Escalator health management system and method based on digital twinning

Abstract

The invention provides an escalator health management system and method based on digital twinning, and the method comprises the steps: constructing a digital twinning model of an escalator through employing a point cloud reconstruction and multi-source sensing data deep integration technology; constructing a multi-level distributed intelligent sensing monitoring network of the escalator, and obtaining first monitoring data according to the first monitoring strategy; performing bidirectional synchronization of the digital twin model and the entity operation state according to the first key index; in combination with the first key index, quantum enhanced equipment state evaluation and prediction are implemented, and a first analysis result is obtained; dynamically optimizing the first monitoring strategy through a deep reinforcement learning framework according to the first analysis result; generating a first maintenance scheme according to the first analysis result; and sending the first analysis result and the first maintenance scheme to a group intelligent learning platform based on privacy calculation. According to the scheme, intelligence and individuation of escalator maintenance can be achieved, so that the service life of the escalator is prolonged, the operation safety is improved, and the maintenance cost is reduced.

Description

Technical Field

[0001] This invention relates to the field of digital twin technology, specifically to a health management system and method for escalators and elevators based on digital twins. Background Technology

[0002] Escalators and elevators are essential transportation tools in buildings such as hospitals, shopping malls, subway stations, and train stations, making their safety management particularly crucial. However, existing escalator safety management technologies often rely on regular manual inspections, which can lead to high labor costs, insufficient management efficiency, a lack of real-time monitoring and intelligent digital analysis capabilities, and the potential for overlooking safety hazards due to human factors. Summary of the Invention

[0003] Based on the above-mentioned problems, this invention proposes a digital twin-based health management system and method for escalators and elevators. The solution of this invention can realize intelligent and personalized maintenance of escalators and elevators, thereby extending the service life of escalators and elevators, improving operational safety, and reducing maintenance costs.

[0004] In view of this, one aspect of the present invention proposes a health management method for escalators and elevators based on digital twins, comprising: A digital twin model of an escalator or elevator is constructed using point cloud reconstruction and deep integration technology of multi-source sensor data. Construct a multi-level distributed intelligent sensing and monitoring network for elevators and escalators, and acquire first monitoring data according to a preset first monitoring strategy; Based on the first key indicator in the first monitoring data, conduct two-way synchronization between the digital twin model and the physical operation status; Based on the first key indicators in the first monitoring data, a state assessment and prediction of the quantum-enhanced device was carried out to obtain the first analysis results; Based on the results of the first analysis, the first monitoring strategy is dynamically optimized using a deep reinforcement learning framework. A first maintenance plan is generated based on the first analysis results; The first analysis results and the first maintenance plan were sent to a swarm intelligence learning platform based on privacy computing.

[0005] Optionally, the step of constructing a digital twin model of the escalator using point cloud reconstruction and deep integration technology of multi-source sensor data includes: Collect three-dimensional spatial information of elevators and escalators; Preprocess the point cloud data in the acquired 3D spatial information; Geometric model reconstruction is performed based on preprocessed point cloud data to obtain the geometric model of escalator; Acquire multi-source sensor data and perform preprocessing; By combining the preprocessed multi-source sensor data, a multi-physics coupling model is constructed to obtain the physical field model of the escalator; By fusing the geometric model of the escalator with the physical field model of the escalator, a digital twin model of the escalator is obtained. To achieve parameter mapping between sensor data and digital twin models; Optimize digital twin model parameters; Establish a time-varying parameter model and integrate it into the digital twin model.

[0006] Optionally, the step of constructing a multi-level distributed intelligent sensing and monitoring network for escalators and elevators, and acquiring first monitoring data according to a preset first monitoring strategy, includes: Identify and plan the layout of key monitoring points for elevators and escalators; A multi-level sensor network architecture is constructed based on the identification results and layout planning of key monitoring points; Deploy edge computing devices; Develop an adaptive multi-level monitoring strategy; Establish an edge intelligence preprocessing mechanism; Establish a sensor data quality control mechanism; Build an edge-cloud collaborative processing architecture; According to the preset first monitoring strategy, the monitoring data is collected and preprocessed, and the processed first monitoring data is transmitted to the digital twin management server.

[0007] Optionally, the step of bidirectionally synchronizing the digital twin model with the entity's operational status based on the first key indicator in the first monitoring data includes: Determine the mapping relationship between key indicators and synchronization parameters; Establish a distributed ledger data verification chain; Perform difference calculations between real-time data and the digital twin model; Establish a dynamic threshold adaptive adjustment mechanism for digital twin models; Based on the results of the difference calculation and the dynamic threshold adaptive adjustment mechanism, a self-optimization algorithm for the digital twin model is implemented. Establish an impact propagation model for digital twin models; Execute reverse guidance from the digital twin model to the entity; Establish a synchronous state assessment and recording mechanism.

[0008] Optionally, the step of combining the first key indicator from the first monitoring data to perform quantum-enhanced device status assessment and prediction to obtain the first analysis result includes: A quantum feature mapping framework is constructed by combining the first key indicator; Build a variable quantum neural network architecture; Implement a quantum-classical hybrid optimization algorithm; Strategies for constructing quantum parallel computing; Establish a quantum-enhanced anomaly detection model; Implement quantum-enhanced fault diagnosis; Develop quantum-enhanced lifetime prediction models; Perform quantum uncertainty quantization and assessment, and generate the first analysis results.

[0009] Optionally, in the step of dynamically optimizing the first monitoring strategy using a deep reinforcement learning framework based on the first analysis result, the reward function of the reinforcement learning algorithm used is: .

[0010] Optionally, k is a dynamic sensitivity adjustment factor; the value of k is calculated according to the following formula: .

[0011] Optionally, the step of generating the first maintenance plan based on the first analysis result includes: A fault risk matrix for elevator and escalator components was constructed based on the results of the first analysis. Based on the fault risk matrix, a list of high-risk components and their associated parts with risk values ​​higher than the first preset risk value are extracted, and the optimal maintenance component combination scheme is determined based on the component dependency graph in the digital twin model. A maintenance simulation sub-model is established in the digital twin model, and the maintenance plan to be executed is implanted into the sub-model. The effectiveness and impact of the maintenance plan are verified through full-dimensional physical characteristic simulation. Based on the maintenance simulation results, a multi-objective optimization function is constructed, which simultaneously considers maintenance costs, equipment downtime, degree of safety risk elimination, and maintenance operation complexity, and solves to obtain the Pareto optimal maintenance scheme set. From the Pareto optimal maintenance solution set, the solution with the highest comprehensive score is selected as the first maintenance solution, and a digital maintenance guidance package is generated; The first maintenance plan is analyzed for resource requirements using edge computing nodes. It is matched with the current availability of maintenance resources. When a resource gap is found, an automatic reservation process is triggered and the maintenance time window is updated.

[0012] Optionally, the step of sending the first analysis result and the first maintenance plan to the privacy-preserving swarm intelligence learning platform includes: Data preprocessing is performed on the first analysis results and the first maintenance plan for the elevators and escalators; A data privacy protection algorithm based on homomorphic encryption is used to encrypt the preprocessed first analysis result and the first maintenance scheme to generate ciphertext data packets; The encrypted data packets are transmitted to a privacy-preserving swarm intelligence learning platform via a secure communication protocol. In the swarm intelligence learning platform, secure multi-party computation technology is used to perform joint computation on encrypted data from multiple escalator and elevator stations while maintaining data encryption. Build a blockchain-based cross-site maintenance knowledge sharing network to ensure the protection of contributors' rights and transparency in data use; Obtain privacy-protected knowledge feedback from swarm intelligence learning platforms; The acquired knowledge feedback is integrated and adapted with the local digital twin model to update the local equipment status assessment model, fault prediction algorithm, and maintenance plan generation strategy, forming a closed-loop optimization.

[0013] Another aspect of the present invention provides a digital twin-based escalator health management system for executing a digital twin-based escalator health management method, comprising: a digital twin management server, an edge computing node, and a swarm intelligence learning platform; The digital twin management server is configured as follows: A digital twin model of an escalator or elevator is constructed using point cloud reconstruction and deep integration technology of multi-source sensor data. Construct a multi-level distributed intelligent sensing and monitoring network for elevators and escalators, and acquire first monitoring data according to a preset first monitoring strategy; Based on the first key indicator in the first monitoring data, conduct two-way synchronization between the digital twin model and the physical operation status; Based on the first key indicators in the first monitoring data, a state assessment and prediction of the quantum-enhanced device was carried out to obtain the first analysis results; Based on the results of the first analysis, the first monitoring strategy is dynamically optimized using a deep reinforcement learning framework. A first maintenance plan is generated based on the first analysis results; The first analysis results and the first maintenance plan were sent to a swarm intelligence learning platform based on privacy computing.

[0014] The technical solution of this invention provides a digital twin-based health management method for escalators and elevators, comprising: constructing a digital twin model of the escalator and elevator using point cloud reconstruction and deep integration of multi-source sensor data; constructing a multi-level distributed intelligent sensor monitoring network for the escalator and elevator, and acquiring first monitoring data according to a first monitoring strategy; performing bidirectional synchronization between the digital twin model and the physical operating status based on a first key indicator; implementing quantum-enhanced equipment status assessment and prediction based on the first key indicator to obtain a first analysis result; dynamically optimizing the first monitoring strategy based on the first analysis result using a deep reinforcement learning framework; generating a first maintenance plan based on the first analysis result; and sending the first analysis result and the first maintenance plan to a privacy-preserving swarm intelligence learning platform. By integrating quantum computing and digital twin technologies, the predictive capabilities of complex systems are significantly improved. Privacy-preserving computing and distributed ledgers solve the coordination problem between equipment swarm intelligence and data security. A virtual-physical linked maintenance verification system effectively reduces operating and maintenance expenses, enabling intelligent and personalized escalator and elevator maintenance, thereby extending the service life of the escalator and elevator, improving operational safety, and reducing maintenance costs. Attached Figure Description

[0015] Figure 1 This is a flowchart of a digital twin-based health management method for escalators and elevators provided in one embodiment of the present invention; Figure 2 This is a schematic block diagram of a digital twin-based health management system for escalators and elevators provided in one embodiment of the present invention. Detailed Implementation

[0016] To better understand the above-mentioned objectives, features, and advantages of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in these embodiments can be combined with each other.

[0017] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and therefore the scope of protection of the invention is not limited to the specific embodiments disclosed below.

[0018] The terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or apparatuses.

[0019] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0020] The following reference Figures 1 to 2 This invention describes a digital twin-based health management system and method for escalators and elevators, provided by some embodiments of the present invention.

[0021] like Figure 1 As shown, one embodiment of the present invention provides a method for health management of escalators and elevators based on digital twins, including: A digital twin model of an escalator or elevator is constructed using point cloud reconstruction and deep integration technology of multi-source sensor data. In this step, a digital twin model of the escalator is constructed, including: building a physical and geometric model of the escalator to characterize its structure and spatial relationships; establishing a dynamic model of the escalator to describe the motion and stress state of its components; establishing a time-varying parameter model of the escalator to characterize the performance changes of its components over time; and creating an electrical control model to describe the electrical control processes, specifications, and operations of its components. The dynamic model integrates the dynamic equations of the transmission chain and a stress calculation engine. The multi-source sensor data includes real-time data from intelligent sensors deployed on the escalator, such as vibration sensors, temperature sensors, sound sensors, image sensors, acceleration sensors, load sensors, deformation sensors, and displacement sensors, as well as current, voltage, and running time data from the escalator's operation control system. This step achieves a high degree of consistency between the physical equipment and the virtual representation, overcoming the technical challenge of limited accuracy in traditional single-parameter modeling.

[0022] Construct a multi-level distributed intelligent sensing and monitoring network for elevators and escalators, and acquire first monitoring data according to a preset first monitoring strategy; In this step, multi-dimensional intelligent sensors, including vibration sensors, temperature sensors, sound sensors, image sensors, acceleration sensors, load sensors, deformation sensors, and displacement sensors, are installed in the core parts of the escalator to collect data in real time (first monitoring data). Edge computing nodes then perform real-time data filtering and feature extraction on the collected first monitoring data, and transmit the first key indicator from the first monitoring data to the digital twin management server. This step significantly reduces the data transmission load, enabling preliminary identification of key anomalies at the microsecond level.

[0023] Based on the first key indicator in the first monitoring data, conduct two-way synchronization between the digital twin model and the physical operation status; In this step, a trusted data verification chain is constructed using distributed ledger technology. Real-time collected parameters are cryptographically verified against virtual model predictions. When the discrepancy exceeds a dynamic threshold, the model's self-optimization algorithm is activated. This step ensures the synchronization of virtual and physical data, solving the problem of "cumulative deviation" in digital twin models over long-term operation.

[0024] Based on the first key indicators in the first monitoring data, a state assessment and prediction of the quantum-enhanced device was carried out to obtain the first analysis results; In this step, a Variable Quantum Neural Network (VQNN) is deployed in the digital twin management server to encode multidimensional sensor data (the first key indicator in the first monitoring data) into qubit states. Features are processed in parallel through parameterized quantum circuits to output the first analysis results (such as Component Remaining Lifetime (RULC) and health index). Compared with traditional deep learning networks, this approach effectively improves the accuracy of health prediction in complex environments.

[0025] Based on the results of the first analysis, the first monitoring strategy is dynamically optimized using a deep reinforcement learning framework. In this step, based on the analysis results from the previous step, the monitoring strategy is dynamically optimized using a deep reinforcement learning framework, including the frequency of thermal imaging capture, the sensitivity of vibration sensors, and the parameters of the visual recognition algorithm. Through this step, monitoring resources are intelligently allocated, and the overall energy consumption of the system is reduced.

[0026] A first maintenance plan is generated based on the first analysis results; In this step, the maintenance plan to be executed is embedded into the digital twin model. The maintenance effect is verified through full-dimensional physical characteristic simulation, generating an immersive mixed reality (MR) technical guidance plan that includes component priority ranking and work path planning. This step can avoid ineffective maintenance interventions and shorten the average maintenance time.

[0027] The first analysis results and the first maintenance plan were sent to a swarm intelligence learning platform based on privacy computing.

[0028] In this step, the diagnostic data (first analysis results and first maintenance plan) of the digital twin models of escalators and elevators at each station are encrypted and submitted to a secure multi-party computing platform. This platform enables collaborative knowledge evolution across the equipment group through intelligent protocols. By ensuring data security, this step significantly improves the accuracy of anomaly identification in newly commissioned equipment.

[0029] The technical solution adopted in this embodiment significantly improves the predictive capabilities of complex systems by integrating quantum computing and digital twin technology. It solves the coordination problem between device swarm intelligence and data security by utilizing privacy computing and distributed ledger. It effectively reduces operation and maintenance expenses through a virtual-physical linkage maintenance and verification system. It can realize intelligent and personalized maintenance of escalators and elevators, thereby extending the service life of escalators and elevators, improving operational safety, and reducing maintenance costs.

[0030] In some possible embodiments of the present invention, the step of constructing a digital twin model of an escalator using point cloud reconstruction and deep integration technology of multi-source sensor data includes: Collect 3D spatial information of escalators and elevators, including: using a high-precision 3D laser scanner to scan the overall structure of the escalator and elevator and its surrounding environment to obtain raw point cloud data with a density of not less than a preset value (such as 1000 points / square meter); collecting precise dimensional parameters of each component of the escalator and elevator, including step width, handrail length, tilt angle and connection structure; and recording environmental parameters of the escalator installation location, including the layout of surrounding buildings, the width of pedestrian passages and the distance between adjacent facilities. The point cloud data in the acquired 3D spatial information is preprocessed, including: density equalization of the original point cloud using a voxel downsampling algorithm to remove redundant points; removal of noise points using a statistical outlier filtering algorithm to improve point cloud quality; and registration and fusion of multi-site scan data using an iterative nearest point (ICP) algorithm to construct a complete 3D point cloud model of the escalator. Based on the preprocessed point cloud data, a geometric model is reconstructed to obtain the escalator geometric model, including: dividing the point cloud data into different component regions using a region growing segmentation algorithm; extracting basic geometric features using the RANSAC algorithm to identify basic shapes such as planes, cylinders, and spheres; reconstructing complex surface shapes using a B-spline surface fitting algorithm; and assembling the segmented and identified components into a complete escalator geometric model through topological relationships. Acquire and preprocess multi-source sensor data, including: collecting time-series data from vibration sensors, temperature and humidity sensors, sound sensors, image sensors, and mechanical sensors; collecting operating parameters such as current, voltage, and speed output from the escalator control system, as well as structural parameters and material properties of the escalator; performing noise reduction, outlier detection, and standardization on various types of sensor data; applying wavelet transform to perform time-frequency analysis on vibration signals to extract characteristic frequencies and amplitudes; and so on. By combining preprocessed multi-source sensor data, a multi-physics coupling model is constructed to obtain the physical field model of the escalator, including: establishing the dynamic equations of the mechanical transmission system to describe the mechanical relationships between components such as gearbox, chain, and steps; constructing an electrical system model, including the electrical characteristics and logical relationships of the motor, controller, and sensor network; establishing a thermodynamic model to describe the heat conduction, heat convection, and thermal stress distribution of each component; and constructing a structural stress analysis model using the finite element method to calculate the stress distribution and deformation of key components. The geometric model and the physical field model of the escalator are fused to obtain a digital twin model of the escalator. This includes: establishing a spatial mapping relationship between the geometric model and the physical model; assigning corresponding physical parameters and material properties to each component in the geometric model; and constructing topological constraint relationships between components to ensure the consistency between the physical model and the geometric model. To achieve parameter mapping between sensor data and digital twin models, the following steps are taken: establishing a spatial mapping between the physical location of the sensor and the corresponding node in the digital twin model; constructing a mathematical transformation relationship between sensor data and model parameters; and designing a data synchronization and update mechanism to ensure that real-time data can be reflected in the model parameters in a timely manner. Optimize the parameters of the digital twin model, including: using a Bayesian optimization algorithm to adjust the model parameters based on historical running data; using a genetic algorithm to optimize key physical parameters in the model so that the model output is consistent with the actual observation data; and establishing a model uncertainty quantification mechanism to calculate parameter sensitivity and confidence intervals. Establish a time-varying parameter model and integrate it into a digital twin model, including: constructing a performance degradation curve model of each key component over time; integrating an environmental factor influence model to consider the impact of temperature, humidity, load, etc. on equipment performance; and establishing an operating condition classification model to adapt parameters to the equipment characteristics under different operating conditions.

[0031] In this embodiment, the constructed multi-physics coupled digital twin model can more comprehensively describe the working state of the escalator and improve the model's prediction accuracy. Through the combination of point cloud reconstruction and parameter optimization, a high-precision mapping between the physical entity and the virtual model is achieved. The dynamic time-varying parameter model can accurately reflect the aging and wear process of the equipment, improving the accuracy of component life prediction. The deep integration technology of multi-source sensor data enables the model to respond to changes in equipment status in real time, greatly shortening the time for abnormal status identification. By establishing an environmental factor influence model, the system can adapt to different working environments, improving the model's versatility and adaptability.

[0032] In some possible embodiments of the present invention, the step of constructing a multi-level distributed intelligent sensing and monitoring network for escalators and elevators, and acquiring first monitoring data according to a preset first monitoring strategy, includes: The key monitoring points for escalators and elevators are identified and their layout is planned, including: identifying key failure points and high-risk areas of escalators and elevators based on Failure Mode and Effects Analysis (FMEA); determining the failure frequency and risk weight of each component of the escalator and elevator using heat map analysis technology; establishing a sensor optimization layout model to calculate the minimum number of sensors and the optimal installation location; and generating a sensor deployment map to clarify the installation coordinates, orientation, and monitoring range of various sensors. Based on the identification results and layout planning of key monitoring points, a multi-level sensor network architecture is constructed, including: deploying vibration sensors and acceleration sensors in the escalator drive system with a sampling frequency set to 10KHz; installing strain sensors and displacement sensors in the load-bearing structure and step system to monitor structural deformation and displacement; installing temperature sensors and thermal imagers in the electrical control cabinet and motor to monitor hotspot distribution; configuring sound sensors in key parts of the transmission system to collect acoustic feature data; and installing high-definition cameras at the escalator entrance and exit to monitor the step operation status and passenger flow in real time. Deploying edge computing devices includes: installing industrial-grade edge computing nodes near escalator control cabinets and configuring AI acceleration chips; establishing low-latency communication links between sensors and edge computing nodes, using the Time-Sensitive Networking (TSN) protocol; deploying a lightweight real-time operating system on the edge nodes to provide millisecond-level response capabilities; and configuring edge storage modules to achieve local data caching and disaster recovery functions. Develop an adaptive multi-level monitoring strategy, including: designing a basic monitoring mode and defining the default sampling frequency, accuracy, and triggering conditions for various sensors; establishing a dynamic sampling mechanism for load sensing, automatically adjusting the sampling frequency according to the operating load of the escalator and elevator; constructing an anomaly-triggered deep monitoring mode, automatically increasing the sampling accuracy and frequency of sensors in relevant areas when anomalies are detected; and developing an energy consumption balance algorithm to minimize the energy consumption of the sensor network while ensuring monitoring effectiveness. An edge-intelligent preprocessing mechanism is established, including: deploying time-domain feature extraction algorithms to calculate characteristic parameters such as root mean square value, peak factor, and impulse factor of vibration signals; applying frequency-domain analysis algorithms to perform fast Fourier transform on vibration and sound signals to extract spectral features; configuring an image preprocessing module to perform object detection and motion analysis on video streams; and implementing data compression and dimensionality reduction processing to reduce transmission bandwidth requirements. Establish a sensor data quality control mechanism, including: designing a sensor self-calibration algorithm to periodically calibrate the sensor's zero point and gain; deploying a data anomaly detection algorithm to identify and mark abnormal sensor data; implementing sensor redundancy design and adopting a multi-sensor cross-validation mechanism for key monitoring points; and establishing a sensor health status self-diagnosis function to monitor the sensor's working status in real time. Constructing an edge-cloud collaborative processing architecture includes: designing a task allocation mechanism between the edge and the cloud (such as cloud servers, digital twin management servers, etc.) and clarifying the execution boundaries of computing tasks; establishing a data hierarchical transmission strategy and determining transmission priority based on data importance; implementing an edge strategy adjustment mechanism based on cloud feedback and dynamically adjusting the monitoring parameters of the edge based on cloud analysis results; and constructing a network outage fault tolerance mechanism to ensure that the edge can still work independently in the event of a network interruption. According to the preset first monitoring strategy, the monitoring data acquisition and preprocessing are performed, including: activating the corresponding sensors to acquire data according to the preset first monitoring strategy; performing noise reduction, filtering and standardization on the acquired raw data; extracting time domain, frequency domain and time-frequency domain features to form a structured feature vector; performing preliminary anomaly detection on the preprocessed data and marking potential anomalies; and transmitting the processed first monitoring data to the digital twin management server.

[0033] The solution in this embodiment features an optimized sensor layout that significantly improves monitoring coverage while reducing the number of sensors compared to traditional solutions, thus substantially lowering hardware costs. Edge intelligent preprocessing technology reduces data transmission bandwidth requirements, alleviating network load and improving system stability. The adaptive multi-level monitoring strategy reduces overall energy consumption compared to fixed-frequency sampling schemes, extending the lifespan of sensing devices. Edge computing enables microsecond-level preliminary identification of abnormal features, detecting potential anomalies earlier than traditional cloud processing solutions, providing valuable time for emergency response. The sensor data quality control mechanism enhances the reliability of monitoring data, effectively reducing the risk of false alarms and missed alarms. The edge-cloud collaborative architecture ensures that the system maintains functional integrity even under network instability, significantly improving the robustness of the escalator monitoring system.

[0034] In some possible embodiments of the present invention, the step of bidirectionally synchronizing the digital twin model with the entity's operational status based on a first key indicator in the first monitoring data includes: Determine the mapping relationship between key indicators and synchronous parameters, including: establishing a mapping table between the first key indicators in the first monitoring data and the parameters of the digital twin model, and determining the model parameter set corresponding to each key indicator; assigning weight coefficients to each type of key indicator to reflect the degree of influence of different indicators on model state updates; constructing a constraint relationship matrix between model parameters to ensure that the consistency of physical laws is maintained when updating parameters; and setting the effective range and rate of change limit for parameter updates to prevent abnormal data from causing drastic fluctuations in model parameters. Establishing a distributed ledger data verification chain includes: deploying a blockchain network using a consortium blockchain architecture, setting up consensus nodes such as escalator maintenance and management parties, manufacturers, and owners; designing a timestamp-based data hash algorithm to generate a unique hash value for each batch of sensor data; constructing a Merkle tree structure for sensor data to optimize data verification efficiency; and developing a data consistency verification mechanism based on smart contracts to automatically execute the data validity verification process. Perform the difference calculation between real-time data and the digital twin model, including: receiving the first key indicator from the first monitoring data transmitted from the edge computing node; running the digital twin model to generate predicted values ​​based on the current parameter set; calculating various statistical indicators between the measured values ​​and the predicted values, including mean absolute error, root mean square error, relative error rate, etc.; and using the dynamic time warping algorithm to compare the measured and predicted time series data and identify differences in time series characteristics. Establish a dynamic threshold adaptive adjustment mechanism for digital twin models, including: constructing a normal error range model for each key indicator based on historical error distribution characteristics; integrating operational condition information to set differentiated threshold standards for different operating conditions; applying seasonal time series analysis to identify and adjust threshold parameters affected by periodic factors; and developing a cumulative error monitoring algorithm to detect long-term trend deviations. Based on the results of the difference calculation and the dynamic threshold adaptive adjustment mechanism, a self-optimization algorithm for the digital twin model is implemented, including: triggering the model parameter update process when the index difference exceeds the dynamic threshold; using the Bayesian parameter estimation method to update the posterior distribution of parameters based on the newly added observation data; applying the gradient descent optimization algorithm to adjust the model parameters to minimize the prediction error; and performing adaptive adjustment of the model structure to dynamically modify the model complexity according to the characteristics of the accumulated error. Establish an impact propagation model for digital twin models, including: analyzing the correlation and causal relationship between parameters and constructing a parameter impact propagation graph; calculating the chain effect of a single parameter change on other parameters based on a graph network propagation algorithm; designing a parameter update priority ranking mechanism to ensure that key parameters are updated first; and implementing a hierarchical parameter update strategy, updating basic parameters first and then adjusting advanced parameters based on the basic parameters. The process of implementing reverse guidance from the digital twin model to the physical entity includes: calculating the optimal operating parameters for the escalator and elevator based on the updated digital twin model; generating recommendations for adjusting operating parameters, including speed, acceleration, and load limits; verifying the rationality and safety of the parameter adjustment recommendations through a safety check mechanism; and sending the verified parameter adjustment instructions to the escalator and elevator control system. Construct a synchronization status assessment and recording mechanism, including: designing a synchronization quality evaluation index system to quantify the consistency level between the model and the entity; establishing a synchronization history database to save the time, parameters, and effects of each synchronization; developing a synchronization effect visualization tool to intuitively display the trend of the consistency between the model and the entity; and constructing a synchronization anomaly early warning mechanism to issue an alarm when the synchronization quality continues to decline.

[0035] In this embodiment, distributed ledger technology ensures the authenticity and immutability of data, enhances data credibility, and provides a high-quality data foundation for model updates; the dynamic threshold adaptive adjustment mechanism improves the system's accuracy in identifying abnormal states while reducing the false alarm rate; the model self-optimization algorithm achieves rapid parameter convergence, shortens model adjustment time, and significantly improves the real-time performance of the digital twin model; the introduction of the influence propagation model solves the coupling problem between parameters, ensures the consistency of simultaneous updates of multiple parameters, and reduces parameter conflicts; the reverse guidance function of the virtual model to the entity improves equipment operating efficiency, reduces energy consumption, and significantly improves the economic indicators of escalators and elevators; the overall synchronization mechanism effectively solves the problem of "cumulative deviation" in long-term operation, enabling the prediction accuracy of the digital twin model to remain stable even after long-term operation.

[0036] In some possible embodiments of the present invention, the step of combining the first key indicator in the first monitoring data to perform quantum-enhanced device status assessment and prediction, and obtaining the first analysis result, includes: A quantum feature mapping framework is constructed based on the first key indicator, including: designing a quantum feature mapping scheme suitable for escalator monitoring data to map the first key indicator in the first monitoring data to the quantum bit space; using an angle encoding method to convert the standardized continuous value features into quantum rotation angles; constructing a quantum amplitude encoder to achieve efficient encoding and information compression of high-dimensional features; and designing a quantum entanglement feature construction module to capture the nonlinear correlation between indicators. The architecture of a variable quantum neural network is constructed, including: designing quantum parameterization circuits, including rotation gates, entanglement gates, and measurement gates; constructing a multi-layer quantum convolution structure to achieve hierarchical abstraction of features; designing quantum pooling operations to reduce the dimensionality of intermediate representations and extract key features; and implementing a hybrid quantum-classical processing architecture to fuse quantum computing results with classical features. Implementing a quantum-classical hybrid optimization algorithm includes: designing a gradient descent-based quantum circuit parameter optimization method; introducing a noise-robust training strategy to improve the stability of the quantum model in real quantum environments; implementing parameter shift rules to reduce the gradient vanishing problem; and employing quantum ensemble learning techniques to combine the prediction results of multiple quantum models. The strategy for constructing quantum parallel computing includes: designing a multi-device feature parallel processing mechanism to process feature data from different device components simultaneously; realizing multi-timescale parallel analysis to process data at different time granularities synchronously; constructing a quantum multi-task learning framework to execute fault diagnosis, lifetime prediction, and performance evaluation tasks in parallel; and designing a quantum circuit parallel execution scheduler to optimize the allocation of computing resources. Establish a quantum-enhanced anomaly detection model, including: designing an anomaly detection algorithm based on the quantum kernel method to identify anomaly patterns in a high-dimensional feature space; implementing a quantum support vector machine to construct an optimal classification hyperplane in the quantum state space; constructing a quantum variational autoencoder to learn the latent representation of normal operating data for anomaly identification; and designing a quantum isolated forest algorithm to identify anomaly points through quantum random projection. Implementing quantum-enhanced fault diagnosis includes: constructing a quantum decision tree ensemble model to achieve high-precision fault classification; designing a quantum probabilistic reasoning framework to calculate the posterior probability distribution of fault causes; implementing a quantum Bayesian network to establish a fault causal relationship model; and constructing a quantum association rule mining algorithm to discover the correlations between faults. Develop quantum-enhanced lifetime prediction models, including: designing a quantum time series prediction framework to capture the performance degradation patterns of equipment; implementing the quantum Monte Carlo method to simulate possible future degradation paths through massive quantum parallel sampling; constructing a quantum risk assessment model to calculate the failure probability of equipment in different time windows; and designing a quantum Weibull distribution parameter estimator to accurately model the remaining lifetime distribution of equipment. Perform quantum uncertainty quantification and assessment, including: designing quantum probability distribution estimation methods to quantify the uncertainty of prediction results; realizing quantum integrated prediction range calculation to provide a reliable prediction range; constructing a quantum sensitivity analysis framework to identify the factors that have the greatest impact on prediction results; and designing interpretable quantum models to provide a physical interpretation of the prediction results. The system generates the first analysis results, including: integrating quantum anomaly detection, fault diagnosis, and lifetime prediction results to form a comprehensive assessment report; generating predicted residual service life (RULC) values ​​for components, including expected values ​​and confidence intervals; outputting health indices for escalator and elevator systems and key components to quantify the current state of the equipment; and providing potential failure risk analysis, including assessments of failure type, probability of occurrence, and severity.

[0037] In this embodiment, the quantum feature mapping framework effectively captures high-order nonlinear relationships between indicators, improving feature extraction capabilities compared to traditional methods and significantly enhancing the fault feature recognition rate. The variable quantum neural network architecture, through quantum state superposition and entanglement characteristics, achieves exponential feature space representation capabilities, enabling the model to handle more complex pattern recognition tasks. The quantum parallel computing strategy significantly improves computational efficiency, increasing model training and real-time inference speeds, meeting the needs of real-time monitoring of escalators and elevators. The quantum-enhanced anomaly detection model improves the anomaly detection rate while reducing the false alarm rate, greatly enhancing the reliability of the early warning system. The quantum-enhanced lifetime prediction model reduces prediction errors, making maintenance plans more accurate and effectively avoiding losses caused by premature or late maintenance. Quantum uncertainty quantification and assessment provide reliable risk assessment methods, reducing decision uncertainty. The overall technical solution improves the prediction accuracy of the escalator and elevator health management system in complex environments, significantly enhancing the system's reliability and practicality.

[0038] In some possible embodiments of the present invention, in the step of dynamically optimizing the first monitoring strategy using a deep reinforcement learning framework based on the first analysis result, the reward function of the reinforcement learning algorithm used is: ; in: S_severity∈[0,1] is the fault level quantization value based on QCNN output (0=no fault, 1=critical failure). η_accuracy=TP / (TP+FP+FN), which is the F1 score calculated from the confusion matrix of the detection results. TP is the number of samples that actually have faults and are correctly identified, FP is the number of samples that are actually without faults but are misclassified as faults, and FN is the number of samples that actually have faults but are not identified; η_min (minimum accuracy threshold) is the minimum acceptable value of the fault detection F1 score. T_threshold is the maximum allowable fault identification delay time, and T_detect is the total time from sensor data acquisition to completion of fault diagnosis. P_current is the real-time total power of the detection system, and P_max is the maximum power limit for safe operation of the system. δ=2ln(P_max / P_nominal), which is the energy efficiency ratio adjustment factor (P_nominal=rated power); λ = 1 / (1 + e^(-k*(T_detect-T_avg))), is the time decay factor (T_avg = historical average delay); k is a dynamic sensitivity adjustment factor; k controls the steepness of the slope of the Sigmoid function, determining the sensitivity of λ to the delay difference (T_detect−T_avg): when k>0, the larger the absolute value of the difference, the faster λ approaches 0 or 1; when k=0, the function degenerates to a fixed value of 0.5 (losing dynamic adjustment capability); k is an engineering parameter mapping the system's tolerance to delay fluctuations: high sensitivity mode (k=2.0): small delay deviations trigger strategy adjustments (suitable for critical scenarios such as hospitals and subways); low sensitivity mode (k=0.5): allows for larger delay fluctuations (suitable for non-emergency scenarios such as shopping malls); I(): Indicator function (takes 1 when η_accuracy<0.95, otherwise takes 0); The dynamic weighting coefficients satisfy the following: α is the fault level weight; β is the weighting coefficient of the time-sensitivity term, used to balance the optimization weight between detection delay and system power consumption; γ=0.5α, ζ=3β, and α+β=1.

[0039] In this embodiment, a product term of fault severity S_severity and detection accuracy η_accuracy is introduced, which allows the algorithm to obtain higher rewards when handling high-risk faults (e.g., when a gear breakage is detected (S=0.9) and the accuracy η=0.98, the contribution value of this term reaches 0.441), improving the response speed to critical faults. The original linear term is replaced by an exponentially decaying form e^(-δ⋅(P_current / P_max)), and a steep penalty is applied when the power consumption approaches P_max (e.g., when P_current=90%*P_max, the energy efficiency term value drops sharply from 0.1 to 0.02), forcing the algorithm to actively reduce the frequency before the power consumption exceeds the limit, which can reduce the peak power consumption under abnormal conditions. A saturation characteristic is constructed using the hyperbolic tangent function tanh(λ*(T_threshold-T_detect)): when T_detect<=T_threshold, the reward growth rate slows down to avoid over-optimization; when T_detect>T_threshold, a superlinear penalty is applied (e.g., the reward decreases by 60% when the delay exceeds the threshold by 10%). Combined with an adaptive factor λ, the time sensitivity under different operating conditions is automatically adjusted (e.g., the λ value is automatically increased by 32% during morning and evening peak hours); a hard threshold η_accuracy (η_min=0.95) is set, and a 3β-fold penalty is applied when the detection accuracy is insufficient, so that the algorithm prioritizes detection reliability.

[0040] In some possible embodiments of the present invention, k is a dynamic sensitivity adjustment factor; the value of k is calculated according to the following formula: ; Where K_base is the base sensitivity (default 0.8); S_max represents the maximum normalized threshold for the severity of the fault; When a severe fault is detected (S_severity>0.7), k is automatically increased to 1.1 to speed up the response.

[0041] In this embodiment, the k value is adaptively updated according to the operating status of the escalator. k is the core parameter that balances the sensitivity and stability of the system. Its dynamic optimization mechanism significantly improves the adaptive capability and energy efficiency ratio of the escalator health management system.

[0042] In some possible embodiments of the present invention, the step of generating a first maintenance plan based on the first analysis result includes: Based on the first analysis results, a fault risk matrix for elevator and escalator components is constructed. The fault risk matrix includes two dimensions: fault probability level and fault impact level. The fault probability level is determined by the remaining service life (RULC) value calculated by quantum enhanced equipment condition assessment and prediction, and the fault impact level is calculated by combining the equipment functional criticality and safety risk coefficient. Based on the fault risk matrix, a list of high-risk components and their associated parts with risk values ​​higher than the first preset risk value is extracted. Based on the component dependency graph in the digital twin model, the optimal maintenance component combination scheme is determined. The component dependency graph includes physical connection relationships, functional dependencies, and fault propagation paths. A maintenance simulation sub-model is established in the digital twin model, and the maintenance plan to be executed is implanted into the sub-model. The effectiveness and impact of the maintenance plan are verified through full-dimensional physical characteristic simulation. The full-dimensional physical characteristic simulation includes: simulation of dynamic parameter changes after component replacement, simulation of stress impact of maintenance operation on adjacent components, and prediction of overall system performance improvement. Based on the maintenance simulation results, a multi-objective optimization function is constructed, which simultaneously considers maintenance costs, equipment downtime, degree of safety risk elimination, and maintenance operation complexity, and solves to obtain the Pareto optimal maintenance scheme set. From the Pareto optimal maintenance solution set, the solution with the highest comprehensive score is selected as the first maintenance solution, and a digital maintenance guidance package is generated. The digital maintenance guidance package includes the following: a component priority list based on risk rating and operational dependence; a work path plan considering spatial constraints and operational sequence; a customized maintenance tool and material list determined according to the actual component status; and an immersive mixed reality (MR) visualization operation guide with 3D dynamic display of operation steps, key points and precautions. The first maintenance plan is analyzed for resource requirements using edge computing nodes. It is matched with the current availability of maintenance resources. When a resource gap is found, an automatic reservation process is triggered and the maintenance time window is updated.

[0043] In this embodiment, by simulating maintenance operations in a digital twin model beforehand, the effectiveness of maintenance plans can be improved, ineffective maintenance interventions can be avoided, and the rate of secondary repairs can be reduced. Immersive MR maintenance guidance based on component priority ranking and operation path planning optimizes the operation process for maintenance personnel, shortens the average maintenance time, and significantly reduces the processing time for complex faults. Through multi-objective optimization, maintenance costs are minimized while ensuring maintenance quality, thereby reducing the total maintenance cost. Resource demand analysis and matching through edge computing nodes enable precise allocation of maintenance resources, reduce resource idle rate, improve resource utilization efficiency, and avoid maintenance delays caused by insufficient resources. The immersive MR technology guidance solution digitizes and standardizes expert maintenance knowledge, making it easier for newly hired maintenance personnel to quickly master maintenance skills and solving the problem of difficult technology transfer in the industry.

[0044] In some possible embodiments of the present invention, the step of sending the first analysis result and the first maintenance plan to the privacy-preserving swarm intelligence learning platform includes: The first analysis results and the first maintenance plan for elevators and escalators are preprocessed, including: standardizing the data format, removing the organization identification information, adding timestamps and equipment type markers, and converting the diagnostic feature vectors and maintenance plan parameters into a standardized learning input format; A data privacy protection algorithm based on homomorphic encryption is used to encrypt the preprocessed first analysis result and the first maintenance scheme to generate a ciphertext data packet. The ciphertext data packet retains the computational properties of the original data in the encrypted state, but cannot be directly deciphered. The encrypted data packets are transmitted to the privacy-preserving swarm intelligence learning platform via a secure communication protocol, which includes end-to-end encryption, data integrity verification, and transmission traffic obfuscation. In the swarm intelligence learning platform, secure multi-party computation technology is used to perform joint computation on encrypted data from multiple escalator and elevator stations while maintaining data encryption. This includes: fault mode clustering analysis based on encrypted data to identify common fault features and unique fault modes; fault-maintenance scheme effect correlation analysis to evaluate the actual effect of different maintenance schemes on similar faults; and fault prediction model parameter update based on differential privacy to enhance the generalization ability of the prediction model. A blockchain-based cross-site maintenance knowledge-sharing network is constructed to ensure the protection of contributors' rights and the transparency of data use. In this network, each escalator / elevator station participates in collaborative verification as a node in the blockchain network; a contribution-based smart contract incentive mechanism is adopted to calculate contribution value based on the quality and quantity of data provided; and all data usage and model update operation logs are recorded to form a complete audit chain. Obtain privacy-protected knowledge feedback from swarm intelligence learning platforms, including: parameter optimization suggestions for local quantum-enhanced device condition assessment models; a novel fault feature fingerprint database validated at multiple sites; and maintenance solution effectiveness evaluation reports and optimization suggestions. The acquired knowledge feedback is integrated and adapted with the local digital twin model to update the local equipment status assessment model, fault prediction algorithm, and maintenance plan generation strategy, forming a closed-loop optimization.

[0045] In this embodiment, homomorphic encryption and secure multi-party computation techniques are used to achieve "usable but invisible" data, solving the data leakage risk in traditional collaborative learning. This allows participants to share knowledge without exposing the original data, meeting strict data privacy compliance requirements. Utilizing the massive data of the swarm intelligence learning platform, newly deployed escalators and elevators can directly inherit experiential knowledge, improving the early accuracy of the fault identification model and significantly shortening the model's cold start cycle. Through secure fusion of multi-site data, the accuracy of identifying low-frequency rare faults is improved, solving the problem that single-site data cannot effectively learn rare fault patterns. For escalators and elevators with different operating environments, load conditions, and usage frequencies, swarm intelligence learning can extract environmental adaptability rules, enhancing the robustness of the prediction model in changing environments and improving the accuracy of the fault prediction time window. Through cross-site maintenance effect analysis, the best maintenance practices are automatically identified, improving the first-time success rate of maintenance plans and further reducing average maintenance costs.

[0046] Please see Figure 2 Another embodiment of the present invention provides a digital twin-based escalator health management system for executing a digital twin-based escalator health management method, comprising: a digital twin management server, an edge computing node, and a swarm intelligence learning platform; The digital twin management server is configured as follows: A digital twin model of an escalator or elevator is constructed using point cloud reconstruction and deep integration technology of multi-source sensor data. Construct a multi-level distributed intelligent sensing and monitoring network for elevators and escalators, and acquire first monitoring data according to a preset first monitoring strategy; Based on the first key indicator in the first monitoring data, conduct two-way synchronization between the digital twin model and the physical operation status; Based on the first key indicators in the first monitoring data, a state assessment and prediction of the quantum-enhanced device was carried out to obtain the first analysis results; Based on the results of the first analysis, the first monitoring strategy is dynamically optimized using a deep reinforcement learning framework. A first maintenance plan is generated based on the first analysis results; The first analysis results and the first maintenance plan were sent to a swarm intelligence learning platform based on privacy computing.

[0047] It should be known that, Figure 2 The block diagram of the digital twin-based escalator health management system shown is for illustrative purposes only, and the number of modules shown does not limit the scope of protection of this invention. The digital twin-based escalator health management system provided in this embodiment can be used to execute various embodiments of the corresponding digital twin-based escalator health management method. For specific implementation details, please refer to the descriptions of the respective method embodiments, which will not be repeated here.

[0048] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0049] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0050] In the several embodiments provided in this application, it should be understood that the disclosed apparatus can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of the units described above is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical or other forms.

[0051] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0052] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0053] If the integrated units described above are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage device (CMD). Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned memory includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0054] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, which may include: flash drive, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc.

[0055] The embodiments of this application have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

[0056] While the present invention has been disclosed above, it is not limited thereto. Any person skilled in the art can easily conceive of variations or substitutions without departing from the spirit and scope of the present invention, and various modifications and alterations can be made, including combinations of the different functions and implementation steps described above, as well as software and hardware implementation methods, all of which are within the protection scope of the present invention.

Claims

1. A health management method for escalators and elevators based on digital twins, characterized in that, include: A digital twin model of an escalator or elevator is constructed using point cloud reconstruction and deep integration technology of multi-source sensor data. Construct a multi-level distributed intelligent sensing and monitoring network for elevators and escalators, and acquire first monitoring data according to a preset first monitoring strategy; Based on the first key indicator in the first monitoring data, conduct two-way synchronization between the digital twin model and the physical operation status; Based on the first key indicators in the first monitoring data, a state assessment and prediction of the quantum-enhanced device was carried out to obtain the first analysis results; Based on the results of the first analysis, the first monitoring strategy is dynamically optimized using a deep reinforcement learning framework. A first maintenance plan is generated based on the first analysis results; The first analysis results and the first maintenance plan were sent to a swarm intelligence learning platform based on privacy computing.

2. The method for health management of escalators and elevators based on digital twins according to claim 1, characterized in that, The steps for constructing a digital twin model of an escalator using point cloud reconstruction and deep integration of multi-source sensor data include: Collect three-dimensional spatial information of elevators and escalators; Preprocess the point cloud data in the acquired 3D spatial information; Geometric model reconstruction is performed based on preprocessed point cloud data to obtain the geometric model of escalator; Acquire multi-source sensor data and perform preprocessing; By combining the preprocessed multi-source sensor data, a multi-physics coupling model is constructed to obtain the physical field model of the escalator; By fusing the geometric model of the escalator with the physical field model of the escalator, a digital twin model of the escalator is obtained. To achieve parameter mapping between sensor data and digital twin models; Optimize digital twin model parameters; Establish a time-varying parameter model and integrate it into the digital twin model.

3. The method for health management of escalators and elevators based on digital twins according to claim 2, characterized in that, The step of constructing a multi-level distributed intelligent sensing and monitoring network for escalators and elevators, and acquiring first monitoring data according to a preset first monitoring strategy, includes: Identify and plan the layout of key monitoring points for elevators and escalators; A multi-level sensor network architecture is constructed based on the identification results and layout planning of key monitoring points; Deploy edge computing devices; Develop an adaptive multi-level monitoring strategy; Establish an edge intelligence preprocessing mechanism; Establish a sensor data quality control mechanism; Build an edge-cloud collaborative processing architecture; According to the preset first monitoring strategy, the monitoring data is collected and preprocessed, and the processed first monitoring data is transmitted to the digital twin management server.

4. The method for health management of escalators and elevators based on digital twins according to claim 3, characterized in that, The step of bidirectionally synchronizing the digital twin model with the entity's operational status based on the first key indicator in the first monitoring data includes: Determine the mapping relationship between key indicators and synchronization parameters; Establish a distributed ledger data verification chain; Perform difference calculations between real-time data and the digital twin model; Establish a dynamic threshold adaptive adjustment mechanism for digital twin models; Based on the results of the difference calculation and the dynamic threshold adaptive adjustment mechanism, a self-optimization algorithm for the digital twin model is implemented. Establish an impact propagation model for digital twin models; Execute reverse guidance from the digital twin model to the entity; Establish a synchronous state assessment and recording mechanism.

5. The method for health management of escalators and elevators based on digital twins according to claim 4, characterized in that, The step of combining the first key indicator from the first monitoring data to perform state assessment and prediction of the quantum-enhanced device and obtain the first analysis result includes: A quantum feature mapping framework is constructed by combining the first key indicator; Build a variable quantum neural network architecture; Implement a quantum-classical hybrid optimization algorithm; Strategies for constructing quantum parallel computing; Establish a quantum-enhanced anomaly detection model; Implement quantum-enhanced fault diagnosis; Develop quantum-enhanced lifetime prediction models; Perform quantum uncertainty quantization and assessment, and generate the first analysis results.

6. The method for health management of escalators and elevators based on digital twins according to claim 5, characterized in that, In the step of dynamically optimizing the first monitoring strategy using a deep reinforcement learning framework based on the first analysis results, the reward function of the reinforcement learning algorithm used is: 。 7. The method for health management of escalators and elevators based on digital twins according to claim 6, characterized in that, k is a dynamic sensitivity adjustment factor; the value of k is calculated according to the following formula: 。 8. The method for health management of escalators and elevators based on digital twins according to claim 7, characterized in that, The step of generating the first maintenance plan based on the first analysis result includes: A fault risk matrix for elevator and escalator components was constructed based on the results of the first analysis. Based on the fault risk matrix, a list of high-risk components and their associated parts with risk values ​​higher than the first preset risk value are extracted, and the optimal maintenance component combination scheme is determined based on the component dependency graph in the digital twin model. A maintenance simulation sub-model is established in the digital twin model, and the maintenance plan to be executed is implanted into the sub-model. The effectiveness and impact of the maintenance plan are verified through full-dimensional physical characteristic simulation. Based on the maintenance simulation results, a multi-objective optimization function is constructed, which simultaneously considers maintenance costs, equipment downtime, degree of safety risk elimination, and maintenance operation complexity, and solves to obtain the Pareto optimal maintenance scheme set. From the Pareto optimal maintenance solution set, the solution with the highest comprehensive score is selected as the first maintenance solution, and a digital maintenance guidance package is generated; The first maintenance plan is analyzed for resource requirements using edge computing nodes. It is matched with the current availability of maintenance resources. When a resource gap is found, an automatic reservation process is triggered and the maintenance time window is updated.

9. The method for health management of escalators and elevators based on digital twins according to claim 8, characterized in that, The step of sending the first analysis result and the first maintenance plan to the privacy-preserving swarm intelligence learning platform includes: Data preprocessing is performed on the first analysis results and the first maintenance plan for the elevators and escalators; A data privacy protection algorithm based on homomorphic encryption is used to encrypt the preprocessed first analysis result and the first maintenance scheme to generate ciphertext data packets; The encrypted data packets are transmitted to a privacy-preserving swarm intelligence learning platform via a secure communication protocol. In the swarm intelligence learning platform, secure multi-party computation technology is used to perform joint computation on encrypted data from multiple escalator and elevator stations while maintaining data encryption. Build a blockchain-based cross-site maintenance knowledge sharing network to ensure the protection of contributors' rights and transparency in data use; Obtain privacy-protected knowledge feedback from swarm intelligence learning platforms; The acquired knowledge feedback is integrated and adapted with the local digital twin model to update the local equipment status assessment model, fault prediction algorithm, and maintenance plan generation strategy, forming a closed-loop optimization.

10. A digital twin-based escalator health management system, used to execute the digital twin-based escalator health management method as described in any one of claims 1 to 9, characterized in that, include: Digital twin management server, edge computing nodes, and swarm intelligence learning platform; The digital twin management server is configured as follows: A digital twin model of an escalator or elevator is constructed using point cloud reconstruction and deep integration technology of multi-source sensor data. Construct a multi-level distributed intelligent sensing and monitoring network for elevators and escalators, and acquire first monitoring data according to a preset first monitoring strategy; Based on the first key indicator in the first monitoring data, conduct two-way synchronization between the digital twin model and the physical operation status; Based on the first key indicators in the first monitoring data, a state assessment and prediction of the quantum-enhanced device was carried out to obtain the first analysis results; Based on the results of the first analysis, the first monitoring strategy is dynamically optimized using a deep reinforcement learning framework. A first maintenance plan is generated based on the first analysis results; The first analysis results and the first maintenance plan were sent to a swarm intelligence learning platform based on privacy computing.

Images