution data to the first dimension, map the deformation field distribution data to the second dimension, and map the damage state distribution data to the third dimension; standardize the third-order tensor to eliminate the dimensional differences between different physical quantities and obtain normalized tensor data; calculate the nuclear norm of the normalized tensor data, extract the principal components of the tensor, and generate the state characteristic matrix of the cableway system; divide the state characteristic matrix into sub-blocks according to the support interval and wire rope section, and establish a topological structure diagram of the cableway system; organize the state characteristic matrix based on the topological structure diagram, construct a digital twin model of the cableway system, and output the digital mapping data of the cableway.

[0034] In terms of functional implementation, the tensor fusion method is used to construct a digital twin model, which overcomes the problem of difficult unified expression of multi-source heterogeneous data in traditional modeling methods. Through tensor decomposition and reconstruction, the correlation characteristics of data in each dimension are effectively retained, which not only reduces data redundancy, but also improves the model's ability to represent the status of the cableway system, laying a good foundation for subsequent feature analysis.

[0035] S2: Build a deep neural network architecture to perform spatial-temporal feature analysis on the cableway digital mapping data, extract key component state parameters through adaptive pooling layers and cross-attention networks, and output the cableway system operation characteristics.

[0036] Specifically, the deep neural network architecture design flow chart is as follows Figure 2 As shown, the following steps are included: S2.1: Parse the cableway digital mapping data into a support vibration feature layer, a steering motion feature layer, a driving force feature layer, and a carrying state feature layer to generate a multi-dimensional feature space.

[0037] It should be noted that the bracket vibration feature layer is constructed based on dynamic stress distribution data, the steering motion feature layer is constructed based on deformation field distribution data, the driving force feature layer and the carrying state feature layer are constructed based on damage state distribution data, and a multi-dimensional feature space is formed through feature reorganization.

[0038] S2.2: Design an enhanced feature extraction unit, use a spatial-temporal dual-stream network structure to process the multidimensional feature space, perform frequency domain decomposition and time domain analysis, and output a spatiotemporal feature matrix.

[0039] S2.3: Configure the adaptive pooling layer, adjust the pooling parameters based on the dynamic threshold mechanism, adaptively select the feature importance for different operating conditions, and generate the core feature matrix.

[0040] Specifically in this embodiment, the execution process of this step is: receiving the spatiotemporal feature matrix, dividing the spatial features and temporal features therein into N sub-feature blocks according to the bracket area, calculating the information entropy value and energy density value of each sub-feature block, and generating a feature statistical vector; calculating the dynamic threshold based on the feature statistical vector, using the hyperbolic tangent function to map the combined features of the information entropy value and the energy density value, comparing the mapping result with the preset benchmark threshold, and outputting the threshold adjustment coefficient; setting the pooling window size and step size parameters according to the threshold adjustment coefficient, using a small pooling window to retain detail information for feature areas above the dynamic threshold, and using a large pooling window to extract summary features for feature areas below the dynamic threshold; inputting the spatial features and temporal features in the spatiotemporal feature matrix into the configured pooling layer respectively, generating the corresponding pooling feature map, and calculating the importance score of each feature dimension; using the attention weighting mechanism, selectively fusing the pooling feature map according to the importance score, and combining the pooling results of different feature dimensions into a core feature matrix.

[0041] It should be noted that the feature importance score is calculated based on the feature information volume and volatility, and the importance weight of each dimension feature is determined by a weighted combination of information gain rate and variance contribution rate.

[0042] In terms of application results, the adaptive pooling mechanism solves the problem of traditional fixed pooling strategies struggling to balance feature preservation and computational efficiency by dynamically adjusting pooling parameters. A small pooling window is used to preserve detailed information in high-importance feature regions, while a large pooling window is used to extract high-level features in low-importance regions. This reduces computational load while ensuring the quality of key feature extraction and improving the targeted nature of feature selection.

[0043] S2.4: Construct a multi-scale cross-attention network, perform local-global feature fusion on the core feature matrix, establish a feature correlation graph, and extract key component state parameters.

[0044] Among them, the multi-scale cross-attention network includes a local feature extraction branch and a global feature extraction branch. It calculates the attention weight by cross-comparison of feature maps to achieve adaptive fusion of multi-scale features.

[0045] S2.5: Construct a cableway system status assessment model based on the status parameters of key components, and output the cableway system operation characteristics including equipment status, operating parameters, and safety risks.

[0046] Specifically, the state parameters of key components are divided into support structure parameter group, traction system parameter group and line component parameter group. A standardized processing matrix is ​​established for each parameter group, and the maximum and minimum normalization method is used to generate normalized feature vectors. A three-layer stacked long short-term memory network is constructed, and the normalized feature vector is input. The historical state information is filtered through the forget gate and input gate, and the timing characteristics are captured by bidirectional propagation. The time-varying state sequence is used to train a Gaussian mixture model to calculate the probability distribution of equipment status. The three equipment status levels of normal, warning and fault are divided into three levels based on expert experience thresholds to generate equipment status assessment results. An improved particle swarm algorithm is used to optimize the parameters of the neural fuzzy system. The time-varying state sequence is input into the optimized neural fuzzy system, and the fuzzy rule base is integrated for inference calculation to output the predicted values ​​of operating parameters such as speed, tension and power. Based on the equipment status assessment results and the predicted values ​​of operating parameters, a Bayesian network structure is constructed to calculate the system failure probability and accident risk level, and generate safety risk assessment indicators. A tensor combination method is used to fuse the equipment status assessment results, the predicted values ​​of operating parameters and the safety risk assessment indicators to construct a cableway system status assessment model, and output the cableway system operation characteristics in real time.

[0047] S3: Build a hierarchical intelligent agent decision-making architecture, receive the operation characteristics of the cableway system, use distributed computing to integrate the expert rule base and deep Q network, and generate the health matrix of key cableway components.

[0048] Specifically, the flow chart for constructing the health matrix of key cableway components is as follows: Figure 3 As shown, the following steps are included: S3.1: Construct a hierarchical intelligent agent decision-making architecture, divide the diagnosis area according to the operating characteristics of the cableway system, and configure the structural intelligent agent, steering intelligent agent, driving intelligent agent, and carrying intelligent agent in the diagnosis area respectively. The intelligent agent adopts a modular heterogeneous structure.

[0049] The diagnostic area includes the support system, steering system, drive system, and carrier system. The diagnostic area is divided based on the equipment status, operating parameters, and safety risks of the cableway system's operational characteristics: the support system area defines its boundaries based on the structural stress distribution and vibration response characteristics; the steering system area is divided based on the stress state of the guide wheel group and the motion trajectory parameters; the drive system area is defined based on the energy flow of the power transmission chain and the equipment vibration characteristics; and the carrier system area determines its monitoring range based on the operating status and load distribution characteristics of the carrier device.

[0050] S3.2: Deploy the intelligent agent diagnostic model. The structural intelligent agent extracts structural stress characteristic indicators, the steering intelligent agent analyzes steering performance characteristics, the driving intelligent agent calculates dynamic response characteristics, and the carrying intelligent agent monitors operating status parameters to generate intelligent agent diagnostic data.

[0051] S3.3: Load the intelligent agent communication module, establish an intelligent agent shared knowledge base based on the intelligent agent diagnosis data, use a two-way verification mechanism to integrate the fault mode data in the expert rule base, and output multi-agent diagnosis opinions.

[0052] Based on the above technical solution, the implementation process is as follows: construct an intelligent agent communication bus, package the intelligent agent diagnosis data of the diagnosis area into data frames at preset time intervals, and write them into the shared memory area through the communication interface; use a graph neural network to model the topological relationship between intelligent agents, construct node feature vectors based on the intelligent agent diagnosis data, calculate the correlation strength coefficient between intelligent agents, and generate an intelligent agent interaction matrix; deploy a knowledge graph engine, parse the fault mode data in the expert rule base into an entity-relationship-attribute structure, establish a fault diagnosis knowledge graph, and mark the fault propagation path and fault severity; design a two-way verification algorithm, use the intelligent agent interaction matrix to verify the fault propagation path in the fault diagnosis knowledge graph, and at the same time correct the intelligent agent diagnosis results based on the fault diagnosis knowledge graph; construct a belief propagation network, take the intelligent agent diagnosis data and the fault diagnosis knowledge graph as input, iteratively update the node status through message passing, and calculate the fault diagnosis confidence distribution; based on the fault diagnosis confidence distribution, combined with the intelligent agent interaction matrix and the fault diagnosis knowledge graph, generate a multi-agent diagnosis opinion containing fault type, fault severity and fault development trend.

[0053] S3.4: Build a reinforcement learning training environment, input multi-agent diagnostic opinions, execute the deep Q-network algorithm based on the actor-critic framework, and use the experience-first replay mechanism to generate agent collaboration strategies.

[0054] In an optional embodiment, this step specifically includes: constructing a reinforcement learning state space, mapping the multi-agent diagnostic opinions into a state vector, including fault feature identifiers, fault severity parameters and fault trend indicators, and establishing a state transition probability matrix; designing a hierarchical action space, setting independent action subspaces for different diagnostic areas, dividing the collaborative behavior of the agents into three levels: monitoring strategy adjustment, maintenance schedule and emergency response plan, and constructing an action mapping table; defining a reward function, constructing a multi-objective evaluation system based on fault diagnosis accuracy, system availability and maintenance cost, setting reward weight coefficients, and calculating the reward value of the collaborative behavior of the agents; building an Actor network and a Critic network, the Actor network adopts a multi-layer perceptron structure to output the action selection probability distribution, and the Critic network adopts temporal difference learning to estimate the state-action value function; constructing an experience replay pool to store samples of the agent interaction sequence, setting priority parameters based on the fault severity and the system impact range, and performing importance sampling to update the training sample distribution; using a deep Q network algorithm to select the optimal action sequence according to the state vector, alternately optimizing parameters through the Actor network and the Critic network, updating the policy gradient in combination with the experience replay pool, and outputting the agent collaborative strategy.

[0055] Optimally, this invention deeply integrates expert experience with real-time monitoring data by building an agent communication bus and a knowledge graph engine, employing a graph neural network to capture inter-agent connections. Compared to traditional approaches that rely on a fixed rule base, the proposed bidirectional verification mechanism and reinforcement learning approach within the actor-critic framework dynamically optimize diagnostic strategies. The introduction of an experience-first replay mechanism improves the system's adaptability and diagnostic accuracy, enabling it to maintain stable diagnostic performance even under complex operating conditions.

[0056] S3.5: Apply the intelligent agent collaborative strategy to calculate the support system structure score, steering system performance score, drive system power score, and carrying system work score respectively, and generate the health matrix of key cableway components.

[0057] S4: Construct a cableway network hierarchical graph model, import the health matrix of key cableway components, apply a hierarchical path search algorithm to calculate candidate detour paths, and output a fault-tolerant detour solution by combining dynamic constraints and path optimization strategies.

[0058] Specifically, the method includes the following steps: S4.1: Construct a cableway network hierarchical graph model, mark the node weights based on the health matrix of key cableway components, set the support node set, steering node set, drive node set, and carrying node set, use the adjacency matrix to represent the connection relationship between nodes, and calculate the maximum traffic capacity between nodes.

[0059] S4.2: Execute the hierarchical path search algorithm, normalize the node weights, construct the path accessibility criterion, search for node combinations that meet the capacity constraints based on the depth-first strategy, and generate a set of candidate detour paths.

[0060] S4.3: Perform preliminary path optimization, map the candidate detour path set to the evaluation index distribution, construct the path evaluation function, design the parameter update rule, and iteratively calculate the initial optimal detour path.

[0061] Specifically, the path evaluation function adopts a multi-objective weighted combination form, including path health component, traffic capacity component and resource consumption component. The calculation formula is as follows: Where p represents a candidate detour path, is the path health component. Based on the health matrix of key cableway components, the combined health status value of each node on the path is calculated and expressed as: , is the health of the i-th node on the path, is the node weight; is the capacity component, which is the path capacity determined by the maximum capacity between nodes; is the resource consumption component, which represents the impact of roundabout operations on equipment loss and energy consumption, and is expressed as: , E is the estimated value of energy consumption, S is the estimated value of equipment loss, is the balance coefficient; is the weight coefficient and satisfies , and dynamically adjusted according to the operating characteristics of the cableway system. The parameters are updated using the gradient descent method. Through iterative optimization, the evaluation function value converges to the local optimal solution and the initial optimal circuitous path is obtained.

[0062] The optimal evaluation function comprehensively assesses the performance of the detour route using three dimensions: health, capacity, and resource consumption. The path health component, a weighted summation of node health, ensures the overall reliability of key components along the detour route and avoids systemic risks caused by single-node failures. The capacity component focuses on the capacity of bottleneck sections along the route, ensuring the actual transport performance of the detour. The resource consumption component incorporates a balancing mechanism between energy consumption and equipment wear, achieving a coordinated optimization of operating costs and equipment lifespan. These three components are combined using dynamic weight coefficients, enabling the evaluation function to adaptively adjust the optimization objective based on the cableway system's operating status, resulting in a detour solution that balances safety, efficiency, and economy.

[0063] S4.4: Perform dynamic calculations on the initial optimal detour path, extract the stress state, motion parameters, and control boundaries of key nodes, evaluate the safety margin of the detour path, and output local optimization parameters.

[0064] S4.5: Update the path evaluation function based on the local optimization parameters, perform path refinement optimization, conduct a local search for the initial optimal detour path, and generate a fault-tolerant detour solution that meets the dynamic constraints.

[0065] S5: Read the fault-tolerant detour plan, configure the fuzzy controller parameters, adjust the cableway system operating parameters, and update the digital mapping status based on the edge nodes.

[0066] Specifically, the method includes the following steps: S5.1: Based on the node force state, motion parameters and control boundaries in the fault-tolerant detour scheme, establish a control priority queue and generate a segmented control timing table.

[0067] S5.2: Configure system operating parameters according to the segmented control timing table, establish state buffers for the drive system, steering system, support system, and carrier system, load the multi-stage control strategy matrix, and set parameter smooth transition thresholds.

[0068] Specifically, the control time domain nodes in the segmented control timing table are parsed, and the control parameter queue is constructed according to the time domain sequence. The control parameter queue is divided into a preparatory stage, an execution stage and a transition stage, and a state buffer space is allocated to each stage; for the preparatory stage, the power reserve coefficient of the drive system, the angle compensation coefficient of the steering system, the stress buffer coefficient of the bracket system, and the speed adjustment coefficient of the carrying system are calculated to construct a preparatory stage parameter matrix; for the execution stage, based on the preparatory stage parameter matrix, the driving force-steering angle coupling parameters, the steering angle-bracket deformation coupling parameters, and the bracket deformation-carrying speed coupling parameters are calculated to form an execution stage parameter matrix; for the transition stage, based on the execution stage parameter matrix, the driving power smoothing factor, the steering displacement smoothing factor, the bracket stress smoothing factor, and the carrying speed smoothing factor are calculated to generate a transition stage parameter matrix; the preparatory stage parameter matrix, the execution stage parameter matrix, and the transition stage parameter matrix are combined into a multi-stage control strategy matrix, a smooth transition function of the parameters of each stage is established, and the system adjustment boundary value and the parameter smooth transition threshold are calibrated.

[0069] Optimally, by configuring system parameters in stages, system oscillations caused by parameter switching can be effectively avoided, ensuring the smooth operation of the cableway system and achieving a smooth parameter transition. Compared to the existing single parameter adjustment method, this solution fully considers the coupling relationship between system components, improving the system's coordinated control capabilities and operational stability.

[0070] S5.3: According to the control priority queue order, based on the control strategy matrix of the state buffer, the driving force adjustment command, steering angle adjustment command, bracket compensation command, and carrying speed adjustment command are issued in sequence, and the working condition data during the adjustment process are collected.

[0071] Furthermore, the preparatory stage parameter matrix in the multi-stage control strategy matrix is ​​read, the driving motor power adjustment amount is calculated based on the power reserve coefficient, the driving force adjustment curve is generated, and the driving force adjustment instruction sequence is output; the execution status of the driving force adjustment instruction sequence is detected, the driving force-steering angle coupling parameters in the execution stage parameter matrix are read, the steering mechanism displacement compensation amount is calculated, the steering angle adjustment curve is generated, and the steering angle adjustment instruction sequence is output; the bracket force data during the steering angle adjustment process is collected, the bracket deformation compensation amount is calculated based on the steering angle-bracket deformation coupling parameters, the bracket compensation curve is generated, and the bracket compensation instruction sequence is output; the execution effect of the bracket compensation instruction is monitored, the cableway running speed correction amount is calculated in combination with the bracket deformation-carrying speed coupling parameters, the carrying speed adjustment curve is generated, and the carrying speed adjustment instruction sequence is output; the transition stage parameter matrix is ​​read, the adjustment instruction sequence is dynamically corrected based on each smoothing factor, the response parameters of each system are collected, and the working condition data is stored in the corresponding state buffer.

[0072] S5.4: Use edge computing nodes to perform segmented verification of operating condition data, evaluate control effects through strain-vibration joint analysis, and generate a parameter deviation matrix.

[0073] S5.5: Update the dynamic stress distribution, deformation field distribution, and damage state distribution in the cableway digital mapping data based on the parameter deviation matrix to form a closed-loop feedback channel.

[0074] Furthermore, this embodiment also provides an intelligent monitoring system for a detour cableway, including a data acquisition module for deploying a monitoring sensor array, configuring a regional sensor network of optical fiber strain sensors, inertial measurement units, and acoustic emission sensors, performing temperature-strain compensation operations on sensor data through edge computing nodes, and generating cableway digital mapping data; a feature extraction module for constructing a deep neural network architecture, performing spatial-temporal feature analysis on the cableway digital mapping data, extracting key component state parameters through an adaptive pooling layer and a cross-attention network, and outputting cableway system operation characteristics; a health assessment module for constructing a hierarchical intelligent agent decision architecture, receiving cableway system operation characteristics, integrating an expert rule base and a deep Q network using a distributed computing method, and generating a health matrix of key cableway components; a path planning module for constructing a cableway network hierarchical graph model, importing the health matrix of key cableway components, applying a hierarchical path search algorithm to calculate candidate detour paths, and outputting a fault-tolerant detour plan by combining dynamic constraints and path optimization strategies; and an execution control module for reading the fault-tolerant detour plan, configuring fuzzy controller parameters, adjusting cableway system operation parameters, and updating the digital mapping state based on edge nodes.

[0075] This embodiment also provides a computer device suitable for the intelligent monitoring method for a circuitous cableway, including a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the intelligent monitoring method for a circuitous cableway proposed in the above embodiment.

[0076] The computer device may be a terminal, comprising a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device comprises a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner may be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display or an electronic ink display screen, and the input device of the computer device may be a touch layer covering the display screen, or a button, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse.

[0077] This embodiment further provides a storage medium having a computer program stored thereon. When the program is executed by a processor, the intelligent monitoring method for a circuitous cableway as proposed in the above embodiment is implemented.

[0078] In summary, the present invention uses a regional sensing network constructed by deploying optical fiber strain sensors, inertial measurement units and acoustic emission sensors, combined with temperature-strain compensation and edge computing technology, to achieve real-time and accurate monitoring of the operating status of the cableway system. Based on the spatial-temporal feature analysis and adaptive pooling layer design of deep neural networks, it is possible to automatically extract the state parameters of key components and improve the accuracy and robustness of feature extraction. By combining a hierarchical intelligent agent decision-making architecture with a deep Q network, the health status of key cableway components can be accurately assessed through multi-agent collaborative diagnosis and reinforcement learning training. By constructing a hierarchical graph model of the cableway network and combining it with dynamic constraints, applying a hierarchical path search algorithm and a multi-stage optimization strategy, a safe and reliable fault-tolerant detour solution can be quickly generated to ensure the continuous operation of the cableway system.

[0079] Example 2, reference Figures 1 to 3 , which is the second embodiment of the present invention, provides an intelligent monitoring method for a circuitous cableway. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments.

[0080] To validate the effectiveness of the proposed intelligent monitoring method for circuitous cableways, a six-month experimental study was conducted on a passenger cableway system at a high mountain tourist attraction. The 2.8-kilometer-long cableway system features 12 support frames and uses 8-passenger detachable gondolas with a designed transport capacity of 2,400 passengers per hour.

[0081] The experimental platform uses an industrial-grade edge computing controller equipped with a quad-core industrial processor running at 2.0 GHz and 4GB of system memory. The sensor network consists of fiber Bragg grating strain sensors, a six-axis inertial measurement unit (IMU), and a piezoelectric acoustic emission sensor. The fiber strain sensor has a measurement range of ±2500 με and a sampling frequency of 100 Hz. The IMU has an acceleration range of ±15 g and an angular velocity range of ±2000° / s, with a sampling frequency of 200 Hz. The acoustic emission sensor operates in a frequency range of 20-400 kHz and is equipped with an adjustable-gain signal conditioning circuit.

[0082] Twenty-four fiber-optic strain sensor measurement points were placed on the support frame structure based on its load characteristics, primarily at the column bases, beam connections, and handrail attachment points. Each support frame is equipped with two inertial measurement units, one located in the middle and one at the top of the beam, to monitor structural vibration and displacement response. An acoustic emission sensor array consisting of 36 sensors, evenly distributed across the contact area between the wire rope and the guide pulley, provides real-time monitoring of mechanical wear and fatigue damage.

[0083] The deep neural network model was trained on a high-performance GPU computing platform using the PyTorch deep learning framework. The baseline threshold of the adaptive pooling layer was set to 0.65, and the pooling window size was dynamically adjusted between 4×4 and 16×16. The crisscross attention network employed a four-layer architecture with 8 attention heads and a hidden layer dimension of 256. Training was performed using the Adam optimizer with an initial learning rate of 0.001, a batch size of 64, and 200 training epochs.

[0084] To fully test the system performance, the experimental team simulated a variety of fault conditions, including excessive support structure stress, abnormal wire rope tension, guide wheel wear, and drive system power fluctuations. By collecting data under different operating conditions, the performance differences between the proposed method and traditional monitoring methods were compared and analyzed, as shown in Table 1. Table 1 Performance differences between the method of the present invention and the traditional monitoring method Evaluation Metrics Method of the present invention Traditional monitoring methods Fault detection rate (%) 95.7 83.5 False alarm rate (%) 3.1 8.2 Positioning accuracy (m) 0.9 0.9 Response time (s) 1.5 1.5 Status assessment accuracy (%) 94.8 94.8 Time to generate a detour plan (s) 4.8 4.8 Experimental data demonstrates that the proposed method significantly outperforms traditional monitoring methods across all performance metrics. In particular, the fault detection rate reached 95.7%, with a false alarm rate reduced to 3.1%. This is due to the deep fusion of multi-dimensional sensor data and the adaptive feature extraction mechanism. During actual operation, the system successfully warned and addressed multiple potential faults, including one abnormal support frame stress and two wire rope tension fluctuations.

[0085] The detour control strategy was highly effective. When the strain value of the beam on support frame No. 8 approached the warning threshold, the system automatically generated a fault-tolerant detour plan. By adjusting the force distribution and operating speed of adjacent support frames, this plan reduced the beam strain value to a safe range within 15 minutes, while maintaining normal cableway operation and unimpaired passenger comfort. Throughout the entire process, the drive system power fluctuation was controlled within ±5% of the rated power, and the wire rope tension variation did not exceed 10%, fully demonstrating the superiority of this invention in ensuring operational safety and service continuity.

[0086] Temperature-strain compensation achieved excellent results, keeping strain measurement errors within reasonable limits within an ambient temperature range of -10°C to 30°C. The digital twin model's state prediction accuracy reached 94.8%, providing a reliable basis for fault diagnosis and detour control. Over six months of operational verification, the system demonstrated excellent stability and reliability, handling 127 fault warning events and successfully avoiding eight potential major maintenance repairs, significantly improving the operational efficiency and safety of the cableway system.

[0087] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. An intelligent monitoring method for a circuitous cableway, characterized by: include, Deploy a monitoring sensor array and configure a regional sensor network consisting of fiber optic strain sensors, inertial measurement units, and acoustic emission sensors. Edge computing nodes perform temperature-strain compensation operations to process sensor data and generate digital mapping data for the cableway. Build a deep neural network architecture to perform spatial-temporal feature analysis on the cableway digital mapping data, extract key component state parameters through adaptive pooling layers and cross-attention networks, and output the cableway system operation characteristics; A hierarchical intelligent agent decision-making architecture is constructed to receive the operating characteristics of the cableway system. A distributed computing approach is used to integrate the expert rule base and the deep Q network to generate a health matrix of key cableway components. Construct a cableway network hierarchical graph model, import the health matrix of key cableway components, apply a hierarchical path search algorithm to calculate candidate detour paths, and output a fault-tolerant detour solution by combining dynamic constraints and path optimization strategies. Read the fault-tolerant detour plan, configure the fuzzy controller parameters, adjust the cableway system operating parameters, and update the digital mapping status based on the edge nodes.

2. The intelligent monitoring method for a circuitous ropeway according to claim 1, characterized in that: Generating cableway digital mapping data comprises the following steps: A fiber optic strain sensor grid, a multi-axis inertial measurement unit group, and an acoustic emission sensor array are arranged in the bracket area to collect bracket structure strain data, cableway motion parameters, and mechanical vibration signals; the cableway motion parameters include acceleration and angular velocity; Perform temperature-strain compensation calculations to compensate the support structure strain data to obtain corrected strain data, process the cableway motion parameters to obtain motion trajectory data, and analyze the mechanical vibration signal to obtain fault characteristic data; Establishing a parameterized cableway model based on the corrected strain data, the motion trajectory data, and the fault characteristic data, and calculating the dynamic stress distribution, deformation field distribution, and damage state distribution of the cableway; The tensor fusion method is used to combine the dynamic stress distribution, deformation field distribution and damage state distribution to construct a digital twin model of the cableway system and generate cableway digital mapping data.

3. The intelligent monitoring method for a circuitous ropeway according to claim 1, characterized in that: The construction of the deep neural network architecture includes the following steps: Parse the cableway digital mapping data into support vibration feature layer, steering motion feature layer, driving force feature layer, and carrying state feature layer to generate a multi-dimensional feature space; Designing an enhanced feature extraction unit, using a space-time dual-stream network structure to process the multidimensional feature space, performing frequency domain decomposition and time domain analysis, and outputting a space-time feature matrix; Configure the adaptive pooling layer, adjust the pooling parameters based on the dynamic threshold mechanism, adaptively select the feature importance for different operating conditions, and generate the core feature matrix; Constructing a multi-scale cross-attention network, performing local-global feature fusion on the core feature matrix, establishing a feature correlation graph, and extracting key component state parameters; A cableway system status assessment model is constructed based on the key component status parameters, and the cableway system operation characteristics including equipment status, operation parameters and safety risks are output.

4. The intelligent monitoring method for a circuitous cableway according to claim 1, characterized in that: Generating the health matrix of key components of the cableway comprises the following steps: A hierarchical agent decision-making architecture is constructed, and the diagnosis area is divided according to the operation characteristics of the cableway system. The diagnosis area is respectively configured with a structural agent, a steering agent, a driving agent, and a carrying agent. The agents adopt a modular heterogeneous structure. Deploy an agent diagnostic model, wherein the structural agent extracts structural stress characteristic indicators, the steering agent analyzes steering performance characteristics, the driving agent calculates dynamic response characteristics, and the transport agent monitors operating status parameters to generate agent diagnostic data; An agent communication module is installed, an agent shared knowledge base is established based on the agent diagnosis data, a two-way verification mechanism is used to integrate the fault mode data in the expert rule base, and a multi-agent diagnosis opinion is output; Build a reinforcement learning training environment, input the multi-agent diagnostic opinions, execute the deep Q-network algorithm based on the actor-critic framework, and use the experience-first replay mechanism to generate the agent collaboration strategy; Applying the intelligent agent collaborative strategy, the support system structure score, steering system performance score, drive system power score, and carrying system work score are calculated respectively to generate the health matrix of key cableway components.

5. The intelligent monitoring method for a circuitous ropeway according to claim 1, characterized in that: The method of combining dynamic constraints and path optimization strategy to output a fault-tolerant detour solution comprises the following steps: Construct a cableway network hierarchical graph model, label node weights based on the health matrix of key cableway components, set support node sets, steering node sets, drive node sets, and carrying node sets, use an adjacency matrix to represent the connection relationship between nodes, and calculate the maximum traffic capacity between nodes; Executing a hierarchical path search algorithm, normalizing the node weights, constructing a path reachability criterion, searching for node combinations that meet the capacity constraints based on a depth-first strategy, and generating a set of candidate detour paths; Performing preliminary path optimization, mapping the candidate circuitous path set into an evaluation index distribution, constructing a path evaluation function, designing parameter update rules, and iteratively calculating an initial optimal circuitous path; Performing dynamic calculations on the initial optimal detour path, extracting the stress state, motion parameters, and control boundaries of key nodes, evaluating the safety margin of the detour path, and outputting local optimization parameters; The path evaluation function is updated based on the local optimization parameters, path refinement optimization is performed, a local search is performed on the initial optimal detour path, and a fault-tolerant detour solution that meets dynamic constraints is generated.

6. The intelligent monitoring method for a circuitous cableway according to claim 1, characterized in that: The adjusting of the operating parameters of the cableway system comprises the following steps: Based on the node force state, motion parameters and control boundaries in the fault-tolerant detour scheme, a control priority queue is established and a segmented control timing table is generated; Configure system operating parameters according to the segmented control timing table, establish state buffers for the drive system, steering system, support system, and carrier system, load the multi-stage control strategy matrix, and set parameter smooth transition thresholds; According to the control priority queue order and based on the control strategy matrix of the state buffer, the driving force adjustment command, steering angle adjustment command, bracket compensation command, and carrying speed adjustment command are issued in sequence, and the working condition data during the adjustment process are collected; Use edge computing nodes to verify the working condition data in sections, evaluate the control effect through strain-vibration joint analysis, and generate a parameter deviation matrix; The dynamic stress distribution, deformation field distribution and damage state distribution in the cableway digital mapping data are updated based on the parameter deviation matrix to form a closed-loop feedback channel.

7. The intelligent monitoring method for a circuitous ropeway according to claim 5, characterized in that: The specific formula of the path evaluation function is as follows: in, is the path health component, is the traffic capacity component, is the resource consumption component, is the weight coefficient, and p represents the candidate detour path.

8. An intelligent monitoring system for a circuitous cableway, based on the intelligent monitoring method for a circuitous cableway according to any one of claims 1 to 7, characterized in that: Also includes, The data acquisition module is used to deploy the monitoring sensor array and configure a regional sensor network of optical fiber strain sensors, inertial measurement units, and acoustic emission sensors. It processes the sensor data through edge computing nodes to perform temperature-strain compensation operations and generate digital mapping data for the cableway. The feature extraction module is used to build a deep neural network architecture to perform spatial-temporal feature analysis on the cableway digital mapping data, extract key component state parameters through adaptive pooling layers and cross-attention networks, and output the cableway system operation characteristics; The health assessment module is used to build a hierarchical intelligent decision-making architecture, receive the operation characteristics of the cableway system, integrate the expert rule base and the deep Q network using distributed computing, and generate the health matrix of the key components of the cableway; The path planning module is used to build a hierarchical graph model of the cableway network, import the health matrix of key cableway components, apply a hierarchical path search algorithm to calculate candidate detour paths, and output a fault-tolerant detour plan by combining dynamic constraints and path optimization strategies; Execution control module, used to read the fault-tolerant detour plan, configure the fuzzy controller parameters, adjust the cableway system operating parameters, and update the digital mapping status based on the edge nodes A computer device comprises a memory and a processor, wherein the memory stores a computer program, and wherein the processor implements the steps of the intelligent monitoring method for a circuitous cableway according to any one of claims 1 to 7 when executing the computer program.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the intelligent monitoring method for a circuitous cableway according to any one of claims 1 to 7 are implemented.

Images