Construction method and system for elevator part performance model and storage medium

Abstract

The invention relates to the technical field of elevator component performance analysis, in particular to a construction method and system for an elevator component performance model and a storage medium. The method comprises the following steps that elevator use records are collected, flow frequency spectrum modeling is carried out to generate start-stop frequency spectrum features and elevator use mode features, then part response analysis is carried out on the use mode features to obtain dynamic characteristic indexes, the independent performance of parts is calculated, and the independent performance of the parts is calculated. And an elevator interaction network is reconstructed according to the independence performance and the use mode characteristics, energy consumption distribution of the elevator is calculated, overall parameters of the elevator are obtained, framework simulation is conducted, framework movement performance loss is calculated, part position projection is conducted on a simulation framework based on part independence performance, interaction performance is evaluated, and therefore an elevator part performance model is constructed. According to the elevator part performance model construction method, elevator start-stop frequency spectrum characteristics and energy consumption distribution can be accurately described, and the response characteristics and interaction performance of the parts can be dynamically evaluated.

Description

Technical Field

[0001] The present invention relates to the technical field of elevator component performance analysis, and in particular to a method, system and storage medium for constructing an elevator component performance model. Background Art

[0002] Traditional elevator maintenance relies heavily on experience and regular inspections, making it difficult to achieve real-time monitoring and efficient management, and unable to effectively respond to sudden failures, which reduces the safety and efficiency of elevator use. To improve the intelligence level and maintenance efficiency of elevators, there is an urgent need for a systematic method to build an elevator component performance model to achieve dynamic monitoring and performance analysis of each component. Although existing elevator performance evaluation technologies have improved the safety and efficiency of elevator operation to a certain extent, they often lack in-depth analysis of the actual use of elevators. A single performance test is difficult to fully reflect the performance of elevators in different usage modes, especially in high-frequency use or complex environments. The performance changes of elevator components are more obvious. Traditional methods cannot capture these changes in a timely manner, resulting in wrong decisions or improper maintenance, increasing operating costs and accident risks. Therefore, it is necessary to establish a more accurate and scientific elevator performance model through comprehensive multi-dimensional data analysis. Summary of the Invention

[0003] Based on this, it is necessary to provide a method, system and storage medium for constructing an elevator component performance model to solve at least one of the above technical problems.

[0004] To achieve the above-mentioned purpose, a method for constructing an elevator component performance model includes the following steps:

[0005] Step S1: Collect elevator usage records; perform flow spectrum modeling on the elevator usage records to generate start-stop spectrum features; perform flow-frequency coordinated fitting to obtain elevator usage pattern features;

[0006] Step S2: Analyze component responses based on elevator usage pattern characteristics to obtain dynamic characteristic indices; and infer component independent performance based on the dynamic characteristic indices.

[0007] Step S3: Reconstructing the elevator interaction network based on the independent performance of the components and the elevator usage pattern characteristics; estimating the elevator energy consumption distribution based on the elevator usage pattern characteristics and the elevator interaction network;

[0008] Step S4: obtaining overall elevator parameters; performing elevator frame simulation on the overall elevator parameters; and calculating frame movement performance loss based on the simulated elevator frame;

[0009] Step S5: Project component positions on the simulated elevator frame based on component independent performance, and evaluate component interaction performance based on component position projection, component independent performance, and elevator energy consumption distribution; construct an elevator component performance model based on component independent performance, component interaction performance, and frame movement performance loss.

[0010] This paper achieves multi-dimensional analysis of elevator operating status by collecting elevator usage records. Flow spectrum modeling provides a basis for identifying elevator start and stop characteristics. The results of flow-frequency synergistic fitting reveal the characteristics of elevator usage patterns, laying a scientific foundation for component performance analysis. The acquisition of dynamic characteristic indicators provides a deep understanding of the response behavior of each component during operation. The calculation of independent component performance provides data support for elevator maintenance and upgrades. Reconstructing the elevator interaction network enhances control over the overall operating status of the elevator system. The energy consumption distribution calculated based on elevator usage pattern characteristics provides a practical basis for energy-saving improvements. The acquisition of overall parameters combined with framework simulation effectively analyzes the performance of the elevator framework, helping to identify and reduce energy loss during operation. Component position projection clearly indicates the relative position of each component in the whole, providing visual data for interactive performance evaluation. The constructed elevator component performance model accurately reflects the performance of each component in actual operation, providing a reliable reference for subsequent fault prediction and performance optimization. Overall, it improves the operating efficiency and safety of the elevator and provides strong support for the intelligent management of the elevator system.

[0011] Preferably, step S1 includes the following steps:

[0012] Step S11: Collecting elevator usage records; extracting elevator usage flow in the elevator usage records;

[0013] Step S12: Divide the elevator usage flow into time windows to obtain the time-divided usage flow; infer the flow periodicity pattern based on the time-divided usage flow;

[0014] Step S13: Perform outlier detection on the traffic periodic pattern, set the outlier threshold to 2.5 times the standard deviation, remove outliers, and generate a purified traffic pattern; extract historical start and stop frequency data from the elevator usage record;

[0015] Step S14: performing statistical filtering on the historical start-stop frequencies and performing frequency domain conversion to obtain start-stop spectrum characteristics;

[0016] Step S15: Identify the peak of the start-stop spectrum characteristics and mark it as a key frequency feature point; perform flow spectrum collaborative analysis on the purification flow pattern and the key frequency feature point to obtain flow-frequency correlation data;

[0017] Step S16: extracting collaborative variable distribution data based on the flow-frequency correlation data, and performing operating condition fitting based on the collaborative variable distribution data to obtain elevator usage pattern characteristics.

[0018] The present invention realizes accurate estimation of elevator usage by collecting elevator usage records and extracting flow data. Time window segmentation effectively identifies usage flow in different time periods. The inference of flow periodic patterns supports scientific analysis of elevator usage behavior. Outlier detection and elimination improve the accuracy and reliability of data. The purified flow pattern provides a clearer basis for subsequent analysis. The extraction of historical start-stop frequency data provides an important basis for understanding the start-stop characteristics of elevators. The application of statistical filtering processing and frequency domain conversion technology effectively extracts the start-stop spectrum characteristics. Peak recognition and marking of key frequency feature points provide important references for analysis. The realization of collaborative analysis of flow spectrum increases the in-depth understanding of the relationship between flow and frequency. The extraction of collaborative variable distribution data provides a comprehensive evaluation of elevator performance under different working conditions. The elevator usage pattern characteristics obtained by working condition fitting provide strong support for further system optimization and performance improvement, and the overall level of intelligence and scientificity of elevator monitoring and management are improved.

[0019] Preferably, step S2 includes the following steps:

[0020] Step S21: performing component load decomposition on the elevator usage pattern characteristics, and querying component response characteristics in the elevator usage pattern characteristics based on the component load data;

[0021] Step S22: identifying the component transfer function based on the component response characteristics and the component load data; performing pole-zero analysis on the component transfer function, and performing dynamic characteristic index mapping;

[0022] Step S23: determining the stability boundary based on the dynamic characteristic index, and dividing the component stability range according to the stability boundary;

[0023] Step S24: identifying key operating parameters of the dynamic characteristic index according to the component stability range to obtain the key operating parameters of the component; performing performance curve fitting on the key operating parameters of the component to generate independent performance of the component.

[0024] By decomposing the component loads based on the elevator usage pattern characteristics, the present invention can deeply analyze the load conditions of various elevator components in actual operation, providing important data basis for identifying component response characteristics. Based on the combination of component response characteristics and load data, it is helpful to accurately identify the transfer function of the component, thereby reflecting its dynamic characteristics. The implementation of pole-zero analysis lays the foundation for an in-depth understanding of dynamic performance. The mapping of dynamic characteristic indicators can provide intuitive expression for component performance evaluation. The determination of stability boundaries ensures a comprehensive analysis of the stability of elevator components under different working conditions. The effectively divided component stability range provides guidance for subsequent performance optimization. The identification of key operating parameters further strengthens the grasp of component operating status. The independent component performance data generated by performance curve fitting provides a reliable quantitative basis for elevator maintenance and upgrades, which improves the accuracy and practicality of the elevator performance model as a whole and promotes the intelligent development and management efficiency improvement of the elevator industry.

[0025] Preferably, the elevator framework simulation of the overall elevator parameters in step S4 includes:

[0026] Expand the characteristic dimensions of the overall elevator parameters to obtain a hierarchical set of parameters;

[0027] Perform constraint coupling integration on parameter hierarchical sets to generate parameter correlation data;

[0028] Perform frame mechanical load projection on the parameter association data to obtain the frame load distribution, where the load types include static load, which ranges from 1.0 to 1.5 times the rated load, dynamic load, which ranges from 0.2 to 0.5 times the rated load, and impact load, which ranges from 0.1 to 0.3 times the rated load;

[0029] Map the component rigidity requirements according to the frame load distribution to obtain the rigidity requirements of the frame components;

[0030] Identify the main load-bearing component parameters of the overall elevator parameters;

[0031] Determine the geometric characteristics of the load-bearing components based on the parameters of the main load-bearing components, and perform stress distribution analysis based on the geometric characteristics of the load-bearing components to obtain load-bearing stress distribution data;

[0032] Perform three-dimensional simulation on the parameters of the main load-bearing components, and perform stress deformation simulation on the simulated load-bearing components based on the load-bearing stress distribution data to generate simulated deformation load-bearing components;

[0033] Identify the outer end contact surface component parameters of the overall elevator parameters;

[0034] determining an outer end contact area of the outer end contact surface component parameters, and selecting protruding component parameters of the outer end contact surface component parameters based on the outer end contact area;

[0035] Elevator component morphology simulation is performed based on the simulated deformation load-bearing components and protruding component parameters to obtain the reconstructed elevator component morphology;

[0036] According to the rigidity requirements of the frame components, the reconstructed elevator component morphology is topologically matched to obtain matching elevator components;

[0037] Elevator frame simulation is performed based on matching elevator components, where the number of nodes in the simulation model is 1000-10000 and the simulation time step is 0.01-0.1 second.

[0038] The present invention expands the characteristic dimensions of the overall parameters of the elevator, and the resulting parameter hierarchical set provides comprehensive data support for subsequent analysis. The constraint coupling integration realizes the efficient expression of the relationship between parameters. The generated parameter correlation data provides a solid foundation for frame mechanics analysis. The application of frame mechanics load projection makes the influence of different types of loads systematically considered. The setting of static load, dynamic load and impact load ensures the effectiveness of the model under various working conditions. The mapping of component rigidity requirements provides a scientific basis for the design of elevator components. The identification of the main load-bearing component parameters ensures the pertinence of the analysis. The determination of geometric features and stress distribution analysis provide detailed data for the performance evaluation of load-bearing components. The three-dimensional simulation and force deformation simulation accurately The performance of components under actual working conditions is reproduced. The identification of the external contact surface component parameters and the determination of the external contact area help to screen out key components. The parameters of these prominent components can better reflect the actual working status in the simulation. The realization of reconstructing the elevator component morphology provides an intuitive model for optimized design. The topological matching of the rigidity requirements of the frame components to the reconstructed morphology ensures the rationality of the elevator components in the overall design. The number of nodes in the simulation model is set between 1000 and 10000 to improve the accuracy of the calculation. The setting of the time step provides the necessary meticulousness for the simulation of dynamic behavior, which improves the overall design efficiency and performance optimization of the elevator system, and provides important theoretical support and practical guidance for the innovative development of future elevator technology.

[0039] Preferably, the calculation of the frame movement performance loss based on the simulated elevator frame in step S4 includes:

[0040] Expand the structural freedom of the simulated elevator frame to obtain the frame freedom data;

[0041] Calibrate the elevator frame motion mode based on the frame degree of freedom data;

[0042] Perform force mapping transformation on the simulated elevator frame according to the elevator frame motion mode to generate a frame motion benchmark;

[0043] Extract the contact surface energy distribution in the frame motion benchmark and infer the frame contact area based on the contact surface energy distribution;

[0044] Perform dynamic fitting of friction force on the frame contact area to generate frame friction force data;

[0045] Perform power consumption reduction on the frame friction matrix to obtain the frame friction response data;

[0046] Perform support component moment distribution on the frame friction response data to obtain the frame support force data;

[0047] Mapping frame resistance characteristics based on frame support force data;

[0048] The performance attenuation is calculated according to the frame resistance characteristics, and the attenuation effect is accumulated based on the performance attenuation data to generate the frame movement performance loss.

[0049] By expanding the structural degrees of freedom of a simulated elevator frame, the present invention can identify the frame's motion capabilities and limitations, providing the necessary data foundation for dynamic performance evaluation. Calibration of the frame's degrees of freedom data facilitates accurate identification of the elevator frame's motion pattern. Implementing force mapping transformation enables a comprehensive analysis of the frame's stress state during motion. The generated frame motion benchmark lays the foundation for subsequent energy distribution extraction. Extracting the contact surface energy distribution helps identify energy loss in the frame's contact area, providing a basis for dynamic fitting of friction forces. The generation of frame friction force data can provide a detailed reflection of the friction loss during elevator operation. Power consumption reduction and the aggregated query of the friction force matrix provide quantitative data for overall performance analysis. Support component torque distribution ensures the rationality and balance of stress on different components. Mapping the frame support force data clearly presents the resistance characteristics of the frame during motion. Performance attenuation calculation based on these resistance characteristics quantifies changes in frame performance. Accumulating attenuation effects provides data support for future design optimization. This overall approach enhances understanding of elevator frame mobility performance and provides important technical guidance for reducing energy consumption and improving operational efficiency.

[0050] Preferably, step S5 includes the following steps:

[0051] Step S51: splitting the elevator component units of the simulated elevator frame; performing component unit matching on the independent performance of the components based on the elevator component units to obtain a matching relationship;

[0052] Step S52: projecting the independent performance of the component onto the simulated elevator frame based on the matching relationship to obtain the component position projection;

[0053] Step S53: Calculate component energy consumption based on component position projection and elevator energy consumption distribution, and derive component network performance based on component energy consumption data and overall elevator parameters; evaluate component interaction performance based on component network performance and component independent performance;

[0054] Step S54: performing functional block splitting on the independent performance data of the component to obtain component functional blocks; performing energy efficiency response fitting on the component functional blocks to obtain component energy efficiency distribution;

[0055] Step S55: performing global correlation aggregation on the component energy efficiency distribution data to obtain component performance data; performing output performance integration based on the component performance data and component interaction performance to obtain integrated output performance;

[0056] Step S56: Constructing an elevator component performance model based on the integrated output performance and the frame movement performance loss.

[0057] The present invention makes component analysis more detailed by splitting the elevator component units of the simulated elevator frame, and provides a reasonable corresponding basis for the independent performance of the components based on the matching relationship of the component units. The projection of the independent performance of the components into the simulation framework can clearly locate the role of each component in the overall structure. The realization of component position projection lays the foundation for the calculation of component energy consumption, and can quantify the energy consumption of each component during operation according to the energy consumption distribution, providing data support for the optimization of the overall performance of the elevator. The evaluation of component network performance helps to understand the interaction of each component in the operation of the elevator. The splitting of functional blocks makes the component performance evaluation more systematic. The implementation of energy efficiency response fitting can reveal the energy efficiency performance of the components under different working conditions. The correlation aggregation of component energy efficiency distribution data ensures a comprehensive evaluation of the comprehensive performance of the components. The generation of integrated output performance provides a key reference for the optimal design of the elevator system. By combining the integrated output performance with the frame movement performance loss, a more accurate elevator component performance model can be constructed, providing scientific guidance for the performance improvement and resource optimization of the elevator system.

[0058] The present invention further provides a system for constructing an elevator component performance model, which is used to execute the method for constructing an elevator component performance model as described above. The system for constructing an elevator component performance model includes:

[0059] The spectrum modeling module is used to collect elevator usage records; perform flow spectrum modeling on the elevator usage records to generate start-stop spectrum characteristics; and perform flow-frequency co-fitting to obtain elevator usage pattern characteristics;

[0060] Dynamic analysis module, used to analyze component responses based on elevator usage pattern characteristics to obtain dynamic characteristic indicators; and to infer component independent performance based on dynamic characteristic indicators;

[0061] The interactive network module is used to reconstruct the elevator interactive network based on the independent performance of components and the characteristics of elevator usage patterns; and to estimate the elevator energy consumption distribution based on the elevator usage pattern characteristics and the elevator interactive network;

[0062] The frame simulation module is used to obtain the overall parameters of the elevator; simulate the elevator frame based on the overall parameters of the elevator; and calculate the frame movement performance loss based on the simulated elevator frame;

[0063] The model building module is used to project the component positions of the simulated elevator frame based on the independent performance of the components, and evaluate the component interaction performance based on the component position projection, component independent performance and elevator energy consumption distribution; and to build an elevator component performance model based on the component independent performance, component interaction performance and frame movement performance loss.

[0064] This invention achieves in-depth exploration of elevator operating characteristics through the collection and analysis of elevator usage records. The generated start-stop spectrum characteristics provide data support for understanding elevator usage patterns. The results of flow-frequency synergistic fitting reveal the elevator's operating patterns, helping to optimize elevator operation strategies. The component response analysis capability of the dynamic analysis module improves understanding of the operating status of each elevator component. The calculation of dynamic characteristic indicators makes component performance evaluation more scientific and systematic. The construction of the interactive network module provides a modeling basis for the interaction between various components of the elevator system. The effectively calculated elevator energy consumption distribution supports energy-saving renovation and optimization design. The frame simulation module evaluates the operating performance of the elevator frame through the acquisition and simulation analysis of overall parameters. The accurately calculated frame movement performance loss provides a basis for elevator operation and maintenance. The model construction module constructs an elevator component performance model by projecting component positions and evaluating interaction performance, covering the mutual influence and comprehensive performance of each component. This provides comprehensive data support and decision-making basis for the intelligent management and optimization of the elevator system, overall improving the efficiency and stability of elevator operation and promoting the continuous improvement and development of the elevator system.

[0065] The present invention further provides a computer-readable storage medium storing a computer program, wherein when the computer program is executed, the method for constructing an elevator component performance model as described in any one of the above items is implemented.

[0066] The present invention realizes the efficient execution of the elevator component performance model construction method by storing computer programs. The executable nature of the program ensures the automation and standardization of the model construction process, improves the operation efficiency and accuracy, and the stored program can flexibly adapt to the characteristics of different elevator systems to realize personalized parameter setting and analysis. With the help of the powerful computing power of the computer system, large-scale data processing and analysis can be carried out, thereby improving the reliability and timeliness of elevator usage data. The modular design of the program makes the connection between various functional modules smoother, facilitating subsequent maintenance and upgrading. The implementation in computer-readable form also supports the compatibility of multiple platforms and devices, improves the application scope and convenience, effectively meets the needs of intelligent elevator management and optimization, and provides a theoretical and practical basis for the efficient operation and sustainable development of the elevator system. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] Figure 1 A schematic flow chart of steps for a method for constructing an elevator component performance model;

[0068] Figure 2 Detailed implementation flow chart of step S2;

[0069] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings and in conjunction with the embodiments. DETAILED DESCRIPTION

[0070] The following is a clear and complete description of the technical method of the present invention in conjunction with the accompanying drawings. It is obvious that the embodiments described are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making any creative efforts are within the scope of protection of the present invention.

[0071] In addition, the accompanying drawings are merely schematic illustrations of the present invention and are not necessarily drawn to scale. Identical reference numerals in the figures denote identical or similar parts, and thus repetitive descriptions thereof will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities that do not necessarily correspond to physically or logically separate entities. These functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor and / or microcontroller approaches.

[0072] It should be understood that although the terms "first," "second," and the like may be used herein to describe various elements, these elements should not be limited by these terms. These terms are used solely to distinguish one element from another. For example, a first element may be referred to as a second element, and similarly, a second element may be referred to as a first element, without departing from the scope of the exemplary embodiments. The term "and / or" as used herein includes any and all combinations of one or more of the listed associated items.

[0073] To achieve this, please refer to Figures 1 to 2 , a method for constructing an elevator component performance model, comprising the following steps:

[0074] Step S1: Collect elevator usage records; perform flow spectrum modeling on the elevator usage records to generate start-stop spectrum features; perform flow-frequency coordinated fitting to obtain elevator usage pattern features;

[0075] Step S2: Analyze component responses based on elevator usage pattern characteristics to obtain dynamic characteristic indices; and infer component independent performance based on the dynamic characteristic indices.

[0076] Step S3: Reconstructing the elevator interaction network based on the independent performance of the components and the elevator usage pattern characteristics; estimating the elevator energy consumption distribution based on the elevator usage pattern characteristics and the elevator interaction network;

[0077] Step S4: obtaining overall elevator parameters; performing elevator frame simulation on the overall elevator parameters; and calculating frame movement performance loss based on the simulated elevator frame;

[0078] Step S5: Project component positions on the simulated elevator frame based on component independent performance, and evaluate component interaction performance based on component position projection, component independent performance, and elevator energy consumption distribution; construct an elevator component performance model based on component independent performance, component interaction performance, and frame movement performance loss.

[0079] This paper achieves multi-dimensional analysis of elevator operating status by collecting elevator usage records. Flow spectrum modeling provides a basis for identifying elevator start and stop characteristics. The results of flow-frequency synergistic fitting reveal the characteristics of elevator usage patterns, laying a scientific foundation for component performance analysis. The acquisition of dynamic characteristic indicators provides a deep understanding of the response behavior of each component during operation. The calculation of independent component performance provides data support for elevator maintenance and upgrades. Reconstructing the elevator interaction network enhances control over the overall operating status of the elevator system. The energy consumption distribution calculated based on elevator usage pattern characteristics provides a practical basis for energy-saving improvements. The acquisition of overall parameters combined with framework simulation effectively analyzes the performance of the elevator framework, helping to identify and reduce energy loss during operation. Component position projection clearly indicates the relative position of each component in the whole, providing visual data for interactive performance evaluation. The constructed elevator component performance model accurately reflects the performance of each component in actual operation, providing a reliable reference for subsequent fault prediction and performance optimization. Overall, it improves the operating efficiency and safety of the elevator and provides strong support for the intelligent management of the elevator system.

[0080] In an embodiment of the present invention, the method for constructing an elevator component performance model includes the following steps:

[0081] Step S1: Collect elevator usage records; perform flow spectrum modeling on the elevator usage records to generate start-stop spectrum features; perform flow-frequency coordinated fitting to obtain elevator usage pattern features;

[0082] In this embodiment, when collecting elevator usage records, a data collection device is set in the core unit of the elevator control system. The collection device includes a high-precision time series recording chip, a data cache module, and a data transmission module. The high-precision time series recording chip performs millisecond-level timestamp calibration for each elevator start and stop, the data cache module performs short-term storage of continuous start and stop data, and the data transmission module transmits the cached data to the central storage unit at fixed time intervals. All recorded data includes but is not limited to elevator start and stop time, elevator running floors, elevator load weight, elevator door opening time, elevator door closing time, and elevator running direction. All data are stored in a standardized format and stored in a database according to time series. At the same time, traffic spectrum modeling is performed on the collected data using short-time Fourier transform (STFT). The short-time Fourier transform (SFT) method is used to decompose the time series of the elevator operating status to obtain the start-stop signals of different frequency components. The time-frequency resolution of the short-time Fourier transform is improved by window function optimization to obtain high-resolution start-stop spectrum features. The start-stop spectrum features are then subjected to flow-frequency co-fitting. First, a Poisson flow model of the elevator start-stop events is established, and the elevator start-stop sequence is mapped into a time window. The start-stop frequency per unit time is calculated. Then, the nonlinear regression method is used to co-optimize the start-stop spectrum features and the flow model to obtain the elevator usage pattern features. The elevator usage pattern features include different categories such as high-frequency start-stop mode, low-frequency start-stop mode, load-sensitive start-stop mode, and inertia start-stop mode. The parameters of each mode are quantified and stored. The hierarchical clustering algorithm is used to classify and organize the elevator usage pattern features, and finally an elevator usage pattern feature library is obtained.

[0083] Step S2: Analyze component responses based on elevator usage pattern characteristics to obtain dynamic characteristic indices; and infer component independent performance based on the dynamic characteristic indices.

[0084] In this embodiment, when analyzing component responses based on elevator usage pattern characteristics, time series data for each pattern in the elevator usage pattern feature library is first extracted and a discrete wavelet transform (DWT) is performed on the data to extract response signals of key elevator components (traction machine, guide rails, car, door system, speed governor, etc.) under each pattern. Time and frequency domain features are extracted from the response signals. The Hilbert-Huang Transform (HHT) method is used to perform envelope analysis on the non-stationary signals to obtain the instantaneous amplitude, instantaneous frequency, and their changing trends of each component. Based on the extracted dynamic response features, a support vector regression (SVR) algorithm is used to establish a mapping relationship between the elevator usage pattern characteristics and the dynamic characteristics of the components. The dynamic characteristic indicators of the components under each pattern are predicted and stored. Subsequently, the independent performance of the components is estimated based on the dynamic characteristic indicators. A time-series recurrent neural network (RNN) is used to predict the independent performance of the components. A recursive neural network (RNN) is used to model the time series of component dynamic characteristic indicators. During training, a long short-term memory (LSTM) network is used for parameter optimization to calculate independent performance parameters such as component operation stability, load adaptability, and fatigue loss rate.

[0085] Step S3: Reconstructing the elevator interaction network based on the independent performance of the components and the elevator usage pattern characteristics; estimating the elevator energy consumption distribution based on the elevator usage pattern characteristics and the elevator interaction network;

[0086] In this embodiment, when reconstructing an elevator interaction network based on the independent performance of components and elevator usage pattern characteristics, a component-pattern adjacency matrix is first constructed. The rows of the matrix represent different elevator components, and the columns represent different usage patterns. The elements in the matrix are the independent performance scores of the components in the respective modes. A graph convolutional neural network (GCN) is used to extract features from the component-pattern adjacency matrix to obtain a high-dimensional representation vector of the component interaction relationships. The interaction strength between components is calculated using cosine similarity to generate an elevator interaction network topology graph. Subsequently, the elevator energy consumption distribution is inferred based on the elevator usage pattern characteristics and the elevator interaction network. First, centrality indicators of different components in the interaction network are calculated based on the topology graph, including degree centrality, betweenness centrality, and eigenvector centrality. Then, a Markov decision process (MDP) is used to model the elevator energy consumption distribution state transition model. Bayesian update is performed on the energy consumption under different states, ultimately obtaining a component-level energy consumption distribution matrix.

[0087] Step S4: obtaining overall elevator parameters; performing elevator frame simulation on the overall elevator parameters; and calculating frame movement performance loss based on the simulated elevator frame;

[0088] In this embodiment, when obtaining the overall parameters of the elevator, basic parameters in the elevator operation database are called, including the elevator rated load, the elevator rated speed, the elevator guide rail spacing, the elevator balance coefficient, the elevator traction ratio, etc., and combined with the independent performance data of the components and the elevator interactive network data, a complete elevator system parameter set is established. Then, the elevator frame simulation is performed on the overall parameters of the elevator, and the stress conditions of the elevator frame are simulated and calculated using the finite element analysis (FEA) method. First, the elevator frame is meshed, and a non-uniform mesh encryption strategy is adopted during the meshing process. The mesh is encrypted in stress concentration areas (such as guide rail connection points and suspension points) to ensure calculation accuracy. Then, the frame movement performance loss is calculated based on the simulated elevator frame. The calculation process takes into account multiple factors such as friction loss, structural deformation loss, and energy conversion loss. The Coulomb friction model is used to calculate the friction loss between the guide rail and the car, the elastic-plastic deformation theory is used to calculate the strain energy loss of the frame structure, and the power balance method is used to calculate the energy loss of the traction system.

[0089] Step S5: Project component positions on the simulated elevator frame based on component independent performance, and evaluate component interaction performance based on component position projection, component independent performance, and elevator energy consumption distribution; construct an elevator component performance model based on component independent performance, component interaction performance, and frame movement performance loss.

[0090] In this embodiment, when projecting the positions of components on a simulated elevator frame based on the independent performance of the components, the simulated elevator frame is first subjected to coordinate transformation, and the positions of the components are mapped to a standardized three-dimensional coordinate system. The weighted centroid method is used to calculate the relative position weights of the components in the frame, and the component position projections are interpolated to ensure the continuity and smoothness of the data. Subsequently, the component interaction performance is evaluated based on the component position projections, the independent performance of the components, and the elevator energy consumption distribution. First, the non-parametric kernel density estimation (KDE, Kernel Density Estimation, kernel density estimation) method is used to calculate the interaction probability density of the components under different operating modes, and the Markov random field (MRF, Markov Random Field, Markov Random Field) method is used to model the synergistic influence relationship between the components. Finally, the component interaction performance matrix is obtained and stored in a database. Subsequently, an elevator component performance model is constructed based on the independent performance of the components, the component interaction performance, and the frame movement performance loss. First, a component feature space is constructed based on the independent performance of the components, and the principal component analysis (PCA, Principal Component Analysis) is used. Component Analysis (Principal Component Analysis) method is used for dimensionality reduction to extract the principal components of component performance. Subsequently, a component collaborative performance model is established based on the component interaction performance matrix. A deep neural network (DNN) is used to train the elevator component performance evaluation model, and the model parameters are corrected based on the framework movement performance loss. Finally, a complete elevator component performance model is output.

[0091] Preferably, step S1 includes the following steps:

[0092] Step S11: Collecting elevator usage records; extracting elevator usage flow in the elevator usage records;

[0093] Step S12: Divide the elevator usage flow into time windows to obtain the time-divided usage flow; infer the flow periodicity pattern based on the time-divided usage flow;

[0094] Step S13: Perform outlier detection on the traffic periodic pattern, set the outlier threshold to 2.5 times the standard deviation, remove outliers, and generate a purified traffic pattern; extract historical start and stop frequency data from the elevator usage record;

[0095] Step S14: performing statistical filtering on the historical start-stop frequencies and performing frequency domain conversion to obtain start-stop spectrum characteristics;

[0096] Step S15: Identify the peak of the start-stop spectrum characteristics and mark it as a key frequency feature point; perform flow spectrum collaborative analysis on the purification flow pattern and the key frequency feature point to obtain flow-frequency correlation data;

[0097] Step S16: extracting collaborative variable distribution data based on the flow-frequency correlation data, and performing working condition fitting based on the collaborative variable distribution data to obtain elevator usage pattern characteristics.

[0098] In this embodiment, when collecting elevator usage records, a data acquisition device is installed in the central unit of the elevator control system. The data acquisition device includes a high-precision time synchronization module, a state sensing module, a data storage module, and a data transmission module. The high-precision time synchronization module uses GPS (Global Positioning System) timing technology to synchronously record the time of elevator state changes with microsecond accuracy. The state sensing module includes a start-stop detection sensor, a load sensor, a door state sensor, and a running direction detection sensor. The sampling frequency of each sensor is set to 200Hz. The data storage module uses a ring buffer storage structure to temporarily store the data of the last 24 hours. The data transmission module uses an Ethernet connection to transmit data to the central storage unit every 5 minutes. The stored data includes elevator start time, elevator stop time, elevator running floors, elevator load weight, elevator door opening time, elevator door closing time, and elevator running direction. The data is stored in a database in a time series format. Then the elevator usage records were parsed to extract the elevator usage flow. The usage flow was defined as the number of starts and stops of the elevator per unit time. The sliding window method was used for calculation. The window length was set to 30 minutes and the sliding step was set to 5 minutes. The number of starts and stops in all time windows was calculated to generate the elevator usage flow data sequence. When the elevator usage flow was divided into time windows, the time window length was first set to 1 hour, and the elevator usage flow sequence was divided into multiple time windows in chronological order. The mean, variance and median of the flow data in each time window were calculated and stored in the database. Then the flow periodic pattern was inferred based on the time period usage flow. Fast Fourier transform (FFT) was used to analyze the flow periodicity. The fast Fourier transform (FT) is used to calculate the dominant frequency component of the traffic sequence. The period corresponding to the highest power spectral density is extracted, and the phase alignment of the traffic sequence within that period is calculated. The period with the highest phase alignment is taken as the dominant period of the elevator traffic periodic pattern. An autoregressive moving average (ARMA) model is then used to fit the periodic pattern and calculate traffic trends over future periods. For outlier detection within the periodic traffic pattern, a local outlier factor (LOF) method is used to calculate the local anomaly score of the traffic data within each time window, with an outlier threshold of 2.5 times the standard deviation, and remove all time windows with scores exceeding the threshold, generate a purified traffic pattern, and store it in the database. Then, extract the historical start and stop frequency data from the elevator usage record. The calculation method is to perform time series statistics on the elevator start and stop events, calculate the start and stop frequency every 1 minute, and store the start and stop frequencies in all time windows to generate a historical start and stop frequency data sequence. When performing statistical filtering on the historical start and stop frequencies, the median filter method is used to eliminate sudden outliers. The median filter window length is set to 5 minutes. The start and stop frequencies of all time windows are filtered, and then frequency domain conversion is performed. Short-time Fourier transform (STFT) is used. Transform (short-time Fourier transform) is used to calculate the time-frequency characteristics of the start-stop frequency in different time windows, the Hanning window function is used to improve the spectral resolution, and the power spectrum density is calculated to obtain the start-stop spectrum characteristics. When identifying the peak of the start-stop spectrum characteristics, the peak detection algorithm is used to calculate the first-order derivative of the power spectrum density curve, and all frequency points where the derivative is zero and the second-order derivative is negative are located. These frequency points are marked as key frequency feature points and stored in the database. Subsequently, the flow spectrum is collaboratively analyzed for the purification flow pattern and the key frequency feature points. The mutual information method is used to calculate the correlation between the purification flow pattern and the key frequency feature points, and the flow-frequency correlation data is calculated based on the correlation. When extracting the collaborative variable distribution data based on the flow-frequency correlation data, the kernel density estimation (KDE, Kernel Density Estimation, kernel density estimation) is used to calculate the probability density function of the flow-frequency correlation data, and the collaborative variable distribution data is generated based on the probability density function. Subsequently, the working condition fitting is performed based on the collaborative variable distribution data, and the Gaussian mixture model (GMM, Gaussian Mixture The Gaussian mixture model (GMM) is used to perform cluster analysis on the distribution data of collaborative variables, and the elevator usage probability under different working conditions is calculated to finally obtain the elevator usage pattern characteristics.

[0099] Preferably, step S2 includes the following steps:

[0100] Step S21: performing component load decomposition on the elevator usage pattern characteristics, and querying component response characteristics in the elevator usage pattern characteristics based on the component load data;

[0101] Step S22: identifying the component transfer function based on the component response characteristics and the component load data; performing pole-zero analysis on the component transfer function, and performing dynamic characteristic index mapping;

[0102] Step S23: determining the stability boundary based on the dynamic characteristic index, and dividing the component stability range according to the stability boundary;

[0103] Step S24: identifying key operating parameters of the dynamic characteristic index according to the component stability range to obtain the key operating parameters of the component; performing performance curve fitting on the key operating parameters of the component to generate independent performance of the component.

[0104] In this embodiment, when performing component load decomposition on the elevator usage pattern characteristics, the flow-frequency correlation data in the elevator usage pattern characteristics is first obtained, and the working status of the key components of the elevator is analyzed based on the elevator operation log. The key components of the elevator include the traction machine, elevator guide rails, elevator brakes and elevator balancing system. A load decomposition model based on operation state recognition is used to convert the elevator usage pattern characteristics into instantaneous load data of each component. The load decomposition model uses rigid body dynamics equations to calculate the force state of each component. Assuming that the mass of the elevator car is 1000kg, the rated load is 800kg, the traction machine drive power is 22kW, and the traction ratio is set The ratio of the load to the load ratio is set to 2:1. The traction machine output torque is calculated. The force state of the elevator guide rails is calculated using the contact stiffness model, the brake force is calculated using the brake friction model, and the tension distribution of the balance system is calculated using the wire rope tension distribution model. The calculated component load data is stored in a database. The component response characteristics of the elevator usage pattern characteristics are then queried based on the component load data. The dynamic coupling relationship between the load changes of each component and the elevator usage pattern characteristics is calculated using time series cross-correlation analysis. The response lag time of each component load change to the elevator start and stop frequency is calculated using sliding window cross-correlation. The calculation accuracy of the lag time is set to 0.1 second. When identifying the component transfer function based on the component response characteristics and component load data, the least squares method is used to estimate the transfer function parameters. The transfer function is represented by a second-order system model in the form of G(s) = (Kω_n^2) / (s^2+2ζω_ns+ω_n^2), where K is the gain coefficient, ω_n is the system natural frequency, and ζ is the damping ratio. The least squares method is used to calculate the optimal estimate of K, ω_n and ζ, and the calculated transfer function is stored in the database. Then, the component transfer function is subjected to pole-zero point analysis, and the Laplace transform is used to calculate the system pole position, the real part and imaginary part of the pole, and the distribution of the pole is analyzed to determine the stability of the system. The state space method is used to calculate the system zero position, analyze the distribution of the system zero, and calculate the degree of influence of the zero on the system response. Then, dynamic characteristic index mapping is performed to extract the component's natural frequency, damping ratio, Resonant frequency and frequency response characteristics. When determining the stability boundary based on dynamic characteristic indicators, the Nyquist criterion is used to calculate the system stability margin. The calculation method is to draw the Nyquist curve of the system and analyze the surrounding situation of the curve relative to the (-1,0) point, calculate the gain margin and phase margin, set the gain margin threshold to 6dB, and the phase margin threshold to 45 degrees, and judge the stability range of the system. The Lyapunov method is used to calculate the stability area of the system. The Lyapunov function V(x) is established, its derivative dV / dt is calculated, and the sign of dV / dt is determined. If dV / dt is less than zero, the system is in a stable state. If dV / dt is greater than zero, the system is in an unstable state. The component stability range is divided according to the stability judgment result. When identifying the key operating parameters of the dynamic characteristic indicators based on the component stability range, support vector regression (SVR) is used. The Support Vector Regression (SVR) method calculates the mapping relationship between dynamic characteristic indicators and stability ranges. The input features are dynamic characteristic indicators, including natural frequency, damping ratio, resonant frequency, and frequency response characteristics, and the output is the stability range. The RBF (Radial Basis Function) kernel function is used for regression calculations to optimize the SVR model parameters and store them in a database. Subsequently, a performance curve is fitted to the key operating parameters of the component. The least squares curve fitting method is used to calculate the relationship between the key operating parameters and the component performance. The order of the fitting curve is set to a third-order polynomial. The fitting parameters are calculated and stored in the database, ultimately generating the independent performance of the component.

[0105] Preferably, the step S3 of reconstructing the elevator interaction network according to the independent performance of the components and the elevator usage pattern characteristics includes:

[0106] Derive elevator operation data based on elevator usage pattern characteristics;

[0107] Split the elevator operation data to obtain the elevator component operation data;

[0108] Perform component dependency analysis on elevator component operation data to obtain inter-component dependency;

[0109] Locate the causal direction of the dependencies between components and identify the component dependencies based on the causal direction;

[0110] The elevator interaction network is reconstructed based on the independent performance and component dependency of components.

[0111] In this embodiment, elevator operation data is derived based on elevator usage pattern characteristics, and a time series data processing method is used to perform time series analysis on the elevator usage pattern characteristics. The time window size is set to 10 minutes, and a sliding window method is used to extract feature sequences. The car operation time distribution is calculated, and the car single operation time range is set to 5-60 seconds. The elevator start and stop frequency is calculated, and the start and stop frequency range is set to 0.1-2 times / minute. Fourier transform is used to calculate the operation cycle characteristics, and the main frequency range is set to 0.01-0.5Hz. The obtained elevator operation data is stored, and the elevator operation data is split to obtain elevator component operation data. A rule-based decomposition method is used, and the decomposition rule is set as a mapping relationship between operation data and component working status. The elevator start and stop signals are parsed, and the start and stop signals are mapped to the control system response. The control system response time range is set to 0.01-0.1 seconds. The car acceleration signal is parsed, and the maximum car acceleration is set to 1.5m / s. 2, analyze the braking system status, set the braking system working current range to 0.5-5A, analyze the component dependency of the elevator component operation data, obtain the dependency between components, use the dynamic correlation analysis method to calculate the temporal correlation between components, set the correlation threshold to 0.8, use the Granger causality analysis method to calculate the causal relationship between components, set the significance level to 0.05, calculate the causal effect of the traction motor power change on the car speed change, calculate the effect of the guide rail friction change on the car vibration, use the directed graph model to build the component dependency network, set the number of network nodes to 10-50, store the dependency between components, locate the causal direction of the dependency between components, and identify the component subordination based on the causal direction, use the structural equation modeling (SEM) method to calculate the causal path, set the path coefficient range to 0.1-0.9, and use the directional Bayesian network (DBN) The causal direction of the change in car mass on the change in traction motor load is inferred by using a directional Bayesian network (DBN) with a confidence level of 95%. The causal direction of the change in car mass on the change in traction motor load is calculated. The impact of the braking system response time on the car stopping accuracy is calculated. The component affiliation is generated. The elevator interaction network is reconstructed based on the independent performance of the components and the component affiliation. The elevator component interaction topology is constructed using a weighted directed graph method with a weight range of 0.1-1. The interaction strength between components is calculated with a threshold of 0.5. The clustering coefficient of the interaction network is calculated using a graph theory method with a range of 0.2-0.8. The network centrality is calculated with a range of 0.1-0.9, and the elevator interaction network is finally generated.

[0112] Of particular importance is that the reconstruction of the elevator interaction network based on the independent performance and dependency relationships of components includes:

[0113] Convert component dependencies into a dependent tree structure;

[0114] The subordinate tree structure and the independent performance of the components are fused and mapped to obtain a hybrid feature space;

[0115] Construct an interaction topology framework based on the hybrid feature space, and analyze the component interaction strength based on the interaction topology framework;

[0116] Directional weight adjustment of component interaction strength is performed based on component subordination to obtain asymmetric interaction strength;

[0117] Sort component links based on asymmetric interaction strength to generate a ranking of key interaction links;

[0118] Connect elevator components based on the ranking of key interaction links, and construct an elevator component interaction topology based on the connected elevator components;

[0119] Calculate the centrality of the elevator component interaction topology to obtain the node importance distribution;

[0120] Identify key elevator components based on node importance distribution;

[0121] The elevator interaction network is reconstructed based on key elevator components and the interaction topology of elevator components.

[0122] In this embodiment, the component subordination is converted into a subordination tree structure. A directed acyclic graph (DAG) is used to model the subordination of elevator components. The root node is set as the elevator frame, and the child nodes include guide rails, car, counterweight, traction machine, control cabinet, door system, etc. The component subordination is classified by a hierarchical clustering method, and the hierarchical depth is set to 3-5 layers. The connection relationship between components is stored by an adjacency matrix, and the matrix dimension is set to N×N, where N is the total number of components. The component subordination tree is traversed by a depth-first search (DFS) method, and the search depth is set to no more than 6 layers. Finally, a component subordination tree structure is generated. The subordination tree structure and the independent performance of the components are fused and mapped to obtain a hybrid feature space. The independent performance of the components is digitized by a feature embedding method, and the embedding dimension is set to 10-50 dimensions. A Gaussian kernel function is used. Function) is used to perform nonlinear mapping on component performance data, and the kernel function width is set to 0.1-1.0. Principal component analysis (PCA) is used to reduce the dimension of the fused feature data, and the principal component contribution rate threshold is set to 85%-95%. Cosine similarity is used to calculate the feature similarity between components, and the similarity threshold is set to 0.75-1.0. Finally, a hybrid feature space is generated, and an interactive topology framework is constructed based on the hybrid feature space. The component interaction strength is analyzed based on the interactive topology framework. A graph neural network (GNN) is used to model the component interaction relationship, and the number of network layers is set to 3-5 layers. The random walk method is used to sample the interaction network, and the walk step size is set to 5-10. The node embedding method is used to encode the component interaction features, and the embedding dimension is set to 16-64 dimensions. The link prediction method is used. The interaction strength of components is calculated using the prediction method, with the interaction strength range set between 0 and 1. This ultimately results in an interaction topology framework. Based on component affiliation, the interaction strength is directional-weighted to obtain asymmetric interaction strength. A directed graph is used to store component interaction data, with the initial edge weight set to 1.0. The importance of components in the interaction topology is calculated using the PageRank algorithm, with a damping coefficient set to 0.85, using normalized mutual information (NMI) to calculate component interaction weights, setting the weight range to 0.1-1.0, using information entropy to adjust the direction of interaction strength, setting the entropy value range to 0.01-0.1, and finally generating asymmetric interaction strength. Component links are sorted according to asymmetric interaction strength, generating key interaction link rankings, using a ranking algorithm to prioritize interaction links, setting the sorting criteria to interaction strength, component importance, structural stability, etc., using the Top-K method to screen key interaction links, setting the K value to 10-50, using the bidirectional sorting method to optimize interaction links, setting the optimization goal to maximum interaction stability, and using the analytic hierarchy process (AHP) to optimize interaction links. The weights of different ranking factors are calculated by the algorithm, and the weight range is set to 0.2-0.8. Finally, the key interaction link ranking is obtained. The elevator components are connected based on the key interaction link ranking, and the elevator component interaction topology is constructed according to the connected elevator components. The Dijkstra shortest path algorithm is used to calculate the optimal connection path between elevator components. The path weight is set to the inverse of the interaction strength. The adjacency list is used to store the connection information of the elevator components. The maximum capacity of the adjacency list is set to N×N, where N is the total number of components. The sparse matrix is used to store the interaction topology data. The sparsity threshold is set to 0.1-0.3. The minimum spanning tree (MST) method is used to optimize the elevator component connection structure. Finally, the elevator component interaction topology is constructed. The centrality of the elevator component interaction topology is calculated to obtain the node importance distribution. The degree centrality is used to calculate the connectivity of each component in the topology structure. The degree range is set to 1-20, and the betweenness centrality is used to calculate the connectivity of each component in the topology structure. Centrality is used to calculate the influence of each component in the interaction path, with the betweenness range set to 0.01-0.5. Eigenvector centrality is used to calculate the global importance of each component, with the eigenvector range set to 0.1-1.0. Closeness is used to calculate the closeness between components, with the calculation range set to 0.05-0.95, and finally obtain the node importance distribution. Based on the node importance distribution, key elevator components are identified, and the K-Means (K-means clustering) method is used to classify the components, with the number of categories set to 3-5. The LOF (local outlier factor) method is used to identify abnormal components, with the outlier factor threshold set to 1.5-2.5. The weighted scoring method is used to calculate the key component scores, with the scoring criteria set to centrality score, interaction strength score, connection stability score, etc. The threshold filtering method is used to select key components, with the screening threshold set to 0.7-1.0, and finally the key elevator components are identified. The elevator interaction network is reconstructed according to the key elevator components and the elevator component interaction topology. The local reconstruction method is used to optimize the elevator interaction topology, with the optimization goal set to minimum transmission loss. The global optimization method is used to calculate the optimal network structure, with the optimization parameters set to include node stability, interaction path redundancy, energy consumption, etc. The dynamic weighting method is used to select key components. The edge weighting method is used to adjust the edge weights in the elevator interaction network, with the dynamic adjustment range set to 0.1-1.0. Connectivity analysis is used to calculate the connectivity of the elevator interaction network, with the connectivity threshold set to 0.8-1.0. Finally, the elevator interaction network is reconstructed.

[0123] Preferably, the step S3 of estimating the elevator energy consumption distribution based on the elevator usage pattern characteristics and the elevator interaction network includes:

[0124] Identify the key connection skeletons of the elevator interaction network;

[0125] Identify redundant paths on the associated connection skeleton to obtain a set of backup paths;

[0126] Construct the elevator interaction network topology based on the key connection skeleton and the collection of backup paths;

[0127] Elevator energy consumption is mapped based on the elevator usage pattern characteristics, and the elevator interaction network topology is connected and energy mapped based on the elevator energy consumption. The energy data is connected to identify the elevator energy consumption distribution.

[0128] In this embodiment, the key connection skeleton of the elevator interaction network is identified, and a principal component analysis (PCA) is performed on the elevator interaction network using a topological dimension reduction method. The threshold of the principal component variance contribution rate is set to 90%, the connection relationship of key components is screened, and a minimum spanning tree (MST) method is used to construct a skeleton connection structure. The weight calculation method is set to the energy flow intensity between elevator components, and the energy flow intensity calculation method is set to the power transmission ratio. The threshold range is set to 0.05-0.5 to obtain the key connection skeleton. Redundant paths are identified for the associated connection skeletons to obtain a set of backup paths. The elevator interaction network is traversed using a breadth-first search (BFS) method. The path search depth is set to 3-5 layers. The maximum flow minimum cut (Max-Flow Minimum Cut) method is used. The path redundancy is calculated using the Min-Cut method, the maximum flow calculation method is set as the component energy transmission rate, the energy transmission rate calculation method is the energy consumption ratio, and the energy consumption ratio range is set to 0.1-0.9. All paths that meet the redundancy conditions are identified, and the set of backup paths is stored. The elevator interaction network topology is constructed based on the key connection skeleton and the set of backup paths. The weighted directed graph method is used for topology modeling, the number of nodes is set to 10-50, the edge weight calculation method is set to energy transmission efficiency, the energy transmission efficiency calculation method is set to input-output power ratio, and the power ratio range is set to 0.6-0.95. The network connectivity is calculated using the shortest path optimization method, the shortest path calculation method is set to the Dijkstra algorithm based on energy loss, and the energy loss range is set to 1-10%. Finally, the elevator interaction network topology is generated, the elevator usage pattern characteristics are mapped to elevator energy consumption, and the elevator interaction network topology is connected based on elevator energy consumption. Time series regression analysis is used. Methods: Calculate the energy consumption trend of the elevator, set the regression window size to 30 minutes, calculate the energy consumption distribution of the traction machine, set the energy consumption range of the traction machine to 1-10kW, calculate the power consumption of the control system, set the power range of the control system to 0.1-2kW, use the power flow analysis method to calculate the energy transmission relationship of each component, set the power transmission efficiency range to 0.7-0.98, generate connection energy data, identify the energy consumption distribution of the elevator based on the connection energy data, and use the power density calculation method to calculate the energy consumption density of the elevator components, set the power density calculation method to the power consumption ratio per unit volume, and set the power consumption ratio range to 0.5-5kW / m 3,The hierarchical clustering method is used to classify the energy consumption of elevator components, and the number of classification layers is set to 3-5 layers. The energy consumption proportion of each layer is calculated, and the energy consumption proportion calculation method is set as the ratio of power consumption of each layer to the total power ratio, and finally the elevator energy consumption distribution data is generated.

[0129] Preferably, the elevator framework simulation of the overall elevator parameters in step S4 includes:

[0130] Expand the characteristic dimensions of the overall elevator parameters to obtain a hierarchical set of parameters;

[0131] Perform constraint coupling integration on parameter hierarchical sets to generate parameter correlation data;

[0132] Perform frame mechanical load projection on the parameter association data to obtain the frame load distribution, where the load types include static load, which ranges from 1.0 to 1.5 times the rated load, dynamic load, which ranges from 0.2 to 0.5 times the rated load, and impact load, which ranges from 0.1 to 0.3 times the rated load;

[0133] Map the component rigidity requirements according to the frame load distribution to obtain the rigidity requirements of the frame components;

[0134] Identify the main load-bearing component parameters of the overall elevator parameters;

[0135] Determine the geometric characteristics of the load-bearing components based on the parameters of the main load-bearing components, and perform stress distribution analysis based on the geometric characteristics of the load-bearing components to obtain load-bearing stress distribution data;

[0136] Perform three-dimensional simulation on the parameters of the main load-bearing components, and perform stress deformation simulation on the simulated load-bearing components based on the load-bearing stress distribution data to generate simulated deformation load-bearing components;

[0137] Identify the outer end contact surface component parameters of the overall elevator parameters;

[0138] determining an outer end contact area of the outer end contact surface component parameters, and selecting protruding component parameters of the outer end contact surface component parameters based on the outer end contact area;

[0139] Elevator component morphology simulation is performed based on the simulated deformation load-bearing components and protruding component parameters to obtain the reconstructed elevator component morphology;

[0140] According to the rigidity requirements of the frame components, the reconstructed elevator component morphology is topologically matched to obtain matching elevator components;

[0141] Elevator frame simulation is performed based on matching elevator components, where the number of nodes in the simulation model is 1000-10000 and the simulation time step is 0.01-0.1 second.

[0142] In this embodiment, the overall parameters of the elevator are expanded in feature dimensions, and a multi-level feature classification method is used to decompose the overall parameters of the elevator into four dimensions: physical properties, electrical characteristics, structural characteristics, and environmental adaptability. The physical properties include material density, elastic modulus, tensile strength, and yield limit; the electrical characteristics include motor power, current, voltage, and control signal type; the structural characteristics include frame size, welding method, connection method; and the environmental adaptability includes temperature tolerance and humidity tolerance. During data processing, principal component analysis (PCA) is performed on the parameters of each dimension to extract a highly correlated parameter set. The importance weights of the parameters are calculated by singular value decomposition (SVD). Parameters with weights greater than 0.8 are used as core features, and a parameter hierarchical set is constructed. The data structure of the hierarchical set is stored in the form of a multidimensional array. Each parameter is stored in the corresponding array layer according to physical, mechanical, electrical, or environmental properties, and fast access is achieved through data indexing. All data are stored in double-precision floating-point numbers to ensure calculation accuracy. During the calculation process, the data normalization range is set so that all parameter values fall into the interval [0,1]. The normalization calculation formula is used. Linear normalization, that is, the parameter value minus the minimum value is divided by the difference between the maximum and minimum values, and the parameter layer set is constrained and coupled to integrate, thereby generating parameter-related data. When performing constraint coupling calculations, the finite constraint optimization method is used to establish a set of constraint equations for the data in each parameter layer set. For example, for the thickness of the frame main beam, the constraints include the material yield limit, the maximum working stress, and the safety factor. The calculation formula is that the main beam thickness is equal to the total load borne by the frame divided by the product of the main beam width and the yield limit, and multiplied by the safety factor. During the constraint calculation process, the Lagrange multiplier method is used to solve the constraints The optimization problem is solved. The parameter constraint matrix is stored in a sparse matrix to reduce computational complexity. QR decomposition is used to improve computational efficiency during the solution process. A correlation matrix is established for all parameters. Each element in the matrix represents the influence weight between two parameters. The weight is calculated by data fitting. The least squares method is used for data fitting. The number of fitting data points is not less than 1000 to ensure computational accuracy. Variance analysis is performed on the fitting error, and parameters with variance exceeding the threshold are eliminated. When the parameter correlation data is projected onto the frame mechanical load, the finite element analysis (FEA) method is used to simulate the load of the elevator frame. First, a finite element model of the elevator frame is established. The model element type is a four-node tetrahedral element (Tetrahedral 4-node), and the mesh size is set to 10 mm. The material parameters of each component are set according to the Young's modulus of 210 GPa, Poisson's ratio of 0.3, and yield strength of 355 MPa for steel structures. Static load, dynamic load, and impact load are applied to the finite element model. The static load range is set to 1.0-1 of the rated load.5 times, applied to the contact surface at the bottom of the car, using a uniformly distributed load mode, the dynamic load range is set to 0.2-0.5 times the rated load, applied to the connection point between the traction wire rope and the car through time history analysis, the impact load range is set to 0.1-0.3 times the rated load, applied to the buffer contact point, using transient analysis method to simulate the impact response, using ANSYS Mechanical simulation calculations are performed, the simulation time step is set to 0.01s, and the total simulation time is set to 10s. During the calculation process, the node stress, displacement, and acceleration data of each time step are recorded and stored in CSV format. Finally, the frame load distribution data is output. When mapping the rigidity requirements of components according to the frame load distribution, the high stress area of the elevator frame is first extracted, and the stress threshold is set to 80% of the material yield strength, that is, the area with stress exceeding 284MPa is regarded as a part with higher rigidity requirements. The interpolation method is used to calculate the rigidity requirement gradient and determine the rigidity requirement level of each component. For areas with high rigidity requirements, the rigidity is improved by thickening the plate or adding reinforcement ribs. For areas with low rigidity requirements, the material usage is optimized to reduce weight and reduce manufacturing costs, and the main parameters of the overall elevator are identified. When determining the parameters of load-bearing components, the primary load-bearing components, including the guide rails, car chassis, counterweight frame, and traction machine base, were first determined based on the frame load distribution. Stress analysis was then used to calculate the load contribution of each component, with a threshold set at 5% of the total load. Components with a load contribution exceeding 5% were considered primary load-bearing components. MATLAB was used for calculations, and the load data for all components was entered into a matrix. Principal component analysis (PCA) was then used to identify the primary load-bearing components. To determine the geometric characteristics of the load-bearing components based on their parameters, the dimensional data for each component was extracted and a 3D CAD model was constructed. Abaqus was used for stress distribution analysis, with a mesh size of 5 mm. Actual operating loads were applied, the stress distribution of each component was calculated, and the stress distribution data was output. Abaqus Explicit was used to perform a 3D simulation of the primary load-bearing components, applying actual operating loads and calculating deformation based on the stress distribution data. The simulation time step was set to 0.005s, iterate the calculation for 1000 steps, and finally output the deformation of the simulated load-bearing component, extract the outer end contact surface data of the overall parameters of the elevator, including the bottom of the car, the door sill, the buffer contact surface, etc., extract all components in contact with the external environment based on the structural model of the elevator, and use the computer vision method (Computer Vision) to automatically identify the contact area, segment the contact area through the image segmentation algorithm, and determine the outer end contact surface component parameters in combination with geometric analysis, determine the outer end contact area of the outer end contact surface component parameters, and screen the protruding component parameters of the outer end contact surface component parameters based on the outer end contact area. First, use a 3D scanner to obtain the surface data of the outer end contact surface component, and use point cloud processing software to reconstruct the surface, calculate the contact area, and set a threshold based on the contact area to screen out the protruding component parameters of the outer end contact surface component. When performing elevator component morphology simulation based on the simulated deformation load-bearing components and protruding component parameters, use CATIA modeling software to construct the elevator components. A morphological model was created and simulated using Abaqus software. Material properties, load conditions, and contact characteristics were defined, ultimately resulting in a reconstructed elevator component morphology. When performing component topology matching based on the rigidity requirements of the frame components, a topology matching method based on constraint optimization was used to optimize the component morphology data. Matching calculations were performed based on the structural rigidity requirements, ultimately resulting in a matched elevator component. When simulating the elevator frame based on the matched elevator components, a finite element simulation model containing 1,000 to 10,000 nodes was established, with a simulation time step set to 0.01 to 0.1 seconds. A dynamics solver was used for time-domain analysis to ultimately obtain the simulation results.

[0143] Preferably, the calculation of the frame movement performance loss based on the simulated elevator frame in step S4 includes:

[0144] Expand the structural freedom of the simulated elevator frame to obtain the frame freedom data;

[0145] Calibrate the elevator frame motion mode based on the frame degree of freedom data;

[0146] Perform force mapping transformation on the simulated elevator frame according to the elevator frame motion mode to generate a frame motion benchmark;

[0147] Extract the contact surface energy distribution in the frame motion benchmark and infer the frame contact area based on the contact surface energy distribution;

[0148] Perform dynamic fitting of friction force on the frame contact area to generate frame friction force data;

[0149] Perform power consumption reduction on the frame friction matrix to obtain the frame friction response data;

[0150] The frame friction response data is used to distribute the supporting component moment to obtain the frame support force data;

[0151] Mapping frame resistance characteristics based on frame support force data;

[0152] The performance attenuation is calculated according to the frame resistance characteristics, and the attenuation effect is accumulated based on the performance attenuation data to generate the frame movement performance loss.

[0153] In this embodiment, the structural freedom degree of the simulated elevator frame is expanded to obtain the frame freedom degree data. The elevator frame is discretized using the finite element analysis (FEA) method. The total number of nodes of the frame is set to a range of 1000-10000. The freedom degree of each node is defined, including three-dimensional displacement freedom and three-dimensional rotation freedom. Constraints are applied to the fixed end, hinge point and sliding support position of the frame. The Lagrange Multiplier method is used to calculate the freedom degree of the frame. Method) calculates the constraint reaction force, and the constraint equation is expressed as: [K]{U}={F}, where [K] is the frame stiffness matrix, {U} is the node displacement vector, and {F} is the external load vector. The calculation convergence error is set to within 0.01 mm, and the frame degree of freedom data is obtained. The elevator frame motion mode is calibrated according to the frame degree of freedom data, and the modal analysis method is used to calculate the natural vibration characteristics of the frame. The natural frequency range is set to 0.5-5 Hz, and the vibration shape of each mode of the frame is calculated. The modal response is Fourier transformed (Fourier Transform), and the frequency domain resolution is set to 0.1 Hz. The main frequency component and modal morphological characteristics of the motion mode are calculated to obtain the elevator frame motion mode data. According to the elevator frame motion mode, the simulated elevator frame is subjected to force mapping transformation to generate a frame motion benchmark. The Newton-Euler equation (Newton-Euler Equation) is used to analyze the stress state of the frame under the action of traction force, braking force and inertia force. The traction force range is set to 1.2-1.8 times the rated load, and the braking force range is set to 0.8-1.5 times the rated load. The force equilibrium state of the frame under different load conditions is calculated, and a force mapping matrix is established. The force distribution data is corresponded to the frame structure to form a frame motion benchmark. The contact surface energy distribution in the frame motion benchmark is extracted, and the frame contact area is inferred based on the contact surface energy distribution. The contact mechanics analysis method is used to calculate the energy transfer between the frame and the guide rail, car bottom and other contact parts. The contact energy range is set to 10-500 joules. The finite element meshing method is used to refine the contact area, and the mesh size range is set to 1-5 mm. The energy density per unit area is calculated, and the contact area boundary is identified based on the energy density gradient. The friction force is dynamically fitted in the frame contact area to generate the frame friction force data. The Coulomb friction model is used to calculate the friction force. The friction force expression is: F f =μN, where μ is the friction coefficient, F fThe friction data is set in the range of 0.05-0.3, N is the normal force, and the Newton-Raphson method is used to calculate the dynamic curve of friction changing with time. The time step range is set to 0.01-0.1 seconds to generate the frame friction data. The power consumption of the frame friction matrix is calculated to obtain the frame friction response data. The energy conservation principle is used to calculate the friction power consumption. The friction power consumption expression is: P f =F f v, where v is the relative sliding velocity of the contact surface, P f The friction power consumption data is set to 1-50 watts, and the total power consumption of the friction matrix is calculated by matrix decomposition method. The calculation error range is set to within 0.01 watts. The frame friction response data is obtained, and the support component torque is distributed to the frame friction response data to obtain the frame support force data. The static equilibrium equation is used to calculate the force state of the support component. The support force range is set to 100-5000 Newtons. The matrix operation method is used to calculate the torque distribution. The torque calculation error is set to within 0.01 Newton meters. The frame support force data is generated. The frame resistance characteristics are mapped based on the frame support force data, and the fluid dynamics method is used to calculate the elevator operation process. The air resistance in the process is set to 1-10 Newtons. The friction resistance between the frame and the guide rail is calculated using the material tribology analysis method, and the friction resistance range is set to 10-500 Newtons. The air resistance, friction resistance and other mechanical resistances are accumulated to form the frame resistance characteristics. The performance attenuation is calculated according to the frame resistance characteristics, and the attenuation effect is accumulated based on the performance attenuation data to generate the frame movement performance loss. The time series analysis method is used to calculate the frame performance attenuation trend, and the time step range is set to 1-10 hours. The integration method is used to calculate the accumulation of attenuation effects, and the accumulation error range is set to 0.01%. Finally, the frame movement performance loss is generated.

[0154] Preferably, step S5 includes the following steps:

[0155] Step S51: splitting the elevator component units of the simulated elevator frame; performing component unit matching on the independent performance of the components based on the elevator component units to obtain a matching relationship;

[0156] Step S52: projecting the independent performance of the component onto the simulated elevator frame based on the matching relationship to obtain the component position projection;

[0157] Step S53: Calculate component energy consumption based on component position projection and elevator energy consumption distribution, and derive component network performance based on component energy consumption data and overall elevator parameters; evaluate component interaction performance based on component network performance and component independent performance;

[0158] Step S54: performing functional block splitting on the independent performance data of the component to obtain component functional blocks; performing energy efficiency response fitting on the component functional blocks to obtain component energy efficiency distribution;

[0159] Step S55: performing global correlation aggregation on the component energy efficiency distribution data to obtain component performance data; performing output performance integration based on the component performance data and component interaction performance to obtain integrated output performance;

[0160] Step S56: Constructing an elevator component performance model based on the integrated output performance and the frame movement performance loss.

[0161] In this embodiment, the elevator component units of the simulated elevator frame are decomposed, and the elevator frame is structurally analyzed using a topological decomposition method. The decomposition threshold is set to a component size greater than 50 mm or a mass greater than 1 kg. The connectivity matrix is used to determine the structural coupling relationship between components. The connection strength of each component is calculated, and the connection strength threshold is set to 0.1-1.0 N·m. The elevator components are decomposed into independent units based on the connection strength values to obtain component unit data. Component unit matching is performed on the independent performance of the components based on the elevator component units. A feature matching algorithm is used to calculate the component independent performance vector, and the vector dimension is set to 10-50 dimensions. Performance indicators such as mass, stiffness, damping, and moment of inertia of the elevator components are standardized. The component matching degree is calculated using a K-nearest neighbor (KNN) classification method, and the K value is set to 3-7. The matching degree threshold is set to 80%-100%. A matching relationship is generated. Based on the matching relationship, the independent performance of the component is projected onto the simulated elevator frame to obtain the component position projection. The geometric transformation matrix is used to calculate the component position projection. Matrix) calculates the spatial transformation relationship of the component relative to the frame, sets the rotation angle range of the transformation matrix to 0-180 degrees, uses the homogeneous coordinate transformation method to calculate the spatial mapping coordinates of the component in the simulated elevator frame, sets the coordinate accuracy to 0.1 mm, obtains the component position projection data, calculates the component energy consumption based on the component position projection and the elevator energy consumption distribution, and derives the component network performance based on the component energy consumption data and the overall parameters of the elevator. The energy balance equation is used to calculate the component energy consumption. The energy balance equation is expressed as: E=P·t, where E is the component energy consumption in joules (J), P is the power in watts (W), and t is the running time in seconds (s). The power measurement range is set to 1-500 watts. The Fourier series expansion is used to calculate the time series characteristics of the overall energy consumption parameters of the elevator. The series expansion order is set to 5-20 orders. The component energy consumption data and the overall parameters of the elevator are combined to derive the component network performance. The graph convolutional network (GCN) is used to The energy transfer path between components is calculated using a network with 2-5 layers and 50-500 nodes. The component network performance data is obtained. The component interaction performance is calculated based on the component network performance and the component independent performance. The energy interaction process between components is simulated using a Markov process, and the state transition probability threshold is set to 0.1-0.9. Obtain component interaction performance data, perform functional block decomposition on component independent performance data to obtain component functional blocks, use functional module decomposition to calculate the functional independence of components in the system, set the functional independence score range to 0-100 points, and set the score threshold to 50 points. For components with scores below the threshold, perform functional block decomposition, use hierarchical clustering to calculate the functional similarity between components, set the similarity threshold to 80%-100%, obtain component functional block data, perform energy efficiency response fitting on component functional blocks, obtain component energy efficiency distribution, and use multiple regression analysis to calculate the functional similarity between components. The relationship between the energy consumption of the component and variables such as load and speed was calculated by regression coefficients ranging from 0.1 to 10.0. The regression equation was expressed as: E = a·L+b·v+c, where E is the energy consumption of the component in joules (J), L is the load in Newtons (N), v is the speed in meters per second (m / s), a, b, and c are regression coefficients. The regression coefficients were solved using the least squares method, and the calculation error range was set to within 0.1%. The component energy efficiency distribution data was obtained, and the component energy efficiency distribution data was globally correlated and aggregated to obtain component performance data. Principal component analysis (PCA) was used. The principal components of component energy efficiency characteristics were calculated using the K-Means clustering algorithm. The principal component contribution threshold was set at 85%-95%. The top 3-5 principal components were extracted for global aggregation. Component performance categories were calculated using the K-Means clustering algorithm, with the number of clusters set between 3 and 7. Component performance data was then obtained. Output performance was integrated based on component performance data and component interaction performance to obtain integrated output performance. A weighted averaging method was used to calculate the overall component performance score, with weights set between 0.1 and 0.9. The Delphi method was used to calculate the weight coefficients, with the expert panel size set between 10 and 30. Weighted scores for different components were calculated and normalized between 0 and 1 to obtain integrated output performance data. An elevator component performance model was constructed based on the integrated output performance and frame mobility performance loss. The performance prediction model was trained using the Support Vector Regression (SVR) method, with the Radial Basis Function (RBF) kernel function and the penalty parameter C set within the range of 0.1-100, the model error is set to within 1%, and the cross-validation method is used to calculate the model generalization ability. The cross-validation fold is set to 5-10 folds, and the elevator component performance model is finally obtained.

[0162] Of particular importance is the calculation of component energy consumption based on component position projection and elevator energy consumption distribution, the joint derivation of component network performance based on component energy consumption data and overall elevator parameters, and the evaluation of component interaction performance based on component network performance and component independent performance, including:

[0163] Perform spatial association on component position projection data to obtain component topological connection data;

[0164] Track energy transfer based on component topology connection data and map elevator energy consumption distribution data based on energy transfer paths;

[0165] Analyze component energy consumption characteristics based on elevator energy consumption distribution data;

[0166] Reconstruct network associations based on component energy consumption characteristics and overall elevator parameters, and perform system performance collaborative mapping to obtain component network performance;

[0167] Decouple the independent performance data of components to obtain component feature vectors;

[0168] Perform topological penetration analysis on component network performance to obtain the network flow structure;

[0169] The component interaction potential is analyzed based on the network flow structure and component eigenvectors, and the component interaction performance is mapped based on the interaction potential.

[0170] In this embodiment, spatial association is performed on the component position projection data to obtain component topological connection data. The spatial position data of the elevator components are modeled using a three-dimensional coordinate mapping method. The coordinate accuracy is set to 0.1 mm. The fixed components of the elevator frame (such as guide rails and shaft walls) and the moving components (such as the car and counterweight) are calibrated. The Delaunay triangulation algorithm (Delaunay The shortest connection path between components is calculated by using KNN (K-NearestNeighbors) method, and the path length threshold is set to 10-1000 mm. The spatial connection relationship of adjacent components is topologically analyzed. The adjacency relationship is set by using K-nearestneighbors method, and the K value range is set to 3-7. The connection data is normalized and the normalization range is set to 0-1. Finally, the topological connection data of the elevator components is obtained. The energy transfer of the component topological connection data is tracked, and the elevator energy consumption distribution data is mapped based on the energy transfer path. The energy conservation principle is used to calculate the energy transfer path between elevator components. The power loss threshold per unit time is set to 1-100 watts (W). The energy transfer rate is calculated by using the finite difference method (FDM), and the time step is set to 1-10 milliseconds. The energy transfer chain is analyzed by using the Markov process, and the dimension of the state transfer matrix is set to 5-20. The energy flow under different working conditions is simulated and calculated by using Monte Carlo simulation. Simulation) method was used to set the number of simulations to 1000-10000 times, and regression analysis was performed on the simulated data. The goodness of fit R was set. 2The threshold is 0.9-1.0, and the energy consumption distribution data of the elevator is finally obtained. The energy consumption characteristics of the components are analyzed based on the energy consumption distribution data of the elevator. The frequency domain characteristics of the component power consumption signal are calculated using the Fourier Transform method, and the transformation order is set to 10-50. The energy distribution of the components is calculated using power spectrum density analysis (PSD), and the frequency resolution is set to 0.1-5 Hz. The distribution characteristics of the component energy consumption data are calculated using mutual information entropy, and the entropy value range is set to 0.1-1.0. Principal component analysis (PCA) is performed on the energy consumption characteristic data, and the threshold of the principal component contribution rate is set to 85%-95%. Finally, the energy consumption characteristics of the elevator components are obtained. The network association is reconstructed according to the component energy consumption characteristics and the overall parameters of the elevator, and the system performance collaborative mapping is performed to obtain the component network performance. The weighted undirected graph (WUG) is used to calculate the energy distribution of the components. Graph) to construct component energy consumption network, set the number of graph nodes to 50-500, the edge weight range to 0.1-1.0, use the shortest path algorithm (Dijkstra Algorithm) to calculate the energy transmission path, set the path cost function to power loss, use hierarchical clustering to classify component energy consumption data, set the number of clustering layers to 2-4 layers, use dynamic time warping (DTW) to calculate the similarity of energy consumption data, set the similarity threshold to 0.8-1.0, perform system performance collaborative mapping on the associated reconstructed energy consumption network, use deep neural network (DNN) to calculate system performance collaborative parameters, set the number of network layers to 3-7 layers, and the number of hidden layer neurons to 100-500, and finally obtain the network performance of elevator components, perform feature decoupling on the independent performance data of components to obtain component feature vectors, and use convolutional neural network (CNN) to perform feature decoupling on the independent performance data of components to obtain component feature vectors. The spatial features of components were extracted using a convolutional neural network (CNN) with a convolution kernel size of 3×3. A long short-term memory (LSTM) network was used to extract the temporal features of components. The number of LSTM units was set to 50-200. A feature selection algorithm was used to calculate the optimal feature subset of the independent performance data of the components. The feature screening threshold was set to 0.05-0.5. Finally, the characteristic vectors of the elevator components are obtained. The network performance of the components is topologically analyzed to obtain the network flow structure. The flow characteristics of the elevator energy flow are calculated using the fluid dynamics (CFD) method. The grid size is set to 0.1-1 mm. The stress distribution of the network topology is calculated using finite element analysis (FEA). The stress threshold is set to 10-100 MPa. The network flow importance of the components is calculated using the PageRank algorithm. The damping factor is set to 0.85. The flow structure data is analyzed in time series. The autoregressive moving average model (ARMA) is used to set the model order to 1-5. Finally, the network flow structure of the elevator components is obtained. The interaction potential energy of the components is analyzed based on the network flow structure and the characteristic vectors of the components. The interaction performance of the components is mapped based on the interaction potential energy. The interaction potential energy of the components is calculated using Hamiltonian mechanics. The total energy conservation condition of the system is set. Lagrangian mechanics is used to calculate the interaction potential energy of the components. The relative motion relationships between components are calculated using a generalized coordinate system (D-Mechanics) algorithm. The generalized coordinate dimensions are set to 3-6. The particle swarm optimization (PSO) algorithm is used to calculate the optimal interaction parameters for the components. The particle swarm size is set to 50-200. Feature extraction is performed on the interaction potential data. Principal component analysis (PCA) is used to select the optimal interaction pattern. The principal component contribution threshold is set to 85%-95%. The interaction performance of the elevator components is finally determined.

[0171] The present invention further provides a system for constructing an elevator component performance model, which is used to execute the method for constructing an elevator component performance model as described above. The system for constructing an elevator component performance model includes:

[0172] The spectrum modeling module is used to collect elevator usage records; perform flow spectrum modeling on the elevator usage records to generate start-stop spectrum characteristics; and perform flow-frequency co-fitting to obtain elevator usage pattern characteristics;

[0173] Dynamic analysis module, used to analyze component responses based on elevator usage pattern characteristics to obtain dynamic characteristic indicators; and to infer component independent performance based on dynamic characteristic indicators;

[0174] The interactive network module is used to reconstruct the elevator interactive network based on the independent performance of components and the characteristics of elevator usage patterns; and to estimate the elevator energy consumption distribution based on the elevator usage pattern characteristics and the elevator interactive network;

[0175] The frame simulation module is used to obtain the overall parameters of the elevator; simulate the elevator frame based on the overall parameters of the elevator; and calculate the frame movement performance loss based on the simulated elevator frame;

[0176] The model building module is used to project the component positions of the simulated elevator frame based on the independent performance of the components, and evaluate the component interaction performance based on the component position projection, component independent performance and elevator energy consumption distribution; and to build an elevator component performance model based on the component independent performance, component interaction performance and frame movement performance loss.

[0177] This invention achieves in-depth exploration of elevator operating characteristics through the collection and analysis of elevator usage records. The generated start-stop spectrum characteristics provide data support for understanding elevator usage patterns. The results of flow-frequency synergistic fitting reveal the operating laws of elevators and help optimize elevator operation strategies. The component response analysis capability of the dynamic analysis module improves understanding of the working status of each elevator component. The calculation of dynamic characteristic indicators makes component performance evaluation more scientific and systematic. The construction of the interactive network module provides a modeling basis for the interaction between various components of the elevator system. The effectively calculated elevator energy consumption distribution supports energy-saving transformation and optimization design. The frame simulation module evaluates the operating performance of the elevator frame through the acquisition and simulation analysis of overall parameters. The accurately calculated frame movement performance loss provides a basis for elevator operation and maintenance. The model construction module constructs an elevator component performance model by projecting component positions and evaluating interaction performance, covering the mutual influence and comprehensive performance of each component. This provides comprehensive data support and decision-making basis for the intelligent management and optimization of the elevator system, overall improving the efficiency and stability of elevator operation and promoting the continuous improvement and development of the elevator system.

[0178] The present invention further provides a computer-readable storage medium storing a computer program, wherein when the computer program is executed, the method for constructing an elevator component performance model as described in any one of the above items is implemented.

[0179] The present invention realizes the efficient execution of the elevator component performance model construction method by storing computer programs. The executable nature of the program ensures the automation and standardization of the model construction process, improves the operation efficiency and accuracy, and the stored program can flexibly adapt to the characteristics of different elevator systems to realize personalized parameter setting and analysis. With the help of the powerful computing power of the computer system, large-scale data processing and analysis can be carried out, thereby improving the reliability and timeliness of elevator usage data. The modular design of the program makes the connection between various functional modules smoother, facilitating subsequent maintenance and upgrading. The implementation in computer-readable form also supports the compatibility of multiple platforms and devices, improves the application scope and convenience, effectively meets the needs of intelligent elevator management and optimization, and provides a theoretical and practical basis for the efficient operation and sustainable development of the elevator system.

Claims

1. A method for constructing an elevator component performance model, characterized in that: The following steps are involved: Step S1: Collect elevator usage records; perform traffic spectrum modeling on the elevator usage records to generate start-stop spectrum features; Perform flow-frequency co-fitting to obtain the elevator usage pattern characteristics; Step S2: Analyze component responses based on elevator usage pattern characteristics to obtain dynamic characteristic indices; and infer component independent performance based on the dynamic characteristic indices. Step S3: Reconstructing the elevator interaction network based on the independent performance of the components and the elevator usage pattern characteristics; estimating the elevator energy consumption distribution based on the elevator usage pattern characteristics and the elevator interaction network; Step S4: obtaining overall elevator parameters; performing elevator frame simulation on the overall elevator parameters; and calculating frame movement performance loss based on the simulated elevator frame; Step S5: Project component positions on the simulated elevator frame based on component independent performance, and evaluate component interaction performance based on component position projection, component independent performance, and elevator energy consumption distribution; construct an elevator component performance model based on component independent performance, component interaction performance, and frame movement performance loss.

2. The method for constructing an elevator component performance model according to claim 1, wherein: Step S1 includes the following steps: Step S11: Collecting elevator usage records; extracting elevator usage flow in the elevator usage records; Step S12: Divide the elevator usage flow into time windows to obtain the time-divided usage flow; infer the flow periodicity pattern based on the time-divided usage flow; Step S13: Perform outlier detection on the traffic periodic pattern, set the outlier threshold to 2.5 times the standard deviation, remove outliers, and generate a purified traffic pattern; extract historical start and stop frequency data from the elevator usage record; Step S14: performing statistical filtering on the historical start-stop frequencies and performing frequency domain conversion to obtain start-stop spectrum characteristics; Step S15: Identify the peak of the start-stop spectrum characteristics and mark it as a key frequency feature point; perform flow spectrum collaborative analysis on the purification flow pattern and the key frequency feature point to obtain flow-frequency correlation data; Step S16: extracting collaborative variable distribution data based on the flow-frequency correlation data, and performing operating condition fitting based on the collaborative variable distribution data to obtain elevator usage pattern characteristics.

3. The method for constructing an elevator component performance model according to claim 1, wherein: Step S2 includes the following steps: Step S21: performing component load decomposition on the elevator usage pattern characteristics, and querying component response characteristics in the elevator usage pattern characteristics based on the component load data; Step S22: identifying the component transfer function based on the component response characteristics and the component load data; performing pole-zero analysis on the component transfer function, and performing dynamic characteristic index mapping; Step S23: determining the stability boundary based on the dynamic characteristic index, and dividing the component stability range according to the stability boundary; Step S24: identifying key operating parameters of the dynamic characteristic index according to the component stability range to obtain the key operating parameters of the component; performing performance curve fitting on the key operating parameters of the component to generate independent performance of the component.

4. The method for constructing an elevator component performance model according to claim 1, wherein: The step S3 of reconstructing the elevator interaction network based on the independent performance of components and the elevator usage pattern characteristics includes: Derive elevator operation data based on elevator usage pattern characteristics; Split the elevator operation data to obtain the elevator component operation data; Perform component dependency analysis on elevator component operation data to obtain inter-component dependency; Locate the causal direction of the dependencies between components and identify the component dependencies based on the causal direction; The elevator interaction network is reconstructed based on the independent performance and component dependency of components.

5. The method for constructing an elevator component performance model according to claim 1, wherein: The calculation of elevator energy consumption distribution using elevator usage pattern characteristics and elevator interaction network in step S3 includes: Identify the key connection skeletons of the elevator interaction network; Identify redundant paths on the associated connection skeleton to obtain a set of backup paths; Construct the elevator interaction network topology based on the key connection skeleton and the collection of backup paths; Elevator energy consumption is mapped based on the elevator usage pattern characteristics, and the elevator interaction network topology is connected and energy mapped based on the elevator energy consumption. The energy data is connected to identify the elevator energy consumption distribution.

6. The method for constructing an elevator component performance model according to claim 1, characterized in that: The elevator framework simulation of the overall elevator parameters in step S4 includes: Expand the characteristic dimensions of the overall elevator parameters to obtain a hierarchical set of parameters; Perform constraint coupling integration on parameter hierarchical sets to generate parameter correlation data; Perform frame mechanical load projection on the parameter association data to obtain the frame load distribution, where the load types include static load, which ranges from 1.0 to 1.5 times the rated load, dynamic load, which ranges from 0.2 to 0.5 times the rated load, and impact load, which ranges from 0.1 to 0.3 times the rated load; Map the component rigidity requirements according to the frame load distribution to obtain the rigidity requirements of the frame components; Identify the main load-bearing component parameters of the overall elevator parameters; Determine the geometric characteristics of the load-bearing components based on the parameters of the main load-bearing components, and perform stress distribution analysis based on the geometric characteristics of the load-bearing components to obtain load-bearing stress distribution data; Perform three-dimensional simulation on the parameters of the main load-bearing components, and perform stress deformation simulation on the simulated load-bearing components based on the load-bearing stress distribution data to generate simulated deformation load-bearing components; Identify the outer end contact surface component parameters of the overall elevator parameters; determining an outer end contact area of the outer end contact surface component parameters, and selecting protruding component parameters of the outer end contact surface component parameters based on the outer end contact area; Elevator component morphology simulation is performed based on the simulated deformation load-bearing components and protruding component parameters to obtain the reconstructed elevator component morphology; According to the rigidity requirements of the frame components, the reconstructed elevator component morphology is topologically matched to obtain matching elevator components; Elevator frame simulation is performed based on matching elevator components, where the number of nodes in the simulation model is 1000-10000 and the simulation time step is 0.01-0.1 second.

7. The method for constructing an elevator component performance model according to claim 1, wherein: The calculation of the frame movement performance loss based on the simulated elevator frame in step S4 includes: Expand the structural freedom of the simulated elevator frame to obtain the frame freedom data; Calibrate the elevator frame motion mode based on the frame degree of freedom data; Perform force mapping transformation on the simulated elevator frame according to the elevator frame motion mode to generate a frame motion benchmark; Extract the contact surface energy distribution in the frame motion benchmark and infer the frame contact area based on the contact surface energy distribution; Perform dynamic fitting of friction force on the frame contact area to generate frame friction force data; Perform power consumption reduction on the frame friction matrix to obtain the frame friction response data; The frame friction response data is used to distribute the supporting component moment to obtain the frame support force data; Mapping frame resistance characteristics based on frame support force data; The performance attenuation is calculated according to the frame resistance characteristics, and the attenuation effect is accumulated based on the performance attenuation data to generate the frame movement performance loss.

8. The method for constructing an elevator component performance model according to claim 1, wherein: Step S5 includes the following steps: Step S51: splitting the elevator component units of the simulated elevator frame; performing component unit matching on the independent performance of the components based on the elevator component units to obtain a matching relationship; Step S52: projecting the independent performance of the component onto the simulated elevator frame based on the matching relationship to obtain the component position projection; Step S53: Calculate component energy consumption based on component position projection and elevator energy consumption distribution, and derive component network performance based on component energy consumption data and overall elevator parameters; evaluate component interaction performance based on component network performance and component independent performance; Step S54: performing functional block splitting on the independent performance data of the component to obtain component functional blocks; performing energy efficiency response fitting on the component functional blocks to obtain component energy efficiency distribution; Step S55: performing global correlation aggregation on the component energy efficiency distribution data to obtain component performance data; performing output performance integration based on the component performance data and component interaction performance to obtain integrated output performance; Step S56: Constructing an elevator component performance model based on the integrated output performance and the frame movement performance loss.

9. A system for constructing an elevator component performance model, characterized in that: For executing the method for constructing an elevator component performance model according to claim 1, the system for constructing an elevator component performance model comprises: The spectrum modeling module is used to collect elevator usage records; perform flow spectrum modeling on the elevator usage records to generate start-stop spectrum characteristics; and perform flow-frequency co-fitting to obtain elevator usage pattern characteristics; Dynamic analysis module, used to analyze component responses based on elevator usage pattern characteristics to obtain dynamic characteristic indicators; and to infer component independent performance based on dynamic characteristic indicators; The interactive network module is used to reconstruct the elevator interactive network based on the independent performance of components and the characteristics of elevator usage patterns; and to estimate the elevator energy consumption distribution based on the elevator usage pattern characteristics and the elevator interactive network; The frame simulation module is used to obtain the overall parameters of the elevator; simulate the elevator frame based on the overall parameters of the elevator; and calculate the frame movement performance loss based on the simulated elevator frame; The model building module is used to project the component positions of the simulated elevator frame based on the independent performance of the components, and evaluate the component interaction performance based on the component position projection, component independent performance and elevator energy consumption distribution; and to build an elevator component performance model based on the component independent performance, component interaction performance and frame movement performance loss.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed, the method for constructing an elevator component performance model according to any one of claims 1 to 8 is implemented.

Images